1 /****************************************************************
2 Copyright (C) Lucent Technologies 1997
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
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
23 ****************************************************************/
25 This file lists all bug fixes, changes, etc., made since the AWK book
26 was sent to the printers in August, 1987.
29 fixed a bug in regular expressions that dates (so help me) from 1977;
30 it's been there from the beginning. an anchored longest match that
31 was longer than the number of states triggered a failure to initialize
32 the machine properly. many thanks to moinak ghosh for not only finding
33 this one but for providing a fix, in some of the most mysterious
36 fixed a storage leak in call() that appears to have been there since
37 1983 or so -- a function without an explicit return that assigns a
38 string to a parameter leaked a Cell. thanks to moinak ghosh for
39 spotting this very subtle one.
42 fixed, thanks to andrey chernov and ruslan ermilov, a bug in lex.c
43 that mis-handled the character 255 in input. (it was being compared
44 to EOF with a signed comparison.)
47 fixed (i think) the long-standing botch that included the beginning of
48 line state ^ for RE's in the set of valid characters; this led to a
49 variety of odd problems, including failure to properly match certain
50 regular expressions in non-US locales. thanks to ruslan for keeping
54 n-th try at getting internationalization right, with thanks to volker
55 kiefel, arnold robbins and ruslan ermilov for advice, though they
56 should not be blamed for the outcome. according to posix, "." is the
57 radix character in programs and command line arguments regardless of
58 the locale; otherwise, the locale should prevail for input and output
59 of numbers. so it's intended to work that way.
61 i have rescinded the attempt to use strcoll in expanding shorthands in
62 regular expressions (cclenter). its properties are much too
63 surprising; for example [a-c] matches aAbBc in locale en_US but abBcC
64 in locale fr_CA. i can see how this might arise by implementation
65 but i cannot explain it to a human user. (this behavior can be seen
66 in gawk as well; we're leaning on the same library.)
68 the issue appears to be that strcoll is meant for sorting, where
69 merging upper and lower case may make sense (though note that unix
70 sort does not do this by default either). it is not appropriate
71 for regular expressions, where the goal is to match specific
72 patterns of characters. in any case, the notations [:lower:], etc.,
73 are available in awk, and they are more likely to work correctly in
76 a moratorium is hereby declared on internationalization changes.
77 i apologize to friends and colleagues in other parts of the world.
78 i would truly like to get this "right", but i don't know what
79 that is, and i do not want to keep making changes until it's clear.
82 fixed bug that permitted non-terminated RE, as in "awk /x".
85 subtle change to split: if source is empty, number of elems
86 is always 0 and the array is not set.
89 added some parens to isblank, in another attempt to make things
90 internationally portable.
93 the internationalization changes, somewhat modified, are now
94 reinstated. in theory awk will now do character comparisons
95 and case conversions in national language, but "." will always
96 be the decimal point separator on input and output regardless
97 of national language. isblank(){} has an #ifndef.
99 this no longer compiles on windows: LC_MESSAGES isn't defined
102 fixed subtle behavior in field and record splitting: if FS is
103 a single character and RS is not empty, \n is NOT a separator.
104 this tortuous reading is found in the awk book; behavior now
105 matches gawk and mawk.
108 for the moment, the internationalization changes of nov 29 are
109 rolled back -- programs like x = 1.2 don't work in some locales,
110 because the parser is expecting x = 1,2. until i understand this
111 better, this will have to wait.
114 modified b.c (with tiny changes in main and run) to support
115 locales, using strcoll and iswhatever tests for posix character
116 classes. thanks to ruslan ermilov (ru@freebsd.org) for code.
117 the function isblank doesn't seem to have propagated to any
118 header file near me, so it's there explicitly. not properly
119 tested on non-ascii character sets by me.
122 modified run/format() and tran/getsval() to do a slightly better
123 job on using OFMT for output from print and CONVFMT for other
124 number->string conversions, as promised by posix and done by
125 gawk and mawk. there are still places where it doesn't work
126 right if CONVFMT is changed; by then the STR attribute of the
127 variable has been irrevocably set. thanks to arnold robbins for
130 fixed subtle bug in format that could get core dump. thanks to
131 Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing.
132 minor cleanup in run.c / format() at the same time.
134 added some tests for null pointers to debugging printf's, which
135 were never intended for external consumption. thanks to dave
136 kerns (dkerns@lucent.com) for pointing this out.
138 GNU compatibility: an empty regexp matches anything (thanks to
139 dag-erling smorgrav, des@ofug.org). subject to reversion if
140 this does more harm than good.
142 pervasive small changes to make things more const-correct, as
143 reported by gcc's -Wwrite-strings. as it says in the gcc manual,
144 this may be more nuisance than useful. provoked by a suggestion
145 and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk
147 minor documentation changes to note that this now compiles out
148 of the box on Mac OS X.
151 changed types in posix chars structure to quiet solaris cc.
154 fflush() or fflush("") flushes all files and pipes.
156 length(arrayname) returns number of elements; thanks to
157 arnold robbins for suggestion.
159 added a makefile.win to make it easier to build on windows.
160 based on dan allen's buildwin.bat.
163 added support for posix character class names like [:digit:],
164 which are not exactly shorter than [0-9] and perhaps no more
165 portable. thanks to dag-erling smorgrav for code.
168 removed -m option; no longer needed, and it was actually
169 broken (noted thanks to volker kiefel).
172 fixed an appalling bug in gettok: any sequence of digits, +,-, E, e,
173 and period was accepted as a valid number if it started with a period.
174 this would never have happened with the lex version.
176 other 1-character botches, now fixed, include a bare $ and a
177 bare " at the end of the input.
180 more (const char *) casts in b.c and tran.c to silence warnings.
183 fixed a bug introduced in august 1997 that caused expressions
184 like $f[1] to be syntax errors. thanks to arnold robbins for
185 noticing this and providing a fix.
188 fixed some nextfile bugs: not handling all cases. thanks to
189 arnold robbins for pointing this out. new regressions added.
191 close() is now a function. it returns whatever the library
192 fclose returns, and -1 for closing a file or pipe that wasn't
196 permit \n explicitly in character classes; won't work right
197 if comes in as "[\n]" but ok as /[\n]/, because of multiple
198 processing of \'s. thanks to arnold robbins.
201 minor fiddles in tran.c to keep compilers happy about uschar.
202 thanks to norman wilson.
205 yet another attempt at making 8-bit input work, with another
206 band-aid in b.c (member()), and some (uschar) casts to head
207 off potential errors in subscripts (like isdigit). also
208 changed HAT to NCHARS-2. thanks again to santiago vila.
210 changed maketab.c to ignore apparently out of range definitions
211 instead of halting; new freeBSD generates one. thanks to
212 jon snader <jsnader@ix.netcom.com> for pointing out the problem.
215 fixed an 8-bit problem in b.c by making several char*'s into
216 unsigned char*'s. not clear i have them all yet. thanks to
217 Santiago Vila <sanvila@unex.es> for the bug report.
220 finally found and fixed a memory leak in function call; it's
221 been there since functions were added ~1983. thanks to
222 jon bentley for the test case that found it.
224 added test in envinit to catch environment "variables" with
225 names begining with '='; thanks to Berend Hasselman.
228 added test in defn() to catch function foo(foo), which
229 otherwise recurses until core dump. thanks to arnold
230 robbins for noticing this.
233 added *bp in gettok in lex.c; appears possible to exit function
234 without terminating the string. thanks to russ cox.
237 added function stdinit() to run to initialize files[] array,
238 in case stdin, etc., are not constants; some compilers care.
241 replaced the ERROR ... FATAL, etc., macros with functions
242 based on vprintf, to avoid problems caused by overrunning
243 fixed-size errbuf array. thanks to ralph corderoy for the
244 impetus, and for pointing out a string termination bug in
248 fixed bug that caused occasional core dumps with commandline
249 variable with value ending in \. (thanks to nelson beebe for
253 with code kindly provided by Bruce Lilly, awk now parses
254 /=/ and similar constructs more sensibly in more places.
255 Bruce also provided some helpful test cases.
258 changed true/false to True/False in run.c to make it
259 easier to compile with C++. Added some casts on malloc
260 and realloc to be honest about casts; ditto. changed
261 ltype int to long in struct rrow to reduce some 64-bit
262 complaints; other changes scattered throughout for the
263 same purpose. thanks to Nelson Beebe for these portability
266 removed some horrible pointer-int casting in b.c and elsewhere
267 by adding ptoi and itonp to localize the casts, which are
268 all benign. fixed one incipient bug that showed up on sgi
271 reset lineno for new source file; include filename in error
272 message. also fixed line number error in continuation lines.
273 (thanks to Nelson Beebe for both of these.)
276 Nelson Beebe notes that irix 5.3 yacc dies with a bogus
277 error; use a newer version or switch to bison, since sgi
278 is unlikely to fix it.
281 changed isnumber to is_number to avoid the problem caused by
282 versions of ctype.h that include the name isnumber.
284 distribution now includes a script for building on a Mac,
288 fixed memory leaks in run.c (call) and tran.c (setfval).
289 thanks to Stephen Nutt for finding these and providing the fixes.
292 replaced srand argument by (unsigned int) in run.c;
293 avoids problem on Mac and potentially on Unix & Windows.
296 added a few (int) casts to silence useless compiler warnings.
297 e.g., errorflag= in run.c jump().
299 added proctab.c to the bundle outout; one less thing
300 to have to compile out of the box.
302 added calls to _popen and _pclose to the win95 stub for
303 pipes (thanks to Steve Adams for this helpful suggestion).
304 seems to work, though properties are not well understood
305 by me, and it appears that under some circumstances the
306 pipe output is truncated. Be careful.
309 fixed a couple of bugs in getrec: could fail to update $0
310 after a getline var; because inputFS wasn't initialized,
311 could split $0 on every character, a misleading diversion.
313 fixed caching bug in makedfa: LRU was actually removing
316 thanks to ross ridge for finding these, and for providing
320 fixed potential bug in readrec: might fail to update record
321 pointer after growing. thanks to dan levy for spotting this
322 and suggesting the fix.
325 added -V to print version number and die.
328 subtle silent bug in lex.c: if the program ended with a number
329 longer than 1 digit, part of the input would be pushed back and
330 parsed again because token buffer wasn't terminated right.
331 example: awk 'length($0) > 10'. blush. at least i found it
335 s/adelete/awkdelete/: SGI uses this in malloc.h.
336 thanks to nelson beebe for pointing this one out.
339 fixed some bugs in sub and gsub when replacement includes \\.
340 this is a dark, horrible corner, but at least now i believe that
341 the behavior is the same as gawk and the intended posix standard.
342 thanks to arnold robbins for advice here.
345 somewhat regretfully, replaced the ancient lex-based lexical
346 analyzer with one written in C. it's longer, generates less code,
347 and more portable; the old one depended too much on mysterious
348 properties of lex that were not preserved in other environments.
349 in theory these recognize the same language.
351 now using strtod to test whether a string is a number, instead of
352 the convoluted original function. should be more portable and
353 reliable if strtod is implemented right.
355 removed now-pointless optimization in makefile that tries to avoid
356 recompilation when awkgram.y is changed but symbols are not.
358 removed most fixed-size arrays, though a handful remain, some
359 of which are unchecked. you have been warned.
362 with some trepidation, replaced the ancient code that managed
363 fields and $0 in fixed-size arrays with arrays that grow on
364 demand. there is still some tension between trying to make this
365 run fast and making it clean; not sure it's right yet.
367 the ill-conceived -mr and -mf arguments are now useful only
368 for debugging. previous dynamic string code removed.
370 numerous other minor cleanups along the way.
373 using code provided by dan levy (to whom profuse thanks), replaced
374 fixed-size arrays and awkward kludges by a fairly uniform mechanism
375 to grow arrays as needed for printf, sub, gsub, etc.
378 falling off the end of a function returns "" and 0, not 0.
379 thanks to arnold robbins.
382 replaced several fixed-size arrays by dynamically-created ones
383 in run.c; added overflow tests to some previously unchecked cases.
384 getline, toupper, tolower.
386 getline code is still broken in that recursive calls may wind
387 up using the same space. [fixed later]
389 increased RECSIZE to 8192 to push problems further over the horizon.
391 added \r to \n as input line separator for programs, not data.
394 modified format() to permit explicit printf("%c", 0) to include
395 a null byte in output. thanks to ken stailey for the fix.
397 added a "-safe" argument that disables file output (print >,
398 print >>), process creation (cmd|getline, print |, system), and
399 access to the environment (ENVIRON). this is a first approximation
400 to a "safe" version of awk, but don't rely on it too much. thanks
401 to joan feigenbaum and matt blaze for the inspiration long ago.
404 fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to
408 fixed awful bug in new field splitting; didn't get all the places
409 where input was done.
412 changed field-splitting to conform to posix definition: fields are
413 split using the value of FS at the time of input; it used to be
414 the value when the field or NF was first referred to, a much less
415 predictable definition. thanks to arnold robbins for encouragement
416 to do the right thing.
419 fixed appalling but apparently unimportant bug in parsing octal
420 numbers in reg exprs.
422 explicit hex in reg exprs now limited to 2 chars: \xa, \xaa.
425 cleaned up some declarations so gcc -Wall is now almost silent.
427 makefile now includes backup copies of ytab.c and lexyy.c in case
428 one makes before looking; it also avoids recreating lexyy.c unless
431 s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes
432 with unwisely-written header files.
434 thanks to jeffrey friedl for several of these.
437 an attempt to rationalize the (unsigned) char issue. almost all
438 instances of unsigned char have been removed; the handful of places
439 in b.c where chars are used as table indices have been hand-crafted.
440 added some latin-1 tests to the regression, but i'm not confident;
441 none of my compilers seem to care much. thanks to nelson beebe for
442 pointing out some others that do care.
445 removed all register declarations.
447 enhanced split(), as in gawk, etc: split(s, a, "") splits s into
448 a[1]...a[length(s)] with each character a single element.
450 made the same changes for field-splitting if FS is "".
452 added nextfile, as in gawk: causes immediate advance to next
453 input file. (thanks to arnold robbins for inspiration and code).
455 small fixes to regexpr code: can now handle []], [[], and
456 variants; [] is now a syntax error, rather than matching
457 everything; [z-a] is now empty, not z. far from complete
458 or correct, however. (thanks to jeffrey friedl for pointing out
459 some awful behaviors.)
462 replaced uchar by uschar everwhere; apparently some compilers
463 usurp this name and this causes conflicts.
465 fixed call to time in run.c (bltin); arg is time_t *.
467 replaced horrible pointer/long punning in b.c by a legitimate
468 union. should be safer on 64-bit machines and cleaner everywhere.
469 (thanks to nelson beebe for pointing out some of these problems.)
471 replaced nested comments by #if 0...#endif in run.c, lib.c.
473 removed getsval, setsval, execute macros from run.c and lib.c.
474 machines are 100x faster than they were when these macros were
477 revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l,
478 y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of
479 portability to nameless systems.
481 "make bundle" now includes yacc and lex output files for recipients
482 who don't have yacc or lex.
485 initialized Cells in setsymtab more carefully; some fields
486 were not set. (thanks to purify, all of whose complaints i
487 think i now understand.)
489 fixed at least one error in gsub that looked at -1-th element
490 of an array when substituting for a null match (e.g., $).
492 delete arrayname is now legal; it clears the elements but leaves
493 the array, which may not be the right behavior.
495 modified makefile: my current make can't cope with the test used
496 to avoid unnecessary yacc invocations.
499 added dynamically growing strings to awk.lx.l and b.c
500 to permit regular expressions to be much bigger.
501 the state arrays can still overflow.
504 detect duplicate arguments in function definitions (mdm).
507 trivial fix to printf to limit string size in sub().
510 fixed yet another subtle self-assignment problem:
511 $1 = $2; $1 = $1 clobbered $1.
513 Regression tests now use private echo, to avoid quoting problems.
516 changed error() to print line number as %d, not %g.
519 cosmetic changes: increased sizes of some arrays,
520 reworded some error messages.
522 added CONVFMT as in posix (just replaced OFMT in getsval)
524 FILENAME is now "" until the first thing that causes a file
528 deleted yyunput and yyoutput from proto.h;
529 different versions of lex give these different declarations.
532 added -mr N and -mf N options: more record and fields.
533 these really ought to adjust automatically.
535 cleaned up some error messages; "out of space" now means
536 malloc returned NULL in all cases.
538 changed rehash so that if it runs out, it just returns;
539 things will continue to run slow, but maybe a bit longer.
542 remove redundant close of stdin when using -f -.
544 got rid of core dump with -d; awk -d just prints date.
547 added explicit check for /dev/std(in,out,err) in redirection.
548 unlike gawk, no /dev/fd/n yet.
550 added (file/pipe) builtin. hard to test satisfactorily.
554 recompile after abortive changes; should be unchanged.
557 die-casting time: converted to ansi C, installed that.
560 fixed storage leak in freefa, failing to recover [N]CCL.
561 thanks to Bill Jones (jones@cs.usask.ca)
564 use RAND_MAX instead of literal in builtin().
567 cranked up some fixed-size arrays in b.c, and added a test for
568 overflow in penter. thanks to mark larsen.
571 increased buffer in gsub. a very crude fix to a general problem.
575 enforce variable name syntax for commandline variables: has to
576 start with letter or _.
579 allow newline after ; in for statements.
582 fixed so that in self-assignment like $1=$1, side effects
583 like recomputing $0 take place. (this is getting subtle.)
586 better test for detecting too-long output record.
589 better defense against very long printf strings.
590 made break and continue illegal outside of loops.
593 removed extra arg on gettemp, tempfree. minor error message rewording.
596 fixed silly bug in hex parsing in hexstr().
597 removed an apparently unnecessary test in isnumber().
598 warn about weird printf conversions.
599 fixed unchecked array overwrite in relex().
601 changed for (i in array) to access elements in sorted order.
602 then unchanged it -- it really does run slower in too many cases.
603 left the code in place, commented out.
606 check error status on all writes, to avoid banging on full disks.
609 awk -f - reads the program from stdin.
612 failed to set numeric state on $0 in cmd|getline context in run.c.
615 fixed sleazy test for integrality in getsval; use modf.
618 fixed sleazy buggy code in lib.c that looked (incorrectly) for
619 too long input lines.
622 fixed the bug on p. 198 in which it couldn't deduce that an
623 argument was an array in some contexts. replaced the error
624 message in intest() by code that damn well makes it an array.
627 fixed horrible bug: types and values were not preserved in
628 some kinds of self-assignment. (in assign().)
631 changed NCHARS to 256 to handle 8-bit characters in strings
632 presented to match(), etc.
635 changed struct rrow (awk.h) to use long instead of int for lval,
636 since cfoll() stores a pointer in it. now works better when int's
637 are smaller than pointers!
640 AVA fixed the grammar so that ! is uniformly of the same precedence as
641 unary + and -. This renders illegal some constructs like !x=y, which
642 now has to be parenthesized as !(x=y), and makes others work properly:
643 !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements.
644 (These problems were pointed out by Bob Lenk of Posix.)
646 Added \x to regular expressions (already in strings).
647 Limited octal to octal digits; \8 and \9 are not octal.
648 Centralized the code for parsing escapes in regular expressions.
649 Added a bunch of tests to T.re and T.sub to verify some of this.
652 fixed null pointer dereference bug in main.c: -F[nothing]. sigh.
654 restored srand behavior: it returns the current seed.
657 srand now returns previous seed value (0 to start).
660 fix potential problem in tran.c -- something was freed,
661 then used in freesymtab.
664 another try to get the max number of open files set with
665 relatively machine-independent code.
667 small fix to input() in case of multiple reads after EOF.
670 FILENAME is now defined in the BEGIN block -- too many old
673 "-" means stdin in getline as well as on the commandline.
675 added a bunch of casts to the code to tell the truth about
676 char * vs. unsigned char *, a right royal pain. added a
677 setlocale call to the front of main, though probably no one
678 has it usefully implemented yet.
681 removed redundant relational tests against nullnode if parse
682 tree already had a relational at that point.
685 fixed bug: commandline variable assignment has to look like
686 var=something. (consider the man page for =, in file =.1)
688 changed number of arguments to functions to static arrays
689 to avoid repeated malloc calls.
692 restored -F (space) separator
695 added -v x=1 y=2 ... for immediate commandline variable assignment;
696 done before the BEGIN block for sure. they have to precede the
697 program if the program is on the commandline.
698 Modified Aug 2 to require a separate -v for each assignment.
701 fixed ref-thru-zero bug in environment code in tran.c
704 add newline to usage message.
707 added some missing ansi printf conversion letters: %i %X %E %G.
708 no sensible meaning for h or L, so they may not do what one expects.
710 made %* conversions work.
712 changed x^y so that if n is a positive integer, it's done
713 by explicit multiplication, thus achieving maximum accuracy.
714 (this should be done by pow() but it seems not to be locally.)
715 done to x ^= y as well.
718 ENVIRON array contains environment: if shell variable V=thing,
719 ENVIRON["V"] is "thing"
721 multiple -f arguments permitted. error reporting is naive.
722 (they were permitted before, but only the last was used.)
724 fixed a really stupid botch in the debugging macro dprintf
726 fixed order of evaluation of commandline assignments to match
727 what the book claims: an argument of the form x=e is evaluated
728 at the time it would have been opened if it were a filename (p 63).
729 this invalidates the suggested answer to ex 4-1 (p 195).
731 removed some code that permitted -F (space) fieldseparator,
732 since it didn't quite work right anyway. (restored aug 2)
735 Line number now accumulated correctly for comment lines.
738 Debugging output now includes a version date,
739 if one compiles it into the source each time.
742 Changed grammar to prohibit constants as 3rd arg of sub and gsub;
743 prevents class of overwriting-a-constant errors. (Last one?)
744 This invalidates the "banana" example on page 43 of the book.
746 Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal),
747 as in ANSI, for strings. Rescinded the sloppiness that permitted
748 non-octal digits in \ooo. Warning: not all compilers and libraries
749 will be able to deal with \x correctly.
752 Fixed bug that caused tempcell list to contain a duplicate.
756 Catches some more commandline errors in main.
757 Removed redundant decl of modf in run.c (confuses some compilers).
758 Warning: there's no single declaration of malloc, etc., in awk.h
759 that seems to satisfy all compilers.
762 Added a bit of code to error printing to avoid printing nulls.
763 (Not clear that it actually would.)
766 With fear and trembling, modified the grammar to permit
767 multiple pattern-action statements on one line without
768 an explicit separator. By definition, this capitulation
769 to the ghost of ancient implementations remains undefined
770 and thus subject to change without notice or apology.
774 Fixed bug in call() that failed to recover storage.
776 A warning is now generated if there are more arguments
777 in the call than in the definition (in lieu of fixing
778 another storage leak).
781 Fixed %c: if expr is numeric, use numeric value;
782 otherwise print 1st char of string value. still
783 doesn't work if the value is 0 -- won't print \0.
785 Added a few more checks for running out of malloc.
788 Fixed bug in call() that freed local arrays twice.
790 Fixed to handle deletion of non-existent array right;
791 complains about attempt to delete non-array element.
794 Now guarantees to evaluate all arguments of built-in
795 functions, as in C; the appearance is that arguments
796 are evaluated before the function is called. Places
797 affected are sub (gsub was ok), substr, printf, and
798 all the built-in arithmetic functions in bltin().
799 A warning is generated if a bltin() is called with
800 the wrong number of arguments.
802 This requires changing makeprof on p167 of the book.
805 setting FILENAME in BEGIN caused core dump, apparently
806 because it was freeing space not allocated by malloc.
809 fixed egregious error in toupper/tolower functions.
810 still subject to rescinding, however.
813 flush stdout before opening file or pipe
816 performance bug in b.c/cgoto(): not freeing some sets of states.
817 partial fix only right now, and the number of states increased
818 to make it less obvious.
821 check error status on close
824 srand returns seed value it's using.
828 Removed limit on depth of function calls.
831 Fixed lib.c to permit _ in commandline variable names.
834 main.c fixed to recognize -- as terminator of command-
835 line options. Illegal options flagged.
836 Error reporting slightly cleaned up.
839 Newer C compilers apply a strict scope rule to extern
840 declarations within functions. Two extern declarations in
841 lib.c and tran.c have been moved to obviate this problem.
844 Reluctantly added toupper and tolower functions.
845 Subject to rescinding without notice.
848 Error-message printer had printf(s) instead of
849 printf("%s",s); got core dumps when the message
853 Very long printf strings caused core dump;
854 fixed aprintf, asprintf, format to catch them.
855 Can still get a core dump in printf itself.