| blob | e611e2a7c9a6103568ddb6903a78b93aef2bbe6d |
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 Feb 8, 2010:
29 i give up. replaced isblank with isspace in b.c; there are
30 no consistent header files.
32 Nov 26, 2009:
33 fixed a long-standing issue with when FS takes effect. a
34 change to FS is now noticed immediately for subsequent splits.
36 changed the name getline() to awkgetline() to avoid yet another
37 name conflict somewhere.
39 Feb 11, 2009:
40 temporarily for now defined HAS_ISBLANK, since that seems to
41 be the best way through the thicket. isblank arrived in C99,
42 but seems to be arriving at different systems at different
43 times.
45 Oct 8, 2008:
46 fixed typo in b.c that set tmpvec wrongly. no one had ever
47 run into the problem, apparently. thanks to alistair crooks.
49 Oct 23, 2007:
50 minor fix in lib.c: increase inputFS to 100, change malloc
51 for fields to n+1.
53 fixed memory fault caused by out of order test in setsval.
55 thanks to david o'brien, freebsd, for both fixes.
57 May 1, 2007:
58 fiddle in makefile to fix for BSD make; thanks to igor sobrado.
60 Mar 31, 2007:
61 fixed some null pointer refs calling adjbuf.
63 Feb 21, 2007:
64 fixed a bug in matching the null RE in sub and gsub. thanks to al aho
65 who actually did the fix (in b.c), and to wolfgang seeberg for finding
66 it and providing a very compact test case.
68 fixed quotation in b.c; thanks to Hal Pratt and the Princeton Dante
69 Project.
71 removed some no-effect asserts in run.c.
73 fiddled maketab.c to not complain about bison-generated values.
75 removed the obsolete -V argument; fixed --version to print the
76 version and exit.
78 fixed wording and an outright error in the usage message; thanks to igor
79 sobrado and jason mcintyre.
81 fixed a bug in -d that caused core dump if no program followed.
83 Jan 1, 2007:
84 dropped mac.code from makefile; there are few non-MacOSX
85 mac's these days.
87 Jan 17, 2006:
88 system() not flagged as unsafe in the unadvertised -safe option.
89 found it while enhancing tests before shipping the ;login: article.
90 practice what you preach.
92 removed the 9-years-obsolete -mr and -mf flags.
94 added -version and --version options.
96 core dump on linux with BEGIN {nextfile}, now fixed.
98 removed some #ifdef's in run.c and lex.c that appear to no
99 longer be necessary.
101 Apr 24, 2005:
102 modified lib.c so that values of $0 et al are preserved in the END
103 block, apparently as required by posix. thanks to havard eidnes
104 for the report and code.
106 Jan 14, 2005:
107 fixed infinite loop in parsing, originally found by brian tsang.
108 thanks to arnold robbins for a suggestion that started me
109 rethinking it.
111 Dec 31, 2004:
112 prevent overflow of -f array in main, head off potential error in
113 call of SYNTAX(), test malloc return in lib.c, all with thanks to
114 todd miller.
116 Dec 22, 2004:
117 cranked up size of NCHARS; coverity thinks it can be overrun with
118 smaller size, and i think that's right. added some assertions to b.c
119 to catch places where it might overrun. the RE code is still fragile.
121 Dec 5, 2004:
122 fixed a couple of overflow problems with ridiculous field numbers:
123 e.g., print $(2^32-1). thanks to ruslan ermilov, giorgos keramidas
124 and david o'brien at freebsd.org for patches. this really should
125 be re-done from scratch.
127 Nov 21, 2004:
128 fixed another 25-year-old RE bug, in split. it's another failure
129 to (re-)initialize. thanks to steve fisher for spotting this and
130 providing a good test case.
132 Nov 22, 2003:
133 fixed a bug in regular expressions that dates (so help me) from 1977;
134 it's been there from the beginning. an anchored longest match that
135 was longer than the number of states triggered a failure to initialize
136 the machine properly. many thanks to moinak ghosh for not only finding
137 this one but for providing a fix, in some of the most mysterious
138 code known to man.
140 fixed a storage leak in call() that appears to have been there since
141 1983 or so -- a function without an explicit return that assigns a
142 string to a parameter leaked a Cell. thanks to moinak ghosh for
143 spotting this very subtle one.
145 Jul 31, 2003:
146 fixed, thanks to andrey chernov and ruslan ermilov, a bug in lex.c
147 that mis-handled the character 255 in input. (it was being compared
148 to EOF with a signed comparison.)
150 Jul 29, 2003:
151 fixed (i think) the long-standing botch that included the beginning of
152 line state ^ for RE's in the set of valid characters; this led to a
153 variety of odd problems, including failure to properly match certain
154 regular expressions in non-US locales. thanks to ruslan for keeping
155 at this one.
157 Jul 28, 2003:
158 n-th try at getting internationalization right, with thanks to volker
159 kiefel, arnold robbins and ruslan ermilov for advice, though they
160 should not be blamed for the outcome. according to posix, "." is the
161 radix character in programs and command line arguments regardless of
162 the locale; otherwise, the locale should prevail for input and output
163 of numbers. so it's intended to work that way.
165 i have rescinded the attempt to use strcoll in expanding shorthands in
166 regular expressions (cclenter). its properties are much too
167 surprising; for example [a-c] matches aAbBc in locale en_US but abBcC
168 in locale fr_CA. i can see how this might arise by implementation
169 but i cannot explain it to a human user. (this behavior can be seen
170 in gawk as well; we're leaning on the same library.)
172 the issue appears to be that strcoll is meant for sorting, where
173 merging upper and lower case may make sense (though note that unix
174 sort does not do this by default either). it is not appropriate
175 for regular expressions, where the goal is to match specific
176 patterns of characters. in any case, the notations [:lower:], etc.,
177 are available in awk, and they are more likely to work correctly in
178 most locales.
180 a moratorium is hereby declared on internationalization changes.
181 i apologize to friends and colleagues in other parts of the world.
182 i would truly like to get this "right", but i don't know what
183 that is, and i do not want to keep making changes until it's clear.
185 Jul 4, 2003:
186 fixed bug that permitted non-terminated RE, as in "awk /x".
188 Jun 1, 2003:
189 subtle change to split: if source is empty, number of elems
190 is always 0 and the array is not set.
192 Mar 21, 2003:
193 added some parens to isblank, in another attempt to make things
194 internationally portable.
196 Mar 14, 2003:
197 the internationalization changes, somewhat modified, are now
198 reinstated. in theory awk will now do character comparisons
199 and case conversions in national language, but "." will always
200 be the decimal point separator on input and output regardless
201 of national language. isblank(){} has an #ifndef.
203 this no longer compiles on windows: LC_MESSAGES isn't defined
204 in vc6++.
206 fixed subtle behavior in field and record splitting: if FS is
207 a single character and RS is not empty, \n is NOT a separator.
208 this tortuous reading is found in the awk book; behavior now
209 matches gawk and mawk.
211 Dec 13, 2002:
212 for the moment, the internationalization changes of nov 29 are
213 rolled back -- programs like x = 1.2 don't work in some locales,
214 because the parser is expecting x = 1,2. until i understand this
215 better, this will have to wait.
217 Nov 29, 2002:
218 modified b.c (with tiny changes in main and run) to support
219 locales, using strcoll and iswhatever tests for posix character
220 classes. thanks to ruslan ermilov (ru@freebsd.org) for code.
221 the function isblank doesn't seem to have propagated to any
222 header file near me, so it's there explicitly. not properly
223 tested on non-ascii character sets by me.
225 Jun 28, 2002:
226 modified run/format() and tran/getsval() to do a slightly better
227 job on using OFMT for output from print and CONVFMT for other
228 number->string conversions, as promised by posix and done by
229 gawk and mawk. there are still places where it doesn't work
230 right if CONVFMT is changed; by then the STR attribute of the
231 variable has been irrevocably set. thanks to arnold robbins for
232 code and examples.
234 fixed subtle bug in format that could get core dump. thanks to
235 Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing.
236 minor cleanup in run.c / format() at the same time.
238 added some tests for null pointers to debugging printf's, which
239 were never intended for external consumption. thanks to dave
240 kerns (dkerns@lucent.com) for pointing this out.
242 GNU compatibility: an empty regexp matches anything (thanks to
243 dag-erling smorgrav, des@ofug.org). subject to reversion if
244 this does more harm than good.
246 pervasive small changes to make things more const-correct, as
247 reported by gcc's -Wwrite-strings. as it says in the gcc manual,
248 this may be more nuisance than useful. provoked by a suggestion
249 and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk
251 minor documentation changes to note that this now compiles out
252 of the box on Mac OS X.
254 Feb 10, 2002:
255 changed types in posix chars structure to quiet solaris cc.
257 Jan 1, 2002:
258 fflush() or fflush("") flushes all files and pipes.
260 length(arrayname) returns number of elements; thanks to
261 arnold robbins for suggestion.
263 added a makefile.win to make it easier to build on windows.
264 based on dan allen's buildwin.bat.
266 Nov 16, 2001:
267 added support for posix character class names like [:digit:],
268 which are not exactly shorter than [0-9] and perhaps no more
269 portable. thanks to dag-erling smorgrav for code.
271 Feb 16, 2001:
272 removed -m option; no longer needed, and it was actually
273 broken (noted thanks to volker kiefel).
275 Feb 10, 2001:
276 fixed an appalling bug in gettok: any sequence of digits, +,-, E, e,
277 and period was accepted as a valid number if it started with a period.
278 this would never have happened with the lex version.
280 other 1-character botches, now fixed, include a bare $ and a
281 bare " at the end of the input.
283 Feb 7, 2001:
284 more (const char *) casts in b.c and tran.c to silence warnings.
286 Nov 15, 2000:
287 fixed a bug introduced in august 1997 that caused expressions
288 like $f[1] to be syntax errors. thanks to arnold robbins for
289 noticing this and providing a fix.
291 Oct 30, 2000:
292 fixed some nextfile bugs: not handling all cases. thanks to
293 arnold robbins for pointing this out. new regressions added.
295 close() is now a function. it returns whatever the library
296 fclose returns, and -1 for closing a file or pipe that wasn't
297 opened.
299 Sep 24, 2000:
300 permit \n explicitly in character classes; won't work right
301 if comes in as "[\n]" but ok as /[\n]/, because of multiple
302 processing of \'s. thanks to arnold robbins.
304 July 5, 2000:
305 minor fiddles in tran.c to keep compilers happy about uschar.
306 thanks to norman wilson.
308 May 25, 2000:
309 yet another attempt at making 8-bit input work, with another
310 band-aid in b.c (member()), and some (uschar) casts to head
311 off potential errors in subscripts (like isdigit). also
312 changed HAT to NCHARS-2. thanks again to santiago vila.
314 changed maketab.c to ignore apparently out of range definitions
315 instead of halting; new freeBSD generates one. thanks to
316 jon snader <jsnader@ix.netcom.com> for pointing out the problem.
318 May 2, 2000:
319 fixed an 8-bit problem in b.c by making several char*'s into
320 unsigned char*'s. not clear i have them all yet. thanks to
321 Santiago Vila <sanvila@unex.es> for the bug report.
323 Apr 21, 2000:
324 finally found and fixed a memory leak in function call; it's
325 been there since functions were added ~1983. thanks to
326 jon bentley for the test case that found it.
328 added test in envinit to catch environment "variables" with
329 names beginning with '='; thanks to Berend Hasselman.
331 Jul 28, 1999:
332 added test in defn() to catch function foo(foo), which
333 otherwise recurses until core dump. thanks to arnold
334 robbins for noticing this.
336 Jun 20, 1999:
337 added *bp in gettok in lex.c; appears possible to exit function
338 without terminating the string. thanks to russ cox.
340 Jun 2, 1999:
341 added function stdinit() to run to initialize files[] array,
342 in case stdin, etc., are not constants; some compilers care.
344 May 10, 1999:
345 replaced the ERROR ... FATAL, etc., macros with functions
346 based on vprintf, to avoid problems caused by overrunning
347 fixed-size errbuf array. thanks to ralph corderoy for the
348 impetus, and for pointing out a string termination bug in
349 qstring as well.
351 Apr 21, 1999:
352 fixed bug that caused occasional core dumps with commandline
353 variable with value ending in \. (thanks to nelson beebe for
354 the test case.)
356 Apr 16, 1999:
357 with code kindly provided by Bruce Lilly, awk now parses
358 /=/ and similar constructs more sensibly in more places.
359 Bruce also provided some helpful test cases.
361 Apr 5, 1999:
362 changed true/false to True/False in run.c to make it
363 easier to compile with C++. Added some casts on malloc
364 and realloc to be honest about casts; ditto. changed
365 ltype int to long in struct rrow to reduce some 64-bit
366 complaints; other changes scattered throughout for the
367 same purpose. thanks to Nelson Beebe for these portability
368 improvements.
370 removed some horrible pointer-int casting in b.c and elsewhere
371 by adding ptoi and itonp to localize the casts, which are
372 all benign. fixed one incipient bug that showed up on sgi
373 in 64-bit mode.
375 reset lineno for new source file; include filename in error
376 message. also fixed line number error in continuation lines.
377 (thanks to Nelson Beebe for both of these.)
379 Mar 24, 1999:
380 Nelson Beebe notes that irix 5.3 yacc dies with a bogus
381 error; use a newer version or switch to bison, since sgi
382 is unlikely to fix it.
384 Mar 5, 1999:
385 changed isnumber to is_number to avoid the problem caused by
386 versions of ctype.h that include the name isnumber.
388 distribution now includes a script for building on a Mac,
389 thanks to Dan Allen.
391 Feb 20, 1999:
392 fixed memory leaks in run.c (call) and tran.c (setfval).
393 thanks to Stephen Nutt for finding these and providing the fixes.
395 Jan 13, 1999:
396 replaced srand argument by (unsigned int) in run.c;
397 avoids problem on Mac and potentially on Unix & Windows.
398 thanks to Dan Allen.
400 added a few (int) casts to silence useless compiler warnings.
401 e.g., errorflag= in run.c jump().
403 added proctab.c to the bundle outout; one less thing
404 to have to compile out of the box.
406 added calls to _popen and _pclose to the win95 stub for
407 pipes (thanks to Steve Adams for this helpful suggestion).
408 seems to work, though properties are not well understood
409 by me, and it appears that under some circumstances the
410 pipe output is truncated. Be careful.
412 Oct 19, 1998:
413 fixed a couple of bugs in getrec: could fail to update $0
414 after a getline var; because inputFS wasn't initialized,
415 could split $0 on every character, a misleading diversion.
417 fixed caching bug in makedfa: LRU was actually removing
418 least often used.
420 thanks to ross ridge for finding these, and for providing
421 great bug reports.
423 May 12, 1998:
424 fixed potential bug in readrec: might fail to update record
425 pointer after growing. thanks to dan levy for spotting this
426 and suggesting the fix.
428 Mar 12, 1998:
429 added -V to print version number and die.
431 Feb 11, 1998:
432 subtle silent bug in lex.c: if the program ended with a number
433 longer than 1 digit, part of the input would be pushed back and
434 parsed again because token buffer wasn't terminated right.
435 example: awk 'length($0) > 10'. blush. at least i found it
436 myself.
438 Aug 31, 1997:
439 s/adelete/awkdelete/: SGI uses this in malloc.h.
440 thanks to nelson beebe for pointing this one out.
442 Aug 21, 1997:
443 fixed some bugs in sub and gsub when replacement includes \\.
444 this is a dark, horrible corner, but at least now i believe that
445 the behavior is the same as gawk and the intended posix standard.
446 thanks to arnold robbins for advice here.
448 Aug 9, 1997:
449 somewhat regretfully, replaced the ancient lex-based lexical
450 analyzer with one written in C. it's longer, generates less code,
451 and more portable; the old one depended too much on mysterious
452 properties of lex that were not preserved in other environments.
453 in theory these recognize the same language.
455 now using strtod to test whether a string is a number, instead of
456 the convoluted original function. should be more portable and
457 reliable if strtod is implemented right.
459 removed now-pointless optimization in makefile that tries to avoid
460 recompilation when awkgram.y is changed but symbols are not.
462 removed most fixed-size arrays, though a handful remain, some
463 of which are unchecked. you have been warned.
465 Aug 4, 1997:
466 with some trepidation, replaced the ancient code that managed
467 fields and $0 in fixed-size arrays with arrays that grow on
468 demand. there is still some tension between trying to make this
469 run fast and making it clean; not sure it's right yet.
471 the ill-conceived -mr and -mf arguments are now useful only
472 for debugging. previous dynamic string code removed.
474 numerous other minor cleanups along the way.
476 Jul 30, 1997:
477 using code provided by dan levy (to whom profuse thanks), replaced
478 fixed-size arrays and awkward kludges by a fairly uniform mechanism
479 to grow arrays as needed for printf, sub, gsub, etc.
481 Jul 23, 1997:
482 falling off the end of a function returns "" and 0, not 0.
483 thanks to arnold robbins.
485 Jun 17, 1997:
486 replaced several fixed-size arrays by dynamically-created ones
487 in run.c; added overflow tests to some previously unchecked cases.
488 getline, toupper, tolower.
490 getline code is still broken in that recursive calls may wind
491 up using the same space. [fixed later]
493 increased RECSIZE to 8192 to push problems further over the horizon.
495 added \r to \n as input line separator for programs, not data.
496 damn CRLFs.
498 modified format() to permit explicit printf("%c", 0) to include
499 a null byte in output. thanks to ken stailey for the fix.
501 added a "-safe" argument that disables file output (print >,
502 print >>), process creation (cmd|getline, print |, system), and
503 access to the environment (ENVIRON). this is a first approximation
504 to a "safe" version of awk, but don't rely on it too much. thanks
505 to joan feigenbaum and matt blaze for the inspiration long ago.
507 Jul 8, 1996:
508 fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to
509 ralph corderoy.
511 Jun 29, 1996:
512 fixed awful bug in new field splitting; didn't get all the places
513 where input was done.
515 Jun 28, 1996:
516 changed field-splitting to conform to posix definition: fields are
517 split using the value of FS at the time of input; it used to be
518 the value when the field or NF was first referred to, a much less
519 predictable definition. thanks to arnold robbins for encouragement
520 to do the right thing.
522 May 28, 1996:
523 fixed appalling but apparently unimportant bug in parsing octal
524 numbers in reg exprs.
526 explicit hex in reg exprs now limited to 2 chars: \xa, \xaa.
528 May 27, 1996:
529 cleaned up some declarations so gcc -Wall is now almost silent.
531 makefile now includes backup copies of ytab.c and lexyy.c in case
532 one makes before looking; it also avoids recreating lexyy.c unless
533 really needed.
535 s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes
536 with unwisely-written header files.
538 thanks to jeffrey friedl for several of these.
540 May 26, 1996:
541 an attempt to rationalize the (unsigned) char issue. almost all
542 instances of unsigned char have been removed; the handful of places
543 in b.c where chars are used as table indices have been hand-crafted.
544 added some latin-1 tests to the regression, but i'm not confident;
545 none of my compilers seem to care much. thanks to nelson beebe for
546 pointing out some others that do care.
548 May 2, 1996:
549 removed all register declarations.
551 enhanced split(), as in gawk, etc: split(s, a, "") splits s into
552 a[1]...a[length(s)] with each character a single element.
554 made the same changes for field-splitting if FS is "".
556 added nextfile, as in gawk: causes immediate advance to next
557 input file. (thanks to arnold robbins for inspiration and code).
559 small fixes to regexpr code: can now handle []], [[], and
560 variants; [] is now a syntax error, rather than matching
561 everything; [z-a] is now empty, not z. far from complete
562 or correct, however. (thanks to jeffrey friedl for pointing out
563 some awful behaviors.)
565 Apr 29, 1996:
566 replaced uchar by uschar everywhere; apparently some compilers
567 usurp this name and this causes conflicts.
569 fixed call to time in run.c (bltin); arg is time_t *.
571 replaced horrible pointer/long punning in b.c by a legitimate
572 union. should be safer on 64-bit machines and cleaner everywhere.
573 (thanks to nelson beebe for pointing out some of these problems.)
575 replaced nested comments by #if 0...#endif in run.c, lib.c.
577 removed getsval, setsval, execute macros from run.c and lib.c.
578 machines are 100x faster than they were when these macros were
579 first used.
581 revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l,
582 y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of
583 portability to nameless systems.
585 "make bundle" now includes yacc and lex output files for recipients
586 who don't have yacc or lex.
588 Aug 15, 1995:
589 initialized Cells in setsymtab more carefully; some fields
590 were not set. (thanks to purify, all of whose complaints i
591 think i now understand.)
593 fixed at least one error in gsub that looked at -1-th element
594 of an array when substituting for a null match (e.g., $).
596 delete arrayname is now legal; it clears the elements but leaves
597 the array, which may not be the right behavior.
599 modified makefile: my current make can't cope with the test used
600 to avoid unnecessary yacc invocations.
602 Jul 17, 1995:
603 added dynamically growing strings to awk.lx.l and b.c
604 to permit regular expressions to be much bigger.
605 the state arrays can still overflow.
607 Aug 24, 1994:
608 detect duplicate arguments in function definitions (mdm).
610 May 11, 1994:
611 trivial fix to printf to limit string size in sub().
613 Apr 22, 1994:
614 fixed yet another subtle self-assignment problem:
615 $1 = $2; $1 = $1 clobbered $1.
617 Regression tests now use private echo, to avoid quoting problems.
619 Feb 2, 1994:
620 changed error() to print line number as %d, not %g.
622 Jul 23, 1993:
623 cosmetic changes: increased sizes of some arrays,
624 reworded some error messages.
626 added CONVFMT as in posix (just replaced OFMT in getsval)
628 FILENAME is now "" until the first thing that causes a file
629 to be opened.
631 Nov 28, 1992:
632 deleted yyunput and yyoutput from proto.h;
633 different versions of lex give these different declarations.
635 May 31, 1992:
636 added -mr N and -mf N options: more record and fields.
637 these really ought to adjust automatically.
639 cleaned up some error messages; "out of space" now means
640 malloc returned NULL in all cases.
642 changed rehash so that if it runs out, it just returns;
643 things will continue to run slow, but maybe a bit longer.
645 Apr 24, 1992:
646 remove redundant close of stdin when using -f -.
648 got rid of core dump with -d; awk -d just prints date.
650 Apr 12, 1992:
651 added explicit check for /dev/std(in,out,err) in redirection.
652 unlike gawk, no /dev/fd/n yet.
654 added (file/pipe) builtin. hard to test satisfactorily.
655 not posix.
657 Feb 20, 1992:
658 recompile after abortive changes; should be unchanged.
660 Dec 2, 1991:
661 die-casting time: converted to ansi C, installed that.
663 Nov 30, 1991:
664 fixed storage leak in freefa, failing to recover [N]CCL.
665 thanks to Bill Jones (jones@cs.usask.ca)
667 Nov 19, 1991:
668 use RAND_MAX instead of literal in builtin().
670 Nov 12, 1991:
671 cranked up some fixed-size arrays in b.c, and added a test for
672 overflow in penter. thanks to mark larsen.
674 Sep 24, 1991:
675 increased buffer in gsub. a very crude fix to a general problem.
676 and again on Sep 26.
678 Aug 18, 1991:
679 enforce variable name syntax for commandline variables: has to
680 start with letter or _.
682 Jul 27, 1991:
683 allow newline after ; in for statements.
685 Jul 21, 1991:
686 fixed so that in self-assignment like $1=$1, side effects
687 like recomputing $0 take place. (this is getting subtle.)
689 Jun 30, 1991:
690 better test for detecting too-long output record.
692 Jun 2, 1991:
693 better defense against very long printf strings.
694 made break and continue illegal outside of loops.
696 May 13, 1991:
697 removed extra arg on gettemp, tempfree. minor error message rewording.
699 May 6, 1991:
700 fixed silly bug in hex parsing in hexstr().
701 removed an apparently unnecessary test in isnumber().
702 warn about weird printf conversions.
703 fixed unchecked array overwrite in relex().
705 changed for (i in array) to access elements in sorted order.
706 then unchanged it -- it really does run slower in too many cases.
707 left the code in place, commented out.
709 Feb 10, 1991:
710 check error status on all writes, to avoid banging on full disks.
712 Jan 28, 1991:
713 awk -f - reads the program from stdin.
715 Jan 11, 1991:
716 failed to set numeric state on $0 in cmd|getline context in run.c.
718 Nov 2, 1990:
719 fixed sleazy test for integrality in getsval; use modf.
721 Oct 29, 1990:
722 fixed sleazy buggy code in lib.c that looked (incorrectly) for
723 too long input lines.
725 Oct 14, 1990:
726 fixed the bug on p. 198 in which it couldn't deduce that an
727 argument was an array in some contexts. replaced the error
728 message in intest() by code that damn well makes it an array.
730 Oct 8, 1990:
731 fixed horrible bug: types and values were not preserved in
732 some kinds of self-assignment. (in assign().)
734 Aug 24, 1990:
735 changed NCHARS to 256 to handle 8-bit characters in strings
736 presented to match(), etc.
738 Jun 26, 1990:
739 changed struct rrow (awk.h) to use long instead of int for lval,
740 since cfoll() stores a pointer in it. now works better when int's
741 are smaller than pointers!
743 May 6, 1990:
744 AVA fixed the grammar so that ! is uniformly of the same precedence as
745 unary + and -. This renders illegal some constructs like !x=y, which
746 now has to be parenthesized as !(x=y), and makes others work properly:
747 !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements.
748 (These problems were pointed out by Bob Lenk of Posix.)
750 Added \x to regular expressions (already in strings).
751 Limited octal to octal digits; \8 and \9 are not octal.
752 Centralized the code for parsing escapes in regular expressions.
753 Added a bunch of tests to T.re and T.sub to verify some of this.
755 Feb 9, 1990:
756 fixed null pointer dereference bug in main.c: -F[nothing]. sigh.
758 restored srand behavior: it returns the current seed.
760 Jan 18, 1990:
761 srand now returns previous seed value (0 to start).
763 Jan 5, 1990:
764 fix potential problem in tran.c -- something was freed,
765 then used in freesymtab.
767 Oct 18, 1989:
768 another try to get the max number of open files set with
769 relatively machine-independent code.
771 small fix to input() in case of multiple reads after EOF.
773 Oct 11, 1989:
774 FILENAME is now defined in the BEGIN block -- too many old
775 programs broke.
777 "-" means stdin in getline as well as on the commandline.
779 added a bunch of casts to the code to tell the truth about
780 char * vs. unsigned char *, a right royal pain. added a
781 setlocale call to the front of main, though probably no one
782 has it usefully implemented yet.
784 Aug 24, 1989:
785 removed redundant relational tests against nullnode if parse
786 tree already had a relational at that point.
788 Aug 11, 1989:
789 fixed bug: commandline variable assignment has to look like
790 var=something. (consider the man page for =, in file =.1)
792 changed number of arguments to functions to static arrays
793 to avoid repeated malloc calls.
795 Aug 2, 1989:
796 restored -F (space) separator
798 Jul 30, 1989:
799 added -v x=1 y=2 ... for immediate commandline variable assignment;
800 done before the BEGIN block for sure. they have to precede the
801 program if the program is on the commandline.
802 Modified Aug 2 to require a separate -v for each assignment.
804 Jul 10, 1989:
805 fixed ref-thru-zero bug in environment code in tran.c
807 Jun 23, 1989:
808 add newline to usage message.
810 Jun 14, 1989:
811 added some missing ansi printf conversion letters: %i %X %E %G.
812 no sensible meaning for h or L, so they may not do what one expects.
814 made %* conversions work.
816 changed x^y so that if n is a positive integer, it's done
817 by explicit multiplication, thus achieving maximum accuracy.
818 (this should be done by pow() but it seems not to be locally.)
819 done to x ^= y as well.
821 Jun 4, 1989:
822 ENVIRON array contains environment: if shell variable V=thing,
823 ENVIRON["V"] is "thing"
825 multiple -f arguments permitted. error reporting is naive.
826 (they were permitted before, but only the last was used.)
828 fixed a really stupid botch in the debugging macro dprintf
830 fixed order of evaluation of commandline assignments to match
831 what the book claims: an argument of the form x=e is evaluated
832 at the time it would have been opened if it were a filename (p 63).
833 this invalidates the suggested answer to ex 4-1 (p 195).
835 removed some code that permitted -F (space) fieldseparator,
836 since it didn't quite work right anyway. (restored aug 2)
838 Apr 27, 1989:
839 Line number now accumulated correctly for comment lines.
841 Apr 26, 1989:
842 Debugging output now includes a version date,
843 if one compiles it into the source each time.
845 Apr 9, 1989:
846 Changed grammar to prohibit constants as 3rd arg of sub and gsub;
847 prevents class of overwriting-a-constant errors. (Last one?)
848 This invalidates the "banana" example on page 43 of the book.
850 Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal),
851 as in ANSI, for strings. Rescinded the sloppiness that permitted
852 non-octal digits in \ooo. Warning: not all compilers and libraries
853 will be able to deal with \x correctly.
855 Jan 9, 1989:
856 Fixed bug that caused tempcell list to contain a duplicate.
857 The fix is kludgy.
859 Dec 17, 1988:
860 Catches some more commandline errors in main.
861 Removed redundant decl of modf in run.c (confuses some compilers).
862 Warning: there's no single declaration of malloc, etc., in awk.h
863 that seems to satisfy all compilers.
865 Dec 7, 1988:
866 Added a bit of code to error printing to avoid printing nulls.
867 (Not clear that it actually would.)
869 Nov 27, 1988:
870 With fear and trembling, modified the grammar to permit
871 multiple pattern-action statements on one line without
872 an explicit separator. By definition, this capitulation
873 to the ghost of ancient implementations remains undefined
874 and thus subject to change without notice or apology.
875 DO NOT COUNT ON IT.
877 Oct 30, 1988:
878 Fixed bug in call() that failed to recover storage.
880 A warning is now generated if there are more arguments
881 in the call than in the definition (in lieu of fixing
882 another storage leak).
884 Oct 20, 1988:
885 Fixed %c: if expr is numeric, use numeric value;
886 otherwise print 1st char of string value. still
887 doesn't work if the value is 0 -- won't print \0.
889 Added a few more checks for running out of malloc.
891 Oct 12, 1988:
892 Fixed bug in call() that freed local arrays twice.
894 Fixed to handle deletion of non-existent array right;
895 complains about attempt to delete non-array element.
897 Sep 30, 1988:
898 Now guarantees to evaluate all arguments of built-in
899 functions, as in C; the appearance is that arguments
900 are evaluated before the function is called. Places
901 affected are sub (gsub was ok), substr, printf, and
902 all the built-in arithmetic functions in bltin().
903 A warning is generated if a bltin() is called with
904 the wrong number of arguments.
906 This requires changing makeprof on p167 of the book.
908 Aug 23, 1988:
909 setting FILENAME in BEGIN caused core dump, apparently
910 because it was freeing space not allocated by malloc.
912 July 24, 1988:
913 fixed egregious error in toupper/tolower functions.
914 still subject to rescinding, however.
916 July 2, 1988:
917 flush stdout before opening file or pipe
919 July 2, 1988:
920 performance bug in b.c/cgoto(): not freeing some sets of states.
921 partial fix only right now, and the number of states increased
922 to make it less obvious.
924 June 1, 1988:
925 check error status on close
927 May 28, 1988:
928 srand returns seed value it's using.
929 see 1/18/90
931 May 22, 1988:
932 Removed limit on depth of function calls.
934 May 10, 1988:
935 Fixed lib.c to permit _ in commandline variable names.
937 Mar 25, 1988:
938 main.c fixed to recognize -- as terminator of command-
939 line options. Illegal options flagged.
940 Error reporting slightly cleaned up.
942 Dec 2, 1987:
943 Newer C compilers apply a strict scope rule to extern
944 declarations within functions. Two extern declarations in
945 lib.c and tran.c have been moved to obviate this problem.
947 Oct xx, 1987:
948 Reluctantly added toupper and tolower functions.
949 Subject to rescinding without notice.
951 Sep 17, 1987:
952 Error-message printer had printf(s) instead of
953 printf("%s",s); got core dumps when the message
954 included a %.
956 Sep 12, 1987:
957 Very long printf strings caused core dump;
958 fixed aprintf, asprintf, format to catch them.
959 Can still get a core dump in printf itself.