FIXES revision 285830
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 28Dec 20, 2012: 29 fiddled makefile to get correct yacc and bison flags. pick yacc 30 (linux) or bison (mac) as necessary. 31 32 added __attribute__((__noreturn__)) to a couple of lines in 33 proto.h, to silence someone's enthusiastic checker. 34 35 fixed obscure call by value bug in split(a[1],a) reported on 36 9fans. the management of temporary values is just a mess; i 37 took a shortcut by making an extra string copy. thanks 38 to paul patience and arnold robbins for passing it on and for 39 proposed patches. 40 41 tiny fiddle in setfval to eliminate -0 results in T.expr, which 42 has irritated me for 20+ years. 43 44Aug 10, 2011: 45 another fix to avoid core dump with delete(ARGV); again, many thanks 46 to ruslan ermilov. 47 48Aug 7, 2011: 49 split(s, a, //) now behaves the same as split(s, a, "") 50 51Jun 12, 2011: 52 /pat/, \n /pat/ {...} is now legal, though bad style to use. 53 54 added checks to new -v code that permits -vnospace; thanks to 55 ruslan ermilov for spotting this and providing the patch. 56 57 removed fixed limit on number of open files; thanks to aleksey 58 cheusov and christos zoulos. 59 60 fixed day 1 bug that resurrected deleted elements of ARGV when 61 used as filenames (in lib.c). 62 63 minor type fiddles to make gcc -Wall -pedantic happier (but not 64 totally so); turned on -fno-strict-aliasing in makefile. 65 66May 6, 2011: 67 added #ifdef for isblank. 68 now allows -ffoo as well as -f foo arguments. 69 (thanks, ruslan) 70 71May 1, 2011: 72 after advice from todd miller, kevin lo, ruslan ermilov, 73 and arnold robbins, changed srand() to return the previous 74 seed (which is 1 on the first call of srand). the seed is 75 an Awkfloat internally though converted to unsigned int to 76 pass to the library srand(). thanks, everyone. 77 78 fixed a subtle (and i hope low-probability) overflow error 79 in fldbld, by adding space for one extra \0. thanks to 80 robert bassett for spotting this one and providing a fix. 81 82 removed the files related to compilation on windows. i no 83 longer have anything like a current windows environment, so 84 i can't test any of it. 85 86May 23, 2010: 87 fixed long-standing overflow bug in run.c; many thanks to 88 nelson beebe for spotting it and providing the fix. 89 90 fixed bug that didn't parse -vd=1 properly; thanks to santiago 91 vila for spotting it. 92 93Feb 8, 2010: 94 i give up. replaced isblank with isspace in b.c; there are 95 no consistent header files. 96 97Nov 26, 2009: 98 fixed a long-standing issue with when FS takes effect. a 99 change to FS is now noticed immediately for subsequent splits. 100 101 changed the name getline() to awkgetline() to avoid yet another 102 name conflict somewhere. 103 104Feb 11, 2009: 105 temporarily for now defined HAS_ISBLANK, since that seems to 106 be the best way through the thicket. isblank arrived in C99, 107 but seems to be arriving at different systems at different 108 times. 109 110Oct 8, 2008: 111 fixed typo in b.c that set tmpvec wrongly. no one had ever 112 run into the problem, apparently. thanks to alistair crooks. 113 114Oct 23, 2007: 115 minor fix in lib.c: increase inputFS to 100, change malloc 116 for fields to n+1. 117 118 fixed memory fault caused by out of order test in setsval. 119 120 thanks to david o'brien, freebsd, for both fixes. 121 122May 1, 2007: 123 fiddle in makefile to fix for BSD make; thanks to igor sobrado. 124 125Mar 31, 2007: 126 fixed some null pointer refs calling adjbuf. 127 128Feb 21, 2007: 129 fixed a bug in matching the null RE in sub and gsub. thanks to al aho 130 who actually did the fix (in b.c), and to wolfgang seeberg for finding 131 it and providing a very compact test case. 132 133 fixed quotation in b.c; thanks to Hal Pratt and the Princeton Dante 134 Project. 135 136 removed some no-effect asserts in run.c. 137 138 fiddled maketab.c to not complain about bison-generated values. 139 140 removed the obsolete -V argument; fixed --version to print the 141 version and exit. 142 143 fixed wording and an outright error in the usage message; thanks to igor 144 sobrado and jason mcintyre. 145 146 fixed a bug in -d that caused core dump if no program followed. 147 148Jan 1, 2007: 149 dropped mac.code from makefile; there are few non-MacOSX 150 mac's these days. 151 152Jan 17, 2006: 153 system() not flagged as unsafe in the unadvertised -safe option. 154 found it while enhancing tests before shipping the ;login: article. 155 practice what you preach. 156 157 removed the 9-years-obsolete -mr and -mf flags. 158 159 added -version and --version options. 160 161 core dump on linux with BEGIN {nextfile}, now fixed. 162 163 removed some #ifdef's in run.c and lex.c that appear to no 164 longer be necessary. 165 166Apr 24, 2005: 167 modified lib.c so that values of $0 et al are preserved in the END 168 block, apparently as required by posix. thanks to havard eidnes 169 for the report and code. 170 171Jan 14, 2005: 172 fixed infinite loop in parsing, originally found by brian tsang. 173 thanks to arnold robbins for a suggestion that started me 174 rethinking it. 175 176Dec 31, 2004: 177 prevent overflow of -f array in main, head off potential error in 178 call of SYNTAX(), test malloc return in lib.c, all with thanks to 179 todd miller. 180 181Dec 22, 2004: 182 cranked up size of NCHARS; coverity thinks it can be overrun with 183 smaller size, and i think that's right. added some assertions to b.c 184 to catch places where it might overrun. the RE code is still fragile. 185 186Dec 5, 2004: 187 fixed a couple of overflow problems with ridiculous field numbers: 188 e.g., print $(2^32-1). thanks to ruslan ermilov, giorgos keramidas 189 and david o'brien at freebsd.org for patches. this really should 190 be re-done from scratch. 191 192Nov 21, 2004: 193 fixed another 25-year-old RE bug, in split. it's another failure 194 to (re-)initialize. thanks to steve fisher for spotting this and 195 providing a good test case. 196 197Nov 22, 2003: 198 fixed a bug in regular expressions that dates (so help me) from 1977; 199 it's been there from the beginning. an anchored longest match that 200 was longer than the number of states triggered a failure to initialize 201 the machine properly. many thanks to moinak ghosh for not only finding 202 this one but for providing a fix, in some of the most mysterious 203 code known to man. 204 205 fixed a storage leak in call() that appears to have been there since 206 1983 or so -- a function without an explicit return that assigns a 207 string to a parameter leaked a Cell. thanks to moinak ghosh for 208 spotting this very subtle one. 209 210Jul 31, 2003: 211 fixed, thanks to andrey chernov and ruslan ermilov, a bug in lex.c 212 that mis-handled the character 255 in input. (it was being compared 213 to EOF with a signed comparison.) 214 215Jul 29, 2003: 216 fixed (i think) the long-standing botch that included the beginning of 217 line state ^ for RE's in the set of valid characters; this led to a 218 variety of odd problems, including failure to properly match certain 219 regular expressions in non-US locales. thanks to ruslan for keeping 220 at this one. 221 222Jul 28, 2003: 223 n-th try at getting internationalization right, with thanks to volker 224 kiefel, arnold robbins and ruslan ermilov for advice, though they 225 should not be blamed for the outcome. according to posix, "." is the 226 radix character in programs and command line arguments regardless of 227 the locale; otherwise, the locale should prevail for input and output 228 of numbers. so it's intended to work that way. 229 230 i have rescinded the attempt to use strcoll in expanding shorthands in 231 regular expressions (cclenter). its properties are much too 232 surprising; for example [a-c] matches aAbBc in locale en_US but abBcC 233 in locale fr_CA. i can see how this might arise by implementation 234 but i cannot explain it to a human user. (this behavior can be seen 235 in gawk as well; we're leaning on the same library.) 236 237 the issue appears to be that strcoll is meant for sorting, where 238 merging upper and lower case may make sense (though note that unix 239 sort does not do this by default either). it is not appropriate 240 for regular expressions, where the goal is to match specific 241 patterns of characters. in any case, the notations [:lower:], etc., 242 are available in awk, and they are more likely to work correctly in 243 most locales. 244 245 a moratorium is hereby declared on internationalization changes. 246 i apologize to friends and colleagues in other parts of the world. 247 i would truly like to get this "right", but i don't know what 248 that is, and i do not want to keep making changes until it's clear. 249 250Jul 4, 2003: 251 fixed bug that permitted non-terminated RE, as in "awk /x". 252 253Jun 1, 2003: 254 subtle change to split: if source is empty, number of elems 255 is always 0 and the array is not set. 256 257Mar 21, 2003: 258 added some parens to isblank, in another attempt to make things 259 internationally portable. 260 261Mar 14, 2003: 262 the internationalization changes, somewhat modified, are now 263 reinstated. in theory awk will now do character comparisons 264 and case conversions in national language, but "." will always 265 be the decimal point separator on input and output regardless 266 of national language. isblank(){} has an #ifndef. 267 268 this no longer compiles on windows: LC_MESSAGES isn't defined 269 in vc6++. 270 271 fixed subtle behavior in field and record splitting: if FS is 272 a single character and RS is not empty, \n is NOT a separator. 273 this tortuous reading is found in the awk book; behavior now 274 matches gawk and mawk. 275 276Dec 13, 2002: 277 for the moment, the internationalization changes of nov 29 are 278 rolled back -- programs like x = 1.2 don't work in some locales, 279 because the parser is expecting x = 1,2. until i understand this 280 better, this will have to wait. 281 282Nov 29, 2002: 283 modified b.c (with tiny changes in main and run) to support 284 locales, using strcoll and iswhatever tests for posix character 285 classes. thanks to ruslan ermilov (ru@freebsd.org) for code. 286 the function isblank doesn't seem to have propagated to any 287 header file near me, so it's there explicitly. not properly 288 tested on non-ascii character sets by me. 289 290Jun 28, 2002: 291 modified run/format() and tran/getsval() to do a slightly better 292 job on using OFMT for output from print and CONVFMT for other 293 number->string conversions, as promised by posix and done by 294 gawk and mawk. there are still places where it doesn't work 295 right if CONVFMT is changed; by then the STR attribute of the 296 variable has been irrevocably set. thanks to arnold robbins for 297 code and examples. 298 299 fixed subtle bug in format that could get core dump. thanks to 300 Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing. 301 minor cleanup in run.c / format() at the same time. 302 303 added some tests for null pointers to debugging printf's, which 304 were never intended for external consumption. thanks to dave 305 kerns (dkerns@lucent.com) for pointing this out. 306 307 GNU compatibility: an empty regexp matches anything (thanks to 308 dag-erling smorgrav, des@ofug.org). subject to reversion if 309 this does more harm than good. 310 311 pervasive small changes to make things more const-correct, as 312 reported by gcc's -Wwrite-strings. as it says in the gcc manual, 313 this may be more nuisance than useful. provoked by a suggestion 314 and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk 315 316 minor documentation changes to note that this now compiles out 317 of the box on Mac OS X. 318 319Feb 10, 2002: 320 changed types in posix chars structure to quiet solaris cc. 321 322Jan 1, 2002: 323 fflush() or fflush("") flushes all files and pipes. 324 325 length(arrayname) returns number of elements; thanks to 326 arnold robbins for suggestion. 327 328 added a makefile.win to make it easier to build on windows. 329 based on dan allen's buildwin.bat. 330 331Nov 16, 2001: 332 added support for posix character class names like [:digit:], 333 which are not exactly shorter than [0-9] and perhaps no more 334 portable. thanks to dag-erling smorgrav for code. 335 336Feb 16, 2001: 337 removed -m option; no longer needed, and it was actually 338 broken (noted thanks to volker kiefel). 339 340Feb 10, 2001: 341 fixed an appalling bug in gettok: any sequence of digits, +,-, E, e, 342 and period was accepted as a valid number if it started with a period. 343 this would never have happened with the lex version. 344 345 other 1-character botches, now fixed, include a bare $ and a 346 bare " at the end of the input. 347 348Feb 7, 2001: 349 more (const char *) casts in b.c and tran.c to silence warnings. 350 351Nov 15, 2000: 352 fixed a bug introduced in august 1997 that caused expressions 353 like $f[1] to be syntax errors. thanks to arnold robbins for 354 noticing this and providing a fix. 355 356Oct 30, 2000: 357 fixed some nextfile bugs: not handling all cases. thanks to 358 arnold robbins for pointing this out. new regressions added. 359 360 close() is now a function. it returns whatever the library 361 fclose returns, and -1 for closing a file or pipe that wasn't 362 opened. 363 364Sep 24, 2000: 365 permit \n explicitly in character classes; won't work right 366 if comes in as "[\n]" but ok as /[\n]/, because of multiple 367 processing of \'s. thanks to arnold robbins. 368 369July 5, 2000: 370 minor fiddles in tran.c to keep compilers happy about uschar. 371 thanks to norman wilson. 372 373May 25, 2000: 374 yet another attempt at making 8-bit input work, with another 375 band-aid in b.c (member()), and some (uschar) casts to head 376 off potential errors in subscripts (like isdigit). also 377 changed HAT to NCHARS-2. thanks again to santiago vila. 378 379 changed maketab.c to ignore apparently out of range definitions 380 instead of halting; new freeBSD generates one. thanks to 381 jon snader <jsnader@ix.netcom.com> for pointing out the problem. 382 383May 2, 2000: 384 fixed an 8-bit problem in b.c by making several char*'s into 385 unsigned char*'s. not clear i have them all yet. thanks to 386 Santiago Vila <sanvila@unex.es> for the bug report. 387 388Apr 21, 2000: 389 finally found and fixed a memory leak in function call; it's 390 been there since functions were added ~1983. thanks to 391 jon bentley for the test case that found it. 392 393 added test in envinit to catch environment "variables" with 394 names beginning with '='; thanks to Berend Hasselman. 395 396Jul 28, 1999: 397 added test in defn() to catch function foo(foo), which 398 otherwise recurses until core dump. thanks to arnold 399 robbins for noticing this. 400 401Jun 20, 1999: 402 added *bp in gettok in lex.c; appears possible to exit function 403 without terminating the string. thanks to russ cox. 404 405Jun 2, 1999: 406 added function stdinit() to run to initialize files[] array, 407 in case stdin, etc., are not constants; some compilers care. 408 409May 10, 1999: 410 replaced the ERROR ... FATAL, etc., macros with functions 411 based on vprintf, to avoid problems caused by overrunning 412 fixed-size errbuf array. thanks to ralph corderoy for the 413 impetus, and for pointing out a string termination bug in 414 qstring as well. 415 416Apr 21, 1999: 417 fixed bug that caused occasional core dumps with commandline 418 variable with value ending in \. (thanks to nelson beebe for 419 the test case.) 420 421Apr 16, 1999: 422 with code kindly provided by Bruce Lilly, awk now parses 423 /=/ and similar constructs more sensibly in more places. 424 Bruce also provided some helpful test cases. 425 426Apr 5, 1999: 427 changed true/false to True/False in run.c to make it 428 easier to compile with C++. Added some casts on malloc 429 and realloc to be honest about casts; ditto. changed 430 ltype int to long in struct rrow to reduce some 64-bit 431 complaints; other changes scattered throughout for the 432 same purpose. thanks to Nelson Beebe for these portability 433 improvements. 434 435 removed some horrible pointer-int casting in b.c and elsewhere 436 by adding ptoi and itonp to localize the casts, which are 437 all benign. fixed one incipient bug that showed up on sgi 438 in 64-bit mode. 439 440 reset lineno for new source file; include filename in error 441 message. also fixed line number error in continuation lines. 442 (thanks to Nelson Beebe for both of these.) 443 444Mar 24, 1999: 445 Nelson Beebe notes that irix 5.3 yacc dies with a bogus 446 error; use a newer version or switch to bison, since sgi 447 is unlikely to fix it. 448 449Mar 5, 1999: 450 changed isnumber to is_number to avoid the problem caused by 451 versions of ctype.h that include the name isnumber. 452 453 distribution now includes a script for building on a Mac, 454 thanks to Dan Allen. 455 456Feb 20, 1999: 457 fixed memory leaks in run.c (call) and tran.c (setfval). 458 thanks to Stephen Nutt for finding these and providing the fixes. 459 460Jan 13, 1999: 461 replaced srand argument by (unsigned int) in run.c; 462 avoids problem on Mac and potentially on Unix & Windows. 463 thanks to Dan Allen. 464 465 added a few (int) casts to silence useless compiler warnings. 466 e.g., errorflag= in run.c jump(). 467 468 added proctab.c to the bundle outout; one less thing 469 to have to compile out of the box. 470 471 added calls to _popen and _pclose to the win95 stub for 472 pipes (thanks to Steve Adams for this helpful suggestion). 473 seems to work, though properties are not well understood 474 by me, and it appears that under some circumstances the 475 pipe output is truncated. Be careful. 476 477Oct 19, 1998: 478 fixed a couple of bugs in getrec: could fail to update $0 479 after a getline var; because inputFS wasn't initialized, 480 could split $0 on every character, a misleading diversion. 481 482 fixed caching bug in makedfa: LRU was actually removing 483 least often used. 484 485 thanks to ross ridge for finding these, and for providing 486 great bug reports. 487 488May 12, 1998: 489 fixed potential bug in readrec: might fail to update record 490 pointer after growing. thanks to dan levy for spotting this 491 and suggesting the fix. 492 493Mar 12, 1998: 494 added -V to print version number and die. 495 496Feb 11, 1998: 497 subtle silent bug in lex.c: if the program ended with a number 498 longer than 1 digit, part of the input would be pushed back and 499 parsed again because token buffer wasn't terminated right. 500 example: awk 'length($0) > 10'. blush. at least i found it 501 myself. 502 503Aug 31, 1997: 504 s/adelete/awkdelete/: SGI uses this in malloc.h. 505 thanks to nelson beebe for pointing this one out. 506 507Aug 21, 1997: 508 fixed some bugs in sub and gsub when replacement includes \\. 509 this is a dark, horrible corner, but at least now i believe that 510 the behavior is the same as gawk and the intended posix standard. 511 thanks to arnold robbins for advice here. 512 513Aug 9, 1997: 514 somewhat regretfully, replaced the ancient lex-based lexical 515 analyzer with one written in C. it's longer, generates less code, 516 and more portable; the old one depended too much on mysterious 517 properties of lex that were not preserved in other environments. 518 in theory these recognize the same language. 519 520 now using strtod to test whether a string is a number, instead of 521 the convoluted original function. should be more portable and 522 reliable if strtod is implemented right. 523 524 removed now-pointless optimization in makefile that tries to avoid 525 recompilation when awkgram.y is changed but symbols are not. 526 527 removed most fixed-size arrays, though a handful remain, some 528 of which are unchecked. you have been warned. 529 530Aug 4, 1997: 531 with some trepidation, replaced the ancient code that managed 532 fields and $0 in fixed-size arrays with arrays that grow on 533 demand. there is still some tension between trying to make this 534 run fast and making it clean; not sure it's right yet. 535 536 the ill-conceived -mr and -mf arguments are now useful only 537 for debugging. previous dynamic string code removed. 538 539 numerous other minor cleanups along the way. 540 541Jul 30, 1997: 542 using code provided by dan levy (to whom profuse thanks), replaced 543 fixed-size arrays and awkward kludges by a fairly uniform mechanism 544 to grow arrays as needed for printf, sub, gsub, etc. 545 546Jul 23, 1997: 547 falling off the end of a function returns "" and 0, not 0. 548 thanks to arnold robbins. 549 550Jun 17, 1997: 551 replaced several fixed-size arrays by dynamically-created ones 552 in run.c; added overflow tests to some previously unchecked cases. 553 getline, toupper, tolower. 554 555 getline code is still broken in that recursive calls may wind 556 up using the same space. [fixed later] 557 558 increased RECSIZE to 8192 to push problems further over the horizon. 559 560 added \r to \n as input line separator for programs, not data. 561 damn CRLFs. 562 563 modified format() to permit explicit printf("%c", 0) to include 564 a null byte in output. thanks to ken stailey for the fix. 565 566 added a "-safe" argument that disables file output (print >, 567 print >>), process creation (cmd|getline, print |, system), and 568 access to the environment (ENVIRON). this is a first approximation 569 to a "safe" version of awk, but don't rely on it too much. thanks 570 to joan feigenbaum and matt blaze for the inspiration long ago. 571 572Jul 8, 1996: 573 fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to 574 ralph corderoy. 575 576Jun 29, 1996: 577 fixed awful bug in new field splitting; didn't get all the places 578 where input was done. 579 580Jun 28, 1996: 581 changed field-splitting to conform to posix definition: fields are 582 split using the value of FS at the time of input; it used to be 583 the value when the field or NF was first referred to, a much less 584 predictable definition. thanks to arnold robbins for encouragement 585 to do the right thing. 586 587May 28, 1996: 588 fixed appalling but apparently unimportant bug in parsing octal 589 numbers in reg exprs. 590 591 explicit hex in reg exprs now limited to 2 chars: \xa, \xaa. 592 593May 27, 1996: 594 cleaned up some declarations so gcc -Wall is now almost silent. 595 596 makefile now includes backup copies of ytab.c and lexyy.c in case 597 one makes before looking; it also avoids recreating lexyy.c unless 598 really needed. 599 600 s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes 601 with unwisely-written header files. 602 603 thanks to jeffrey friedl for several of these. 604 605May 26, 1996: 606 an attempt to rationalize the (unsigned) char issue. almost all 607 instances of unsigned char have been removed; the handful of places 608 in b.c where chars are used as table indices have been hand-crafted. 609 added some latin-1 tests to the regression, but i'm not confident; 610 none of my compilers seem to care much. thanks to nelson beebe for 611 pointing out some others that do care. 612 613May 2, 1996: 614 removed all register declarations. 615 616 enhanced split(), as in gawk, etc: split(s, a, "") splits s into 617 a[1]...a[length(s)] with each character a single element. 618 619 made the same changes for field-splitting if FS is "". 620 621 added nextfile, as in gawk: causes immediate advance to next 622 input file. (thanks to arnold robbins for inspiration and code). 623 624 small fixes to regexpr code: can now handle []], [[], and 625 variants; [] is now a syntax error, rather than matching 626 everything; [z-a] is now empty, not z. far from complete 627 or correct, however. (thanks to jeffrey friedl for pointing out 628 some awful behaviors.) 629 630Apr 29, 1996: 631 replaced uchar by uschar everywhere; apparently some compilers 632 usurp this name and this causes conflicts. 633 634 fixed call to time in run.c (bltin); arg is time_t *. 635 636 replaced horrible pointer/long punning in b.c by a legitimate 637 union. should be safer on 64-bit machines and cleaner everywhere. 638 (thanks to nelson beebe for pointing out some of these problems.) 639 640 replaced nested comments by #if 0...#endif in run.c, lib.c. 641 642 removed getsval, setsval, execute macros from run.c and lib.c. 643 machines are 100x faster than they were when these macros were 644 first used. 645 646 revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l, 647 y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of 648 portability to nameless systems. 649 650 "make bundle" now includes yacc and lex output files for recipients 651 who don't have yacc or lex. 652 653Aug 15, 1995: 654 initialized Cells in setsymtab more carefully; some fields 655 were not set. (thanks to purify, all of whose complaints i 656 think i now understand.) 657 658 fixed at least one error in gsub that looked at -1-th element 659 of an array when substituting for a null match (e.g., $). 660 661 delete arrayname is now legal; it clears the elements but leaves 662 the array, which may not be the right behavior. 663 664 modified makefile: my current make can't cope with the test used 665 to avoid unnecessary yacc invocations. 666 667Jul 17, 1995: 668 added dynamically growing strings to awk.lx.l and b.c 669 to permit regular expressions to be much bigger. 670 the state arrays can still overflow. 671 672Aug 24, 1994: 673 detect duplicate arguments in function definitions (mdm). 674 675May 11, 1994: 676 trivial fix to printf to limit string size in sub(). 677 678Apr 22, 1994: 679 fixed yet another subtle self-assignment problem: 680 $1 = $2; $1 = $1 clobbered $1. 681 682 Regression tests now use private echo, to avoid quoting problems. 683 684Feb 2, 1994: 685 changed error() to print line number as %d, not %g. 686 687Jul 23, 1993: 688 cosmetic changes: increased sizes of some arrays, 689 reworded some error messages. 690 691 added CONVFMT as in posix (just replaced OFMT in getsval) 692 693 FILENAME is now "" until the first thing that causes a file 694 to be opened. 695 696Nov 28, 1992: 697 deleted yyunput and yyoutput from proto.h; 698 different versions of lex give these different declarations. 699 700May 31, 1992: 701 added -mr N and -mf N options: more record and fields. 702 these really ought to adjust automatically. 703 704 cleaned up some error messages; "out of space" now means 705 malloc returned NULL in all cases. 706 707 changed rehash so that if it runs out, it just returns; 708 things will continue to run slow, but maybe a bit longer. 709 710Apr 24, 1992: 711 remove redundant close of stdin when using -f -. 712 713 got rid of core dump with -d; awk -d just prints date. 714 715Apr 12, 1992: 716 added explicit check for /dev/std(in,out,err) in redirection. 717 unlike gawk, no /dev/fd/n yet. 718 719 added (file/pipe) builtin. hard to test satisfactorily. 720 not posix. 721 722Feb 20, 1992: 723 recompile after abortive changes; should be unchanged. 724 725Dec 2, 1991: 726 die-casting time: converted to ansi C, installed that. 727 728Nov 30, 1991: 729 fixed storage leak in freefa, failing to recover [N]CCL. 730 thanks to Bill Jones (jones@cs.usask.ca) 731 732Nov 19, 1991: 733 use RAND_MAX instead of literal in builtin(). 734 735Nov 12, 1991: 736 cranked up some fixed-size arrays in b.c, and added a test for 737 overflow in penter. thanks to mark larsen. 738 739Sep 24, 1991: 740 increased buffer in gsub. a very crude fix to a general problem. 741 and again on Sep 26. 742 743Aug 18, 1991: 744 enforce variable name syntax for commandline variables: has to 745 start with letter or _. 746 747Jul 27, 1991: 748 allow newline after ; in for statements. 749 750Jul 21, 1991: 751 fixed so that in self-assignment like $1=$1, side effects 752 like recomputing $0 take place. (this is getting subtle.) 753 754Jun 30, 1991: 755 better test for detecting too-long output record. 756 757Jun 2, 1991: 758 better defense against very long printf strings. 759 made break and continue illegal outside of loops. 760 761May 13, 1991: 762 removed extra arg on gettemp, tempfree. minor error message rewording. 763 764May 6, 1991: 765 fixed silly bug in hex parsing in hexstr(). 766 removed an apparently unnecessary test in isnumber(). 767 warn about weird printf conversions. 768 fixed unchecked array overwrite in relex(). 769 770 changed for (i in array) to access elements in sorted order. 771 then unchanged it -- it really does run slower in too many cases. 772 left the code in place, commented out. 773 774Feb 10, 1991: 775 check error status on all writes, to avoid banging on full disks. 776 777Jan 28, 1991: 778 awk -f - reads the program from stdin. 779 780Jan 11, 1991: 781 failed to set numeric state on $0 in cmd|getline context in run.c. 782 783Nov 2, 1990: 784 fixed sleazy test for integrality in getsval; use modf. 785 786Oct 29, 1990: 787 fixed sleazy buggy code in lib.c that looked (incorrectly) for 788 too long input lines. 789 790Oct 14, 1990: 791 fixed the bug on p. 198 in which it couldn't deduce that an 792 argument was an array in some contexts. replaced the error 793 message in intest() by code that damn well makes it an array. 794 795Oct 8, 1990: 796 fixed horrible bug: types and values were not preserved in 797 some kinds of self-assignment. (in assign().) 798 799Aug 24, 1990: 800 changed NCHARS to 256 to handle 8-bit characters in strings 801 presented to match(), etc. 802 803Jun 26, 1990: 804 changed struct rrow (awk.h) to use long instead of int for lval, 805 since cfoll() stores a pointer in it. now works better when int's 806 are smaller than pointers! 807 808May 6, 1990: 809 AVA fixed the grammar so that ! is uniformly of the same precedence as 810 unary + and -. This renders illegal some constructs like !x=y, which 811 now has to be parenthesized as !(x=y), and makes others work properly: 812 !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements. 813 (These problems were pointed out by Bob Lenk of Posix.) 814 815 Added \x to regular expressions (already in strings). 816 Limited octal to octal digits; \8 and \9 are not octal. 817 Centralized the code for parsing escapes in regular expressions. 818 Added a bunch of tests to T.re and T.sub to verify some of this. 819 820Feb 9, 1990: 821 fixed null pointer dereference bug in main.c: -F[nothing]. sigh. 822 823 restored srand behavior: it returns the current seed. 824 825Jan 18, 1990: 826 srand now returns previous seed value (0 to start). 827 828Jan 5, 1990: 829 fix potential problem in tran.c -- something was freed, 830 then used in freesymtab. 831 832Oct 18, 1989: 833 another try to get the max number of open files set with 834 relatively machine-independent code. 835 836 small fix to input() in case of multiple reads after EOF. 837 838Oct 11, 1989: 839 FILENAME is now defined in the BEGIN block -- too many old 840 programs broke. 841 842 "-" means stdin in getline as well as on the commandline. 843 844 added a bunch of casts to the code to tell the truth about 845 char * vs. unsigned char *, a right royal pain. added a 846 setlocale call to the front of main, though probably no one 847 has it usefully implemented yet. 848 849Aug 24, 1989: 850 removed redundant relational tests against nullnode if parse 851 tree already had a relational at that point. 852 853Aug 11, 1989: 854 fixed bug: commandline variable assignment has to look like 855 var=something. (consider the man page for =, in file =.1) 856 857 changed number of arguments to functions to static arrays 858 to avoid repeated malloc calls. 859 860Aug 2, 1989: 861 restored -F (space) separator 862 863Jul 30, 1989: 864 added -v x=1 y=2 ... for immediate commandline variable assignment; 865 done before the BEGIN block for sure. they have to precede the 866 program if the program is on the commandline. 867 Modified Aug 2 to require a separate -v for each assignment. 868 869Jul 10, 1989: 870 fixed ref-thru-zero bug in environment code in tran.c 871 872Jun 23, 1989: 873 add newline to usage message. 874 875Jun 14, 1989: 876 added some missing ansi printf conversion letters: %i %X %E %G. 877 no sensible meaning for h or L, so they may not do what one expects. 878 879 made %* conversions work. 880 881 changed x^y so that if n is a positive integer, it's done 882 by explicit multiplication, thus achieving maximum accuracy. 883 (this should be done by pow() but it seems not to be locally.) 884 done to x ^= y as well. 885 886Jun 4, 1989: 887 ENVIRON array contains environment: if shell variable V=thing, 888 ENVIRON["V"] is "thing" 889 890 multiple -f arguments permitted. error reporting is naive. 891 (they were permitted before, but only the last was used.) 892 893 fixed a really stupid botch in the debugging macro dprintf 894 895 fixed order of evaluation of commandline assignments to match 896 what the book claims: an argument of the form x=e is evaluated 897 at the time it would have been opened if it were a filename (p 63). 898 this invalidates the suggested answer to ex 4-1 (p 195). 899 900 removed some code that permitted -F (space) fieldseparator, 901 since it didn't quite work right anyway. (restored aug 2) 902 903Apr 27, 1989: 904 Line number now accumulated correctly for comment lines. 905 906Apr 26, 1989: 907 Debugging output now includes a version date, 908 if one compiles it into the source each time. 909 910Apr 9, 1989: 911 Changed grammar to prohibit constants as 3rd arg of sub and gsub; 912 prevents class of overwriting-a-constant errors. (Last one?) 913 This invalidates the "banana" example on page 43 of the book. 914 915 Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal), 916 as in ANSI, for strings. Rescinded the sloppiness that permitted 917 non-octal digits in \ooo. Warning: not all compilers and libraries 918 will be able to deal with \x correctly. 919 920Jan 9, 1989: 921 Fixed bug that caused tempcell list to contain a duplicate. 922 The fix is kludgy. 923 924Dec 17, 1988: 925 Catches some more commandline errors in main. 926 Removed redundant decl of modf in run.c (confuses some compilers). 927 Warning: there's no single declaration of malloc, etc., in awk.h 928 that seems to satisfy all compilers. 929 930Dec 7, 1988: 931 Added a bit of code to error printing to avoid printing nulls. 932 (Not clear that it actually would.) 933 934Nov 27, 1988: 935 With fear and trembling, modified the grammar to permit 936 multiple pattern-action statements on one line without 937 an explicit separator. By definition, this capitulation 938 to the ghost of ancient implementations remains undefined 939 and thus subject to change without notice or apology. 940 DO NOT COUNT ON IT. 941 942Oct 30, 1988: 943 Fixed bug in call() that failed to recover storage. 944 945 A warning is now generated if there are more arguments 946 in the call than in the definition (in lieu of fixing 947 another storage leak). 948 949Oct 20, 1988: 950 Fixed %c: if expr is numeric, use numeric value; 951 otherwise print 1st char of string value. still 952 doesn't work if the value is 0 -- won't print \0. 953 954 Added a few more checks for running out of malloc. 955 956Oct 12, 1988: 957 Fixed bug in call() that freed local arrays twice. 958 959 Fixed to handle deletion of non-existent array right; 960 complains about attempt to delete non-array element. 961 962Sep 30, 1988: 963 Now guarantees to evaluate all arguments of built-in 964 functions, as in C; the appearance is that arguments 965 are evaluated before the function is called. Places 966 affected are sub (gsub was ok), substr, printf, and 967 all the built-in arithmetic functions in bltin(). 968 A warning is generated if a bltin() is called with 969 the wrong number of arguments. 970 971 This requires changing makeprof on p167 of the book. 972 973Aug 23, 1988: 974 setting FILENAME in BEGIN caused core dump, apparently 975 because it was freeing space not allocated by malloc. 976 977July 24, 1988: 978 fixed egregious error in toupper/tolower functions. 979 still subject to rescinding, however. 980 981July 2, 1988: 982 flush stdout before opening file or pipe 983 984July 2, 1988: 985 performance bug in b.c/cgoto(): not freeing some sets of states. 986 partial fix only right now, and the number of states increased 987 to make it less obvious. 988 989June 1, 1988: 990 check error status on close 991 992May 28, 1988: 993 srand returns seed value it's using. 994 see 1/18/90 995 996May 22, 1988: 997 Removed limit on depth of function calls. 998 999May 10, 1988: 1000 Fixed lib.c to permit _ in commandline variable names. 1001 1002Mar 25, 1988: 1003 main.c fixed to recognize -- as terminator of command- 1004 line options. Illegal options flagged. 1005 Error reporting slightly cleaned up. 1006 1007Dec 2, 1987: 1008 Newer C compilers apply a strict scope rule to extern 1009 declarations within functions. Two extern declarations in 1010 lib.c and tran.c have been moved to obviate this problem. 1011 1012Oct xx, 1987: 1013 Reluctantly added toupper and tolower functions. 1014 Subject to rescinding without notice. 1015 1016Sep 17, 1987: 1017 Error-message printer had printf(s) instead of 1018 printf("%s",s); got core dumps when the message 1019 included a %. 1020 1021Sep 12, 1987: 1022 Very long printf strings caused core dump; 1023 fixed aprintf, asprintf, format to catch them. 1024 Can still get a core dump in printf itself. 1025 1026 1027