1 Xapian-core 1.3.4 (2016-01-01):
5 * Update to Unicode 8.0.0. Fixes #680.
7 * Overhaul database compaction API. Add a Xapian::Database::compact() method,
8 with the Database object specifying the source database(s).
9 Xapian::Compactor is now just a functor to use if you want to control
10 progress reporting and/or the merging of user metadata. The existing API
11 has been reimplemented using the new one, but is marked as deprecated.
13 * Add support for a default value when sorting. Fixes #452, patch from
16 * Make all functor objects non-copyable. Previously some were, some weren't,
17 but it's hard to correctly make use of this ability. Fixes #681.
19 * Fix use after free with WILDCARD_LIMIT_MOST_FREQUENT. If we tried to open a
20 postlist after processing such a wildcard, the postlist hint could be
21 pointing to a PostList object which had been deleted. Fixes #696, reported
24 * Add support for optional reference counting of MatchSpy objects.
26 * Improve Document::get_description() - the output is now always valid UTF-8,
27 doesn't contain implementation details like "Document::Internal", and more
28 clearly reports if the document is linked to a database.
30 * Remove XAPIAN_CONST_FUNCTION marker from sortable_serialise_() helper, as it
31 writes to the passed in buffer, so it isn't const or pure. Fixes
32 decvalwtsource2 testcase failure when compiled with clang.
34 * Make PostingSource::set_maxweight() public - it's hard to wrap for the
35 bindings as a protected method. Fixes #498, reported by Richard Boulton.
39 * Add unit test for internal C_isupper(), etc functions.
43 * Optimise value range which is a superset of the bounds. If the value
44 frequency is equal to the doccount, such a range is equivalent to MatchAll,
45 and we now avoid having to read the valuestream at all.
47 * Optimise OP_VALUE_RANGE when the upper bound can't be exceeded. In this
48 case, we now use ValueGePostList instead of ValueRangePostList.
52 * Shave 2 bytes of every Btree item (which will probably typically reduce
53 database size by several percent).
55 * More compact item format for branch blocks - 2 bytes per item smaller. This
56 means each branch block can branch more ways, reducing the number of Btree
57 levels needed, which is especially helpful for cold-cache search times.
59 * Track an upper bound on spelling word frequency. This isn't currently used,
60 but will be useful for improving the spelling algorithm, and we want to
61 stabilise the glass backend format. See #225, reported by Philip Neustrom.
63 * Support 64-bit docids in the glass backend on-disk format. This changes the
64 encoding used by pack_uint_preserving_sort() to one which supports 64 bit
65 values, and is a byte smaller for values 16384-32767, and the same size for
66 all other 32 bit values. Fixes #686, from original report by James Aylett.
68 * Use memcpy() not memmove() when no risk of overlap.
70 * Store length of just the key data itself, allowing keys to be up to 255 bytes
71 long - the previous limit was 252.
73 * Change glass to store DB stats in the version file. Previously we stored
74 them in a special item in the postlist table, but putting them in the version
75 file reduces the number of block reads required to open the database, is
76 simpler to deal with, and means we can potentially recalculate tight upper
77 and lower bounds for an existing database without having to commit a new
80 * Add support for a single-file variant for glass. Currently such databases
81 can only be opened for reading - to create one you need to use
82 xapian-compact (or its API equivalent). You can embed such databases within
83 another file, and open them by passing in a file descriptor open on that file
84 and positioned at the offset the database starts at). Database::check() also
85 supports them. Fixes #666, reported by Will Greenberg (and previously
86 suggested on xapian-discuss by Emmanuel Engelhart).
88 * Avoid potential DB corruption with full-compaction when using 64K blocks.
90 * Where posix_fadvise() is available, use it to prefetch postlist Btree blocks
91 from the level below the root block which will be needed for postlists of
92 terms in the query, and similarly for the docdata table when MSet::fetch() is
93 called. Based on patch by Will Greenberg in #671.
97 * Where posix_fadvise() is available, use it to prefetch postlist Btree blocks
98 from the level below the root block which will be needed for postlists of
99 terms in the query, and similarly for the record table when MSet::fetch() is
100 called. Based on patch by Will Greenberg in #671.
104 * Fix hook for remote support of user weighting schemes. The commented-out
105 code used entirely the wrong class - now we use the server object we have
106 access to, and forward the method to the class which needs it.
110 * New configure options --enable-64bit-docid and --enable-64bit-termcount,
111 which control the size of these types. Because these types are used in
112 the API, libraries built with different combinations of them won't be ABI
113 compatible. Based heavily on patch from James Aylett and Dylan Griffith.
116 * Sort out hiding most of the internal symbols which had public visibility
117 for various reason. Mostly addresses #63.
121 * xapian-inspect: We no longer install this - it's really an aid to Xapian
122 development rather than a user tool.
126 * Minimum supported GCC version is now documented as GCC 4.7, for C++11
127 support. Previously we documented 4.7 as the oldest known to work.
129 * Use CLOCK_REALTIME with timer_create() on Cygwin.
131 * Don't include winsock headers on Cygwin. Instead include <arpa/inet.h> for
134 * Handle AI_ADDRCONFIG not being defined by some mingw versions.
136 * Fix to handle mingw now providing a nanosleep() function.
138 * Use WSAAddressToString instead of inet_ntop under __WIN32__ - at least under
139 mingw we don't seem to have inet_ntop().
141 * Fix testsuite to compile when S_ISSOCK() isn't defined.
145 * Add missing parameters to debug logging for a few methods.
147 Xapian-core 1.3.3 (2015-06-01):
149 This release includes all changes from 1.2.20-1.2.21 which are relevant.
155 + Add new flag Xapian::DB_RETRY_LOCK which allows opening a database for
156 writing to wait until it can get a write lock. (Fixes #275, reported by
159 + Fix Database::get_doclength_lower_bound() over multiple databases when some
160 are empty or consist only of zero-length documents. Previously this would
161 report a lower bound of zero, now it reports the same lowest bound as a
162 single database containing all the same documents.
164 + Database::check(): When checking a single table, handle the ".glass"
165 extension on glass database tables, and use the extension to guide the
166 decision of which backend the table is from.
170 + Add new OP_WILDCARD query operator, which expands wildcards lazily, so now
171 we create the PostList tree for a wildcard directly, rather than creating
172 an intermediate Query tree. OP_WILDCARD offers a choice of ways to limit
173 wildcard expansion (no limit, throw an exception, use the first N by term
174 name, or use the most frequent N). (See tickets #48 and #608).
178 + Add new set_max_expansion() method which provides access to OP_WILDCARD's
179 choice of ways to limit expansion and can set limits for partial terms as
180 well as for wildcards. Partial terms now default to the 100 most frequent
181 matching terms. (Completes #608, reported by boomboo).
183 + Deprecate set_max_wildcard_expansion() in favour of set_max_expansion().
185 * Add support for optional reference counting of FieldProcessor and
186 ValueRangeProcessor objects.
190 * If command line option --verbose/-v isn't specified, set the verbosity level
191 from environmental variable VERBOSE.
193 * Re-enable replicate3 for glass, as it no longer fails.
195 * Add more test coverage for get_unique_terms().
197 * Don't leave an extra fd open when starting xapian-tcpsrv for remotetcp tests.
201 * When reporting freelist errors during a database check, distinguish between a
202 block in use and in the freelist, and a block in the freelist more than once.
204 * Fix compaction and database checking for the change to the format of keys
205 in the positionlist table which happened in 1.3.2.
207 * After splitting a block, we always insert the new block in the parent right
208 after the block it was split from - there's no need to binary chop.
210 * Avoid infinite recursion when we hit the end of the freelist block we're
211 reading and the end of the block we're writing at the same time.
213 * Fix freelist handling to allow for the newly loaded first block of the
214 freelist being already used up.
218 * Fix problems with get_unique_terms() on a modified chert database.
220 * Fix xapian-check on a single chert table, which seg faulted in 1.3.2.
224 * Avoid dividing zero by zero when calculating the average length for an empty
229 * Merge generate-allsnowballheaders script into collate-sbl.
233 * A compiler with good support for C++11 is now required to build Xapian.
234 Most of the actively developed C++ compilers already have decent support,
235 or are close to having it, and it makes development easier and more
236 efficient. Currently known to work: GCC >= 4.7, recent versions of clang
237 (3.5 works). Solaris Studio 12.4 compiles the code, but tests currently
238 fail. IBM's xlC doesn't support enough of C++11 yet. HP's aCC hasn't
239 been tested, but its documentation suggests it also doesn't support enough
242 * Drop workarounds and special cases for old versions of various compilers
243 which don't support C++11.
245 * Use C++11's static_assert() and unique_ptr instead of custom implementations
246 of equivalent functionality.
248 * Building on OS/2 with EMX is no longer supported - EMX was last updated in
249 2001 and comes with GCC 3.2.1, which is much too old to support C++11.
251 * Building with SGI's and Compaq's C++ compilers is no longer supported -
252 both seem to have ceased development, and don't support C++11.
254 * Building with STLport is no longer supported - STLport was last released in
255 2008, so it's no longer actively developed and won't support C++11.
257 * Building on IRIX is no longer supported, because IRIX has reached end of
260 * Disable "<FUNCTION> is expected to return a value" warning from Sun's C++
261 compiler, as it fires for functions which end in a "throw" statement.
262 Genuine instances of missing return values will be caught by compilers with
263 superior warning machinery.
265 * Fix warning from GCC 5.1 where template expansion leads to the comparison
266 (bool_value < 255) which is always true. Warning introduced by changes in
269 * Use getaddrinfo() instead of gethostbyname(), since the latter may not be
270 thread-safe, and as a step towards IPv6 support (see #374), but currently we
271 still only look for IPv4 addresses.
273 * timer_create() seems to always fail on AIX with EAGAIN, so just skip the
274 matchtimelimit1 testcase there.
276 * Under __WIN32__, we need to specify Vista as the minimum supported version to
277 get the AI_ADDRCONFIG flag. Older versions seem to all be out of support
280 * Change configure probe for log2() to check for a declaration in <cmath>
281 to get it to fix build on Solaris with Sun C++. C++11 compilers should all
282 provide log2(), but let's not rely on that just yet as it's easy to provide a
283 fallback implementation.
285 * Use scalbn() instead of ldexp() where possible (which we can in all cases
286 when FLT_RADIX == 2, as it is on pretty much all current platforms). On
287 overflow and underflow ldexp() sets errno, which it seems better to avoid
290 * The list of stemmers is now in the same static const struct as the version
291 info, and Stem::get_available_languages() is just an inlined wrapper which
292 fetches this structure and returns the appropriate member. This saves a
293 relocation, reducing library load time a little.
295 * Remove "pure" attribute from API functions which could throw an exception.
296 These functions aren't really pure, and while we're happy for calls to them
297 to be CSE-ed or eliminated entirely, the compiler might make more assumptions
298 than that about a pure function - clang seems to assume pure => nothrow and
299 an exception from such a function can't be caught.
301 * Remove "pure" attribute from sortable_unserialise(), which can raise floating
302 point exceptions FE_OVERFLOW and FE_UNDERFLOW.
304 * Add "nothrow" attribute to more API functions which will never throw an
307 * Make sortable_serialise() an inlined wrapper around a function which won't
308 throw and can be flagged with attribute 'const'.
310 * Tweak sortable_unserialise() not to compare with a fixed string by
311 constructing a temporary std::string object (which could throw
312 std::bad_alloc), and mark it as XAPIAN_NOTHROW.
316 * Only enable assertions in sortable_serialise() and sortable_unserialise() in
317 the testsuite (since these functions shouldn't throw exceptions), and move
318 the tests of these functions from queryparsertest to unittest to facilitate
321 * Add more assertions to the glass backend code.
323 Xapian-core 1.3.2 (2014-11-24):
325 This release includes all changes from 1.2.16-1.2.19 which are relevant.
329 * Update Unicode character database to Unicode 7.0.0.
331 * New Xapian::Snipper class from Mihai Bivol's GSOC 2012 project. (mostly
334 * Fix all get_description() methods to always return UTF-8 text. (fixes #620)
338 + Alter to take its "out" parameter as a pointer to std::ostream instead of a
339 reference, and make passing NULL mean "do not produce output", and make
340 the second and third parameters optional, defaulting to a quiet check.
342 + Escape invalid UTF-8 data in keys and tags reported by xapian-check, using
343 the same code we use to clean up strings returned by get_description()
346 + Correct failure message which talks above the root block when it's actually
349 + Rename DBCHECK_SHOW_BITMAP to DBCHECK_SHOW_FREELIST (old name still
350 provided for now, but flagged as deprecated - DBCHECK_SHOW_BITMAP was new
351 in 1.3.0, so will likely be removed before 1.4.0).
353 * Methods and functions which take a string to unserialise now consistently
354 call that parameter "serialised".
356 * Weight: Make number of distinct terms indexing each document and the
357 collection frequency of the term available to subclasses. Patch from
358 Gaurav Arora's Language Modelling branch.
360 * WritableDatabase: Add support for multiple subdatabases, and support opening
361 a stub database containing multiple subdatabases as a WritableDatabase.
363 * WritableDatabase can now be constructed from just a pathname (defaulting to
364 opening the database with DB_CREATE_OR_OPEN).
366 * WritableDatabase: Add flags which can be bitwise OR-ed into the second
367 argument when constructing:
369 + Xapian::DB_NO_SYNC: to disable use of fsync, etc
371 + Xapian::DB_DANGEROUS: to enable in-place updates
373 + Xapian::DB_BACKEND_CHERT: if creating, create a chert database
375 + Xapian::DB_BACKEND_GLASS: if creating, create a glass database
377 + Xapian::DB_NO_TERMLIST: create a database without a termlist (see #181)
379 + Xapian::DB_FULL_SYNC flag - if this is set for a database, we use the Mac
380 OS X F_FULL_SYNC instead of fdatasync()/fsync()/etc on the version file
383 * Database: Add optional flags argument to constructor - the following can be
384 bitwise OR-ed into it:
386 + Xapian::DB_BACKEND_CHERT (only open a chert database)
388 + Xapian::DB_BACKEND_GLASS (only open a glass database)
390 + Xapian::DB_BACKEND_STUB (only open a stub database)
392 * Xapian::Auto::open_stub() and Xapian::Chert::open() are now deprecated in
393 favour of these new flags.
395 * Add LMWeight class, which implements the Unigram Language Modelling weighting
396 scheme. Patch from Gaurav Arora.
398 * Add implementations of a number of DfR weighting schemes (BB2, DLH, DPH,
399 IfB2, IneB2, InL2, PL2). Patches from Aarsh Shah.
401 * Add support for the Bo1 query expansion scheme. Patch from Aarsh Shah.
403 * Add Enquire::set_time_limit() method which sets a timelimit after which
404 check_at_least will be disabled.
406 * Database: Trying to perform operations on a database with no subdatabases now
407 throws InvalidOperationError not DocNotFoundError.
409 * Query: Implement new OP_MAX query operator, which returns the maximum weight
410 of any of its subqueries. (see #360)
412 * Query: Add methods to allow introspection on Query objects - currently you
413 can read the leaf type/operator, how many subqueries there are, and get a
414 particular subquery. For a query which is a term, Query::get_terms_begin()
415 allows you to get the term. (see #159)
417 * Query: Only simplify OP_SYNONYM with a single subquery if that subquery is a
420 * Avoid two vector copies when storing term positions in most common cases.
422 * Reimplement version functions to use a single function in libxapian which
423 returns a pointer to a static const struct containing the version
424 information, with inline wrappers in the API header which call this. This
425 means we only need one relocation instead of 4, reducing library load time a
428 * Make TermGenerator flags an anonymous enum, and typedef TermGenerator::flags
429 to int for backward compatibility with existing user code which uses it.
431 * Stem: Fix incorrect Unicode codepoints for o-double-acute and u-double-acute
432 in the Hungarian Snowball stemmer. Reported by Tom Lane to snowball-discuss.
434 * Stem: Add an early english stemmer.
436 * Provide the stopword lists from Snowball plus an Arabic one, installed in
437 ${prefix}/share/xapian-core/stopwords/. Patch from Assem Chelli, fixes #269.
439 * Improve check for direct inclusion of Xapian subheaders in user code to
442 * Add simple API to help with creating language-idiomatic iterator wrappers
443 in <xapian/iterator.h>.
447 * Extend checkstatsweight1 to check that Weight::get_collection_freq() returns
448 the same number as Database::get_collection_freq().
450 * queryparsertest: Add testcase for FieldProcessor on boolean prefix with
453 * queryparsertest: Enable some disabled cases which actually work (in some
454 cases with slightly tweaked expected answers which are equivalent to those
457 * Make use of the new writable multidatabase feature to simplify the
458 multi-database handling in the test harness.
460 * Change querypairwise1_helper to repeat the query build 100 times, as with a
461 fast modern machine we were sometimes trying with so many subqueries that we
462 would run out of stack.
464 * apitest: Use Xapian::Database::check() in cursordelbug1. (partly addresses
467 * apitest: Test Query ops with a single MatchAll subquery.
469 * apitest: New testcase readonlyparentdir1 to ensure that commit works with a
470 read-only parent directory.
474 * Streamline collation of statistics for use by weighting schemes - tests show
475 a 2% or so increase in speed in some cases.
477 * If a term matches all documents and its weight doesn't depend on its wdf, we
478 can optimise it to MatchAll (the previous requirement that maxpart == 0 was
479 unnecessarily strict).
481 * Fix the check for a term which matches all documents to use the sub-db
482 termfreq, not the combined db termfreq.
484 * When we optimise a postlist for a term which matches all documents to use
485 MatchAll, we still need to set a weight object on it to get percentages
486 calculated correctly.
490 * 'brass' backend renamed to 'glass' - we decided to use names in ascending
491 alphabetical order to make it easier to understand which backend is newest,
492 and since 'flint' was used recently, we skipped over 'd', 'e' and 'f'.
494 * Change positionlist keys to be ordered by term first rather than docid first,
495 which helps phrase searching significantly. For more efficient indexing,
496 positionlist changes are now batched up in memory and written out in key
499 * Use a separate cursor for each position list - now we're ordering the
500 position B-tree by term first, phrase matching would cause a single cursor
501 to cycle between disparate areas of the B-tree and reread the same blocks
504 * Reference count blocks in the btree cursor, so cursors can cheaply share
505 blocks. This can significantly reduce the amount of memory used by cursors
506 for queries which contain a lot of terms (e.g. wildcards which expand to a
509 * Under glass, optimise the turning of a query into a postlist to reuse the
510 cursor blocks which are the same as the previous term's postlist. This is
511 particularly effective for a wildcard query which expands to a lot of terms.
513 * Keep track of unused blocks in the Btrees using freelists rather than
516 * Eliminate the base files, and instead store the root block and freelist
517 pointers in the "iamglass" file.
519 * When compacting, sync all the tables together at the end.
521 * In DB_DANGEROUS mode, update the version file in-place.
523 * Only actually store the document data if it is non-empty. The table which
524 holds the document data is now lazily created, so won't exist if you never
525 set the document data.
529 * Improve DBCHECK_FIX:
531 + if fixing a whole database, we now take the revision from the first table
532 we successfully look at, which should be correct in most cases, and is
533 definitely better than trying to determine the revision of each broken
536 + handle a zero-sized .DB file.
538 + After we successfully regenerate baseA, remove any empty baseB file to
539 prevent it causing problems. Tracked down with help from Phil Hands.
543 * Bump remote protocol version to 38.0, due to extra statistics being tracked
546 * Make Weight::Internal track if any max_part values are set, so we don't need
547 to serialise them when they've not been set.
551 * Fix conditional for enabling replication code - if chert is disabled but
552 glass isn't, we should still enable it.
554 * configure: Add hint for which package to install for rst2html
558 * Don't build, ship or install PDF versions of the API docs by default, but
559 provide an easy way for people to build it for themselves if they want it.
561 * Convert equations in rst docs to use LaTeX via the math role and directive.
563 * Actually ship, process and install geospatial.rst.
565 * postingsource.rst: Use a modern class in postingsource example. (Noted by
568 * Move the protocol docs for the remote and replication protocols into the net/
571 * Remove the dir_contents files and all the machinery to handle them.
573 * HACKING: Note we now use doxygen 1.8.8 for 1.3.x snapshots and releases.
575 * HACKING: Now using libtool 2.4.3 to bootstrap snapshots and 1.3.x releases.
577 * HACKING: Now using automake 1.14.1 to bootstrap snapshots and 1.3.x releases.
579 * HACKING: Drop note about needing git-svn if you're using git - bootstrap now
580 only uses git-svn if your Xapian tree was checked out using git-svn.
582 * HACKING: Need sphinx-doc to generate API docs for Python and Python 3 bindings.
584 * HACKING: Note that MacTeX seems to be the best option if using homebrew.
588 * Don't pass an integer argument to log(), to avoid ambiguity errors with xlC
589 and Sun's C++ compiler. (fixes #627)
591 * Fix compilations issues with Sun's C++ compiler (mostly missing library
594 * Implement RealTime::now() using clock_gettime() where it's available, since
595 it can provide nanosecond resolution.
597 * Implement RealTime::sleep() using nanosleep() where it's available, since it
598 has a simpler API and a finer resolution than select().
600 * Use lround() instead of round() in geospatial code, since we want the result
601 as an int. GCC 4.4.3 seems to optimise to use lround() anyway, but other
604 * Include <math.h> for lround()/round(). (fixes #628)
606 * Drop code supporting Microsoft Windows 9x which reached EOL in 2006.
608 * Under C++11, use unique_ptr for AutoPtr.
610 * Stop using a reference where we may end up passing *NULL, as that's invalid.
611 Thanks Nick Lewycky and ubsan for helping track this down.
613 * In DLHWeight and DPHWeight, avoid dividing by zero when the collection size
618 * Fix assertion failure when built with --enable-assertions. The behaviour
619 when built without assertions happened to be correct.
621 * Fix assertion in BitReader::decode(), and remove 'Assert(rd);' in two places
622 where rd is no longer a pointer.
624 Xapian-core 1.3.1 (2013-05-03):
626 This release includes all changes from 1.2.10-1.2.15 which are relevant.
630 * Give an compilation error if user code tries to include API headers other
631 than xapian.h directly - these other headers are an internal implementation
632 detail, but experience has shown that some people try to include them
633 directly. Please just use '#include <xapian.h>' instead.
635 * Update Unicode character database to Unicode 6.2.0.
637 * Add FieldProcessor class (ticket#128) - currently marked as an experimental
638 API while we sort out how best to sort out exactly how it interacts with
639 other QueryParser features.
641 * Add implementation of several TF-IDF weighting schemes via a new TfIdfWeight
644 * Add ExpandDeciderFilterPrefix class which only return terms with a particular
647 * QueryParser: Adjust handling of Unicode opening/closing double quotes - if a
648 quoted boolean term was started with ASCII double quote, then only ASCII
649 double quote can end it, as otherwise it's impossible to quote a term
650 containing Unicode double quotes.
652 * Database::check(): If the database can't be opened, don't emit a bogus
653 warning about there being too many documents to cross-check doclens.
655 * TradWeight,BM25Weight: Throw SerialisationError instead of NetworkError if
658 * QueryParser: Change the default stemming strategy to STEM_SOME, to eliminate
659 the API gotcha that setting a stemmer is ignored until you also set a
662 * Deprecate Xapian::ErrorHandler. (ticket#3)
664 * Stem: Generate a compact and efficient table to decode language names. This
665 is both faster and smaller than the approach we were using, with the added
666 benefit that the table is auto-generated.
670 + Add check for Qt headers being included before us and defining
671 'slots' as a macro - if they are, give a clear error advising how to work
672 around this (previously compilation would fail with a confusing error).
674 + Add a similar check for Wt headers which also define 'slots' as a macro
679 * tests/generate-api_generated: Test that the string returned by a
680 get_description() method isn't empty.
682 * Use git commit hash in title of test coverage reports generated from a git
687 * Drop MatchNothing subqueries in OR-like situations in add_subquery() rather
688 than adding them and then handling it later.
690 * Handle the left side of AND_NOT and AND_MAYBE being MatchNothing in
691 add_subquery() rather than in done().
693 * Handle QueryAndLike with a MatchNothing subquery in add_subquery() rather
696 * Query: Multi-way operators now store their subquery pointers in a custom
697 class rather than std::vector<Xapian::Query>. The custom class take the
698 same amount of space, or often less. It's particularly efficient when
699 there are two subqueries, which is very desirable as we no longer flatten a
700 subtree of the same operator as we build the query.
702 * Optimise an unweighted query term which matches all the documents in a
703 subdatabase to use the "MatchAll" postlist. (ticket#387)
707 * Iterating positional data now decodes it lazily, which should speed up
708 phrases which include common words.
710 * Compress changesets in brass replication. Increments the changeset version.
713 * Restore two missing lines in database checking where we report a block with
716 * When checking if a block was newly allocated in this revision, just look
717 at its revision number rather than consulting the base file's bitmap.
721 * Iterating positional data now decodes it lazily, which should speed up
722 phrases which include common words.
726 * Prefix compress list of terms and metadata keys in the remote protocol.
727 This requires a remote protocol major version bump.
731 * Fix the 'libxapian' to be 'libxapian-1.3' and 'xapian.m4' to be
732 'xapian-1.3.m4' (this was supposed to be the case for 1.3.0, but the
733 change wasn't made correctly).
735 * Remove support for 'configure --enable-quiet', 'make QUIET=' and 'make
736 QUIET=y' - automake now supports 'configure --enable-silent-rules', 'make
737 V=1' and 'make V=0' which are broadly equivalent and more standard.
739 * configure: If we fail to find a function needed for the remote backend, don't
740 autodisable it - it's more helpful to error out so the use can decide if they
741 want to pass --disable-backend-remote to disable it, or work out what values
742 to pass for LIBS, etc to make it work. This also matches what we do for the
745 * automake 1.13.1 is now used to generate snapshots and releases.
747 * Add check-syntax make target to support editor syntax checks.
749 * Fix to build when configured with --disable-backend-brass
750 --disable-backend-chert. (ticket#586)
752 * Generate a check for compatible _DEBUG settings if built with MSVC.
755 * If you run "make coverage-check" by hand, the previous default of compressed
756 HTML is unhelpful, so don't default to passing --html-gzip to genhtml, but
757 instead add support for GENHTML_ARGS.
759 * API methods and functions are now marked as 'const', 'pure', or 'nothrow'
760 allowing compilers which support such annotations to generate more efficient
761 code. (tickets #151, #454)
765 * HACKING: Note which MacPorts are needed for development work.
767 * docs/remote_protocol.rst: Correct error in documentation of REPLY_DOCDATA
772 * xapian-check: Add "fix" option, which currently will regenerate iamchert if
773 it isn't valid, and will regenerate base files from the .DB files (only
774 really tested on databases which have just been compacted).
778 * Fix warning with GCC in build with assertions enabled.
780 * common/fileutils.cc: Add safeunistd.h for mkdir, required by GCC 4.7
781 (reported by Gaurav Arora).
783 * backends/brass/brass_databasereplicator.cc: Use new/delete to avoid variable
784 length array gcc extension and comply with c++98
786 * Mark file descriptors as close-on-exec where supported.
788 * api/queryinternal.cc: Need <functional> for mem_fun().
790 * Work around Apple's OS X SDK defining a check() macro.
792 * Add an option to use a flock() based locking implementation for brass and
793 chert - this is much simpler than using fcntl() due to saner semantics around
794 releasing locks when closing other descriptors on the same file (at least on
795 platforms where flock() isn't just a compatibility wrapper around fcntl()).
796 Sadly we can't simply switch to this without breaking locking compatibility
797 with previous releases, but it's useful for platforms without fcntl()
798 locking (it's enabled for DJGPP) and may be useful for custom builds for
803 * xapian-core.spec: Remove xapian-chert-update.
807 * Building with --enable-log works once again.
809 Xapian-core 1.3.0 (2012-03-14):
813 * Update Unicode character database to Unicode 6.1.0. (ticket#497)
815 * TermIterator returned by Enquire::get_matching_terms_begin(),
816 Query::get_terms_begin(), Database::synonyms_begin(),
817 QueryParser::stoplist_begin(), and QueryParser::unstem_begin() now stores the
818 list of terms to iterate much more compactly.
822 + Allow Unicode curly double quote characters to start and/or end phrases.
824 + The set_default_op() method will now reject operators which don't make
825 sense to set. The operators which are allowed are now explicitly
826 documented in the API docs.
828 * Query: The internals have been completely reimplemented (ticket#280). The
831 + Query objects are smaller and should be faster.
833 + More readable format for Query::get_description().
835 + More compact serialisation format for Query objects.
837 + Query operators are no longer flattened as you build up a tree (but the
838 query optimiser still combines groups of the same operator). This means
839 that Query objects are truly immutable, and so we don't need to copy Query
840 objects when composing them. This should also fix a few O(n*n) cases when
841 building up an n-way query pair-wise. (ticket#273)
843 + The Query optimiser can do a few extra optimisations.
845 * There's now explicit support for geospatial search (this API is currently
846 marked as experimental). (ticket#481)
848 * There's now an API (currently experimental) for checking the integrity of
849 databases (partly addresses ticket#238).
851 * Database::reopen() now returns true if the database may have been reopened
852 (previously it returned void). (ticket#548)
854 * Deprecate Xapian::timeout in favour of POSIX type useconds_t.
856 * Deprecate Xapian::percent and use int instead in the API and our own code.
858 * Deprecate Xapian::weight typedef in favour of just using double and change
859 all uses in the API and our own code. (ticket#560)
861 * Rearrange members of Xapian::Error to reduce its size (from 48 to 40 bytes on
864 * Assignment operators for PositionIterator and TermIterator now return *this
867 * PositionIterator, PostingIterator, TermIterator and ValueIterator now
868 handle their reference counts in hand-crafted code rather than using
869 intrusive_ptr/RefCntPtr, which means the compiler can inline the destructor
870 and default constructor, so a comparison to an end iterator should now
871 optimise to a simple NULL pointer check, but without the issues which the
872 ValueIteratorEnd_ proxy class approach had (such as not working in templates
873 or some cases of overload resolution).
877 + Previously, Enquire::get_matching_terms_begin() threw InvalidArgumentError
878 if the query was empty. Now we just return an end iterator, which is more
879 consistent with how empty queries behave elsewhere.
881 + Remove the deprecated old-style match spy approach of using a MatchDecider.
883 * Remove deprecated Sorter class and MultiValueSorter subclass.
887 + Add stemmers for Armenian (hy), Basque (eu), and Catalan (ca).
889 + Stem::operator= now returns a reference to the assigned-to object.
893 * Make unittest use the test harness, so it gets all the valgrind and fd leak
894 checks, and other handy features all the other tests have.
896 * Improve test coverage in several places.
898 * Compress generated HTML files in coverage report.
902 * Remove flint backend.
906 * When propagating exceptions from a remote backend server, the protocol now
907 sends a numeric code to represent which exception is being propagated, rather
908 than the name of the type, as a number can be turned back into an exception
909 with a simple switch statement and is also less data to transfer.
912 * Remote protocol (these changes require a protocol major version bump):
914 + Unify REPLY_GREETING and REPLY_UPDATE.
916 + Send (last_docid - doccount) instead of last_docid and (doclen_ubound -
917 doclen_lbound) instead of doclen_ubound.
919 * Remove special check which gives a more helpful error message when a modern
920 client is used against a remote server running Xapian <= 0.9.6.
924 * Various changes allow us to now remove XAPIAN_VISIBILITY_DEFAULT from a
925 number of functions which aren't in the public API (partly addresses
928 * configure: For this development series, the library gets a -1.3 suffix and
929 include files are installed with an extra /xapian-1.3 component to make
930 parallel installs easier.
932 * configure: Enable -fshow-column for GCC - things like vim's quickfix mode
933 will then jump to the appropriate column for a compiler error or warning, not
934 just the appropriate line.
936 * Snowball compiler now reports "FILE:LINE:" before each error so tools like
937 vim's quickfix mode can parse this and bring up the line with the error
940 * docs/doxygen_api.conf.in: Don't generate XML from doxygen for the bindings -
941 the bindings now do this for themselves. (ticket#262)
945 * INSTALL: Update GCC details - we now recommend 4.3 or newer (was 4.1), and
946 note that while 3.1 is the hard minimum requirement, the oldest we've tested
947 with at all recently was 3.3.
949 * docs/deprecation.rst: Updated.
955 + Move delve from examples to bin and rename to xapian-delve.
957 + Send errors to stderr not stdout.
959 * xapian-check: Now reports useful descriptions rather than cryptic numeric
960 codes for B-tree errors.
964 * Add assertions that the index is in range when dereferencing MSetIterator and
967 * Fix various errors in debug logging statements.
969 * Add QUERY category for debug logging.
971 Xapian-core 1.2.22 (2015-12-29):
975 * Add FLAG_CJK_NGRAM for QueryParser and TermGenerator. Has the same effect as
976 setting the environment variable XAPIAN_CJK_NGRAM. Fixes #180, reported by
977 Richard Boulton, with contributions from Pavel Strashkin, Mikkel Kamstrup
978 Erlandsen and Brandon Schaefer.
980 * Fix bug parsing multiple non-exclusive filter terms - previously this could
981 result in such filters effectively being ignored.
983 * Fix Database::get_doclength_lower_bound() over multiple databases when some
984 are empty or consist only of zero-length documents. Previously this would
985 report a lower bound of zero, now it reports the same lowest bound as a
986 single database containing all the same documents.
988 * Make Database::get_wdf_upper_bound("") return 0.
990 * Mark constructors taking a single argument as "explicit" to avoid unwanted
991 implicit conversions.
995 * If command line option --verbose/-v isn't specified, set the verbosity level
996 from environmental variable VERBOSE.
998 * Skip timed tests if $AUTOMATED_TESTING is set. Fixes #553, reported by
1001 * Don't leave an extra fd open when starting xapian-tcpsrv for remotetcp tests.
1003 * apitest: Revert disabling of part of adddoc5 for clang - the test failure was
1004 in fact due to a bug in 1.3.x, and 1.2.x was never affected.
1006 * apitest: Tweak bounds checks in dbstats1 testcase - multi backends should
1011 * Format limit on docid now correctly imposed when sizeof(int) > 4.
1013 * Avoid potential DB corruption with full-compaction when using 64K blocks.
1017 * Format limit on docid now correctly imposed when sizeof(int) > 4.
1019 * Avoid potential DB corruption with full-compaction when using 64K blocks.
1023 * Format limit on docid now correctly imposed when sizeof(int) > 4.
1025 * Avoid potential DB corruption with full-compaction when using 64K blocks.
1029 * Fix to handle total document length exceeding 34,359,738,368. (Fixes #678,
1032 * Avoid dividing by zero when getting the average length for an empty database.
1034 * Stop apparent error from remote server when read-only client disconnects. A
1035 read-only client just closes the connection when done, but the server
1036 previously reported "Got exception NetworkError: Received EOF", which sounds
1037 like there was a problem. Now we just say "Connection closed" here, and
1038 "Connection closed unexpectedly" if the client connects in the middle of an
1039 exchange. Possibly fixes #654, reported by German M. Bravo.
1041 * Give a clearer error message when the client and server remote protocol
1042 versions aren't compatible.
1044 * Check length of key in MSG_SETMETADATA.
1048 * pkg-config: Fix library name in .pc file to say "xapian" not "xapian-core".
1049 Reported by Eric Lindblad to the xapian-devel list.
1051 * Private symbol decode_length() is no longer visible outside the library.
1055 * Stop maintaining ChangeLog files. They make merging patches harder, and stop
1056 'git cherry-pick' from working as it should. The git repo history should be
1057 sufficient for complying with GPLv2 2(a).
1059 * Strip out "quickstart" examples which are out of date and rather redundant
1060 with the "simple" examples.
1062 * Correct documentation of Enquire::get_query(). If no query has been set,
1063 the documentation said Xapian::InvalidArgumentError was thrown, but in
1064 fact we just return a default initialised Query object (i.e. Query()). This
1065 seems reasonable behaviour and has been the case since Xapian 0.9.0.
1067 * Document xapian-compact --blocksize takes an argument.
1069 * Update snowball website link to snowballstem.org.
1073 * xapian-replicate: Fix replication for files > 4GB on 32-bit platforms.
1074 Previously replication would fail to copy a file whose size didn't fit in
1075 size_t. Fixes #685, reported by Josh Elsasser.
1077 * xapian-tcpsrv: Better error if -p/--port not specified
1079 * quest: Support `-f cjk_ngram`.
1083 * xapian-metadata: Extend "list" subcommand to take optional key prefix.
1087 * Fix new warnings from recent versions of GCC and clang.
1089 * Add spaces between literal strings and macros which expand to literal strings
1090 for C++11 compatibility in __WIN32__-specific code.
1092 * Need <unistd.h> for unlink() on FreeBSD, reported by Germán M. Bravo via
1095 * Fix testsuite to build when S_ISSOCK() isn't defined.
1097 * Don't provide our own implementation of sleep() under __WIN32__ if there
1098 already is one - mingw provides one, and in some situations it seems to clash
1099 with ours. Reported to xapian-discuss by John Alveris.
1101 * Add missing '#include <arpa/inet.h>' to htons(). Seems to be implicitly
1102 included on most platforms, but Interix needs it. Reported by Eric Lindblad
1105 * Disable "<FUNCTION> is expected to return a value" warning from Sun's C++
1106 compiler, as it fires for functions ending in a "throw" statement. Genuine
1107 instances will be caught by compilers with superior warning machinery.
1109 * Prefer scalbn() to ldexp() where possible, since the former doesn't ever set
1112 * '#include <config.h>' in the "simple" examples, as when compiling with xlC on
1113 AIX, _LARGE_FILES gets defined by AC_SYS_LARGEFILE to enable large file
1114 support, and defining this changes the ABI of std::string, so it also needs
1115 to be defined when compiling code using Xapian.
1117 * On cygwin, include <arpa/inet.h> instead of winsock headers for htons() and
1120 * Include <cygwin/version.h> for CYGWIN_VERSION_API_MAJOR.
1122 * Avoid referencing static members via an object in that object's own
1123 definition, as this doesn't work with all compilers (noted with GCC 3.3), and
1124 is a bit of an odd construct anyway. Reported by Eric Lindblad on
1127 * GCC < 3.4.2 lacks operator<< overloads for unsigned long long on some
1128 platforms, so simply work around this by using str(), as this isn't
1129 performance sensitive code. Reported by Eric Lindblad on xapian-discuss.
1131 * Fix delete which should be delete[] in brass backend cursor code.
1133 Xapian-core 1.2.21 (2015-05-20):
1137 * QueryParser: Extend the set of characters allowed in the start of a range to
1138 be anything except for '(' and characters <= ' '. This better matches what's
1139 accepted for a range end (anything except for ')' and characters <= ' ').
1140 Reported by Jani Nikula.
1144 * Reimplement OP_PHRASE for non-exact phrases. The previous implementation was
1145 buggy, giving both false positives and false negatives in rare cases when
1146 three or more terms were involved. Fixes #653, reported by Jean-Francois
1149 * Reimplement OP_NEAR - the new implementation consistently requires the terms
1150 to occur at different positions, and fixes some previously missed matches.
1152 * Fix a reversed check for picking the shorter position list for an exact
1153 phrase of two terms. The difference this makes isn't dramatic, but can be
1154 measured (at least with cachegrind). Thanks to kbwt for spotting this.
1156 * When matching an exact phrase, if a term doesn't occur where we want, use
1157 its actual position to advance the anchor term, rather than just checking
1158 the next position of the anchor term.
1162 * Fix cursor versioning to consider cancel() and reopen() as events where
1163 the cursor version may need incrementing, and flag the current cursor version
1164 as used when a cursor is rebuilt. Fixes #675, reported by Germán M. Bravo.
1166 * Avoid using file descriptions < 3 for writable database tables, as it risks
1167 corruption if some code in the same process tries to write to stdout or
1168 stderr without realising it is closed. (Partly addresses #651)
1172 * Fix cursor versioning to consider cancel() and reopen() as events where
1173 the cursor version may need incrementing, and flag the current cursor version
1174 as used when a cursor is rebuilt. Fixes #675, reported by Germán M. Bravo.
1176 * Avoid using file descriptions < 3 for writable database tables, as it risks
1177 corruption if some code in the same process tries to write to stdout or
1178 stderr without realising it is closed. (Partly addresses #651)
1182 * Fix cursor versioning to consider cancel() and reopen() as events where
1183 the cursor version may need incrementing, and flag the current cursor version
1184 as used when a cursor is rebuilt. Fixes #675, reported by Germán M. Bravo.
1188 * Fix sort by value when multiple databases are in use and one or more are
1189 remote. This change necessitated a minor version bump in the remote
1190 protocol. Fixes #674, reported by Dylan Griffith. If you are upgrading a
1191 live system which uses the remote backend, upgrade the servers before the
1196 * The compiler ABI check in the public API headers now issues a warning
1197 (instead of an error) for an ABI mismatch for ABI versions 2 and later
1198 (which means GCC >= 3.4). The changes in these ABI versions are bug fixes
1199 for corner cases, so there's a good chance of things working - e.g. building
1200 xapian-bindings with GCC 5.1 (which defaults to ABI version 8) against
1201 xapian-core built with GCC 4.9 (which defaults to ABI version 2) seems to
1202 work OK. A warning is still useful as a clue to what is going on if linking
1203 fails due to a missing symbol.
1205 * xapian-config,xapian-core.pc: When compiling with xlC on AIX, the reported
1206 --cxxflags/--cflags now include -D_LARGE_FILES=1 as this is defined for the
1207 library, and defining it changes the ABI of std::string with this compiler,
1208 so it must also be defined when building code using the Xapian API.
1210 * xapian-core.pc: Include --enable-runtime-pseudo-reloc in --libs output for
1211 mingw and cygwin, like xapian-config does.
1213 * xapian-core.pc: Fix include directory reported by `pkg-config --cflags`.
1214 This bug was harmless if xapian-core was installed to a directory which was
1215 on the default header search path (such as /usr/include).
1217 * xapian-config: Fix typo so cached result of test in is_uninstalled() is
1218 actually used on subsequent calls. Fixes #676, reported (with patch) by Ryan
1221 * configure: Changes in 1.2.19 broke the custom macro we use to probe for
1222 supported compiler flags such that the flags never got used. This release
1225 * configure: Set default value for AUTOM4TE before AC_OUTPUT so the default
1226 will actually get used. Only relevant when building in maintainer mode
1229 * soaktest: Link with libtool's '-no-install' or '-no-fast-install', like we
1230 already do for other test programs, which means that libtool doesn't need to
1231 generate shell script wrappers for them on most platforms.
1235 * API documentation: Minor wording tweaks and formatting improvements.
1237 * docs/deprecation.rst: Add deprecation of omindex --preserve-nonduplicates
1238 which happened in 1.2.4.
1240 * HACKING: Update URL.
1242 * HACKING: libtool 2.4.6 is now used for bootstrapping snapshots and releases.
1246 * xapian-compact: Make sure we open all the tables of input databases at the
1247 same revision. (Fixes #649)
1249 * xapian-metadata: Add 'list' subcommand to list all the metadata keys.
1251 * xapian-replicate: Fix connection timeout to be 10 seconds rather than 10000
1252 seconds (the incorrect timeout has been the case since 1.2.3).
1254 * xapian-replicate: Set SO_KEEPALIVE for xapian-replicate's connection to the
1255 master, and add command line option to allow setting socket-level timeouts
1256 (SO_RCVTIMEO and SO_SNDTIMEO) on platforms that support them. Fixes #546,
1259 * xapian-replicate-server: Avoid potentially reading uninitialised data if a
1260 changeset file is truncated.
1264 * Add spaces between literal strings and macros which expand to literal strings
1265 for C++11 compatibility.
1267 * ValueCountMatchSpy::top_values_begin(): Fix the comparison function not to
1268 return true for two equal elements, which manifests as incorrect sorting in
1269 some cases when using clang's libc++ (which recent OS X versions do).
1271 * apitest: The adddoc5 testcase fails under clang due to an exception handling
1272 bug, so just #ifdef out the problematic part of the testcase when building
1275 * Fix clang warnings on OS X. Reported by Germán M. Bravo.
1277 * Fix examples to build with IBM's xlC compiler on AIX - they were failing due
1278 to _LARGE_FILES being defined for the library build but not for the examples,
1279 and defining this changes the ABI of std::string with this compiler.
1281 * configure: Improve the probe for whether the test harness can use RTTI to
1282 work for IBM's xlC compiler (which defaults to not generating RTTI).
1284 * Fix to build with Sun's C++ compiler.
1286 * Use F_DUPFD where available to dup to a file descriptor which is >= 2, rather
1287 than calling dup() until we get one.
1289 * When unserialising a double, avoid reading one byte past the end of the
1290 serialised value. In practice this was harmless on most platforms, as
1291 dbl_max_mantissa is 255 for IEEE-754 format doubles, and at least GCC's
1292 std::string keeps the buffer nul-terminated. Reported by Germán M. Bravo in
1295 * When unserialising a double, add missing cast to unsigned char when we check
1296 if the value will fit in the double type. On machines with IEEE-754 doubles
1297 (which is most current platforms) this happened to work OK before. It would
1298 also have been fine on machines where char is unsigned by default.
1300 * Fix incorrect use of "delete" which should be "delete []". This is
1301 undefined behaviour in C++, though the type is POD, so in practice this
1302 probably worked OK on many platforms.
1306 * Fix some overly strict assertions in flint, which caused apitest's
1307 cursordelbug1 to fail with assertions on.
1309 Xapian-core 1.2.20 (2015-03-04):
1313 * After splitting a block, we always insert the new block in the parent right
1314 after the block it was split from - there's no need to binary chop.
1318 * Generate and install a file for pkg-config. (Fixes#540)
1320 * configure: Update link to cygwin FAQ in error message.
1324 * include/xapian/weight.h: Document the enum stat_flags values.
1326 * docs/postingsource.rst: Use a modern class in postingsource example. (Noted
1329 * docs/deprecation.rst,docs/replication.rst: Fix typos.
1331 * Update doxygen configuration files to avoid warnings about obsolete tags from
1332 newer doxygen versions.
1334 * HACKING: Update details of building Xapian packages.
1338 * xapian-check: For chert and brass, cross-check the position and postlist
1339 tables to detect positional data for non-existent documents.
1343 * When locking a database for writing, use F_OFD_SETLK where available, which
1344 avoids having to fork() a child process to hold the lock. This currently
1345 requires Linux kernel >= 3.15, but it has been submitted to POSIX so
1346 hopefully will be widely supported eventually. Thanks to Austin Clements for
1347 pointing out this now exists.
1349 * Fix detection of fdatasync(), which appears to have been broken practically
1350 forever - this means we've probably been using fsync() instead, which
1351 probably isn't a big additional overhead. Thanks to Vlad Shablinsky for
1352 helping with Mac OS X portability of this fix.
1354 * configure: Define MINGW_HAS_SECURE_API under mingw to get _putenv_s()
1355 declared in stdlib.h.
1357 * Use POSIX O_NONBLOCK in preference to O_NDELAY - the semantics of the latter
1358 differ between BSD and System V.
1360 * According to POSIX, strerror() may not be thread safe, so use alternative
1361 thread-safe ways to translate errno values where possible.
1363 * On Microsoft Windows, avoid defining EADDRINUSE, etc if they're already
1364 defined, and use WSAE* constants un-negated - they start from a high value
1365 so won't collide with E* constants.
1369 * Add more assertions to the chert backend code.
1371 Xapian-core 1.2.19 (2014-10-21):
1375 * Xapian::BM25Weight:
1377 + Improve BM25 upper bound in the case when our wdf upper bound > our
1378 document length lower bound. Thanks to Craig Macdonald for pointing out
1381 + Pre-multiply termweight by (param_k1 + 1) rather than doing it for
1382 every weighted term in every document considered.
1386 * Don't report apparent leaks of fds opened on /dev/urandom - at least on
1387 Linux, something in the C library seems to lazily open it, and the report of
1388 a possible leak followed by assurance that it's OK really is just noise we
1393 * Fix false matches reported for non-exact phrases in some cases. Fixes the
1394 reduced testcase in #657, reported by Jean-Francois Dockes.
1398 * Only full sync after writing the final base file (only affects Max OS X).
1402 * Only full sync after writing the final base file (only affects Max OS X).
1406 * Only full sync after writing the final base file (only affects Max OS X).
1410 * For Sun's C++ compiler, pass -library=Crun separately since libtool looks for
1411 " -library=stlport4 " (with the spaces). (fixes#650)
1413 * Remove .replicatmp (created by the test suite) upon "make clean".
1417 * include/xapian/compactor.h: Fix formatting of doxygen comment.
1419 * HACKING: freecode no longer accepts updates, so drop that item from the
1422 * docs/overview.rst: Add missing database path to example of using
1423 xapian-progsrv in a stub database file.
1427 * Suppress unused typedef warnings from debugging logging macros, which occur
1428 in functions which always exit via throwing an exception when compiling with
1429 recent versions of GCC or clang.
1431 * Fix debug logging code to compile with clang. (fixes #657, reported by
1436 * Add missing RETURN() markup for debug logging in a few places, highlighted by
1437 warnings from recent GCC.
1439 * Fix incorrect return types in debug logging annotations so that code compiles
1440 when configured with --enable-log.
1442 Xapian-core 1.2.18 (2014-06-22):
1446 * Document: Fix get_docid() to return the docid for the sub-database (as it
1447 is explicitly documented to) for Document objects passed to functors like
1448 KeyMaker during the match. (fixes#636, reported by Jeff Rand).
1450 * Document: Don't store the termname in OmDocumentTerm - we were only using it
1451 in get_description() output and an exception message. Speeds up indexing
1452 etext.txt using simpleindex by 0.4%, and should reduce memory usage a bit
1453 too. (Change inspired by comments from Vishesh Handa on xapian-devel).
1455 * Database: Iterating the values in a particular slot is now a bit more
1456 efficient for inmemory and remote backends (but still slow compared to
1457 flint, chert and brass).
1461 * apitest: Expand crashrecovery1 to check that the expected base files exist
1462 and ones which shouldn't exist don't.
1464 * queryparsertest: Fix testcase for empty wildcard followed by negation to
1465 enable FLAG_LOVEHATE so the negation is actually parsed. Fortunately the
1466 fixed testcase passes.
1470 * OP_SYNONYM: avoid fetching the doclength if the weighting scheme doesn't
1471 need it and the calculated wdf for the synonym is <= doclength_lower_bound
1472 for the current subdatabase. (fixes #360)
1476 * Releases are now bootstrapped with libtool 2.4.2 instead of 2.4, and with
1477 config.guess and config.sub updated to the latest versions.
1481 * Add an example of initializing SimpleStopper using a file listing stopwords.
1482 (Patch from Assem Chelli)
1484 * Improve the descriptions of the stem_strategy values in the API docs.
1485 (Reported by "oilap" on #xapian)
1487 * docs/sorting.rst: Fix incorrect parameter types in Xapian::Weight
1490 * docs/glossary.rst: Add definition of "collection frequency".
1494 + makeindex is now in Debian package texlive-binaries.
1496 + Replace a link to the outdated autotools "goat book" with a link to the
1497 "Portable Shell" chapter of the autoconf manual.
1499 * include/xapian/base.h: Remove very out of date comments talking about atomic
1500 assignment and locking - since 0.5.0 we've adopted a "user locks" policy.
1501 (Reported by Jean-Francois Dockes)
1507 + Add -A <prefix> option to list all terms with a particular prefix.
1509 + Send errors to stderr not stdout.
1511 + If -v is specified more than once, show even more info in some cases.
1512 (NEWS file claimed this was backported in 1.2.15, but it actually wasn't).
1516 + Add --default-op option.
1518 + Add --weight option to allow the weighting scheme to be specified.
1522 * Explicitly '#include <algorithm>' for std::max(), fixing build with VS2013.
1523 (Fixes#641, reported by "boomboo").
1525 * Fix testcase blocksize1 not to try to delete an open database, which isn't
1526 possible under Windows. (Fixes #643, reported by Chris Olds)
1528 * docs/quickstart.rst: Split --cxxflags and --libs for portability (noted by
1529 "Hurricane Tong" on xapian-devel).
1531 * Fix warnings with clang 5.0.
1535 * Add assertions that weighting scheme upper bounds aren't exceeded.
1537 Xapian-core 1.2.17 (2014-01-29):
1541 * Enquire::set_sort_by_relevance_then_value() and
1542 Enquire::set_sort_by_relevance_then_key(): Fix sense of reverse parameter.
1543 Reported by "boomboo" on IRC.
1545 * BM25Weight: Fix case where (k1 == 0 || b == 0) but k2 != 0. Reported by
1548 * Unicode::tolower(): Fix to give correct results for U+01C5, U+01C8, U+01CB,
1549 and U+01F2 (previously these were left unchanged).
1553 * Automatically probe for and hook in eatmydata to the testsuite using the
1554 wrapper script it now includes.
1556 * Fix apitest to build when brass, chert or flint are disabled.
1560 * Fix handling of invalid block sizes passed to Xapian::Brass::open() - the
1561 size gets fixed as documented, but the uncorrected size was passed to the
1562 base file (and abort() was called if 0 was passed).
1564 * Validate "dir_end" when reading a block. (fixes #592)
1568 * Fix handling of invalid block sizes passed to Xapian::Chert::open() - the
1569 size gets fixed as documented, but the uncorrected size was passed to the
1570 base file (and abort() was called if 0 was passed).
1572 * Validate "dir_end" when reading a block. (fixes #592)
1576 * Fix handling of invalid block sizes passed to Xapian::Flint::open() - the
1577 size gets fixed as documented, but the uncorrected size was passed to the
1578 base file (and abort() was called if 0 was passed).
1580 * Validate "dir_end" when reading a block. (fixes #592)
1584 * configure: Improve reporting of GCC version.
1586 * Use -no-fast-install on platforms where -no-install causes libtool to emit a
1589 * docs/Makefile.am: Fix handling of MAINTAINER_NO_DOCS.
1591 * Include UnicodeData.txt and the script to generate the unicode tables from
1596 * postingsource.rst: Clarify a couple of points (reported by "vHanda" on IRC).
1600 * Protect the ValueIterator::check() method against Mac OS X SDK headers
1601 which define a check() macro.
1603 * Fix warning from xlC compiler.
1605 * Avoid use of grep -e in configure, as /usr/bin/grep on Solaris doesn't
1608 * Fix check for flags which might be needed for ANSI mode for compilers called
1611 * configure: Improve handling of Sun's C++ compiler - trick libtool into not
1612 adding -library=Cstd, and prefer -library=stdcxx4 if supported. Explicitly
1613 add -library=Crun which seems to be required, even though the documentation
1616 Xapian-core 1.2.16 (2013-12-04):
1620 * PositionIterator,PostingIterator,TermIterator,ValueIterator: Don't segfault
1621 if skip_to() or check() is called on an iterator which is already at_end().
1622 Reported by David Bremner.
1624 * ValueCountMatchSpy: get_description() on a default-constructed
1625 ValueCountMatchSpy object no longer fails when xapian-core is built with
1628 * ValueCountMatchSpy: get_total() on a default-constructed ValueCountMatchSpy
1629 object now returns 0 rather than segfaulting.
1633 * If -v/--verbose is specified more than once to a test program, show the
1634 diagnostic output for passing tests as well as failing/skipped ones.
1636 * queryparsertest: Change qp_scale1 to time 5 repetitions of the large query to
1637 help average out variations.
1639 * queryparsertest: Add test coverage for explicit synonym of a term with a
1640 prefix (e.g. ~foo:search).
1642 * apitest: Remove code from registry* testcases which tries to test the
1643 consequences of throwing an exception from a destructor - it's complex to
1644 ensure we don't leak memory while doing this (it seems GCC doesn't release
1645 the object in this case, but clang does), and it's generally frowned upon,
1646 plus C++11 makes destructors noexcept by default.
1648 * Fix "make check" to actually removed cached databases first, as is
1653 * When moving a cursor on a read-only table, check if the block we want is in
1654 the internal cursor. We already do this for a writable table, as it is
1655 necessary for correctness, but it's a cheap check and may avoid asking the
1656 OS for a block we actually already have.
1658 * Correctly report the database as closed rather than 'Bad file descriptor'
1661 * Reuse a cursor for reading values from valuestreams rather than creating
1662 a new one each time. This can dramatically reduce the number of blocks
1663 redundantly reread when sorting by value. The rereads will generally get
1664 served from VM cache, but there's still an overhead to that.
1668 * When moving a cursor on a read-only table, check if the block we want is in
1669 the internal cursor. We already do this for a writable table, as it is
1670 necessary for correctness, but it's a cheap check and may avoid asking the
1671 OS for a block we actually already have.
1673 * Correctly report the database as closed rather than 'Bad file descriptor'
1676 * Reuse a cursor for reading values from valuestreams rather than creating
1677 a new one each time. This can dramatically reduce the number of blocks
1678 redundantly reread when sorting by value. The rereads will generally get
1679 served from VM cache, but there's still an overhead to that.
1683 * When moving a cursor on a read-only table, check if the block we want is in
1684 the internal cursor. We already do this for a writable table, as it is
1685 necessary for correctness, but it's a cheap check and may avoid asking the
1686 OS for a block we actually already have.
1688 * Correctly report the database as closed rather than 'Bad file descriptor'
1693 * Compress source tarballs with xz instead of gzip.
1695 * Split XAPIAN_LIBS out of XAPIAN_LDFLAGS so that -l flags for libraries
1696 configure detects are needed appear after -L flags specified by the user
1697 that may be needed to find such libraries. (fixes#626)
1699 * XO_LIB_XAPIAN now handles the user specifying a relative path in
1700 XAPIAN_CONFIG, e.g.: "./configure XAPIAN_CONFIG=../xapian-core/xapian-config"
1702 * Adjust XO_LIB_XAPIAN to strip _gitNNN suffix from snapshot versions.
1704 * configure: Handle git snapshot naming when calculating REVISION.
1706 * configure: Enable -fshow-column for GCC - things like vim's quickfix mode
1707 will then jump to the appropriate column for a compiler error or warning, not
1708 just the appropriate line.
1710 * configure: Report GCC version in configure output.
1714 * The API documentation shipped with the release is now generated with
1715 doxygen 1.8.5 instead of 1.5.9, which is most evident in the different
1716 HTML styling newer doxygen uses.
1718 * Document how Utf8Iterator handles invalid UTF-8 in API documentation.
1720 * Improve how descriptions of deprecated features appear in the API
1723 * docs/remote_protocol.rst: Correct error in documentation of REPLY_DOCDATA
1726 * docs/overview.rst: Correct documentation for how to specify "prog" remote
1727 databases in stub files.
1729 * Direct users to git in preference to SVN - we'll be switching entirely in
1734 * xapian-chert-update: Fix -b to work rather than always segfaulting (reported
1735 in http://bugs.debian.org/716484).
1737 * xapian-chert-update: The documented alias --blocksize for -b has never
1738 actually been supported, so just drop mentions of it from --help and the man
1743 + Fix chert database check that first docid in each doclength chunk is more
1744 than the last docid in the previous chunk - previously this didn't actually
1747 + Fix database check not to falsely report "position table: Junk after
1748 position data" whenever there are 7 unused bits (7 is OK, *more* than 7
1751 + Fix to report block numbers correctly for links within the B-tree.
1753 + If the METAINFO key is missing, only report it once per table.
1755 + Fix database consistency checking to always open all the tables at the same
1756 revision - not doing this could lead to false errors being reported after a
1757 commit interrupted by the process being killed or the machine crashing.
1758 Reported by Joey Hess in http://bugs.debian.org/724610
1762 * quest: Add --check-at-least option.
1766 * configure: clang doesn't support -Wstrict-null-sentinel or -Wlogical-op, so
1767 don't pass it these options.
1769 * Fix build errors and warnings with mingw.
1771 * Suppress "unused local typedef" warnings from GCC 4.8.
1773 * If the compiler supports C++11, use static_assert to implement
1776 * tests/zlib-vg.c: Fix two warnings when compiled with clang.
1778 * Fix failure when built with -D_GLIBCXX_DEBUG - we were modifying the top()
1779 element of a heap before calling pop(), such that the heap comparison
1780 operation (which is called when -D_GLIBCXX_DEBUG is on to verify the heap is
1781 valid) would read off the end of the data. In a normal build, this issue
1782 would likely never manifest.
1784 * configure: When generating ABI compatibility checks in xapian/version.h, pass
1785 $CXXFLAGS and $CPPFLAGS to $CXXCPP as they could contain options which affect
1786 the ABI (such as -fabi-version for GCC). (Fixes #622)
1788 * Microsoft GUIDs in binary form have reversed byte order in the first three
1789 components compared to standard UUIDs, so the same database would report a
1790 different UUID on Windows to on other platforms. We now swap the bytes to
1791 match the standard order. With this fix, the UUIDs of existing databases
1792 will appear to change on Windows (except in rare "palindronic" cases).
1794 * Fix a couple of issues to get Xapian to build and work on AIX.
1796 * common/safeuuid.h: Remove bogus take-address-of from uuid handling code for
1799 * Under cygwin, use cygwin_conv_path() if using a new enough cygwin version,
1800 rather than the now deprecated cygwin_conv_to_win32_path(). Reported by
1801 "Haroogan" on the xapian-devel mailing list.
1803 * common/safeuuid.h: Add missing '#include <cstdlib>' and qualify free with std.
1805 * Fix 'unused label' warning when chert backend is disabled.
1807 * xapian.h: Add check for Wt headers being included before us and defining
1808 'slots' as a macro - if they are, give a clear error advising how to work
1809 around this (previously compilation would fail with a confusing error).
1813 * Fix assertion failure for when an OrPostList decays to an AndPostList - the
1814 ordering of the subqueries by estimated termfreq may not be the same as it
1815 was when the OrPostList was constructed, as the subqueries may themselves
1816 have decayed. Reported by Michel Pelletier.
1818 * Fix -Wcast-qual warning from GCC 4.7 when configured with --enable-log.
1820 Xapian-core 1.2.15 (2013-04-16):
1824 * QueryParser/TermGenerator: Don't include CJK codepoints which are
1825 punctuation in N-grams.
1827 * TermGenerator: Fix bug where we failed to generate the first bigram
1828 from the second sequence of N-grammable CJK characters in a piece of text.
1832 * Call fdatasync()/fsync() when creating the "iambrass" file.
1836 * Call fdatasync()/fsync() when creating the "iamchert" file.
1840 * Call fdatasync()/fsync() when creating the "iamflint" file.
1844 * XO_LIB_XAPIAN now handles the user specifying XAPIAN_CONFIG without a path,
1845 for example: ./configure XAPIAN_CONFIG=xapian-config-1.3
1849 * delve: If -v is specified more than once, show even more info in some cases.
1853 * Fix warning due to needlessly casting away const-ness in debug logging.
1855 * Fix pointer truncation bug in lemon parser generator, which probably affects
1856 regenerating the query parser on WIN64.
1860 * Fix to build when configured with --enable-log.
1862 Xapian-core 1.2.14 (2013-03-14):
1866 * MSet::get_document(): Don't cache retrieved Document objects unless they
1867 were requested with fetch(). This avoids using a lot of memory when many
1868 MSet entries are retrieved. (Fixes #604)
1872 * apitest: Improved test coverage.
1876 * Check if a candidate document has at least the minimum weight needed
1877 before checking positional information, which speeds up slow phrase
1878 searches (partly addresses #394).
1882 * Fix multipass compaction not to damage document values, and to merge the
1883 database stats correctly. (fixes #615)
1887 * Fix multipass compaction not to damage document values, and to merge the
1888 database stats correctly. (fixes #615)
1892 * Fix multipass compaction bug. (fixes #615)
1898 + Fix handling of delays between replication events - the subtraction of the
1899 target time and the current time was reversed, so we wouldn't sleep when
1900 before the deadline, but would sleep after it for the amount we'd missed it
1903 + On Microsoft Windows, we no longer sleep for more than 43 years if the
1904 target time for a replication event had already passed. (Fixes #472)
1908 * matcher/queryoptimiser.cc: Need <functional> for mem_fun().
1910 * tests/harness/testsuite.cc: Don't provide explicit template types to
1911 make_pair - it isn't useful, and breaks with C++11. Fixes build error with
1914 * examples/quest.cc: Fix to build with Sun Studio 12 compiler. (ticket#611)
1916 Xapian-core 1.2.13 (2013-01-09):
1920 * TermGenerator: Add new method TermGenerator::set_max_word_length() to allow
1921 this limit to be adjusted by the user.
1923 * QueryParser: Implicitly close any unclosed brackets at the end of the query
1924 string. Patch from Sehaj Singh Kalra.
1926 * DateValueRangeProcessor: Add extra constructor overloaded form so that in
1927 DateValueRangeProcessor(1, "date:"), the const char * gets interpreted as
1928 std::string rather than bool.
1932 * apitest: Assorted test coverage improvements.
1934 * When reporting valgrind errors, skip any warnings before the error in the
1939 * Improved fix for #590 - count all matching LeafPostList objects with a Weight
1940 object rather than trying to prune at the MultiAndPostList level based on
1941 max_wt (if wdf is always zero for a term, BM25 gives max_wt of 0, which lead
1942 to us never counting that subquery.
1944 * Fix calculation of 0.0/0.0 in some cases. This then got used as a minimum
1945 weight, but it seems this gives -nan (at least on x86-64 Linux) so it may
1946 have been harmless in practice.
1948 * We no longer use the highest weighted MSet entry to calculate percentages, so
1949 remove code which finds it.
1953 * Close excess file handles before we get the fcntl lock, which avoids the
1954 lock being released again if one is open on the lock file. Notably this
1955 avoids a situation where multiple threads in the same process could succeed
1956 in locking a database concurrently.
1960 * Close excess file handles before we get the fcntl lock, which avoids the
1961 lock being released again if one is open on the lock file. Notably this
1962 avoids a situation where multiple threads in the same process could succeed
1963 in locking a database concurrently.
1967 * Close excess file handles before we get the fcntl lock, which avoids the
1968 lock being released again if one is open on the lock file. Notably this
1969 avoids a situation where multiple threads in the same process could succeed
1970 in locking a database concurrently.
1974 * Improve the UnimplementedError message for a MatchSpy subclass which doesn't
1975 implement name() so it's clearer that it is this particular subclass which
1976 can't be used remotely, rather than all MatchSpy objects.
1980 * The build system is now generated with automake 1.11.6 rather than 1.11.1,
1981 which fixes a security issue in "make distcheck" (not something users will
1982 usually run, but it seems worth addressing).
1984 * Use user-specified LIBS for configure tests, which is what you'd expect to
1985 happen, and provides a way for the user to tell configure where to find
1986 library functions which configure can't find for itself.
1988 * INCLUDES is now deprecated in automake, so use AM_CPPFLAGS instead.
1990 * Test coverage rules now assume lcov 1.10 which allows them to be simpler
1991 and not to require a patched version of lcov.
1995 * valueranges.html: Update documentation to reflect change in Xapian 1.1.2 -
1996 DateValueRangeProcessor and StringValueRangeProcessor now support a prefix or
1999 * Clarify that the "reverse" parameter of set_sort_by_relevance_then_value()
2000 and set_sort_by_relevance_then_key() only affects the ordering of the
2001 value/key part of the sort.
2003 * docs/quickstart.html: Fix seriously outdated statement that Xapian doesn't
2004 create the database directory - that changed in 0.7.2 (released 2003-07-11).
2006 * HACKING: Try to make it clearer we're looking for a dual-licence on submitted
2013 + Add a --full-copy option to force a full copy to be sent. (ticket#436)
2015 + Add --quiet option, and be a little more verbose by default.
2017 + Allow files > 32G to be be copied by replication.
2019 + Fix "if (fd > 0)" tests in some replication code to be "if (fd >= 0)".
2020 In practice this is unlikely to actually have caused problems since
2021 stdin is typically still open and using fd 0.
2023 + Simplify how we open the .DB file on the replication slave to just call
2024 open() once with O_CREAT, rather than once without, than stat() if that
2025 fails, and then again with O_CREAT|O_TRUNC if stat() doesn't show an
2026 ordinary file exists.
2032 + New --flags command line option to allow setting arbitrary QueryParser
2035 + Align option descriptions in --help output, and make the initial letter of
2036 such descriptions consistently lowercase.
2040 * Fix testsuite harness to compile with GCC 4.7.
2042 * On platforms with the F_MAXFD fcntl but without closefrom(), we were failing
2043 to close the highest numbered open fd in our closefrom() replacement.
2045 * Our closefrom() replacement on Linux now works around valgrind not hiding
2046 some extra fds it has open, but then complaining if we try to close them.
2048 + Pass O_BINARY when opening replication related files in some cases where we
2049 weren't before, which will probably help solve ticket #472.
2051 * configure: socketpair() needs -lnetwork on Haiku.
2053 * Micro-optimisation in Unicode handling - GCC doesn't currently optimise the
2054 arithmetic shift right idiom we use, but it documents that signed right shift
2055 does sign extension so we now just use a right shift for GCC.
2059 * Preserve errno over debug logging calls, so they can safely be added to code
2060 which expects errno not to change.
2062 Xapian-core 1.2.12 (2012-06-27):
2066 * 1.2.11 had its library version information incorrectly set. This resulted in
2067 the shared library having an incorrect SONAME - e.g. on Linux,
2068 libxapian.so.21 instead of libxapian.so.22. This release has been made to
2073 * AUTHORS: Add the GSoC students.
2075 Xapian-core 1.2.11 (2012-06-26):
2079 * Add new QueryParser::STEM_ALL_Z stemming strategy, which stems all terms and
2080 adds a Z prefix. (Patch from Sehaj Singh Kalra, fixes ticket#562)
2082 * Add TermGenerator::set_stemming_strategy() method, with strategies which
2083 correspond to those of QueryParser. Based on patch from Sehaj Singh Kalra,
2084 with some tweaks for adding term positions in more cases. (Fixes ticket#563)
2086 * Correct "BM25Weight" to "TradWeight" in exception message from TradWeight.
2088 * We were failing to call init() for user-defined Weight objects providing the
2089 term-independent weight. These now get called with init(0.0).
2091 * Xapian::Auto::open_stub() now throws a Xapian::DatabaseOpeningError exception
2092 if the stub file can't be opened. Previously we failed to check for this
2093 condition, which resulted in us treating the file as empty.
2097 * When the testsuite is using valgrind, we used to run remote servers under
2098 valgrind too (but with --tool=none) to get consistent behaviour as valgrind's
2099 emulation of x87 excess precision isn't exact. Now we only do this if x87 FP
2100 instructions are actually in use (which means x86 architecture and configure
2101 run with --disable-sse).
2103 * Make sure XAPIAN_MAX_CHANGESETS gets unset after replication testcases which
2104 set it, so further testcases don't waste time generating changesets.
2106 * Improved test coverage (including more tests for closed databases -
2111 * After closing the database, methods which try to use the termlist would throw
2112 FeatureUnavailableError with message "Database has no termlist", assuming
2113 that the termlist table not being open meant it wasn't present. Fix to check
2114 if the postlist_table is open to determine which case we're in.
2118 * After closing the database, methods which try to use the termlist would throw
2119 FeatureUnavailableError with message "Database has no termlist", assuming
2120 that the termlist table not being open meant it wasn't present. Fix to check
2121 if the postlist_table is open to determine which case we're in.
2125 * Check if the database is closed in metadata_keys_begin() for InMemory
2130 * xapian-config: Don't interpret a missing .la file as meaning that we only
2131 have static libraries.
2135 * Fix API documentation for Query constructors - both XOR and ELITE_SET can
2136 take any number of subqueries, not only exactly two.
2138 * Backport missing API documentation comments for operator++ and operator*
2139 methods or PositionIterator, PostingIterator and TermGenerator.
2141 * docs/replication.rst: Update documentation - since 1.2.5, the value of
2142 XAPIAN_MAX_CHANGESETS determines how many changesets we keep.
2144 * docs/admin_notes.rst: Correction - we don't "create a lock file", we "lock a
2147 * Fix API documentation for TradWeight constructor - "k1" should be "k".
2151 * configure: Overhaul handling of compilers which pretend to be GCC. Clang
2152 is now detected, and we only pass it warning flags it actually understands.
2153 And we now check for symbol visibility support with Intel's compiler.
2155 * configure: Solaris automatically pulls in library dependencies, so set
2156 link_all_deplibs_CXX=no there.
2158 * configure: We now check -Bsymbolic-functions for all compilers.
2160 * configure: Enable -Wdouble-promotion for GCC >= 4.6.
2162 * Pass -ldl last when compiling zlib-vg.so, as that seems to be needed on
2165 * Fix incorrect use of "delete" which should be "delete []". This is
2166 undefined behaviour in C++, though the type is POD, so in practice this
2167 probably worked OK on many platforms.
2169 * In BM25Weight when k1 or b is zero (not the default), we used to multiply
2170 an uninitialised double by zero, which is undefined behaviour, but in
2171 practice will often give zero, leading to the desired results.
2173 * xapian.h: Add check for Qt headers being included before us and defining
2174 'slots' as a macro - if they are, give a clear error advising how to work
2175 around this (previously compilation would fail with a confusing error).
2177 Xapian-core 1.2.10 (2012-05-09):
2183 * apitest: Extend tradweight1 to test that TradWeight(0) means that wdf and
2184 document length don't affect the weight of a term.
2186 * termgentest: Check that TermGenerator discards words > 64 bytes.
2190 * Don't count unweighted subqueries of MultiAndPostList in percentage
2191 calculations, as OP_FILTER maps to MultiAndPostList now. (ticket#590)
2195 * When compacting, if the output database is empty, don't write out a metainfo
2196 tag. Take care not to divide by zero when computing the percentage size
2201 * When compacting, if the output database is empty, don't write out a metainfo
2202 tag. Take care not to divide by zero when computing the percentage size
2207 * API documentation:
2209 + Note version when Database::close() was added.
2211 + Fix switched lower and upper in API documentation for Weight methods
2212 get_doclength_lower_bound() and get_doclength_upper_bound(). Correct
2213 maximum to minimum in get_doclength_lower_bound() comment and note that this
2214 excludes zero length documents. Fix "An lower" to "A lower".
2216 * docs/admin_notes.html: Mention that postlist and termlist tables also hold
2217 value info for chert. Mention that xapian-chert-update was removed in 1.3.0.
2218 Mention that you need to use copydatabase from 1.2.x to convert flint to
2221 * HACKING: Update section on patches to mention git (git diff and git
2222 format-patch), and using "-r" with normal diff, and also that ptardiff offers
2223 a nice way to diff against an unpacked tarball.
2227 * Fix use of AssertEq() on NULL, which doesn't compile, at least with recent
2230 Xapian-core 1.2.9 (2012-03-08):
2234 * QueryParser: Fix FLAG_AUTO_SYNONYMS not to enable auto multi-word synonyms
2235 too (but in a different way to trunk so as to not break the ABI).
2239 * Fix issue with running AND, OR and XOR queries against a database with no
2240 documents in it - this was leading to a divide by zero, which led to
2241 MSet::get_matches_estimated() reporting 2147483648 on i386.
2245 * Remove configure's --with-stlport and --with-stlport-compiler options, as
2246 they don't allow you to actually specify what you need to (at least to use
2247 the Debian STLport package), and instead document what to pass to configure
2248 to enable building with STLport (though it seems to no longer be actively
2249 maintained, and the debug mode (which is probably the most interesting
2250 feature now) doesn't seem to work on Debian stable).
2254 * Document that OP_ELITE_SET with non-term subqueries might pick subqueries
2255 which don't match anything. Closes ticket#49.
2257 * Document that you can define a static operator delete method in your subclass
2258 if deallocation needs to be handled specially. (Closes ticket#554)
2260 * Assorted minor documentation improvements.
2264 * Address new warnings from GCC 4.6.
2266 * Fix argument order when linking xapian-check to fix mingw build.
2269 * Add some missing explicit header includes to fix build with STLport.
2271 Xapian-core 1.2.8 (2011-12-13):
2275 * Add support to TermGenerator and QueryParser for indexing and searching CJK
2276 text using n-grams. Currently this is only enabled when the environmental
2277 variable XAPIAN_CJK_NGRAM is set to a non-empty value.
2281 * Add link from index page to apidoc.pdf.
2283 * quickstart.html: Correct link which was to quickstartsearch.cc.html but
2284 should be to quickstartindex.cc.html.
2286 * overview.html,quickstart.html: Fix several factual errors.
2288 * API documentation:
2290 + Improve documentation comments for several methods.
2292 + Add documentation for function parameters which didn't have it.
2294 + Remove bogus paragraph in WritableDatabase::replace_document()
2295 documentation comment which had been cut and pasted from delete_document()
2296 documentation comment. (Fixes ticket#579)
2298 + Explicitly document which value slot numbers are valid. (Fixes ticket#555)
2300 + Escape < and > in doxygen comments so "<foo>" doesn't get eaten by doxygen.
2304 + Some fixes for warnings when cross-compiling to mingw.
2306 * tests/soaktest/soaktest.cc: With Sun's compiler, random() and srandom()
2307 aren't in <cstdlib> so we need to use <stdlib.h> instead.
2309 Xapian-core 1.2.7 (2011-08-10):
2313 * Document objects now track whether any document positions have been modified
2314 so that replacing a modified document can completely skip considering
2315 updating positions if none have changed. Currently the flint, chert, and
2316 brass backends implement this optimisation. A common case this speeds up is
2317 adding and/or removing boolean filter terms to/from existing documents - for
2318 example this gives an 18% speedup for adding tags in notmuch.
2322 * Make sure that perftest isn't run with libeatmydata preloaded, as making
2323 fsync() a no-op makes performance tests rather bogus.
2327 * Remove unnecessary call to reopen() in the remote servers in a case where
2328 either we had just called it or we are using a writable database and so
2329 reopen() doesn't do anything.
2333 * configure: -Wshadow gives bogus warnings with 4.0 (at least on Mac OS X), so
2334 disable it for GCC < 4.1 (like the comments already said we did!)
2338 * Improve the documentation comment for Database::close(). (ticket#504)
2340 * Fix typo in documentation comment for Enquire constructor which reversed the
2341 intended sense (though the text was fairly obviously wrong before).
2343 * Improve documentation of QueryParser::add_boolean_prefix()'s exclusive
2344 parameter to talk about terms and prefixes rather than values and fields
2345 (which was confusing since "document value" has a particular meaning in
2348 * docs/facets.html: Expand descriptions for indexing and finding facets.
2349 Fix errors in example code.
2351 * docs/index.html: Add links to Omega and bindings documentation.
2353 * docs/remote_protocol.html: Fixed typo which reversed the intended sense.
2355 * xapian-check --help: Document that checking a whole database performs
2356 additional cross-checks between the tables.
2358 * docs/admin_notes.html: Add note about xapian-chert-update.
2360 * docs/deprecation.html: Note here that WritableDatabase::flush() is
2361 deprecated in favour of WritableDatabase::commit().
2365 * Fix -Wshadow warnings from GCC 4.6.
2367 * Fix warning from GCC 3.3.
2371 * Fix some problems with the templates used to implement output of parameters
2372 and return values in debug logging.
2374 Xapian-core 1.2.6 (2011-06-12):
2380 + Add new set_max_wildcard_expansion() method to allow limiting the number of
2381 terms a wildcard can expand to. (ticket#350)
2383 + If default_op is OP_NEAR or OP_PHRASE then disable stemming of the terms,
2384 since we don't index positional information for stemmed terms by default.
2386 * Spelling correction was failing to correctly handle words which had the same
2387 trigram in an even number of times.
2391 * We now actually include the soaktest code in the release tarballs.
2395 * Eliminate some vector copies when handling phrase subqueries in the query
2400 * Kill the child process which holds the lock with SIGKILL as that can't be
2401 ignored, whereas SIGHUP can be in some cases.
2405 * Kill the child process which holds the lock with SIGKILL as that can't be
2406 ignored, whereas SIGHUP can be in some cases.
2410 * Kill the child process which holds the lock with SIGKILL as that can't be
2411 ignored, whereas SIGHUP can be in some cases.
2415 * The HTML documentation is now maintained in reStructured Text format.
2417 * docs/queryparser.html: Document the precedence order of operators.
2419 * docs/scalability.html: Bring up-to-date.
2421 * docs/overview.html: Document "remote" in stub databases.
2423 * docs/postingsource.html: Add PostingSource example. (ticket#503)
2425 * include/xapian/database.h: Add @exception InvalidArgumentError for
2426 Database::get_document() (ticket#542).
2428 * Ship ChangeLog.0 in the tarball.
2430 * Assorted minor improvements.
2434 * examples/delve: Report has_positions().
2436 * examples/simpleindex: Add short description to usage message.
2440 * Fix to build for mingw.
2442 Xapian-core 1.2.5 (2011-04-04):
2446 * Enquire::get_eset() now accepts a min_wt argument to allow the minimum wanted
2447 weight to be specified. Default is 0, which gives the previous behaviour.
2449 * QueryParser: Handle NEAR/<offset> and ADJ/<offset> where offset isn't an
2450 integer the same way at the end of the query as in the middle.
2454 + Only keep $XAPIAN_MAX_CHANGESETS changeset files when generating a new one
2455 (previously this variable only controlled if we generated changesets or
2456 not). Closes ticket#278.
2458 + $XAPIAN_MAX_CHANGESETS is reread each time, rather than only when the
2461 + If you build Xapian with DANGEROUS mode enabled, changeset files now
2462 actually have the appropriate flag set (the reader will currently throw an
2463 exception, but that's better than quietly handling them incorrectly).
2467 * Compaction tests which generate stub files now close them before performing
2468 the actual compaction, to avoid issues on Microsoft Windows (ticket#525).
2470 * Improve test coverage.
2474 * Fix memory leak if an exception is thrown during the match.
2478 * Bumped format version number (we now store the oldest revision for which we
2479 might have a replication changeset).
2481 * Optimise not to read the bitmaps from the base files when opening a database
2482 for reading (cross-port of equivalent change to chert).
2484 * Optimise not to update doclength when it hasn't changed (cross-port of
2485 equivalent change to chert).
2487 * If we try to delete an old base file and it isn't there, just continue rather
2488 than throwing an exception. We wanted to get rid of it anyway, and it may be
2489 NFS issues telling us the wrong thing. In particular, DatabaseCorruptError
2490 was rather a pessimistic assessment.
2494 * Optimise not to read the bitmaps from the base files when opening a database
2497 * Optimise not to update doclength when it hasn't changed.
2499 * xapian-chert-update: Fix to handle larger databases, and databases which
2502 * If we try to delete an old base file and it isn't there, just continue rather
2503 than throwing an exception. We wanted to get rid of it anyway, and it may be
2504 NFS issues telling us the wrong thing. In particular, DatabaseCorruptError
2505 was rather a pessimistic assessment.
2509 * Optimise not to read the bitmaps from the base files when opening a database
2510 for reading (cross-port of equivalent change to chert).
2512 * Optimise not to update doclength when it hasn't changed (cross-port of
2513 equivalent change to chert).
2515 * If we try to delete an old base file and it isn't there, just continue rather
2516 than throwing an exception. We wanted to get rid of it anyway, and it may be
2517 NFS issues telling us the wrong thing. In particular, DatabaseCorruptError
2518 was rather a pessimistic assessment.
2522 * xapian-tcpsrv: If we can't bind to the specified port because it is a
2523 privileged one, exit with code 77 (EX_NOPERM) to make it easier to
2524 automatically handle failure when starting the server from a script.
2528 * Snapshots and releases are now bootstrapped with autoconf 2.68 and libtool
2531 * configure: -Wstrict-null-sentinel was added in GCC 4.0.1 and so doesn't work
2532 with GCC 4.0.0. For simplicity, only enable it for GCC >= 4.1.
2536 * INSTALL: Note how to build for a non-default arch on a multi-arch platform.
2538 * include/xapian/enquire.h: Fix doxygen markup so alternative overloaded forms
2539 of Enquire::get_mset() appear in the API documentation.
2541 * collapsing.html: Add missing document (written some time ago, but never
2542 actually added to builds).
2544 * replication.html: Update documentation to make it clear that users shouldn't
2545 create the destination directory for replication themselves.
2547 * docs/intro_ir.html: Update link to a paper. Update text about book "to be
2550 * docs/deprecation.html:
2552 + PostingSource now offers a replacement for Enquire::set_bias().
2554 + OmegaScript: $set{spelling,true} is now deprecated.
2556 + Add note about botched removal of Enquire.get_matching_terms from Python
2557 bindings (now fully removed).
2559 + Note removal of "if idx in mset" from Python bindings.
2561 + Deprecate MSet.items and ESet.items from Python bindings (ticket#531).
2563 * docs/admin_notes.html: Update for 1.2.5.
2565 * Updates to documentation of internals.
2569 * xapian-replicate-server: Fix race condition between checking if a file
2570 exists and opening it to replicate it.
2572 * xapian-replicate: Complain unless host name and port number are specified -
2573 previously these defaulted to an empty string and 0, which resulted in
2574 potentially confusing error messages.
2576 * xapian-replicate: If --master isn't specified, default to DATABASE.
2580 * quest: Report any spelling correction (requires the database contains
2581 spelling data of course).
2583 * copydatabase: Add --no-renumber option.
2587 * api/compactor.cc: Add missing header <ctime> for time() (ticket#530).
2589 * api/compactor.cc: Use msvc_posix_rename() under __WIN32__ to atomically
2590 update stub file after compaction (ticket#525).
2592 * Fix uninitialised variable warnings with gcc -O3.
2594 * Eliminate std::string member of global static object used when compiled with
2595 --enable-log which was causes problems on Mac OS X.
2597 * Fix some issues highlighted by clang++ warnings.
2599 Xapian-core 1.2.4 (2010-12-19):
2605 + Avoid a double free if Query construction throws an exception in a
2606 particular case. Fixes ticket#515.
2608 + Allow phrase generators between a probabilistic prefix and the term itself
2609 (e.g. path:/usr/local).
2611 + The correct window size wasn't being set in some cases when default_op was
2614 * Enquire::get_mset():
2616 + Avoid pointlessly trying to allocate lots of memory if the first document
2617 requested is larger than the size of the database.
2619 + An empty query now returns an MSet with firstitem set correctly -
2620 previously firstitem was always 0 in this case.
2622 * Document: Initialise docid to 0 when creating a document from
2623 scratch, as documented.
2627 + Move the database compaction and merging functionality into this new class,
2628 and make xapian-compact a simple wrapper around this class. (ticket#175)
2630 + Inputs can now be stub database directories or files, in which case the
2631 databases in the stub are used as inputs.
2633 + Add support for compacting to a stub database, which can be one of the
2634 inputs (for atomic update).
2636 + If spellings and/or synonyms were only present in some source databases,
2637 they weren't copied to the output database, but now they are.
2641 * Improve test coverage (particularly for Xapian::Utf8Iterator and
2644 * Add zlib-vg.c to distribution tarballs.
2646 * tests/runtest: Add XAPIAN_TESTSUITE_LD_PRELOAD hook to allow libeatmydata to
2647 easily be used to speed up testsuite runs.
2651 * The matcher wasn't recalculating the max possible weight after a subquery of
2652 XOR reached its end. This caused an assertion failure in debug builds, and
2653 is a missed optimisation opportunity.
2655 * Implement SelectPostList::check() so that check() on OP_NEAR and OP_PHRASE
2656 subqueries will just check a single document, not a potentially huge numbers
2659 * BM25Weight: Fix calculation order to avoid inconsistent weights due to
2660 rounding when certain non-default parameter combinations are used.
2662 * TradWeight: Fix calculation order to avoid inconsistent weights due to
2663 rounding with TradWeight(0).
2665 * Fix regression in speed of OP_OR queries in certain cases due to optimisation
2666 added in 1.0.21/1.2.1.
2668 * In the query optimiser, use value range bounds to detect value ranges which
2673 * Add support for iterating metadata keys with the remote backend. This change
2674 necessitated an increase in the minor version of the remote protocol. If you
2675 are upgrading a live system which uses the remote backend, upgrade the
2676 servers before the clients.
2680 * xapian-config: Add --static option which makes other options report values
2683 * xapian-config is now removed by "make distclean" not "make clean".
2685 * configure: FreeBSD and OpenBSD don't need explicit dependency libraries, so
2686 set link_all_deplibs_CXX=no there.
2688 * This release uses autoconf 2.67 rather than 2.65.
2692 * INSTALL: Raise recommended GCC version from 3.3 to 4.1, since that's the
2693 oldest we regularly test with.
2695 * replication.html: Update and improve in various ways.
2697 * Remove lingering "experimental" marker from PostingSource and
2698 ValueCountMatchSpy API documentation.
2700 * index.html: Add links to replication and facets documents, and fix typo in
2701 serialisation document link.
2703 * internals.html: Add link to replication protocol.
2705 * Change the categorisation document to talk about facets, since that's the
2706 terminology that seems to be most widely used these days, and
2707 "categorisation" can also mean automatically assigning categories to
2708 documents. Also update to reflect the final API.
2710 * deprecation.html: Add guidelines for supporting other software.
2712 * Document cases where QueryParser's FLAG_WILDCARD and FLAG_PARTIAL aren't
2713 currently supported.
2715 * PLATFORMS: Move PLATFORMS information to the wiki and replace with a pointer.
2719 * xapian-compact: Fix access to empty priority_queue while merging synonyms.
2720 This could have caused problems, though we've had no reports of any (the
2721 bug was found with _GLIBCXX_DEBUG).
2723 * xapian-compact: Add --quiet/-q option to suppress progress output.
2726 * xapian-replicate: If a full copy was attempted, but was not put live, display
2727 an explanatory message (in verbose mode).
2731 * examples/quest: Add command line options to allow prefixes to be specified
2732 for the QueryParser.
2734 * examples/delve: Add '-z' option to count zero-length documents.
2736 * examples/simplesearch: Fix cut-and-paste errors in usage message and
2741 * configure: Add support for --enable-sse=sse and --enable-sse=sse2 to allow
2742 control of which SSE instructions to use.
2744 * configure: Enable use of SSE maths on x86 by default with Sun's compiler.
2746 * configure: Beef up the test for whether -lm is required and add a special
2747 case to force it to be for Sun's C++ compiler - there's some interaction with
2748 libtool and/or shared objects which means that the previous configure test
2749 didn't think -lm is needed here when it is.
2751 * Fix to build on OpenBSD 4.5 with GCC 3.3.5.
2753 * Need to avoid excess precision on m68k when targeting models 68010, 68020,
2754 68030 as well as 68000.
2756 * Fix compilation with Sun's C++ compiler.
2758 * Fix testsuite to build on Solaris < 10.
2760 Xapian-core 1.2.3 (2010-08-24):
2764 * Database::get_spelling_suggestion() will now suggest a correction even if the
2765 passed word is in the dictionary, provided the correction has at least the
2766 same frequency. Partly addresses #225.
2770 + Fix handling of groups of terms which are all stopwords - in situations
2771 where this causes a problem we now disable stopword checks for such groups.
2774 + Fix to be smarter about handling a boolean filter term containing ".." in
2775 the presence of valuerangeprocessors.
2779 * New "unittest" program for testing low level functions directly. Currently
2780 this has tests for the internal resolve_relative_path() function.
2785 * Retry select() if it fails with EINTR while waiting for connect(), and
2786 discriminate cases with same failure message to aid debugging.
2790 * Fix documentation comment for Xapian::timeout type - it holds a time interval
2791 in milliseconds not microseconds (the API docs for the methods which use it
2792 explicitly correctly document that the timeouts are in milliseconds).
2794 * libuuid moved from e2fsprogs to util-linux-ng about a year ago, so update
2795 documentation, comments, and configure error messages to reflect this.
2799 * configure: Don't pass -mtune=generic unless GCC >= 4.2 is in use
2802 * Fix handling of some obscure cases of resolving relative paths on Microsoft
2803 Windows. (ticket#243).
2805 * Optimise closing of all unwanted file descriptors after forking by using
2806 closefrom() if available, and otherwise providing our own implementation
2807 (optimised to some extent for many platforms).
2809 * Fix test harness to build under Microsoft Windows (ticket#495).
2813 * xapian-core.spec: Add xapian-metadata and cmake related files to RPM
2816 * xapian-core.spec: Update BuildRequires to specify libuuid-devel instead of
2821 * Improve logging of function parameter placeholder strings.
2823 Xapian-core 1.2.2 (2010-06-27):
2827 * Sync changes from each Btree table to disk right after syncing changes to
2828 its base file, which allows more time for the table changes to be written
2829 and may also be more efficient with some Linux kernel versions.
2833 * Sync changes from each Btree table to disk right after syncing changes to
2834 its base file, which allows more time for the table changes to be written
2835 and may also be more efficient with some Linux kernel versions.
2839 * xapian-check: Don't try to check document lengths are consistent between the
2840 postlist and termlist tables if it would use more than 1GB of memory, and
2841 handle std::bad_alloc or std::length_error when trying to allocate space
2842 for this. This issue affected sup users, as sup allocates docids such that
2843 they are sparse and large docids can easily occur.
2847 * delve: Show the database's UUID.
2851 * Revert 1.2.1 change to visibility of Xapian::Weight's copy constructor as
2852 it making it private broke compilation with GCC 4.1 (which seems to be a
2853 bug in this compiler version).
2855 * tests/harness/testsuite.cc: Need <cstdio> for sprintf(). Fixes compilation
2856 error which was masked if valgrind was installed. (ticket#489)
2860 * xapian-core.spec: Update for 1.2.x - add e2fsprogs-devel to BuildRequires and
2861 add new files to install.
2863 Xapian-core 1.2.1 (2010-06-22):
2865 This release includes all changes from 1.0.21 which are relevant.
2869 * QueryParser: Add support for open-ended ranges (ticket#480).
2871 * Add new optional parameter to QueryParser::add_boolean_prefix() to allow the
2872 user to indicate a prefix isn't "exclusive" and that multiple instances
2873 should be combined with OP_AND rather than OP_OR. Fixes ticket#402. This
2874 change should also improve efficiency as it avoids copying the lists of
2875 prefixes and compares them more efficiently.
2877 * You can now specify a custom stemming algorithm by subclassing
2878 Xapian::StemImplementation, mostly based on patch from Evgeny Sizikov in
2881 * Fix replication bug: when multiple commits were made to the master database
2882 while a client was performing a full copy, the client would only apply the
2883 first changeset and then try to make the database live, but fail due to
2884 trying to set the wrong revision number.
2886 * Replication no longer sleeps between applying changesets to an offline
2887 database. It's only necessary to sleep for a live database (to allow readers
2888 to complete a search without getting DatabaseModifiedErrror.
2890 * xapian-replicate: Add new "-r" command line option to specify how long
2891 replication sleeps for between applying changesets to a live database.
2893 * If a Btree table doesn't exist when applying a replication changeset, create
2894 it. This fixes replicating a revision where a lazy table is created.
2899 * zlib can produce "uninitialised" output from "initialised" input - the
2900 output does decode to the input, so this is presumably just some unused bits
2901 in the output, so we use an LD_PRELOAD hack to get valgrind to check the
2902 input is initialised and then tell it that the output is initialised.
2904 * Don't pass NULL to closedir(), which fixes test harness failures on platforms
2905 without /proc/self/fd.
2907 * Use safesyswait.h, fixing build failure on "make check" on FreeBSD.
2909 * Check is SA_SIGINFO is defined before using it as it isn't available
2910 everywhere. Fixes testsuite build failure on GNU Hurd.
2912 * Add a "soaktest" testsuite, intended to contain long-running tests with
2913 random data. Currently contains a single test which builds and runs random
2914 queries, checking that the results returned are consistent when asking for
2915 different result ranges.
2917 * Test UUID returned by Database::get_uuid() is 36 characters long.
2921 * Xapian no longer forces the wdf_max value to be at least one in
2922 BM25Weight::get_maxpart(). We used to do this so that a non-existent term in
2923 the query would cause it not to achieve 100%, but now we calculate
2924 percentages based on the number of matching subqueries, and it is more
2925 natural for a non-existent term to get zero weight (ditto for a term which
2928 * OP_VALUE_RANGE and OP_VALUE_GE now use value streams directly which is much
2929 more efficient for chert (the default backend in 2.2.x). As an example, a
2930 range query testcase which previously took 29 seconds now takes 0.4 seconds
2931 (70 times faster). (ticket#432)
2933 * The term statistics from multiple databases are now gathered in a simpler
2934 way which is a bit faster and uses less memory.
2938 * Install headers under PREFIX/include not PREFIX/include/xapian. If you used
2939 XO_LIB_XAPIAN or xapian-config in your build system, the headers would still
2942 * Releases and snapshots are now generated with libtool 2.2.10 instead of
2945 * Fix build failures with some combinations of backends disabled (partially
2946 addresses ticket#361 - some combinations still fail).
2948 * Add check to configure that GCC actually supports visibility for the platform
2949 being built for, which fixes compiler warnings with platforms which don't
2950 (such as Mac OS X and mingw).
2954 * Update documentation - replication and PostingSource aren't experimental in
2959 * Make use of built-in UUID API on FreeBSD and NetBSD. (ticket#470)
2965 * Add new pretty printer for values reported by calls and returns in debug
2966 logging - in particular, strings are now reported with non-printable
2969 * Debug logging should have less runtime overhead when built in but not in use.
2971 * Drop support for --enable-log=profile - dedicated profiling tools are likely
2972 to return more useful results.
2974 Xapian-core 1.2.0 (2010-04-28):
2976 This release includes all changes from 1.0.20 which are relevant.
2980 * Fix --abort-on-error to actually work.
2982 * Exit with status 1 not 0 if we caught an exception from the harness itself.
2984 Xapian-core 1.1.5 (2010-04-16):
2986 This release includes all changes from 1.0.19 which are relevant.
2990 * Database replication now handles an exception while applying a changeset
2993 * If environment variable XAPIAN_MAX_CHANGESETS is set on a replication client
2994 then any changesets read are saved so the replicated copy can itself be
2999 * Use sigsetjmp() and siglongjmp() where available so that the set of blocked
3000 signals get restored and the test harness can catch a second incidence of a
3001 particular signal in a run. Use sigaction() instead of signal() where
3002 available, which allows us to report the address associated with SIGSEGV,
3003 SIGFPE, SIGILL, and SIGBUS.
3005 * Add machinery to check for leaked file descriptors. Currently this requires
3006 /proc/self/fd to work (which is present on Linux and some other platforms).
3007 Remove the crude ulimit in runtest which has caused problems on some Debian
3010 * The test harness now explicitly catches const char * exceptions and reports
3015 * Ensure that the wdf upper bound is correctly updated when replacing
3018 * xapian-compact: Now sets lastdocid correctly when using --no-renumber.
3022 * Ensure that the wdf upper bound is correctly updated when replacing
3025 * xapian-compact: Now sets lastdocid correctly when using --no-renumber.
3027 * xapian-check: Check that the initial doclen chunk exists.
3031 * xapian-compact: Now sets lastdocid correctly when using --no-renumber.
3035 * Add remote backend support for WritableDatabase::add_spelling() and
3036 WritableDatabase::remove_spelling(). This bumps the remote protocol to
3037 version 35.0 (so both client and servers will need updating). Suggesting
3038 spelling corrections isn't yet supported. (ticket#178)
3042 * XO_LIB_XAPIAN: Give a more specific error message for the cases where
3043 XAPIAN_CONFIG isn't found, is a directory, or isn't executable.
3049 + If any documents are specified with "-d<docid>", "-V<slot>" now only show
3050 values for those documents.
3052 + Remove undocumented -k option, which has been a compatibility alias for -V
3053 since 0.9.10. Just use -V instead.
3055 * xapian-metadata: Add new example program which allows you to get and set
3056 individual user metadata entries.
3058 Xapian-core 1.1.4 (2010-02-15):
3060 This release includes all changes from 1.0.18 which are relevant.
3064 * Xapian::TermGenerator,Xapian::QueryParser,Xapian::Unicode::is_wordchar():
3065 Add ENCLOSING_MARK and COMBINING_SPACING_MARK categories to is_wordchar(),
3066 which is used by TermGenerator and QueryParser. Also make TermGenerator and
3067 QueryParser ignore several zero-width space characters. This is a better
3068 but less compatible version of a fix in 1.0.18.
3070 * Implement support for iterating valuestreams for multidatabases.
3072 * Xapian::Stem: Update the german and german2 stemming algorithms to the latest
3073 versions from Snowball. These add an extra rule for the "-nisse" ending.
3075 * Xapian::ValueCountMatchSpy: Replace get_values() with values_begin() and
3078 * Xapian::MatchSpy: Provide an iterator for accessing the top values found
3079 instead of taking a vector by reference to return them in.
3081 * Xapian::NumericRanges: Remove experimental API we aren't happy with yet.
3083 * Xapian::DatabaseReplica, Xapian::DatabaseMaster: Remove experimental
3084 API we aren't happy with. Replication is still supported via the
3085 command line programs. (ticket#347)
3087 * Xapian::score_evenness(): Remove as it turns out not to be useful in practice.
3090 * Xapian::ValueWeightPostingSource: A ValueWeightPostingSource with no entries
3091 would report -infinity as its upper bound, which could cause no results to be
3092 incorrectly returned for some queries involving such an object.
3094 * Xapian::WritableDatabase::close() fixed to commit() changes (unless a
3095 transaction is in progress).
3099 * apitest: Improve test coverage in various places.
3103 * Uses of values during the match (sorting by value or Sorter, MatchSpy,
3104 MatchDecider, and collapsing) now use value stream iteration which is
3105 a lot more efficient for chert and brass (but may be slower for flint).
3109 * New development backend. Changes over chert:
3111 + Batched posting list changes during indexing use significantly less memory.
3113 + Instead of using complex code to iterate modified posting lists and
3114 documents length lists, brass can flush individual such lists to disk
3115 and then iterates them from there.
3117 + To iterate all terms, chert flushes all pending postlist changes. In the
3118 case where a prefix is specified, brass only flushes postlist changes for
3119 terms starting with the specified prefix, and doesn't flush document length
3124 * Promote chert to being the stable backend.
3126 * Change the packing of integers and strings into sortable keys, which reduces
3127 database size by 2.5% in tests. This means an incompatible change in the
3128 chert format. You can use the new xapian-chert-update utility to update a
3129 chert database from the old format to the new format. It works much like
3130 xapian-compact so should take a similar amount of time (and results in a
3135 + Prune unused docids off the end of each database when merging multiple
3136 databases with renumbering.
3138 + Extend --no-renumber to support merging databases, but only if they have
3139 disjoint ranges of used document ids.
3141 + Ensure that the resultant database has a fresh UUID (previously chert
3142 copied the UUID from the first input).
3146 + Fix checking of the METAINFO key in chert. For small databases, the
3147 statistics fit in few enough bytes that incorrect check appeared to
3148 succeed and no errors were reported, but for larger databases an
3149 error was incorrectly reported.
3151 + Rework the checking of postlist chunks to use a cleaner approach which
3152 should report errors better.
3154 + Use a type wider than 32 bits to keep count of items in a table.
3155 Previously xapian-check would report the number of entries modulo
3158 * When iterating a value stream, skip_to() now only assigns the value to a
3159 std::string when it reaches its target. This saves a lot of unnecessary
3160 string copying - in a real-world test it improved the time for 100 queries
3161 from 3.66s to 3.10s.
3163 * When skipping through a chunk of postings to find the one we want, don't
3164 bother to unpack the wdf values we're skipping over. This should save a
3165 significant amount of time in certain cases where the profile data shows
3166 about a third of the time is spent in the function where this happens.
3168 * Report locking failure due to running out of file descriptors better.
3174 + Prune unused docids off the end of each database when merging multiple
3175 databases with renumbering.
3177 + Ensure that the resultant database has a fresh UUID (previously flint
3178 didn't set a UUID so one would be generated on demand when next requested,
3179 but only if the database was writable).
3181 * Report locking failure due to running out of file descriptors better.
3185 * Add support for WritableDatabase::set_metadata() and Database::get_metadata()
3186 to the remote backend (based largely on patch in #178).
3190 * Read the document data and values lazily for the inmemory backend like we do
3191 for other backends. They're much less costly to fetch than if a disk or
3192 network access is involved, but it avoids copying potentially large data
3193 which may not be needed. Consistency here also makes things easier to
3194 understand for both users and developers.
3198 * This release uses autoconf 2.65 rather than 2.64.
3202 * docs/replication.html: Add note about not using reopen() with databases being
3203 updated by the replication client.
3205 * docs/admin_notes.html: Update for chert and other recent changes.
3207 * Remove out-of-date reference in the API documentation comment to an
3208 add_slot() method. This no longer exists - you need to use multiple
3209 ValueCountMatchSpy objects to monitor more than one slot.
3213 * simpleexpand,simpleindex,simplesearch: Handle --help and --version.
3217 * The debug log now reports boolean values as "true" and "false" (instead of
3220 Xapian-core 1.1.3 (2009-09-18):
3222 This release includes all changes from 1.0.15-1.0.17 which are relevant.
3226 * Update Unicode character database to Unicode 5.2. (ticket#351)
3228 * Rename Xapian::Sorter to Xapian::KeyMaker, paving the way for using it to
3229 build collapse keys too. Xapian::Sorter remains for compatibility (and is
3230 now a subclass of Xapian::KeyMaker) but is deprecated.
3232 * Resolve the inconsistency in MultiValueSorter::add()'s "forward" parameter
3233 versus the "reverse" parameters which the Enquire sorting functions now take
3234 by replacing the class with MultiKeyMaker with a renamed method add_value()
3235 with a "reverse" parameter. MultiValueSorter remains with the old semantics
3236 for compatibility but is deprecated. (ticket#359)
3238 * QueryParser: Don't apply spelling correction to wildcarded terms, or to terms
3239 at the end of the query which we expand under FLAG_PARTIAL.
3241 * Add new Error subclass SerialisationError which we throw for serialisation
3242 related errors (which previously mostly threw NetworkError.
3244 * Rename Xapian::SerialisationContext to Xapian::Registry.
3246 * Add DecreasingValueWeightPostingSource class, which reads weights from a
3247 value slot in which a significant range of the values are in decreasing
3248 order. This functions similarly to ValueWeightPostingSource, but can be much
3251 * Add new Xapian::MatchSpy class:
3253 + This replaces the use of Xapian::MatchDecider as a "matchspy", which is now
3254 deprecated. The new class only inspects, and can't reject. It can work
3255 with remote databases, with the results being serialised to return them
3258 + Add subclass ValueCountMatchSpy, which counts the occurrences of each value
3259 in a slot in the search results seen (useful for faceted or categorisation
3260 systems). The results can be grouped into ranges using the NumericRange
3261 and NumericRanges classes, and the score_evenness() function. This API is
3262 currently experimental.
3264 * Remove default implementation of Weight::clone() which returns NULL. We
3265 always need clone() to be implemented because it's called for every term
3266 in the query, not just used for the remote backend.
3270 * Rewrite the low level packing and unpacking functions more efficiently. As
3271 well as being generally faster, the pack functions now take a reference to a
3272 string to append to, which avoids creating a lot of temporary string objects.
3273 Indexing HTML files with omindex is 5-10% faster. Searching for "The" on
3274 gmane (which results in a lot of unpacking of postings and document lengths)
3275 is about 35% faster. (ticket#326)
3277 * xapian-compact: Don't report an absent lazy input table as 0 size.
3279 * Fix ChertModifiedPostList to skip added-but-then-deleted-before-flush
3280 documents. (ticket#392)
3282 * Fix WritableDatabase::get_doclength() to work properly after a call to commit
3283 for the chert backend (ticket#397).
3285 * Fix to work with the metainfo key stored in the latest format of chert
3288 * Avoid doing pointless work by trying to delete non-existent lists of values
3289 when we're just adding documents.
3291 * Fix code to find the first docid in the next chunk (ticket#399).
3293 * Add support for chert databases without a termlist table (ticket#181).
3294 Currently the only way to create such a database is to create a chert
3295 database and do "rm termlist.*".
3299 * xapian-compact: Don't report an absent lazy input table as 0 size.
3303 * Remote protocol major version has changed to support serialising MatchSpy
3306 * Fixed not to sometimes read off the end of the returned matches when
3307 searching multiple databases, some of which are remote, and when the primary
3308 ordering is by relevance.
3312 * This release uses autoconf 2.64 rather than 2.63. This means configure now
3313 makes use of shell functions, which makes it ~13% smaller, and should also
3314 make it execute faster.
3316 * configure: Send stderr output from ldconfig to config.log.
3318 * Add optional third parameter to XO_LIB_XAPIAN autoconf macro which specifies
3319 the basename for the "xapian-config" script (defaults to "xapian-config" to
3320 give the current behaviour).
3322 * This release uses doxygen 1.5.9 to generate the API documentation.
3326 * Minor improvements to the formatting of the collated API documentation.
3330 * Fix code to compile with Sun's C++ compiler.
3332 * Fix our uuid_unparse_lower() replacement for older libuuid to actually
3333 compile (really fixes ticket#368).
3335 * Fix xapian-config to work with Solaris 10 /bin/sh. (ticket#405)
3339 * Use C++ syntax for NULL with a type in log output.
3341 Xapian-core 1.1.2 (2009-07-23):
3343 This release includes all changes from 1.0.14 which are relevant.
3347 * Move support for a prefix/suffix from NumberValueRangeProcessor to
3348 StringValueRangeProcessor, and change NumberValueRangeProcessor and
3349 DateValueRangeProcessor to inherit from StringValueRangeProcessor so all
3350 three now support a prefix/suffix. (ticket#220)
3352 * Query: Trim 4 bytes off the internals. (ticket#280)
3354 * QueryParser: If default_op is OP_NEAR or OP_PHRASE then make the window size
3355 (9 + no_of_terms) to match the default for an explicit NEAR or PHRASE.
3360 * Sort out the clash between two different patches to fix leaking file
3361 descriptors when running tests with the remotetcp backend (broken by
3366 * If the highest weighted document doesn't match all the terms in the query,
3367 its percentage weight is now calculated by simply counting how many weighted
3368 leaf subqueries match it instead of scaling by the proportion of the weight
3369 which matches (which required accessing the termlist for that document).
3372 * XOR with a SYNONYM subquery could previously achieve 100% - this has been
3377 * Backport the lazy update changes from chert to flint:
3379 WritableDatabase::replace_document() now updates the database lazily in
3380 simple cases - for example, if you just change a document's values and
3381 replace it with the same docid, then the terms and document data aren't
3382 needlessly rewritten. Caveats: currently we only check if you've looked at
3383 the values/terms/data, not if they've actually been modified, and only keep
3384 track of the last document read.
3388 * Update to always use C++ forms for ISO C standard headers (ticket#330).
3390 * Fix several places where Xapian::doccount is used instead of
3391 Xapian::termcount, and similar issues. It's still not possible to make
3392 these types different sizes, but we're now closer to this goal.
3397 * Note that PostingSource and Weight objects returned by clone() and
3398 unserialise() methods will be deallocated with "delete".
3402 * Fix debug logging not to segfault on NULL Query::Internal pointers.
3404 Xapian-core 1.1.1 (2009-06-09):
3406 This release includes all changes from 1.0.13 which are relevant.
3410 * New Query::OP_SYNONYM operator, which matches the same documents as OP_OR,
3411 but attempts to weight as if the all the subqueries were a single term with
3412 their combined wdf, which should give better relevance weights.
3414 * QueryParser's synonym, wildcard, and partial query features now use
3415 the new OP_SYNONYM operator.
3417 * PostingSource: Add new set_maxweight() method to allow subclasses to tell
3418 the matcher that their maximum weight has decreased. Make get_maxweight()
3419 a non-virtual method of the baseclass which returns the last set maxweight
3420 (which will require updates to most user subclasses. (ticket#340)
3422 * DatabaseReplica: Fix SEGV when calling get_description() on a default
3423 constructed DatabaseReplica.
3425 * Make Query::MatchAll and Query::MatchNothing const since they're immutable.
3426 All the public methods of Query are const, so this should be completely API
3429 * Methods returning an end iterator for a ValueIterator now actually return a
3430 proxy object which silently converts to ValueIterator if required. This
3431 proxy object allows a comparison with an "_end()" method to be optimised
3432 better so that it just ends up comparing the internal member of the iterator
3433 class with NULL (previously a call to ValueIterator's destructor remained).
3434 This should be API compatible, but note that it is definitely now more
3435 efficient just to compare against the return value of the relevant _end()
3436 method than to store the end iterator explicitly.
3440 * Testcase valuestats4 requires transactions, so indicate that and remove the
3441 explicit SKIP for inmemory.
3443 * Testcase changemaxweightsource1 uses ChangeMaxweightPostingSource, which
3444 doesn't work with multi or remote, so mark the test accordingly.
3446 * We've decided that "going back" with skip_to() or check() should have
3447 unspecified behaviour, so stop testing how this case behaves!
3451 * Subclass MultiPostList directly from PostList instead of from LeafPostList.
3452 This gets rid of two unused data members per MultiPostList in exchange for
3453 having to define 5 extra "never called" methods, but 4 of these just
3456 * Store termfreqs and reltermfreqs for query terms in a single map rather than
3457 one map for each, which saves is more compact and likely to be faster.
3461 * xapian-check: For chert, check value stats are the correct format and that
3462 the streamed values are consistent with their stats (ticket#277).
3464 * xapian-check: Chert doesn't store termlist entries for documents without
3465 terms, which resulted in us reporting an error when we found document ids in
3466 the doclength "postlist" which were greater than any with an entry in the
3467 termlist. Instead compare these entries against db.get_last_docid() if we
3468 are checking a whole db and the db can be opened. If not, suppress this
3473 * When serialising stats, serialise the termfreq and reltermfreq together,
3474 rather than in separate lists. This gives a smaller serialised form, and
3475 matches these both being stored in the same map now. This is an incompatible
3476 remote protocol change, so bump the major version to 32. (ticket#362)
3480 * Some build failures with --disable-backend-XXX options have been fixed, but
3481 we haven't exhaustively tested all combinations.
3483 * Ship common/win32_uuid.cc and common/win32_uuid.h (ticket#367).
3487 * Update PostingSource documentation to describe how init() is called again if
3488 a PostingSource is reused. Fixes #352.
3492 * Fixed to build with GCC 4.4.
3494 * Drop support for GCC 2.95.3 and 3.0.x - we now require at least 3.1 as doing
3495 so eliminates some preprocessor conditionals which we aren't able to test
3496 regularly as we don't have easy access to such old GCC versions. GCC 3.1 is
3497 nearly 7 years old now, and GCC3 didn't get widespread use until later
3498 versions anyway. If you still need to use GCC < 3.1, Xapian 1.0.x should
3499 build with 2.95.3 or newer.
3501 * Older versions of libuuid don't have uuid_unparse_lower() so probe for it in
3502 configure, and if it isn't present provide an inline version in safeuuid.h
3505 * Fixed to build with MSVC (ticket#379).
3507 * Add static_cast<char>() to str(bool) overload to suppress bogus MSVC warning
3512 * common/debuglog.h: Add missing initialisation of uncaught_exception variable
3513 in a couple of places.
3515 Xapian-core 1.1.0 (2009-04-22):
3519 * All deprecated xapian-core features listed for removal in 1.1.0 have been
3520 removed. See deprecation.html for details, and suggested updates.
3522 * The Unicode character categorisation functions have been updated from
3525 * Add NON_SPACING_MARK to is_wordchar() for better tokenisation of languages
3526 which use such marks - for example, Arabic. This is better than the stop-gap
3527 fix in 1.0 of treating NON_SPACING_MARK as a phrase-generator character
3528 when parsing queries, but it does mean that databases built from data
3529 containing such characters will need to be rebuilt. (ticket#355)
3531 * The details of how to subclass Xapian::Weight to implement your own
3532 weighting scheme have changed incompatibly to allow user weighting schemes
3533 to have access to the same statistics as built-in schemes (ticket#213)
3534 If you have a existing subclass of Xapian::Weight you'll need to update it.
3536 * New Database methods get_doclength_upper_bound(), get_doclength_lower_bound()
3537 and get_wdf_upper_bound(), primarily intended for allowing weighting schemes
3538 to calculate tighter upper bounds on weights (which BM25Weight and TradWeight
3539 now do) which allows matcher weight-based optimisations to be more effective.
3540 Chert actually tracks doclength bounds and a global (rather than per term)
3541 upper bound on wdf; other backends return much less tight bounds, but these
3542 still lead to better upper bounds on weights.
3544 * Enquire::get_eset() now uses an unmodified of probabilistic formula, and
3545 doesn't return terms which would get a negative weight from it (since that
3546 means they are expected to be harmful not helpful).
3548 * Add Database::close() method, which will release system resources (in
3549 particular, close filehandles) held by a database. This is particularly
3550 useful when wrapping the API for languages with garbage collection.
3552 * Change Database::positionlist_begin() not to throw exceptions if the term or
3553 document doesn't exist.
3555 * Xapian databases now have a UUID, readable with Database::get_uuid().
3557 * A new Database replication API has been added (currently experimental).
3559 * MSet::get_termfreq() will now fall back to looking up the term frequency in
3560 the database rather than raising an exception if a term wasn't present in
3563 * Calling RSet:add_document() with argument 0 now throws InvalidArgumentError.
3565 * QueryParser sped up (new version of lemon); queryparsertest runs 2.2% faster.
3567 * Add ValueSetMatchDecider, which is a matchdecider which is intended to be
3568 passed a set of values to look for in documents, and selects documents based
3569 on the presence of those values.
3571 * Add new Xapian::PostingSource class to allow passing custom sources of
3572 postings and weights to the matcher. Built-in PostingSource subclasses:
3573 FixedWeightPostingSource, ValueMapPostingSource, ValuePostingSource, and
3574 ValueWeightPostingSource. (Currently experimental).
3576 * Database: Add get_value_freq(), get_value_lower_bound() and
3577 get_value_upper_bound() methods to get statistics about the values stored in
3578 a slot. Add support for the value statistics methods to chert, inmemory,
3579 multi and remote databases.
3581 * Enquire::get_eset() now faster for large ESet size.
3583 * Xapian::Document objects now have a reduced memory footprint.
3585 * Enquire::set_collapse_key() now allows you to specify a maximum number of
3586 matches with each collapse key to keep (which defaults to 1, giving the
3587 previous behaviour). Enquire can now report bounds and an estimate of what
3588 the total number of matches would have been if collapsing wasn't in use.
3590 * WritableDatabase::commit() is a new, preferred alias for
3591 WritableDatabase::flush(). (ticket#266)
3593 * Add methods for serialising documents and queries to strings, and
3594 unserialising back from strings. (ticket#206)
3598 * stemtest: No longer checks environment variables OM_STEMTEST_SKIP_RANDOM,
3599 OM_STEMTEST_LANGUAGES, and OM_STEMTEST_SEED.
3601 * perftest: New performance testsuite. This is intended to contain intended to
3602 contain potentially time-consuming performance tests, which log output to
3603 an XML file for later analysis. It's not run by "make check" - use "make
3604 check-perf" to run it.
3606 * apitest: Now runs tests over both flint and chert for multi, remotetcp, and
3609 * Wait for subprocesses to finish at end of tests with remotetcp backend, to
3610 avoid test failures when the same database is used for the next testcase.
3614 * Internally, pass around non-normalised document lengths as Xapian::termcount
3615 (unsigned integer) not Xapian::doclength (double). This gives a 3% speedup
3616 for 10 term OR queries!
3620 * New development backend. Use Chert::open() to explicitly create a chert
3621 format database, or set XAPIAN_PREFER_CHERT=1 in the environment to
3622 prefer chert when creating a new database without an explicit type.
3624 * Quartz and Flint stored the document length alongside every posting list
3625 entry. Chert instead stores a chunked list of all the document lengths
3626 which saves a lot of space, and is a big win for large queries or those
3627 which don't need the document lengths. This structure is used to
3628 implement much faster iteration (six times faster in a test) over all
3629 document ids (which speeds up queries using unary NOT, e.g. `NOT apples'),
3630 and to test for the existence of documents (instead of checking the record
3631 table for an entry).
3633 * Document values are now stored in a chunked stream for each slot for
3634 efficient access to the same slot in lots of documents. This makes
3635 operations like sort by value much more efficient.
3637 * WritableDatabase::replace_document() now updates the database lazily in
3638 simple cases - for example, if you just change a document's values and
3639 replace it with the same docid, then the terms and document data aren't
3640 needlessly rewritten. Caveats: currently we only check if you've looked at
3641 the values/terms/data, not if they've actually been modified, and only keep
3642 track of the last document read.
3646 * If we can't obtain a write lock while trying to create a new database
3647 we now report the lock failure with DatabaseLockError, not
3648 DatabaseOpeningError - it's more useful to know that the lock attempt failed
3651 * Improve reporting of failures to obtain lock due to unexpected errors.
3653 * xapian-check: Don't stop checking a table after an error in certain cases -
3654 instead increment the error counter and try to continue checking from the
3659 * The remote database protocol major version has been increased, allowing
3660 a significant amount of compatibility code to be removed. This change means
3661 that new clients won't work with old servers, and old clients won't work
3662 with new servers. If upgrading a live system, you will need to take this
3665 * The remote servers now always default to opening a Database and the client
3666 has to send a protocol message to explicitly request write access. This
3667 allows a single server to support multiple readers and one writer
3668 simultaneously. (ticket#145)
3670 * Database::get_document() no longer does an unnecessary copy of the document's
3673 * Change serialisation of queries to be more compact and easier to parse.
3677 * Stub databases used to assume that any relative paths were relative to the
3678 current working directory. They now assume that relative paths are
3679 relative to the directory holding the stub database file.
3681 * Stub database lines which begin with a '#' character are now ignored,
3682 allowing comments in stub database files.
3684 * New "stub directory" database type - this is a directory containing a stub
3685 database file named "XAPIANDB".
3687 * Don't just ignore lines with no spaces in a stub database file.
3689 * Bad lines in a stub file were being ignored after we'd seen a good entry.
3691 * Add new Auto::open_stub() overload which opens a stub database file
3692 containing a single entry as a WritableDatabase.
3694 * Add support for "inmemory" to stub database (which is useful now that stub
3695 databases can be opened for writing).
3697 * A stub database file is now allowed to contain no database entries, which
3698 results in an empty Database object (this avoids user code having to special
3699 case to handle "0 or more" databases).
3703 * To allow installations of Xapian 1.0 and 1.1 to easily coexist, the library
3704 is now libxapian-1.1; xapian.m4 is now xapian-1.1.m4; headers are now
3705 installed in $prefix/include/xapian-1.1. If you use XO_LIB_XAPIAN or
3706 xapian-config as we recommend, this should all be transparent. Also
3707 programs and scripts have a default program suffix to -1.1 unless overridden
3708 using the --program-suffix argument to configure (if you really want no
3709 suffix, "./configure --program-suffix=" will achieve this).
3711 * On Linux and k*bsd-gnu, override libtool's link_all_deplibs_CXX to "no".
3713 * On Linux, override libtool's sys_lib_dlsearch_path_spec to a list generated
3714 in a more reliable way which includes all the default directories.
3716 * configure: --enable-debug and --enable-debug-verbose have been deprecated
3717 since 1.0.0, so remove specific errors pointing to the replacements.
3721 * Disable "JAVADOC_AUTOBRIEF" in doxygen configuration since we always try to
3722 write a brief description explicitly, and JAVADOC_AUTOBRIEF causes problems
3725 * docs/deprecation.html: Describe what "experimental" features are, and why
3726 replication and posting sources are currently experimental.
3728 * docs/deprecation.html: Deprecate Stem_get_available_languages() from the
3733 * Use C++ forms of C headers in examples (ticket#330).
3737 * xapian-core.spec: We no longer need to run autoreconf to work around
3738 libtool's incomplete sys_lib_dlsearch_path_spec or to pick up distro-specific
3739 patches for link_all_deplibs.
3743 * Report get_description() rather than the pointer value for
3744 Xapian::Query::Internal* parameters to internal functions.
3746 * The debug logging framework has been overhauled. See HACKING for details
3747 of how it now works.
3749 * Faster integer to string functions inside the library (this is a general
3750 improvement, but will particularly speed up debug logging as that converts a
3751 lot of integers to strings).
3753 Xapian-core 1.0.23 (2011-01-14):
3757 * QueryParser: Avoid a double free if Query construction throws an exception
3758 in a particular case. Fixes ticket#515.
3760 * QueryParser: Handle NEAR/<offset> and ADJ/<offset> where offset isn't an
3761 integer the same way at the end of the query as in the middle.
3763 * Enquire::get_mset(): Avoid pointlessly trying to allocate lots of memory
3764 if the first document requested is larger than the size of the database.
3766 * Enquire::get_mset(): An empty query now returns an MSet with firstitem set
3767 correctly - previously firstitem was always 0 in this case.
3771 * The matcher wasn't recalculating the max possible weight after a subquery of
3772 XOR reached its end. This caused an assertion failure in debug builds, and
3773 is a missed optimisation opportunity.
3777 * xapian-compact: Fix access to empty priority_queue while merging synonyms.
3778 This could have caused problems, though we've had no reports of any (the
3779 bug was found with _GLIBCXX_DEBUG).
3781 Xapian-core 1.0.22 (2010-10-03):
3785 * Xapian::Document: Initialise docid to 0 when creating a document from
3786 scratch, as documented.
3788 * Xapian::QueryParser: Allow phrase generators between a probabilistic prefix
3789 and the term itself (e.g. path:/usr/local).
3793 * Back out the OP_OR efficiency improvement made in 1.0.21 since this change
3794 slows down some other common cases. We'll address this fully in 1.2.4, but
3795 that fix is more invasive than we are comfortable with for 1.0.x at this
3800 * xapian-config: Add --static option which makes other options report values
3805 * deprecation.html: Add guidelines for supporting other software.
3807 * Document cases where QueryParser's FLAG_WILDCARD and FLAG_PARTIAL aren't
3808 currently supported.
3810 * Fix documentation for Xapian::timeout type - it holds a time interval in
3811 milliseconds not microseconds (the API docs for the methods which use it
3812 explicitly correctly document that the timeouts are in milliseconds).
3816 * configure: Don't pass -mtune=generic unless GCC >= 4.2 is in use
3819 * configure: Add support for --enable-sse=sse and --enable-sse=sse2 to allow
3820 control of which SSE instructions to use.
3822 * configure: Enable use of SSE maths on x86 by default with Sun's compiler.
3824 * configure: Beef up the test for whether -lm is required and add a special
3825 case to force it to be for Sun's C++ compiler - there's some interaction with
3826 libtool and/or shared objects which means that the previous configure test
3827 didn't think -lm is needed here when it is.
3829 * Fix test harness to build under Microsoft Windows (ticket#495).
3831 * Fix to build on OpenBSD 4.5 with GCC 3.3.5.
3833 * Need to avoid excess precision on m68k when targeting models 68010, 68020,
3834 68030 as well as 68000.
3838 * xapian-core.spec: Add cmake related files to RPM packaging.
3840 Xapian-core 1.0.21 (2010-06-18):
3844 * Xapian::Stem now recognises "nb" and "nn" as additional codes for the
3847 * Xapian::QueryParser now correctly parses a wildcarded term in between two
3848 other terms (ticket#484).
3852 * Improve test coverage of OP_VALUE_RANGE and MSet::get_percent().
3856 * OP_OR could skip a matching document if it decayed to OP_AND or OP_AND_MAYBE
3857 during the match in some cases. Fixes ticket#476.
3859 * OP_XOR with non-leaf subqueries could skip matching documents in some cases,
3860 and OP_XOR of three or more sub-queries could return incorrect weights.
3863 * OP_OR is now more efficient if a subquery is potentially expensive (e.g.
3864 ValueRangePostList, OP_NEAR, OP_PHRASE). A 10-fold speed-up with
3865 ValueRangePostList has been observed.
3869 * When iterating a table, if the table changes underneath we could end up
3870 returning the same entry twice. (Debian#579951)
3872 * A cancelled transaction (or a failing operation implicitly cancelling
3873 pending changes) now marks the tables as unmodified, which fixes an exception
3874 trying to read block 0 if one of the tables is empty on disk.
3878 * When iterating a table, if the table changes underneath we could end up
3879 returning the same entry twice. (Debian#579951)
3883 * When daemonising, read the max fd to close with sysconf() instead of using
3884 a hardcoded value of 256, and work even if stdin and stdout have been closed.
3888 * Install files to make Xapian easier to use with cmake.
3892 * Update the list of languages that the Xapian::Stem constructor recognises.
3894 * Assorted minor improvements to the collated API documentation.
3898 * On x86 processors, Xapian now defaults to using SSE2 FP instructions. This
3899 avoids issues with excess precision and it a bit faster too. If you need
3900 to support processors without SSE2 (this means pre-Pentium4 for Intel) then
3901 configure with --disable-sse. (ticket#387)
3903 * Fix warning when compiling for mingw with GCC 4.2.1.
3905 * Remove mutable from a couple of reference class members - mutable doesn't
3906 make sense for a reference and some compilers warn about it.
3908 Xapian-core 1.0.20 (2010-04-27):
3912 * MSet: Fix incorrect values reported by get_matches_estimated(),
3913 get_matches_lower_bound(), and get_matches_upper_bound() in certain cases
3914 when sorting and collapsing (ticket#464).
3918 * deprecation.html: Note how to disable deprecation warnings. (ticket#393)
3922 * delve: Add -a option to list all terms in a database.
3924 * delve: -d and -V command line options now report out of range and invalid
3929 * The getopt warning fix for Cygwin in 1.0.19 caused build failures on Mac OS X
3930 (and probably some other platforms with non-GNU getopt implementations), so
3931 replace with a fix which is only enabled for Cygwin. (ticket#469)
3933 Xapian-core 1.0.19 (2010-04-15):
3937 * QueryParser: Fix leak if Xapian::Database throws an exception during parsing
3942 * Explicitly flush after indexing for quartz and flint, so we see any
3943 exceptions from the flush (the implicit flush from the destructor swallows
3946 * apitest: Add databasemodified1 testcase to provide some test coverage for
3947 DatabaseModifiedError.
3951 * When updating a document, rather than decoding the old positions, comparing
3952 with the new, and then encoding the new if different, we now just encode the
3953 new and then compare the encoded forms. (ticket#428)
3955 * Avoid trying to delete the document positions when we know there aren't any.
3957 * Fix memory leak if Database::allterms_begin() throws an exception
3960 * xapian-check: Report document id for document length mismatch.
3962 * Fix potential issues with iterators over a WritableDatabase which is modified
3963 during iteration. No problems have actually been observed with flint, only
3964 in 1.1.4 with chert in cases which don't occur in flint, but it seems likely
3965 the issue can manifest for flint in other situations. Fixes ticket#455.
3967 * Initialise zlib z_stream structure members zalloc, zfree, and opaque with
3968 Z_NULL rather than 0 cast to the appropriate type, as that's what the zlib
3969 documentation says to do. Add missing initialisation of opaque for the
3970 inflate z_stream which the zlib docs say is needed (reading the zlib code,
3971 this isn't true for current versions, so this improves robustness rather
3972 than fixing an observable bug).
3974 * Don't memcpy() a block to itself - it's a waste of effort, and (probably)
3975 undefined behaviour (as a block overlaps itself).
3979 * Fix potential issues with iterators over a WritableDatabase which is modified
3980 during iteration. No problems have actually been observed with quartz, only
3981 in 1.1.4 with chert in cases which don't occur in quartz, but it seems likely
3982 the issue can manifest for quartz in other situations. Fixes ticket#455.
3984 * Don't memcpy() a block to itself - it's a waste of effort, and (probably)
3985 undefined behaviour (as a block overlaps itself).
3989 * Force -fno-strict-aliasing for GCC 4.2 to avoid bad code being generated due
3990 to a bug in that compiler version. Fixes ticket#449. This issue hasn't been
3991 observed to affect Xapian 1.0.x, but it seems prudent to backport the fix.
3995 * INSTALL: Correct description of --enable-assertions. It does NOT enable
3996 debugging symbols, and shouldn't control checks on bad data passed to API
3997 calls (if it does anywhere, that's a bug). Note that Xapian will run more
3998 slowly with assertions on.
4002 + Add section on indexing.
4004 + Add a note about removing automatically added spelling dictionary entries.
4006 + Move the "algorithm" section to the end, as it is really just background
4007 information for the curious.
4009 * include/xapian/queryparser.h: Document the possible exception messages from
4010 QueryParser::parse_query().
4012 * include/xapian/termgenerator.h: Note how TermGenerator handles stopwords.
4016 * delve: Display the lastdocid value when displaying general database
4019 * simpleindex: Explicitly call flush() on the database, as that is good
4020 practice (since you see any exceptions).
4024 * Fix compilation failure in testsuite on OpenBSD, introduced by new regression
4025 test in 1.0.18. Fixes ticket#458.
4027 * Fix getopt-related warning on Cygwin.
4029 Xapian-core 1.0.18 (2009-02-14):
4033 * Document: Add new add_boolean_term() method, which is an alias for add_term()
4038 + Add support for quoting boolean terms so they can contain arbitrary
4039 characters (partly addresses ticket#128).
4041 + Add ENCLOSING_MARK and COMBINING_SPACING_MARK categories, plus several
4042 zero-width space characters, as phrase generators. This mirrors a better
4043 fix in 1.1.4, but without losing compatibility with existing databases.
4045 + Fix handling of an explicit AND before a hated term (foo AND -bar).
4048 * TermIterator: Only include trailing '+' or '#' on a term if it isn't followed
4049 by a word character (makes more sense and matches QueryParser's behaviour).
4052 * Database: Fix many methods to behave better on a database with no
4053 subdatabases, such as is constructed by Database(). Fixes ticket#415.
4057 * Add test coverage for xapian-compact, and improve coverage for
4058 WritableDatabase::replace_document().
4060 * apitest: Rename matchfunctor<n> to matchdecider<n> to match current
4065 * When updating documents, don't update posting entries which haven't changed.
4066 Largely fixes ticket #250.
4068 * If the number of entries in the position table happened to be 4294967296 or
4069 an exact multiple, Xapian would ignore positional data for that table when
4070 running queries, and xapian-compact wouldn't copy its contents.
4072 * Iterating all the terms in the database with a prefix is now slightly more
4075 * Fix locking code to work if stdin and/or stdout have been closed.
4077 * If a document is replaced with itself unmodified, we no longer increase the
4078 automatic flush counter.
4080 * When iterating a posting list modified since the last flush(), the reported
4081 wdf is now correct (previously it was too high by its old value).
4083 * Replacing a document deleted since the last flush failed to update the
4084 collection frequency and wdf, and caused an assertion failure when assertions
4087 * WritableDatabase::replace_document() didn't always remove old positional
4088 data (the only effect is that the position table was bloated by unwanted
4093 + New "until" command which shows entries until a specified key is reached.
4095 + New "open" command which allows easy switching between tables.
4097 * xapian-compact: Fix typos in --help output.
4101 * Replacing a document deleted since the last flush failed to update the
4102 collection frequency and wdf, and caused an assertion failure when assertions
4105 * WritableDatabase::replace_document() didn't always remove old positional
4106 data (the only effect is that the position table was bloated by unwanted
4111 * Throw UnimplementedError if a MatchDecider is used with the remote backend.
4112 Previously Xapian returned incorrect results in this case.
4116 * configure: With --enable-maintainer-mode, enable -Werror for GCC >= 4.1
4117 rather than >= 4.0 as Apple's GCC 4.0 gives bogus uninitialised variable
4122 * The API documentation now includes Xapian::Error and subclasses, and doesn't
4123 mention Xapian::Query::Internal.
4125 * Make clear in the Xapian::Document API documentation that this class is a
4126 lazy handle and discuss the issues this can cause.
4128 * INSTALL: Improve text about zlib dependency.
4130 * HACKING: Add details of our licensing policy for accepting patches.
4134 * quest: If no database is specified, still parse the query and report
4135 Query::get_description() to provide an easy way to check how a query parses.
4139 * Fix GCC 4.2 warning.
4141 xapian-core 1.0.17 (2009-11-18):
4147 + Fix handling of a group of two or more terms which are all stopwords which
4148 notably caused issues when default_op was OP_AND, but could probably
4149 manifest in other cases too. Fixes ticket#406.
4151 + Fix interaction of FLAG_PARTIAL and FLAG_SYNONYM. (ticket#407)
4153 * Database: A database created via the default constructor no longer causes a
4154 segfault when the methods get_metadata() or metadata_keys_begin() are called.
4158 * Don't try to close the fd one more than the maximum allowable when locking
4159 the database. Harmless, except it causes a warning when running under
4160 valgrind. (ticket#408)
4164 * Xapian::Sorter isn't supported with the remote backend so throw
4165 UnimplementedError rather than giving incorrect results. (ticket#384)
4167 * Fix potential reading off the end of the MSet which is returned internally
4168 by the remote server.
4172 * Various documentation comment improvements for the Database class.
4176 * examples/quest.cc: Tighten up the type of the error we catch to detect an
4177 unknown stemming language.
4181 * xapian-config: Need to quote ^ for Solaris /bin/sh.
4183 * configure: Actually use any flags we determine are needed to switch the
4184 compiler to proper ANSI C++ mode, when building xapian-core - this stopped
4185 working in 1.0.12, breaking support for HP's aCC, Compaq's cxx, Sun's CC, and
4188 Xapian-core 1.0.16 (2009-09-10):
4192 * Fix a typo which stopped this fix in 1.0.12 from working (ticket #398):
4194 If we fail to get the lock after we spawn the child lock process (the common
4195 case is because the database is already open for writing) then we now clean
4196 up the child process properly.
4200 * Improve API documentation of QueryParser::set_default_op() and
4201 QueryParser::get_default_op().
4205 * Fix build failure on Mac OS X 10.6.
4207 Xapian-core 1.0.15 (2009-08-26):
4211 * Fix the test harness not to report heaps of bogus errors when using valgrind
4216 * Backport the lazy update changes from 1.1.2:
4218 WritableDatabase::replace_document() now updates the database lazily in
4219 simple cases - for example, if you just change a document's values and
4220 replace it with the same docid, then the terms and document data aren't
4221 needlessly rewritten. Caveats: currently we only check if you've looked at
4222 the values/terms/data, not if they've actually been modified, and only keep
4223 track of the last document read.
4225 * Fix PostingIterator::skip_to() on an unflushed WritableDatabase to skip
4226 documents which were added and deleted since the last flush. (ticket#392)
4230 * Overhaul the doxygen options we use and tweak various documentation comments
4231 to improve the generated API documentation.
4233 * Explicitly document that an empty prefix argument to
4234 QueryParser::add_prefix() means "no prefix".
4236 * Update the documentation comments for Enable::set_sort_by_value(),
4237 set_sort_by_value_then_relevance(), and set_sort_by_relevance_then_value() to
4238 mention sortable_serialise() as a good way to store numeric values for
4241 Xapian-core 1.0.14 (2009-07-21):
4245 * When using more than one ValueRangeProcessor, QueryParser didn't reset the
4246 begin and end strings to ignore any changes made by a ValueRangeProcessor
4247 which returned false, so further ValueRangeProcessors would see any changes
4248 it had made. This is now fixed, and test coverage improved.
4252 * The test harness code which launches xapian-tcpsrv child processes was
4253 failing to close a file descriptor for each one launched due to a bug in
4254 the code which is meant to track them. This was causing apitest to fail
4255 on OpenBSD (ticket#382). Also wait between testcases for any spawned
4256 xapian-tcpsrv processes to exit to avoid spurious failures when a database is
4257 reused by the next testcase.
4259 * tests/runtest.in: Use "ulimit -n" where available to limit the number of
4260 available file descriptors to 64 so we catch file descriptor leaks sooner.
4262 * When measuring CPU time used for scalability tests, we no longer try to
4263 include the CPU time used by child processes, as we can only get that for
4264 child processes which have exited and it's hard to ensure that they have
4265 with the current framework. Although this means we only tests the
4266 client-side scaling for remote tests, the local backend tests cover most of
4267 the work done by the server part of the remote backend.
4269 * apitest: In testcase topercent2, don't expect max_attained or max_possible to
4270 be exact as rounding errors in different ways of calculating can cause small
4271 variations. On trunk we already have similar code because the new weighting
4272 scheme stuff gives different bounds in the different cases. This should fix
4273 testsuite failures seen on some of the Debian and Ubuntu buildds.
4275 * The test harness now always reports the full exception message (was
4276 conditional on --verbose), and output for different exception types and
4277 other causes of failure is now more consistent.
4279 * For scalability tests, the test harness now increases the number of
4280 repetitions until the first run takes more than 0.001 seconds, to avoid
4281 trying to base calculations on a length of time we probably can't reliably
4282 measure to start with.
4284 * Add test coverage for Stem::get_description() for each supported language.
4286 * queryparsertest: Reenable tests which require the inmemory backend to be
4287 enabled by fixing typo XAPIAN_HAS_BACKEND_INMEMORY ->
4288 XAPIAN_HAS_INMEMORY_BACKEND.
4292 * Use F_FULLFSYNC where available (Mac OS X currently) to ensure that changes
4293 have been committed to disk. (ticket#288)
4297 * Fix handling of percentage weights in various cases when we're searching
4298 multiple remote databases or a mix of local and remote databases.
4302 * configure: -Wshadow produces false positives with GCC 4.0, so only enable it
4303 for >= 4.1 since we enable -Werror for maintainer-mode builds for GCC >= 4.0.
4305 * configure: Check that we can find the valgrind/memcheck.h header as well as
4306 the valgrind binary.
4308 * Change how snowball generates the data used by its among operation - instead
4309 of using pointers to the strings in struct among, store an offset into a
4310 constant pool, as this reduces the number of relocations by about 2300, which
4311 should decrease the time taken by the dynamic linker when loading the
4312 library. This also reduces the size of the shared library significantly
4313 (on x86-64 Linux, the stripped shared library is 4% smaller).
4315 Xapian-core 1.0.13 (2009-05-23):
4319 * Xapian::Document no longer ever stores empty values explicitly. This
4320 wasn't intentional behaviour, and how this case was handled wasn't
4321 documented. The amended behaviour is consistent with how user metadata
4322 is handled. This change isn't observable using Document::get_value(),
4323 but can be noticed when iterating with Document::values_begin(), using
4324 Document::values_count(), or trying to delete the value with
4325 Document::remove_value().
4329 * Fix testcase scaleweight4 not to fail on x86 when compiled with -O0. The
4330 problem was in the testcase code, and was caused by excess precision in
4331 intermediate FP values.
4333 * Testcases which check that operations have the expected O(...) behaviour now
4334 check CPU time instead of wallclock time on most platforms, which should
4335 eliminate occasional failures due to load spikes from other processes.
4338 * Fix test failures due to SKIP_TEST_FOR_BACKEND("inmemory") not skipping when
4339 it should due to comparing char * strings with == (on trunk the return value
4340 being tested is std::string rather than const char *).
4342 * Improve test coverage in several corner cases.
4344 * Fix testcase consistency2 to actually be run (fortunately it passes).
4346 * In the generated testcases, call get_description() on the default
4347 constructed object of each class to make sure that works (and doesn't try to
4348 dereference NULL, or fail some assertion, etc). All currently checked
4349 classes are fine - this is to avoid future regressions or such problems with
4352 * In the test coverage build, use "--coverage" instead of "-fprofile-arcs
4355 * The test harness now has the inmemory backend flagged as supporting
4356 user-specified metadata (apart from iteration over metadata keys).
4360 * If a query contains a MatchAll subquery, check for it before checking the
4361 other terms so that the loop which checks how many terms match can exit
4362 early if they all match.
4364 * When an OR or ANY_MAYBE decayed to an AND, we were carefully swapping the
4365 children for maximum efficiency, but the condition was reversed so we were
4366 in fact making things worse. This was noticed because it was resulting in
4367 the same query running faster when more results were asked for!
4369 * Only build the termname to termfreq and weight map for the first subdatabase
4370 instead of rebuilding it for each one. Also don't copy this map to return
4371 it. This should speed up searches a little, especially those over multiple
4374 * If a submatcher fails but ErrorHandler tells us to continue without it, we
4375 just use a NULL pointer to stand in rather than allocating a special dummy
4376 place-holder object.
4378 * Remove AndPostList, in favour of MultiAndPostList. AndPostList was only used
4379 as a decay product (by AndMaybePostList and OrPostList), and doesn't appear
4380 to be any faster. Removing it reduces CPU cache pressure, and is less code
4383 * Call check() instead of skip_to() on the optional branch of AND_MAYBE.
4387 * Fix a bug in TermIterator::skip_to() over metadata keys.
4391 * Fix xapian-tcpsrv --interface option to work on MacOS X (ticket#373).
4393 * Fix typo which caused us to return the docid instead of the maximum weight
4394 a document from a remote match could return! This could have led to wrong
4395 results when searching multiple databases with the remote backend, but
4396 probably usually didn't matter as with BM25 the weights are generally small
4397 (often all < 1) while docids are inevitably >= 1.
4401 * The inmemory backend doesn't support iterating over metadata keys. Trying
4402 to do so used to give an empty iteration, but has now been fixed to throw
4403 UnimplementedError (and this limitation has now been documented).
4407 * Remove a lot of unused header inclusions and some unused code which should
4408 make the build faster and slightly smaller.
4410 * Fix to compile under --disable-backend-flint, --disable-backend-remote, and
4411 --disable-backend-inmemory.
4413 * Don't remove any built sources in "make clean" even under
4414 --make-maintainer-mode as that breaks switching a tree away from
4415 maintainer-mode with: make distclean;./configure
4417 * configure: Enable more GCC warnings - "-Woverloaded-virtual" for all
4418 versions, "-Wstrict-null-sentinel" for 4.0+, "-Wlogical-op
4419 -Wmissing-declarations" for 4.3+. Notably "-Wmissing-declarations" caught
4420 that consistency2 wasn't being run.
4422 * Internally, fix the few places where we pass std::string by value to pass
4423 by const reference instead (except where we need a modifiable copy anyway) as
4424 benchmarking shows that const reference is slightly faster and generates
4425 less code with GCC's reference counted std::string implementation - with a
4426 non-reference counted implementation, const reference should be much faster.
4431 * INSTALL: We no longer regularly test build with GCC 2.95.4 and we're raising
4432 the minimum GCC version required to 3.1 for Xapian 1.1.x.
4434 * Document what passing maxitems=0 to Enquire::get_mset() does.
4436 * docs/queryparser.html: Add examples of using a prefix on a phrase or
4439 * Correct doxygen comments for user metadata functions:
4440 Database::get_metadata() can't throw UnimplementedError but
4441 WritableDatabase::set_metadata() can.
4443 * Document that Database::metadata_keys_begin() returns an end iterator if the
4444 backend doesn't support metadata.
4446 * HACKING: Update the list of Debian/Ubuntu packages needed for a development
4451 * Fix build with --enable-debug.
4453 * Added some more assertions.
4455 Xapian-core 1.0.12 (2009-04-19):
4459 * WritableDatabase::remove_spelling() now works properly.
4461 * The QueryParser now treats NON_SPACING_MARK Unicode characters as phrase
4462 generators, which improves handling of Arabic. This is a stop-gap solution
4463 for 1.0.x which will work with existing databases without requiring
4464 reindexing - in 1.1.0, NON_SPACING_MARK will be regarded as part of a word.
4467 * Fix undefined behaviour in distribution of OP_NEAR and OP_PHRASE over a
4468 non-leaf subquery (indentified by valgrind on testcase nearsubqueries1).
4471 * Enhance distribution of OP_NEAR/OP_PHRASE over non-leaf subqueries to work
4472 when there are multiple non-leaf subqueries (ticket#201).
4474 * Enquire::get_mset() no longer needlessly checks if the documents exist.
4476 * PostingIterator::get_description() output improved visually in some cases.
4480 * Add make targets to assist generating a testsuite code coverage report with
4481 lcov. See HACKING for details.
4483 * Improved test coverage in a number of places and removed some used code as
4484 shown by lcov's coverage report.
4490 + Now handles databases which contains no documents but have user metadata
4493 + Fix test for the total document length overflowing.
4495 * Release the database lock if the database is closed due to an unrecoverable
4496 error during modifications. (ticket#354)
4498 * If we fail to get the lock after we spawn the child lock process (the common
4499 case is because the database is already open for writing) then we now clean
4500 up the child process properly.
4504 * Overriding CXXFLAGS at make-time (e.g. "make CXXFLAGS=-Os") no longer
4505 overrides any flags configure detected to be required to make the compiler
4506 accept ISO C++ (for GCC, no such flags are required, so this doesn't
4511 * Update documentation and code comments to reflect that 1.1 will be a
4512 development series, and 1.2 the next release series.
4514 * docs/admin_notes.html: Document the child process used for locking which
4515 exec-s "cat" (ticket #258).
4517 * include/xapian/unicode.h: Fix documentation comment typos.
4519 * include/xapian/matchspy.h: Removed currently unused header to stop doxygen
4520 from generating documentation for it.
4522 Xapian-core 1.0.11 (2009-03-15):
4526 * Enquire::get_mset():
4528 + Now throws UnimplementedError if there's a percentage cutoff and sorting is
4529 primarily by value - this has never been correctly supported and it's
4530 better to warn people than give incorrect results.
4532 + No longer needlessly copies the results internally.
4534 + When searching multiple databases, now recalculates the maximum attainable
4535 weight after each database which may allow it to terminate earlier.
4538 + Fix inconsistent percentage scores when sorting primarily by value, except
4539 when a MatchDecider is also being used; document this remaining problem
4542 * Enquire::set_sort_by_value() (and similar methods): Rename the wrongly named
4543 "ascending" parameter to "reverse", and note that its value should always be
4544 explicitly given since defaulting to "reverse=true" is confusing and the
4545 default will be deprecated in 1.1.0. (ticket#311)
4547 * Database::allterms_begin(): Fix memory leak when iterating all terms from
4548 more than one database.
4550 * Query::get_terms_begin(): Don't return "" from the TermIterator (happened
4551 when the query contained or was Query::MatchAll).
4553 * Add QueryParser::FLAG_DEFAULT to make it easier to add flags to those set by
4558 * The testsuite now reports problems detected by valgrind with newer valgrind
4559 versions. Drop support for running the testsuite under valgrind < 3.3.0
4560 (well over a year old) as this greatly simplifies the configure tests.
4562 * Fix usage message for options which take arguments in --help output from test
4563 programs - "-x=foo" doesn't work, the correct syntax is "-x foo".
4565 * If comparing MSet percentages fails, report the differing percentages if in
4568 * Add test that backends don't truncate total document length to 32 bits.
4570 * Disable lockfileumask1 (regression testcase added in 1.0.10) on Cygwin and on
4575 * The configure test for pread() and pwrite() got accidentally disabled in
4576 0.8.4 and we've always been using llseek() followed by read() or write()
4577 since then. The configure test is now fixed, and gives a slight speedup
4578 (3% measured for searching).
4580 * The child process used to implement WritableDatabase locking now changes
4581 directory to / so that it doesn't block unmounting of any partitions and
4582 closes any open file descriptors which aren't relating to locking so that
4583 if those files are closed by our parent and deleted the disk space gets
4584 released right away.
4586 * We now reuse the same zlib zstream structures rather than using a fresh
4587 one for each operation. This doesn't make a measurable difference in
4588 our own tests on Linux but reportedly is measurably faster on some
4589 systems. (ticket #325)
4593 * The pread()/pwrite() fix also speeds up quartz.
4597 * Avoid copying Query::Internal objects needlessly when unserialising Query
4602 * Store the (non-normalised) document lengths as Xapian::termcount (unsigned
4603 int) rather than Xapian::doclength (double) which saves 4 bytes per document.
4607 * configure: The output of g++ --version changed format (again) with GCC 4.3
4608 which meant configure got "g++" for the version. Instead use the (hopefully)
4609 more robust technique of using g++ -E to pull out __GNUC__ and
4614 * API documentation:
4616 + WritableDatabase::flush() can't throw DatabaseLockError.
4618 + WritableDatabase's constructor can throw at least DatabaseCorruptError or
4621 + Document how to get all matches from Enquire::get_mset().
4623 + Other minor improvements.
4625 * docs/sorting.html: Clarify meaning.
4629 * Fix "#line" directives in generated file queryparser/queryparser_internal.cc
4630 to give a relative path - previously they had a full path when generated by a
4631 VPATH build (as release tarballs are), and this confused GCC 2.95 and
4634 * Fix for compiling with Sun's compiler (untested as we no longer have access
4637 Xapian-core 1.0.10 (2008-12-23):
4641 * Composing an OP_NEAR query with two non-term subqueries now throws
4642 UnimplementedError instead of AssertionError (in a --enable-assertions build)
4643 or leading to unexpected results (otherwise). This partly addresses bug#201.
4645 * Using a MultiValueSorter with no values set no longer causes a hang or
4646 segmentation fault (but it is still rather pointless!)
4650 * If we're using values for sorting and for another purpose, cache the
4651 Document::Internal object created to get the value for sorting, like we do
4656 * If the disk became full while flushing database changes to disk, the
4657 WritableDatabase object would throw a DatabaseError exception but be left in
4658 an inconsistent state such that further use could lead to the database on
4659 disk ending up in a "corrupt" state (theoretically fixable, but no tool
4660 to fix such a database exists). Now we try to ensure that the object is
4661 left in a consistent state, but if doing so throws a further exception, we
4662 put the WritableDatabase object in a "closed" state such that further
4663 attempts to use it throw an exception.
4665 * Create the lockfile "flintlock" with permissions 0666 so that the umask is
4666 honoured just like we do for the other files (previously we used 0600).
4667 Previously it wasn't possible to lock a database for update if it was
4668 owned by another user, even if you otherwise had sufficient permissions via
4671 * Fix garbled exception message when a base file can't be reread.
4675 * Fix garbled exception message when a base file can't be reread.
4679 * xapian-tcpsrv and xapian-progsrv now accept -w as a short form of --writable,
4680 as was always intended.
4684 * This release now uses newer versions of the autotools (autoconf 2.62 ->
4685 2.63; automake 1.10.1 -> 1.10.2).
4689 * INSTALL: Add new paragraphs about HP's aCC and IRIX (adapted from footnotes
4692 * PLATFORMS: HP testdrive has been shut down, so all mark all those machines as
4693 "no longer available". Update atreus' build report to 1.0.10.
4695 * docs/queryparser.html: Add link to valueranges.html.
4699 * delve: Add missing "and" to --help output. Report termfreq and collection
4700 freq for each term we're asked about.
4704 * Fix to build with GCC 4.4 snapshot.
4706 Xapian-core 1.0.9 (2008-10-31):
4710 * Database::get_spelling_suggestion() is now faster (15% speed up for parsing
4711 queries with FLAG_SPELLING_CORRECTION set in a test on real world data).
4713 * Fix OP_ELITE_SET segmentation fault due to excess floating point precision
4714 on x86 Linux (and possibly other platforms).
4716 * Database::allterms_begin() over multiple databases now gives a TermIterator
4717 with operations O(log(n)) rather than potentially O(n) in the number of
4720 * Add new Database methods metadata_keys_begin() and metadata_keys_end() to
4721 allow the complete list of metadata in a database to be retrieved (this
4722 API addition is needed so that copydatabase can copy database metadata).
4726 * Remove the cached test databases before running the testsuite.
4728 * apitest: Fix cursordelbug1 to work on Microsoft Windows (bug#301).
4730 * apitest,queryparsertest: Skip tests which fail because the timer granularity
4731 is too coarse to measure how long the test took. In practice, this is only
4732 an issue on Microsoft Windows (bug#300 and bug#308).
4736 * Adjust percent cutoff calculations in the matcher in a way which corresponds
4737 to the change to percentage calculations made in 1.0.7 to allow for excess
4740 * Query::MatchAll no longer gives match results ranked by increasing document
4745 * xapian-compact: Fix crash while compacting spelling table for a single
4746 database when built with MSVC, and probably other platforms, though Linux
4747 got lucky and happened to work (bug#305).
4751 * configure: Disable -Wconversion for now - it's not useful for older GCC and
4752 is buggy in GCC 4.3.
4754 * configure: Set -Wstrict-overflow to 1 instead of 5, to avoid unreasonable
4755 warnings under GCC 4.3.
4759 * Minor improvements to API documentation, including documenting the
4760 XAPIAN_FLUSH_THRESHOLD environmental variable in WriteableDatabase::flush()
4763 * valueranges.html: Fix typos in example code, and drop superfluous empty
4764 destructor from ValueRangeProcessor subclass.
4766 * HACKING: Several improvements.
4770 * copydatabase: Also copy user metadata.
4772 Xapian-core 1.0.8 (2008-09-04):
4776 * Fix output of RSet::get_description
4780 * Report subtotals per backend, rather than per testgroup per backend to make
4781 the output easier to read.
4785 * Fix WritableDatabase::add_document() and replace_document() not to be O(n*n)
4786 in the number of values in the new document.
4788 * Fix handling of a table created lazily after the database has had commits,
4789 and which is then cursored while still in sequential mode.
4791 * Fix failure to remove all the Btree entries in some cases when all the
4792 postings for a term are removed. (bug#287)
4794 * xapian-inspect: Show the help message on start-up. Correct the documented
4795 alias for next from ' ' to ''. Avoid reading outside of input string when it
4800 * Backport fix from flint for WritableDatabase::add_document() and
4801 replace_document() not to be O(n*n) in the number of values in the new
4806 * configure: Report bug report URL in --help output.
4808 * xapian-config: Report bug report URL in --help output.
4810 * configure: Fix deprecation error for --enable-debug=full to say to instead
4811 use '--enable-assertions --enable-log' not '--enable-debug --enable-log'.
4815 * valueranges.html: Expand on some sections.
4819 * quest: Fix to catch QueryParserError instead of const char * which
4820 QueryParser threw in Xapian < 1.0.0.
4822 * copydatabase: Use C++ forms of C headers. Only treat '\' as a directory
4823 separator on platforms where it is. Update counter every 13 counting up to
4824 the end so that the digits all "rotate" and the counter ends up on the exact
4829 * Eliminate literal top-bit-set characters in testsuite source code.
4831 Xapian-core 1.0.7 (2008-07-15):
4835 * OP_VALUE_RANGE, OP_VALUE_GE, and OP_VALUE_LE:
4837 + If there were gaps in the document id numbering, these operators could
4838 return document ids which weren't present in the database. This has been
4841 + These operators are now more efficient when there are a lot of "missing"
4842 document ids (bug#270).
4844 + Optimise Query(OP_VALUE_GE, <n>, "") to Query::MatchAll.
4846 * Xapian::QueryParser:
4848 + QueryParser now stops parsing immediately when it hits a syntax error.
4849 This doesn't change behaviour, but does mean failing to parse queries is
4852 + Cases of O(N*N) behaviour have been fixed.
4854 * Xapian::Stem now recognises "nl" as an alias for "dutch" (debian bug 484458).
4856 * Setting sort by value was being ignored by a Xapian::Enquire object which had
4857 previously had a Xapian::Sorter set (bug#256).
4861 * Improved test coverage in a few places.
4865 * When using a MatchDecider, we weren't reducing matches_lower_bound unless
4866 all the potential results were retrieved, which led to the lower bound
4867 being too high in some such cases.
4869 * We now track how many documents were tested by a MatchDecider and how many
4870 of those it rejected, and set matches_estimated based on this rate. Also,
4871 matches_upper_bound is reduced by the number of rejected documents.
4873 * Fixed matches_upper_bound in some cases when collapsing and using a
4876 * Fixed matches_lower_bound when collapsing and using a percentage cutoff.
4878 * When using two or more of a MatchDecider, collapsing, or a percentage
4879 cutoff, we now only round the scaled estimate once, and we also round it to
4880 the nearest rather than always rounding down. Hopefully this should
4881 improve the estimate a little in such cases.
4883 * Fix problem on x86 with the top match getting 99% rather than 100% (caused
4884 by excess precision in an intermediate value).
4888 * If Database::reopen() is called and the database revision on disk hasn't
4889 changed, then do as little work as possible. Even if it has changed, don't
4890 bother to recheck the version file (bug#261).
4894 + Fix check for user metadata key to not match other key types we may add in
4895 the future. When compacting, we can't assume how we should handle them.
4897 + If the same user metadata key is present in more than one source database
4898 with different tag values, issue a warning and copy an arbitrary tag value.
4900 + Fix potential SEGV when compacting database(s) with user metadata but no
4903 + In error message, refer to "iamflint" as the "version file", not the
4908 + Print top-bit-set characters as escaped hex forms as they often won't be
4909 valid UTF-8 sequences.
4911 + If we're passed a database directory rather than a single table, issue a
4912 special error message since this is an obvious mistake for users to make.
4914 * Fix cursor handling for a modified table which has previously only had
4915 sequential updates which usually manifested as zlib errors (bug#259).
4919 * Fix cursor handling for a modified table which has previously only had
4920 sequential updates which usually manifested as incorrect data being returned
4923 * Calling skip_to() as the first operation on an all-documents PostingIterator
4924 now works correctly.
4928 * Improve performance of matches with multiple databases at least one of which
4929 is remote, and when the top hit is from a remote database (bug#279).
4931 * When remote protocol version doesn't match, the error message displayed
4932 now shows the minor version number supplied by the server correctly.
4934 * We now wait for the connection to close after sending MSG_SHUTDOWN for a
4935 WritableDatabase, which ensures that changes have been written to disk
4936 and the lock released before the WritableDatabase destructor returns
4937 (as is the case with a local database).
4939 * We no longer ever send MSG_SHUTDOWN for a read-only Database - just closing
4940 the connection is enough (and is protocol compatible).
4944 * Fix bug which resulted in the values not being stored correctly when
4945 replacing an existing document, or if there are gaps in the document id
4950 * This release now uses newer versions of the autotools (autoconf 2.61 ->
4951 2.62; automake 1.10 -> 1.10.1; libtool 1.5.24 -> 1.5.26). The newer
4952 autoconf reportedly results in a faster configure script, and warns about
4953 use of unrecognised configure options.
4955 * Fix configure to recognise --enable-log=profile and fix build problems when
4958 * "make up" in the "tests" subdirectory now does "make" in the top-level.
4960 * Fix "make distcheck" by using dist-hook to install generated files from
4961 either srcdir or builddir, with the appropriate dependency to generate them
4962 automatically in maintainer mode builds.
4966 * intro_ir.html: Improve wording a bit.
4968 * The documentation now links to trac instead of bugzilla. For links to the
4969 main website, we now prefer xapian.org to www.xapian.org.
4971 * Doxygen-generated API documentation:
4973 + Improved documentation in several places.
4975 + The helper macro XAPIAN_VISIBILITY_DEFAULT no longer appears in the output.
4977 + Header and directory relationship graphs are no longer generated as they
4978 aren't actually informative here.
4980 * HACKING: Numerous updates and improvements.
4984 * quest: Output get_description() of the parsed query.
4988 * Fix build with GCC 2.95.3.
4990 * Fix build with GCC 4.3.
4992 * Newer libtool features improved support for Mac OS X Leopard and added
4993 support for AIX 6.1.
4997 * Database::get_spelling_suggestion() now debug logs with category APICALL
4998 rather than SPELLING, for consistency with all other API methods.
5000 * Added APICALL logging to a few Database methods which didn't have it.
5002 * Remove debug log tracing from get_description() methods since logging for
5003 other methods calls get_description() methods on parameters, so logging these
5004 calls just makes for more confusing debug logs. A get_description() method
5005 should have no side-effects so it's not very interesting even when explicitly
5008 Xapian-core 1.0.6 (2008-03-17):
5012 * Add new query operators OP_VALUE_LE and OP_VALUE_GE which perform "single
5013 ended" range checks, and a corresponding new Query constructor.
5015 * Add Unicode::toupper() to complement Unicode::tolower().
5017 * Xapian::Stem has been further optimised - stemtest now runs ~2.5% faster.
5021 * tests/runtest: Fixed to handle test programs with a ".exe" extension.
5023 * tests/queryparsertest: Add a couple more testcases which already work to
5024 improve test coverage.
5026 * tests/apitest: Add caseconvert1 testcase to test Unicode::tolower() and
5031 * xapian-check: Fix not to report an error for a database containing no
5032 postings but some user metadata.
5034 * Update the base files atomically to avoid problems with reading processes
5035 finding partially written ones.
5037 * Create lazy tables with the correct revision to avoid producing a database
5038 which we later report as "corrupt" (bug#232).
5040 * xapian-compact: Fix compaction for databases which contain user metadata
5045 * Update the base files atomically to avoid problems with reading processes
5046 finding partially written ones.
5050 * The addition of OP_VALUE_LE and OP_VALUE_GE required an update to the Query
5051 serialisation, which required a minor remote protocol version bump.
5053 * Fix to actually set the writing half as the connection as non-blocking when
5054 a timeout is specified. This would have prevented timeouts from operating
5055 correctly in some situations.
5059 * configure: GCC warning flag overhaul: Stop passing "-Wno-multichar" since
5060 any multi-character character literal is bound to be a typo (I believe we
5061 were only passing it after misinterpreting its sense!) Pass
5062 "-Wformat-security", and "-Wconversion" for all GCC versions. Add
5063 "-Winit-self" and "-Wstrict-overflow=5" for GCC >= 4.2. The latter might
5064 prove too aggressive, but seems reasonable so far. Fix some minor niggles
5065 revealed by "-Wconversion" and "-Wstrict-overflow=5".
5067 * Add XAPIAN_NORETURN() annotations to functions and non-virtual methods which
5072 * docs/intro_ir.html: Briefly mention how pure boolean retrieval is supported.
5074 * docs/valueranges.html: Fix example of using multiple VRPs to come out as a
5077 * include/xapian/queryparser.h: Fix incorrect example in doccomment.
5079 * docs/quickstart.html: Remove information covered by INSTALL since
5080 there's no good reason to repeat it and two copies just risks one
5081 getting out of date (as has happened here!)
5083 * docs/quickstart.html: Fix very out of date reference to MSet::items
5086 * PLATFORMS: Remove reports for 0.8.x as they're too old to be interesting.
5087 Separate out 0.9.x reports. Add Solaris 9 and 10 success reports from James
5088 Aylett. Update from Debian buildd logs.
5092 * Now builds on OS/2, thanks to a patch by Yuri Dario.
5094 * Fix testsuite to build on mingw (broken by changes in 1.0.5).
5098 * Fix --enable-assertions build, broken by changes in 1.0.5.
5100 Xapian-core 1.0.5 (2007-12-21):
5104 * More sophisticated sorting of results is now possible by defining a
5105 functor subclassing Xapian::Sorter (bug#100).
5107 * Xapian::Enquire now provides a public copy constructor and assignment
5110 * Xapian::Document::values_begin() didn't ensure that values had been read
5111 when working on a Document read from a database. However, values_end() did
5112 (and so did values_count()) so this wasn't generally a problem in practice.
5114 * Xapian::PostingIterator::skip_to() now works correctly when running over
5117 * Xapian::Database::postlist_begin() no longer adds a "MultiPostList" wrapper
5118 for the common case when there's only one subdatabase.
5120 * Xapian::TradWeight now avoids division by zero in the (rare) situation of the
5121 average document length being zero (which can only happen if all documents
5122 are empty or only have terms with wdf 0).
5124 * Calling Xapian::WritableDatabase methods when we don't have exactly one
5125 subdatabase now throws InvalidOperationError.
5131 + Testcases now describe the conditions they need to run, and are
5132 automatically collated by a Perl script. This makes it significantly
5133 easier to add a new testcase.
5135 + The test harness's "BackendManager" has been overhauled to allow
5136 cleaner implementations of testcases which are currently hard to
5137 write cleanly, and to make it easier to add new backend settings.
5139 + Add a "multi" backend setting which runs suitable tests over two
5140 subdatabases combined. There's a corresponding new make target
5143 + Add more feature tests of document values.
5145 + sortrel1 now runs for inmemory too.
5147 + Add simple feature test for TradWeight being used to run a query.
5149 + Fix spell3 to work on Microsoft Windows (bug#177).
5151 + API classes are now tested to check they have copy constructors and
5152 assignment operators, and also that most have a default constructor.
5154 + quartztest testcases adddoc2 and adddoc3 have been reworked as apitest
5155 testcases adddoc5 and adddoc6, which run for other backends.
5157 + stubdb1 now explicitly creates the database it needs - generally this
5158 bug didn't manifest because an earlier test has already created it.
5160 * queryparsertest: Add feature tests to check that ':' is being inserted
5161 between prefix and term when it should be.
5163 * Fix extracting of valgrind error messages in the test harness.
5165 * tests/valgrind.supp: Add more variants of the zlib suppressions.
5169 * Xapian::Enquire: When the "first" parameter to get_mset() is non-zero, avoid
5170 copying all the wanted items after performing the match.
5172 * Fix bug in handling a pure boolean match over more than one database under
5173 set_docid_order(ASCENDING) - we used to exit early which isn't correct.
5175 * When collapsing on a value, give a better lower bound on the number of
5176 matches by keeping track of the number of empty collapse values seen.
5178 * Xapian::BM25Weight: Fix bug when k2 is non-zero: a non-initialised value
5179 influenced the weight calculations. By default k2 is zero, so this bug
5180 probably won't have affected most users.
5182 * The mechanism used to collate term statistics across multiple databases has
5183 been greatly simplified (bug#45).
5189 + Update to handle flint databases produced by Xapian 1.0.3 and later.
5191 + Fix not to go into an infinite loop if certain checks fail.
5195 * quartzcompact: Fix equality testing of C strings to use strcmp() rather than
5196 '=='! In practice, using '==' often gives the desired effect due to pooling
5197 of constant strings, but this may have resulted in a bug on some platforms.
5201 * If we're doing a match with only one database which is remote then just
5202 return the unserialised MSet from the remote match. This requires an
5203 update to the MSet serialisation, which requires a minor remote protocol
5208 * XO_LIB_XAPIAN now hooks LT_INIT as well as AC_PROG_LIBTOOL and
5211 * Distribute preautoreconf, dir_contents, docs/dir_contents and
5214 * Fix preautoreconf to correctly handle all the sources passed to doxygen to
5215 create the collated internal source documentation, and to work in a VPATH
5220 * sorting.html: New document on the topic of sorting match results.
5222 * HACKING,admin_notes.html,bm25.html,glossary.html,intro_ir.html,overview.html,
5223 quickstart.html,scalability.html,termgenerator,html,synonyms.html: Assorted
5226 * valueranges.html: State explicitly that Xapian::sortable_serialise() is used
5227 to encode values at index time, and give an example of how it is called.
5229 * API documentation:
5231 + Clarify get_wdf() versus get_termfreq().
5233 + We now use pngcrush to reduce the size of PNG files in the HTML version.
5235 + The HTML version no longer includes various intermediate files which doxygen
5238 + Hide the v102 namespace from Doxygen as it isn't user visible.
5240 + Stop describing get_description() as an "Introspection method", as this
5241 doesn't help to explain what it does, and get_description() doesn't really
5242 fall under common formal definitions of "introspection".
5244 * index.html: Add a list of documents on particular topics and include links to
5245 previously unlinked-to documents. Weed down the top navigation bar which had
5246 grown to unwieldy length.
5248 * PLATFORMS: Update for Debian buildds.
5250 * Improve documentation comment for Document::termlist_count().
5252 * admin_notes.html: Note that this document is up-to-date for 1.0.5.
5254 * INSTALL: zlib 1.2.0 apparently fixes a memory leak in deflateInit2(), which
5255 we use, so that's another reason to prefer 1.2.x.
5259 * Add explicit includes of C headers needed to build with the latest snapshots
5260 of GCC 4.3. Fix new warnings.
5262 * xapian-config: On platforms which we know don't need explicit dependencies,
5263 --ltlibs now gives the same output as --libs.
5265 * The minimum supported GCC version is now 2.95.3 (rather than 2.95) as 2.95.3
5266 added support for '#include <sstream>' which means we no longer need to
5267 maintain our own version.
5269 * Fix build with SGI's compiler on IRIX.
5271 * Fix or suppress some MSVC warnings.
5275 * Remove incorrect assertion in MultiAndPostList (bug#209).
5277 * Fix build when configured with "--enable-log --disable-assertions".
5279 Xapian-core 1.0.4 (2007-10-30):
5285 + Add OP_SCALE_WEIGHT operator (and a corresponding constructor which
5286 takes a single subquery and a parameter of type "double"). This
5287 multiplies the weights from the subquery by the parameter, allowing
5288 adjustment of the importance of parts of the query tree.
5290 + Deprecate the essentially useless constructor Query(Query::op, Query).
5294 + A field prefix can now be set to expand to more than one term prefix.
5295 Similarly, multiple term prefixes can now be applied by default. This is
5296 done by calling QueryParser::add_boolean_prefix() or
5297 QueryParser::add_prefix() more than once with the same field name but a
5298 different term prefix (previously subsequent calls with the same field name
5301 + Trying to set the same field as probabilistic and boolean now throws
5302 InvalidOperationError.
5304 + Fix parsing of `term1 site:example.org term2', broken by changes in 1.0.2.
5306 + Drop special treatment for unmatched ')' at the start of the query, as it
5307 seems rather arbitrary and not particularly useful and was causing us to
5308 parse `(site:example.org) -term' incorrectly.
5310 + The QueryParser now generates pure boolean Query objects for strings such
5311 as `site:example.org' by applying OP_SCALE_WEIGHT with a factor of 0.0.
5313 + Fix handling of `"quoted phrase" +term' and `"quoted phrase" -term'.
5315 + Fix handling of `site:example.org -term'.
5317 + Fix problem with spelling correction of hyphenated terms (or other terms
5318 joined with phrase generators): the position of the start of the term
5319 wasn't being reset for the second term in the generated phrase, resulting
5320 in out of bounds errors when substituting the new value in the corrected
5323 + The parser stack is now a std::vector<> rather than a fixed size, so it
5324 will typically use less memory, and can't hit the fixed limit.
5326 + Fix handling of STEM_ALL and update the documentation comment for
5327 QueryParser::set_stemming_strategy() to explain how it works clearly.
5329 * PostingIterator: positionlist_begin() and get_wdf() should now always
5330 throw InvalidOperationError where they aren't meaningful (before in some
5331 cases UnimplementedError was thrown).
5335 * Add tests for new features.
5337 * Add another valgrind suppression for a slightly different error from zlib
5340 * Remove quartztest's test_postlist1 and test_postlist2, replacing the coverage
5341 lost by extending and adding tests which work with other backends as well.
5343 * If a test throws a subclass of std::exception, the test harness now
5344 reports the class name and the extra information returned by std::exception's
5349 * Several performance improvements have been made, mainly to the handling
5350 of OP_AND and related operations (OP_FILTER, OP_NEAR, and OP_PHRASE).
5351 In combination, these are likely to speed up searching significantly
5352 for most users - in tests on real world data we've seen savings of 15-55%
5353 in search times). These improvements are:
5355 + OP_AND of 3 or more sub-queries is now processed more efficiently.
5357 + Sub-queries from adjacent OP_AND, OP_FILTER, OP_NEAR, and OP_PHRASE are now
5358 combined into a single multi-way OP_AND operation, and the filters which
5359 implement the near/phrase restrictions are hoisted above this so they need
5360 to check fewer documents (bug#23).
5362 + If an OP_OR or OP_AND_MAYBE decays to OP_AND, we now ensure that the less
5363 frequent sub-query is on the left, which OP_AND is optimised to expect.
5365 * When the Enquire::get_mset() parameter checkatleast is set, and we're sorting
5366 by relevance with forward ordering by docid, and the query is pure boolean,
5367 the matcher was deciding it was done before the checkatleast requirement was
5368 satisfied. Then the adjustments made to the estimated and max statistics
5369 based on checkatleast meant the results claimed there were exactly msize
5370 results. This bug has now been fixed.
5372 * Queries involving an OP_VALUE_RANGE filter now run around 3.5 times faster
5375 * The calculations behind MSet::get_matches_estimated() were always rounding
5376 down fractions, but now round to the nearest integer. Due to cumulative
5377 rounding, this could mean that the estimate is now a few documents higher in
5378 some cases (and hopefully a better estimate).
5380 * Implement explicit swap() methods for internal classes MSetItem and ESetItem
5381 which should make the final sort of the MSet and ESet a little more
5386 * Fixed a bug introduced in 1.0.3 - trying to open a flint database for reading
5387 no longer fails if it isn't writable.
5389 * We no longer use member function pointers in the Btree implementation which
5390 seems to speed up searching a little.
5394 * The remote protocol minor version has been increased (to accommodate
5395 OP_SCALE_WEIGHT). If you are upgrading a live system which uses the
5396 remote backend, upgrade the servers before the clients.
5400 * Added macro machinery to allow branch prediction hints to be specified and
5401 used by compilers which support this (current GCC and Intel C++).
5403 * In a developer build, look for rst2html.py if rst2html isn't found as some
5404 Linux distros have it installed under with an extension.
5408 * In the API documentation, explicitly note that Database::get_metadata()
5409 returns an empty string when the backend doesn't support user-specified
5410 metadata, and that WritableDatabase::set_metadata() throws UnimplementedError
5411 in this case. Also describe the current behaviour with multidatabases.
5413 * README: Remove the ancient history lesson - this material is better left to
5414 the history page on the website.
5418 + Deprecate the non-pythonic iterators in favour of the pythonic ones.
5420 + Move "Stem::stem_word(word)" in the bindings to the right section (it was
5421 done in 1.0.0, as already indicated).
5423 + Improve formatting.
5425 * When running rst2html, using "--verbose" was causing "info" messages to be
5426 included in the HTML output, so drop this option and really fix this issue
5427 (which was thought to have been fixed by changes in 1.0.3).
5429 * install.html: Reworked - this document now concentrates on giving
5430 a brief overview of building which should be suitable for most common cases,
5431 and defers to the INSTALL document in each tarball for more details.
5433 * PLATFORMS: Update from tinderbox and buildbot.
5435 * remote.html: xapian-tcpsrv has been able to handle concurrent read
5436 access since 0.3.1 (7 years ago) so update the very out-of-date information
5437 here. Also, note that some newer features aren't supported by the remote
5440 * HACKING: Note specifically that std::list::size() is O(n) for GCC.
5442 * intro_ir.html: Add link to the forthcoming book "Introduction to
5443 Information Retrieval", which can be read online.
5445 * scalability.html: Update size of gmane.
5447 * quartzdesign.html: Note that Quartz is now deprecated.
5451 * The debug assertion code has been rewritten from scratch to be cleaner and
5452 pull in fewer other headers.
5454 Xapian-core 1.0.3 (2007-09-28):
5458 * Add support for user specified metadata (bug#143). Currently supported by
5459 the flint and inmemory backends.
5461 * Deprecate Enquire::register_match_decider() which has always been a no-op.
5463 * Improve the lower bound on the number of matching documents for an AND query
5464 - if the sum of the lower bounds for the two sides is greater than the
5465 number of documents in the database, then some of them must have both terms.
5467 * Spelling correction: Fix off-by-one error in loop bounds when initialising
5470 * If the check_at_least parameter to Enquire::get_mset() is used, but there
5471 aren't that many results, then MSet::get_matches_lower_bound() and
5472 MSet::get_matches_upper_bound() weren't always reported as equal - this
5475 * When sorting by value, and using the check_at_least parameter to
5476 Enquire::get_mset(), some potential matches weren't being counted.
5478 * Failing to create a flint or quartz database because we couldn't create the
5479 directory for it now throws DatabaseCreateError not DatabaseOpeningError.
5483 * Fix display of valgrind output when a test fails because valgrind detected
5486 * Add another version of valgrind suppression for the zlib end condition check
5487 as this gives a different backtrace for zlib in Ubuntu gutsy.
5491 * The Flint database format has been extended to support user metadata, and
5492 each termlist entry is now a byte shorter (before compression). As a
5493 result, Xapian 1.0.2 and earlier won't be able to read Xapian 1.0.3
5494 databases. However, Xapian 1.0.3 can read older databases. If you open an
5495 older flint database for writing with Xapian 1.0.3, it will be upgraded
5496 such that it cannot then be read by Xapian 1.0.2 and earlier.
5498 * Zlib compression wasn't being used for the spelling or synonym tables (due
5499 to a typo - Z_DEFAULT_COMPRESSION where it should be Z_DEFAULT_STRATEGY).
5501 * xapian-check: Allow "db/record." and "db/record.DB" as arguments.
5503 * Fix "key too long" exception message by substituting FLINT_BTREE_MAX_KEY_LEN
5504 with its numeric value.
5506 * Assorted minor efficiency improvements.
5508 * If we reach the flush threshold during a transaction, we now write out the
5509 postlist changes, but don't actually commit them.
5511 * Check length of new terms is at most 245 bytes for flint in add_document()
5512 and replace_document() so that the API user gets an error there rather
5513 than when flush() is called (explicitly or implicitly). Fixes bug#44.
5515 * Flint used to read the value of the environmental variable
5516 XAPIAN_FLUSH_THRESHOLD when the first WritableDatabase was opened and would
5517 then cache this value. However the program using Xapian may have changed
5518 it, so we now reread it each time a WritableDatabase is opened.
5520 * Implement TermIterator::positionlist_count() for the flint backend.
5524 * Fix the result of MSet::get_matches_lower_bound() when using the
5525 check_at_least parameter to get_mset().
5529 * Implement TermIterator::positionlist_count() for the inmemory backend.
5533 * xapian-config: We always need to include dependency_libs in the output of
5534 `xapian-config --libs` if shared libraries are disabled.
5536 * Distribution tarballs are now in the POSIX "ustar" format. This supports
5537 pathnames longer than 99 characters (which we now have a few instances of
5538 in the doxygen generated documentation) and also results in a distribution
5539 tarball that is about half the size! This format should be readable by any
5540 tar program in current use - if your tar program doesn't support it, we'd
5541 like to know (but note that the GNU tar tarball is smaller than the size
5542 reduction in the xapian-core tarball...)
5544 * configure no longer generates msvc/version.h - this is now entirely handled
5545 by the MSVC-specific makefiles.
5551 * docs/stemming.html: Reorder the initial paragraphs so we actually answer the
5552 question "What is a stemming algorithm?" up front.
5554 * When running rst2html, use "--exit-status=warning" rather than "--strict".
5555 The former actually gives a non-zero exit status for a warning or worse,
5556 while the former doesn't, but does include any "info" messages in the output
5559 * docs/deprecation.rst: Add "Database::positionlist_begin() throwing
5560 RangeError and DocNotFoundError".
5562 * valueranges.rst: Correct out-of-date reference to float_to_string.
5564 * HACKING: Document a few more "coding standards".
5566 * PLATFORMS: Updated.
5568 * docs/overview.html: Restore HTML header accidentally deleted in November
5571 * Fix several typos.
5575 * Add missing instances of "#include <string.h>" to fix compilation with recent
5578 * Fix some warnings for various compilers and platforms.
5580 Xapian-core 1.0.2 (2007-07-05):
5584 * Xapian now offers spelling correction, based on a dynamically maintained
5585 list of spelling "target" words. This is currently supported by the
5586 flint backend, and works when searching multiple databases.
5588 * Xapian now offers search-time synonym expansion, based on an externally
5589 provided synonym dictionary. This is currently supported by the flint
5590 backend, and works when searching multiple databases.
5592 * TermGenerator: now offers support for generating spelling correction
5597 + New flag FLAG_SPELLING_CORRECTION to enable spelling correction, and a new
5598 method, "get_corrected_query_string()" to get the spelling corrected
5601 + New flags have been added to allow the new synonym expansion feature to be
5602 enabled and controlled. Synonym expansion can either be automatic, or only
5603 for terms explicitly indicated in the query string by the new "~" operator.
5605 + The precedence of the boolean operators has been adjusted to match their
5606 usual precedence in mathematics and programming languages. "NOT" now binds
5607 as tightly as "AND" (previously "AND NOT" would bind like "AND", but just
5608 "NOT" would bind like "OR"!) Also "XOR" now binds more tightly than "OR",
5609 but less tightly than "AND" (previously it bound just like "OR").
5611 + '+' and '-' have been fixed to work on bracketed subexpressions as
5614 + If the stemmer is "none", no longer put a Z prefix on terms; this now
5615 matches the output of TermGenerator.
5617 * Add new Xapian::sortable_serialise() and Xapian::sortable_unserialise()
5618 functions which serialise and unserialise numbers (currently only
5619 doubles) to a string representation which sorts in numeric order. Small
5620 integers have a short representation.
5622 * NumberValueRangeProcessor has been changed to work usefully. Previously
5623 the numbers had to be the same length; now numbers are serialised to
5624 strings such that a string sort on the string orders the numbers correctly.
5625 Negative and floating point numbers are also supported now. The old
5626 NumberValueRangeProcessor is still present in the library to preserve
5627 ABI compatibility, but code linking against 1.0.2 or later will pick
5628 up the new implementation, which really lives in a sub-namespace.
5630 * Documents now have a get_docid() method, to get the document ID from the
5631 database they came from.
5633 * Add support for a new type of match decider, called a "matchspy". Unlike
5634 the old deciders, this will reliably be tested on every candidate
5635 document, so can be used to tally statistics on them.
5637 * Fixed a segfault when getting a description for a MatchNothing query
5638 joined with AND_NOT (bug #176).
5640 * Header files have been tidied up to remove some unnecessary includes.
5641 Applications using "#include <xapian.h>" will not be affected. We don't
5642 intend to support direct inclusion of individual header files from the xapian
5643 directory, but if you do that, you may have to update you code.
5647 * Feature tests added for all new features.
5649 * Improved test coverage in queryparsertest. Some tests in queryparsertest
5650 now use flint databases, so the test now ensures that the .flint
5651 subdirectory exists.
5653 * The test harness no longer creates <dbdir>/log for flint (flint doesn't
5654 create a log like quartz does).
5656 * apitest: "-bremote" must now be "-bremoteprog" (to better match
5657 "-bremotetcp"); "-bvoid" must now be "-bnone" (to better describe not
5658 using a database backend).
5660 * To complement "make check-flint", "make check-quartz", and "make
5661 check-remote", you can now run tests for the remotetcp backend with
5662 "make check-remotetcp", for the remoteprog backend with "make
5663 check-remoteprog", for the inmemory backend with "make check-inmemory", and
5664 tests not requiring a backend with "make check-none".
5666 * Several extra tests of the check_at_least parameter supplied to
5667 get_mset() were added.
5669 * Fix memory leak and fd leak in remotetcp handling, so apitest now passes
5672 * quartztest: no longer test QuartzPostList::get_collection_freq(), which
5675 * Add regression test emptyquery2 for bug #176.
5677 * Add regression test matchall1 for bug with MatchAll queries.
5679 * Enhanced test coverage of match functor, to check that it returns all
5684 * Fix bug when check_at_least was supplied - the matches after the
5685 requested MSet size were being returned to the user. The parameter is
5686 also now handled in a more efficient way - no extra memory is required
5687 (previously, extra memory proportional to the value of check_at_least was
5690 * Fix bug which used incorrect statistics, and caused assertion failures,
5691 when performing a search using a MatchAll query.
5693 * Optimisation for single term queries: we don't need to look at the top
5694 document's termlist to determine that it matches all the query terms.
5698 * The value and position tables are now only created if there is anything to
5699 add to them. So if you never use document values, there's no value.DB,
5700 value.baseA, or value.baseB. This means the table doesn't need to be opened
5701 for searching (saving a file handle and a number of syscalls) and when
5702 flushing changes, we don't need to update baseA/baseB just to keep the
5703 revisions in step. The flint database version has been increased, but the
5704 new code will happily open and read/update flint databases from Xapian 1.0.0
5705 and 1.0.1. Xapian 1.0.2 flint databases can't be read by Xapian 1.0.1 or
5708 * Two new optional tables are now supported: "spelling", which is used to
5709 store information for spelling correction, and "synonym", which is used
5710 to store synonym information.
5712 * xapian-compact: Now compacts and merges spelling and synonym tables.
5713 Also has a new option "--no-renumber" to preserve document ids from
5716 * xapian-check: Now checks the spelling and synonym tables (only the Btree
5717 structure is currently checked, not the information inside).
5719 * Database::term_exists(), Database::get_termfreq(), and
5720 Database::get_collection_freq() are now slightly more efficient for flint
5723 * New utility 'xapian-inspect' which allowing interactive inspection of key/tag
5724 pairs in a flint Btree. Useful for development and debugging, and an
5725 approximate equivalent to quartzdump.
5727 * WritableDatabase::delete_document() no longer cancels pending changes if the
5728 document doesn't exist.
5730 * Fix handling of exceptions during commit - previously, this could result
5731 in tables getting out-of-sync, perhaps even resulting in a corrupt database.
5733 * Optimise iteration of all documents in the case where all the document
5734 IDs up to lastdocid are used; in this case, we no longer need to access disk
5735 to get the document IDs.
5739 * WritableDatabase::delete_document() no longer cancels pending changes if the
5740 document doesn't exist.
5742 * We no longer create a postlist just to find the termfreq or collection
5747 * Calling WritableDatabase::delete_document() on a non-existent document now
5748 correctly propagates DocNotFoundError.
5750 * The minor remote protocol version has increased (to fix the previous issue).
5751 You should be able to cleanly upgrade a live system by upgrading servers
5752 first and then clients.
5754 * progclient: Reopen stderr on the child process to /dev/null rather than
5755 closing it. This fixes apitest with the remoteprog backend to pass when run
5756 under valgrind (it failed in this case in 1.0.0 and 1.0.1). It probably
5757 has no effect otherwise.
5759 * check_at_least is now passed to the remote server to reduce the work
5760 needed to produce the match, and the serialised size of the returned MSet.
5764 * Bug fix: using replace_document() to add a document with a specific
5765 document id above the highest currently used would create empty documents
5766 for all document ids in between.
5770 * Work around an apparent bug in automake which causes the entries in .libs
5771 subdirectories generated for targets of bin_PROGRAMS not to be removed on
5772 make clean. This was causing make distcheck to fail.
5774 * Snapshots and releases are now bootstrapped with automake 1.10, and
5777 * HTML documentation generated from RST files is now installed.
5781 * The API documentation is now generated with Doxygen 1.5.2, which fixes the
5782 missing docs for Xapian::Query.
5784 * Ship and install internals.html.
5786 * Generating the doxygen-collated documentation of the library internals (with
5787 "make doxygen_source_docs") now only tries to generate an HTML version. The
5788 PDF version kept exceeding TeX limits, and HTML is a more useful format for
5791 * API docs for Xapian::QueryParser now make it clear that the default value for
5792 the stemming strategy is STEM_NONE.
5794 * API docs now describe the NumberValueRangeProcessor more clearly.
5796 * Several typo fixes and assorted wording improvements.
5798 * queryparser.html: Mention "AND NOT" as an alternative way to write "NOT",
5799 and document synonym expansion.
5801 * admin_notes.html: Updated for changes in this release, and corrected a
5804 * spelling.rst: New file, documenting the spelling correction feature.
5806 * synonyms.rst: New file, documenting the synonyms expansion feature.
5808 * valueranges.rst: The NumberValueRangeProcessor is now documented.
5810 * HACKING: Mention new libtool, and more details about preferring
5811 pre-increment. Also add a note about 2 space indentation of protection
5812 level declarations in classes.
5814 * INSTALL: note that zlib must be installed before you can build.
5818 * copydatabase: Now copies synonym and spelling data. Also, fix a cosmetic
5819 bug with progress output when a specified database directory has a trailing
5824 * Fix to build on with OpenBSD's zlib (xapian-core 1.0.0 and 1.0.1 didn't).
5826 * Fixed to build with older zlib such as zlib 1.1.5 which Solaris apparently
5827 uses (xapian-core 1.0.0 and 1.0.1 didn't). However, we recommend using zlib
5828 1.2.x as decompressing is apparently about 20% faster.
5830 * msvc/version.h.in: Generated version.h for MSVC build no longer has the
5831 remote backend marked as disabled.
5833 * Fix warnings from Intel's C++ compiler.
5835 * Fixes for compilation with gcc-2.95 and GCC 4.3 snapshots.
5841 + Rename xapian.spec to xapian-core.spec to match tarball name.
5843 + Append the user name to BuildRoot.
5847 * Better debug logging from the queryparser internals.
5849 Xapian-core 1.0.1 (2007-06-11):
5855 + Make Error::error_string member std::string rather than char * to avoid
5856 problems with double free() with copied Error objects. Unfortunately
5857 this mean an incompatible ABI change which we had hoped to avoid until
5858 1.1.0, but in this case there didn't seem to be a sane way to fix the
5859 problem without an ABI change.
5861 + Error::get_description() now converts my_errno to error_string if it hasn't
5862 been already rather than not including any error description in this case.
5864 + Add new method "get_description()" to get a string describing the error
5865 object. This is used in various examples and scripts, improving their
5868 * Xapian::Database: Add new form of allterms_begin() and allterms_end()
5869 which allow iterating of all terms with a particular prefix. This
5870 is easier to use than checking the end condition yourself, and is
5871 more efficiently implemented for the remote backend (fixes bug#153).
5873 * Xapian::Enquire: Passing an uninitialised Database object to Enquire will
5874 now cause InvalidArgumentError to be thrown, rather than causing a segfault
5875 when you call Enquire::get_mset(). If you really want an empty database,
5876 you can use Xapian::InMemory::open() to create one.
5878 * Xapian::QueryParser: Multiple boolean prefixed terms with the same term
5879 prefix are now combined with OR before such groups are combined with AND
5880 (bug#157). Multiple value ranges on the same value are handled similarly.
5882 * Xapian::Query OP_VALUE_RANGE: Avoid calling db->get_lastdocid() repeatedly
5883 as we know the answer won't change - this reduces the run time of a
5884 particular test case by 25%.
5888 * Add test for serialisation of error strings.
5890 * Improved output in various situations:
5892 + Quote strings in TEST_STRINGS_EQUAL().
5894 + queryparsertest: Use TEST_STRINGS_EQUAL when comparing query descriptions
5895 against their expected output, since this makes it much easier to see the
5898 + Report whole message for exceptions, rather than a truncated version, in
5901 + Make use of Xapian::Error::get_description(), giving better error
5904 * queryparsertest: New test of custom ValueRangeProcessor subclass
5905 (qp_value_customrange1).
5907 * apitest: flintdatabaseformaterror1 and flintdatabaseformaterror2 now use a
5908 genuine Xapian 0.9.9 flint database for their tests, and more cases are
5909 tested. The two tests have also been split into 3 now.
5911 * Fix test harness not to invoke undefined behaviour in cases where a paragraph
5912 of test data contains two or fewer characters.
5914 * Implement a better fix for the MSVC ifstream issue which was fixed in 1.0.0.
5915 This fixes an unintentional side-effect of the previous fix which meant that
5916 apitest's consistency1 wasn't working as intended (it now has a regression
5917 test to make sure it is testing what we intend).
5921 * xapian-compact: Don't uncompress and recompress tags when compacting a
5922 database. This speeds up xapian-compact rather a lot (by more than 50% in a
5925 * If the docid counter wraps, Flint now throws DatabaseError (fixes bug#152).
5927 * Remove the special case error message for pre-0.6 databases since they'll
5928 be quartz format (the check is only in flint because this code was taken from
5933 * If the docid counter wraps, Quartz now throws DatabaseError (fixes bug#152).
5937 * The remote protocol now has a minor version number. If the major
5938 version number is the same, a client can work with any server with
5939 the same or higher minor version number, which makes upgrading live
5940 systems easier for most remote protocol changes - just upgrade the servers
5943 * When a read-only remote database is closed, the client no longer sends a
5944 (totally bogus) MSG_FLUSH to the server, and the reply is also eliminated.
5945 This reduces the time taken to close a remote database a little (fixes
5950 * skip_to() on an allterms TermIterator from an InMemory Database can no longer
5953 * An allterms TermIterator now initialises lazily, which can save some work if
5954 the first operation is a skip_to() (as it often will be).
5958 * Fix VPATH compilation in maintainer mode with gcc-2.95.
5960 * Fix multiple target rule for generating the queryparser source files in
5963 * Distribute missing stub Makefiles for "bin", "examples", and
5968 * Document the design flaw with NumberValueRangeProcessor and why it shouldn't
5971 * ValueRangeProcessor and subclasses now have API documentation and an overview
5974 * Expand documentation of value range Query constructor.
5976 * Improved API documentation for the TermGenerator class.
5978 * docs/deprecation.rst:
5980 + Fix copy and paste error - set_sort_forward() should be changed to
5983 + Improve entry for QueryParserError.
5985 * PLATFORMS: Updated from tinderbox.
5989 * copydatabase: Rewritten to use the ability to iterate over all the documents
5990 in a database. Should be much more efficient for databases with sparsely
5991 distributed document IDs.
5993 * simpleindex: Rewritten to use the TermGenerator class, which eliminates a
5994 lot of non-Xapian related code and is more typical of what a user is likely
5997 * simplesearch,simpleexpand: Rewritten to use the QueryParser class, which
5998 is more typical of what a user is likely to want to do.
6002 * xapian-config: Add special case check for host_os matching linux* or
6003 k*bsd-gnu since vanilla libtool doesn't correctly probe link_all_deplibs=no
6008 * RPMs: Add "# norootforbuild" comment which SuSE's build scripts look for.
6009 Rename "Source0:" to "Source:" as there's only one tarball now. Add gcc-c++
6010 and zlib-devel to "Build-Requires:".
6012 * The required automake version has been lowered to 1.8.3, so RPMs can now be
6013 built on RHEL 4 and SLES 9.
6015 Xapian-core 1.0.0 (2007-05-17):
6021 + The Database(const std::string &) constructor has been marked as "explicit".
6022 Hopefully this won't affect real code, but it's possible. Instead of
6023 passing a std::string where a Xapian::Database is expected, you'll now
6024 have to explicitly write `Xapian::Database(path)' instead of `path'.
6026 + Fixed problem when calling skip_to() on an allterms iterator over multiple
6027 databases which could cause a debug assertion in debug builds, and possible
6028 misbehaviour in normal builds.
6032 + The constructors of Error subclasses which take a `const std::string &'
6033 parameter are now explicit. This is very unlikely to affect any real code
6034 but if it does, just write `Xapian::Error(msg)' instead of `msg'.
6036 + Xapian::Error::get_type() now returns const char* rather than std::string.
6037 Generally existing code will just work (only one change was required in
6038 Xapian itself) - the simplest change is to write `std::string(e.get_type())'
6039 instead of `e.get_type()'.
6041 + Previously, the errno value was lost when an error was propagated from
6042 a remote server to the client, because errno values aren't portable
6043 between platforms. To fix this, Error::get_errno() is now deprecated and
6044 you should use Error::get_error_string() instead, which returns a string
6045 expanded from the errno value (or other system error code).
6047 * Xapian::QueryParser:
6049 + Now assumes input text is encoded as UTF-8.
6051 + We've made several changes to term generation strategy. Most notably:
6052 Unicode support has been added; '_' now counts as a word character; numbers
6053 and version numbers are now parsed as a single term; single apostrophes are
6054 now included in a term; we now store unstemmed forms of all terms; and we
6055 no longer try to "normalise" accents.
6057 + parse_query() now throws the new Xapian::Error subclass QueryParserError
6058 instead of throwing const char * (bug#101).
6060 + Pure NOT queries are now supported (for example, `NOT apples' will match
6061 all documents not indexed by the stemmed form of `apples'). You need
6062 to enable this feature by passing QueryParser::FLAG_PURE_NOT in flags
6063 to QueryParser::parse_query().
6065 + We now clear the stoplist when we parse a new query.
6067 + Queries such as `+foo* bar', where no terms in the database match the
6068 wildcard `foo*', now match no documents, even if `bar' exists. Handling
6069 of `-foo*' has also been fixed.
6071 + Now supports wildcarding the last term of a query to provide better support
6072 for incremental searching. Enabled by QueryParser::FLAG_PARTIAL.
6074 + The default prefix can now be specified to parse_query() to allow parsing
6075 of text entry boxes for particular fields.
6077 + QueryParser::set_stemming_options() has been deprecated since 0.9.0 and
6078 has now been removed.
6082 + Now assumes input text is encoded as UTF-8.
6084 + We've updated to the latest version of the Snowball stemmers. This means
6085 that a small number of words produce different (and generally better)
6086 stems and that some new stemmers are supported: german2 (like german but
6087 normalises umlauts), hungarian, kraaij_pohlmann (a different Dutch
6088 stemmer), romanian, and turkish.
6090 * Xapian::TermGenerator:
6092 + New class which generates terms from a piece of text.
6096 + The Enquire(const Database &) constructor has been marked as "explicit".
6097 This probably won't affect real code - certainly no Xapian API methods
6098 or functions take an Enquire object as a parameter - but calls to user
6099 methods or functions taking an Enquire object could be affected. In
6100 such cases, you'll now have to explicitly write `Xapian::Enquire(db)'
6103 + Enquire::get_eset() now produces better results when used with multiple
6104 databases - without USE_EXACT_TERMFREQ they should be much more similar to
6105 results from an equivalent single database; with USE_EXACT_TERMFREQ they
6106 should be identical.
6108 + Track the minimum weight required to be considered for the MSet separately
6109 from the minimum item which could be considered. Trying to combine the two
6110 caused several subtle bugs (bug#86).
6112 + Enquire::get_query() is now `const'. Should have no effect on user code.
6114 + Enquire::get_mset() now handles the common case of an "exact" phrase search
6115 (where the window size is equal to the number of terms) specially.
6117 + Enquire::include_query_terms and Enquire::use_exact_termfreq are now
6118 deprecated in favour of capitalised versions Enquire::INCLUDE_QUERY_TERMS
6119 and Enquire::USE_EXACT_TERMFREQ (for consistency with our other manifest
6120 constants, and general C/C++ conventions).
6124 + RSet::contains(MSetIterator) is now `const'. Should have no effect on user
6127 * Xapian::SimpleStopper::add() now takes `const std::string &' not `const
6128 std::string'. Should have no effect on user code.
6132 + We now only perform internal validation on a Query object when it's either
6133 constructed or changed, to avoid O(n^2) behaviour in some cases.
6135 + Xapian::Query::MatchAll (an alias for Query("")) matches all terms in the
6136 document (useful for "pure NOT" queries) and Xapian::Query:MatchNothing
6137 is now a more memorable alias for Query().
6139 * Instead of explicitly checking that a term exists before opening its
6140 postlist, we now do both in one operation, which is more efficient.
6142 * MatchDecider::operator() now returns `bool' not `int'.
6144 * ExpandDecider::operator() now returns `bool' not `int'.
6146 * Xapian::TermIterator::get_termfreq() now throws InvalidOperationError
6147 if called on a TermIterator from a freshly created Document (since
6148 there's no meaningful term frequency as there's no Database for
6151 * <xapian/output.h> is no longer available as an externally visible header.
6152 It's not been included by <xapian.h> since 0.7.0. Instead of using
6153 `cout << obj;' use `cout << obj.get_description();'.
6155 * New constant Xapian::BAD_VALUENO which is -1 cast to Xapian::valueno.
6157 * New Xapian::ValueRangeProcessor hierarchy: DateValueRangeProcessor,
6158 NumberValueRangeProcessor, and StringValueRangeProcessor. In
6159 conjunction with the new QueryParser::add_valuerangeprocessor()
6160 method and the new Query::OP_VALUE_RANGE op these allow you to
6161 implement ranges in the query parser, such as `$50..100',
6162 `10..20kg', `01/02/2007..03/04/2007'.
6166 * Many new and improved testcases in various areas.
6168 * If a test throws an unknown exception, say so in the test failure message.
6169 If it throws std::string, report the first 40 characters (or first line if
6170 less than 40 characters) of the string even in non-verbose mode.
6172 * Use of valgrind improved:
6174 + The test harness now only hooks into valgrind if environment variable
6175 XAPIAN_TESTSUITE_VALGRIND is set, which makes it easy to run test programs
6176 under valgrind in the normal way. The runtest script sets this
6179 + runtest now passes "--leak-resolution=high" to valgrind to prevent
6180 unrelated leak reports related to STL classes from being combined.
6182 + configure tests for valgrind improved and streamlined.
6184 + New runsrv script to run xapian-tcpsrv and xapian-progsrv. We need to
6185 run these under valgrind to avoid issues with excess numerical precision
6186 in valgrind's FP handling, but we can use "--tool=none" which is a lot
6187 faster than running them under valgrind's default memcheck tool.
6189 * The test harness now starts xapian-tcpsrv in a more reliable way - it will
6190 try sequentially higher port numbers, rather than failing because a
6191 xapian-tcpsrv (or something else) is already using the default port.
6192 It also no longer leaks file descriptors (which was causing later tests
6193 to fail on some platforms), and if xapian-tcpsrv fails to start, the error
6194 message is now reported.
6196 * remotetest has been removed and its testcases have either been added to
6197 apitest or just removed if redundant with tests already in apitest.
6199 * termgentest is a new test program which tests the Xapian::TermGenerator
6202 * TEST_EQUAL_DOUBLE() now uses a slightly less stringent threshold -
6203 DBL_EPSILON is too strict for calculations which include multiple
6204 steps. Also, we now use it instead of doubles_are_equal_enough() and
6205 weights_are_equal_enough() which try to perform the same job.
6207 * New macro TEST_STRINGS_EQUAL() which displays the strings on separate lines
6208 so the differences can be clearly seen.
6210 * Test programs are now linked with '-no-install' which means that libtool
6211 doesn't need to generate shell script wrappers for them on most platforms.
6213 * runtest: Now turns on MALLOC_CHECK_ and MALLOC_PERTURB_ for glibc if
6214 valgrind isn't being used.
6216 * Better support for Microsoft Windows:
6218 + test_emptyterm2 no longer tries to delete a database from disk while a
6219 WritableDatabase object still exists for it, since this isn't supported
6220 under Microsoft Windows.
6222 + Fallback handling when srcdir isn't specified how takes into account .exe
6223 extensions and different path separators.
6227 * Flint is now the default backend.
6229 * xapian-check: New program which performs consistency checks on a flint
6232 * xapian-compact: Now prunes unused docids off the start of each source
6233 database's range of docids.
6235 * Positional information is now encoded using a highly optimised fls()
6236 implementation, which is much faster than the FP code 0.9.x used.
6237 Unfortunately the old encoding could occasionally add extra bits
6238 on some architectures, which was harmless except the databases
6239 wouldn't be portable. Because of this, the flint format has had to
6240 be changed incompatibly.
6242 * The lock file is now called "flintlock" rather than "flicklock" (which
6245 * Flint now releases its lock correctly if there's an error in
6246 WritableDatabase's constructor. Previously the lock would remain until
6249 * Flint now throws new Xapian::Error subclass DatabaseVersionError instead of
6250 DatabaseOpeningError when it fails to open a database because it has an
6251 unsupported version. DatabaseVersionError is a subclass of
6252 DatabaseOpeningError so existing code should continue to work, but it's
6253 now much easier to determine if the problem is that a database needs
6256 * If you try to open a flint database with an older or newer version than
6257 flint understands, the exception message now gives the version understood,
6258 rather than "I only understand FLINT_VERSION" (literally).
6260 * If we fail to obtain the lock, report why in the exception message.
6262 * Flint now compresses tags in the record and termlist tables using zlib.
6264 * More robust code to handle the flint locking child process, in case of
6267 * If a document was replaced more than once between flushes, the document
6268 length wouldn't be updated after the first change.
6272 * Quartz is still supported, but use in new projects is deprecated (use Flint
6273 instead). Quartz will be removed eventually.
6275 * quartzcheck: Test if this is a quartz database by looking at "meta" not
6276 "record_DB". If "record_DB" is >= 2GB and we don't have a LFS aware stat
6277 function then stat can fail even though the file is there. Also open the
6278 database explicitly as a Quartz database for extra robustness.
6280 * If a document was replaced more than once between flushes, the document
6281 length wouldn't be updated after the first change.
6285 * The remote backend is now supported under Microsoft Windows.
6287 * Open a fresh copy of the database(s) on each connection to a xapian-tcpsrv
6288 rather than relying on being able to share a database across fork() or
6289 between threads (which we don't promise will work).
6291 * xapian-tcpsrv: New "--interface" option allows the hostname or address of the
6292 interface to listen on to be specified (the default is the previous behaviour
6293 of listening on all interfaces).
6295 * If name lookup fails, report the h_errno code from gethostbyname() rather
6296 than whatever value errno happens to currently have!
6298 * Fix bugs in query unserialisation.
6300 * The remote backend now supports all operations (get_lastdocid(), and
6301 postlist_begin() have now been implemented).
6303 * Currently a read-only server can be opened as a WritableDatabase (which is
6304 a minor bug we plan to fix). In this case, operations which write will fail
6305 and the exception is now InvalidOperationError not NetworkError.
6307 * If a remote server catches NetworkTimeoutError then it will now only
6308 propagate it if we can send it right away (since the connection is
6309 probably unhappy). After that (and for any other NetworkError) we now
6310 just rethrow it locally to close the connection and let it be logged if
6313 * The timeout parameter to RemoteDatabase wasn't being used, instead the
6314 client would wait indefinitely for the server to respond.
6316 * A timeout of zero to the remote backend now means "never timeout". This
6317 is now the default idle timeout for WritableDatabase (the connection
6318 timeout default is now 10 seconds, rather than defaulting to the idle
6321 * Fix handling of the document length in remote termlists.
6323 * The remote backend now checks when decoding serialised string that the
6324 length isn't more than the amount of data available (bug#117).
6326 * The remote backend now handles the unique term variants of delete_document
6327 and replace_document on the server side.
6329 * The RSet serialisation now encodes deltas between docids (rather than the
6330 docids themselves) which greatly reduces the size of the encoding of a
6331 sparse RSet for a large database.
6333 * We now encode deltas between term positions when sending data after calling
6334 positionlist_begin() on a remote database.
6336 * When using a MatchDecider with remote database(s), don't rerun the
6337 MatchDecider on documents which a remote server has already checked.
6339 * Apply the "decreasing weights with remote database" optimisation which we use
6340 in the sort_by_relevance case in the sort_by_relevance_then_value case too.
6342 * We now throw NetworkError rather than InternalError for invalid data received
6343 over the remote protocol.
6345 * We now close stderr of the spawned backend program when using the "prog" form
6346 of the remote backend. Previously stderr output would go to the client
6347 application's stderr.
6351 * Support for the old Muscat 3.6 backends has been completely removed. It's
6352 still possible to convert Muscat 3.6 databases to Xapian databases by
6353 building 0.9.10 and using copydatabase to create a quartz database, which can
6354 then be read by 1.0.0 (and converted to a flint database using copydatabase
6359 * We've added GCC visibility annotations to the library, which when using GCC
6360 version 4.0 or later reduce the size and load time of the library and
6361 increase the runtime speed a little. Under x86_64, the stripped library is
6362 6.4% smaller (1.5% smaller with debug information).
6364 * configure: If using GCC, use -Bsymbolic-functions if it is supported
6365 (it requires a very recent version of ld currently). This option reduces the
6366 size and load time of the shared library by resolving references within the
6367 library when it's created.
6369 * We automatically define _FORTIFY_SOURCE in config.h if GCC is in use
6370 and it's not already set (you can override this as documented in INSTALL).
6371 This adds some checking (mostly at compile time) that important return
6372 values aren't ignored and that array bounds aren't exceeded.
6374 * `./configure --enable-quiet' already allows you to specify at configure time
6375 to pass `--quiet' to libtool. Now you can override this at make-time by
6376 using `make QUIET=' (to turn off `--quiet') or `make QUIET=y' (to turn on
6379 * In non-maintainer mode, we don't need the tools required to rebuild some of
6380 the documentation, so speed up configure by not even probing for them in
6383 * The makefiles now use non-recursive make in all directories except "docs" and
6384 "tests". For users, this means that the build is faster and requires less
6385 disk space (bug#97).
6387 * configure: Add proper detection for SGI's C++ (check stderr output of
6388 "CC -v") and automatically pass -ptused in CXXFLAGS for xapian-core and any
6389 applications using xapian-config --cxxflags since it seems to be required to
6390 avoid template linking errors.
6392 * XO_LIB_XAPIAN now checks for the case where XAPIAN_CONFIG wasn't specified
6393 and xapian-config wasn't found, but the library appears to be installed -
6394 this almost certainly means that the user has installed xapian-core from
6395 a package, but hasn't installed the -dev or -devel package, so include
6396 that advice in the error message.
6398 * `./configure --with-stlport-compiler' now requires a compiler name as an
6401 * configure: Disable probes for f77, gcj, and rc completely by preventing
6402 the probe code from even appearing in configure - this reduces the size of
6403 configure by 209KB (~25%) and should speed it up significantly.
6405 * configure: Suppress more unhelpful warnings and "remarks" for HP's aCC, and
6406 turn on "+wlint", which seems useful.
6408 * A number of cases of unnecessary header inclusions have been addressed,
6409 which should speed up compilation (fewer headers to parse when compiling
6410 many source files). This also reduces dependencies within the source code,
6411 and thus the number of files which need to be rebuilt when a header is
6414 * configure: Cache the results of some of our custom tests.
6418 * The documentation has all been updated for changes in Xapian 1.0.0.
6420 * Many of the documentation comments in the API headers (which are collated
6421 using doxygen to generated the API reference) have been improved, and some
6422 missing ones added. Also, internal classes, members, and methods are now all
6423 marked as such so that none should appear in the generated documentation. In
6424 particular, the class inheritance graphs should be a lot clearer. A few other
6425 problems have also been addressed.
6427 * docs/internals.html: New separate index page for the "internal"
6430 * docs/deprecated.html: New document describing deprecation policy. This
6431 includes lists of features which have been removed, or which are deprecated
6432 and scheduled for removal, along with suggested replacements.
6434 * docs/admin_notes.html: New document introducing Xapian for sysadmins.
6436 * docs/termgenerator.html: New document describing the new term generation
6437 strategy implemented by the Term::Generator class.
6439 * docs/bm25.html,docs/intro_ir.html: These have been overhauled to make them
6440 fit better with the rest of the documentation, and with Xapian itself.
6442 * docs/overview.html: Fixed links to error classes in generated API
6445 * HACKING,INSTALL: Many updates and improvements.
6447 * xapian-config: Improve --version output so that help2man produces a better
6450 * PLATFORMS: Remove reports for 0.7.* and demote reports for 0.8.* to "older
6451 reports" status. All SF compilefarm machines are now "no longer available",
6452 so update the symbols and key to reflect this. Update with recent success
6453 reports from the tinderbox and other sources.
6455 * AUTHORS: Thanks several bug reporters I missed before, as well as recent
6458 * docs/code_structure.html now looks nicer and includes links to
6461 * docs/remote_protocol.html: Fixed several typos and other errors, and document
6462 all the new messages.
6464 * We no longer include docs/apidoc/latex/* in the xapian-core tarballs since
6465 it's just useless bloat.
6471 + Report the exception error string if open a database fails.
6473 + Rename "-k" to "-V" since "keys" were renamed to "values" long ago. Keep
6474 "-k" as an alias for now, but don't advertise it. Add handling so "-V3"
6475 shows value #3 for every document in the database.
6477 + No longer stems terms by default. Add "-s/--stemmer" option to allow a
6478 stemmer to be specified.
6480 * quest: Add "--stemmer" option to allow stemming language to be set, or
6481 stemming to be disabled.
6485 * Fix compilation with GCC 4.3 snapshot.
6487 * Always use pid_t not int for holding a process id, and use AC_TYPE_PID_T to
6488 `#define pid_t int' if <sys/types.h> doesn't provide pid_t.
6490 * Pass the 4th parameter of setsockopt() as char* which works whether the
6491 function actually takes char* or void* (since C++ allows implicit conversion
6492 from char* to void*).
6494 * Most warnings in the MSVC build have been fixed.
6496 * Refactored most portability workarounds into safeXXXX.h headers.
6498 * Building for mingw in a cygwin environment should work better now.
6504 + Updated for the changes in this release.
6506 + ChangeLog.examples is now packaged.
6510 * Rename --enable-debug* configure options - conflating the options to "turn on
6511 assertions" and "turn on logging" is confusing. `--enable-debug[=partial]'
6512 becomes `--enable-assertions'; `--enable-debug-verbose' becomes
6513 `--enable-log' and `--enable-debug=full' becomes `--enable-assertions
6514 --enable-log'. For now the old options give an error telling you the new
6517 * Debug logging from expand is now all of type EXPAND (some was of types
6518 MATCHER and WTCALC before).
6520 * Hook the debug tracing in the lemon generated parser into Xapian's debug
6523 * New assertion types: AssertEqParanoid() and AssertNeParanoid().
6525 * Retry write() if it fails when writing a debug log entry to ensure to avoid
6526 the risk of a partial write.
6528 Xapian-core 0.9.10 (2007-03-04):
6532 * Fix WritableDatabase::replace_document() not to lose positional information
6533 for a document if it is replaced with itself with unmodified postings.
6535 * QueryParser: Add entries to the "unstem" map for prefixed boolean filters
6538 * Fix inconsistent ordering of documents between pages with
6539 Enquire::set_sort_by_value_then_relevance (fixes bug#110).
6543 * Workaround apparent bug in MSVC's ifstream class.
6545 flint and quartz backends:
6547 * Fix possible double-free after a transaction fails.
6549 * Fix code for recovering from failing to open a table for reading
6550 mid-modification. If modifications are so frequent that opening for reading
6551 fails 100 times in a row, throw DatabaseModifiedError not
6552 DatabaseOpeningError.
6554 * Don't call std::string::append(ptr, 0) when ptr may be uninitialised
6555 or NULL (rather suspect, and reported to cause SEGV-like behaviour with
6558 * Ensure both_bases is set to false if we don't have both bases when
6559 opening a table using an existing object.
6561 * Use MS Windows API calls to delete files and open files we might want to
6562 delete while they are still open (i.e. the flint and quartz btree base
6563 files). This fixes a problem when a writer can't discard an old revision at
6564 the exact moment a reader is opening it (bug #108).
6568 * Fix WritableDatabase::has_positions() to refetch the cached value if it
6569 might be out of date.
6571 * Fix incorrect serialisation of a query with non-default termpositions.
6575 * If replace_document is used to set the docid of a newly added document which
6576 has previously existed, ensure we mark that document as valid.
6580 * Assorted improvements to API documentation.
6582 * docs/Makefile.am: The larger pool_size we set in 0.9.9 for building
6583 sourcedoc.pdf was a bit marginal, so increase it further.
6585 * docs/stemming.html,docs/install.html: Correct 2 references to "CVS" to say
6588 * HACKING: Update the release checklist.
6592 * Fix flint and quartz to allow 2GB+ B-tree tables when compiling with MSVC.
6596 * RPMs: Remove "." from end of "Summary:". Package the new man page for
6599 Xapian-core 0.9.9 (2006-11-09):
6603 * Use popen() to run xapian-tcpsrv and wait for "Listening..." before returning
6604 rather than just sleeping for 1 second and hoping that's enough.
6606 * If we can't start xapian-tcpsrv because the port is in use, try higher
6611 * xapian-tcpsrv: If the port requested is in use, exit with code 69
6612 (EX_UNAVAILABLE) which is useful if you're trying to automate launching of
6613 xapian-tcpsrv instances.
6615 * xapian-tcpsrv: Output "Listening..." once the socket is open and read for
6616 connections (this allows the testsuite to wait until xapian-tcpsrv is ready
6617 before connecting to it).
6619 * xapian-progsrv: Now supports --help, --version, and has a man page. Fixes
6622 * Turn on TCP_NODELAY for the TCP variant of the remote backend which
6623 dramatically improves the latency of operations on the database.
6627 * internaltest: Disable serialiselength1 and serialisedoc1 when the remote
6628 backend is disabled to fix build error in this case.
6630 * Move libbtreecheck.la from testsuite/ to backends/quartz/.
6632 * Move the testsuite harness from testsuite/ to tests/harness/.
6636 * Ship our custom INSTALL file rather than the generic one from autoconf which
6637 we've accidentally been shipping instead since 0.9.5.
6639 * docs/Makefile.am: Building sourcedoc.pdf needs a larger pool_size now we're
6642 * HACKING: Update debian packaging checklist.
6644 * PLATFORMS: Updated with results from tinderbox.
6648 * Create "safefcntl.h" as a replacement for <fcntl.h> instead of using
6649 "utils.h" for this purpose, since "utils.h" pulls in many other things we
6654 * RPMs: Prevent binaries getting an rpath for /usr/lib64 on FC6.
6656 Xapian-core 0.9.8 (2006-11-02):
6660 * QueryParser: Don't require a prefixed boolean term to start with an
6661 alphanumeric - allow the same set of characters as we do for the second
6662 and subsequent characters.
6666 * Only force a flush on WritableDatabase::allterms_begin() if there are
6667 actually pending changes.
6671 * Only force a flush on WritableDatabase::allterms_begin() if there are
6672 actually pending changes.
6674 * quartzcheck: Avoid dying because of an unhandled exception if the Btree
6675 checking code finds an error in the low-level Btree structure. Add a
6676 catch for any other unknown exceptions.
6680 * When building with GCC, turn on warning flag -Wshadow even when not in
6681 maintainer mode (provided it is supported by the GCC version being used).
6683 * testsuite/backendmanager.cc: Fix compilation when valgrind is detected by
6686 * If generating apidoc.pdf fails, display the logfile pdflatex generates since
6687 that is likely to show what failed.
6691 * Produce a PDF for apidoc rather than PostScript, since the PDF is smaller,
6692 plus at least as easy to print and easier to view for most users. Use
6693 pdflatex to generate the PDF directly rather than going via a DVI file which
6694 apparently produces a better result and also avoids problems on some Linux
6695 distros where latex is a symlink to pdfelatex (bug#81, bug#95).
6697 * HACKING: Mention automake 1.10 is out but we've not tested it yet.
6699 * HACKING: Add entries to release checklist: make sure new API methods
6700 are wrapped by the bindings, and that bug submitters are thanked.
6702 * HACKING: Note that on Debian, tetex-extra is needed for
6705 * HACKING: Note that dch can be used to update debian/changelog.
6707 * docs/code_structure.html: Document backends/remote.
6709 * PLATFORMS: Update from tinderbox.
6713 * configure: When checking if we need -lm, don't use a constant argument to
6714 log() as the compiler might simply evaluate the whole expression at compile
6717 * configure: Redhat's GCC 2.96 doesn't support -Wundef even though real GCC
6718 version before and after it do!
6720 * configure: Avoid use of double quotes in double-quoted backticks since
6721 it causes problems on some platforms.
6723 * backends/flint/flint_io.cc: Fix compilation on windows (needs to
6724 #include "safewindows.h" to get definition of SSIZE_T).
6726 * Fix our implementation of om_ostringstream to compile so that the build
6727 works once more on older compilers without <sstream> (regression probably
6728 introduced in 0.9.7).
6732 * xapian.spec: Package xapian-progsrv.
6734 Xapian-core 0.9.7 (2006-10-10):
6740 + Allow a distance to be optionally specified for NEAR - e.g.
6741 "cats NEAR/3 dogs" (bug#92).
6743 + Implement "ADJ" operator - like "NEAR" except the terms must
6744 appear in matching documents in the same order as in the query.
6746 + Fix bug in how we handle prefixed quoted phrases and prefixed brackets.
6748 + Fix parsing of loved and hated prefixed phrases and bracketted expressions.
6750 + Fix handling of stopwords in boolean expressions.
6752 + Don't ignore a stopword if it's the only query term.
6754 * Document::add_value() failed to replace an existing value with the same
6755 number, contrary to what the documentation says (bug #82).
6757 * Enquire::set_sort_by_value(): Don't fetch the document data when fetching
6758 the value to sort on. Simple benchmarking showed this to speed up sort by
6759 value by a factor of between 3 and 9!
6761 * Implement transactions for flint and quartz. Also supported are "unflushed"
6762 transactions, which provided an efficient way to atomically group a number
6763 of database modifications.
6765 * The Xapian::Error and Xapian::ErrorHandler classes have been reimplemented.
6766 The new versions have better, clearer documentation comments and are cleaner
6769 * Change how doubles are serialised by TradWeight, BM25Weight, and in the
6770 remote backend protocol. The new encoding allows us to transfer any double
6771 value which can be represented by both machines precisely and compactly.
6775 * Add targets "check-flint", "check-quartz", and "check-remote" in tests and at
6776 the top level which run the subset of tests which test the respective backend.
6778 * apitest: Run tests on flint if flint is enabled, rather than if quartz is
6781 * apitest: Speed up deldoc4 when run in verbose mode - some stringstream
6782 implementations are very inefficient when the string grows long.
6784 * Turn on GLIBCXX_FORCE_NEW when running tests under valgrind to stop the GNU
6785 C++ STL from using a pooling allocator. This helps make velgrind's leak
6786 tracking more reliable.
6788 * Probe for required valgrind logging options at configure time rather than
6789 when running the test program. This saves about 2 seconds per test program
6792 * Fix testsuite harness to show valgrind output when a test fails (when running
6793 under valgrind in verbose mode). This had stopped working, probably due to
6794 changes in valgrind 3.
6796 * internaltest: Check that the destructor on a temporary object gets called
6797 at the correct time (Sun C++ deliberately gets this wrong by default, and it
6798 would be good to catch any other compilers which do the same).
6800 * apitest: When running tests on the remote backend and running under valgrind,
6801 run xapian-tcpsrv and xapian-progsrv under valgrind too to avoid issues
6802 with the precision of doubles (bug#94).
6806 * Retry on EINTR from fcntl or waitpid when creating or releasing the flint
6809 * xapian-compact: Add --blocksize option to allow the blocksize to be set
6810 (default is 8K as before.)
6812 * WritableDatabase::replace_document(did, doc) was double-incrementing the
6813 "changes" counter when document did didn't exist so it would flush twice
6816 * WritableDatabase::postlist_begin(): Remove forced flush when iterating the
6817 posting list of a term which has modified postings pending.
6821 * quartzcompact: Add --blocksize option to allow the blocksize to be set
6822 (default is 8K as before.)
6824 * WritableDatabase::replace_document(did, doc) was double-incrementing the
6825 "changes" counter when document did didn't exist so it would flush twice
6830 * Most of the remote backend has been rewritten. It now supports most
6831 operations which a local database does (including writing!), the protocol
6832 used is more compact, and a number of layers of classes have been eliminated
6833 and the sequences of method calls simplified, so the code should be easier to
6834 understand and maintain despite doing more. A number of bugs have been fixed
6837 * xapian-tcpsrv: Report errno if we catch a Xapian::Error which has it set.
6839 * xapian-tcpsrv: Fix memory leak in query unserialisation.
6843 * Now using autoconf 2.60 for snapshots and releases. Also now using a
6844 libtool patch which improves support for Sun C++'s -library=stlport4 option.
6846 * configure: Fix generation of version.h to work with Solaris sed.
6848 * automake adds suitable rules for rebuilding doxygen_api_conf and
6849 doxygen_source_conf, so remove our less accurate versions. Also fix
6850 dependencies for regenerating the doxygen documentation, and make the
6851 documentation build work with parallel make.
6853 * Make use of the dist_ prefix to avoid having to list files in EXTRA_DIST as
6854 well as in *_DATA and man_MANS.
6856 * Removed a few unused #include-s.
6858 * include/xapian/error.h: Add hook to allow SWIG bindings to be built using
6859 GCC's visibility support.
6861 * configure: Turn on automake's -Wportability to help ensure our Makefile.am's
6862 are written in a portable way.
6864 * configure: Disable probing and short-cut tests for a FORTRAN compiler. We
6865 don't use one, but current libtool versions always check for it regardless.
6867 * xapian-config: Prune -L/usr/lib from output of `xapian-config --libs'.
6871 * docs/scalability.html: quartzcompact and xapian-compact now allow you to set
6872 the blocksize, so there's no need to use copydatabase if you want to migrate
6873 a database to a larger blocksize. Mention gmane. Other minor tweaks.
6875 * Eliminate "XAPIAN_DEPRECATED" from generated documentation.
6877 * PLATFORMS: Added success report for Nexenta (alpha 5), MSVC, and sparc linux.
6878 Updated other results from tinderbox.
6880 * Add links to the wiki from README and the documentation index.
6882 * docs/overview.html: Add discussion of uses of terms vs values.
6884 * docs/overview.html: Rewrite the section on Xapian::Document to remove some
6885 very out-of-date information and make it clearer.
6887 * include/xapian/database.h: Note that automatically allocated document IDs
6888 don't reuse IDs from deleted documents.
6890 * include/xapian/enquire.h: Note that "set_sort_by_relevance" is the default
6893 * docs/queryparser.html,include/xapian/queryparser.h: Add note that
6894 FLAG_WILDCARD requires you to call set_database.
6896 * HACKING: Add some advice regarding debugging using -D_GLIBCXX_DEBUG,
6899 * HACKING: Give URL to Alexandre Duret-Lutz's autotools tutorial, which is much
6900 more up-to-date than the "goat book".
6902 * HACKING: Update and expand the information about the debian packaging.
6904 * Add missing dir_contents files.
6908 * xapian/version.h: Add a check that _GLIBCXX_DEBUG is set compatibly if we're
6909 compiling with GNU C++ 3.4 or newer.
6911 * Add configure check to see if "-lm" is needed to get maths functions since
6912 newer versions of Sun's C++ compiler seem to require this.
6914 * Automatically put Sun's C++ compiler into "ANSI C++ compliant library" mode
6915 (using -library=stlport4). This allows us to remove most of the special
6916 case bits of code we've accumulated for just this compiler, which improves
6919 * Sun's C++ compiler implements non-standards-conforming lifetimes for
6920 temporary objects by default. This means database locks don't get released
6921 when they should, so we now always pass "-features=tmplife" for Sun C++
6922 which selects the behaviour specified by the C++ standard.
6924 Xapian-core 0.9.6 (2006-05-15):
6928 * Rename Xapian::xapian_version_string() and companions to
6929 Xapian::version_string(), etc. Keep the old functions as aliases which are
6930 marked as deprecated.
6932 * QueryParser: Add rules to handle a boolean filter with a "+" in front (such
6933 as +site:xapian.org).
6937 * queryparsertest: Add another prefix testcase to improve coverage.
6941 * configure: Simpler check for VALGRIND being set to empty value.
6943 * include/Makefile.am: Add xapian/version.h.timestamp as a dependency on
6944 all-local so that xapian/version.h actually gets regenerated when required.
6946 * Eliminate XAPIAN_BUILD_BACKEND_* from config.h and just use
6947 XAPIAN_HAS_*_BACKEND from xapian/version.h instead.
6951 * remote_protocol.html: Document keep-alive messages.
6953 * xapian/enquire.h: Remove bogus documentation for a parameter which doesn't
6956 * PLATFORMS: Added a summary. Updated and pruned old entries for which we
6957 have a newer close match.
6959 * HACKING: Expand on details of what's required when changing Xapian (discuss
6960 documentation requirements, and more on why feature tests are vital).
6962 * HACKING: Update section on building debian packages.
6966 * The tarball is generated with a patched version of libtool 1.5.22 which
6967 fixes libtool bugs on HP-UX and some BSD platforms.
6969 * configure: Fix problems with test for snprintf which affected cygwin, and
6970 possibly some other platforms.
6972 * configure: Tweak version.h generation to cope with CXXCPP putting carriage
6973 returns into its output as can happen on cygwin.
6975 * Fix renaming of "iamflint.tmp" for MS Windows where you can't rename an open
6978 * Fixed MSVC7 warnings.
6980 * Added workaround for newlib header bug.
6982 Xapian-core 0.9.5 (2006-04-08):
6988 + Fix FLAG_BOOLEAN_ANY_CASE to really allow any case combination - previously
6989 it only allowed all uppercase or all lowercase.
6991 + Fix QueryParser's handling of terms with trailing "#", "+", or "-" when
6992 set_database has been called and the term doesn't exist in the database
6995 * Add mechanism to allow xapian-bindings to override deprecation warnings so
6996 we can continue to wrap deprecated methods without lots of warnings.
6998 * Move Enquire::get_matching_terms_end() and Document::termlist_end() inline in
7001 * Database::termlist_begin(): Eliminate the MultiTermList wrapper in the common
7002 case where we're only dealing with a single database.
7004 * Fix TermIterator::positionlist_begin() to work on TermIterator from
7005 Database::termlist_begin(). Make TermList::positionlist_begin() pure
7006 virtual and put dummy implementations in BranchTermList and other
7007 subclasses which can't (or don't) implement it. This makes it hard to
7008 accidentally fail to implement it in a backend's TermList subclass.
7010 * TermIterator::positionlist_begin() with the remote backend now throws
7011 UnimplementedError instead of InvalidOperationError.
7013 * Implement Enquire::set_sort_by_relevance_then_value().
7017 * Added missing feature test for QueryParser::FLAG_BOOLEAN_ANY_CASE.
7019 * remotetest: Check mset size in tcpmatch1.
7023 * xapian-compact: Fixed segfault from passing an unknown option (e.g.
7024 "xapian-compact --foo").
7028 * quartzdump,quartzcompact: Fixed segfault from passing an unknown option
7029 (e.g. "quartzdump --foo").
7033 * xapian-tcpsrv: Don't perform a name lookup on the IP address which an
7034 incoming connection is from as that could easily slow down the search
7035 response - instead just print the IP address itself if output is verbose.
7037 * xapian-tcpsrv: Allow up to 5 connections in the listen queue instead of just
7042 * Removed unused code from the matcher and the remote, quartz, and flint
7047 * All installed binaries now support --help and --version and have a man page
7048 (which is generated using help2man).
7050 * docs/overview.html: Bring up to date.
7052 * docs/remote_protocol.html: Document messages for requesting and sending a
7053 termlist and a document.
7055 * PLATFORMS, AUTHORS: Updated.
7057 * INSTALL: Improve wording.
7059 * HACKING: Note that we now use a lightly patched version of libtool 1.5.22.
7061 * HACKING: aclocal is part of automake, not autoconf.
7065 * Added some tweaks to help support compilation with MSVC.
7069 * RPMs: package the new man pages.
7073 * Add missing spaces in some debug output.
7075 Xapian-core 0.9.4 (2006-02-21):
7079 * Flag deprecated methods such that the compiler gives a warning, for compilers
7080 which support such a feature (most notably GCC >= 3.1).
7082 * Correct typo in name of definition of function xapian_revision().
7086 * Updated uses of deprecated methods in the testsuite.
7090 * xapian-config: Set exec_prefix and prefix at top of script so that
7091 xapian-config works after xapian-core is installed.
7095 * Add documentation comment for Enquire::set_sort_by_value_then_relevance().
7097 * README: Add pointer to HACKING. Change "CVS access" to "SVN access".
7099 * PLATFORMS: Updated from tinderbox.
7101 * COPYING: Update second occurrence of old FSF address.
7103 Xapian-core 0.9.3 (2006-02-16):
7107 * Added 4 functions to report version information for the library version being
7108 used (which may not be the same as that compiled against if shared libraries
7109 are in use): xapian_version_string(), xapian_major_version(),
7110 xapian_minor_version(), xapian_revision().
7112 * Xapian::QueryParser:
7114 + Fix handling of "+" terms in a query when the default query operator is
7115 AND. Added regression test for this.
7117 + Added "AND NOT" as a synonym for "NOT". Added feature tests for this.
7119 * Fix prototype for ESet::operator[] to take parameter of type termcount
7120 instead of doccount (doccount and termcount are both typedefs to the same
7121 type so this really just makes the prototype more consistent).
7123 * Xapian::Stem: Check for malloc and calloc failing to allocate memory and
7124 throw an exception. Richard has fixed this upstream in snowball, so this is
7125 a temporary fix until we import a new version of snowball.
7127 * Xapian::Database: Trying to open a database for reading which doesn't exist
7128 now fails with DatabaseOpeningError instead of FeatureUnavailableError.
7129 Added regression test for this.
7131 * Add Stopper::get_description() and SimpleStopper::get_description().
7135 * Fixed testsuite harness to work with valgrind on 64 bit platforms.
7137 * Merged the "running tests" section of docs/tests.html into the similar
7138 section in HACKING, and make docs/tests.html refer the reader to HACKING for
7141 * Tidied and enhanced environmental variables which the test suite harness
7144 + OM_TEST_BACKEND: Removed support since the "-b" switch to apitest allows
7145 you control which backend is used, making OM_TEST_BACKEND pretty much
7148 + XAPIAN_SIG_DFL: Renamed to XAPIAN_TESTSUITE_SIG_DFL.
7150 + XAPIAN_TESTSUITE_OUTPUT: New environmental variable to control use of
7151 ANSI colour escape sequences in test output (set to "plain" to disable
7152 them, unset, empty, or "auto" to check if stdout is a tty, or anything
7153 else to force colour).
7157 * xapian-compact: Added "--multipass" option to merge postlists in pairs or
7158 triples until all are merged. Generally this is faster than an N-way merge,
7159 but it does require more disk space for temporary files so it's not the
7164 * quartzcheck: If the database is too broken to open, emit a warning message
7165 and bump the error count.
7169 * Now generate snapshots and releases with automake 1.9.6 (was 1.9.5) and
7170 libtool 1.5.22 (was 1.5.18).
7172 * configure: If not cross-compiling, try to actually run a test program built
7173 with the C++ compiler, not just link one.
7175 * configure: Fix to actually skip the check for valgrind if VALGRIND is set to
7178 * configure: Add sanity check for MS Windows that "find" is Unix-like find, not
7181 * Fix conditional compilation of flint backend - it was being disabled when
7182 quartz was, not when flint was supposed to be.
7186 * INSTALL,README: Updated.
7188 * Give pointer to replacements for the deprecated Enquire sorting methods
7189 in the doxygen collated documentation.
7191 * PLATFORMS: Added success reports for ppc64 linux and Fedora Core 4. Updated
7194 * HACKING: Note platforms valgrind now has solid support for; Improve
7195 phrasing in a few places.
7197 * Upgrade to using doxygen 1.4.6 for generating API documentation.
7199 * Change title of the "full source" documentation to "Internal Source
7200 Documentation" rather than "Full source documentation" to make it
7201 clearer it's only useful if you want to modify Xapian itself.
7203 * Fix documentation comments for the values of QueryParser::feature_flag so
7204 doxygen actually pulls out the documentation for them. Add documentation for
7205 the parameters of QueryParser::parse_query().
7207 * queryparser.html: Document wildcards.
7211 * Fix compilation with GCC 4.0.1 and later (need to forward declare class
7212 InMemoryDatabase) (bug #69).
7214 * Fix compilation under cygwin (broken in 0.9.2).
7216 * Don't pass NULL for the second parameter of execl() - the Linux man page
7217 says execl takes "one or more pointers to null-terminated strings". Also
7218 cast the NULL to (void*) to avoid "missing sentinel" warning from GCC4.
7220 * Use snprintf instead of sprintf where available (we were attempting to
7221 do this in some places before, but the configure test was broken so
7222 sprintf was always being used).
7224 * Enable more warnings under aCC and fix minor issues highlighted. Suppress
7225 "Entire translation unit was empty" warning which isn't useful to us.
7227 * Write top-bit set characters in the source using \xXX notation to avoid
7228 warnings from Intel's C++ compiler.
7230 * configure: TYPE_SOCKLEN_T fails hard, so only run it if we've successfully
7231 run other socket tests.
7233 * queryparser/accentnormalisingitor.h: #include <limits.h> for CHAR_BIT.
7235 * bin/xapian-compact.cc: Fix printf type mismatch on 64 bit platforms.
7237 * Replace pair<bool, string> with a simple class BoolAndString - the pair
7238 results in a 4328 byte symbol on HP-UX which gets truncated (to 4000 bytes).
7239 Most likely this is harmless, but it causes a warning.
7241 * configure: Disable flint backend by default if building for djgpp or msdos.
7243 * xapian-config: Previously when linking without libtool we've always thrown
7244 in dependency_libs, even though only some platforms need it (because it's
7245 generally pretty harmless). However some Linux distros have an unhelpful
7246 policy of not packaging .la files, so libxapian.la isn't available to
7247 extract dependency_libs from. Linux is a platform which doesn't require
7248 dependency_libs to be explicitly linked, so extend xapian-config to not
7249 pull in dependency_libs if libtool's link_all_deplibs_CXX=no.
7251 * xapian-config: If the current platform needs dependency_libs and
7252 libxapian.la's dependency_libs contains another .la file, transform it into a
7253 pair of -L and -l options, and recursively expand its dependency_libs (if
7256 * Don't pass functions with C++ linkage to places wanting pointers to functions
7257 with C linkage. So far this has worked for us, but it causes warnings with
7258 some compilers, and may not be portable.
7260 * Compaq C++ 7.1 doesn't suffer from the problem which previously prevented
7261 it from building Xapian. This release includes workarounds for some
7262 oddities with errno.h support in this compiler, but currently the build
7263 fails when trying to link a binary with the library.
7267 * RPM: Invoke %setup correctly in xapian.spec.
7271 * Add missing '#include <iostream>' when TIMING_PATCH is defined.
7273 Xapian-core 0.9.2 (2005-07-15):
7279 + Added optional "flags" argument to parse_query method.
7281 + Add flag FLAG_BOOLEAN_ANY_CASE which tells the QueryParser that boolean
7282 operators such as "AND", "OR", and "NEAR" should be recognised even if
7283 they aren't fully capitalised (so "and", "And", "aNd", etc will work too).
7285 + Add flag FLAG_WILDCARD which tells the QueryParser to allow right
7286 truncation e.g. "xap*".
7288 + Fixed to handle "-site:microsoft.com" where site is a boolean prefix.
7289 Added testcases for this.
7293 * The test harness was incorrectly creating a quartz database when a flint one
7294 was requested, which meant tests weren't being run against flint and so it
7295 had bugs rendering it pretty much unusable.
7297 * Added regression test longpositionlist1 (to check encoding/decoding a long
7298 position list, which flint had problems with).
7302 * Bumped format version number.
7304 * Added new "xapian-compact" program which can compact and merge flint
7305 databases in a similar way to how quartzcompact does for quartz databases.
7307 * Fixed to auto-detect database type when opening an existing Flint database
7308 as a WritableDatabase.
7310 * The code to encode the position list size, first entry, and last entry
7311 didn't match the code to decode them! Reworked both to match, using a
7312 slightly more compact encoding.
7314 * We were failing to append "DB" to the path when opening a table for reading.
7316 * Rewrite of FlintAllTermsList with several fewer member variables. The
7317 rewrite fixes a bug too - the old version wasn't ignoring the metainfo
7318 entry which is now in the postlist table.
7320 * It seems we need to explicitly kill the child process used for locking.
7321 Otherwise when we have two databases locked just closing the connection
7322 doesn't cause the child to die. I don't understand why it's needed, but this
7323 fix is at least clean.
7327 * quartzcompact: Fix mis-repacking of keys in positionlist table when merging
7330 * Disable assertion in allterms iteration which is incorrect in a corner case.
7331 This is only a problem if a termname contains zero bytes and you're using a
7332 debug build. Add regression test test_specialterms2.
7336 * Implement sorting on a value with the remote backend.
7340 * Pass automake options to AM_INIT_AUTOMAKE rather than specifying them in
7341 Makefile.am. This way, the version requirements for autoconf and automake
7342 are stated close together.
7344 * configure: -Wshadow causes false positives with GCC 3.0.4, so only enable it
7347 * configure: Eliminate use of "ln -s" when generating include/xapian/version.h
7348 since it seems to cause problems on Solaris in some setups and isn't really
7351 * Add dependency mechanism so version.h gets regenerated when the template is
7354 * configure: Check for spaces in build directory, source directory, or install
7355 prefix and die with a helpful message.
7357 * Add dependency to generate queryparser_token.h.
7359 * Eliminated TOP_SRCDIR and TOP_BUILDDIR - it's better to just use top_srcdir
7360 and top_builddir directly.
7362 * configure: Generate the list of source files to feed to doxygen by inspecting
7363 all the Makefile.am files prior to running autoreconf rather than by using
7364 "find" when the user runs ./configure. This speeds up configure, avoids
7365 generating docs for random .cc and .h files which aren't part of xapian-core,
7366 and avoids problems with picking up FIND.EXE on MS Windows.
7370 * Expanded explanation of the "descending docid with boolean weighting" trick
7371 for fast date ordered searching in Enquire::set_docid_order() API docs.
7373 * docs/intro_ir.html: Citeseer has moved, so update link.
7375 * testsuite/testsuite.cc: Update URL for valgrind FAQ in comment.
7377 * COPYING: Update FSF address.
7379 * HACKING: Minor updates to release checklist.
7383 * Assorted tweaks towards allowing compilation with MSVC.
7387 * xapian.spec.in: Package xapian-compact.
7389 Xapian-core 0.9.1 (2005-06-06):
7393 * Fix SEGV on get_terms_begin() on an empty Query object. This was causing
7394 a SEGV in Omega with an empty query.
7396 * Put Query::get_terms_end() inline in header.
7400 * Added the new "flint" backend, which starts out as a copy of the quartz
7401 backend plus some modifications and replacements. When creating a database
7402 without a specified backend, quartz is still used unless the environmental
7403 variable XAPIAN_PREFER_FLINT is set to a non-empty value.
7405 * apitest now runs tests on flint as well as the other backends.
7407 * Removed undocumented (and hence the little used) quartz "log" feature.
7409 * Implement new fork+fcntl+exec based locking (for Unix) and CreateFile based
7410 locking (for Windows - currently untested).
7412 * Move the special key/tag pair holding the total document length and doc id
7413 high water mark from the record table to the postlist table. This means that
7414 when appending documents, the insertion point will now always be at the end
7415 of the record table which is more efficient. We need to jump around the
7416 postlist table to merge postings in anyway.
7418 * Changed metafile magic to be different from quartz, and make the metafile
7419 version a datestamp which we'll change each time the format changes.
7421 * Check the return value of close() when writing the metafile.
7423 * Flint position list table now stores entries using interpolative coding
7424 (which is significantly more compact).
7428 * quartzcheck: Fixed corner case where you couldn't check a single Btree table
7429 which was just the DB and baseA/baseB files in a directory (Xapian doesn't
7430 produce anything like this, but btreetest does while unit testing the
7435 * Releases are now created using libtool 1.5.18 and automake 1.9.5.
7437 * configure: Pass more -W flags to g++ (including -Wundef which caught the
7438 getopt problem fixed in this release). Fixed new GCC warnings from these new
7441 * Fixed a lingering DOXYGEN_HAVE_DOT reference.
7443 * Fixed accidentally pruned #define which meant that getopt code was being
7444 included even on systems which use glibc (on such systems, we should use
7445 the glibc copy of the code instead).
7447 * queryparser/queryparser.lemony: Add missing '#include <config.h>'.
7451 * Added missing documentation comments for a QueryParser methods added in
7454 * docs/quartzdesign.html: Removed warning that quartz is still in development.
7456 * PLATFORMS: Updated from tinderbox.
7458 * configure: Describe CC_FOR_BUILD in configure --help output.
7460 * HACKING: Updated release instructions to refer to SVN, and note that release
7461 tarballs are now built specially rather than being copies of snapshots.
7462 Update information about the SVN tag name to use for debian files.
7464 * HACKING: Add "email Fabrice" to the release checklist so that RPM
7465 spec files don't lag behind.
7467 * Fixed a few spelling mistakes.
7471 * xapian.spec: Remove bogus %setup line left over from when we packaged
7472 xapian-core and xapian-examples together from separate tarballs.
7476 * api/omqueryinternal.cc: Fixed compilation with --enable-debug.
7478 * common/omdebug.h: Replace C style cast with static_cast<> which reveals that
7479 we were discarding const (harmlessly though).
7481 Xapian-core 0.9.0 (2005-05-13):
7485 * Query objects really need to be immutable after construction (otherwise we
7486 need a copy-on-write mechanism). To achieve this the following API changes
7489 + Remove Query::set_length() in favour of an optional length
7490 parameter to Enquire::set_query().
7492 + Eliminated Query::set_elite_set_size() in favour of optional parameter
7495 + Eliminated Query::set_window() in favour of an optional parameter to the
7498 * Removed OP_WEIGHT_CUTOFF, since it doesn't actually seem to add useful
7499 functionality over using Enquire::set_cutoff().
7501 * MSet::max_size() (which only exists so that MSet is an STL container) now
7502 returns MSet::size() and is inlined from the header.
7504 * Added ESet::max_size() (for STL compatibility).
7506 * Fixed Xapian::RSet to have the same "it's a handle" copy semantics as most of
7509 * Rewritten QueryParser class:
7511 + Uses Lemon instead of Bison to generate the parser, which enables us to
7512 stop using static data, so this class is at last reentrant.
7514 + QueryParser now uses a PIMPL style with reference counted internals like
7515 most of the other Xapian classes.
7517 + Direct access to member variables has gone, which unfortunately forces an
7518 API change (but this fixes bug #39). Instead of accessing
7519 QueryParser::termlist member variable, iterate over terms using
7520 Query::get_terms_begin() and get_terms_end() on the returned Query object.
7521 Direct access to stoplist is replaced by QueryParser::get_stoplist_begin()
7522 and get_stoplist_end(); and to unstem by get_unstem_begin() and
7525 + The rewrite parses many real world examples better than the old version.
7527 + Now allow searches for C#, etc. If a database has been set, for this and +
7528 and - suffixes, check if the term actually exists, and if not, ignore the
7529 suffix if the unsuffixed term exists.
7531 + Added QueryParser::get_description() method (not very descriptive yet!)
7533 + Added backward compatibility wrapper for old version of
7534 QueryParser::set_stemming_options().
7536 + xapian.h now automatically includes xapian/queryparser.h. Directly
7537 including xapian/queryparser.h will continue to work for now, but is
7540 + QueryParser::parse_query() was failing to clear termlist and unstem
7541 - the rewrite fixes this.
7543 + New QueryParser parses "term prefix:(term2 term3)" correctly.
7545 * Added Xapian::SimpleStopper which just stops terms specified by a pair of
7546 iterators. This should be sufficient for the majority of uses.
7548 * Tidied up the Enquire sorting API and added ability to reverse sort on a
7549 value. Removed sort_bands support.
7551 * Enquire::get_description() improved.
7553 * Methods which return an end iterator where the internals are just NULL are
7554 now inline in the header for efficiency. Should we ever need to change an
7555 implementation, we can easily move methods back into the library and bump the
7556 library version suitably.
7558 * Added Stem::operator() as preferred alternative to Stem::stem_word().
7560 * Simplified Stem internal design by restructuring to eliminate a few internal
7563 * BM25Weight: Avoid fetching document length if we're simply going to multiply
7568 * Fixed TEST_EQUAL_DOUBLE to use DBL_EPSILON correctly.
7570 * Rewrite of index_utils test harness code, removing unused and unusual
7571 features. Data files for tests are now easier to write. These changes
7572 also fix the bug that ^x didn't actually decode hex values correctly.
7574 * tests/testdata/etext.txt: Stripped carriage returns.
7576 * apitest: Extended stemlang1 to check that trying to create
7577 a stemmer for a non-existent language throws InvalidArgumentError.
7581 + Moved into tests/ subdirectory.
7583 + Reworked to use the standard testsuite harness.
7585 + Added tests for new features in the rewritten QueryParser.
7589 * quartzcheck: Now checks the structure of all the tables, not
7590 just the postlist table, and cross-checks doclen values between
7591 termlist and postlist tables. Recognises "--help" option. Should
7592 now continue after an error (typically it would crash before), and
7593 counts the number of errors found. Now exits with non-zero status
7594 if any errors were found. More readable output.
7596 * quartzcompact: Extended to allow merging several quartz
7597 databases to produce a single compact quartz database. This
7598 allows for faster building - simple index in chunks, then merge
7601 * quartzcompact: Made full compaction a tiny bit more compact.
7603 * quartzcompact: Added "fuller compaction" mode, which ignores the usual "at
7604 least 4 items per block" rule. This achieves slightly tighter compaction,
7605 though it's probably not advisable to use this option if you plan to update
7606 the compacted database.
7608 * Improved compaction by a few % in non-full case. Tighter bound on amount of
7609 memory to reserve to read the tag into.
7611 * Fix skip_to on an allterms TermIterator to set the current term when the
7612 skip_to-ed term is in the database. Add regression test for this
7615 * Values are stored in sorted order so we can stop unpacking the list once we
7616 get to one after the one we're looking for (in the case where the one we're
7617 looking for doesn't exist).
7621 * configure: Check that the C++ compiler can actually link a program.
7622 AC_LANG_CXX doesn't, and if it can't find a C++ compiler it'll just return
7623 "g++" which just leads to a later configure test failing in a confusing way.
7625 * configure: corrected configure output of "none known for yes" or "none known
7626 for no" to "none known for g++-3.2" or similar.
7628 * include/xapian/version.h: Define XAPIAN_HAS_xxx_BACKEND for each backend
7629 which is enabled. The bindings need this, and user code might find it useful
7632 * include/xapian/database.h: Don't declare the backend factory functions if the
7633 corresponding backend has been disabled. This means that trying to use a
7634 disabled backend will be caught at compile time rather than link time.
7636 * configure: Enhanced valgrind test to (a) see if --tool=memcheck
7637 is needed and (b) see if valgrind actually works (we don't want to
7638 try to use an x86 valgrind on an x86_64 box).
7640 * configure: Suppress 2 Intel C++ warnings which we can't easily code around,
7641 and enable -Werror automatically with --enable-maintainer-mode.
7643 * Clearer make rules for building Postscript doxygen docs.
7645 * Removed some no longer used code.
7647 * Moved a number of method definitions out of headers because they are virtual,
7648 or too large to be sensible candidates for inlining.
7650 * Eliminated the extra library for the queryparser - it's tiny compared to the
7651 main library and having it around just complicates things.
7653 * configure: We no longer need Bison, but we do need CC_FOR_BUILD to compile
7656 * Snapshot generator now appends _svn6789 or similar to the version string.
7657 Adjusted configure and XO_LIB_XAPIAN macro to take this into account.
7659 * configure: If any tools needed for documentation are missing
7660 and we're in maintainer mode, die with a suitable error in
7661 configure rather than with strange errors when building the
7664 * docs/Makefile.am: Explicitly set the pool_size for latex, because we
7665 now seem to overflow the default setting on some systems.
7667 * docs/Makefile.am: Use $(MAKE) instead of make.
7671 * Numerous improvements to documentation comments. Added documentation
7672 comments for QueryParser class.
7674 * HACKING: Added better description of how reference-counted API
7675 classes are structured.
7677 * HACKING: Note that '#include <limits>' isn't supported by GCC 2.95,
7678 and other assorted minor tweaks.
7680 * HACKING: Note how to disable use of VALGRIND on the make check
7681 command line, or when using runtest directly.
7683 * Updated all documentation mentions of CVS to talk about Subversion
7686 * PLATFORMS: Updated from tinderbox and other sources.
7688 * PLATFORMS: Added minimal testcase which fails to compile with
7689 Compaq's C++ compiler (cxx).
7691 * INSTALL,README: Updated.
7693 * docs/queryparser.html: Note that + and - work on phrases and
7694 bracketed expressions.
7696 * docs/intro_ir.html: Corrected two errors.
7698 * docs/stemming.html: Stemming appears to be applicable to Japanese
7699 so don't say it isn't!
7703 * Moved xapian-examples module to examples subdirectory of xapian-core.
7705 * quest: Added stopword handling.
7709 * configure: autoconf identifies Intel's C++ compiler as GCC, so probe for
7710 which we actually have.
7712 * Xapian will now compile cleanly with Intel C++ 8.1 on ia64 Linux and
7715 * backends/quartz/btree.cc: Fixed GCC compilation warning.
7717 * tests/api_db.cc: Fixed warning from Sun's C++ compiler.
7719 * configure: Automatically enable ANSI C++ mode for SGI's compiler
7720 with '-LANG:std'; check that any automatically determined flags
7721 for ANSI C++ mode actually allow us to compile a trivial program
7722 - if they don't it probably means the compiler isn't the one we
7723 were expecting, but one installed with the same name, so we now
7724 drop the flags in this case.
7726 * The compile on IRIX with SGI compiler is now warning free, apart from two
7727 "unused variable" warnings in Snowball generated code.
7729 * On WIN32, don't define NOMINMAX if it is already defined.
7733 * xapian.spec: Don't say "%makeinstall" in a comment since rpm
7734 tries to expand it and explodes.
7736 * xapian.spec: '/usr/share' -> '%{_datadir}'.
7738 * xapian.spec: Put the .so in the -devel package (it's only useful
7739 for linking to - the .so.* files are all that's needed at runtime).
7743 * net/socketserver.cc: Fixed typo in debug code.
7745 Xapian-core 0.8.5 (2004-12-23):
7749 * quartzcompact: When full_compaction is enabled, don't fill the last few bytes
7750 of a block if that would mean we needed an extra item and the overhead for
7751 that item would use up more of the next block than we save. This reduces the
7752 table size after full compaction by up to 0.2% in my tests!
7754 * quartzcompact: Tables sizes will always be a whole number of Kbytes, since
7755 the blocksize is, so report the size in K. Also report the change in size as
7756 well as the before and after sizes.
7758 * quartzcompact: Added missing '#include <config.h>' so that largefile support
7759 is enabled when we call stat() and we report compression statistics for
7762 * quartzcompact: Added --no-full / -n option to disable full compaction. This
7763 may be useful if you want to update the database after compacting it (need to
7764 test to see if this option is actually useful).
7766 * Renamed Btree::compress() to Btree::compact() for consistency with
7767 "full_compaction" and "quartzcompact". Also, "compress" is confusing since
7768 we use that term in the zlib patch.
7772 * xapian-config: Fixed --libs output to not include libxapian.la.
7774 * Added missing '#include <config.h>' to various .cc files (the omissions were
7775 probably harmless, but config.h should be included as the first thing any
7784 * RPM spec file: %makeinstall puts the wrong paths in the .la files so use
7785 "make DESTDIR=... install" instead.
7789 * Fixed to build with AssertParanoid enabled.
7791 Xapian-core 0.8.4 (2004-12-08):
7795 * Added constructors to Database and WritableDatabase which fulfil the role
7796 that the Auto::open() factory functions currently do. Auto::open() is
7799 * Removed the ability to write a Xapian object to an ostream directly, as
7800 it's little used and potentially dangerous ('cout << mset[i];' will
7801 compile, but you almost certainly meant 'cout << *mset[i];'). You can
7802 get the old effect by writing 'cout << obj->get_description();' instead
7803 of 'cout << obj;'. Note that including xapian.h no longer pulls in
7804 fstream, which code may have been implicitly relying on - if this is
7805 a problem add '#include <fstream>' after '#include <xapian.h>'.
7807 * QueryParser: Be smarter about when to add a ':' when adding a term prefix.
7809 * BoolWeight::unserialise() now returns BoolWeight*, and similarly for
7810 TradWeight and BM25Weight. BoolWeight::clone() now returns BoolWeight *.
7812 * If a database contains no positional information, change NEAR and PHRASE
7813 queries into AND queries (as otherwise they'd return no matches at all)
7814 (bug #56). Added feature test phraseorneartoand1.
7816 * Renamed BM25 parameters to match standard naming in papers and elsewhere
7817 (A->k3, B->k1, C->k2, D->b), eliminated the extra factor of 2 which our C
7818 had, and reordered the parameters to k1, k2, k3. This is an incompatible API
7819 change for BM25Weight(), so if you are using custom parameters for BM25
7820 you'll need to update your code.
7822 * During query expansion, if we estimate the term frequency, ensure it has a
7823 sane value (>= r and <= N - R + r) rather than bodging around the problem
7826 * TradWeight, BM25Weight: termfreq is always exact for matching (we only
7827 approximate it for query expansion) so replace code to work around bad
7828 approximations with Assert() to make sure this never happens.
7832 * runtest: Enhanced to allow it to run test programs under valgrind and other
7833 tools (gdb was already supported).
7835 * runtest: now works with valgrind 2.1.2 and later (valgrind's --logfile-fd
7836 option was renamed to --log-fd).
7838 * runtest: Allow VALGRIND environmental variable to override the value we got
7841 * Added a dependency so "make check" regenerates runtest if necessary.
7843 * The test programs now point the user to the runtest script if srcdir can't
7844 be guessed. And they no longer look for the test program in the tests
7845 subdirectory of the current directory.
7847 * btreetest: Fixed memory leaks in test_cursor1 (the testcase itself was
7848 causing the leak, not the library).
7850 * apitest: Fixed mset_range_is_same() and mset_range_is_same_weights() helper
7851 functions which were only comparing the first item in the range. Thankfully
7852 the tests still all pass so this wasn't hiding any bugs.
7854 * apitest: A modified version of changequery1 fails - the bug is obscure and
7855 subtle, and the fix is tricky so set the modified test to SKIP for now.
7857 * apitest: Added test_weight1 which tests the built-in Xapian::Weight
7858 subclasses and test_userweight1 which tests user defined weighting schemes
7861 * quartztest: Test with DB_CREATE_OR_OPEN in writelock1.
7865 * An interrupted update could cause any further updates to fail with "New
7866 revision too low" because the new revision was being calculated incorrectly -
7869 * Fixed Bcursor::del() which didn't always leave the cursor on the next item
7870 like it should. This may have been causing problems when trying to remove
7871 the last references to a particular term.
7873 * Fixed ultra-obscure bug in the code which finds a key suitable to
7874 discriminating between two blocks in a B-tree branch (discovered by reading
7875 the code). Comparing the keys didn't consider the length of the second, so
7876 it is possible the code would miscompare. But in reality this is extremely
7877 unlikely to happen, and even then would probably just mean that the
7878 discriminating key wouldn't be as short as it could be (wasting a few bytes
7879 but otherwise harmless).
7881 * If we're removing a posting list entirely, often there will only be one
7882 chunk, so avoid creating a Bcursor in this case.
7884 * Simplified Btree::compare_keys() by removing the last case which was dead
7885 code as it was covered by an earlier case.
7887 * Check that any user specified block size is a power of 2. If the block
7888 size passed is invalid, use the default of 8192 rather than throwing an
7891 * Started to refactor the Btree manager by introducing Item and Key classes
7892 which take care of handling the on-disk format, and eliminated duplicated
7893 tag reading code in Btree and Bcursor. These changes will pave the way for
7894 improvements to the on disk format.
7896 * Applied the Quartz "DANGEROUS" patch, but disabled for now. This way it
7897 won't keep being broken by changes to the code.
7899 * quartzcompact: Added --help and --version; Check that the source path and
7900 desitination path aren't the same; Report each table name when we start
7901 compacting it, and some simple stats on the compaction achieved when we
7906 * Removed a default parameter value from one variant of
7907 Xapian::Muscat36::open_db() so that there's only one candidate for
7912 * xapian-config: If flags are needed to select ANSI mode with the current
7913 compiler, then make xapian-config --cxxflags include them so that Xapian
7914 users don't have to jump through the same hoops we do.
7916 * xapian-config: Added --swigflags option for use with SWIG.
7918 * XO_LIB_XAPIAN now passes ac_top_srcdir to xapian-config which uses it
7919 (if provided) to say "configure.ac" or "configure.in" rather than
7920 "configure.in (or configure.ac)" in the "Add AC_PROG_LIBTOOL"
7923 * Cleaned up the build system in a few places.
7925 * Removed a few totally unneeded header includes.
7927 * Moved a number of functions and methods out of headers because they're not
7928 good inlining candidates (too big or virtual methods).
7930 * Changed C style casts to C++ style. The syntax is ugly, but they do make the
7931 intent clearer which is a good thing. Note this as a coding style guideline
7934 * configure.ac: Automatically add -Werror to CFLAGS and CXXFLAGS if
7935 maintainer mode is enabled and we're using GCC3 or newer. Don't do
7936 this for older GCCs as GCC 2.95 issues spurious warnings.
7938 * Reworked how include/xapian/version.h is generated so that it works
7939 better with compilers other than GCC, and with HP-UX sed.
7941 * XAPIAN_VERSION is now a string (e.g. "0.8.4").
7943 * Added new #define XAPIAN_REVISION (which is 4 for version 0.8.4).
7947 * docs/bm25.html,docs/intro_ir.html: Reworked to talk about Xapian
7948 rather than Muscat. Also improved the appearance of the formulae.
7950 * HACKING: Valgrind now supports x86 FreeBSD and PowerPC Linux.
7952 * Documented parameters of Enquire::register_match_decider().
7954 * We now use doxygen 1.3.8 to build documentation for snapshots and releases.
7956 * PLATFORMS: Updated from the tinderbox (which now runs builds on machines
7957 available in HP's testdrive scheme) and other assorted reports.
7959 * PLATFORMS: Removed reports from versions prior to 0.7.0. So much
7960 has changed that these are of little value.
7962 * docs/scalability.html: Added note warning about benchmarking from cold.
7964 * Assorted other minor documentation improvements.
7968 * configure.ac: Improved snprintf configure test to actually
7969 check that it works (older implementations may have different
7970 semantics for the return value, and at least one ignores the length
7971 restriction entirely!)
7973 * Reworked the GNU getopt source we use so that the header is clean and
7974 suitable for use from a reasonably ISO-conforming C++ compiler instead of
7975 being full of cruft for working around quirky C compilers which C++ compilers
7976 tend to stumble over.
7978 * Use SOCKLEN_T for the type we need to pass to various socket calls, since
7979 HPUX defines socklen_t yet wants int in those calls. Reworked the
7980 TYPE_SOCKLEN_T test we use.
7982 * On Windows, we want winsock2.h instead of sys/socket.h. Mingw doesn't seem
7983 to even have the latter, so I think previously we've been compiling by
7984 picking one up from somewhere random!
7986 * Change the small number of C sources we have to be C++ so we can compile
7987 everything with the C++ compiler. This way we don't need to worry about
7988 configure choosing a mismatching pair of compilers, or about whether
7989 configure tests with the C compiler don't apply to the C++ compiler, or vice
7992 * Compiles and passes testsuite with HP's aCC (we have to compile in
7993 ANSI mode, so we automatically add -AA to CXXFLAGS).
7995 * If the link test detects pread and pwrite are present, get configure to try
7996 out prototypes for pread and pwrite. This is much cleaner than trying to
7997 find the right combination of preprocessor defines to get each platform's
7998 system headers to provide prototypes.
8000 * configure: Disable probing for pread/pwrite on HP-UX as they're present but
8001 don't work when LFS (Large File Support) is enabled, and we definitely want
8004 * Fixed some warnings from Sun's C++ compiler.
8006 * Provide our own C_isalpha(), etc replacements for isalpha(), etc
8007 which always work in the C locale and avoid signed char problems.
8009 * For mingw/cygwin, pass -no-undefined when linking libxapianqueryparser.la
8010 so libtool builds a shared library. Also pass the magic linker flag
8011 -Wl,--enable-runtime-pseudo-reloc if configure has determined it is needed.
8013 * For cygwin, use the underlying MoveFile API call for locking, as link()
8014 doesn't work on FAT partitions. And don't rely on HAVE_LINK to control
8015 whether we use link() otherwise - if the configure test somehow misfires, a
8016 compilation error is better than using rename() on Unix as that would cause a
8017 second writer to smash the lock of the first.
8019 * Closer to building with Compaq C++ - add "-std strict_ansi" to CXXFLAGS, and
8020 tweaked the code in several places. It currently dies trying to compile
8021 the PIMPL smart pointer template code which looks hard to fix.
8025 * HACKING: Document that %% in XAPIAN_DEBUG_LOG is substituted with
8026 the process-id, and that setting XAPIAN_DEBUG_FLAGS to -1 enables
8029 * Removed compatibility code for checking environment variables OM_DEBUG_FILE
8032 Xapian-core 0.8.3 (2004-09-20):
8036 * Fixed bug which caused a segmentation fault or odd "Document not found"
8037 exceptions when new check_at_least parameter to Enquire::get_mset() was used
8038 and there weren't many matches (regression test checkatleast1).
8042 * Renamed omtcpsrv to xapian-tcpsrv and omprogsrv to xapian-progsrv.
8046 * RPM packaging now has a separate package for the runtime libraries to
8047 allow 32 and 64 bit versions to be installed concurrently.
8049 * RPM for xapian-core now includes binaries from xapian-examples.
8053 * Fixed to compile with debug tracing enabled.
8055 Xapian-core 0.8.2 (2004-09-13):
8059 * Removed the compatibility layer which allowed programs written against the
8060 pre-0.7.0 API to be compiled.
8062 * Added new ESet methods swap(), back() and operator[].
8064 * Xapian::WritableDatabase::replace_document can now be used
8065 to add a document with a specific docid (to allow keeping docids
8066 in sync with numeric UIDs from another system).
8068 * Added Xapian::WritableDatabase::replace_document and
8069 delete_document variants which take a unique id term name rather
8072 * Enquire::get_mset(): If a matchdecider is specified and no matches
8073 are requested, the lower bound on the number of matches must be 0
8074 (since the matchdecider could reject all the matches).
8076 * Renamed Query::is_empty() to Query::empty() for consistency. Keep
8077 Query::is_empty() for now as a deprecated alias.
8079 * Enquire::set_sorting() now takes an optional third parameter which allows
8080 you to specify a sort by value, then relevance, then docid instead of
8081 by value then docid.
8083 * Enquire::get_mset() now takes an optional "check_at_least" parameter
8084 which allows Omega's MIN_HITS functionality to be implemented in the matcher
8085 (where it can be done a bit more efficiently).
8089 * Reworked quartztest's positionlist1 into a generic api test as apitest's
8092 * apitest: Reenabled allterms2, but with the iterator copying parts removed -
8093 TermIterator is an input_iterator so that part was invalid.
8095 * Overhauled btreetest and quartztest - tests at the Btree level are now all
8096 in btreetest. Those at the QuartzDatabase level are in quartztest.
8098 * Split api_db.cc into 3 files as it has grown rather large.
8100 * tests/runtest: Added support for easily running gdb on a test program,
8101 automatically sorting out srcdir and libtool.
8105 * Refactored the quartz backend code to reduce the number of layered classes
8106 and eliminate unnecessary buffering, reducing memory usage so that more
8107 posting list changes can be batched together (see next change) and database
8108 building can be done several times faster.
8110 * Added tunable flush threshold - set XAPIAN_FLUSH_THRESHOLD=50000 to flush
8111 every 50000 documents. The default is now every 10000 documents (was
8112 every 1000 documents previously). The optimum value will most likely
8113 depend on your data and hardware.
8115 * WritableDatabase::get_document() no longer forces pending changes to be
8116 flushed. The document will read things lazily from the database, and that
8117 reading may trigger a forced flush).
8119 * WritableDatabase::get_avlength() no longer forces pending changes to be
8120 flushed. This means you can now search a modified WritableDatabase without
8121 causing a flush unless the search includes a term whose postlist has pending
8124 * Reduced quartz postlist chunk threshold from "2048 or a few bytes more" to
8125 "2000 or a few bytes more" so that full size chunks won't get split by the
8128 * Improved the "Db block overwritten" message. The DatabaseCorruptError
8129 version now suggests multiple writers may be the cause, while the
8130 DatabaseModifiedError version uses less alarming wording and says to call
8133 * QuartzWritableDatabase now stores the total document length and the last
8134 docid itself rather than tallying added and removed document length and
8135 writing the last docid back every time a document is added. This gives
8136 cleaner code and a small performance win.
8138 * Make the first key null for blocks more than 1 away from the leaves.
8139 It saves disk space for a tiny CPU and RAM cost so is bound to be
8142 * matcher/localmatch.cc: Fixed problems handling termweights in queries with
8143 the same term repeated (bug #37) and added regression test (qterminfo2).
8145 * Sped up iteration over all the terms in a database (QuartzCursor now only
8146 reads the tag from the Btree if asked to).
8148 * Cancelling an operation is now implemented more efficiently.
8152 * Fixed bugs with deleting a document while a PostingIterator over it is
8157 * Fixed to compile now that internal_end_session() has gone (broken in 0.8.1).
8161 * Fixed to compile when configured with --disable-inmemory (bug #33).
8163 * XO_LIB_XAPIAN now AC_SUBSTs XAPIAN_VERSION so your application's build
8164 system can easily check for a particular version of Xapian.
8166 * When compiling with GCC, we check that the compiler used to compile the
8167 library and the compiler used to compile the application have compatible
8168 C++ ABI versions. Unfortunately GCC 3.1 incorrectly reports the same
8169 ABI version as GCC 3.0, so we now special case that test.
8171 * Bumped the versions of the autotools we require for bootstrapping, and
8172 updated the documentation of these in the HACKING document.
8174 * Quote macro names to fix warnings from newer aclocal.
8178 * Improved API documentation for Xapian::WritableDatabase::replace_document and
8181 * Added documentation comments for MSet methods size(), empty(), swap(),
8182 begin(), end(), back().
8184 * Removed bogus documentation comments saying that some Enquire methods can
8185 throw DatabaseOpeningError.
8187 * Updated quartz design docs to reflect recent changes. Also pulled
8188 out the Btree and Bcursor API docs and slotted them in as doxygen
8189 documentation comments - this way they're much more likely to
8192 * Corrected multiple occurrences of "an Xapian::XXX" to "a Xapian::XXX"
8193 (presumably these all resulted from replacing "Om" with "Xapian::").
8195 * Various minor updates and improvements.
8199 * Reworked how we cope with fcntl.h #define-ing open on Solaris. This change
8200 finally allows Sun's C++ compiler to produce a working Xapian build on
8203 * configure.ac: Don't define DATADIR - we no longer use it and clashes
8204 with more recent mingw headers.
8206 * matcher/andpostlist.cc: Initialise lmax and rmax to 0. This cures
8207 the SIGFPE on apitest's qterminfo2 on alpha linux.
8209 Xapian-core 0.8.1 (2004-06-30):
8213 * New method Xapian::Database::get_lastdocid which returns the highest used
8214 document id for a database (useful for re-synchronizing an indexer which
8215 was interrupted). Implemented for quartz and inmemory.
8217 * Xapian::MSet::get_matches_*() methods now take collapsing into account, and
8218 the documentation has been clarified to state explicitly that collapsing and
8219 cutoffs are taken into account (bug#31).
8221 * Xapian::MSet: Need to adjust index by firstitem when indexing into items
8224 * MSetIterator and ESetIterator are now bidirectional iterators (rather than
8225 just input iterators)
8227 * Fixed post-increment forms of PostingIterator, TermIterator,
8228 PositionIterator, and ValueIterator so that *i++ works (as it must for them
8229 to be true input iterators).
8231 * Xapian::QueryParser: If we fail to parse a query, try stripping out
8232 non-alphanumerics (except '.') and reparsing.
8234 * Fixed memory leaked upon Xapian::QueryParser destruction.
8236 * Removed several unused Xapian::Error subclasses (these were used by the
8237 indexer framework which we decided was a failed experiment).
8241 * queryparsertest: Pruned near-duplicate queryparsertest testcases.
8243 * queryparsertest: Added test case for `term NOT "a phrase'.
8245 * remotetest: Use 127.0.0.1 instead of localhost so that tcpmatch1 doesn't fail
8246 just because the network setup is broken.
8248 * apitest: Make emptyquery1 check that Query("") causes an InvalidArgumentError
8253 * Fixed bug which meant we sometimes failed to remove a posting when deleting
8254 or replacing a document.
8256 * Fixed PostlistChunkReader to take a copy of the postlist data being read to
8257 avoid problems with reading data from a string that's been deleted.
8259 * Fixed bug in postlist merging which could occasionally extend a postlist
8260 chunk to overlap the docid range of the next chunk.
8262 * Eliminated the split cursor in each Btree object - we only actually need a
8263 single block buffer to handle splitting blocks. This reduces the memory
8264 overhead of each Bcursor (and hence each QuartzPostList).
8266 * Changed 2 calls to abort() to throw Xapian::DatabaseCorruptError instead,
8268 * If Btree is writable, throw DatabaseCorruptError if we detect overwritten.
8270 * Check the return value of fdatasync()/fsync()/_commit() and raise an error.
8271 If they fail, we really want to know as it could cause data corruption.
8273 * Assorted clean ups, improved comments, debug tracing, assertions.
8275 * When merging in postlist changes, removed an unneeded call to
8276 QuartzBufferedTable::get_or_make_tag() in a case when we're using a cursor
8277 which has already fetched the tag.
8279 * Added SON_OF_QUARTZ define to disable incompatible changes to database
8280 formats by default, and use it to control the docid encoding for keys such
8281 that we're always inserting at the end of the table when added new documents.
8283 * Reopening the readonly version of a writable Btree is now more efficient
8284 (we used to close and reopen all the files and destroy and recreate a lot
8285 of objects and buffers).
8287 * Share file descriptors between the read and write Btree objects so that a
8288 quartz WritableDatabase now uses 5 fds rather than 10.
8290 * Added configure test for glibc, because otherwise we need to include a header
8291 before we can check for glibc in order to define something we should be
8292 defining before we include any headers! Defining _XOPEN_SOURCE on OpenBSD
8293 seems to do the opposite to Linux and *disable* pread and pwrite!
8297 * Stripped out the session machinery - all that is actually required is to
8298 ensure that any unflushed changes are flushed when the destructor runs.
8300 * A few other backend interface cleanups.
8304 * Unified the shlib version numbers (the small benefit of tracking them
8305 individually makes it hard to justify the extra work required, and having one
8306 version simplifies debian packaging too).
8308 * configure.in: Fix typo (STLPORT_CXXLAGS -> STLPORT_CXXFLAGS)
8310 * Removed trivial m4/Makefile.am and autoconf/Makefile.am and do the work
8311 from the top level Makefile.am instead. It's easier to see the structure
8312 this way, and it also removes a couple of recursive make invocations which
8313 will speed up builds a little.
8317 * HACKING: Added a list of subtasks when doing a release.
8318 Currently it's always me that does this, but it may not always be
8319 and anyhow it'll help me to have a list to run through.
8321 * include/xapian/database.h: Remove references to sessions in doxygen
8324 * docs/quickstart.html: Corrected lingering reference to "om.h" and
8325 note that we need <iostream>.
8327 * docs/quickstartindex.cc.html,docs/quickstartexpand.cc.html,
8328 docs/quickstartsearch.cc.html: Add <iostream>.
8330 * PLATFORMS,AUTHORS: Updated.
8332 * docs/quartzdesign.html: Corrected various pieces of out of date
8333 information, and improved wording in a couple of places.
8335 * docs/scalability.html: Removed the reference to the Quartz update bottleneck
8336 "currently being addressed for Xapian 0.8" as it's now been addressed! Also
8337 reworded to remove use of first person (it was originally a message sent to
8340 Xapian-core 0.8.0 (2004-04-19):
8342 * Omega, xapian-examples and xapian-bindings now have their own NEWS files.
8346 * Throw an exception when an empty query is used to build in the binary
8347 operator Query constructor (previously this caused a segfault. Added
8350 * Made the TradWeight constructor explicit. This is technically an API change
8351 as before you could pass a double where a Xapian::Weight was required - now
8352 you must pass Xapian::TradWeight(2.0) instead of 2.0. That seems desirable,
8353 and it's unlikely any existing code will be affected.
8355 * Added "explicit" qualifier to constructors for internal use which take a
8358 * Renamed Xapian::Document::add_term_nopos to Xapian::Document::add_term
8359 (with forwarding wrapper method for compatibility with existing code).
8361 * The reference counting mechanism used by most API classes now handles
8362 creating a new object slightly more efficiently.
8364 * Xapian::QueryParser: Don't use a raw term for a term which starts with a
8369 * apitest, quartztest: Added a couple of tests, and commented out some test
8370 lines which fail in debug builds.
8372 * quartztest: cause a test to fail if there's still a directory after a call
8373 to rmdir(), or if there isn't a directory after calling mkdir().
8375 * apitest: Check returned docids are the expected values in a couple more
8376 cases. Improved wording of a comment.
8380 * We now merge a batch of changes into a posting list in a single pass which
8381 relieves an update bottleneck in previous versions.
8383 * When storing the termlist, pack the wdf into the same byte as the reuse
8384 length when possible - doing so typically makes the termlist 14% smaller!
8385 This change is backward compatible (0.7 database will work with 0.8, but
8386 databases built or updated with 0.8 won't work with 0.7).
8388 * quartzcheck: Check the structure within the postlist Btree as well as
8389 the Btree structures themselves.
8391 * Reduced code duplication in the btree manager and btreechecking code.
8393 * quartzdump: Backslash escape space and backslash in output rather than hex
8394 encoding them; renamed start-term and end-term to start-key and end-key;
8395 removed rather pointless "Calling next" message; if there's an error, write
8396 it to stderr not stdout, and exit with return code 1.
8398 * Corrected a number of comments in the source.
8400 * Removed several needless inclusions of quartz_table_entries.h.
8402 * Removed OLD_TERMLIST_FORMAT code - it has been disabled for since 0.6.0.
8404 * Removed all the quartz lexicon code and docs. It's been disabled for ages,
8405 and we've not missed it.
8409 * XO_LIB_XAPIAN autoconf macro can now be called without arguments in the
8410 common case where you want the test to fail if Xapian isn't found.
8412 * Fixed the configure test for valgrind - it wasn't working correctly when
8413 valgrind was installed but was too a version to support VALGRIND_COUNT_ERRORS
8414 and VALGRIND_COUNT_LEAKS.
8416 * GCC 2.95 supported -Wno-long-long and is our minimum recommended version, so
8417 unconditionally use -Wno-long-long with GCC, and don't test for it on other
8418 compilers (the old test incorrectly decided to use it with SGI's compiler
8419 resulting in a warning for every file compiled).
8423 * Updated the quickstart tutorial and removed the warning that "this
8424 document isn't up to date".
8426 * docs/intro_ir.html: Added a link to "Information Retrieval" by Keith van
8427 Rijsbergen which can be downloaded from his website!
8429 * docs/quartzdesign.html: Some minor improvements.
8431 * docs/matcherdesign.html: Merged in more details from a message sent to the
8434 * docs/queryparser.html: Grammar fixes.
8436 * Doxygen wasn't picking up the documentation for PostingIterator and
8437 PositionListIterator - fixed. Added doxygen comments for Xapian::Stopper
8438 and Xapian::QueryParser.
8440 * PLATFORMS: Updated with many results from tinderbox and from users.
8442 * AUTHORS: Updated the list of contributors.
8444 * HACKING: XAPIAN_DEBUG_TYPES should be XAPIAN_DEBUG_FLAGS.
8446 * HACKING: Updated to mention that building from CVS requires
8447 `./configure --enable-maintainer-mode' (or use bootstrap).
8449 * HACKING: Added notes about using "using", and pointers to a couple of useful
8454 * Solaris: Code tweaks for compiling with Sun's C++ compiler.
8456 * IRIX: Code tweaks for compiling with SGI's C++ compiler.
8458 * NetBSD mkdir() doesn't cope with a trailing / on the path - fixed our code to
8461 * mingw/cygwin: Only use O_SYNC (on the debug log) if the headers define it.
8463 * backends/quartz/quartz_table_manager.cc: Fix for building on mingw.
8465 * mingw: Added configure test for link() to avoid infinite loop in our C++
8468 * mingw and cygwin both need -Wl,--enable-runtime-pseudo-reloc passing when
8469 linking. Arrange for xapian-config to include this, and check that the ld
8470 installed is a new enough version (or at least that it was at configure
8471 time). Also pass to programs linked as part of the xapian-core build.
8473 * cygwin: Close a QuartzDatabase or QuartzWritableDatabase before trying to
8474 overwrite it - cygwin doesn't allow use to delete open/locked files...
8476 * backends/quartz/quartz_termlist.cc: Use Xapian::doccount instead of
8477 unsigned int in set_entries().
8479 * Database::Internal::Internal::keep_alive() should be
8480 Database::Internal::keep_alive().
8482 * Make Xapian::Weight::Weight() protected rather than private as we want to be
8483 able to call it from derived classes (GCC 3.4 flags this, other compilers
8488 * Open debug log with flag O_WRONLY so that we can actually write to it!
8490 * backends/quartz/quartz_values.cc: Fixed problem with dereferencing
8491 a pointer to the end of a string in debug output.
8493 Xapian 0.7.5 (2003-11-26):
8497 * Xapian::QueryParser now supports prefixes on phrases and expressions (e.g.
8498 author:(twain OR poe) subject:"space flight").
8500 * Added missing default constructors for TermIterator, PostingIterator, and
8501 PositionIterator classes.
8503 * Fixed PositionIterator assignment operator.
8507 * queryparsertest: Added testcase for new phrase and expression prefix support.
8509 * apitest: Added regression tests for API fixes.
8513 * quartzcompact: Fix the name that the meta file gets copied to (was
8514 /path/to/dbdirmeta rather than /path/to/dbdir/meta).
8518 * Changed to using AM_MAINTAINER_MODE. If you're doing development work on
8519 Xapian itself, you should configure with "--enable-maintainer-mode" and
8520 ideally use GNU make.
8522 * Fixed configure test for fdatasync to work (I suspect a change in a recent
8523 autoconf broke it as it relied on autoconf internal naming).
8525 * Fully updated to reflect move of libbtreecheck.la from backends/quartz
8526 to testsuite. btreetest and quartzcheck should build correctly now.
8530 * Added first cut of documentation for Xapian::QueryParser query syntax.
8532 * Fixed incorrectly formatted doxygen documentation comments which resulted in
8533 some missing text in the collated API and internal classes documentation.
8535 * Documented --enable-maintainer-mode and problems with BSD make in HACKING.
8537 * Fixed typo in docs/scalability.html.
8539 * PLATFORMS: Updated from the tinderbox.
8543 * omega: Parsing of the probabilistic query is now delayed until we need some
8544 information from it. This means that we can now use options set by the
8545 omegascript template to control the behaviour of the query parser.
8546 $set{stemmer,...} now controls the stemming language (e.g. $set{stemmer,fr})
8547 and $setmap{prefix,...} now sets the QueryParser prefix map (e.g.
8548 $setmap{prefix,subject,XT,abstract,XA}).
8550 * omega: Fixed $setmap not to add bogus entries.
8552 * docs/omegascript.txt: Expanded documentation of $set and $setmap to list
8553 values which Omega itself makes use of.
8555 * omega: Cleaned up the start up code quite a bit.
8557 * omega: Removed the unfinished code for caching omegascript command
8558 expansions. Added code to cache $dbsize. The only other value correctly
8559 marked for caching is already being cached!
8561 Xapian 0.7.4 (2003-10-02):
8565 * Fixed small memory leak if Xapian::Enquire::set_query() is called more than
8568 * Xapian::ESet now has reference counted internals (library interface version
8569 bumped because of this).
8571 * Removed unused OmDocumentTerm::termfreq member variable.
8573 * OmDocumentTerm ctor now takes wdf, and replaced set_wdf() with inc_wdf() and
8576 * Removed unused open_document() method from SubMatch and derived classes.
8578 * Calls made by the matcher to Document::Internal::open_document() now use the
8579 lazy flag provided for precisely this purpose, but apparently never used -
8580 this should give quite a speed boost to any matcher options which use values
8581 (e.g. sort, collapse).
8585 * Finished off support for running tests under valgrind to check for memory
8586 leaks and access to uninitialised variables.
8588 * apitest: Sped up deldoc4.
8590 * btreetest: Removed superfluous `/'s from constructed paths.
8592 * quartztest: adddoc2 now checks that there weren't any extra values created.
8596 * quartz: don't start the document's TermIterator from scratch on every
8597 iteration in replace_document(). Should be a small performance win.
8599 * quartz: Pass 0 for the lexicon/postlist table when creating a termlist just
8600 to find the doc length.
8602 * quartz: quartz_table_entries.cc: Removed rather unnecessary use of
8605 * quartz: quartz_table.cc: Removed unused variable.
8607 * quartz: Improved encapsulation of class Btree.
8611 * libbtreecheck.la now has an explicit dependency on libxapian.la.
8613 * We now set the dependencies for libxapian correctly so that linking
8614 applications will pull in other required libraries.
8616 * matcher/Makefile.am: Ship networkmatch.cc even if "make dist" is run from a
8617 tree with the remote backend disabled.
8619 * configure.in: Sorted out tests for gethostbyname and gethostbyaddr using
8620 standard autoconf macros.
8622 * configure.in: If fork is found, but socketpair isn't, automatically disable
8623 the remote backend rather than configure dying with an error.
8625 * autoconf/: Removed various unused autoconf macros.
8629 * xapian-config.in: Link with libxapianqueryparser before libxapian, since
8630 that's the dependency order.
8632 * Removed or replaced uses of <iostream> and <iosfwd> in the library sources
8633 - we don't need or want the library to pull in cin and friends.
8635 * extra/queryparser.yy: Fixed to build with Sun's C++ compiler.
8637 * Make the dummy source file C++ rather than C so that automake tells libtool
8638 that this is a C++ library - vital for correct linking on some platforms.
8640 * Makefile.am: Pass -no-undefined to libtool so that we can build build a DLL
8643 * configure.in: Fixed check for socketpair - we were automatically disabling
8644 the remote backend on platforms where socketpair is in libsocket
8647 * Use O_BINARY for binary I/O if it exists.
8649 * common/utils.h: mkdir() only takes one argument on mingw.
8651 * common/utils.h,testsuite/backendmanager.cc: Touch file using open() rather
8654 * common/utils.cc: Fixed to compile if snprintf isn't available.
8658 * docs/scalability.html: Fixed slip (32GB should be 32TB); Added note about
8659 Linux 2.4 and ext2 filesize limits.
8661 * PLATFORMS: Updated.
8663 * NEWS: Fixed a few typos.
8667 * xapian.i: using namespace std in SWIG parsed segment to sort out typemaps.
8671 * Updated RPM packaging.
8675 * omega: $topdoc now ensures the match has been run; $date no longer ensures
8676 the match has been run.
8678 * omega: Fixed to build with Sun's C++ compiler.
8680 Xapian 0.7.3 (2003-08-08):
8684 * MSetIterator: Fixed MSetIterator::get_document() to work when get_mset() was
8685 called with first != 0 (regression test msetiterator3).
8689 * internaltest: Changed test exception1 to actually test something (hopefully
8690 what was originally intended!)
8692 * Added long option support to the testsuite programs (and quartzdump).
8694 * Testsuite now builds on platforms for which we use our own stringstream
8697 * Only use \r in test output if the output is a tty.
8699 * Increased default timeout used by tests running on the remote backend from 10
8700 seconds to 5 minutes to avoid tests failing just because the machine running
8701 them is slow and/or busy.
8703 * Fixed check for broken exception handling - we were getting "Xapian::"
8704 prefixed to one version and not on the other.
8706 * tests/runtest: Set srcdir if it isn't already to make it easy to manually run
8707 test programs from a VPATH build.
8709 * apitest: Check termfreq in allterms4.
8713 * quartz: Fixed allterms TermIterator to not give duplicate terms when a
8714 posting list is chunked; added regression test (allterms4).
8716 * quartz: Check for EINTR when reading or writing blocks and retry the
8717 operation. This should mean quartz won't fail falsely if a signal is
8718 received (e.g. if alarm() is used).
8722 * Renamed libomqueryparser to libxapianqueryparser - for backward compatibility
8723 we still provide a library with the old name for now.
8725 * xapian.m4: Added XO_LIB_XAPIAN to replace OM_PATH_XAPIAN. XO_LIB_XAPIAN will
8726 automagically enable use of "xapian-config --ltlibs" if A[CM]_PROG_LIBTOOL is
8727 used in configure.in.
8729 * xapian-config: Now supports linking with libtool - using libtool means that
8730 the run-time library path is set and that you can now link with an
8731 uninstalled libxapian. Also xapian-config will now work once xapian-core's
8732 configure has been run, rather than only after "make all".
8734 * xapian-config: Now automatically tries to link libxapianqueryparser too.
8736 * bootstrap: Removed bootstrap scripts in favour of top-level bootstrap which
8737 creates a top-level configure you can optionally use to configure all checked
8738 out Xapian modules with one command, and which creates a top level Makefile
8739 to build all checked out Xapian modules with one command.
8741 * Added versioning information to libxapian and libxapianqueryparser.
8743 * xapian-example/omega: Use libtool and XO_LIB_XAPIAN so we can link with an
8744 uninstalled Xapian, and so the run time load path gets built into the
8745 binaries (no need to set LD_LIBRARY_PATH just because you install Xapian with
8746 a non-standard prefix).
8748 * configure: Stop the API documentation from being regenerated when
8749 include/xapian/version.h changes (since it's generated by configure).
8751 * Fixed "make dist" in VPATH builds.
8755 * common/getopt.h: #include <stdlib.h>, <stdio.h>, and <unistd.h> before
8756 defining getopt as a macro - this avoids problems with clobbering prototypes
8757 of getopt() in system headers.
8759 * bin/quartzcompact.cc: Need stdio.h for rename().
8761 * languages/Makefile.am: Fixed compilation for compilers other than GCC.
8763 * Moved rset serialisation into a method of RSet::Internal, so
8764 omrset_to_string() is now just glue code. This eliminates the need for it to
8765 be a friend of RSet::Internal which Sun's C++ compiler didn't seem to be able
8770 * Fix incorrect documentation comment for Enquire::set_set_forward(). (Looked
8771 like a cut&paste error)
8773 * COPYING: Updated FSF address, and reinstated missing section: "How to Apply
8774 These Terms to Your New Programs"
8776 * PLATFORMS: Updated some linux results: RH7.3 on x86, and Debian on alpha and
8777 arm; Updated FreeBSD success report; Updated with results from the tinderbox.
8779 * docs/mkdoc.pl: Don't choke on a comment at the end of the DIST_SUBDIRS line
8782 * HACKING: Improved note about why libtool 1.5 is needed.
8784 * HACKING: Added note about additional tools needed for building a
8789 * Fixed VPATH builds.
8791 * python: Fixed to link with libomqueryparser.
8793 * guile,tcl8: Updated typemaps to SWIG 1.3 style.
8797 * omindex.cc: Added missing `#include <errno.h>'.
8799 * omindex/scriptindex: Fixed signed character issue in accent normalisation.
8801 * omindex: fixed memory and file descriptor leak on indexing a zero-sized file.
8803 * omindex: Fixed sense of test for unreadable files.
8805 * omindex: Improved log messages to distinguish re-indexed/added.
8807 * omindex,omega,scriptindex: Fixed to compile with mingw.
8809 * omindex: Fixed to compile with GNU getopt so we can build on non-glibc
8814 * msearch: Quick fix to get mingw building going.
8816 * getopt: Copied over our fixes for better C++ compatibility.
8818 * simplesearch: Stem search terms.
8820 * simpleindex: Fixed not to run words together between lines.
8822 * simpleindex: Create database if it doesn't exist.
8824 Xapian 0.7.2 (2003-07-11):
8828 * Fixed NULL pointer dereference when a test threw an unexpected exception.
8832 * Quartz: When asked to create a quartz database, try to create the directory
8833 if it doesn't already exist. Then we don't have to do it in every single
8834 Xapian program which wants to create a database...
8838 * common/getopt.h: Fixed to work better with C++ compilers on non-glibc
8841 * common/utils.h: missing #include <ctype.h>
8843 * Quartz: Defined _XOPEN_SOURCE=500 for GLIBC so we get pread() and pwrite().
8845 * common/utils.h: Improved mingw implementation of rmdir().
8849 * PLATFORMS: Added MacOS X 10.2 success report.
8851 * Improvements to doxygen-generated documentation.
8855 * Moved to separate xapian-bindings module.
8857 * Added configure check for SWIG version (require at least 1.3.14).
8859 * bindings/swig/xapian.i: Fixed over-enthusiastic automatic conversion of
8860 termname to std::string.
8862 * PHP4 bindings much closer to working once again; updated guile and tcl8
8867 * omega: If the same database is listed more than once, only search the first
8870 * omega: use snprintf to help guard against buffer overflows.
8872 Xapian 0.7.1 (2003-07-08):
8876 * Fixed testsuite programs to not try to use "rm -rf" under mingw.
8880 * Quartz: Use pread() and pwrite() on platforms which support them. Doing so
8881 avoids one syscall per block read/write.
8883 * Quartz block count is now unsigned, which should nearly double the size of
8884 database for a given block size. Not tested this yet.
8888 * omindex: Fixed compilation problem in 0.7.0.
8892 * Added new document discussing scalability issues.
8894 * PLATFORMS: Updated.
8896 Xapian 0.7.0 (2003-07-03):
8900 * Moved everything into a Xapian namespace, which the main header now being
8901 xapian.h (rather than om/om.h).
8903 * Three classes have been renamed for better naming consistency:
8904 OmOpeningError is now Xapian::DatabaseOpeningError, OmPostListIterator is
8905 now Xapian::PostingIterator, and OmPositionListIterator is now
8906 Xapian::PositionIterator.
8908 * xapian.h includes <iosfwd> rather than <iostream> - if you were relying on
8909 the implicit inclusion, you'll need to add an explicit "#include <iostream>".
8911 * Replaced om_termname with explicit use of std::string - om_termname was just
8912 a typedef for std::string and the typedef doesn't really buy us anything.
8914 * Older code can be compiled by continuing to use om/om.h which uses #define
8915 and other tricks to map the old names onto the new ones.
8917 * Define XAPIAN_VERSION (e.g. 0.7.0), XAPIAN_MAJOR_VERSION (e.g. 0), and
8918 XAPIAN_MINOR_VERSION (e.g. 7).
8920 * Updated omega and xapian-examples to use Xapian namespace.
8924 * Xapian::QueryParser: Accent normalisation added; Improved error reporting;
8925 Fixed to handle the most common examples found in the wild which used to give
8930 * Python bindings brought up to date - use ./configure --enable-bindings to
8931 build them. Requires Python >= 2.0 - may require Python >= 2.1.
8933 * Enabled optional building of bindings as part of normal build process. Old
8934 Perl and Java bindings dropped; for Perl, use Search::Xapian from CPAN; Java
8935 JNI bindings will be replaced with a SWIG-based implmentation.
8937 internal implementation changes:
8939 * Removed one wrapper layer from the internal implementation of most API
8942 * Xapian::Stem now uses reference counted internals.
8944 * Internally a lot of cases of unnecessary header inclusion have been removed
8945 or replaced with forward declarations of classes. This should speed up
8946 compilation and recompilation of the Xapian library.
8948 * Suppress warnings in Snowball generated C code.
8950 * Reworked query serialisation in the remote backend so that the code is now
8951 all in one place. The serialisation is now rather more compact and no longer
8952 relies on flex for parsing.
8956 * Moved all the core library tests to tests subdirectory.
8958 * apitest now allows backend to be specified with "-b" rather than having to
8959 mess with environmental variables.
8961 * Testsuite programs can now hook into valgrind for leak checking, undefined
8962 variable checking, etc.
8966 * Fixed parsing of port number in remote stub databases.
8968 * Quartz: Improved error message when asked to open a pre-0.6 Quartz database.
8970 * Quartz backend: Workaround for shared_level problem turns out to
8971 be arguably the better approach, so made it permanent and tidied up
8976 * Build system fixed to never leave partial files in place of the expected
8977 output if a build is interrupted.
8979 * quartzcheck, quartzdump, and quartzcompact are now built by "make" rather
8980 than only by "make check".
8982 * xapian-config: Removed --prefix and --exec-prefix - you can't reliably
8983 install Xapian with a different prefix to the one it was configured with,
8984 yet these options give the impression you can.
8988 * Fixed sending debug output to a file with XAPIAN_DEBUG_LOG with a value which
8989 didn't contain "%%" (%% expands to the current PID).
8991 * Fixed Xapian::MSetIterator::get_collapse_count() to work as intended.
8995 * omindex,scriptindex: Normalise accents in probabilistic terms.
8997 * omindex: Read output from pstotext and pdftotext via pipes rather
8998 than temporary files to side-step the whole problem of secure temporary file
8999 creation; Use pdfinfo to get the title and keywords from when indexing a PDF;
9000 Safe filename escaping tweaked to not escape common safe punctuation.
9002 * omindex: Implement an upper limit on the length of URL terms - this is a
9003 slightly conservative 240 characters. If the URL term would be longer than
9004 this, its last few bytes are replaced by a hash of the tail of the URL. This
9005 means that (apart from hopefully very rare collisions) urlterms should still
9006 be unique ids for documents. This is forward and backward compatible for
9007 URLs less than 240 characters.
9009 * omindex: Clean up processing of HTML documents:
9010 - Ignore the contents of <script> and <style> tags in HTML.
9011 - Strip initial whitespace in each tag in an HTML document.
9012 - Try not to split words in half when truncating title and summary.
9014 * query.cc: Set STEM_LANGUAGE near the start of the file so it's easy
9015 for users to change until we get better configurability.
9017 * omega: Replaced half-hearted logging support with flexible OmegaScript-based
9018 approach with new $log command. Also added $now to allow the current
9019 date/time to be logged.
9021 * templates/xml: added collapse info to xml template.
9025 * Assorted minor documentation improvements.
9027 * PLATFORMS: Updated.
9031 * Improved RPM packaging of xapian-core and omega.
9033 Xapian 0.6.5 (2003-04-10):
9035 * OmEnquire: optimised the handling when sort_bands == 1 and fixed incorrect
9036 results in this and some other sorting cases; added some sorting testcases.
9038 * OmMSetIterator: added get_collapse_count() which returns a lower bound on
9039 the number of items which were removed by collapsing onto the current item.
9041 * OmStem: added default OmStem constructor and "none" language. Both of these
9042 give a stemmer object which leaves terms unchanged which should allow for
9043 simpler logic in programs using Xapian. The default constructor also removes
9044 the need to mess with pointers in some cases.
9046 * Automatically disable the remote backend if we don't have fork() since the
9047 remote backend requires it in several places.
9049 * Fixed to build with debug enabled.
9051 * testsuite: fixed to still build when some backends are disabled.
9053 * extra/parsequerytest.cc: Fixed to build with GCC 2.95.
9055 * Testsuite: Added regression test for Quartz bug which caused problems with
9056 long terms on machines with signed chars.
9058 * testsuite/index_utils.cc: Handling of ^x was just downright wrong due to a
9061 * Improved portability: Fix for 64 bit machines. Fixed btreetest to build with
9062 older compilers lacking <sstream>. Xapian is now much closer to building
9063 with Sun's CFront-based Sun Pro C++ compiler, and with a Linux to mingw
9066 * PLATFORMS: Updated with the results of many test builds.
9068 * Improved RPM packaging of xapian-core and omega.
9070 * Documentation: Use http://www.doxygen.org/ as URL for doxygen; Fixed bad link
9071 to our own website in overview.html; code_structure.html now only includes
9072 directories in the build system.
9076 * Removed bugs/todo.xml, TODO, TODO.release, docs/todo.html, and
9077 docs/todo-release.html from the distribution. Bugs and todo items will be
9078 tracked in Bugzilla instead.
9080 * Install docs in /usr/share/doc/xapian-core instead of /usr/share/xapian-core.
9082 * omega: If xP and P are both empty, there may be a boolean query, so don't
9083 force first page of hits.
9085 * omega: Fixed off-by-one error in rounding down topdoc - it was possible to
9086 get to an empty page of hits if there were exactly a multiple of HITSPERPAGE
9087 matches and the matcher over-estimated the number of matches and Omega
9088 displayed page links.
9090 * omega: Fixed handling of multiple DB parameters to be as documented.
9092 * omega: Added $collapsed to report get_collapse_count() for the current hit.
9094 * omega: Added $transform{} which does regexp manipulation (currently disabled
9095 until configure tests for regexp library are added)
9097 * omega: Added $uniq{} to eliminate duplicates from a sorted list.
9099 * omega: Don't force page 1 for a query with repeated terms!
9101 * omega: removed duplicates from terms listed in term frequencies.
9103 * omega: Added cgi parameter COLLAPSE to collapse on key values
9105 * omega: Added $value{key[,docid]} support to omegascript
9107 * omega: Renamed DATE1, DATE2, and DAYSMINUS to the more meaningful START, END,
9108 and SPAN (NB SPAN is days before END, or after START, or before today -
9109 whereas SPAN was before *DATE1* or before today). The old parameters names
9110 are supported (with the original semantics) for now.
9112 * omega: Actually install documentation!
9114 * templates/query: propagate B boolean filters
9116 * templates/godmode: removed link to EuroFerret image
9118 * templates/godmode: added value dumping, for values from 0-255
9120 * omindex: Report correct version number (was hard-wired to 1.0!)
9122 * scriptindex: Allow '_' in fieldnames. Diagnose bad characters in fieldnames
9125 * dbi2omega: Added DBUSER and DBPASSWD environmental variable support so that
9126 password protected DBs can easily be used
9128 * scriptindex.cc: added missing "#include <stdio.h>" which caused builds
9129 to fail for some platforms.
9131 Xapian 0.6.4 (2002-12-24):
9133 * Quartz backend: Fixed double setting of position list when updating a
9134 document with term position information (overall result was correct, just
9135 inefficient); when deleting a position_list, don't check if it's empty,
9136 just ask the layer below to delete it and let it handle the case when
9137 there's nothing to delete; Fixed unpacking of termlist on platforms where
9140 * OmQueryParser: Added support for searching probabilistic fields (using
9141 <field>:<term>); the unstem multimap now includes "." on the end of a
9142 term if it was there in the query.
9144 * Don't include "om.h" as a dependency for the api docs since it's generated
9145 a configure time and the dependency was forcing users to regenerate the
9146 documentation, which requires doxygen to be installed.
9148 * Bindings: Python bindings updated to work with the updated API (still
9149 disabled by default).
9151 * Muscat 3.6 backend: Fixed to build with the new database factory functions;
9152 fixed compilation warnings; Muscat 3.6 DA and DB databases don't support
9153 positional information. Instead of throwing an exception when we try to
9154 access it, return an empty position list (like a quartz database with no
9155 position information would). This allows copydatabase to be used to convert
9156 a Muscat 3.6 database to a quartz one.
9158 * Documentation: quartzdesign and todo list updated.
9160 * quartzcheck: default mode changed to "v" rather than "+", since "+" is too
9161 verbose for a btree of any size; if you pass a quartz database directory,
9162 quartzcheck will now check all the tables which make up a quartz database.
9164 * quartzcompact: new tool which makes a copy of a quartz database with full
9165 compaction turned on - this results in a smaller database which is faster
9166 to search. The next update will result in a lot of block splitting though
9167 (since all blocks are as full as possible).
9169 * omega: Added $unstem to map a stemmed term to the form(s) used in the query;
9170 $queryterms now only includes the first occurrence of each stemmed form;
9171 $prettyterm makes use of the unstem map; prefer MINHITS to MIN_HITS and
9172 RAWSEARCH to RAW_SEARCH since none of the other CGI parameter names have
9173 _ separating words (continue to support old names for now); fixed default
9174 template to not generate topterms twice, and fixed topterms to not stick
9175 outside the green box; corrected omegascript docs - it's $setrelevant
9178 * scriptindex: index=nopos with new indexnopos action; index and indexnopos now
9179 take an optional prefix argument; index=nopos is handled specially for
9180 backwards compatibility; added new data action to generate terms for date
9183 Xapian 0.6.3 (2002-12-14):
9185 * Updated PLATFORMS and todo list. Noted in HACKING that Bison 1.50 seems to
9188 * OmQueryParser now creates an "unstem" multimap to allow probabilistic
9189 query terms to be converted back to the form the user originally typed.
9191 * Updated documentation for remote protocol description and the quickstart
9192 tutorial which were both very out of date.
9194 * No longer use OmSettings to pass matcher parameters. This completes the
9195 removal of OmSettings.
9197 * Added workaround for problem with cursors sharing levels in the btree.
9198 This should fix sporadic problems with large databases (small databases
9199 have fewer btree levels so aren't affected).
9201 * Stub databases now work again, though with a different format. The new
9202 format allows multiple databases to be specified in the stub file.
9204 * OmEnquire::get_eset() now takes a flags argument of bit constants |-ed
9205 together instead of 2 bools.
9207 * Applied Martin Porter's better fix for the btree sequential addition bug
9208 which Richard fixed a few months ago. Richard's fix resulted in a correct
9209 btree, but didn't always utilise space as efficiently as possible.
9211 * Fixed the remote backend to handle weighting schemes after the OmSettings
9212 changes. You can now even implement your own weighting scheme and use it
9213 with the remote backend provided you register it with SocketServer at
9214 runtime (this feature has been on the todo list for ages).
9216 Xapian 0.6.2 (2002-12-07):
9218 * Set env var XAPIAN_SIG_DFL to stop the testsuite installing its
9219 signal handler (may be useful with some debugging tools).
9221 * backends/quartz/btree.cc: max_item_size wasn't being set due to
9222 some over-zealous code pruning. It was defaulting to 0, and
9223 was causing the code to write off the end of allocated memory
9226 * matcher/localmatch.cc: fixed handling of wtscheme() - we were
9227 trying to use it for the extra weights, and then double
9230 * common/omdebug.cc,common/omdebug.h: Fixed permissions on newly
9231 created log file (was getting 000!); Simplified class internals;
9232 Renamed env vars: OM_DEBUG_FILE is now XAPIAN_DEBUG_LOG,
9233 OM_DEBUG_TYPES is now XAPIAN_DEBUG_FLAGS (old versions still work
9236 * testsuite/testsuite.cc: Fixed so running "gdb .libs/apitest"
9237 finds srcdir (for an in-tree build at least).
9239 * Fixed to compile with --enable-debug=full.
9241 * docs/remote.html: Updated from OmSettings to factory functions.
9243 * PLATFORMS: ixion is actually Linux 2.2.
9245 * OmWritableDatabase now has a default constructor.
9247 * Weighting scheme now specified by passing OmWeight object to OmEnquire.
9248 This also allows user weighting schemes (just subclass OmWeight and
9249 pass in an instance of this new class). [This doesn't currently work
9250 with the remote backend.]
9252 * No longer use OmSettings to specify parameters for constructing databases.
9253 Instead there's a factory function for each database type - temporary naming
9254 scheme is OmXxx__open(), mostly because it's easy to grep for later.
9255 Instead of create and overwrite flags, we pass in a value - a new possible
9256 opening mode is "create or open". [At present stub databases and the
9257 machinery in InMemory to allow the multierrhandler1 test aren't working.
9258 Everything else should be.]
9260 * OmEnquire::get_eset() takes parameters instead of an OmSettings object.
9262 * Fixed reversed sense of use_query_terms (and fixed reversed sense test in
9263 apitest which meant this wasn't spotted).
9265 * Documentation: Link to annotated class lists in doxygen generated
9266 documentation instead of the rather empty index pages; added doxygen
9267 markup so that apidoc now documents header files; updated todo list.
9269 * Documentation: intro doc thing was very out of date in places - fixed.
9271 * Omega: index .php files as HTML, with the PHP code stripped out; omindex
9272 return non-zero return code if an unexpected exception is caught; fixed
9273 HTML parser to not read one character past the end of the document in
9274 some cases; updated in line with OmSettings related changes to the API;
9275 Fixed $dbname to return "default" for the default database instead of "";
9276 templates/query: Removed now unused xDEFAULTOP hidden field, and superfluous
9277 "}"; dbi2omega now more efficient and can be restricted to listed fields.
9279 Xapian 0.6.1 (2002-11-28):
9281 * Fixed to compile with GCC 3.0.
9283 * PLATFORMS: Updated.
9285 Xapian 0.6.0 (2002-11-27):
9287 * Quartz database backend: lexicon disabled (./configure CXXFLAGS=-DUSE_LEXICON
9288 to reenable it), and encoding schemes simplified and made more compact;
9289 extended and added test cases; minimum block size is now 2048 bytes (as
9290 documented before, but now we actually enforce this); btree checking code
9291 split off and only linked in when required; tidied up btreetest's output.
9293 * Replaced our stemmers with those from Snowball. These give better results,
9294 and are actively maintained by Martin Porter (who wrote the original Xapian
9295 stemmers too). It also means that Xapian now has stemmers for Finnish,
9296 and Russian, and an implementation of Lovins' English stemmer.
9298 * Assorted improvements to the documentation, especially the documentation
9299 of the internals of the Quartz backend.
9301 * Removed the three uses of RTTI (typeid() and dynamic_cast<>) - one was
9302 totally superfluous, and the other two easily avoided.
9304 * Omega and simpleindex example: limit probabilistic term length to 64
9305 characters to stop the index filling up with junk terms which nobody will
9308 * Omega: Added dbi2omega perl script to dump any database which perl DBI can
9309 access into the dump format expected by scriptindex.
9311 Xapian 0.5.5 (2002-12-04):
9313 * Fixed compilation with --enable-debug.
9315 * Minor documentation updates.
9317 * Omega: Fixed paging on default database; removed xDEFAULTOP from the query
9318 template as it's no longer used; removed bogus unmatched '}' from query
9319 template; added dbi2omega perl script to dump any database which perl DBI
9320 can access into the dump format expected by scriptindex; limit length of
9321 probabilistic terms generated to 64 characters.
9323 Xapian 0.5.4 (2002-10-16):
9325 * Fixed a compilation error with "make check" when using GCC 3.2.
9327 * PLATFORMS: checked 0.5.3 works on OpenBSD and Solaris 7.
9329 Xapian 0.5.3 (2002-10-12):
9331 Notable changes: Improvements to the test suite, and internal code cleanups:
9333 * Internal code cleanups on Quartz Btree implementation.
9335 * Minor documentation updates (TODO and PLATFORMS updated; Martin Porter's
9336 stemming paper removed - see the Snowball site for background stemmer
9339 * Implemented QuartzAllTermsList::get_approx_size().
9341 * Removed a couple of occurrences of "using std::XXX;" from externally
9344 * With GCC, add warning flags "-Wall -W" rather than "-Wall -Wunused" (-Wall
9345 implies -Wunused anyway). Fixed all the warnings this throws up, except in
9346 languages/ (that code is to be replaced with Snowball soon).
9348 * Test suite: Disable colour test output if stdout isn't a terminal and
9349 reworked check for broken exception handling as the previous version never
9350 seemed to fire. Other assorted minor improvements.
9352 * include/om/om.h is now removed on "make distclean" rather than "make clean".
9354 Xapian 0.5.2 (2002-10-06):
9356 Further improvements to documentation and portability:
9358 * docs/: converted all text docs to HTML (except omsettings which will
9359 has odd markup (LaTeX?) and will probably soon be obsolete anyway).
9361 * remote backend: Fixed handling of timeouts which are now in the past - fixes
9362 test failures with redhat/x86.
9364 * quartz backend: now works on 64 bit platforms.
9366 * test suite: try to spot mishandled exceptions and stop them causing bogus
9369 Xapian 0.5.1 (2002-10-02):
9371 This release fixes features improved documentation and some build system
9374 * PLATFORMS: updated with more test results.
9376 * docs/: tidied up layout of HTML documentation; converted the notes about
9377 BM25 into HTML; updated stemmer docs to reflect intention to use Snowball
9378 instead; included HTML versions of quickstart*.cc.
9380 * automake 1.6.3 and autoconf 2.54 are now required for those working
9381 from CVS to fix a problem with the generated Makefiles and Solaris
9384 * net/Makefile.am: Fixed building of readquery.cc from readquery.ll.
9386 * buildall script is now deprecated - use the new streamlined bootstrap script
9389 Xapian 0.5.0 (2002-09-20):
9391 The last release of the software that is now known as Xapian was OmSee 0.4.1 on
9392 November 24th 2000, not far from 2 years ago.
9394 There's been a significant amount of development in this time, so we've
9395 summarised the most notable changes and improvements:
9397 * The project is now called "Xapian". We've renamed the modules in the light
9400 + "om" is now "xapian-core"
9401 + "om-examples" is now "xapian-examples", and now contains small,
9402 instructive examples which demonstrate how to use Xapian to implement
9403 particularly features.
9404 + Added "xapian-applications" which contains larger sample applications
9406 * Much improved build system - should now build "out of the box" on many Unix
9407 platforms. Can now VPATH build with vendor tools on most platforms. Builds
9408 as cleanly as we can achieve with GCC 2.95.* (some bogus warnings due to
9409 compiler bugs). Should build without warnings on GCC 3.0, 3.1, and 3.2.
9411 * If using GCC, om/om.h now contains a check that the compiler used to build
9412 Xapian and the compiler used to build the application have compatible C++
9413 ABIs. So you get a clear error message early from the first attempt to
9414 compile a file rather than a confusing error from the linker near the end
9417 * RPM packages are now available. We intend to prepare Debian packages in the
9420 * xapian-config no longer support "--uninst". It's hard to make this work
9421 reliably and portably, and the effort is better expended elsewhere.
9422 Configure with a prefix and install to a temporary directory instead.
9424 * Xapian can now work with files > 2Gb on OSes which support them.
9426 * Restructured and reworked documentation.
9428 * Removed thread locks. We intend to be "thread-friendly" so different
9429 threads can access different objects without problems. In the rare event
9430 that you want to concurrently call methods on the same object from
9431 different threads you need to create a mutex and lock it. Thus the thread
9432 lock overhead is only incurred when it's necessary.
9434 * Indexgraph removed from core library. It will reappear as an add-on library
9437 * Omega's query parser has now been reworked as a separate library.
9439 * Terminology change - "keys" are now known as "values" to avoid confusion,
9440 since they're not like keys in a relational database. The exception is when
9441 a value is used as a key in some operation, e.g. "match_collapse_key".
9443 * Database backends:
9445 + Auto backend: can now be used to create a new database.
9446 + Auto backend: added support for "stub" databases - a text file
9447 specifying the settings for the database to be opened (particularly
9448 useful for allowing easy access to specific remote databases).
9449 + Quartz backend: many fixes and improvements, and the code has been
9450 cleaned up a lot. Implemented deleting of items from postlists.
9451 + Remote backend: implemented term_exists() and get_termfreq();
9452 + Multi-backend: the document length is now fetched from the sub-postlist
9453 rather than the database, which provides a huge speed-up in some cases.
9454 + Sleepycat backend: this experimental backend has been removed.
9455 + Muscat 3.6 backends: now disabled by default.
9459 + Test cases added for most bug fixes and new features.
9460 + stemtest: rewritten in C++ rather than part C++, part perl. Now 15%
9462 + includetest: removed - it's no longer useful now the code has matured.
9463 + Removed problematic leak checking from testsuite. We plan to use
9464 valgrind instead soon.
9468 + Fixed several matcher bugs which could cause incorrect results in some
9470 + Fix bug in expander due to nth_element being called on the wrong
9472 + Added sorting within relevance bands to the matcher.
9473 + Matcher now calculates percentages differently, such that 100%
9474 relevance is actually achievable.
9475 + Matcher now uses a min-heap rather than nth-element to maintain the
9476 proto-mset. This is cleaner and more efficient.
9477 + New operator OP_ELITE_SET replaces match_max_or_terms option.
9478 + Implemented multiple XOR queries.
9479 + Add a new query operator, OP_WEIGHT_CUTOFF, which returns only those
9480 documents from a query which have a weight greater than a specified
9482 + Removed OmBatchEnquire from system: it may return at a later date, but
9483 for now it is simply out of date and a maintenance liability, and
9484 gives no significant advantage.
9485 + Added experimental match bias functors.
9487 * The API has been cleaned up in various places:
9489 + OmDocumentContents and OmIndexDoc merged to become OmDocument
9490 + OmQuery interface cleaned up
9491 + OmData and OmKey removed - methods which used them now just pass a
9493 + OmESetItem replaced by OmESetIterator; OmMSetItem by OmMSetIterator;
9494 om_termname_list by OmTermIterator
9495 + OmDocumentTerm and OmDocumentParams removed
9496 + OmMSet::mbound replaced by OmMSet::matches_
9497 {lower_bound,estimated,upper_bound}, giving more information
9498 + Xapian iterators now have default constructors
9499 + Most API classes now have reference counted internals, so assignment
9500 and copying are cheap
9501 + OmStem now has copy constructor and assignment operator