| blob | 2f39d4827c101951ac5d9eacdd7abdaf9e75fbf2 |
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 Oct 23, 2007:
29 minor fix in lib.c: increase inputFS to 100, change malloc
30 for fields to n+1.
32 fixed memory fault caused by out of order test in setsval.
34 thanks to david o'brien, freebsd, for both fixes.
36 May 1, 2007:
37 fiddle in makefile to fix for BSD make; thanks to igor sobrado.
39 Mar 31, 2007:
40 fixed some null pointer refs calling adjbuf.
42 Feb 21, 2007:
43 fixed a bug in matching the null RE in sub and gsub. thanks to al aho
44 who actually did the fix (in b.c), and to wolfgang seeberg for finding
45 it and providing a very compact test case.
47 fixed quotation in b.c; thanks to Hal Pratt and the Princeton Dante
48 Project.
50 removed some no-effect asserts in run.c.
52 fiddled maketab.c to not complain about bison-generated values.
54 removed the obsolete -V argument; fixed --version to print the
55 version and exit.
57 fixed wording and an outright error in the usage message; thanks to igor
58 sobrado and jason mcintyre.
60 fixed a bug in -d that caused core dump if no program followed.
62 Jan 1, 2007:
63 dropped mac.code from makefile; there are few non-MacOSX
64 mac's these days.
66 Jan 17, 2006:
67 system() not flagged as unsafe in the unadvertised -safe option.
68 found it while enhancing tests before shipping the ;login: article.
69 practice what you preach.
71 removed the 9-years-obsolete -mr and -mf flags.
73 added -version and --version options.
75 core dump on linux with BEGIN {nextfile}, now fixed.
77 removed some #ifdef's in run.c and lex.c that appear to no
78 longer be necessary.
80 Apr 24, 2005:
81 modified lib.c so that values of $0 et al are preserved in the END
82 block, apparently as required by posix. thanks to havard eidnes
83 for the report and code.
85 Jan 14, 2005:
86 fixed infinite loop in parsing, originally found by brian tsang.
87 thanks to arnold robbins for a suggestion that started me
88 rethinking it.
90 Dec 31, 2004:
91 prevent overflow of -f array in main, head off potential error in
92 call of SYNTAX(), test malloc return in lib.c, all with thanks to
93 todd miller.
95 Dec 22, 2004:
96 cranked up size of NCHARS; coverity thinks it can be overrun with
97 smaller size, and i think that's right. added some assertions to b.c
98 to catch places where it might overrun. the RE code is still fragile.
100 Dec 5, 2004:
101 fixed a couple of overflow problems with ridiculous field numbers:
102 e.g., print $(2^32-1). thanks to ruslan ermilov, giorgos keramidas
103 and david o'brien at freebsd.org for patches. this really should
104 be re-done from scratch.
106 Nov 21, 2004:
107 fixed another 25-year-old RE bug, in split. it's another failure
108 to (re-)initialize. thanks to steve fisher for spotting this and
109 providing a good test case.
111 Nov 22, 2003:
112 fixed a bug in regular expressions that dates (so help me) from 1977;
113 it's been there from the beginning. an anchored longest match that
114 was longer than the number of states triggered a failure to initialize
115 the machine properly. many thanks to moinak ghosh for not only finding
116 this one but for providing a fix, in some of the most mysterious
117 code known to man.
119 fixed a storage leak in call() that appears to have been there since
120 1983 or so -- a function without an explicit return that assigns a
121 string to a parameter leaked a Cell. thanks to moinak ghosh for
122 spotting this very subtle one.
124 Jul 31, 2003:
125 fixed, thanks to andrey chernov and ruslan ermilov, a bug in lex.c
126 that mis-handled the character 255 in input. (it was being compared
127 to EOF with a signed comparison.)
129 Jul 29, 2003:
130 fixed (i think) the long-standing botch that included the beginning of
131 line state ^ for RE's in the set of valid characters; this led to a
132 variety of odd problems, including failure to properly match certain
133 regular expressions in non-US locales. thanks to ruslan for keeping
134 at this one.
136 Jul 28, 2003:
137 n-th try at getting internationalization right, with thanks to volker
138 kiefel, arnold robbins and ruslan ermilov for advice, though they
139 should not be blamed for the outcome. according to posix, "." is the
140 radix character in programs and command line arguments regardless of
141 the locale; otherwise, the locale should prevail for input and output
142 of numbers. so it's intended to work that way.
144 i have rescinded the attempt to use strcoll in expanding shorthands in
145 regular expressions (cclenter). its properties are much too
146 surprising; for example [a-c] matches aAbBc in locale en_US but abBcC
147 in locale fr_CA. i can see how this might arise by implementation
148 but i cannot explain it to a human user. (this behavior can be seen
149 in gawk as well; we're leaning on the same library.)
151 the issue appears to be that strcoll is meant for sorting, where
152 merging upper and lower case may make sense (though note that unix
153 sort does not do this by default either). it is not appropriate
154 for regular expressions, where the goal is to match specific
155 patterns of characters. in any case, the notations [:lower:], etc.,
156 are available in awk, and they are more likely to work correctly in
157 most locales.
159 a moratorium is hereby declared on internationalization changes.
160 i apologize to friends and colleagues in other parts of the world.
161 i would truly like to get this "right", but i don't know what
162 that is, and i do not want to keep making changes until it's clear.
164 Jul 4, 2003:
165 fixed bug that permitted non-terminated RE, as in "awk /x".
167 Jun 1, 2003:
168 subtle change to split: if source is empty, number of elems
169 is always 0 and the array is not set.
171 Mar 21, 2003:
172 added some parens to isblank, in another attempt to make things
173 internationally portable.
175 Mar 14, 2003:
176 the internationalization changes, somewhat modified, are now
177 reinstated. in theory awk will now do character comparisons
178 and case conversions in national language, but "." will always
179 be the decimal point separator on input and output regardless
180 of national language. isblank(){} has an #ifndef.
182 this no longer compiles on windows: LC_MESSAGES isn't defined
183 in vc6++.
185 fixed subtle behavior in field and record splitting: if FS is
186 a single character and RS is not empty, \n is NOT a separator.
187 this tortuous reading is found in the awk book; behavior now
188 matches gawk and mawk.
190 Dec 13, 2002:
191 for the moment, the internationalization changes of nov 29 are
192 rolled back -- programs like x = 1.2 don't work in some locales,
193 because the parser is expecting x = 1,2. until i understand this
194 better, this will have to wait.
196 Nov 29, 2002:
197 modified b.c (with tiny changes in main and run) to support
198 locales, using strcoll and iswhatever tests for posix character
199 classes. thanks to ruslan ermilov (ru@freebsd.org) for code.
200 the function isblank doesn't seem to have propagated to any
201 header file near me, so it's there explicitly. not properly
202 tested on non-ascii character sets by me.
204 Jun 28, 2002:
205 modified run/format() and tran/getsval() to do a slightly better
206 job on using OFMT for output from print and CONVFMT for other
207 number->string conversions, as promised by posix and done by
208 gawk and mawk. there are still places where it doesn't work
209 right if CONVFMT is changed; by then the STR attribute of the
210 variable has been irrevocably set. thanks to arnold robbins for
211 code and examples.
213 fixed subtle bug in format that could get core dump. thanks to
214 Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing.
215 minor cleanup in run.c / format() at the same time.
217 added some tests for null pointers to debugging printf's, which
218 were never intended for external consumption. thanks to dave
219 kerns (dkerns@lucent.com) for pointing this out.
221 GNU compatibility: an empty regexp matches anything (thanks to
222 dag-erling smorgrav, des@ofug.org). subject to reversion if
223 this does more harm than good.
225 pervasive small changes to make things more const-correct, as
226 reported by gcc's -Wwrite-strings. as it says in the gcc manual,
227 this may be more nuisance than useful. provoked by a suggestion
228 and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk
230 minor documentation changes to note that this now compiles out
231 of the box on Mac OS X.
233 Feb 10, 2002:
234 changed types in posix chars structure to quiet solaris cc.
236 Jan 1, 2002:
237 fflush() or fflush("") flushes all files and pipes.
239 length(arrayname) returns number of elements; thanks to
240 arnold robbins for suggestion.
242 added a makefile.win to make it easier to build on windows.
243 based on dan allen's buildwin.bat.
245 Nov 16, 2001:
246 added support for posix character class names like [:digit:],
247 which are not exactly shorter than [0-9] and perhaps no more
248 portable. thanks to dag-erling smorgrav for code.
250 Feb 16, 2001:
251 removed -m option; no longer needed, and it was actually
252 broken (noted thanks to volker kiefel).
254 Feb 10, 2001:
255 fixed an appalling bug in gettok: any sequence of digits, +,-, E, e,
256 and period was accepted as a valid number if it started with a period.
257 this would never have happened with the lex version.
259 other 1-character botches, now fixed, include a bare $ and a
260 bare " at the end of the input.
262 Feb 7, 2001:
263 more (const char *) casts in b.c and tran.c to silence warnings.
265 Nov 15, 2000:
266 fixed a bug introduced in august 1997 that caused expressions
267 like $f[1] to be syntax errors. thanks to arnold robbins for
268 noticing this and providing a fix.
270 Oct 30, 2000:
271 fixed some nextfile bugs: not handling all cases. thanks to
272 arnold robbins for pointing this out. new regressions added.
274 close() is now a function. it returns whatever the library
275 fclose returns, and -1 for closing a file or pipe that wasn't
276 opened.
278 Sep 24, 2000:
279 permit \n explicitly in character classes; won't work right
280 if comes in as "[\n]" but ok as /[\n]/, because of multiple
281 processing of \'s. thanks to arnold robbins.
283 July 5, 2000:
284 minor fiddles in tran.c to keep compilers happy about uschar.
285 thanks to norman wilson.
287 May 25, 2000:
288 yet another attempt at making 8-bit input work, with another
289 band-aid in b.c (member()), and some (uschar) casts to head
290 off potential errors in subscripts (like isdigit). also
291 changed HAT to NCHARS-2. thanks again to santiago vila.
293 changed maketab.c to ignore apparently out of range definitions
294 instead of halting; new freeBSD generates one. thanks to
295 jon snader <jsnader@ix.netcom.com> for pointing out the problem.
297 May 2, 2000:
298 fixed an 8-bit problem in b.c by making several char*'s into
299 unsigned char*'s. not clear i have them all yet. thanks to
300 Santiago Vila <sanvila@unex.es> for the bug report.
302 Apr 21, 2000:
303 finally found and fixed a memory leak in function call; it's
304 been there since functions were added ~1983. thanks to
305 jon bentley for the test case that found it.
307 added test in envinit to catch environment "variables" with
308 names beginning with '='; thanks to Berend Hasselman.
310 Jul 28, 1999:
311 added test in defn() to catch function foo(foo), which
312 otherwise recurses until core dump. thanks to arnold
313 robbins for noticing this.
315 Jun 20, 1999:
316 added *bp in gettok in lex.c; appears possible to exit function
317 without terminating the string. thanks to russ cox.
319 Jun 2, 1999:
320 added function stdinit() to run to initialize files[] array,
321 in case stdin, etc., are not constants; some compilers care.
323 May 10, 1999:
324 replaced the ERROR ... FATAL, etc., macros with functions
325 based on vprintf, to avoid problems caused by overrunning
326 fixed-size errbuf array. thanks to ralph corderoy for the
327 impetus, and for pointing out a string termination bug in
328 qstring as well.
330 Apr 21, 1999:
331 fixed bug that caused occasional core dumps with commandline
332 variable with value ending in \. (thanks to nelson beebe for
333 the test case.)
335 Apr 16, 1999:
336 with code kindly provided by Bruce Lilly, awk now parses
337 /=/ and similar constructs more sensibly in more places.
338 Bruce also provided some helpful test cases.
340 Apr 5, 1999:
341 changed true/false to True/False in run.c to make it
342 easier to compile with C++. Added some casts on malloc
343 and realloc to be honest about casts; ditto. changed
344 ltype int to long in struct rrow to reduce some 64-bit
345 complaints; other changes scattered throughout for the
346 same purpose. thanks to Nelson Beebe for these portability
347 improvements.
349 removed some horrible pointer-int casting in b.c and elsewhere
350 by adding ptoi and itonp to localize the casts, which are
351 all benign. fixed one incipient bug that showed up on sgi
352 in 64-bit mode.
354 reset lineno for new source file; include filename in error
355 message. also fixed line number error in continuation lines.
356 (thanks to Nelson Beebe for both of these.)
358 Mar 24, 1999:
359 Nelson Beebe notes that irix 5.3 yacc dies with a bogus
360 error; use a newer version or switch to bison, since sgi
361 is unlikely to fix it.
363 Mar 5, 1999:
364 changed isnumber to is_number to avoid the problem caused by
365 versions of ctype.h that include the name isnumber.
367 distribution now includes a script for building on a Mac,
368 thanks to Dan Allen.
370 Feb 20, 1999:
371 fixed memory leaks in run.c (call) and tran.c (setfval).
372 thanks to Stephen Nutt for finding these and providing the fixes.
374 Jan 13, 1999:
375 replaced srand argument by (unsigned int) in run.c;
376 avoids problem on Mac and potentially on Unix & Windows.
377 thanks to Dan Allen.
379 added a few (int) casts to silence useless compiler warnings.
380 e.g., errorflag= in run.c jump().
382 added proctab.c to the bundle outout; one less thing
383 to have to compile out of the box.
385 added calls to _popen and _pclose to the win95 stub for
386 pipes (thanks to Steve Adams for this helpful suggestion).
387 seems to work, though properties are not well understood
388 by me, and it appears that under some circumstances the
389 pipe output is truncated. Be careful.
391 Oct 19, 1998:
392 fixed a couple of bugs in getrec: could fail to update $0
393 after a getline var; because inputFS wasn't initialized,
394 could split $0 on every character, a misleading diversion.
396 fixed caching bug in makedfa: LRU was actually removing
397 least often used.
399 thanks to ross ridge for finding these, and for providing
400 great bug reports.
402 May 12, 1998:
403 fixed potential bug in readrec: might fail to update record
404 pointer after growing. thanks to dan levy for spotting this
405 and suggesting the fix.
407 Mar 12, 1998:
408 added -V to print version number and die.
410 Feb 11, 1998:
411 subtle silent bug in lex.c: if the program ended with a number
412 longer than 1 digit, part of the input would be pushed back and
413 parsed again because token buffer wasn't terminated right.
414 example: awk 'length($0) > 10'. blush. at least i found it
415 myself.
417 Aug 31, 1997:
418 s/adelete/awkdelete/: SGI uses this in malloc.h.
419 thanks to nelson beebe for pointing this one out.
421 Aug 21, 1997:
422 fixed some bugs in sub and gsub when replacement includes \\.
423 this is a dark, horrible corner, but at least now i believe that
424 the behavior is the same as gawk and the intended posix standard.
425 thanks to arnold robbins for advice here.
427 Aug 9, 1997:
428 somewhat regretfully, replaced the ancient lex-based lexical
429 analyzer with one written in C. it's longer, generates less code,
430 and more portable; the old one depended too much on mysterious
431 properties of lex that were not preserved in other environments.
432 in theory these recognize the same language.
434 now using strtod to test whether a string is a number, instead of
435 the convoluted original function. should be more portable and
436 reliable if strtod is implemented right.
438 removed now-pointless optimization in makefile that tries to avoid
439 recompilation when awkgram.y is changed but symbols are not.
441 removed most fixed-size arrays, though a handful remain, some
442 of which are unchecked. you have been warned.
444 Aug 4, 1997:
445 with some trepidation, replaced the ancient code that managed
446 fields and $0 in fixed-size arrays with arrays that grow on
447 demand. there is still some tension between trying to make this
448 run fast and making it clean; not sure it's right yet.
450 the ill-conceived -mr and -mf arguments are now useful only
451 for debugging. previous dynamic string code removed.
453 numerous other minor cleanups along the way.
455 Jul 30, 1997:
456 using code provided by dan levy (to whom profuse thanks), replaced
457 fixed-size arrays and awkward kludges by a fairly uniform mechanism
458 to grow arrays as needed for printf, sub, gsub, etc.
460 Jul 23, 1997:
461 falling off the end of a function returns "" and 0, not 0.
462 thanks to arnold robbins.
464 Jun 17, 1997:
465 replaced several fixed-size arrays by dynamically-created ones
466 in run.c; added overflow tests to some previously unchecked cases.
467 getline, toupper, tolower.
469 getline code is still broken in that recursive calls may wind
470 up using the same space. [fixed later]
472 increased RECSIZE to 8192 to push problems further over the horizon.
474 added \r to \n as input line separator for programs, not data.
475 damn CRLFs.
477 modified format() to permit explicit printf("%c", 0) to include
478 a null byte in output. thanks to ken stailey for the fix.
480 added a "-safe" argument that disables file output (print >,
481 print >>), process creation (cmd|getline, print |, system), and
482 access to the environment (ENVIRON). this is a first approximation
483 to a "safe" version of awk, but don't rely on it too much. thanks
484 to joan feigenbaum and matt blaze for the inspiration long ago.
486 Jul 8, 1996:
487 fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to
488 ralph corderoy.
490 Jun 29, 1996:
491 fixed awful bug in new field splitting; didn't get all the places
492 where input was done.
494 Jun 28, 1996:
495 changed field-splitting to conform to posix definition: fields are
496 split using the value of FS at the time of input; it used to be
497 the value when the field or NF was first referred to, a much less
498 predictable definition. thanks to arnold robbins for encouragement
499 to do the right thing.
501 May 28, 1996:
502 fixed appalling but apparently unimportant bug in parsing octal
503 numbers in reg exprs.
505 explicit hex in reg exprs now limited to 2 chars: \xa, \xaa.
507 May 27, 1996:
508 cleaned up some declarations so gcc -Wall is now almost silent.
510 makefile now includes backup copies of ytab.c and lexyy.c in case
511 one makes before looking; it also avoids recreating lexyy.c unless
512 really needed.
514 s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes
515 with unwisely-written header files.
517 thanks to jeffrey friedl for several of these.
519 May 26, 1996:
520 an attempt to rationalize the (unsigned) char issue. almost all
521 instances of unsigned char have been removed; the handful of places
522 in b.c where chars are used as table indices have been hand-crafted.
523 added some latin-1 tests to the regression, but i'm not confident;
524 none of my compilers seem to care much. thanks to nelson beebe for
525 pointing out some others that do care.
527 May 2, 1996:
528 removed all register declarations.
530 enhanced split(), as in gawk, etc: split(s, a, "") splits s into
531 a[1]...a[length(s)] with each character a single element.
533 made the same changes for field-splitting if FS is "".
535 added nextfile, as in gawk: causes immediate advance to next
536 input file. (thanks to arnold robbins for inspiration and code).
538 small fixes to regexpr code: can now handle []], [[], and
539 variants; [] is now a syntax error, rather than matching
540 everything; [z-a] is now empty, not z. far from complete
541 or correct, however. (thanks to jeffrey friedl for pointing out
542 some awful behaviors.)
544 Apr 29, 1996:
545 replaced uchar by uschar everywhere; apparently some compilers
546 usurp this name and this causes conflicts.
548 fixed call to time in run.c (bltin); arg is time_t *.
550 replaced horrible pointer/long punning in b.c by a legitimate
551 union. should be safer on 64-bit machines and cleaner everywhere.
552 (thanks to nelson beebe for pointing out some of these problems.)
554 replaced nested comments by #if 0...#endif in run.c, lib.c.
556 removed getsval, setsval, execute macros from run.c and lib.c.
557 machines are 100x faster than they were when these macros were
558 first used.
560 revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l,
561 y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of
562 portability to nameless systems.
564 "make bundle" now includes yacc and lex output files for recipients
565 who don't have yacc or lex.
567 Aug 15, 1995:
568 initialized Cells in setsymtab more carefully; some fields
569 were not set. (thanks to purify, all of whose complaints i
570 think i now understand.)
572 fixed at least one error in gsub that looked at -1-th element
573 of an array when substituting for a null match (e.g., $).
575 delete arrayname is now legal; it clears the elements but leaves
576 the array, which may not be the right behavior.
578 modified makefile: my current make can't cope with the test used
579 to avoid unnecessary yacc invocations.
581 Jul 17, 1995:
582 added dynamically growing strings to awk.lx.l and b.c
583 to permit regular expressions to be much bigger.
584 the state arrays can still overflow.
586 Aug 24, 1994:
587 detect duplicate arguments in function definitions (mdm).
589 May 11, 1994:
590 trivial fix to printf to limit string size in sub().
592 Apr 22, 1994:
593 fixed yet another subtle self-assignment problem:
594 $1 = $2; $1 = $1 clobbered $1.
596 Regression tests now use private echo, to avoid quoting problems.
598 Feb 2, 1994:
599 changed error() to print line number as %d, not %g.
601 Jul 23, 1993:
602 cosmetic changes: increased sizes of some arrays,
603 reworded some error messages.
605 added CONVFMT as in posix (just replaced OFMT in getsval)
607 FILENAME is now "" until the first thing that causes a file
608 to be opened.
610 Nov 28, 1992:
611 deleted yyunput and yyoutput from proto.h;
612 different versions of lex give these different declarations.
614 May 31, 1992:
615 added -mr N and -mf N options: more record and fields.
616 these really ought to adjust automatically.
618 cleaned up some error messages; "out of space" now means
619 malloc returned NULL in all cases.
621 changed rehash so that if it runs out, it just returns;
622 things will continue to run slow, but maybe a bit longer.
624 Apr 24, 1992:
625 remove redundant close of stdin when using -f -.
627 got rid of core dump with -d; awk -d just prints date.
629 Apr 12, 1992:
630 added explicit check for /dev/std(in,out,err) in redirection.
631 unlike gawk, no /dev/fd/n yet.
633 added (file/pipe) builtin. hard to test satisfactorily.
634 not posix.
636 Feb 20, 1992:
637 recompile after abortive changes; should be unchanged.
639 Dec 2, 1991:
640 die-casting time: converted to ansi C, installed that.
642 Nov 30, 1991:
643 fixed storage leak in freefa, failing to recover [N]CCL.
644 thanks to Bill Jones (jones@cs.usask.ca)
646 Nov 19, 1991:
647 use RAND_MAX instead of literal in builtin().
649 Nov 12, 1991:
650 cranked up some fixed-size arrays in b.c, and added a test for
651 overflow in penter. thanks to mark larsen.
653 Sep 24, 1991:
654 increased buffer in gsub. a very crude fix to a general problem.
655 and again on Sep 26.
657 Aug 18, 1991:
658 enforce variable name syntax for commandline variables: has to
659 start with letter or _.
661 Jul 27, 1991:
662 allow newline after ; in for statements.
664 Jul 21, 1991:
665 fixed so that in self-assignment like $1=$1, side effects
666 like recomputing $0 take place. (this is getting subtle.)
668 Jun 30, 1991:
669 better test for detecting too-long output record.
671 Jun 2, 1991:
672 better defense against very long printf strings.
673 made break and continue illegal outside of loops.
675 May 13, 1991:
676 removed extra arg on gettemp, tempfree. minor error message rewording.
678 May 6, 1991:
679 fixed silly bug in hex parsing in hexstr().
680 removed an apparently unnecessary test in isnumber().
681 warn about weird printf conversions.
682 fixed unchecked array overwrite in relex().
684 changed for (i in array) to access elements in sorted order.
685 then unchanged it -- it really does run slower in too many cases.
686 left the code in place, commented out.
688 Feb 10, 1991:
689 check error status on all writes, to avoid banging on full disks.
691 Jan 28, 1991:
692 awk -f - reads the program from stdin.
694 Jan 11, 1991:
695 failed to set numeric state on $0 in cmd|getline context in run.c.
697 Nov 2, 1990:
698 fixed sleazy test for integrality in getsval; use modf.
700 Oct 29, 1990:
701 fixed sleazy buggy code in lib.c that looked (incorrectly) for
702 too long input lines.
704 Oct 14, 1990:
705 fixed the bug on p. 198 in which it couldn't deduce that an
706 argument was an array in some contexts. replaced the error
707 message in intest() by code that damn well makes it an array.
709 Oct 8, 1990:
710 fixed horrible bug: types and values were not preserved in
711 some kinds of self-assignment. (in assign().)
713 Aug 24, 1990:
714 changed NCHARS to 256 to handle 8-bit characters in strings
715 presented to match(), etc.
717 Jun 26, 1990:
718 changed struct rrow (awk.h) to use long instead of int for lval,
719 since cfoll() stores a pointer in it. now works better when int's
720 are smaller than pointers!
722 May 6, 1990:
723 AVA fixed the grammar so that ! is uniformly of the same precedence as
724 unary + and -. This renders illegal some constructs like !x=y, which
725 now has to be parenthesized as !(x=y), and makes others work properly:
726 !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements.
727 (These problems were pointed out by Bob Lenk of Posix.)
729 Added \x to regular expressions (already in strings).
730 Limited octal to octal digits; \8 and \9 are not octal.
731 Centralized the code for parsing escapes in regular expressions.
732 Added a bunch of tests to T.re and T.sub to verify some of this.
734 Feb 9, 1990:
735 fixed null pointer dereference bug in main.c: -F[nothing]. sigh.
737 restored srand behavior: it returns the current seed.
739 Jan 18, 1990:
740 srand now returns previous seed value (0 to start).
742 Jan 5, 1990:
743 fix potential problem in tran.c -- something was freed,
744 then used in freesymtab.
746 Oct 18, 1989:
747 another try to get the max number of open files set with
748 relatively machine-independent code.
750 small fix to input() in case of multiple reads after EOF.
752 Oct 11, 1989:
753 FILENAME is now defined in the BEGIN block -- too many old
754 programs broke.
756 "-" means stdin in getline as well as on the commandline.
758 added a bunch of casts to the code to tell the truth about
759 char * vs. unsigned char *, a right royal pain. added a
760 setlocale call to the front of main, though probably no one
761 has it usefully implemented yet.
763 Aug 24, 1989:
764 removed redundant relational tests against nullnode if parse
765 tree already had a relational at that point.
767 Aug 11, 1989:
768 fixed bug: commandline variable assignment has to look like
769 var=something. (consider the man page for =, in file =.1)
771 changed number of arguments to functions to static arrays
772 to avoid repeated malloc calls.
774 Aug 2, 1989:
775 restored -F (space) separator
777 Jul 30, 1989:
778 added -v x=1 y=2 ... for immediate commandline variable assignment;
779 done before the BEGIN block for sure. they have to precede the
780 program if the program is on the commandline.
781 Modified Aug 2 to require a separate -v for each assignment.
783 Jul 10, 1989:
784 fixed ref-thru-zero bug in environment code in tran.c
786 Jun 23, 1989:
787 add newline to usage message.
789 Jun 14, 1989:
790 added some missing ansi printf conversion letters: %i %X %E %G.
791 no sensible meaning for h or L, so they may not do what one expects.
793 made %* conversions work.
795 changed x^y so that if n is a positive integer, it's done
796 by explicit multiplication, thus achieving maximum accuracy.
797 (this should be done by pow() but it seems not to be locally.)
798 done to x ^= y as well.
800 Jun 4, 1989:
801 ENVIRON array contains environment: if shell variable V=thing,
802 ENVIRON["V"] is "thing"
804 multiple -f arguments permitted. error reporting is naive.
805 (they were permitted before, but only the last was used.)
807 fixed a really stupid botch in the debugging macro dprintf
809 fixed order of evaluation of commandline assignments to match
810 what the book claims: an argument of the form x=e is evaluated
811 at the time it would have been opened if it were a filename (p 63).
812 this invalidates the suggested answer to ex 4-1 (p 195).
814 removed some code that permitted -F (space) fieldseparator,
815 since it didn't quite work right anyway. (restored aug 2)
817 Apr 27, 1989:
818 Line number now accumulated correctly for comment lines.
820 Apr 26, 1989:
821 Debugging output now includes a version date,
822 if one compiles it into the source each time.
824 Apr 9, 1989:
825 Changed grammar to prohibit constants as 3rd arg of sub and gsub;
826 prevents class of overwriting-a-constant errors. (Last one?)
827 This invalidates the "banana" example on page 43 of the book.
829 Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal),
830 as in ANSI, for strings. Rescinded the sloppiness that permitted
831 non-octal digits in \ooo. Warning: not all compilers and libraries
832 will be able to deal with \x correctly.
834 Jan 9, 1989:
835 Fixed bug that caused tempcell list to contain a duplicate.
836 The fix is kludgy.
838 Dec 17, 1988:
839 Catches some more commandline errors in main.
840 Removed redundant decl of modf in run.c (confuses some compilers).
841 Warning: there's no single declaration of malloc, etc., in awk.h
842 that seems to satisfy all compilers.
844 Dec 7, 1988:
845 Added a bit of code to error printing to avoid printing nulls.
846 (Not clear that it actually would.)
848 Nov 27, 1988:
849 With fear and trembling, modified the grammar to permit
850 multiple pattern-action statements on one line without
851 an explicit separator. By definition, this capitulation
852 to the ghost of ancient implementations remains undefined
853 and thus subject to change without notice or apology.
854 DO NOT COUNT ON IT.
856 Oct 30, 1988:
857 Fixed bug in call() that failed to recover storage.
859 A warning is now generated if there are more arguments
860 in the call than in the definition (in lieu of fixing
861 another storage leak).
863 Oct 20, 1988:
864 Fixed %c: if expr is numeric, use numeric value;
865 otherwise print 1st char of string value. still
866 doesn't work if the value is 0 -- won't print \0.
868 Added a few more checks for running out of malloc.
870 Oct 12, 1988:
871 Fixed bug in call() that freed local arrays twice.
873 Fixed to handle deletion of non-existent array right;
874 complains about attempt to delete non-array element.
876 Sep 30, 1988:
877 Now guarantees to evaluate all arguments of built-in
878 functions, as in C; the appearance is that arguments
879 are evaluated before the function is called. Places
880 affected are sub (gsub was ok), substr, printf, and
881 all the built-in arithmetic functions in bltin().
882 A warning is generated if a bltin() is called with
883 the wrong number of arguments.
885 This requires changing makeprof on p167 of the book.
887 Aug 23, 1988:
888 setting FILENAME in BEGIN caused core dump, apparently
889 because it was freeing space not allocated by malloc.
891 July 24, 1988:
892 fixed egregious error in toupper/tolower functions.
893 still subject to rescinding, however.
895 July 2, 1988:
896 flush stdout before opening file or pipe
898 July 2, 1988:
899 performance bug in b.c/cgoto(): not freeing some sets of states.
900 partial fix only right now, and the number of states increased
901 to make it less obvious.
903 June 1, 1988:
904 check error status on close
906 May 28, 1988:
907 srand returns seed value it's using.
908 see 1/18/90
910 May 22, 1988:
911 Removed limit on depth of function calls.
913 May 10, 1988:
914 Fixed lib.c to permit _ in commandline variable names.
916 Mar 25, 1988:
917 main.c fixed to recognize -- as terminator of command-
918 line options. Illegal options flagged.
919 Error reporting slightly cleaned up.
921 Dec 2, 1987:
922 Newer C compilers apply a strict scope rule to extern
923 declarations within functions. Two extern declarations in
924 lib.c and tran.c have been moved to obviate this problem.
926 Oct xx, 1987:
927 Reluctantly added toupper and tolower functions.
928 Subject to rescinding without notice.
930 Sep 17, 1987:
931 Error-message printer had printf(s) instead of
932 printf("%s",s); got core dumps when the message
933 included a %.
935 Sep 12, 1987:
936 Very long printf strings caused core dump;
937 fixed aprintf, asprintf, format to catch them.
938 Can still get a core dump in printf itself.