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 28Sep 06, 2023: 29 Fix edge case where FS is changed on commandline. Thanks to 30 Gordon Shephard and Miguel Pineiro Jr. 31 32 Fix regular expression clobbering in the lexer, where lexer does 33 not make a copy of regexp literals. also makedfa memory leaks have 34 been plugged. Thanks to Miguel Pineiro Jr. 35 36Dec 15, 2022: 37 Force hex escapes in strings to be no more than two characters, 38 as they already are in regular expressions. This brings internal 39 consistency, as well as consistency with gawk. Thanks to 40 Arnold Robbins. 41 42Sep 12, 2022: 43 adjbuf minlen error (cannot be 0) in cat, resulting in NULL pbuf. 44 discovered by todd miller. also use-after-free issue with 45 tempfree in cat, thanks to Miguel Pineiro Jr and valgrind. 46 47Aug 30, 2022: 48 Various leaks and use-after-free issues plugged/fixed. 49 Thanks to Miguel Pineiro Jr. <mpj@pineiro.cc>. 50 51May 23, 2022: 52 Memory leak when assigning a string to some of the built-in 53 variables. allocated string erroneously marked DONTFREE. 54 Thanks to Miguel Pineiro Jr. <mpj@pineiro.cc>. 55 56Mar 14, 2022: 57 Historic bug: command-line "name=value" assignment had been 58 truncating its entry in ARGV. (circa 1989) Thanks to 59 Miguel Pineiro Jr. <mpj@pineiro.cc>. 60 61Mar 3, 2022: 62 Fixed file management memory leak that appears to have been 63 there since the files array was first initialized with stdin, 64 stdout, and stderr (circa 1992). Thanks to Miguel Pineiro Jr. 65 <mpj@pineiro.cc>. 66 67December 8, 2021: 68 The error handling in closefile and closeall was mangled. Long 69 standing warnings had been made fatal and some fatal errors went 70 undetected. Thanks to Miguel Pineiro Jr. <mpj@pineiro.cc>. 71 72Nov 03, 2021: 73 getline accesses uninitialized data after getrec() 74 returns 0 on EOF and leaves the contents of buf unchanged. 75 Thanks to Volodymyr Gubarkov, and Todd C Miller. 76 77Oct 12, 2021: 78 The fix for #83 changed the code to insert 2 chars, but the 79 call to adjbuf just above it only allows for 1 char. This can 80 cause a heap buffer overflow. 81 82July 27, 2021: 83 As per IEEE Std 1003.1-2008, -F "str" is now consistent with 84 -v FS="str" when str is null. Thanks to Warner Losh. 85 86July 24, 2021: 87 Fix readrec's definition of a record. This fixes an issue 88 with NetBSD's RS regular expression support that can cause 89 an infinite read loop. Thanks to Miguel Pineiro Jr. 90 91 Fix regular expression RS ^-anchoring. RS ^-anchoring needs to 92 know if it is reading the first record of a file. This change 93 restores a missing line that was overlooked when porting NetBSD's 94 RS regex functionality. Thanks to Miguel Pineiro Jr. 95 96 Fix size computation in replace_repeat() for special case 97 REPEAT_WITH_Q. Thanks to Todd C. Miller. 98 99February 15, 2021: 100 Small fix so that awk will compile again with g++. Thanks to 101 Arnold Robbins. 102 103January 06, 2021: 104 Fix a decision bug with trailing stuff in lib.c:is_valid_number 105 after recent changes. Thanks to Ozan Yigit. 106 107December 18, 2020: 108 Fix problems converting inf and NaN values in lib.c:is_valid_number. 109 Enhance number to string conversion to do the right thing for 110 NaN and inf values. Things are now pretty much the same as in 111 gawk. (Found a gawk bug while we're at it.) Added a torture 112 test for these values. Thanks to Arnold Robbins. Allows closing 113 of PR #101. 114 115December 15, 2020: 116 Merge PR #99, which gets the right header for strcasecmp. 117 Thanks to GitHub user michaelforney. 118 119December 8, 2020: 120 Merge PR #98: Disallow hex data. Allow only +nan, -nan, 121 +inf, -inf (case independent) to give NaN and infinity values. 122 Improve things so that string to double conversion is only 123 done once, yielding something of a speedup. This obviate 124 PR #95. Thanks to Arnold Robbins. 125 126December 3, 2020: 127 Fix to argument parsing to avoid printing spurious newlines. 128 Thanks to Todd Miller. Merges PR #97. 129 130October 13, 2020: 131 Add casts before all the calls to malloc/calloc/realloc in order 132 to get it to compile with g++. Thanks to Arnold Robbins. 133 134August 16, 2020: 135 Additional fixes for DJGPP. Thanks to Eli Zaretskii for 136 the testing. 137 138August 7, 2020: 139 Merge PR #93, which adds casts to (void*) for debug prints 140 using the %p format specifier. Thanks to GitHub user YongHaoWu 141 ("Chris") for the fixes. 142 143August 4, 2020: 144 In run.c, use non-restartable multibyte routines to attain 145 portability to DJGPP. Should fix Issue 92. Thanks to Albert Wik 146 for the report and to Todd Miller for the suggested fix. 147 148July 30, 2020: 149 Merge PRs 88-91 which fix small bugs. Thanks to Todd Miller and 150 Tim van der Molen for the fixes. 151 152 In order to make life easier, we move exclusively to bison 153 as the parser generator. 154 155July 2, 2020: 156 Merge PRs 85 and 86 which fix regressions. Thanks to 157 Tim van der Molen for the fixes. 158 159June 25, 2020: 160 Merge PRs 82 and 84. The latter fixes issue #83. Thanks to 161 Todd Miller and awkfan77. 162 163June 12, 2020: 164 Clear errno before calling errcheck to avoid any spurious errors 165 left over from previous calls that may have set it. Thanks to 166 Todd Miller for the fix, from PR #80. 167 168 Fix Issue #78 by allowing \r to follow floating point numbers in 169 lib.c:is_number. Thanks to GitHub user ajcarr for the report 170 and to Arnold Robbins for the fix. 171 172June 5, 2020: 173 In fldbld(), make sure that inputFS is set before trying to 174 use it. Thanks to Steffen Nurpmeso <steffen@sdaoden.eu> 175 for the report. 176 177May 5, 2020: 178 Fix checks for compilers that can handle noreturn. Thanks to 179 GitHub user enh-google for pointing it out. Closes Issue #79. 180 181April 16, 2020: 182 Handle old compilers that don't support C11 (for noreturn). 183 Thanks to Arnold Robbins. 184 185April 5, 2020: 186 Use <stdnoreturn.h> and noreturn instead of GCC attributes. 187 Thanks to GitHub user awkfan77. Closes PR #77. 188 189February 28, 2020: 190 More cleanups from Christos Zoulas: notably backslash continuation 191 inside strings removes the newline and a fix for RS = "^a". 192 Fix for address sanitizer-found problem. Thanks to GitHub user 193 enh-google. 194 195February 19, 2020: 196 More small cleanups from Christos Zoulas. 197 198February 18, 2020: 199 Additional cleanups from Christos Zoulas. It's no longer necessary 200 to use the -y flag to bison. 201 202February 6, 2020: 203 Additional small cleanups from Christos Zoulas. awk is now 204 a little more robust about reporting I/O errors upon exit. 205 206January 31, 2020: 207 Merge PR #70, which avoids use of variable length arrays. Thanks 208 to GitHub user michaelforney. Fix issue #60 ({0} in interval 209 expressions doesn't work). Also get all tests working again. 210 Thanks to Arnold Robbins. 211 212January 24, 2020: 213 A number of small cleanups from Christos Zoulas. Add the close 214 on exec flag to files/pipes opened for redirection; courtesy of 215 Arnold Robbins. 216 217January 19, 2020: 218 If POSIXLY_CORRECT is set in the environment, then sub and gsub 219 use POSIX rules for multiple backslashes. This fixes Issue #66, 220 while maintaining backwards compatibility. 221 222January 9, 2020: 223 Input/output errors on closing files are now fatal instead of 224 mere warnings. Thanks to Martijn Dekker <martijn@inlv.org>. 225 226January 5, 2020: 227 Fix a bug in the concatentation of two string constants into 228 one done in the grammar. Fixes GitHub issue #61. Thanks 229 to GitHub user awkfan77 for pointing out the direction for 230 the fix. New test T.concat added to the test suite. 231 Fix a few memory leaks reported by valgrind, as well. 232 233December 27, 2019: 234 Fix a bug whereby a{0,3} could match four a's. Thanks to 235 "Anonymous AWK fan" for the report. 236 237December 11, 2019: 238 Further printf-related fixes for 32 bit systems. 239 Thanks again to Christos Zoulas. 240 241December 8, 2019: 242 Fix the return value of sprintf("%d") on 32 bit systems. 243 Thanks to Jim Lowe for the report and to Christos Zoulas 244 for the fix. 245 246November 10, 2019: 247 Convert a number of Boolean integer variables into 248 actual bools. Convert compile_time variable into an 249 enum and simplify some of the related code. Thanks 250 to Arnold Robbins. 251 252November 8, 2019: 253 Fix from Ori Bernstein to get UTF-8 characters instead of 254 bytes when FS = "". This is currently the only bit of 255 the One True Awk that understands multibyte characters. 256 From Arnold Robbins, apply some cleanups in the test suite. 257 258October 25, 2019: 259 More fixes and cleanups from NetBSD, courtesy of Christos 260 Zoulas. Merges PRs 54 and 55. 261 262October 24, 2019: 263 Import second round of code cleanups from NetBSD. Much thanks 264 to Christos Zoulas (GitHub user zoulasc). Merges PR 53. 265 Add an optimization for string concatenation, also from 266 Christos. 267 268October 17, 2019: 269 Import code cleanups from NetBSD. Much thanks to Christos 270 Zoulas (GitHub user zoulasc). Merges PR 51. 271 272October 6, 2019: 273 Import code from NetBSD awk that implements RS as a regular 274 expression. 275 276September 10, 2019: 277 Fixes for various array / memory overruns found via gcc's 278 -fsanitize=unknown. Thanks to Alexander Richardson (GitHub 279 user arichardson). Merges PRs 47 and 48. 280 281July 28, 2019: 282 Import grammar optimization from NetBSD: Two string constants 283 concatenated together get turned into a single string. 284 285July 26, 2019: 286 Support POSIX-specified C-style escape sequences "\a" (alarm) 287 and "\v" (vertical tab) in command line arguments and regular 288 expressions, further to the support for them in strings added on 289 Apr 9, 1989. These now no longer match as literal "a" and "v" 290 characters (as they don't on other awk implementations). 291 Thanks to Martijn Dekker. 292 293July 17, 2019: 294 Pull in a number of code cleanups and minor fixes from 295 Warner Losh's bsd-ota branch. The only user visible change 296 is the use of random(3) as the random number generator. 297 Thanks to Warner Losh for collecting all these fixes in 298 one easy place to get them from. 299 300July 16, 2019: 301 Fix field splitting to use FS value as of the time a record 302 was read or assigned to. Thanks to GitHub user Cody Mello (melloc) 303 for the fix. (Merged from his branch, via PR #42.) Updated 304 testdir/T.split per said PR as well. 305 306June 24, 2019: 307 Extract awktest.tar into testdir directory. Add some very 308 simple mechanics to the makefile for running the tests and 309 for cleaning up. No changes to awk itself. 310 311June 17, 2019: 312 Disallow deleting SYMTAB and its elements, which creates 313 use-after-free bugs. Thanks to GitHub user Cody Mello (melloc) 314 for the fix. (Merged from PR #43.) 315 316June 5, 2019: 317 Allow unmatched right parenthesis in a regular expression to 318 be treated literally. Fixes Issue #40. Thanks to GitHub user 319 Warner Losh (bsdimp) for the report. Thanks to Arnold Robbins 320 for the fix. 321 322May 29,2019: 323 Fix check for command line arguments to no longer require that 324 first character after '=' not be another '='. Reverts change of 325 August 11, 1989. Thanks to GitHub user Jamie Landeg Jones for 326 pointing out the issue; from Issue #38. 327 328Apr 7, 2019: 329 Update awktest.tar(p.50) to use modern options to sort. Needed 330 for Android development. Thanks to GitHub user mohd-akram (Mohamed 331 Akram). From Issue #33. 332 333Mar 12, 2019: 334 Added very simplistic support for cross-compiling in the 335 makefile. We are NOT going to go in the direction of the 336 autotools, though. Thanks to GitHub user nee-san for 337 the basic change. (Merged from PR #34.) 338 339Mar 5, 2019: 340 Added support for POSIX-standard interval expressions (a.k.a. 341 bounds, a.k.a. repetition expressions) in regular expressions, 342 backported (via NetBSD) from Apple awk-24 (20070501). 343 Thanks to Martijn Dekker <martijn@inlv.org> for the port. 344 (Merged from PR #30.) 345 346Mar 3, 2019: 347 Merge PRs as follows: 348 #12: Avoid undefined behaviour when using ctype(3) functions in 349 relex(). Thanks to GitHub user iamleot. 350 #31: Make getline handle numeric strings, and update FIXES. Thanks 351 to GitHub user Arnold Robbins (arnoldrobbins) 352 #32: maketab: support build systems with read-only source. Thanks 353 to GitHub user enh. 354 355Jan 25, 2019: 356 Make getline handle numeric strings properly in all cases. 357 (Thanks, Arnold.) 358 359Jan 21, 2019: 360 Merged a number of small fixes from GitHub pull requests. 361 Thanks to GitHub users Arnold Robbins (arnoldrobbins), 362 Cody Mello (melloc) and Christoph Junghans (junghans). 363 PR numbers: 13-21, 23, 24, 27. 364 365Oct 25, 2018: 366 Added test in maketab.c to prevent generating a proctab entry 367 for YYSTYPE_IS_DEFINED. It was harmless but some gcc settings 368 generated a warning message. Thanks to Nan Xiao for report. 369 370Aug 27, 2018: 371 Disallow '$' in printf formats; arguments evaluated in order 372 and printed in order. 373 374 Added some casts to silence warnings on debugging printfs. 375 (Thanks, Arnold.) 376 377Aug 23, 2018: 378 A long list of fixes courtesy of Arnold Robbins, 379 to whom profound thanks. 380 381 1. ofs-rebuild: OFS value used to rebuild the record was incorrect. 382 Fixed August 19, 2014. Revised fix August 2018. 383 384 2. system-status: Instead of a floating-point division by 256, use 385 the wait(2) macros to create a reasonable exit status. 386 Fixed March 12, 2016. 387 388 3. space: Use provided xisblank() function instead of ispace() for 389 matching [[:blank:]]. 390 391 4. a-format: Add POSIX standard %a and %A to supported formats. Check 392 at runtime that this format is available. 393 394 5. decr-NF: Decrementing NF did not change $0. This is a decades-old 395 bug. There are interactions with the old and new value of OFS as well. 396 Most of the fix came from the NetBSD awk. 397 398 6. string-conv: String conversions of scalars were sticky. Once a 399 conversion to string happened, even with OFMT, that value was used until 400 a new numeric value was assigned, even if OFMT differed from CONVFMT, 401 and also if CONVFMT changed. 402 403 7. unary-plus: Unary plus on a string constant returned the string. 404 Instead, it should convert the value to numeric and give that value. 405 406 Also added Arnold's tests for these to awktest.tar as T.arnold. 407 408Aug 15, 2018: 409 fixed mangled awktest.tar (thanks, Arnold), posted all 410 current (very minor) fixes to github / onetrueawk 411 412Jun 7, 2018: 413 (yes, a long layoff) 414 Updated some broken tests (beebe.tar, T.lilly) 415 [thanks to Arnold Robbins] 416 417Mar 26, 2015: 418 buffer overflow in error reporting; thanks to tobias ulmer 419 and john-mark gurney for spotting it and the fix. 420 421Feb 4, 2013: 422 cleaned up a handful of tests that didn't seem to actually 423 test for correct behavior: T.latin1, T.gawk. 424 425Jan 5, 2013: 426 added ,NULL initializer to static Cells in run.c; not really 427 needed but cleaner. Thanks to Michael Bombardieri. 428 429Dec 20, 2012: 430 fiddled makefile to get correct yacc and bison flags. pick yacc 431 (linux) or bison (mac) as necessary. 432 433 added __attribute__((__noreturn__)) to a couple of lines in 434 proto.h, to silence someone's enthusiastic checker. 435 436 fixed obscure call by value bug in split(a[1],a) reported on 437 9fans. the management of temporary values is just a mess; i 438 took a shortcut by making an extra string copy. thanks 439 to paul patience and arnold robbins for passing it on and for 440 proposed patches. 441 442 tiny fiddle in setfval to eliminate -0 results in T.expr, which 443 has irritated me for 20+ years. 444 445Aug 10, 2011: 446 another fix to avoid core dump with delete(ARGV); again, many thanks 447 to ruslan ermilov. 448 449Aug 7, 2011: 450 split(s, a, //) now behaves the same as split(s, a, "") 451 452Jun 12, 2011: 453 /pat/, \n /pat/ {...} is now legal, though bad style to use. 454 455 added checks to new -v code that permits -vnospace; thanks to 456 ruslan ermilov for spotting this and providing the patch. 457 458 removed fixed limit on number of open files; thanks to aleksey 459 cheusov and christos zoulos. 460 461 fixed day 1 bug that resurrected deleted elements of ARGV when 462 used as filenames (in lib.c). 463 464 minor type fiddles to make gcc -Wall -pedantic happier (but not 465 totally so); turned on -fno-strict-aliasing in makefile. 466 467May 6, 2011: 468 added #ifdef for isblank. 469 now allows -ffoo as well as -f foo arguments. 470 (thanks, ruslan) 471 472May 1, 2011: 473 after advice from todd miller, kevin lo, ruslan ermilov, 474 and arnold robbins, changed srand() to return the previous 475 seed (which is 1 on the first call of srand). the seed is 476 an Awkfloat internally though converted to unsigned int to 477 pass to the library srand(). thanks, everyone. 478 479 fixed a subtle (and i hope low-probability) overflow error 480 in fldbld, by adding space for one extra \0. thanks to 481 robert bassett for spotting this one and providing a fix. 482 483 removed the files related to compilation on windows. i no 484 longer have anything like a current windows environment, so 485 i can't test any of it. 486 487May 23, 2010: 488 fixed long-standing overflow bug in run.c; many thanks to 489 nelson beebe for spotting it and providing the fix. 490 491 fixed bug that didn't parse -vd=1 properly; thanks to santiago 492 vila for spotting it. 493 494Feb 8, 2010: 495 i give up. replaced isblank with isspace in b.c; there are 496 no consistent header files. 497 498Nov 26, 2009: 499 fixed a long-standing issue with when FS takes effect. a 500 change to FS is now noticed immediately for subsequent splits. 501 502 changed the name getline() to awkgetline() to avoid yet another 503 name conflict somewhere. 504 505Feb 11, 2009: 506 temporarily for now defined HAS_ISBLANK, since that seems to 507 be the best way through the thicket. isblank arrived in C99, 508 but seems to be arriving at different systems at different 509 times. 510 511Oct 8, 2008: 512 fixed typo in b.c that set tmpvec wrongly. no one had ever 513 run into the problem, apparently. thanks to alistair crooks. 514 515Oct 23, 2007: 516 minor fix in lib.c: increase inputFS to 100, change malloc 517 for fields to n+1. 518 519 fixed memory fault caused by out of order test in setsval. 520 521 thanks to david o'brien, freebsd, for both fixes. 522 523May 1, 2007: 524 fiddle in makefile to fix for BSD make; thanks to igor sobrado. 525 526Mar 31, 2007: 527 fixed some null pointer refs calling adjbuf. 528 529Feb 21, 2007: 530 fixed a bug in matching the null RE in sub and gsub. thanks to al aho 531 who actually did the fix (in b.c), and to wolfgang seeberg for finding 532 it and providing a very compact test case. 533 534 fixed quotation in b.c; thanks to Hal Pratt and the Princeton Dante 535 Project. 536 537 removed some no-effect asserts in run.c. 538 539 fiddled maketab.c to not complain about bison-generated values. 540 541 removed the obsolete -V argument; fixed --version to print the 542 version and exit. 543 544 fixed wording and an outright error in the usage message; thanks to igor 545 sobrado and jason mcintyre. 546 547 fixed a bug in -d that caused core dump if no program followed. 548 549Jan 1, 2007: 550 dropped mac.code from makefile; there are few non-MacOSX 551 mac's these days. 552 553Jan 17, 2006: 554 system() not flagged as unsafe in the unadvertised -safe option. 555 found it while enhancing tests before shipping the ;login: article. 556 practice what you preach. 557 558 removed the 9-years-obsolete -mr and -mf flags. 559 560 added -version and --version options. 561 562 core dump on linux with BEGIN {nextfile}, now fixed. 563 564 removed some #ifdef's in run.c and lex.c that appear to no 565 longer be necessary. 566 567Apr 24, 2005: 568 modified lib.c so that values of $0 et al are preserved in the END 569 block, apparently as required by posix. thanks to havard eidnes 570 for the report and code. 571 572Jan 14, 2005: 573 fixed infinite loop in parsing, originally found by brian tsang. 574 thanks to arnold robbins for a suggestion that started me 575 rethinking it. 576 577Dec 31, 2004: 578 prevent overflow of -f array in main, head off potential error in 579 call of SYNTAX(), test malloc return in lib.c, all with thanks to 580 todd miller. 581 582Dec 22, 2004: 583 cranked up size of NCHARS; coverity thinks it can be overrun with 584 smaller size, and i think that's right. added some assertions to b.c 585 to catch places where it might overrun. the RE code is still fragile. 586 587Dec 5, 2004: 588 fixed a couple of overflow problems with ridiculous field numbers: 589 e.g., print $(2^32-1). thanks to ruslan ermilov, giorgos keramidas 590 and david o'brien at freebsd.org for patches. this really should 591 be re-done from scratch. 592 593Nov 21, 2004: 594 fixed another 25-year-old RE bug, in split. it's another failure 595 to (re-)initialize. thanks to steve fisher for spotting this and 596 providing a good test case. 597 598Nov 22, 2003: 599 fixed a bug in regular expressions that dates (so help me) from 1977; 600 it's been there from the beginning. an anchored longest match that 601 was longer than the number of states triggered a failure to initialize 602 the machine properly. many thanks to moinak ghosh for not only finding 603 this one but for providing a fix, in some of the most mysterious 604 code known to man. 605 606 fixed a storage leak in call() that appears to have been there since 607 1983 or so -- a function without an explicit return that assigns a 608 string to a parameter leaked a Cell. thanks to moinak ghosh for 609 spotting this very subtle one. 610 611Jul 31, 2003: 612 fixed, thanks to andrey chernov and ruslan ermilov, a bug in lex.c 613 that mis-handled the character 255 in input. (it was being compared 614 to EOF with a signed comparison.) 615 616Jul 29, 2003: 617 fixed (i think) the long-standing botch that included the beginning of 618 line state ^ for RE's in the set of valid characters; this led to a 619 variety of odd problems, including failure to properly match certain 620 regular expressions in non-US locales. thanks to ruslan for keeping 621 at this one. 622 623Jul 28, 2003: 624 n-th try at getting internationalization right, with thanks to volker 625 kiefel, arnold robbins and ruslan ermilov for advice, though they 626 should not be blamed for the outcome. according to posix, "." is the 627 radix character in programs and command line arguments regardless of 628 the locale; otherwise, the locale should prevail for input and output 629 of numbers. so it's intended to work that way. 630 631 i have rescinded the attempt to use strcoll in expanding shorthands in 632 regular expressions (cclenter). its properties are much too 633 surprising; for example [a-c] matches aAbBc in locale en_US but abBcC 634 in locale fr_CA. i can see how this might arise by implementation 635 but i cannot explain it to a human user. (this behavior can be seen 636 in gawk as well; we're leaning on the same library.) 637 638 the issue appears to be that strcoll is meant for sorting, where 639 merging upper and lower case may make sense (though note that unix 640 sort does not do this by default either). it is not appropriate 641 for regular expressions, where the goal is to match specific 642 patterns of characters. in any case, the notations [:lower:], etc., 643 are available in awk, and they are more likely to work correctly in 644 most locales. 645 646 a moratorium is hereby declared on internationalization changes. 647 i apologize to friends and colleagues in other parts of the world. 648 i would truly like to get this "right", but i don't know what 649 that is, and i do not want to keep making changes until it's clear. 650 651Jul 4, 2003: 652 fixed bug that permitted non-terminated RE, as in "awk /x". 653 654Jun 1, 2003: 655 subtle change to split: if source is empty, number of elems 656 is always 0 and the array is not set. 657 658Mar 21, 2003: 659 added some parens to isblank, in another attempt to make things 660 internationally portable. 661 662Mar 14, 2003: 663 the internationalization changes, somewhat modified, are now 664 reinstated. in theory awk will now do character comparisons 665 and case conversions in national language, but "." will always 666 be the decimal point separator on input and output regardless 667 of national language. isblank(){} has an #ifndef. 668 669 this no longer compiles on windows: LC_MESSAGES isn't defined 670 in vc6++. 671 672 fixed subtle behavior in field and record splitting: if FS is 673 a single character and RS is not empty, \n is NOT a separator. 674 this tortuous reading is found in the awk book; behavior now 675 matches gawk and mawk. 676 677Dec 13, 2002: 678 for the moment, the internationalization changes of nov 29 are 679 rolled back -- programs like x = 1.2 don't work in some locales, 680 because the parser is expecting x = 1,2. until i understand this 681 better, this will have to wait. 682 683Nov 29, 2002: 684 modified b.c (with tiny changes in main and run) to support 685 locales, using strcoll and iswhatever tests for posix character 686 classes. thanks to ruslan ermilov (ru@freebsd.org) for code. 687 the function isblank doesn't seem to have propagated to any 688 header file near me, so it's there explicitly. not properly 689 tested on non-ascii character sets by me. 690 691Jun 28, 2002: 692 modified run/format() and tran/getsval() to do a slightly better 693 job on using OFMT for output from print and CONVFMT for other 694 number->string conversions, as promised by posix and done by 695 gawk and mawk. there are still places where it doesn't work 696 right if CONVFMT is changed; by then the STR attribute of the 697 variable has been irrevocably set. thanks to arnold robbins for 698 code and examples. 699 700 fixed subtle bug in format that could get core dump. thanks to 701 Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing. 702 minor cleanup in run.c / format() at the same time. 703 704 added some tests for null pointers to debugging printf's, which 705 were never intended for external consumption. thanks to dave 706 kerns (dkerns@lucent.com) for pointing this out. 707 708 GNU compatibility: an empty regexp matches anything (thanks to 709 dag-erling smorgrav, des@ofug.org). subject to reversion if 710 this does more harm than good. 711 712 pervasive small changes to make things more const-correct, as 713 reported by gcc's -Wwrite-strings. as it says in the gcc manual, 714 this may be more nuisance than useful. provoked by a suggestion 715 and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk 716 717 minor documentation changes to note that this now compiles out 718 of the box on Mac OS X. 719 720Feb 10, 2002: 721 changed types in posix chars structure to quiet solaris cc. 722 723Jan 1, 2002: 724 fflush() or fflush("") flushes all files and pipes. 725 726 length(arrayname) returns number of elements; thanks to 727 arnold robbins for suggestion. 728 729 added a makefile.win to make it easier to build on windows. 730 based on dan allen's buildwin.bat. 731 732Nov 16, 2001: 733 added support for posix character class names like [:digit:], 734 which are not exactly shorter than [0-9] and perhaps no more 735 portable. thanks to dag-erling smorgrav for code. 736 737Feb 16, 2001: 738 removed -m option; no longer needed, and it was actually 739 broken (noted thanks to volker kiefel). 740 741Feb 10, 2001: 742 fixed an appalling bug in gettok: any sequence of digits, +,-, E, e, 743 and period was accepted as a valid number if it started with a period. 744 this would never have happened with the lex version. 745 746 other 1-character botches, now fixed, include a bare $ and a 747 bare " at the end of the input. 748 749Feb 7, 2001: 750 more (const char *) casts in b.c and tran.c to silence warnings. 751 752Nov 15, 2000: 753 fixed a bug introduced in august 1997 that caused expressions 754 like $f[1] to be syntax errors. thanks to arnold robbins for 755 noticing this and providing a fix. 756 757Oct 30, 2000: 758 fixed some nextfile bugs: not handling all cases. thanks to 759 arnold robbins for pointing this out. new regressions added. 760 761 close() is now a function. it returns whatever the library 762 fclose returns, and -1 for closing a file or pipe that wasn't 763 opened. 764 765Sep 24, 2000: 766 permit \n explicitly in character classes; won't work right 767 if comes in as "[\n]" but ok as /[\n]/, because of multiple 768 processing of \'s. thanks to arnold robbins. 769 770July 5, 2000: 771 minor fiddles in tran.c to keep compilers happy about uschar. 772 thanks to norman wilson. 773 774May 25, 2000: 775 yet another attempt at making 8-bit input work, with another 776 band-aid in b.c (member()), and some (uschar) casts to head 777 off potential errors in subscripts (like isdigit). also 778 changed HAT to NCHARS-2. thanks again to santiago vila. 779 780 changed maketab.c to ignore apparently out of range definitions 781 instead of halting; new freeBSD generates one. thanks to 782 jon snader <jsnader@ix.netcom.com> for pointing out the problem. 783 784May 2, 2000: 785 fixed an 8-bit problem in b.c by making several char*'s into 786 unsigned char*'s. not clear i have them all yet. thanks to 787 Santiago Vila <sanvila@unex.es> for the bug report. 788 789Apr 21, 2000: 790 finally found and fixed a memory leak in function call; it's 791 been there since functions were added ~1983. thanks to 792 jon bentley for the test case that found it. 793 794 added test in envinit to catch environment "variables" with 795 names beginning with '='; thanks to Berend Hasselman. 796 797Jul 28, 1999: 798 added test in defn() to catch function foo(foo), which 799 otherwise recurses until core dump. thanks to arnold 800 robbins for noticing this. 801 802Jun 20, 1999: 803 added *bp in gettok in lex.c; appears possible to exit function 804 without terminating the string. thanks to russ cox. 805 806Jun 2, 1999: 807 added function stdinit() to run to initialize files[] array, 808 in case stdin, etc., are not constants; some compilers care. 809 810May 10, 1999: 811 replaced the ERROR ... FATAL, etc., macros with functions 812 based on vprintf, to avoid problems caused by overrunning 813 fixed-size errbuf array. thanks to ralph corderoy for the 814 impetus, and for pointing out a string termination bug in 815 qstring as well. 816 817Apr 21, 1999: 818 fixed bug that caused occasional core dumps with commandline 819 variable with value ending in \. (thanks to nelson beebe for 820 the test case.) 821 822Apr 16, 1999: 823 with code kindly provided by Bruce Lilly, awk now parses 824 /=/ and similar constructs more sensibly in more places. 825 Bruce also provided some helpful test cases. 826 827Apr 5, 1999: 828 changed true/false to True/False in run.c to make it 829 easier to compile with C++. Added some casts on malloc 830 and realloc to be honest about casts; ditto. changed 831 ltype int to long in struct rrow to reduce some 64-bit 832 complaints; other changes scattered throughout for the 833 same purpose. thanks to Nelson Beebe for these portability 834 improvements. 835 836 removed some horrible pointer-int casting in b.c and elsewhere 837 by adding ptoi and itonp to localize the casts, which are 838 all benign. fixed one incipient bug that showed up on sgi 839 in 64-bit mode. 840 841 reset lineno for new source file; include filename in error 842 message. also fixed line number error in continuation lines. 843 (thanks to Nelson Beebe for both of these.) 844 845Mar 24, 1999: 846 Nelson Beebe notes that irix 5.3 yacc dies with a bogus 847 error; use a newer version or switch to bison, since sgi 848 is unlikely to fix it. 849 850Mar 5, 1999: 851 changed isnumber to is_number to avoid the problem caused by 852 versions of ctype.h that include the name isnumber. 853 854 distribution now includes a script for building on a Mac, 855 thanks to Dan Allen. 856 857Feb 20, 1999: 858 fixed memory leaks in run.c (call) and tran.c (setfval). 859 thanks to Stephen Nutt for finding these and providing the fixes. 860 861Jan 13, 1999: 862 replaced srand argument by (unsigned int) in run.c; 863 avoids problem on Mac and potentially on Unix & Windows. 864 thanks to Dan Allen. 865 866 added a few (int) casts to silence useless compiler warnings. 867 e.g., errorflag= in run.c jump(). 868 869 added proctab.c to the bundle outout; one less thing 870 to have to compile out of the box. 871 872 added calls to _popen and _pclose to the win95 stub for 873 pipes (thanks to Steve Adams for this helpful suggestion). 874 seems to work, though properties are not well understood 875 by me, and it appears that under some circumstances the 876 pipe output is truncated. Be careful. 877 878Oct 19, 1998: 879 fixed a couple of bugs in getrec: could fail to update $0 880 after a getline var; because inputFS wasn't initialized, 881 could split $0 on every character, a misleading diversion. 882 883 fixed caching bug in makedfa: LRU was actually removing 884 least often used. 885 886 thanks to ross ridge for finding these, and for providing 887 great bug reports. 888 889May 12, 1998: 890 fixed potential bug in readrec: might fail to update record 891 pointer after growing. thanks to dan levy for spotting this 892 and suggesting the fix. 893 894Mar 12, 1998: 895 added -V to print version number and die. 896 897[notify dave kerns, dkerns@dacsoup.ih.lucent.com] 898 899Feb 11, 1998: 900 subtle silent bug in lex.c: if the program ended with a number 901 longer than 1 digit, part of the input would be pushed back and 902 parsed again because token buffer wasn't terminated right. 903 example: awk 'length($0) > 10'. blush. at least i found it 904 myself. 905 906Aug 31, 1997: 907 s/adelete/awkdelete/: SGI uses this in malloc.h. 908 thanks to nelson beebe for pointing this one out. 909 910Aug 21, 1997: 911 fixed some bugs in sub and gsub when replacement includes \\. 912 this is a dark, horrible corner, but at least now i believe that 913 the behavior is the same as gawk and the intended posix standard. 914 thanks to arnold robbins for advice here. 915 916Aug 9, 1997: 917 somewhat regretfully, replaced the ancient lex-based lexical 918 analyzer with one written in C. it's longer, generates less code, 919 and more portable; the old one depended too much on mysterious 920 properties of lex that were not preserved in other environments. 921 in theory these recognize the same language. 922 923 now using strtod to test whether a string is a number, instead of 924 the convoluted original function. should be more portable and 925 reliable if strtod is implemented right. 926 927 removed now-pointless optimization in makefile that tries to avoid 928 recompilation when awkgram.y is changed but symbols are not. 929 930 removed most fixed-size arrays, though a handful remain, some 931 of which are unchecked. you have been warned. 932 933Aug 4, 1997: 934 with some trepidation, replaced the ancient code that managed 935 fields and $0 in fixed-size arrays with arrays that grow on 936 demand. there is still some tension between trying to make this 937 run fast and making it clean; not sure it's right yet. 938 939 the ill-conceived -mr and -mf arguments are now useful only 940 for debugging. previous dynamic string code removed. 941 942 numerous other minor cleanups along the way. 943 944Jul 30, 1997: 945 using code provided by dan levy (to whom profuse thanks), replaced 946 fixed-size arrays and awkward kludges by a fairly uniform mechanism 947 to grow arrays as needed for printf, sub, gsub, etc. 948 949Jul 23, 1997: 950 falling off the end of a function returns "" and 0, not 0. 951 thanks to arnold robbins. 952 953Jun 17, 1997: 954 replaced several fixed-size arrays by dynamically-created ones 955 in run.c; added overflow tests to some previously unchecked cases. 956 getline, toupper, tolower. 957 958 getline code is still broken in that recursive calls may wind 959 up using the same space. [fixed later] 960 961 increased RECSIZE to 8192 to push problems further over the horizon. 962 963 added \r to \n as input line separator for programs, not data. 964 damn CRLFs. 965 966 modified format() to permit explicit printf("%c", 0) to include 967 a null byte in output. thanks to ken stailey for the fix. 968 969 added a "-safe" argument that disables file output (print >, 970 print >>), process creation (cmd|getline, print |, system), and 971 access to the environment (ENVIRON). this is a first approximation 972 to a "safe" version of awk, but don't rely on it too much. thanks 973 to joan feigenbaum and matt blaze for the inspiration long ago. 974 975Jul 8, 1996: 976 fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to 977 ralph corderoy. 978 979Jun 29, 1996: 980 fixed awful bug in new field splitting; didn't get all the places 981 where input was done. 982 983Jun 28, 1996: 984 changed field-splitting to conform to posix definition: fields are 985 split using the value of FS at the time of input; it used to be 986 the value when the field or NF was first referred to, a much less 987 predictable definition. thanks to arnold robbins for encouragement 988 to do the right thing. 989 990May 28, 1996: 991 fixed appalling but apparently unimportant bug in parsing octal 992 numbers in reg exprs. 993 994 explicit hex in reg exprs now limited to 2 chars: \xa, \xaa. 995 996May 27, 1996: 997 cleaned up some declarations so gcc -Wall is now almost silent. 998 999 makefile now includes backup copies of ytab.c and lexyy.c in case 1000 one makes before looking; it also avoids recreating lexyy.c unless 1001 really needed. 1002 1003 s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes 1004 with unwisely-written header files. 1005 1006 thanks to jeffrey friedl for several of these. 1007 1008May 26, 1996: 1009 an attempt to rationalize the (unsigned) char issue. almost all 1010 instances of unsigned char have been removed; the handful of places 1011 in b.c where chars are used as table indices have been hand-crafted. 1012 added some latin-1 tests to the regression, but i'm not confident; 1013 none of my compilers seem to care much. thanks to nelson beebe for 1014 pointing out some others that do care. 1015 1016May 2, 1996: 1017 removed all register declarations. 1018 1019 enhanced split(), as in gawk, etc: split(s, a, "") splits s into 1020 a[1]...a[length(s)] with each character a single element. 1021 1022 made the same changes for field-splitting if FS is "". 1023 1024 added nextfile, as in gawk: causes immediate advance to next 1025 input file. (thanks to arnold robbins for inspiration and code). 1026 1027 small fixes to regexpr code: can now handle []], [[], and 1028 variants; [] is now a syntax error, rather than matching 1029 everything; [z-a] is now empty, not z. far from complete 1030 or correct, however. (thanks to jeffrey friedl for pointing out 1031 some awful behaviors.) 1032 1033Apr 29, 1996: 1034 replaced uchar by uschar everywhere; apparently some compilers 1035 usurp this name and this causes conflicts. 1036 1037 fixed call to time in run.c (bltin); arg is time_t *. 1038 1039 replaced horrible pointer/long punning in b.c by a legitimate 1040 union. should be safer on 64-bit machines and cleaner everywhere. 1041 (thanks to nelson beebe for pointing out some of these problems.) 1042 1043 replaced nested comments by #if 0...#endif in run.c, lib.c. 1044 1045 removed getsval, setsval, execute macros from run.c and lib.c. 1046 machines are 100x faster than they were when these macros were 1047 first used. 1048 1049 revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l, 1050 y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of 1051 portability to nameless systems. 1052 1053 "make bundle" now includes yacc and lex output files for recipients 1054 who don't have yacc or lex. 1055 1056Aug 15, 1995: 1057 initialized Cells in setsymtab more carefully; some fields 1058 were not set. (thanks to purify, all of whose complaints i 1059 think i now understand.) 1060 1061 fixed at least one error in gsub that looked at -1-th element 1062 of an array when substituting for a null match (e.g., $). 1063 1064 delete arrayname is now legal; it clears the elements but leaves 1065 the array, which may not be the right behavior. 1066 1067 modified makefile: my current make can't cope with the test used 1068 to avoid unnecessary yacc invocations. 1069 1070Jul 17, 1995: 1071 added dynamically growing strings to awk.lx.l and b.c 1072 to permit regular expressions to be much bigger. 1073 the state arrays can still overflow. 1074 1075Aug 24, 1994: 1076 detect duplicate arguments in function definitions (mdm). 1077 1078May 11, 1994: 1079 trivial fix to printf to limit string size in sub(). 1080 1081Apr 22, 1994: 1082 fixed yet another subtle self-assignment problem: 1083 $1 = $2; $1 = $1 clobbered $1. 1084 1085 Regression tests now use private echo, to avoid quoting problems. 1086 1087Feb 2, 1994: 1088 changed error() to print line number as %d, not %g. 1089 1090Jul 23, 1993: 1091 cosmetic changes: increased sizes of some arrays, 1092 reworded some error messages. 1093 1094 added CONVFMT as in posix (just replaced OFMT in getsval) 1095 1096 FILENAME is now "" until the first thing that causes a file 1097 to be opened. 1098 1099Nov 28, 1992: 1100 deleted yyunput and yyoutput from proto.h; 1101 different versions of lex give these different declarations. 1102 1103May 31, 1992: 1104 added -mr N and -mf N options: more record and fields. 1105 these really ought to adjust automatically. 1106 1107 cleaned up some error messages; "out of space" now means 1108 malloc returned NULL in all cases. 1109 1110 changed rehash so that if it runs out, it just returns; 1111 things will continue to run slow, but maybe a bit longer. 1112 1113Apr 24, 1992: 1114 remove redundant close of stdin when using -f -. 1115 1116 got rid of core dump with -d; awk -d just prints date. 1117 1118Apr 12, 1992: 1119 added explicit check for /dev/std(in,out,err) in redirection. 1120 unlike gawk, no /dev/fd/n yet. 1121 1122 added (file/pipe) builtin. hard to test satisfactorily. 1123 not posix. 1124 1125Feb 20, 1992: 1126 recompile after abortive changes; should be unchanged. 1127 1128Dec 2, 1991: 1129 die-casting time: converted to ansi C, installed that. 1130 1131Nov 30, 1991: 1132 fixed storage leak in freefa, failing to recover [N]CCL. 1133 thanks to Bill Jones (jones@cs.usask.ca) 1134 1135Nov 19, 1991: 1136 use RAND_MAX instead of literal in builtin(). 1137 1138Nov 12, 1991: 1139 cranked up some fixed-size arrays in b.c, and added a test for 1140 overflow in penter. thanks to mark larsen. 1141 1142Sep 24, 1991: 1143 increased buffer in gsub. a very crude fix to a general problem. 1144 and again on Sep 26. 1145 1146Aug 18, 1991: 1147 enforce variable name syntax for commandline variables: has to 1148 start with letter or _. 1149 1150Jul 27, 1991: 1151 allow newline after ; in for statements. 1152 1153Jul 21, 1991: 1154 fixed so that in self-assignment like $1=$1, side effects 1155 like recomputing $0 take place. (this is getting subtle.) 1156 1157Jun 30, 1991: 1158 better test for detecting too-long output record. 1159 1160Jun 2, 1991: 1161 better defense against very long printf strings. 1162 made break and continue illegal outside of loops. 1163 1164May 13, 1991: 1165 removed extra arg on gettemp, tempfree. minor error message rewording. 1166 1167May 6, 1991: 1168 fixed silly bug in hex parsing in hexstr(). 1169 removed an apparently unnecessary test in isnumber(). 1170 warn about weird printf conversions. 1171 fixed unchecked array overwrite in relex(). 1172 1173 changed for (i in array) to access elements in sorted order. 1174 then unchanged it -- it really does run slower in too many cases. 1175 left the code in place, commented out. 1176 1177Feb 10, 1991: 1178 check error status on all writes, to avoid banging on full disks. 1179 1180Jan 28, 1991: 1181 awk -f - reads the program from stdin. 1182 1183Jan 11, 1991: 1184 failed to set numeric state on $0 in cmd|getline context in run.c. 1185 1186Nov 2, 1990: 1187 fixed sleazy test for integrality in getsval; use modf. 1188 1189Oct 29, 1990: 1190 fixed sleazy buggy code in lib.c that looked (incorrectly) for 1191 too long input lines. 1192 1193Oct 14, 1990: 1194 fixed the bug on p. 198 in which it couldn't deduce that an 1195 argument was an array in some contexts. replaced the error 1196 message in intest() by code that damn well makes it an array. 1197 1198Oct 8, 1990: 1199 fixed horrible bug: types and values were not preserved in 1200 some kinds of self-assignment. (in assign().) 1201 1202Aug 24, 1990: 1203 changed NCHARS to 256 to handle 8-bit characters in strings 1204 presented to match(), etc. 1205 1206Jun 26, 1990: 1207 changed struct rrow (awk.h) to use long instead of int for lval, 1208 since cfoll() stores a pointer in it. now works better when int's 1209 are smaller than pointers! 1210 1211May 6, 1990: 1212 AVA fixed the grammar so that ! is uniformly of the same precedence as 1213 unary + and -. This renders illegal some constructs like !x=y, which 1214 now has to be parenthesized as !(x=y), and makes others work properly: 1215 !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements. 1216 (These problems were pointed out by Bob Lenk of Posix.) 1217 1218 Added \x to regular expressions (already in strings). 1219 Limited octal to octal digits; \8 and \9 are not octal. 1220 Centralized the code for parsing escapes in regular expressions. 1221 Added a bunch of tests to T.re and T.sub to verify some of this. 1222 1223Feb 9, 1990: 1224 fixed null pointer dereference bug in main.c: -F[nothing]. sigh. 1225 1226 restored srand behavior: it returns the current seed. 1227 1228Jan 18, 1990: 1229 srand now returns previous seed value (0 to start). 1230 1231Jan 5, 1990: 1232 fix potential problem in tran.c -- something was freed, 1233 then used in freesymtab. 1234 1235Oct 18, 1989: 1236 another try to get the max number of open files set with 1237 relatively machine-independent code. 1238 1239 small fix to input() in case of multiple reads after EOF. 1240 1241Oct 11, 1989: 1242 FILENAME is now defined in the BEGIN block -- too many old 1243 programs broke. 1244 1245 "-" means stdin in getline as well as on the commandline. 1246 1247 added a bunch of casts to the code to tell the truth about 1248 char * vs. unsigned char *, a right royal pain. added a 1249 setlocale call to the front of main, though probably no one 1250 has it usefully implemented yet. 1251 1252Aug 24, 1989: 1253 removed redundant relational tests against nullnode if parse 1254 tree already had a relational at that point. 1255 1256Aug 11, 1989: 1257 fixed bug: commandline variable assignment has to look like 1258 var=something. (consider the man page for =, in file =.1) 1259 1260 changed number of arguments to functions to static arrays 1261 to avoid repeated malloc calls. 1262 1263Aug 2, 1989: 1264 restored -F (space) separator 1265 1266Jul 30, 1989: 1267 added -v x=1 y=2 ... for immediate commandline variable assignment; 1268 done before the BEGIN block for sure. they have to precede the 1269 program if the program is on the commandline. 1270 Modified Aug 2 to require a separate -v for each assignment. 1271 1272Jul 10, 1989: 1273 fixed ref-thru-zero bug in environment code in tran.c 1274 1275Jun 23, 1989: 1276 add newline to usage message. 1277 1278Jun 14, 1989: 1279 added some missing ansi printf conversion letters: %i %X %E %G. 1280 no sensible meaning for h or L, so they may not do what one expects. 1281 1282 made %* conversions work. 1283 1284 changed x^y so that if n is a positive integer, it's done 1285 by explicit multiplication, thus achieving maximum accuracy. 1286 (this should be done by pow() but it seems not to be locally.) 1287 done to x ^= y as well. 1288 1289Jun 4, 1989: 1290 ENVIRON array contains environment: if shell variable V=thing, 1291 ENVIRON["V"] is "thing" 1292 1293 multiple -f arguments permitted. error reporting is naive. 1294 (they were permitted before, but only the last was used.) 1295 1296 fixed a really stupid botch in the debugging macro dprintf 1297 1298 fixed order of evaluation of commandline assignments to match 1299 what the book claims: an argument of the form x=e is evaluated 1300 at the time it would have been opened if it were a filename (p 63). 1301 this invalidates the suggested answer to ex 4-1 (p 195). 1302 1303 removed some code that permitted -F (space) fieldseparator, 1304 since it didn't quite work right anyway. (restored aug 2) 1305 1306Apr 27, 1989: 1307 Line number now accumulated correctly for comment lines. 1308 1309Apr 26, 1989: 1310 Debugging output now includes a version date, 1311 if one compiles it into the source each time. 1312 1313Apr 9, 1989: 1314 Changed grammar to prohibit constants as 3rd arg of sub and gsub; 1315 prevents class of overwriting-a-constant errors. (Last one?) 1316 This invalidates the "banana" example on page 43 of the book. 1317 1318 Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal), 1319 as in ANSI, for strings. Rescinded the sloppiness that permitted 1320 non-octal digits in \ooo. Warning: not all compilers and libraries 1321 will be able to deal with \x correctly. 1322 1323Jan 9, 1989: 1324 Fixed bug that caused tempcell list to contain a duplicate. 1325 The fix is kludgy. 1326 1327Dec 17, 1988: 1328 Catches some more commandline errors in main. 1329 Removed redundant decl of modf in run.c (confuses some compilers). 1330 Warning: there's no single declaration of malloc, etc., in awk.h 1331 that seems to satisfy all compilers. 1332 1333Dec 7, 1988: 1334 Added a bit of code to error printing to avoid printing nulls. 1335 (Not clear that it actually would.) 1336 1337Nov 27, 1988: 1338 With fear and trembling, modified the grammar to permit 1339 multiple pattern-action statements on one line without 1340 an explicit separator. By definition, this capitulation 1341 to the ghost of ancient implementations remains undefined 1342 and thus subject to change without notice or apology. 1343 DO NOT COUNT ON IT. 1344 1345Oct 30, 1988: 1346 Fixed bug in call() that failed to recover storage. 1347 1348 A warning is now generated if there are more arguments 1349 in the call than in the definition (in lieu of fixing 1350 another storage leak). 1351 1352Oct 20, 1988: 1353 Fixed %c: if expr is numeric, use numeric value; 1354 otherwise print 1st char of string value. still 1355 doesn't work if the value is 0 -- won't print \0. 1356 1357 Added a few more checks for running out of malloc. 1358 1359Oct 12, 1988: 1360 Fixed bug in call() that freed local arrays twice. 1361 1362 Fixed to handle deletion of non-existent array right; 1363 complains about attempt to delete non-array element. 1364 1365Sep 30, 1988: 1366 Now guarantees to evaluate all arguments of built-in 1367 functions, as in C; the appearance is that arguments 1368 are evaluated before the function is called. Places 1369 affected are sub (gsub was ok), substr, printf, and 1370 all the built-in arithmetic functions in bltin(). 1371 A warning is generated if a bltin() is called with 1372 the wrong number of arguments. 1373 1374 This requires changing makeprof on p167 of the book. 1375 1376Aug 23, 1988: 1377 setting FILENAME in BEGIN caused core dump, apparently 1378 because it was freeing space not allocated by malloc. 1379 1380July 24, 1988: 1381 fixed egregious error in toupper/tolower functions. 1382 still subject to rescinding, however. 1383 1384July 2, 1988: 1385 flush stdout before opening file or pipe 1386 1387July 2, 1988: 1388 performance bug in b.c/cgoto(): not freeing some sets of states. 1389 partial fix only right now, and the number of states increased 1390 to make it less obvious. 1391 1392June 1, 1988: 1393 check error status on close 1394 1395May 28, 1988: 1396 srand returns seed value it's using. 1397 see 1/18/90 1398 1399May 22, 1988: 1400 Removed limit on depth of function calls. 1401 1402May 10, 1988: 1403 Fixed lib.c to permit _ in commandline variable names. 1404 1405Mar 25, 1988: 1406 main.c fixed to recognize -- as terminator of command- 1407 line options. Illegal options flagged. 1408 Error reporting slightly cleaned up. 1409 1410Dec 2, 1987: 1411 Newer C compilers apply a strict scope rule to extern 1412 declarations within functions. Two extern declarations in 1413 lib.c and tran.c have been moved to obviate this problem. 1414 1415Oct xx, 1987: 1416 Reluctantly added toupper and tolower functions. 1417 Subject to rescinding without notice. 1418 1419Sep 17, 1987: 1420 Error-message printer had printf(s) instead of 1421 printf("%s",s); got core dumps when the message 1422 included a %. 1423 1424Sep 12, 1987: 1425 Very long printf strings caused core dump; 1426 fixed aprintf, asprintf, format to catch them. 1427 Can still get a core dump in printf itself. 1428 1429 1430