1/**************************************************************** 2Copyright (C) Lucent Technologies 1997 3All Rights Reserved 4 5Permission to use, copy, modify, and distribute this software and 6its documentation for any purpose and without fee is hereby 7granted, provided that the above copyright notice appear in all 8copies and that both that the copyright notice and this 9permission notice and warranty disclaimer appear in supporting 10documentation, and that the name Lucent Technologies or any of 11its entities not be used in advertising or publicity pertaining 12to distribution of the software without specific, written prior 13permission. 14 15LUCENT DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, 16INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. 17IN NO EVENT SHALL LUCENT OR ANY OF ITS ENTITIES BE LIABLE FOR ANY 18SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES 19WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER 20IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, 21ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF 22THIS SOFTWARE. 23****************************************************************/ 24 25This file lists all bug fixes, changes, etc., made since the AWK book 26was sent to the printers in August, 1987. 27 28July 2, 2020: 29 Merge PRs 85 and 86 which fix regressions. Thanks to 30 Tim van der Molen for the fixes. 31 32June 25, 2020: 33 Merge PRs 82 and 84. The latter fixes issue #83. Thanks to 34 Todd Miller and awkfan77. 35 36June 12, 2020: 37 Clear errno before calling errcheck to avoid any spurious errors 38 left over from previous calls that may have set it. Thanks to 39 Todd Miller for the fix, from PR #80. 40 41 Fix Issue #78 by allowing \r to follow floating point numbers in 42 lib.c:is_number. Thanks to GitHub user ajcarr for the report 43 and to Arnold Robbins for the fix. 44 45June 5, 2020: 46 In fldbld(), make sure that inputFS is set before trying to 47 use it. Thanks to Steffen Nurpmeso <steffen@sdaoden.eu> 48 for the report. 49 50May 5, 2020: 51 Fix checks for compilers that can handle noreturn. Thanks to 52 GitHub user enh-google for pointing it out. Closes Issue #79. 53 54April 16, 2020: 55 Handle old compilers that don't support C11 (for noreturn). 56 Thanks to Arnold Robbins. 57 58April 5, 2020: 59 Use <stdnoreturn.h> and noreturn instead of GCC attributes. 60 Thanks to GitHub user awkfan77. Closes PR #77. 61 62February 28, 2020: 63 More cleanups from Christos Zoulas: notably backslash continuation 64 inside strings removes the newline and a fix for RS = "^a". 65 Fix for address sanitizer-found problem. Thanks to GitHub user 66 enh-google. 67 68February 19, 2020: 69 More small cleanups from Christos Zoulas. 70 71February 18, 2020: 72 Additional cleanups from Christos Zoulas. It's no longer necessary 73 to use the -y flag to bison. 74 75February 6, 2020: 76 Additional small cleanups from Christos Zoulas. awk is now 77 a little more robust about reporting I/O errors upon exit. 78 79January 31, 2020: 80 Merge PR #70, which avoids use of variable length arrays. Thanks 81 to GitHub user michaelforney. Fix issue #60 ({0} in interval 82 expressions doesn't work). Also get all tests working again. 83 Thanks to Arnold Robbins. 84 85January 24, 2020: 86 A number of small cleanups from Christos Zoulas. Add the close 87 on exec flag to files/pipes opened for redirection; courtesy of 88 Arnold Robbins. 89 90January 19, 2020: 91 If POSIXLY_CORRECT is set in the environment, then sub and gsub 92 use POSIX rules for multiple backslashes. This fixes Issue #66, 93 while maintaining backwards compatibility. 94 95January 9, 2020: 96 Input/output errors on closing files are now fatal instead of 97 mere warnings. Thanks to Martijn Dekker <martijn@inlv.org>. 98 99January 5, 2020: 100 Fix a bug in the concatentation of two string constants into 101 one done in the grammar. Fixes GitHub issue #61. Thanks 102 to GitHub user awkfan77 for pointing out the direction for 103 the fix. New test T.concat added to the test suite. 104 Fix a few memory leaks reported by valgrind, as well. 105 106December 27, 2019: 107 Fix a bug whereby a{0,3} could match four a's. Thanks to 108 "Anonymous AWK fan" for the report. 109 110December 11, 2019: 111 Further printf-related fixes for 32 bit systems. 112 Thanks again to Christos Zoulas. 113 114December 8, 2019: 115 Fix the return value of sprintf("%d") on 32 bit systems. 116 Thanks to Jim Lowe for the report and to Christos Zoulas 117 for the fix. 118 119November 10, 2019: 120 Convert a number of Boolean integer variables into 121 actual bools. Convert compile_time variable into an 122 enum and simplify some of the related code. Thanks 123 to Arnold Robbins. 124 125November 8, 2019: 126 Fix from Ori Bernstein to get UTF-8 characters instead of 127 bytes when FS = "". This is currently the only bit of 128 the One True Awk that understands multibyte characters. 129 From Arnold Robbins, apply some cleanups in the test suite. 130 131October 25, 2019: 132 More fixes and cleanups from NetBSD, courtesy of Christos 133 Zoulas. Merges PRs 54 and 55. 134 135October 24, 2019: 136 Import second round of code cleanups from NetBSD. Much thanks 137 to Christos Zoulas (GitHub user zoulasc). Merges PR 53. 138 Add an optimization for string concatenation, also from 139 Christos. 140 141October 17, 2019: 142 Import code cleanups from NetBSD. Much thanks to Christos 143 Zoulas (GitHub user zoulasc). Merges PR 51. 144 145October 6, 2019: 146 Import code from NetBSD awk that implements RS as a regular 147 expression. 148 149September 10, 2019: 150 Fixes for various array / memory overruns found via gcc's 151 -fsanitize=unknown. Thanks to Alexander Richardson (GitHub 152 user arichardson). Merges PRs 47 and 48. 153 154July 28, 2019: 155 Import grammar optimization from NetBSD: Two string constants 156 concatenated together get turned into a single string. 157 158July 26, 2019: 159 Support POSIX-specified C-style escape sequences "\a" (alarm) 160 and "\v" (vertical tab) in command line arguments and regular 161 expressions, further to the support for them in strings added on 162 Apr 9, 1989. These now no longer match as literal "a" and "v" 163 characters (as they don't on other awk implementations). 164 Thanks to Martijn Dekker. 165 166July 17, 2019: 167 Pull in a number of code cleanups and minor fixes from 168 Warner Losh's bsd-ota branch. The only user visible change 169 is the use of random(3) as the random number generator. 170 Thanks to Warner Losh for collecting all these fixes in 171 one easy place to get them from. 172 173July 16, 2019: 174 Fix field splitting to use FS value as of the time a record 175 was read or assigned to. Thanks to GitHub user Cody Mello (melloc) 176 for the fix. (Merged from his branch, via PR #42.) Updated 177 testdir/T.split per said PR as well. 178 179June 24, 2019: 180 Extract awktest.tar into testdir directory. Add some very 181 simple mechanics to the makefile for running the tests and 182 for cleaning up. No changes to awk itself. 183 184June 17, 2019: 185 Disallow deleting SYMTAB and its elements, which creates 186 use-after-free bugs. Thanks to GitHub user Cody Mello (melloc) 187 for the fix. (Merged from PR #43.) 188 189June 5, 2019: 190 Allow unmatched right parenthesis in a regular expression to 191 be treated literally. Fixes Issue #40. Thanks to GitHub user 192 Warner Losh (bsdimp) for the report. Thanks to Arnold Robbins 193 for the fix. 194 195May 29,2019: 196 Fix check for command line arguments to no longer require that 197 first character after '=' not be another '='. Reverts change of 198 August 11, 1989. Thanks to GitHub user Jamie Landeg Jones for 199 pointing out the issue; from Issue #38. 200 201Apr 7, 2019: 202 Update awktest.tar(p.50) to use modern options to sort. Needed 203 for Android development. Thanks to GitHub user mohd-akram (Mohamed 204 Akram). From Issue #33. 205 206Mar 12, 2019: 207 Added very simplistic support for cross-compiling in the 208 makefile. We are NOT going to go in the direction of the 209 autotools, though. Thanks to GitHub user nee-san for 210 the basic change. (Merged from PR #34.) 211 212Mar 5, 2019: 213 Added support for POSIX-standard interval expressions (a.k.a. 214 bounds, a.k.a. repetition expressions) in regular expressions, 215 backported (via NetBSD) from Apple awk-24 (20070501). 216 Thanks to Martijn Dekker <martijn@inlv.org> for the port. 217 (Merged from PR #30.) 218 219Mar 3, 2019: 220 Merge PRs as follows: 221 #12: Avoid undefined behaviour when using ctype(3) functions in 222 relex(). Thanks to GitHub user iamleot. 223 #31: Make getline handle numeric strings, and update FIXES. Thanks 224 to GitHub user arnoldrobbins. 225 #32: maketab: support build systems with read-only source. Thanks 226 to GitHub user enh. 227 228Jan 25, 2019: 229 Make getline handle numeric strings properly in all cases. 230 (Thanks, Arnold.) 231 232Jan 21, 2019: 233 Merged a number of small fixes from GitHub pull requests. 234 Thanks to GitHub users Arnold Robbins (arnoldrobbins), 235 Cody Mello (melloc) and Christoph Junghans (junghans). 236 PR numbers: 13-21, 23, 24, 27. 237 238Oct 25, 2018: 239 Added test in maketab.c to prevent generating a proctab entry 240 for YYSTYPE_IS_DEFINED. It was harmless but some gcc settings 241 generated a warning message. Thanks to Nan Xiao for report. 242 243Aug 27, 2018: 244 Disallow '$' in printf formats; arguments evaluated in order 245 and printed in order. 246 247 Added some casts to silence warnings on debugging printfs. 248 (Thanks, Arnold.) 249 250Aug 23, 2018: 251 A long list of fixes courtesy of Arnold Robbins, 252 to whom profound thanks. 253 254 1. ofs-rebuild: OFS value used to rebuild the record was incorrect. 255 Fixed August 19, 2014. Revised fix August 2018. 256 257 2. system-status: Instead of a floating-point division by 256, use 258 the wait(2) macros to create a reasonable exit status. 259 Fixed March 12, 2016. 260 261 3. space: Use provided xisblank() function instead of ispace() for 262 matching [[:blank:]]. 263 264 4. a-format: Add POSIX standard %a and %A to supported formats. Check 265 at runtime that this format is available. 266 267 5. decr-NF: Decrementing NF did not change $0. This is a decades-old 268 bug. There are interactions with the old and new value of OFS as well. 269 Most of the fix came from the NetBSD awk. 270 271 6. string-conv: String conversions of scalars were sticky. Once a 272 conversion to string happened, even with OFMT, that value was used until 273 a new numeric value was assigned, even if OFMT differed from CONVFMT, 274 and also if CONVFMT changed. 275 276 7. unary-plus: Unary plus on a string constant returned the string. 277 Instead, it should convert the value to numeric and give that value. 278 279 Also added Arnold's tests for these to awktest.tar as T.arnold. 280 281Aug 15, 2018: 282 fixed mangled awktest.tar (thanks, Arnold), posted all 283 current (very minor) fixes to github / onetrueawk 284 285Jun 7, 2018: 286 (yes, a long layoff) 287 Updated some broken tests (beebe.tar, T.lilly) 288 [thanks to Arnold Robbins] 289 290Mar 26, 2015: 291 buffer overflow in error reporting; thanks to tobias ulmer 292 and john-mark gurney for spotting it and the fix. 293 294Feb 4, 2013: 295 cleaned up a handful of tests that didn't seem to actually 296 test for correct behavior: T.latin1, T.gawk. 297 298Jan 5, 2013: 299 added ,NULL initializer to static Cells in run.c; not really 300 needed but cleaner. Thanks to Michael Bombardieri. 301 302Dec 20, 2012: 303 fiddled makefile to get correct yacc and bison flags. pick yacc 304 (linux) or bison (mac) as necessary. 305 306 added __attribute__((__noreturn__)) to a couple of lines in 307 proto.h, to silence someone's enthusiastic checker. 308 309 fixed obscure call by value bug in split(a[1],a) reported on 310 9fans. the management of temporary values is just a mess; i 311 took a shortcut by making an extra string copy. thanks 312 to paul patience and arnold robbins for passing it on and for 313 proposed patches. 314 315 tiny fiddle in setfval to eliminate -0 results in T.expr, which 316 has irritated me for 20+ years. 317 318Aug 10, 2011: 319 another fix to avoid core dump with delete(ARGV); again, many thanks 320 to ruslan ermilov. 321 322Aug 7, 2011: 323 split(s, a, //) now behaves the same as split(s, a, "") 324 325Jun 12, 2011: 326 /pat/, \n /pat/ {...} is now legal, though bad style to use. 327 328 added checks to new -v code that permits -vnospace; thanks to 329 ruslan ermilov for spotting this and providing the patch. 330 331 removed fixed limit on number of open files; thanks to aleksey 332 cheusov and christos zoulos. 333 334 fixed day 1 bug that resurrected deleted elements of ARGV when 335 used as filenames (in lib.c). 336 337 minor type fiddles to make gcc -Wall -pedantic happier (but not 338 totally so); turned on -fno-strict-aliasing in makefile. 339 340May 6, 2011: 341 added #ifdef for isblank. 342 now allows -ffoo as well as -f foo arguments. 343 (thanks, ruslan) 344 345May 1, 2011: 346 after advice from todd miller, kevin lo, ruslan ermilov, 347 and arnold robbins, changed srand() to return the previous 348 seed (which is 1 on the first call of srand). the seed is 349 an Awkfloat internally though converted to unsigned int to 350 pass to the library srand(). thanks, everyone. 351 352 fixed a subtle (and i hope low-probability) overflow error 353 in fldbld, by adding space for one extra \0. thanks to 354 robert bassett for spotting this one and providing a fix. 355 356 removed the files related to compilation on windows. i no 357 longer have anything like a current windows environment, so 358 i can't test any of it. 359 360May 23, 2010: 361 fixed long-standing overflow bug in run.c; many thanks to 362 nelson beebe for spotting it and providing the fix. 363 364 fixed bug that didn't parse -vd=1 properly; thanks to santiago 365 vila for spotting it. 366 367Feb 8, 2010: 368 i give up. replaced isblank with isspace in b.c; there are 369 no consistent header files. 370 371Nov 26, 2009: 372 fixed a long-standing issue with when FS takes effect. a 373 change to FS is now noticed immediately for subsequent splits. 374 375 changed the name getline() to awkgetline() to avoid yet another 376 name conflict somewhere. 377 378Feb 11, 2009: 379 temporarily for now defined HAS_ISBLANK, since that seems to 380 be the best way through the thicket. isblank arrived in C99, 381 but seems to be arriving at different systems at different 382 times. 383 384Oct 8, 2008: 385 fixed typo in b.c that set tmpvec wrongly. no one had ever 386 run into the problem, apparently. thanks to alistair crooks. 387 388Oct 23, 2007: 389 minor fix in lib.c: increase inputFS to 100, change malloc 390 for fields to n+1. 391 392 fixed memory fault caused by out of order test in setsval. 393 394 thanks to david o'brien, freebsd, for both fixes. 395 396May 1, 2007: 397 fiddle in makefile to fix for BSD make; thanks to igor sobrado. 398 399Mar 31, 2007: 400 fixed some null pointer refs calling adjbuf. 401 402Feb 21, 2007: 403 fixed a bug in matching the null RE in sub and gsub. thanks to al aho 404 who actually did the fix (in b.c), and to wolfgang seeberg for finding 405 it and providing a very compact test case. 406 407 fixed quotation in b.c; thanks to Hal Pratt and the Princeton Dante 408 Project. 409 410 removed some no-effect asserts in run.c. 411 412 fiddled maketab.c to not complain about bison-generated values. 413 414 removed the obsolete -V argument; fixed --version to print the 415 version and exit. 416 417 fixed wording and an outright error in the usage message; thanks to igor 418 sobrado and jason mcintyre. 419 420 fixed a bug in -d that caused core dump if no program followed. 421 422Jan 1, 2007: 423 dropped mac.code from makefile; there are few non-MacOSX 424 mac's these days. 425 426Jan 17, 2006: 427 system() not flagged as unsafe in the unadvertised -safe option. 428 found it while enhancing tests before shipping the ;login: article. 429 practice what you preach. 430 431 removed the 9-years-obsolete -mr and -mf flags. 432 433 added -version and --version options. 434 435 core dump on linux with BEGIN {nextfile}, now fixed. 436 437 removed some #ifdef's in run.c and lex.c that appear to no 438 longer be necessary. 439 440Apr 24, 2005: 441 modified lib.c so that values of $0 et al are preserved in the END 442 block, apparently as required by posix. thanks to havard eidnes 443 for the report and code. 444 445Jan 14, 2005: 446 fixed infinite loop in parsing, originally found by brian tsang. 447 thanks to arnold robbins for a suggestion that started me 448 rethinking it. 449 450Dec 31, 2004: 451 prevent overflow of -f array in main, head off potential error in 452 call of SYNTAX(), test malloc return in lib.c, all with thanks to 453 todd miller. 454 455Dec 22, 2004: 456 cranked up size of NCHARS; coverity thinks it can be overrun with 457 smaller size, and i think that's right. added some assertions to b.c 458 to catch places where it might overrun. the RE code is still fragile. 459 460Dec 5, 2004: 461 fixed a couple of overflow problems with ridiculous field numbers: 462 e.g., print $(2^32-1). thanks to ruslan ermilov, giorgos keramidas 463 and david o'brien at freebsd.org for patches. this really should 464 be re-done from scratch. 465 466Nov 21, 2004: 467 fixed another 25-year-old RE bug, in split. it's another failure 468 to (re-)initialize. thanks to steve fisher for spotting this and 469 providing a good test case. 470 471Nov 22, 2003: 472 fixed a bug in regular expressions that dates (so help me) from 1977; 473 it's been there from the beginning. an anchored longest match that 474 was longer than the number of states triggered a failure to initialize 475 the machine properly. many thanks to moinak ghosh for not only finding 476 this one but for providing a fix, in some of the most mysterious 477 code known to man. 478 479 fixed a storage leak in call() that appears to have been there since 480 1983 or so -- a function without an explicit return that assigns a 481 string to a parameter leaked a Cell. thanks to moinak ghosh for 482 spotting this very subtle one. 483 484Jul 31, 2003: 485 fixed, thanks to andrey chernov and ruslan ermilov, a bug in lex.c 486 that mis-handled the character 255 in input. (it was being compared 487 to EOF with a signed comparison.) 488 489Jul 29, 2003: 490 fixed (i think) the long-standing botch that included the beginning of 491 line state ^ for RE's in the set of valid characters; this led to a 492 variety of odd problems, including failure to properly match certain 493 regular expressions in non-US locales. thanks to ruslan for keeping 494 at this one. 495 496Jul 28, 2003: 497 n-th try at getting internationalization right, with thanks to volker 498 kiefel, arnold robbins and ruslan ermilov for advice, though they 499 should not be blamed for the outcome. according to posix, "." is the 500 radix character in programs and command line arguments regardless of 501 the locale; otherwise, the locale should prevail for input and output 502 of numbers. so it's intended to work that way. 503 504 i have rescinded the attempt to use strcoll in expanding shorthands in 505 regular expressions (cclenter). its properties are much too 506 surprising; for example [a-c] matches aAbBc in locale en_US but abBcC 507 in locale fr_CA. i can see how this might arise by implementation 508 but i cannot explain it to a human user. (this behavior can be seen 509 in gawk as well; we're leaning on the same library.) 510 511 the issue appears to be that strcoll is meant for sorting, where 512 merging upper and lower case may make sense (though note that unix 513 sort does not do this by default either). it is not appropriate 514 for regular expressions, where the goal is to match specific 515 patterns of characters. in any case, the notations [:lower:], etc., 516 are available in awk, and they are more likely to work correctly in 517 most locales. 518 519 a moratorium is hereby declared on internationalization changes. 520 i apologize to friends and colleagues in other parts of the world. 521 i would truly like to get this "right", but i don't know what 522 that is, and i do not want to keep making changes until it's clear. 523 524Jul 4, 2003: 525 fixed bug that permitted non-terminated RE, as in "awk /x". 526 527Jun 1, 2003: 528 subtle change to split: if source is empty, number of elems 529 is always 0 and the array is not set. 530 531Mar 21, 2003: 532 added some parens to isblank, in another attempt to make things 533 internationally portable. 534 535Mar 14, 2003: 536 the internationalization changes, somewhat modified, are now 537 reinstated. in theory awk will now do character comparisons 538 and case conversions in national language, but "." will always 539 be the decimal point separator on input and output regardless 540 of national language. isblank(){} has an #ifndef. 541 542 this no longer compiles on windows: LC_MESSAGES isn't defined 543 in vc6++. 544 545 fixed subtle behavior in field and record splitting: if FS is 546 a single character and RS is not empty, \n is NOT a separator. 547 this tortuous reading is found in the awk book; behavior now 548 matches gawk and mawk. 549 550Dec 13, 2002: 551 for the moment, the internationalization changes of nov 29 are 552 rolled back -- programs like x = 1.2 don't work in some locales, 553 because the parser is expecting x = 1,2. until i understand this 554 better, this will have to wait. 555 556Nov 29, 2002: 557 modified b.c (with tiny changes in main and run) to support 558 locales, using strcoll and iswhatever tests for posix character 559 classes. thanks to ruslan ermilov (ru@freebsd.org) for code. 560 the function isblank doesn't seem to have propagated to any 561 header file near me, so it's there explicitly. not properly 562 tested on non-ascii character sets by me. 563 564Jun 28, 2002: 565 modified run/format() and tran/getsval() to do a slightly better 566 job on using OFMT for output from print and CONVFMT for other 567 number->string conversions, as promised by posix and done by 568 gawk and mawk. there are still places where it doesn't work 569 right if CONVFMT is changed; by then the STR attribute of the 570 variable has been irrevocably set. thanks to arnold robbins for 571 code and examples. 572 573 fixed subtle bug in format that could get core dump. thanks to 574 Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing. 575 minor cleanup in run.c / format() at the same time. 576 577 added some tests for null pointers to debugging printf's, which 578 were never intended for external consumption. thanks to dave 579 kerns (dkerns@lucent.com) for pointing this out. 580 581 GNU compatibility: an empty regexp matches anything (thanks to 582 dag-erling smorgrav, des@ofug.org). subject to reversion if 583 this does more harm than good. 584 585 pervasive small changes to make things more const-correct, as 586 reported by gcc's -Wwrite-strings. as it says in the gcc manual, 587 this may be more nuisance than useful. provoked by a suggestion 588 and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk 589 590 minor documentation changes to note that this now compiles out 591 of the box on Mac OS X. 592 593Feb 10, 2002: 594 changed types in posix chars structure to quiet solaris cc. 595 596Jan 1, 2002: 597 fflush() or fflush("") flushes all files and pipes. 598 599 length(arrayname) returns number of elements; thanks to 600 arnold robbins for suggestion. 601 602 added a makefile.win to make it easier to build on windows. 603 based on dan allen's buildwin.bat. 604 605Nov 16, 2001: 606 added support for posix character class names like [:digit:], 607 which are not exactly shorter than [0-9] and perhaps no more 608 portable. thanks to dag-erling smorgrav for code. 609 610Feb 16, 2001: 611 removed -m option; no longer needed, and it was actually 612 broken (noted thanks to volker kiefel). 613 614Feb 10, 2001: 615 fixed an appalling bug in gettok: any sequence of digits, +,-, E, e, 616 and period was accepted as a valid number if it started with a period. 617 this would never have happened with the lex version. 618 619 other 1-character botches, now fixed, include a bare $ and a 620 bare " at the end of the input. 621 622Feb 7, 2001: 623 more (const char *) casts in b.c and tran.c to silence warnings. 624 625Nov 15, 2000: 626 fixed a bug introduced in august 1997 that caused expressions 627 like $f[1] to be syntax errors. thanks to arnold robbins for 628 noticing this and providing a fix. 629 630Oct 30, 2000: 631 fixed some nextfile bugs: not handling all cases. thanks to 632 arnold robbins for pointing this out. new regressions added. 633 634 close() is now a function. it returns whatever the library 635 fclose returns, and -1 for closing a file or pipe that wasn't 636 opened. 637 638Sep 24, 2000: 639 permit \n explicitly in character classes; won't work right 640 if comes in as "[\n]" but ok as /[\n]/, because of multiple 641 processing of \'s. thanks to arnold robbins. 642 643July 5, 2000: 644 minor fiddles in tran.c to keep compilers happy about uschar. 645 thanks to norman wilson. 646 647May 25, 2000: 648 yet another attempt at making 8-bit input work, with another 649 band-aid in b.c (member()), and some (uschar) casts to head 650 off potential errors in subscripts (like isdigit). also 651 changed HAT to NCHARS-2. thanks again to santiago vila. 652 653 changed maketab.c to ignore apparently out of range definitions 654 instead of halting; new freeBSD generates one. thanks to 655 jon snader <jsnader@ix.netcom.com> for pointing out the problem. 656 657May 2, 2000: 658 fixed an 8-bit problem in b.c by making several char*'s into 659 unsigned char*'s. not clear i have them all yet. thanks to 660 Santiago Vila <sanvila@unex.es> for the bug report. 661 662Apr 21, 2000: 663 finally found and fixed a memory leak in function call; it's 664 been there since functions were added ~1983. thanks to 665 jon bentley for the test case that found it. 666 667 added test in envinit to catch environment "variables" with 668 names beginning with '='; thanks to Berend Hasselman. 669 670Jul 28, 1999: 671 added test in defn() to catch function foo(foo), which 672 otherwise recurses until core dump. thanks to arnold 673 robbins for noticing this. 674 675Jun 20, 1999: 676 added *bp in gettok in lex.c; appears possible to exit function 677 without terminating the string. thanks to russ cox. 678 679Jun 2, 1999: 680 added function stdinit() to run to initialize files[] array, 681 in case stdin, etc., are not constants; some compilers care. 682 683May 10, 1999: 684 replaced the ERROR ... FATAL, etc., macros with functions 685 based on vprintf, to avoid problems caused by overrunning 686 fixed-size errbuf array. thanks to ralph corderoy for the 687 impetus, and for pointing out a string termination bug in 688 qstring as well. 689 690Apr 21, 1999: 691 fixed bug that caused occasional core dumps with commandline 692 variable with value ending in \. (thanks to nelson beebe for 693 the test case.) 694 695Apr 16, 1999: 696 with code kindly provided by Bruce Lilly, awk now parses 697 /=/ and similar constructs more sensibly in more places. 698 Bruce also provided some helpful test cases. 699 700Apr 5, 1999: 701 changed true/false to True/False in run.c to make it 702 easier to compile with C++. Added some casts on malloc 703 and realloc to be honest about casts; ditto. changed 704 ltype int to long in struct rrow to reduce some 64-bit 705 complaints; other changes scattered throughout for the 706 same purpose. thanks to Nelson Beebe for these portability 707 improvements. 708 709 removed some horrible pointer-int casting in b.c and elsewhere 710 by adding ptoi and itonp to localize the casts, which are 711 all benign. fixed one incipient bug that showed up on sgi 712 in 64-bit mode. 713 714 reset lineno for new source file; include filename in error 715 message. also fixed line number error in continuation lines. 716 (thanks to Nelson Beebe for both of these.) 717 718Mar 24, 1999: 719 Nelson Beebe notes that irix 5.3 yacc dies with a bogus 720 error; use a newer version or switch to bison, since sgi 721 is unlikely to fix it. 722 723Mar 5, 1999: 724 changed isnumber to is_number to avoid the problem caused by 725 versions of ctype.h that include the name isnumber. 726 727 distribution now includes a script for building on a Mac, 728 thanks to Dan Allen. 729 730Feb 20, 1999: 731 fixed memory leaks in run.c (call) and tran.c (setfval). 732 thanks to Stephen Nutt for finding these and providing the fixes. 733 734Jan 13, 1999: 735 replaced srand argument by (unsigned int) in run.c; 736 avoids problem on Mac and potentially on Unix & Windows. 737 thanks to Dan Allen. 738 739 added a few (int) casts to silence useless compiler warnings. 740 e.g., errorflag= in run.c jump(). 741 742 added proctab.c to the bundle outout; one less thing 743 to have to compile out of the box. 744 745 added calls to _popen and _pclose to the win95 stub for 746 pipes (thanks to Steve Adams for this helpful suggestion). 747 seems to work, though properties are not well understood 748 by me, and it appears that under some circumstances the 749 pipe output is truncated. Be careful. 750 751Oct 19, 1998: 752 fixed a couple of bugs in getrec: could fail to update $0 753 after a getline var; because inputFS wasn't initialized, 754 could split $0 on every character, a misleading diversion. 755 756 fixed caching bug in makedfa: LRU was actually removing 757 least often used. 758 759 thanks to ross ridge for finding these, and for providing 760 great bug reports. 761 762May 12, 1998: 763 fixed potential bug in readrec: might fail to update record 764 pointer after growing. thanks to dan levy for spotting this 765 and suggesting the fix. 766 767Mar 12, 1998: 768 added -V to print version number and die. 769 770[notify dave kerns, dkerns@dacsoup.ih.lucent.com] 771 772Feb 11, 1998: 773 subtle silent bug in lex.c: if the program ended with a number 774 longer than 1 digit, part of the input would be pushed back and 775 parsed again because token buffer wasn't terminated right. 776 example: awk 'length($0) > 10'. blush. at least i found it 777 myself. 778 779Aug 31, 1997: 780 s/adelete/awkdelete/: SGI uses this in malloc.h. 781 thanks to nelson beebe for pointing this one out. 782 783Aug 21, 1997: 784 fixed some bugs in sub and gsub when replacement includes \\. 785 this is a dark, horrible corner, but at least now i believe that 786 the behavior is the same as gawk and the intended posix standard. 787 thanks to arnold robbins for advice here. 788 789Aug 9, 1997: 790 somewhat regretfully, replaced the ancient lex-based lexical 791 analyzer with one written in C. it's longer, generates less code, 792 and more portable; the old one depended too much on mysterious 793 properties of lex that were not preserved in other environments. 794 in theory these recognize the same language. 795 796 now using strtod to test whether a string is a number, instead of 797 the convoluted original function. should be more portable and 798 reliable if strtod is implemented right. 799 800 removed now-pointless optimization in makefile that tries to avoid 801 recompilation when awkgram.y is changed but symbols are not. 802 803 removed most fixed-size arrays, though a handful remain, some 804 of which are unchecked. you have been warned. 805 806Aug 4, 1997: 807 with some trepidation, replaced the ancient code that managed 808 fields and $0 in fixed-size arrays with arrays that grow on 809 demand. there is still some tension between trying to make this 810 run fast and making it clean; not sure it's right yet. 811 812 the ill-conceived -mr and -mf arguments are now useful only 813 for debugging. previous dynamic string code removed. 814 815 numerous other minor cleanups along the way. 816 817Jul 30, 1997: 818 using code provided by dan levy (to whom profuse thanks), replaced 819 fixed-size arrays and awkward kludges by a fairly uniform mechanism 820 to grow arrays as needed for printf, sub, gsub, etc. 821 822Jul 23, 1997: 823 falling off the end of a function returns "" and 0, not 0. 824 thanks to arnold robbins. 825 826Jun 17, 1997: 827 replaced several fixed-size arrays by dynamically-created ones 828 in run.c; added overflow tests to some previously unchecked cases. 829 getline, toupper, tolower. 830 831 getline code is still broken in that recursive calls may wind 832 up using the same space. [fixed later] 833 834 increased RECSIZE to 8192 to push problems further over the horizon. 835 836 added \r to \n as input line separator for programs, not data. 837 damn CRLFs. 838 839 modified format() to permit explicit printf("%c", 0) to include 840 a null byte in output. thanks to ken stailey for the fix. 841 842 added a "-safe" argument that disables file output (print >, 843 print >>), process creation (cmd|getline, print |, system), and 844 access to the environment (ENVIRON). this is a first approximation 845 to a "safe" version of awk, but don't rely on it too much. thanks 846 to joan feigenbaum and matt blaze for the inspiration long ago. 847 848Jul 8, 1996: 849 fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to 850 ralph corderoy. 851 852Jun 29, 1996: 853 fixed awful bug in new field splitting; didn't get all the places 854 where input was done. 855 856Jun 28, 1996: 857 changed field-splitting to conform to posix definition: fields are 858 split using the value of FS at the time of input; it used to be 859 the value when the field or NF was first referred to, a much less 860 predictable definition. thanks to arnold robbins for encouragement 861 to do the right thing. 862 863May 28, 1996: 864 fixed appalling but apparently unimportant bug in parsing octal 865 numbers in reg exprs. 866 867 explicit hex in reg exprs now limited to 2 chars: \xa, \xaa. 868 869May 27, 1996: 870 cleaned up some declarations so gcc -Wall is now almost silent. 871 872 makefile now includes backup copies of ytab.c and lexyy.c in case 873 one makes before looking; it also avoids recreating lexyy.c unless 874 really needed. 875 876 s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes 877 with unwisely-written header files. 878 879 thanks to jeffrey friedl for several of these. 880 881May 26, 1996: 882 an attempt to rationalize the (unsigned) char issue. almost all 883 instances of unsigned char have been removed; the handful of places 884 in b.c where chars are used as table indices have been hand-crafted. 885 added some latin-1 tests to the regression, but i'm not confident; 886 none of my compilers seem to care much. thanks to nelson beebe for 887 pointing out some others that do care. 888 889May 2, 1996: 890 removed all register declarations. 891 892 enhanced split(), as in gawk, etc: split(s, a, "") splits s into 893 a[1]...a[length(s)] with each character a single element. 894 895 made the same changes for field-splitting if FS is "". 896 897 added nextfile, as in gawk: causes immediate advance to next 898 input file. (thanks to arnold robbins for inspiration and code). 899 900 small fixes to regexpr code: can now handle []], [[], and 901 variants; [] is now a syntax error, rather than matching 902 everything; [z-a] is now empty, not z. far from complete 903 or correct, however. (thanks to jeffrey friedl for pointing out 904 some awful behaviors.) 905 906Apr 29, 1996: 907 replaced uchar by uschar everywhere; apparently some compilers 908 usurp this name and this causes conflicts. 909 910 fixed call to time in run.c (bltin); arg is time_t *. 911 912 replaced horrible pointer/long punning in b.c by a legitimate 913 union. should be safer on 64-bit machines and cleaner everywhere. 914 (thanks to nelson beebe for pointing out some of these problems.) 915 916 replaced nested comments by #if 0...#endif in run.c, lib.c. 917 918 removed getsval, setsval, execute macros from run.c and lib.c. 919 machines are 100x faster than they were when these macros were 920 first used. 921 922 revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l, 923 y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of 924 portability to nameless systems. 925 926 "make bundle" now includes yacc and lex output files for recipients 927 who don't have yacc or lex. 928 929Aug 15, 1995: 930 initialized Cells in setsymtab more carefully; some fields 931 were not set. (thanks to purify, all of whose complaints i 932 think i now understand.) 933 934 fixed at least one error in gsub that looked at -1-th element 935 of an array when substituting for a null match (e.g., $). 936 937 delete arrayname is now legal; it clears the elements but leaves 938 the array, which may not be the right behavior. 939 940 modified makefile: my current make can't cope with the test used 941 to avoid unnecessary yacc invocations. 942 943Jul 17, 1995: 944 added dynamically growing strings to awk.lx.l and b.c 945 to permit regular expressions to be much bigger. 946 the state arrays can still overflow. 947 948Aug 24, 1994: 949 detect duplicate arguments in function definitions (mdm). 950 951May 11, 1994: 952 trivial fix to printf to limit string size in sub(). 953 954Apr 22, 1994: 955 fixed yet another subtle self-assignment problem: 956 $1 = $2; $1 = $1 clobbered $1. 957 958 Regression tests now use private echo, to avoid quoting problems. 959 960Feb 2, 1994: 961 changed error() to print line number as %d, not %g. 962 963Jul 23, 1993: 964 cosmetic changes: increased sizes of some arrays, 965 reworded some error messages. 966 967 added CONVFMT as in posix (just replaced OFMT in getsval) 968 969 FILENAME is now "" until the first thing that causes a file 970 to be opened. 971 972Nov 28, 1992: 973 deleted yyunput and yyoutput from proto.h; 974 different versions of lex give these different declarations. 975 976May 31, 1992: 977 added -mr N and -mf N options: more record and fields. 978 these really ought to adjust automatically. 979 980 cleaned up some error messages; "out of space" now means 981 malloc returned NULL in all cases. 982 983 changed rehash so that if it runs out, it just returns; 984 things will continue to run slow, but maybe a bit longer. 985 986Apr 24, 1992: 987 remove redundant close of stdin when using -f -. 988 989 got rid of core dump with -d; awk -d just prints date. 990 991Apr 12, 1992: 992 added explicit check for /dev/std(in,out,err) in redirection. 993 unlike gawk, no /dev/fd/n yet. 994 995 added (file/pipe) builtin. hard to test satisfactorily. 996 not posix. 997 998Feb 20, 1992: 999 recompile after abortive changes; should be unchanged. 1000 1001Dec 2, 1991: 1002 die-casting time: converted to ansi C, installed that. 1003 1004Nov 30, 1991: 1005 fixed storage leak in freefa, failing to recover [N]CCL. 1006 thanks to Bill Jones (jones@cs.usask.ca) 1007 1008Nov 19, 1991: 1009 use RAND_MAX instead of literal in builtin(). 1010 1011Nov 12, 1991: 1012 cranked up some fixed-size arrays in b.c, and added a test for 1013 overflow in penter. thanks to mark larsen. 1014 1015Sep 24, 1991: 1016 increased buffer in gsub. a very crude fix to a general problem. 1017 and again on Sep 26. 1018 1019Aug 18, 1991: 1020 enforce variable name syntax for commandline variables: has to 1021 start with letter or _. 1022 1023Jul 27, 1991: 1024 allow newline after ; in for statements. 1025 1026Jul 21, 1991: 1027 fixed so that in self-assignment like $1=$1, side effects 1028 like recomputing $0 take place. (this is getting subtle.) 1029 1030Jun 30, 1991: 1031 better test for detecting too-long output record. 1032 1033Jun 2, 1991: 1034 better defense against very long printf strings. 1035 made break and continue illegal outside of loops. 1036 1037May 13, 1991: 1038 removed extra arg on gettemp, tempfree. minor error message rewording. 1039 1040May 6, 1991: 1041 fixed silly bug in hex parsing in hexstr(). 1042 removed an apparently unnecessary test in isnumber(). 1043 warn about weird printf conversions. 1044 fixed unchecked array overwrite in relex(). 1045 1046 changed for (i in array) to access elements in sorted order. 1047 then unchanged it -- it really does run slower in too many cases. 1048 left the code in place, commented out. 1049 1050Feb 10, 1991: 1051 check error status on all writes, to avoid banging on full disks. 1052 1053Jan 28, 1991: 1054 awk -f - reads the program from stdin. 1055 1056Jan 11, 1991: 1057 failed to set numeric state on $0 in cmd|getline context in run.c. 1058 1059Nov 2, 1990: 1060 fixed sleazy test for integrality in getsval; use modf. 1061 1062Oct 29, 1990: 1063 fixed sleazy buggy code in lib.c that looked (incorrectly) for 1064 too long input lines. 1065 1066Oct 14, 1990: 1067 fixed the bug on p. 198 in which it couldn't deduce that an 1068 argument was an array in some contexts. replaced the error 1069 message in intest() by code that damn well makes it an array. 1070 1071Oct 8, 1990: 1072 fixed horrible bug: types and values were not preserved in 1073 some kinds of self-assignment. (in assign().) 1074 1075Aug 24, 1990: 1076 changed NCHARS to 256 to handle 8-bit characters in strings 1077 presented to match(), etc. 1078 1079Jun 26, 1990: 1080 changed struct rrow (awk.h) to use long instead of int for lval, 1081 since cfoll() stores a pointer in it. now works better when int's 1082 are smaller than pointers! 1083 1084May 6, 1990: 1085 AVA fixed the grammar so that ! is uniformly of the same precedence as 1086 unary + and -. This renders illegal some constructs like !x=y, which 1087 now has to be parenthesized as !(x=y), and makes others work properly: 1088 !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements. 1089 (These problems were pointed out by Bob Lenk of Posix.) 1090 1091 Added \x to regular expressions (already in strings). 1092 Limited octal to octal digits; \8 and \9 are not octal. 1093 Centralized the code for parsing escapes in regular expressions. 1094 Added a bunch of tests to T.re and T.sub to verify some of this. 1095 1096Feb 9, 1990: 1097 fixed null pointer dereference bug in main.c: -F[nothing]. sigh. 1098 1099 restored srand behavior: it returns the current seed. 1100 1101Jan 18, 1990: 1102 srand now returns previous seed value (0 to start). 1103 1104Jan 5, 1990: 1105 fix potential problem in tran.c -- something was freed, 1106 then used in freesymtab. 1107 1108Oct 18, 1989: 1109 another try to get the max number of open files set with 1110 relatively machine-independent code. 1111 1112 small fix to input() in case of multiple reads after EOF. 1113 1114Oct 11, 1989: 1115 FILENAME is now defined in the BEGIN block -- too many old 1116 programs broke. 1117 1118 "-" means stdin in getline as well as on the commandline. 1119 1120 added a bunch of casts to the code to tell the truth about 1121 char * vs. unsigned char *, a right royal pain. added a 1122 setlocale call to the front of main, though probably no one 1123 has it usefully implemented yet. 1124 1125Aug 24, 1989: 1126 removed redundant relational tests against nullnode if parse 1127 tree already had a relational at that point. 1128 1129Aug 11, 1989: 1130 fixed bug: commandline variable assignment has to look like 1131 var=something. (consider the man page for =, in file =.1) 1132 1133 changed number of arguments to functions to static arrays 1134 to avoid repeated malloc calls. 1135 1136Aug 2, 1989: 1137 restored -F (space) separator 1138 1139Jul 30, 1989: 1140 added -v x=1 y=2 ... for immediate commandline variable assignment; 1141 done before the BEGIN block for sure. they have to precede the 1142 program if the program is on the commandline. 1143 Modified Aug 2 to require a separate -v for each assignment. 1144 1145Jul 10, 1989: 1146 fixed ref-thru-zero bug in environment code in tran.c 1147 1148Jun 23, 1989: 1149 add newline to usage message. 1150 1151Jun 14, 1989: 1152 added some missing ansi printf conversion letters: %i %X %E %G. 1153 no sensible meaning for h or L, so they may not do what one expects. 1154 1155 made %* conversions work. 1156 1157 changed x^y so that if n is a positive integer, it's done 1158 by explicit multiplication, thus achieving maximum accuracy. 1159 (this should be done by pow() but it seems not to be locally.) 1160 done to x ^= y as well. 1161 1162Jun 4, 1989: 1163 ENVIRON array contains environment: if shell variable V=thing, 1164 ENVIRON["V"] is "thing" 1165 1166 multiple -f arguments permitted. error reporting is naive. 1167 (they were permitted before, but only the last was used.) 1168 1169 fixed a really stupid botch in the debugging macro dprintf 1170 1171 fixed order of evaluation of commandline assignments to match 1172 what the book claims: an argument of the form x=e is evaluated 1173 at the time it would have been opened if it were a filename (p 63). 1174 this invalidates the suggested answer to ex 4-1 (p 195). 1175 1176 removed some code that permitted -F (space) fieldseparator, 1177 since it didn't quite work right anyway. (restored aug 2) 1178 1179Apr 27, 1989: 1180 Line number now accumulated correctly for comment lines. 1181 1182Apr 26, 1989: 1183 Debugging output now includes a version date, 1184 if one compiles it into the source each time. 1185 1186Apr 9, 1989: 1187 Changed grammar to prohibit constants as 3rd arg of sub and gsub; 1188 prevents class of overwriting-a-constant errors. (Last one?) 1189 This invalidates the "banana" example on page 43 of the book. 1190 1191 Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal), 1192 as in ANSI, for strings. Rescinded the sloppiness that permitted 1193 non-octal digits in \ooo. Warning: not all compilers and libraries 1194 will be able to deal with \x correctly. 1195 1196Jan 9, 1989: 1197 Fixed bug that caused tempcell list to contain a duplicate. 1198 The fix is kludgy. 1199 1200Dec 17, 1988: 1201 Catches some more commandline errors in main. 1202 Removed redundant decl of modf in run.c (confuses some compilers). 1203 Warning: there's no single declaration of malloc, etc., in awk.h 1204 that seems to satisfy all compilers. 1205 1206Dec 7, 1988: 1207 Added a bit of code to error printing to avoid printing nulls. 1208 (Not clear that it actually would.) 1209 1210Nov 27, 1988: 1211 With fear and trembling, modified the grammar to permit 1212 multiple pattern-action statements on one line without 1213 an explicit separator. By definition, this capitulation 1214 to the ghost of ancient implementations remains undefined 1215 and thus subject to change without notice or apology. 1216 DO NOT COUNT ON IT. 1217 1218Oct 30, 1988: 1219 Fixed bug in call() that failed to recover storage. 1220 1221 A warning is now generated if there are more arguments 1222 in the call than in the definition (in lieu of fixing 1223 another storage leak). 1224 1225Oct 20, 1988: 1226 Fixed %c: if expr is numeric, use numeric value; 1227 otherwise print 1st char of string value. still 1228 doesn't work if the value is 0 -- won't print \0. 1229 1230 Added a few more checks for running out of malloc. 1231 1232Oct 12, 1988: 1233 Fixed bug in call() that freed local arrays twice. 1234 1235 Fixed to handle deletion of non-existent array right; 1236 complains about attempt to delete non-array element. 1237 1238Sep 30, 1988: 1239 Now guarantees to evaluate all arguments of built-in 1240 functions, as in C; the appearance is that arguments 1241 are evaluated before the function is called. Places 1242 affected are sub (gsub was ok), substr, printf, and 1243 all the built-in arithmetic functions in bltin(). 1244 A warning is generated if a bltin() is called with 1245 the wrong number of arguments. 1246 1247 This requires changing makeprof on p167 of the book. 1248 1249Aug 23, 1988: 1250 setting FILENAME in BEGIN caused core dump, apparently 1251 because it was freeing space not allocated by malloc. 1252 1253July 24, 1988: 1254 fixed egregious error in toupper/tolower functions. 1255 still subject to rescinding, however. 1256 1257July 2, 1988: 1258 flush stdout before opening file or pipe 1259 1260July 2, 1988: 1261 performance bug in b.c/cgoto(): not freeing some sets of states. 1262 partial fix only right now, and the number of states increased 1263 to make it less obvious. 1264 1265June 1, 1988: 1266 check error status on close 1267 1268May 28, 1988: 1269 srand returns seed value it's using. 1270 see 1/18/90 1271 1272May 22, 1988: 1273 Removed limit on depth of function calls. 1274 1275May 10, 1988: 1276 Fixed lib.c to permit _ in commandline variable names. 1277 1278Mar 25, 1988: 1279 main.c fixed to recognize -- as terminator of command- 1280 line options. Illegal options flagged. 1281 Error reporting slightly cleaned up. 1282 1283Dec 2, 1987: 1284 Newer C compilers apply a strict scope rule to extern 1285 declarations within functions. Two extern declarations in 1286 lib.c and tran.c have been moved to obviate this problem. 1287 1288Oct xx, 1987: 1289 Reluctantly added toupper and tolower functions. 1290 Subject to rescinding without notice. 1291 1292Sep 17, 1987: 1293 Error-message printer had printf(s) instead of 1294 printf("%s",s); got core dumps when the message 1295 included a %. 1296 1297Sep 12, 1987: 1298 Very long printf strings caused core dump; 1299 fixed aprintf, asprintf, format to catch them. 1300 Can still get a core dump in printf itself. 1301 1302 1303