| blob | 01a54ef8ce440f3d17f5dff01816451bc4448886 |
1 /****************************************************************
2 Copyright (C) Lucent Technologies 1997
3 All Rights Reserved
5 Permission to use, copy, modify, and distribute this software and
6 its documentation for any purpose and without fee is hereby
7 granted, provided that the above copyright notice appear in all
8 copies and that both that the copyright notice and this
9 permission notice and warranty disclaimer appear in supporting
10 documentation, and that the name Lucent Technologies or any of
11 its entities not be used in advertising or publicity pertaining
12 to distribution of the software without specific, written prior
13 permission.
15 LUCENT DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
16 INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS.
17 IN NO EVENT SHALL LUCENT OR ANY OF ITS ENTITIES BE LIABLE FOR ANY
18 SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
19 WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER
20 IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
21 ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
22 THIS SOFTWARE.
23 ****************************************************************/
25 This file lists all bug fixes, changes, etc., made since the AWK book
26 was sent to the printers in August, 1987.
28 May 1, 2007:
29 fiddle in makefile to fix for BSD make; thanks to igor sobrado.
31 Mar 31, 2007:
32 fixed some null pointer refs calling adjbuf.
34 Feb 21, 2007:
35 fixed a bug in matching the null RE in sub and gsub. thanks to al aho
36 who actually did the fix (in b.c), and to wolfgang seeberg for finding
37 it and providing a very compact test case.
39 fixed quotation in b.c; thanks to Hal Pratt and the Princeton Dante
40 Project.
42 removed some no-effect asserts in run.c.
44 fiddled maketab.c to not complain about bison-generated values.
46 removed the obsolete -V argument; fixed --version to print the
47 version and exit.
49 fixed wording and an outright error in the usage message; thanks to igor
50 sobrado and jason mcintyre.
52 fixed a bug in -d that caused core dump if no program followed.
54 Jan 1, 2007:
55 dropped mac.code from makefile; there are few non-MacOSX
56 mac's these days.
58 Jan 17, 2006:
59 system() not flagged as unsafe in the unadvertised -safe option.
60 found it while enhancing tests before shipping the ;login: article.
61 practice what you preach.
63 removed the 9-years-obsolete -mr and -mf flags.
65 added -version and --version options.
67 core dump on linux with BEGIN {nextfile}, now fixed.
69 removed some #ifdef's in run.c and lex.c that appear to no
70 longer be necessary.
72 Apr 24, 2005:
73 modified lib.c so that values of $0 et al are preserved in the END
74 block, apparently as required by posix. thanks to havard eidnes
75 for the report and code.
77 Jan 14, 2005:
78 fixed infinite loop in parsing, originally found by brian tsang.
79 thanks to arnold robbins for a suggestion that started me
80 rethinking it.
82 Dec 31, 2004:
83 prevent overflow of -f array in main, head off potential error in
84 call of SYNTAX(), test malloc return in lib.c, all with thanks to
85 todd miller.
87 Dec 22, 2004:
88 cranked up size of NCHARS; coverity thinks it can be overrun with
89 smaller size, and i think that's right. added some assertions to b.c
90 to catch places where it might overrun. the RE code is still fragile.
92 Dec 5, 2004:
93 fixed a couple of overflow problems with ridiculous field numbers:
94 e.g., print $(2^32-1). thanks to ruslan ermilov, giorgos keramidas
95 and david o'brien at freebsd.org for patches. this really should
96 be re-done from scratch.
98 Nov 21, 2004:
99 fixed another 25-year-old RE bug, in split. it's another failure
100 to (re-)initialize. thanks to steve fisher for spotting this and
101 providing a good test case.
103 Nov 22, 2003:
104 fixed a bug in regular expressions that dates (so help me) from 1977;
105 it's been there from the beginning. an anchored longest match that
106 was longer than the number of states triggered a failure to initialize
107 the machine properly. many thanks to moinak ghosh for not only finding
108 this one but for providing a fix, in some of the most mysterious
109 code known to man.
111 fixed a storage leak in call() that appears to have been there since
112 1983 or so -- a function without an explicit return that assigns a
113 string to a parameter leaked a Cell. thanks to moinak ghosh for
114 spotting this very subtle one.
116 Jul 31, 2003:
117 fixed, thanks to andrey chernov and ruslan ermilov, a bug in lex.c
118 that mis-handled the character 255 in input. (it was being compared
119 to EOF with a signed comparison.)
121 Jul 29, 2003:
122 fixed (i think) the long-standing botch that included the beginning of
123 line state ^ for RE's in the set of valid characters; this led to a
124 variety of odd problems, including failure to properly match certain
125 regular expressions in non-US locales. thanks to ruslan for keeping
126 at this one.
128 Jul 28, 2003:
129 n-th try at getting internationalization right, with thanks to volker
130 kiefel, arnold robbins and ruslan ermilov for advice, though they
131 should not be blamed for the outcome. according to posix, "." is the
132 radix character in programs and command line arguments regardless of
133 the locale; otherwise, the locale should prevail for input and output
134 of numbers. so it's intended to work that way.
136 i have rescinded the attempt to use strcoll in expanding shorthands in
137 regular expressions (cclenter). its properties are much too
138 surprising; for example [a-c] matches aAbBc in locale en_US but abBcC
139 in locale fr_CA. i can see how this might arise by implementation
140 but i cannot explain it to a human user. (this behavior can be seen
141 in gawk as well; we're leaning on the same library.)
143 the issue appears to be that strcoll is meant for sorting, where
144 merging upper and lower case may make sense (though note that unix
145 sort does not do this by default either). it is not appropriate
146 for regular expressions, where the goal is to match specific
147 patterns of characters. in any case, the notations [:lower:], etc.,
148 are available in awk, and they are more likely to work correctly in
149 most locales.
151 a moratorium is hereby declared on internationalization changes.
152 i apologize to friends and colleagues in other parts of the world.
153 i would truly like to get this "right", but i don't know what
154 that is, and i do not want to keep making changes until it's clear.
156 Jul 4, 2003:
157 fixed bug that permitted non-terminated RE, as in "awk /x".
159 Jun 1, 2003:
160 subtle change to split: if source is empty, number of elems
161 is always 0 and the array is not set.
163 Mar 21, 2003:
164 added some parens to isblank, in another attempt to make things
165 internationally portable.
167 Mar 14, 2003:
168 the internationalization changes, somewhat modified, are now
169 reinstated. in theory awk will now do character comparisons
170 and case conversions in national language, but "." will always
171 be the decimal point separator on input and output regardless
172 of national language. isblank(){} has an #ifndef.
174 this no longer compiles on windows: LC_MESSAGES isn't defined
175 in vc6++.
177 fixed subtle behavior in field and record splitting: if FS is
178 a single character and RS is not empty, \n is NOT a separator.
179 this tortuous reading is found in the awk book; behavior now
180 matches gawk and mawk.
182 Dec 13, 2002:
183 for the moment, the internationalization changes of nov 29 are
184 rolled back -- programs like x = 1.2 don't work in some locales,
185 because the parser is expecting x = 1,2. until i understand this
186 better, this will have to wait.
188 Nov 29, 2002:
189 modified b.c (with tiny changes in main and run) to support
190 locales, using strcoll and iswhatever tests for posix character
191 classes. thanks to ruslan ermilov (ru@freebsd.org) for code.
192 the function isblank doesn't seem to have propagated to any
193 header file near me, so it's there explicitly. not properly
194 tested on non-ascii character sets by me.
196 Jun 28, 2002:
197 modified run/format() and tran/getsval() to do a slightly better
198 job on using OFMT for output from print and CONVFMT for other
199 number->string conversions, as promised by posix and done by
200 gawk and mawk. there are still places where it doesn't work
201 right if CONVFMT is changed; by then the STR attribute of the
202 variable has been irrevocably set. thanks to arnold robbins for
203 code and examples.
205 fixed subtle bug in format that could get core dump. thanks to
206 Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing.
207 minor cleanup in run.c / format() at the same time.
209 added some tests for null pointers to debugging printf's, which
210 were never intended for external consumption. thanks to dave
211 kerns (dkerns@lucent.com) for pointing this out.
213 GNU compatibility: an empty regexp matches anything (thanks to
214 dag-erling smorgrav, des@ofug.org). subject to reversion if
215 this does more harm than good.
217 pervasive small changes to make things more const-correct, as
218 reported by gcc's -Wwrite-strings. as it says in the gcc manual,
219 this may be more nuisance than useful. provoked by a suggestion
220 and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk
222 minor documentation changes to note that this now compiles out
223 of the box on Mac OS X.
225 Feb 10, 2002:
226 changed types in posix chars structure to quiet solaris cc.
228 Jan 1, 2002:
229 fflush() or fflush("") flushes all files and pipes.
231 length(arrayname) returns number of elements; thanks to
232 arnold robbins for suggestion.
234 added a makefile.win to make it easier to build on windows.
235 based on dan allen's buildwin.bat.
237 Nov 16, 2001:
238 added support for posix character class names like [:digit:],
239 which are not exactly shorter than [0-9] and perhaps no more
240 portable. thanks to dag-erling smorgrav for code.
242 Feb 16, 2001:
243 removed -m option; no longer needed, and it was actually
244 broken (noted thanks to volker kiefel).
246 Feb 10, 2001:
247 fixed an appalling bug in gettok: any sequence of digits, +,-, E, e,
248 and period was accepted as a valid number if it started with a period.
249 this would never have happened with the lex version.
251 other 1-character botches, now fixed, include a bare $ and a
252 bare " at the end of the input.
254 Feb 7, 2001:
255 more (const char *) casts in b.c and tran.c to silence warnings.
257 Nov 15, 2000:
258 fixed a bug introduced in august 1997 that caused expressions
259 like $f[1] to be syntax errors. thanks to arnold robbins for
260 noticing this and providing a fix.
262 Oct 30, 2000:
263 fixed some nextfile bugs: not handling all cases. thanks to
264 arnold robbins for pointing this out. new regressions added.
266 close() is now a function. it returns whatever the library
267 fclose returns, and -1 for closing a file or pipe that wasn't
268 opened.
270 Sep 24, 2000:
271 permit \n explicitly in character classes; won't work right
272 if comes in as "[\n]" but ok as /[\n]/, because of multiple
273 processing of \'s. thanks to arnold robbins.
275 July 5, 2000:
276 minor fiddles in tran.c to keep compilers happy about uschar.
277 thanks to norman wilson.
279 May 25, 2000:
280 yet another attempt at making 8-bit input work, with another
281 band-aid in b.c (member()), and some (uschar) casts to head
282 off potential errors in subscripts (like isdigit). also
283 changed HAT to NCHARS-2. thanks again to santiago vila.
285 changed maketab.c to ignore apparently out of range definitions
286 instead of halting; new freeBSD generates one. thanks to
287 jon snader <jsnader@ix.netcom.com> for pointing out the problem.
289 May 2, 2000:
290 fixed an 8-bit problem in b.c by making several char*'s into
291 unsigned char*'s. not clear i have them all yet. thanks to
292 Santiago Vila <sanvila@unex.es> for the bug report.
294 Apr 21, 2000:
295 finally found and fixed a memory leak in function call; it's
296 been there since functions were added ~1983. thanks to
297 jon bentley for the test case that found it.
299 added test in envinit to catch environment "variables" with
300 names beginning with '='; thanks to Berend Hasselman.
302 Jul 28, 1999:
303 added test in defn() to catch function foo(foo), which
304 otherwise recurses until core dump. thanks to arnold
305 robbins for noticing this.
307 Jun 20, 1999:
308 added *bp in gettok in lex.c; appears possible to exit function
309 without terminating the string. thanks to russ cox.
311 Jun 2, 1999:
312 added function stdinit() to run to initialize files[] array,
313 in case stdin, etc., are not constants; some compilers care.
315 May 10, 1999:
316 replaced the ERROR ... FATAL, etc., macros with functions
317 based on vprintf, to avoid problems caused by overrunning
318 fixed-size errbuf array. thanks to ralph corderoy for the
319 impetus, and for pointing out a string termination bug in
320 qstring as well.
322 Apr 21, 1999:
323 fixed bug that caused occasional core dumps with commandline
324 variable with value ending in \. (thanks to nelson beebe for
325 the test case.)
327 Apr 16, 1999:
328 with code kindly provided by Bruce Lilly, awk now parses
329 /=/ and similar constructs more sensibly in more places.
330 Bruce also provided some helpful test cases.
332 Apr 5, 1999:
333 changed true/false to True/False in run.c to make it
334 easier to compile with C++. Added some casts on malloc
335 and realloc to be honest about casts; ditto. changed
336 ltype int to long in struct rrow to reduce some 64-bit
337 complaints; other changes scattered throughout for the
338 same purpose. thanks to Nelson Beebe for these portability
339 improvements.
341 removed some horrible pointer-int casting in b.c and elsewhere
342 by adding ptoi and itonp to localize the casts, which are
343 all benign. fixed one incipient bug that showed up on sgi
344 in 64-bit mode.
346 reset lineno for new source file; include filename in error
347 message. also fixed line number error in continuation lines.
348 (thanks to Nelson Beebe for both of these.)
350 Mar 24, 1999:
351 Nelson Beebe notes that irix 5.3 yacc dies with a bogus
352 error; use a newer version or switch to bison, since sgi
353 is unlikely to fix it.
355 Mar 5, 1999:
356 changed isnumber to is_number to avoid the problem caused by
357 versions of ctype.h that include the name isnumber.
359 distribution now includes a script for building on a Mac,
360 thanks to Dan Allen.
362 Feb 20, 1999:
363 fixed memory leaks in run.c (call) and tran.c (setfval).
364 thanks to Stephen Nutt for finding these and providing the fixes.
366 Jan 13, 1999:
367 replaced srand argument by (unsigned int) in run.c;
368 avoids problem on Mac and potentially on Unix & Windows.
369 thanks to Dan Allen.
371 added a few (int) casts to silence useless compiler warnings.
372 e.g., errorflag= in run.c jump().
374 added proctab.c to the bundle outout; one less thing
375 to have to compile out of the box.
377 added calls to _popen and _pclose to the win95 stub for
378 pipes (thanks to Steve Adams for this helpful suggestion).
379 seems to work, though properties are not well understood
380 by me, and it appears that under some circumstances the
381 pipe output is truncated. Be careful.
383 Oct 19, 1998:
384 fixed a couple of bugs in getrec: could fail to update $0
385 after a getline var; because inputFS wasn't initialized,
386 could split $0 on every character, a misleading diversion.
388 fixed caching bug in makedfa: LRU was actually removing
389 least often used.
391 thanks to ross ridge for finding these, and for providing
392 great bug reports.
394 May 12, 1998:
395 fixed potential bug in readrec: might fail to update record
396 pointer after growing. thanks to dan levy for spotting this
397 and suggesting the fix.
399 Mar 12, 1998:
400 added -V to print version number and die.
402 Feb 11, 1998:
403 subtle silent bug in lex.c: if the program ended with a number
404 longer than 1 digit, part of the input would be pushed back and
405 parsed again because token buffer wasn't terminated right.
406 example: awk 'length($0) > 10'. blush. at least i found it
407 myself.
409 Aug 31, 1997:
410 s/adelete/awkdelete/: SGI uses this in malloc.h.
411 thanks to nelson beebe for pointing this one out.
413 Aug 21, 1997:
414 fixed some bugs in sub and gsub when replacement includes \\.
415 this is a dark, horrible corner, but at least now i believe that
416 the behavior is the same as gawk and the intended posix standard.
417 thanks to arnold robbins for advice here.
419 Aug 9, 1997:
420 somewhat regretfully, replaced the ancient lex-based lexical
421 analyzer with one written in C. it's longer, generates less code,
422 and more portable; the old one depended too much on mysterious
423 properties of lex that were not preserved in other environments.
424 in theory these recognize the same language.
426 now using strtod to test whether a string is a number, instead of
427 the convoluted original function. should be more portable and
428 reliable if strtod is implemented right.
430 removed now-pointless optimization in makefile that tries to avoid
431 recompilation when awkgram.y is changed but symbols are not.
433 removed most fixed-size arrays, though a handful remain, some
434 of which are unchecked. you have been warned.
436 Aug 4, 1997:
437 with some trepidation, replaced the ancient code that managed
438 fields and $0 in fixed-size arrays with arrays that grow on
439 demand. there is still some tension between trying to make this
440 run fast and making it clean; not sure it's right yet.
442 the ill-conceived -mr and -mf arguments are now useful only
443 for debugging. previous dynamic string code removed.
445 numerous other minor cleanups along the way.
447 Jul 30, 1997:
448 using code provided by dan levy (to whom profuse thanks), replaced
449 fixed-size arrays and awkward kludges by a fairly uniform mechanism
450 to grow arrays as needed for printf, sub, gsub, etc.
452 Jul 23, 1997:
453 falling off the end of a function returns "" and 0, not 0.
454 thanks to arnold robbins.
456 Jun 17, 1997:
457 replaced several fixed-size arrays by dynamically-created ones
458 in run.c; added overflow tests to some previously unchecked cases.
459 getline, toupper, tolower.
461 getline code is still broken in that recursive calls may wind
462 up using the same space. [fixed later]
464 increased RECSIZE to 8192 to push problems further over the horizon.
466 added \r to \n as input line separator for programs, not data.
467 damn CRLFs.
469 modified format() to permit explicit printf("%c", 0) to include
470 a null byte in output. thanks to ken stailey for the fix.
472 added a "-safe" argument that disables file output (print >,
473 print >>), process creation (cmd|getline, print |, system), and
474 access to the environment (ENVIRON). this is a first approximation
475 to a "safe" version of awk, but don't rely on it too much. thanks
476 to joan feigenbaum and matt blaze for the inspiration long ago.
478 Jul 8, 1996:
479 fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to
480 ralph corderoy.
482 Jun 29, 1996:
483 fixed awful bug in new field splitting; didn't get all the places
484 where input was done.
486 Jun 28, 1996:
487 changed field-splitting to conform to posix definition: fields are
488 split using the value of FS at the time of input; it used to be
489 the value when the field or NF was first referred to, a much less
490 predictable definition. thanks to arnold robbins for encouragement
491 to do the right thing.
493 May 28, 1996:
494 fixed appalling but apparently unimportant bug in parsing octal
495 numbers in reg exprs.
497 explicit hex in reg exprs now limited to 2 chars: \xa, \xaa.
499 May 27, 1996:
500 cleaned up some declarations so gcc -Wall is now almost silent.
502 makefile now includes backup copies of ytab.c and lexyy.c in case
503 one makes before looking; it also avoids recreating lexyy.c unless
504 really needed.
506 s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes
507 with unwisely-written header files.
509 thanks to jeffrey friedl for several of these.
511 May 26, 1996:
512 an attempt to rationalize the (unsigned) char issue. almost all
513 instances of unsigned char have been removed; the handful of places
514 in b.c where chars are used as table indices have been hand-crafted.
515 added some latin-1 tests to the regression, but i'm not confident;
516 none of my compilers seem to care much. thanks to nelson beebe for
517 pointing out some others that do care.
519 May 2, 1996:
520 removed all register declarations.
522 enhanced split(), as in gawk, etc: split(s, a, "") splits s into
523 a[1]...a[length(s)] with each character a single element.
525 made the same changes for field-splitting if FS is "".
527 added nextfile, as in gawk: causes immediate advance to next
528 input file. (thanks to arnold robbins for inspiration and code).
530 small fixes to regexpr code: can now handle []], [[], and
531 variants; [] is now a syntax error, rather than matching
532 everything; [z-a] is now empty, not z. far from complete
533 or correct, however. (thanks to jeffrey friedl for pointing out
534 some awful behaviors.)
536 Apr 29, 1996:
537 replaced uchar by uschar everywhere; apparently some compilers
538 usurp this name and this causes conflicts.
540 fixed call to time in run.c (bltin); arg is time_t *.
542 replaced horrible pointer/long punning in b.c by a legitimate
543 union. should be safer on 64-bit machines and cleaner everywhere.
544 (thanks to nelson beebe for pointing out some of these problems.)
546 replaced nested comments by #if 0...#endif in run.c, lib.c.
548 removed getsval, setsval, execute macros from run.c and lib.c.
549 machines are 100x faster than they were when these macros were
550 first used.
552 revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l,
553 y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of
554 portability to nameless systems.
556 "make bundle" now includes yacc and lex output files for recipients
557 who don't have yacc or lex.
559 Aug 15, 1995:
560 initialized Cells in setsymtab more carefully; some fields
561 were not set. (thanks to purify, all of whose complaints i
562 think i now understand.)
564 fixed at least one error in gsub that looked at -1-th element
565 of an array when substituting for a null match (e.g., $).
567 delete arrayname is now legal; it clears the elements but leaves
568 the array, which may not be the right behavior.
570 modified makefile: my current make can't cope with the test used
571 to avoid unnecessary yacc invocations.
573 Jul 17, 1995:
574 added dynamically growing strings to awk.lx.l and b.c
575 to permit regular expressions to be much bigger.
576 the state arrays can still overflow.
578 Aug 24, 1994:
579 detect duplicate arguments in function definitions (mdm).
581 May 11, 1994:
582 trivial fix to printf to limit string size in sub().
584 Apr 22, 1994:
585 fixed yet another subtle self-assignment problem:
586 $1 = $2; $1 = $1 clobbered $1.
588 Regression tests now use private echo, to avoid quoting problems.
590 Feb 2, 1994:
591 changed error() to print line number as %d, not %g.
593 Jul 23, 1993:
594 cosmetic changes: increased sizes of some arrays,
595 reworded some error messages.
597 added CONVFMT as in posix (just replaced OFMT in getsval)
599 FILENAME is now "" until the first thing that causes a file
600 to be opened.
602 Nov 28, 1992:
603 deleted yyunput and yyoutput from proto.h;
604 different versions of lex give these different declarations.
606 May 31, 1992:
607 added -mr N and -mf N options: more record and fields.
608 these really ought to adjust automatically.
610 cleaned up some error messages; "out of space" now means
611 malloc returned NULL in all cases.
613 changed rehash so that if it runs out, it just returns;
614 things will continue to run slow, but maybe a bit longer.
616 Apr 24, 1992:
617 remove redundant close of stdin when using -f -.
619 got rid of core dump with -d; awk -d just prints date.
621 Apr 12, 1992:
622 added explicit check for /dev/std(in,out,err) in redirection.
623 unlike gawk, no /dev/fd/n yet.
625 added (file/pipe) builtin. hard to test satisfactorily.
626 not posix.
628 Feb 20, 1992:
629 recompile after abortive changes; should be unchanged.
631 Dec 2, 1991:
632 die-casting time: converted to ansi C, installed that.
634 Nov 30, 1991:
635 fixed storage leak in freefa, failing to recover [N]CCL.
636 thanks to Bill Jones (jones@cs.usask.ca)
638 Nov 19, 1991:
639 use RAND_MAX instead of literal in builtin().
641 Nov 12, 1991:
642 cranked up some fixed-size arrays in b.c, and added a test for
643 overflow in penter. thanks to mark larsen.
645 Sep 24, 1991:
646 increased buffer in gsub. a very crude fix to a general problem.
647 and again on Sep 26.
649 Aug 18, 1991:
650 enforce variable name syntax for commandline variables: has to
651 start with letter or _.
653 Jul 27, 1991:
654 allow newline after ; in for statements.
656 Jul 21, 1991:
657 fixed so that in self-assignment like $1=$1, side effects
658 like recomputing $0 take place. (this is getting subtle.)
660 Jun 30, 1991:
661 better test for detecting too-long output record.
663 Jun 2, 1991:
664 better defense against very long printf strings.
665 made break and continue illegal outside of loops.
667 May 13, 1991:
668 removed extra arg on gettemp, tempfree. minor error message rewording.
670 May 6, 1991:
671 fixed silly bug in hex parsing in hexstr().
672 removed an apparently unnecessary test in isnumber().
673 warn about weird printf conversions.
674 fixed unchecked array overwrite in relex().
676 changed for (i in array) to access elements in sorted order.
677 then unchanged it -- it really does run slower in too many cases.
678 left the code in place, commented out.
680 Feb 10, 1991:
681 check error status on all writes, to avoid banging on full disks.
683 Jan 28, 1991:
684 awk -f - reads the program from stdin.
686 Jan 11, 1991:
687 failed to set numeric state on $0 in cmd|getline context in run.c.
689 Nov 2, 1990:
690 fixed sleazy test for integrality in getsval; use modf.
692 Oct 29, 1990:
693 fixed sleazy buggy code in lib.c that looked (incorrectly) for
694 too long input lines.
696 Oct 14, 1990:
697 fixed the bug on p. 198 in which it couldn't deduce that an
698 argument was an array in some contexts. replaced the error
699 message in intest() by code that damn well makes it an array.
701 Oct 8, 1990:
702 fixed horrible bug: types and values were not preserved in
703 some kinds of self-assignment. (in assign().)
705 Aug 24, 1990:
706 changed NCHARS to 256 to handle 8-bit characters in strings
707 presented to match(), etc.
709 Jun 26, 1990:
710 changed struct rrow (awk.h) to use long instead of int for lval,
711 since cfoll() stores a pointer in it. now works better when int's
712 are smaller than pointers!
714 May 6, 1990:
715 AVA fixed the grammar so that ! is uniformly of the same precedence as
716 unary + and -. This renders illegal some constructs like !x=y, which
717 now has to be parenthesized as !(x=y), and makes others work properly:
718 !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements.
719 (These problems were pointed out by Bob Lenk of Posix.)
721 Added \x to regular expressions (already in strings).
722 Limited octal to octal digits; \8 and \9 are not octal.
723 Centralized the code for parsing escapes in regular expressions.
724 Added a bunch of tests to T.re and T.sub to verify some of this.
726 Feb 9, 1990:
727 fixed null pointer dereference bug in main.c: -F[nothing]. sigh.
729 restored srand behavior: it returns the current seed.
731 Jan 18, 1990:
732 srand now returns previous seed value (0 to start).
734 Jan 5, 1990:
735 fix potential problem in tran.c -- something was freed,
736 then used in freesymtab.
738 Oct 18, 1989:
739 another try to get the max number of open files set with
740 relatively machine-independent code.
742 small fix to input() in case of multiple reads after EOF.
744 Oct 11, 1989:
745 FILENAME is now defined in the BEGIN block -- too many old
746 programs broke.
748 "-" means stdin in getline as well as on the commandline.
750 added a bunch of casts to the code to tell the truth about
751 char * vs. unsigned char *, a right royal pain. added a
752 setlocale call to the front of main, though probably no one
753 has it usefully implemented yet.
755 Aug 24, 1989:
756 removed redundant relational tests against nullnode if parse
757 tree already had a relational at that point.
759 Aug 11, 1989:
760 fixed bug: commandline variable assignment has to look like
761 var=something. (consider the man page for =, in file =.1)
763 changed number of arguments to functions to static arrays
764 to avoid repeated malloc calls.
766 Aug 2, 1989:
767 restored -F (space) separator
769 Jul 30, 1989:
770 added -v x=1 y=2 ... for immediate commandline variable assignment;
771 done before the BEGIN block for sure. they have to precede the
772 program if the program is on the commandline.
773 Modified Aug 2 to require a separate -v for each assignment.
775 Jul 10, 1989:
776 fixed ref-thru-zero bug in environment code in tran.c
778 Jun 23, 1989:
779 add newline to usage message.
781 Jun 14, 1989:
782 added some missing ansi printf conversion letters: %i %X %E %G.
783 no sensible meaning for h or L, so they may not do what one expects.
785 made %* conversions work.
787 changed x^y so that if n is a positive integer, it's done
788 by explicit multiplication, thus achieving maximum accuracy.
789 (this should be done by pow() but it seems not to be locally.)
790 done to x ^= y as well.
792 Jun 4, 1989:
793 ENVIRON array contains environment: if shell variable V=thing,
794 ENVIRON["V"] is "thing"
796 multiple -f arguments permitted. error reporting is naive.
797 (they were permitted before, but only the last was used.)
799 fixed a really stupid botch in the debugging macro dprintf
801 fixed order of evaluation of commandline assignments to match
802 what the book claims: an argument of the form x=e is evaluated
803 at the time it would have been opened if it were a filename (p 63).
804 this invalidates the suggested answer to ex 4-1 (p 195).
806 removed some code that permitted -F (space) fieldseparator,
807 since it didn't quite work right anyway. (restored aug 2)
809 Apr 27, 1989:
810 Line number now accumulated correctly for comment lines.
812 Apr 26, 1989:
813 Debugging output now includes a version date,
814 if one compiles it into the source each time.
816 Apr 9, 1989:
817 Changed grammar to prohibit constants as 3rd arg of sub and gsub;
818 prevents class of overwriting-a-constant errors. (Last one?)
819 This invalidates the "banana" example on page 43 of the book.
821 Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal),
822 as in ANSI, for strings. Rescinded the sloppiness that permitted
823 non-octal digits in \ooo. Warning: not all compilers and libraries
824 will be able to deal with \x correctly.
826 Jan 9, 1989:
827 Fixed bug that caused tempcell list to contain a duplicate.
828 The fix is kludgy.
830 Dec 17, 1988:
831 Catches some more commandline errors in main.
832 Removed redundant decl of modf in run.c (confuses some compilers).
833 Warning: there's no single declaration of malloc, etc., in awk.h
834 that seems to satisfy all compilers.
836 Dec 7, 1988:
837 Added a bit of code to error printing to avoid printing nulls.
838 (Not clear that it actually would.)
840 Nov 27, 1988:
841 With fear and trembling, modified the grammar to permit
842 multiple pattern-action statements on one line without
843 an explicit separator. By definition, this capitulation
844 to the ghost of ancient implementations remains undefined
845 and thus subject to change without notice or apology.
846 DO NOT COUNT ON IT.
848 Oct 30, 1988:
849 Fixed bug in call() that failed to recover storage.
851 A warning is now generated if there are more arguments
852 in the call than in the definition (in lieu of fixing
853 another storage leak).
855 Oct 20, 1988:
856 Fixed %c: if expr is numeric, use numeric value;
857 otherwise print 1st char of string value. still
858 doesn't work if the value is 0 -- won't print \0.
860 Added a few more checks for running out of malloc.
862 Oct 12, 1988:
863 Fixed bug in call() that freed local arrays twice.
865 Fixed to handle deletion of non-existent array right;
866 complains about attempt to delete non-array element.
868 Sep 30, 1988:
869 Now guarantees to evaluate all arguments of built-in
870 functions, as in C; the appearance is that arguments
871 are evaluated before the function is called. Places
872 affected are sub (gsub was ok), substr, printf, and
873 all the built-in arithmetic functions in bltin().
874 A warning is generated if a bltin() is called with
875 the wrong number of arguments.
877 This requires changing makeprof on p167 of the book.
879 Aug 23, 1988:
880 setting FILENAME in BEGIN caused core dump, apparently
881 because it was freeing space not allocated by malloc.
883 July 24, 1988:
884 fixed egregious error in toupper/tolower functions.
885 still subject to rescinding, however.
887 July 2, 1988:
888 flush stdout before opening file or pipe
890 July 2, 1988:
891 performance bug in b.c/cgoto(): not freeing some sets of states.
892 partial fix only right now, and the number of states increased
893 to make it less obvious.
895 June 1, 1988:
896 check error status on close
898 May 28, 1988:
899 srand returns seed value it's using.
900 see 1/18/90
902 May 22, 1988:
903 Removed limit on depth of function calls.
905 May 10, 1988:
906 Fixed lib.c to permit _ in commandline variable names.
908 Mar 25, 1988:
909 main.c fixed to recognize -- as terminator of command-
910 line options. Illegal options flagged.
911 Error reporting slightly cleaned up.
913 Dec 2, 1987:
914 Newer C compilers apply a strict scope rule to extern
915 declarations within functions. Two extern declarations in
916 lib.c and tran.c have been moved to obviate this problem.
918 Oct xx, 1987:
919 Reluctantly added toupper and tolower functions.
920 Subject to rescinding without notice.
922 Sep 17, 1987:
923 Error-message printer had printf(s) instead of
924 printf("%s",s); got core dumps when the message
925 included a %.
927 Sep 12, 1987:
928 Very long printf strings caused core dump;
929 fixed aprintf, asprintf, format to catch them.
930 Can still get a core dump in printf itself.