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