Annotation of src/usr.bin/awk/FIXES, Revision 1.24
1.24 ! millert 1: /* $OpenBSD: FIXES,v 1.23 2020/06/10 21:01:50 millert Exp $ */
1.1 tholo 2: /****************************************************************
1.6 kstailey 3: Copyright (C) Lucent Technologies 1997
1.1 tholo 4: All Rights Reserved
5:
6: Permission to use, copy, modify, and distribute this software and
7: its documentation for any purpose and without fee is hereby
8: granted, provided that the above copyright notice appear in all
9: copies and that both that the copyright notice and this
10: permission notice and warranty disclaimer appear in supporting
1.6 kstailey 11: documentation, and that the name Lucent Technologies or any of
12: its entities not be used in advertising or publicity pertaining
13: to distribution of the software without specific, written prior
14: permission.
15:
16: LUCENT DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
17: INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS.
18: IN NO EVENT SHALL LUCENT OR ANY OF ITS ENTITIES BE LIABLE FOR ANY
19: SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
20: WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER
21: IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
22: ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
23: THIS SOFTWARE.
1.1 tholo 24: ****************************************************************/
25:
26: This file lists all bug fixes, changes, etc., made since the AWK book
27: was sent to the printers in August, 1987.
1.23 millert 28:
1.24 ! millert 29: June 17, 2019:
! 30: Disallow deleting SYMTAB and its elements, which creates
! 31: use-after-free bugs. Thanks to GitHub user Cody Mello (melloc)
! 32: for the fix. (Merged from PR #43.)
! 33:
! 34: June 5, 2019:
! 35: Allow unmatched right parenthesis in a regular expression to
! 36: be treated literally. Fixes Issue #40. Thanks to GitHub user
! 37: Warner Losh (bsdimp) for the report. Thanks to Arnold Robbins
! 38: for the fix.
! 39:
! 40: May 29,2019:
! 41: Fix check for command line arguments to no longer require that
! 42: first character after '=' not be another '='. Reverts change of
! 43: August 11, 1989. Thanks to GitHub user Jamie Landeg Jones for
! 44: pointing out the issue; from Issue #38.
! 45:
! 46: Apr 7, 2019:
! 47: Update awktest.tar(p.50) to use modern options to sort. Needed
! 48: for Android development. Thanks to GitHub user mohd-akram (Mohamed
! 49: Akram). From Issue #33.
! 50:
! 51: Mar 12, 2019:
! 52: Added very simplistic support for cross-compiling in the
! 53: makefile. We are NOT going to go in the direction of the
! 54: autotools, though. Thanks to GitHub user nee-san for
! 55: the basic change. (Merged from PR #34.)
! 56:
1.23 millert 57: Mar 5, 2019:
58: Added support for POSIX-standard interval expressions (a.k.a.
59: bounds, a.k.a. repetition expressions) in regular expressions,
60: backported (via NetBSD) from Apple awk-24 (20070501).
61: Thanks to Martijn Dekker <martijn@inlv.org> for the port.
62: (Merged from PR #30.)
63:
64: Mar 3, 2019:
65: Merge PRs as follows:
66: #12: Avoid undefined behaviour when using ctype(3) functions in
67: relex(). Thanks to GitHub user iamleot.
68: #31: Make getline handle numeric strings, and update FIXES. Thanks
1.24 ! millert 69: to GitHub user arnoldrobbins.
1.23 millert 70: #32: maketab: support build systems with read-only source. Thanks
71: to GitHub user enh.
1.22 millert 72:
73: Jan 25, 2019:
74: Make getline handle numeric strings properly in all cases.
75: (Thanks, Arnold.)
76:
77: Jan 21, 2019:
78: Merged a number of small fixes from GitHub pull requests.
79: Thanks to GitHub users Arnold Robbins (arnoldrobbins),
80: Cody Mello (melloc) and Christoph Junghans (junghans).
81: PR numbers: 13-21, 23, 24, 27.
1.21 millert 82:
83: Oct 25, 2018:
84: Added test in maketab.c to prevent generating a proctab entry
85: for YYSTYPE_IS_DEFINED. It was harmless but some gcc settings
86: generated a warning message. Thanks to Nan Xiao for report.
1.20 millert 87:
88: Aug 27, 2018:
89: Disallow '$' in printf formats; arguments evaluated in order
90: and printed in order.
91:
92: Added some casts to silence warnings on debugging printfs.
93: (Thanks, Arnold.)
1.18 millert 94:
1.19 millert 95: Aug 23, 2018:
96: A long list of fixes courtesy of Arnold Robbins,
97: to whom profound thanks.
98:
99: 1. ofs-rebuild: OFS value used to rebuild the record was incorrect.
100: Fixed August 19, 2014. Revised fix August 2018.
101:
102: 2. system-status: Instead of a floating-point division by 256, use
103: the wait(2) macros to create a reasonable exit status.
104: Fixed March 12, 2016.
105:
106: 3. space: Use provided xisblank() function instead of ispace() for
107: matching [[:blank:]].
108:
109: 4. a-format: Add POSIX standard %a and %A to supported formats. Check
110: at runtime that this format is available.
111:
112: 5. decr-NF: Decrementing NF did not change $0. This is a decades-old
113: bug. There are interactions with the old and new value of OFS as well.
114: Most of the fix came from the NetBSD awk.
115:
116: 6. string-conv: String conversions of scalars were sticky. Once a
117: conversion to string happened, even with OFMT, that value was used until
118: a new numeric value was assigned, even if OFMT differed from CONVFMT,
119: and also if CONVFMT changed.
120:
121: 7. unary-plus: Unary plus on a string constant returned the string.
122: Instead, it should convert the value to numeric and give that value.
123:
124: Also added Arnold's tests for these to awktest.tar as T.arnold.
125:
126: Aug 15, 2018:
127: fixed mangled awktest.tar (thanks, Arnold), posted all
128: current (very minor) fixes to github / onetrueawk
129:
1.18 millert 130: Jun 7, 2018:
131: (yes, a long layoff)
132: Updated some broken tests (beebe.tar, T.lilly)
133: [thanks to Arnold Robbins]
134:
135: Mar 26, 2015:
136: buffer overflow in error reporting; thanks to tobias ulmer
137: and john-mark gurney for spotting it and the fix.
138:
139: Feb 4, 2013:
140: cleaned up a handful of tests that didn't seem to actually
141: test for correct behavior: T.latin1, T.gawk.
142:
143: Jan 5, 2013:
144: added ,NULL initializer to static Cells in run.c; not really
145: needed but cleaner. Thanks to Michael Bombardieri.
1.17 millert 146:
147: Dec 20, 2012:
148: fiddled makefile to get correct yacc and bison flags. pick yacc
149: (linux) or bison (mac) as necessary.
150:
151: added __attribute__((__noreturn__)) to a couple of lines in
152: proto.h, to silence someone's enthusiastic checker.
153:
154: fixed obscure call by value bug in split(a[1],a) reported on
155: 9fans. the management of temporary values is just a mess; i
156: took a shortcut by making an extra string copy. thanks
157: to paul patience and arnold robbins for passing it on and for
158: proposed patches.
159:
160: tiny fiddle in setfval to eliminate -0 results in T.expr, which
161: has irritated me for 20+ years.
1.16 millert 162:
163: Aug 10, 2011:
164: another fix to avoid core dump with delete(ARGV); again, many thanks
165: to ruslan ermilov.
166:
167: Aug 7, 2011:
168: split(s, a, //) now behaves the same as split(s, a, "")
169:
170: Jun 12, 2011:
171: /pat/, \n /pat/ {...} is now legal, though bad style to use.
172:
173: added checks to new -v code that permits -vnospace; thanks to
174: ruslan ermilov for spotting this and providing the patch.
175:
176: removed fixed limit on number of open files; thanks to aleksey
177: cheusov and christos zoulos.
178:
179: fixed day 1 bug that resurrected deleted elements of ARGV when
180: used as filenames (in lib.c).
181:
182: minor type fiddles to make gcc -Wall -pedantic happier (but not
183: totally so); turned on -fno-strict-aliasing in makefile.
184:
185: May 6, 2011:
186: added #ifdef for isblank.
187: now allows -ffoo as well as -f foo arguments.
188: (thanks, ruslan)
189:
190: May 1, 2011:
191: after advice from todd miller, kevin lo, ruslan ermilov,
192: and arnold robbins, changed srand() to return the previous
193: seed (which is 1 on the first call of srand). the seed is
194: an Awkfloat internally though converted to unsigned int to
195: pass to the library srand(). thanks, everyone.
196:
197: fixed a subtle (and i hope low-probability) overflow error
198: in fldbld, by adding space for one extra \0. thanks to
199: robert bassett for spotting this one and providing a fix.
200:
201: removed the files related to compilation on windows. i no
202: longer have anything like a current windows environment, so
203: i can't test any of it.
1.15 millert 204:
205: May 23, 2010:
206: fixed long-standing overflow bug in run.c; many thanks to
207: nelson beebe for spotting it and providing the fix.
208:
209: fixed bug that didn't parse -vd=1 properly; thanks to santiago
210: vila for spotting it.
211:
212: Feb 8, 2010:
213: i give up. replaced isblank with isspace in b.c; there are
214: no consistent header files.
215:
216: Nov 26, 2009:
217: fixed a long-standing issue with when FS takes effect. a
218: change to FS is now noticed immediately for subsequent splits.
219:
220: changed the name getline() to awkgetline() to avoid yet another
221: name conflict somewhere.
222:
223: Feb 11, 2009:
224: temporarily for now defined HAS_ISBLANK, since that seems to
225: be the best way through the thicket. isblank arrived in C99,
226: but seems to be arriving at different systems at different
227: times.
228:
229: Oct 8, 2008:
230: fixed typo in b.c that set tmpvec wrongly. no one had ever
231: run into the problem, apparently. thanks to alistair crooks.
1.14 millert 232:
233: Oct 23, 2007:
234: minor fix in lib.c: increase inputFS to 100, change malloc
235: for fields to n+1.
236:
237: fixed memory fault caused by out of order test in setsval.
238:
239: thanks to david o'brien, freebsd, for both fixes.
240:
241: May 1, 2007:
242: fiddle in makefile to fix for BSD make; thanks to igor sobrado.
243:
244: Mar 31, 2007:
245: fixed some null pointer refs calling adjbuf.
246:
247: Feb 21, 2007:
248: fixed a bug in matching the null RE in sub and gsub. thanks to al aho
249: who actually did the fix (in b.c), and to wolfgang seeberg for finding
250: it and providing a very compact test case.
251:
252: fixed quotation in b.c; thanks to Hal Pratt and the Princeton Dante
253: Project.
254:
255: removed some no-effect asserts in run.c.
256:
257: fiddled maketab.c to not complain about bison-generated values.
258:
259: removed the obsolete -V argument; fixed --version to print the
260: version and exit.
261:
262: fixed wording and an outright error in the usage message; thanks to igor
263: sobrado and jason mcintyre.
264:
265: fixed a bug in -d that caused core dump if no program followed.
266:
267: Jan 1, 2007:
268: dropped mac.code from makefile; there are few non-MacOSX
269: mac's these days.
270:
271: Jan 17, 2006:
272: system() not flagged as unsafe in the unadvertised -safe option.
273: found it while enhancing tests before shipping the ;login: article.
274: practice what you preach.
275:
276: removed the 9-years-obsolete -mr and -mf flags.
277:
278: added -version and --version options.
279:
280: core dump on linux with BEGIN {nextfile}, now fixed.
281:
282: removed some #ifdef's in run.c and lex.c that appear to no
283: longer be necessary.
284:
285: Apr 24, 2005:
286: modified lib.c so that values of $0 et al are preserved in the END
287: block, apparently as required by posix. thanks to havard eidnes
288: for the report and code.
289:
290: Jan 14, 2005:
291: fixed infinite loop in parsing, originally found by brian tsang.
292: thanks to arnold robbins for a suggestion that started me
293: rethinking it.
294:
295: Dec 31, 2004:
296: prevent overflow of -f array in main, head off potential error in
297: call of SYNTAX(), test malloc return in lib.c, all with thanks to
298: todd miller.
1.13 millert 299:
300: Dec 22, 2004:
301: cranked up size of NCHARS; coverity thinks it can be overrun with
302: smaller size, and i think that's right. added some assertions to b.c
303: to catch places where it might overrun. the RE code is still fragile.
304:
305: Dec 5, 2004:
306: fixed a couple of overflow problems with ridiculous field numbers:
307: e.g., print $(2^32-1). thanks to ruslan ermilov, giorgos keramidas
308: and david o'brien at freebsd.org for patches. this really should
309: be re-done from scratch.
310:
311: Nov 21, 2004:
312: fixed another 25-year-old RE bug, in split. it's another failure
313: to (re-)initialize. thanks to steve fisher for spotting this and
314: providing a good test case.
315:
316: Nov 22, 2003:
317: fixed a bug in regular expressions that dates (so help me) from 1977;
318: it's been there from the beginning. an anchored longest match that
319: was longer than the number of states triggered a failure to initialize
320: the machine properly. many thanks to moinak ghosh for not only finding
321: this one but for providing a fix, in some of the most mysterious
322: code known to man.
323:
324: fixed a storage leak in call() that appears to have been there since
325: 1983 or so -- a function without an explicit return that assigns a
326: string to a parameter leaked a Cell. thanks to moinak ghosh for
327: spotting this very subtle one.
328:
329: Jul 31, 2003:
330: fixed, thanks to andrey chernov and ruslan ermilov, a bug in lex.c
331: that mis-handled the character 255 in input. (it was being compared
332: to EOF with a signed comparison.)
333:
334: Jul 29, 2003:
335: fixed (i think) the long-standing botch that included the beginning of
336: line state ^ for RE's in the set of valid characters; this led to a
337: variety of odd problems, including failure to properly match certain
338: regular expressions in non-US locales. thanks to ruslan for keeping
339: at this one.
340:
341: Jul 28, 2003:
342: n-th try at getting internationalization right, with thanks to volker
343: kiefel, arnold robbins and ruslan ermilov for advice, though they
344: should not be blamed for the outcome. according to posix, "." is the
345: radix character in programs and command line arguments regardless of
346: the locale; otherwise, the locale should prevail for input and output
347: of numbers. so it's intended to work that way.
348:
349: i have rescinded the attempt to use strcoll in expanding shorthands in
350: regular expressions (cclenter). its properties are much too
351: surprising; for example [a-c] matches aAbBc in locale en_US but abBcC
352: in locale fr_CA. i can see how this might arise by implementation
353: but i cannot explain it to a human user. (this behavior can be seen
354: in gawk as well; we're leaning on the same library.)
355:
356: the issue appears to be that strcoll is meant for sorting, where
357: merging upper and lower case may make sense (though note that unix
358: sort does not do this by default either). it is not appropriate
359: for regular expressions, where the goal is to match specific
360: patterns of characters. in any case, the notations [:lower:], etc.,
361: are available in awk, and they are more likely to work correctly in
362: most locales.
363:
364: a moratorium is hereby declared on internationalization changes.
365: i apologize to friends and colleagues in other parts of the world.
366: i would truly like to get this "right", but i don't know what
367: that is, and i do not want to keep making changes until it's clear.
368:
369: Jul 4, 2003:
370: fixed bug that permitted non-terminated RE, as in "awk /x".
371:
372: Jun 1, 2003:
373: subtle change to split: if source is empty, number of elems
374: is always 0 and the array is not set.
375:
376: Mar 21, 2003:
377: added some parens to isblank, in another attempt to make things
378: internationally portable.
379:
380: Mar 14, 2003:
381: the internationalization changes, somewhat modified, are now
382: reinstated. in theory awk will now do character comparisons
383: and case conversions in national language, but "." will always
384: be the decimal point separator on input and output regardless
385: of national language. isblank(){} has an #ifndef.
386:
387: this no longer compiles on windows: LC_MESSAGES isn't defined
388: in vc6++.
389:
390: fixed subtle behavior in field and record splitting: if FS is
391: a single character and RS is not empty, \n is NOT a separator.
392: this tortuous reading is found in the awk book; behavior now
393: matches gawk and mawk.
1.11 millert 394:
395: Dec 13, 2002:
396: for the moment, the internationalization changes of nov 29 are
397: rolled back -- programs like x = 1.2 don't work in some locales,
398: because the parser is expecting x = 1,2. until i understand this
399: better, this will have to wait.
400:
401: Nov 29, 2002:
402: modified b.c (with tiny changes in main and run) to support
403: locales, using strcoll and iswhatever tests for posix character
404: classes. thanks to ruslan ermilov (ru@freebsd.org) for code.
405: the function isblank doesn't seem to have propagated to any
406: header file near me, so it's there explicitly. not properly
407: tested on non-ascii character sets by me.
408:
409: Jun 28, 2002:
410: modified run/format() and tran/getsval() to do a slightly better
411: job on using OFMT for output from print and CONVFMT for other
412: number->string conversions, as promised by posix and done by
413: gawk and mawk. there are still places where it doesn't work
414: right if CONVFMT is changed; by then the STR attribute of the
415: variable has been irrevocably set. thanks to arnold robbins for
416: code and examples.
417:
418: fixed subtle bug in format that could get core dump. thanks to
419: Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing.
420: minor cleanup in run.c / format() at the same time.
421:
422: added some tests for null pointers to debugging printf's, which
423: were never intended for external consumption. thanks to dave
424: kerns (dkerns@lucent.com) for pointing this out.
425:
426: GNU compatibility: an empty regexp matches anything (thanks to
427: dag-erling smorgrav, des@ofug.org). subject to reversion if
428: this does more harm than good.
429:
430: pervasive small changes to make things more const-correct, as
431: reported by gcc's -Wwrite-strings. as it says in the gcc manual,
432: this may be more nuisance than useful. provoked by a suggestion
433: and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk
434:
435: minor documentation changes to note that this now compiles out
436: of the box on Mac OS X.
437:
438: Feb 10, 2002:
439: changed types in posix chars structure to quiet solaris cc.
440:
441: Jan 1, 2002:
442: fflush() or fflush("") flushes all files and pipes.
443:
444: length(arrayname) returns number of elements; thanks to
445: arnold robbins for suggestion.
446:
447: added a makefile.win to make it easier to build on windows.
448: based on dan allen's buildwin.bat.
449:
450: Nov 16, 2001:
451: added support for posix character class names like [:digit:],
452: which are not exactly shorter than [0-9] and perhaps no more
453: portable. thanks to dag-erling smorgrav for code.
454:
455: Feb 16, 2001:
456: removed -m option; no longer needed, and it was actually
457: broken (noted thanks to volker kiefel).
458:
459: Feb 10, 2001:
460: fixed an appalling bug in gettok: any sequence of digits, +,-, E, e,
461: and period was accepted as a valid number if it started with a period.
462: this would never have happened with the lex version.
463:
464: other 1-character botches, now fixed, include a bare $ and a
465: bare " at the end of the input.
466:
467: Feb 7, 2001:
468: more (const char *) casts in b.c and tran.c to silence warnings.
1.8 millert 469:
1.10 millert 470: Nov 15, 2000:
471: fixed a bug introduced in august 1997 that caused expressions
472: like $f[1] to be syntax errors. thanks to arnold robbins for
473: noticing this and providing a fix.
474:
475: Oct 30, 2000:
476: fixed some nextfile bugs: not handling all cases. thanks to
477: arnold robbins for pointing this out. new regressions added.
478:
479: close() is now a function. it returns whatever the library
480: fclose returns, and -1 for closing a file or pipe that wasn't
481: opened.
482:
483: Sep 24, 2000:
484: permit \n explicitly in character classes; won't work right
485: if comes in as "[\n]" but ok as /[\n]/, because of multiple
486: processing of \'s. thanks to arnold robbins.
487:
488: July 5, 2000:
489: minor fiddles in tran.c to keep compilers happy about uschar.
490: thanks to norman wilson.
491:
492: May 25, 2000:
493: yet another attempt at making 8-bit input work, with another
494: band-aid in b.c (member()), and some (uschar) casts to head
495: off potential errors in subscripts (like isdigit). also
496: changed HAT to NCHARS-2. thanks again to santiago vila.
497:
498: changed maketab.c to ignore apparently out of range definitions
499: instead of halting; new freeBSD generates one. thanks to
500: jon snader <jsnader@ix.netcom.com> for pointing out the problem.
501:
502: May 2, 2000:
503: fixed an 8-bit problem in b.c by making several char*'s into
504: unsigned char*'s. not clear i have them all yet. thanks to
505: Santiago Vila <sanvila@unex.es> for the bug report.
506:
507: Apr 21, 2000:
508: finally found and fixed a memory leak in function call; it's
509: been there since functions were added ~1983. thanks to
510: jon bentley for the test case that found it.
511:
512: added test in envinit to catch environment "variables" with
1.12 otto 513: names beginning with '='; thanks to Berend Hasselman.
1.10 millert 514:
515: Jul 28, 1999:
516: added test in defn() to catch function foo(foo), which
517: otherwise recurses until core dump. thanks to arnold
518: robbins for noticing this.
519:
1.9 millert 520: Jun 20, 1999:
521: added *bp in gettok in lex.c; appears possible to exit function
522: without terminating the string. thanks to russ cox.
523:
524: Jun 2, 1999:
525: added function stdinit() to run to initialize files[] array,
526: in case stdin, etc., are not constants; some compilers care.
527:
528: May 10, 1999:
529: replaced the ERROR ... FATAL, etc., macros with functions
530: based on vprintf, to avoid problems caused by overrunning
531: fixed-size errbuf array. thanks to ralph corderoy for the
532: impetus, and for pointing out a string termination bug in
533: qstring as well.
534:
535: Apr 21, 1999:
536: fixed bug that caused occasional core dumps with commandline
537: variable with value ending in \. (thanks to nelson beebe for
538: the test case.)
539:
1.8 millert 540: Apr 16, 1999:
541: with code kindly provided by Bruce Lilly, awk now parses
542: /=/ and similar constructs more sensibly in more places.
1.10 millert 543: Bruce also provided some helpful test cases.
1.8 millert 544:
545: Apr 5, 1999:
546: changed true/false to True/False in run.c to make it
547: easier to compile with C++. Added some casts on malloc
548: and realloc to be honest about casts; ditto. changed
549: ltype int to long in struct rrow to reduce some 64-bit
550: complaints; other changes scattered throughout for the
551: same purpose. thanks to Nelson Beebe for these portability
552: improvements.
553:
554: removed some horrible pointer-int casting in b.c and elsewhere
555: by adding ptoi and itonp to localize the casts, which are
556: all benign. fixed one incipient bug that showed up on sgi
557: in 64-bit mode.
558:
559: reset lineno for new source file; include filename in error
560: message. also fixed line number error in continuation lines.
561: (thanks to Nelson Beebe for both of these.)
562:
563: Mar 24, 1999:
564: Nelson Beebe notes that irix 5.3 yacc dies with a bogus
565: error; use a newer version or switch to bison, since sgi
566: is unlikely to fix it.
1.1 tholo 567:
1.7 millert 568: Mar 5, 1999:
1.9 millert 569: changed isnumber to is_number to avoid the problem caused by
570: versions of ctype.h that include the name isnumber.
1.7 millert 571:
572: distribution now includes a script for building on a Mac,
573: thanks to Dan Allen.
574:
575: Feb 20, 1999:
576: fixed memory leaks in run.c (call) and tran.c (setfval).
577: thanks to Stephen Nutt for finding these and providing the fixes.
578:
579: Jan 13, 1999:
580: replaced srand argument by (unsigned int) in run.c;
581: avoids problem on Mac and potentially on Unix & Windows.
582: thanks to Dan Allen.
583:
584: added a few (int) casts to silence useless compiler warnings.
585: e.g., errorflag= in run.c jump().
586:
587: added proctab.c to the bundle outout; one less thing
588: to have to compile out of the box.
589:
590: added calls to _popen and _pclose to the win95 stub for
591: pipes (thanks to Steve Adams for this helpful suggestion).
592: seems to work, though properties are not well understood
593: by me, and it appears that under some circumstances the
594: pipe output is truncated. Be careful.
595:
596: Oct 19, 1998:
597: fixed a couple of bugs in getrec: could fail to update $0
598: after a getline var; because inputFS wasn't initialized,
599: could split $0 on every character, a misleading diversion.
600:
601: fixed caching bug in makedfa: LRU was actually removing
602: least often used.
603:
604: thanks to ross ridge for finding these, and for providing
605: great bug reports.
606:
607: May 12, 1998:
608: fixed potential bug in readrec: might fail to update record
609: pointer after growing. thanks to dan levy for spotting this
610: and suggesting the fix.
611:
612: Mar 12, 1998:
613: added -V to print version number and die.
1.19 millert 614:
615: [notify dave kerns, dkerns@dacsoup.ih.lucent.com]
1.7 millert 616:
617: Feb 11, 1998:
618: subtle silent bug in lex.c: if the program ended with a number
619: longer than 1 digit, part of the input would be pushed back and
620: parsed again because token buffer wasn't terminated right.
621: example: awk 'length($0) > 10'. blush. at least i found it
622: myself.
623:
624: Aug 31, 1997:
625: s/adelete/awkdelete/: SGI uses this in malloc.h.
626: thanks to nelson beebe for pointing this one out.
627:
628: Aug 21, 1997:
629: fixed some bugs in sub and gsub when replacement includes \\.
630: this is a dark, horrible corner, but at least now i believe that
631: the behavior is the same as gawk and the intended posix standard.
632: thanks to arnold robbins for advice here.
633:
634: Aug 9, 1997:
635: somewhat regretfully, replaced the ancient lex-based lexical
636: analyzer with one written in C. it's longer, generates less code,
637: and more portable; the old one depended too much on mysterious
638: properties of lex that were not preserved in other environments.
639: in theory these recognize the same language.
640:
641: now using strtod to test whether a string is a number, instead of
642: the convoluted original function. should be more portable and
643: reliable if strtod is implemented right.
644:
645: removed now-pointless optimization in makefile that tries to avoid
646: recompilation when awkgram.y is changed but symbols are not.
647:
648: removed most fixed-size arrays, though a handful remain, some
649: of which are unchecked. you have been warned.
650:
651: Aug 4, 1997:
652: with some trepidation, replaced the ancient code that managed
653: fields and $0 in fixed-size arrays with arrays that grow on
654: demand. there is still some tension between trying to make this
655: run fast and making it clean; not sure it's right yet.
656:
657: the ill-conceived -mr and -mf arguments are now useful only
658: for debugging. previous dynamic string code removed.
659:
660: numerous other minor cleanups along the way.
1.1 tholo 661:
1.7 millert 662: Jul 30, 1997:
663: using code provided by dan levy (to whom profuse thanks), replaced
664: fixed-size arrays and awkward kludges by a fairly uniform mechanism
665: to grow arrays as needed for printf, sub, gsub, etc.
1.1 tholo 666:
1.7 millert 667: Jul 23, 1997:
668: falling off the end of a function returns "" and 0, not 0.
669: thanks to arnold robbins.
1.1 tholo 670:
1.7 millert 671: Jun 17, 1997:
672: replaced several fixed-size arrays by dynamically-created ones
673: in run.c; added overflow tests to some previously unchecked cases.
674: getline, toupper, tolower.
1.1 tholo 675:
1.7 millert 676: getline code is still broken in that recursive calls may wind
677: up using the same space. [fixed later]
1.1 tholo 678:
1.7 millert 679: increased RECSIZE to 8192 to push problems further over the horizon.
1.1 tholo 680:
1.7 millert 681: added \r to \n as input line separator for programs, not data.
682: damn CRLFs.
1.1 tholo 683:
1.7 millert 684: modified format() to permit explicit printf("%c", 0) to include
685: a null byte in output. thanks to ken stailey for the fix.
1.1 tholo 686:
1.7 millert 687: added a "-safe" argument that disables file output (print >,
688: print >>), process creation (cmd|getline, print |, system), and
689: access to the environment (ENVIRON). this is a first approximation
690: to a "safe" version of awk, but don't rely on it too much. thanks
691: to joan feigenbaum and matt blaze for the inspiration long ago.
1.1 tholo 692:
1.7 millert 693: Jul 8, 1996:
694: fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to
695: ralph corderoy.
1.1 tholo 696:
1.7 millert 697: Jun 29, 1996:
698: fixed awful bug in new field splitting; didn't get all the places
699: where input was done.
1.1 tholo 700:
1.7 millert 701: Jun 28, 1996:
702: changed field-splitting to conform to posix definition: fields are
703: split using the value of FS at the time of input; it used to be
704: the value when the field or NF was first referred to, a much less
705: predictable definition. thanks to arnold robbins for encouragement
706: to do the right thing.
1.1 tholo 707:
1.7 millert 708: May 28, 1996:
709: fixed appalling but apparently unimportant bug in parsing octal
710: numbers in reg exprs.
1.1 tholo 711:
1.7 millert 712: explicit hex in reg exprs now limited to 2 chars: \xa, \xaa.
1.1 tholo 713:
1.7 millert 714: May 27, 1996:
715: cleaned up some declarations so gcc -Wall is now almost silent.
1.1 tholo 716:
1.7 millert 717: makefile now includes backup copies of ytab.c and lexyy.c in case
718: one makes before looking; it also avoids recreating lexyy.c unless
719: really needed.
1.1 tholo 720:
1.7 millert 721: s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes
722: with unwisely-written header files.
1.1 tholo 723:
1.7 millert 724: thanks to jeffrey friedl for several of these.
1.1 tholo 725:
1.7 millert 726: May 26, 1996:
727: an attempt to rationalize the (unsigned) char issue. almost all
728: instances of unsigned char have been removed; the handful of places
729: in b.c where chars are used as table indices have been hand-crafted.
730: added some latin-1 tests to the regression, but i'm not confident;
731: none of my compilers seem to care much. thanks to nelson beebe for
732: pointing out some others that do care.
1.1 tholo 733:
1.7 millert 734: May 2, 1996:
735: removed all register declarations.
1.1 tholo 736:
1.7 millert 737: enhanced split(), as in gawk, etc: split(s, a, "") splits s into
738: a[1]...a[length(s)] with each character a single element.
1.1 tholo 739:
1.7 millert 740: made the same changes for field-splitting if FS is "".
1.1 tholo 741:
1.7 millert 742: added nextfile, as in gawk: causes immediate advance to next
743: input file. (thanks to arnold robbins for inspiration and code).
1.1 tholo 744:
1.7 millert 745: small fixes to regexpr code: can now handle []], [[], and
746: variants; [] is now a syntax error, rather than matching
747: everything; [z-a] is now empty, not z. far from complete
748: or correct, however. (thanks to jeffrey friedl for pointing out
749: some awful behaviors.)
1.1 tholo 750:
1.7 millert 751: Apr 29, 1996:
1.12 otto 752: replaced uchar by uschar everywhere; apparently some compilers
1.7 millert 753: usurp this name and this causes conflicts.
1.1 tholo 754:
1.7 millert 755: fixed call to time in run.c (bltin); arg is time_t *.
1.1 tholo 756:
1.7 millert 757: replaced horrible pointer/long punning in b.c by a legitimate
758: union. should be safer on 64-bit machines and cleaner everywhere.
759: (thanks to nelson beebe for pointing out some of these problems.)
1.1 tholo 760:
1.7 millert 761: replaced nested comments by #if 0...#endif in run.c, lib.c.
1.1 tholo 762:
1.7 millert 763: removed getsval, setsval, execute macros from run.c and lib.c.
764: machines are 100x faster than they were when these macros were
765: first used.
1.1 tholo 766:
1.7 millert 767: revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l,
768: y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of
769: portability to nameless systems.
1.1 tholo 770:
1.7 millert 771: "make bundle" now includes yacc and lex output files for recipients
772: who don't have yacc or lex.
1.1 tholo 773:
1.7 millert 774: Aug 15, 1995:
775: initialized Cells in setsymtab more carefully; some fields
776: were not set. (thanks to purify, all of whose complaints i
777: think i now understand.)
1.1 tholo 778:
1.7 millert 779: fixed at least one error in gsub that looked at -1-th element
780: of an array when substituting for a null match (e.g., $).
1.1 tholo 781:
1.7 millert 782: delete arrayname is now legal; it clears the elements but leaves
783: the array, which may not be the right behavior.
1.1 tholo 784:
1.7 millert 785: modified makefile: my current make can't cope with the test used
786: to avoid unnecessary yacc invocations.
1.1 tholo 787:
1.7 millert 788: Jul 17, 1995:
789: added dynamically growing strings to awk.lx.l and b.c
790: to permit regular expressions to be much bigger.
791: the state arrays can still overflow.
1.1 tholo 792:
1.7 millert 793: Aug 24, 1994:
794: detect duplicate arguments in function definitions (mdm).
1.1 tholo 795:
1.7 millert 796: May 11, 1994:
797: trivial fix to printf to limit string size in sub().
1.1 tholo 798:
1.7 millert 799: Apr 22, 1994:
800: fixed yet another subtle self-assignment problem:
801: $1 = $2; $1 = $1 clobbered $1.
1.1 tholo 802:
1.7 millert 803: Regression tests now use private echo, to avoid quoting problems.
1.1 tholo 804:
1.7 millert 805: Feb 2, 1994:
806: changed error() to print line number as %d, not %g.
1.1 tholo 807:
1.7 millert 808: Jul 23, 1993:
809: cosmetic changes: increased sizes of some arrays,
810: reworded some error messages.
1.1 tholo 811:
1.7 millert 812: added CONVFMT as in posix (just replaced OFMT in getsval)
1.1 tholo 813:
1.7 millert 814: FILENAME is now "" until the first thing that causes a file
815: to be opened.
1.1 tholo 816:
1.7 millert 817: Nov 28, 1992:
818: deleted yyunput and yyoutput from proto.h;
819: different versions of lex give these different declarations.
1.1 tholo 820:
1.7 millert 821: May 31, 1992:
822: added -mr N and -mf N options: more record and fields.
823: these really ought to adjust automatically.
1.1 tholo 824:
1.7 millert 825: cleaned up some error messages; "out of space" now means
826: malloc returned NULL in all cases.
1.1 tholo 827:
1.7 millert 828: changed rehash so that if it runs out, it just returns;
829: things will continue to run slow, but maybe a bit longer.
1.1 tholo 830:
1.7 millert 831: Apr 24, 1992:
832: remove redundant close of stdin when using -f -.
1.1 tholo 833:
1.7 millert 834: got rid of core dump with -d; awk -d just prints date.
1.1 tholo 835:
1.7 millert 836: Apr 12, 1992:
837: added explicit check for /dev/std(in,out,err) in redirection.
838: unlike gawk, no /dev/fd/n yet.
1.1 tholo 839:
1.7 millert 840: added (file/pipe) builtin. hard to test satisfactorily.
841: not posix.
1.1 tholo 842:
1.7 millert 843: Feb 20, 1992:
844: recompile after abortive changes; should be unchanged.
1.1 tholo 845:
1.7 millert 846: Dec 2, 1991:
847: die-casting time: converted to ansi C, installed that.
1.1 tholo 848:
1.7 millert 849: Nov 30, 1991:
850: fixed storage leak in freefa, failing to recover [N]CCL.
851: thanks to Bill Jones (jones@cs.usask.ca)
1.1 tholo 852:
1.7 millert 853: Nov 19, 1991:
854: use RAND_MAX instead of literal in builtin().
1.1 tholo 855:
1.7 millert 856: Nov 12, 1991:
857: cranked up some fixed-size arrays in b.c, and added a test for
858: overflow in penter. thanks to mark larsen.
1.1 tholo 859:
1.7 millert 860: Sep 24, 1991:
861: increased buffer in gsub. a very crude fix to a general problem.
862: and again on Sep 26.
1.1 tholo 863:
1.7 millert 864: Aug 18, 1991:
865: enforce variable name syntax for commandline variables: has to
866: start with letter or _.
1.1 tholo 867:
1.7 millert 868: Jul 27, 1991:
869: allow newline after ; in for statements.
1.1 tholo 870:
1.7 millert 871: Jul 21, 1991:
872: fixed so that in self-assignment like $1=$1, side effects
873: like recomputing $0 take place. (this is getting subtle.)
1.1 tholo 874:
1.7 millert 875: Jun 30, 1991:
876: better test for detecting too-long output record.
1.1 tholo 877:
1.7 millert 878: Jun 2, 1991:
879: better defense against very long printf strings.
880: made break and continue illegal outside of loops.
1.1 tholo 881:
1.7 millert 882: May 13, 1991:
883: removed extra arg on gettemp, tempfree. minor error message rewording.
1.1 tholo 884:
885: May 6, 1991:
886: fixed silly bug in hex parsing in hexstr().
887: removed an apparently unnecessary test in isnumber().
888: warn about weird printf conversions.
889: fixed unchecked array overwrite in relex().
890:
891: changed for (i in array) to access elements in sorted order.
892: then unchanged it -- it really does run slower in too many cases.
893: left the code in place, commented out.
894:
1.7 millert 895: Feb 10, 1991:
896: check error status on all writes, to avoid banging on full disks.
1.1 tholo 897:
1.7 millert 898: Jan 28, 1991:
899: awk -f - reads the program from stdin.
1.1 tholo 900:
1.7 millert 901: Jan 11, 1991:
902: failed to set numeric state on $0 in cmd|getline context in run.c.
1.1 tholo 903:
1.7 millert 904: Nov 2, 1990:
905: fixed sleazy test for integrality in getsval; use modf.
1.1 tholo 906:
1.7 millert 907: Oct 29, 1990:
908: fixed sleazy buggy code in lib.c that looked (incorrectly) for
909: too long input lines.
1.1 tholo 910:
1.7 millert 911: Oct 14, 1990:
912: fixed the bug on p. 198 in which it couldn't deduce that an
913: argument was an array in some contexts. replaced the error
914: message in intest() by code that damn well makes it an array.
1.1 tholo 915:
1.7 millert 916: Oct 8, 1990:
917: fixed horrible bug: types and values were not preserved in
918: some kinds of self-assignment. (in assign().)
1.1 tholo 919:
1.7 millert 920: Aug 24, 1990:
921: changed NCHARS to 256 to handle 8-bit characters in strings
922: presented to match(), etc.
1.1 tholo 923:
1.7 millert 924: Jun 26, 1990:
925: changed struct rrow (awk.h) to use long instead of int for lval,
926: since cfoll() stores a pointer in it. now works better when int's
927: are smaller than pointers!
1.1 tholo 928:
1.7 millert 929: May 6, 1990:
930: AVA fixed the grammar so that ! is uniformly of the same precedence as
931: unary + and -. This renders illegal some constructs like !x=y, which
932: now has to be parenthesized as !(x=y), and makes others work properly:
933: !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements.
934: (These problems were pointed out by Bob Lenk of Posix.)
1.1 tholo 935:
1.7 millert 936: Added \x to regular expressions (already in strings).
937: Limited octal to octal digits; \8 and \9 are not octal.
938: Centralized the code for parsing escapes in regular expressions.
939: Added a bunch of tests to T.re and T.sub to verify some of this.
1.1 tholo 940:
1.7 millert 941: Feb 9, 1990:
942: fixed null pointer dereference bug in main.c: -F[nothing]. sigh.
1.1 tholo 943:
1.7 millert 944: restored srand behavior: it returns the current seed.
1.1 tholo 945:
1.7 millert 946: Jan 18, 1990:
947: srand now returns previous seed value (0 to start).
1.1 tholo 948:
1.7 millert 949: Jan 5, 1990:
950: fix potential problem in tran.c -- something was freed,
951: then used in freesymtab.
1.1 tholo 952:
1.7 millert 953: Oct 18, 1989:
954: another try to get the max number of open files set with
955: relatively machine-independent code.
1.1 tholo 956:
1.7 millert 957: small fix to input() in case of multiple reads after EOF.
1.1 tholo 958:
1.7 millert 959: Oct 11, 1989:
960: FILENAME is now defined in the BEGIN block -- too many old
961: programs broke.
1.1 tholo 962:
1.7 millert 963: "-" means stdin in getline as well as on the commandline.
1.1 tholo 964:
1.7 millert 965: added a bunch of casts to the code to tell the truth about
966: char * vs. unsigned char *, a right royal pain. added a
967: setlocale call to the front of main, though probably no one
968: has it usefully implemented yet.
1.1 tholo 969:
1.7 millert 970: Aug 24, 1989:
971: removed redundant relational tests against nullnode if parse
972: tree already had a relational at that point.
1.1 tholo 973:
1.7 millert 974: Aug 11, 1989:
975: fixed bug: commandline variable assignment has to look like
976: var=something. (consider the man page for =, in file =.1)
1.1 tholo 977:
1.7 millert 978: changed number of arguments to functions to static arrays
979: to avoid repeated malloc calls.
1.1 tholo 980:
1.7 millert 981: Aug 2, 1989:
982: restored -F (space) separator
1.1 tholo 983:
1.7 millert 984: Jul 30, 1989:
985: added -v x=1 y=2 ... for immediate commandline variable assignment;
986: done before the BEGIN block for sure. they have to precede the
987: program if the program is on the commandline.
988: Modified Aug 2 to require a separate -v for each assignment.
1.1 tholo 989:
1.7 millert 990: Jul 10, 1989:
991: fixed ref-thru-zero bug in environment code in tran.c
1.1 tholo 992:
1.7 millert 993: Jun 23, 1989:
994: add newline to usage message.
1.1 tholo 995:
1.7 millert 996: Jun 14, 1989:
997: added some missing ansi printf conversion letters: %i %X %E %G.
998: no sensible meaning for h or L, so they may not do what one expects.
1.1 tholo 999:
1.7 millert 1000: made %* conversions work.
1.1 tholo 1001:
1.7 millert 1002: changed x^y so that if n is a positive integer, it's done
1003: by explicit multiplication, thus achieving maximum accuracy.
1004: (this should be done by pow() but it seems not to be locally.)
1005: done to x ^= y as well.
1.1 tholo 1006:
1.7 millert 1007: Jun 4, 1989:
1008: ENVIRON array contains environment: if shell variable V=thing,
1009: ENVIRON["V"] is "thing"
1.1 tholo 1010:
1.7 millert 1011: multiple -f arguments permitted. error reporting is naive.
1012: (they were permitted before, but only the last was used.)
1.1 tholo 1013:
1.7 millert 1014: fixed a really stupid botch in the debugging macro dprintf
1.1 tholo 1015:
1.7 millert 1016: fixed order of evaluation of commandline assignments to match
1017: what the book claims: an argument of the form x=e is evaluated
1018: at the time it would have been opened if it were a filename (p 63).
1019: this invalidates the suggested answer to ex 4-1 (p 195).
1.1 tholo 1020:
1.7 millert 1021: removed some code that permitted -F (space) fieldseparator,
1022: since it didn't quite work right anyway. (restored aug 2)
1.1 tholo 1023:
1.7 millert 1024: Apr 27, 1989:
1025: Line number now accumulated correctly for comment lines.
1.1 tholo 1026:
1.7 millert 1027: Apr 26, 1989:
1028: Debugging output now includes a version date,
1029: if one compiles it into the source each time.
1.1 tholo 1030:
1.7 millert 1031: Apr 9, 1989:
1032: Changed grammar to prohibit constants as 3rd arg of sub and gsub;
1033: prevents class of overwriting-a-constant errors. (Last one?)
1034: This invalidates the "banana" example on page 43 of the book.
1.1 tholo 1035:
1.7 millert 1036: Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal),
1037: as in ANSI, for strings. Rescinded the sloppiness that permitted
1038: non-octal digits in \ooo. Warning: not all compilers and libraries
1039: will be able to deal with \x correctly.
1.1 tholo 1040:
1.7 millert 1041: Jan 9, 1989:
1042: Fixed bug that caused tempcell list to contain a duplicate.
1043: The fix is kludgy.
1.1 tholo 1044:
1.7 millert 1045: Dec 17, 1988:
1046: Catches some more commandline errors in main.
1047: Removed redundant decl of modf in run.c (confuses some compilers).
1048: Warning: there's no single declaration of malloc, etc., in awk.h
1049: that seems to satisfy all compilers.
1.1 tholo 1050:
1.7 millert 1051: Dec 7, 1988:
1052: Added a bit of code to error printing to avoid printing nulls.
1053: (Not clear that it actually would.)
1.1 tholo 1054:
1.7 millert 1055: Nov 27, 1988:
1056: With fear and trembling, modified the grammar to permit
1057: multiple pattern-action statements on one line without
1058: an explicit separator. By definition, this capitulation
1059: to the ghost of ancient implementations remains undefined
1060: and thus subject to change without notice or apology.
1061: DO NOT COUNT ON IT.
1.1 tholo 1062:
1.7 millert 1063: Oct 30, 1988:
1064: Fixed bug in call() that failed to recover storage.
1.1 tholo 1065:
1.7 millert 1066: A warning is now generated if there are more arguments
1067: in the call than in the definition (in lieu of fixing
1068: another storage leak).
1.2 millert 1069:
1.7 millert 1070: Oct 20, 1988:
1071: Fixed %c: if expr is numeric, use numeric value;
1072: otherwise print 1st char of string value. still
1073: doesn't work if the value is 0 -- won't print \0.
1.2 millert 1074:
1.7 millert 1075: Added a few more checks for running out of malloc.
1.2 millert 1076:
1.7 millert 1077: Oct 12, 1988:
1078: Fixed bug in call() that freed local arrays twice.
1.5 kstailey 1079:
1.7 millert 1080: Fixed to handle deletion of non-existent array right;
1081: complains about attempt to delete non-array element.
1.6 kstailey 1082:
1.7 millert 1083: Sep 30, 1988:
1084: Now guarantees to evaluate all arguments of built-in
1085: functions, as in C; the appearance is that arguments
1086: are evaluated before the function is called. Places
1087: affected are sub (gsub was ok), substr, printf, and
1088: all the built-in arithmetic functions in bltin().
1089: A warning is generated if a bltin() is called with
1090: the wrong number of arguments.
1.6 kstailey 1091:
1.7 millert 1092: This requires changing makeprof on p167 of the book.
1.6 kstailey 1093:
1.7 millert 1094: Aug 23, 1988:
1095: setting FILENAME in BEGIN caused core dump, apparently
1096: because it was freeing space not allocated by malloc.
1.6 kstailey 1097:
1.7 millert 1098: July 24, 1988:
1099: fixed egregious error in toupper/tolower functions.
1100: still subject to rescinding, however.
1.6 kstailey 1101:
1.7 millert 1102: July 2, 1988:
1103: flush stdout before opening file or pipe
1.6 kstailey 1104:
1.7 millert 1105: July 2, 1988:
1106: performance bug in b.c/cgoto(): not freeing some sets of states.
1107: partial fix only right now, and the number of states increased
1108: to make it less obvious.
1.6 kstailey 1109:
1.7 millert 1110: June 1, 1988:
1111: check error status on close
1.6 kstailey 1112:
1.7 millert 1113: May 28, 1988:
1114: srand returns seed value it's using.
1115: see 1/18/90
1.6 kstailey 1116:
1.7 millert 1117: May 22, 1988:
1118: Removed limit on depth of function calls.
1.6 kstailey 1119:
1.7 millert 1120: May 10, 1988:
1121: Fixed lib.c to permit _ in commandline variable names.
1.6 kstailey 1122:
1.7 millert 1123: Mar 25, 1988:
1124: main.c fixed to recognize -- as terminator of command-
1125: line options. Illegal options flagged.
1126: Error reporting slightly cleaned up.
1.6 kstailey 1127:
1.7 millert 1128: Dec 2, 1987:
1129: Newer C compilers apply a strict scope rule to extern
1130: declarations within functions. Two extern declarations in
1131: lib.c and tran.c have been moved to obviate this problem.
1.6 kstailey 1132:
1.7 millert 1133: Oct xx, 1987:
1134: Reluctantly added toupper and tolower functions.
1135: Subject to rescinding without notice.
1.6 kstailey 1136:
1.7 millert 1137: Sep 17, 1987:
1138: Error-message printer had printf(s) instead of
1139: printf("%s",s); got core dumps when the message
1140: included a %.
1.6 kstailey 1141:
1.7 millert 1142: Sep 12, 1987:
1143: Very long printf strings caused core dump;
1144: fixed aprintf, asprintf, format to catch them.
1145: Can still get a core dump in printf itself.
1.6 kstailey 1146:
1147: