FIXES revision 118194
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 28Jul 29, 2003: 29 fixed (i think) the long-standing botch that included the beginning of 30 line state ^ for RE's in the set of valid characters; this led to a 31 variety of odd problems, including failure to properly match certain 32 regular expressions in non-US locales. thanks to ruslan for keeping 33 at this one. 34 35Jul 28, 2003: 36 n-th try at getting internationalization right, with thanks to volker 37 kiefel, arnold robbins and ruslan ermilov for advice, though they 38 should not be blamed for the outcome. according to posix, "." is the 39 radix character in programs and command line arguments regardless of 40 the locale; otherwise, the locale should prevail for input and output 41 of numbers. so it's intended to work that way. 42 43 i have rescinded the attempt to use strcoll in expanding shorthands in 44 regular expressions (cclenter). its properties are much too 45 surprising; for example [a-c] matches aAbBc in locale en_US but abBcC 46 in locale fr_CA. i can see how this might arise by implementation 47 but i cannot explain it to a human user. (this behavior can be seen 48 in gawk as well; we're leaning on the same library.) 49 50 the issue appears to be that strcoll is meant for sorting, where 51 merging upper and lower case may make sense (though note that unix 52 sort does not do this by default either). it is not appropriate 53 for regular expressions, where the goal is to match specific 54 patterns of characters. in any case, the notations [:lower:], etc., 55 are available in awk, and they are more likely to work correctly in 56 most locales. 57 58 a moratorium is hereby declared on internationalization changes. 59 i apologize to friends and colleagues in other parts of the world. 60 i would truly like to get this "right", but i don't know what 61 that is, and i do not want to keep making changes until it's clear. 62 63Jul 4, 2003: 64 fixed bug that permitted non-terminated RE, as in "awk /x". 65 66Jun 1, 2003: 67 subtle change to split: if source is empty, number of elems 68 is always 0 and the array is not set. 69 70Mar 21, 2003: 71 added some parens to isblank, in another attempt to make things 72 internationally portable. 73 74Mar 14, 2003: 75 the internationalization changes, somewhat modified, are now 76 reinstated. in theory awk will now do character comparisons 77 and case conversions in national language, but "." will always 78 be the decimal point separator on input and output regardless 79 of national language. isblank(){} has an #ifndef. 80 81 this no longer compiles on windows: LC_MESSAGES isn't defined 82 in vc6++. 83 84 fixed subtle behavior in field and record splitting: if FS is 85 a single character and RS is not empty, \n is NOT a separator. 86 this tortuous reading is found in the awk book; behavior now 87 matches gawk and mawk. 88 89Dec 13, 2002: 90 for the moment, the internationalization changes of nov 29 are 91 rolled back -- programs like x = 1.2 don't work in some locales, 92 because the parser is expecting x = 1,2. until i understand this 93 better, this will have to wait. 94 95Nov 29, 2002: 96 modified b.c (with tiny changes in main and run) to support 97 locales, using strcoll and iswhatever tests for posix character 98 classes. thanks to ruslan ermilov (ru@freebsd.org) for code. 99 the function isblank doesn't seem to have propagated to any 100 header file near me, so it's there explicitly. not properly 101 tested on non-ascii character sets by me. 102 103Jun 28, 2002: 104 modified run/format() and tran/getsval() to do a slightly better 105 job on using OFMT for output from print and CONVFMT for other 106 number->string conversions, as promised by posix and done by 107 gawk and mawk. there are still places where it doesn't work 108 right if CONVFMT is changed; by then the STR attribute of the 109 variable has been irrevocably set. thanks to arnold robbins for 110 code and examples. 111 112 fixed subtle bug in format that could get core dump. thanks to 113 Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing. 114 minor cleanup in run.c / format() at the same time. 115 116 added some tests for null pointers to debugging printf's, which 117 were never intended for external consumption. thanks to dave 118 kerns (dkerns@lucent.com) for pointing this out. 119 120 GNU compatibility: an empty regexp matches anything (thanks to 121 dag-erling smorgrav, des@ofug.org). subject to reversion if 122 this does more harm than good. 123 124 pervasive small changes to make things more const-correct, as 125 reported by gcc's -Wwrite-strings. as it says in the gcc manual, 126 this may be more nuisance than useful. provoked by a suggestion 127 and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk 128 129 minor documentation changes to note that this now compiles out 130 of the box on Mac OS X. 131 132Feb 10, 2002: 133 changed types in posix chars structure to quiet solaris cc. 134 135Jan 1, 2002: 136 fflush() or fflush("") flushes all files and pipes. 137 138 length(arrayname) returns number of elements; thanks to 139 arnold robbins for suggestion. 140 141 added a makefile.win to make it easier to build on windows. 142 based on dan allen's buildwin.bat. 143 144Nov 16, 2001: 145 added support for posix character class names like [:digit:], 146 which are not exactly shorter than [0-9] and perhaps no more 147 portable. thanks to dag-erling smorgrav for code. 148 149Feb 16, 2001: 150 removed -m option; no longer needed, and it was actually 151 broken (noted thanks to volker kiefel). 152 153Feb 10, 2001: 154 fixed an appalling bug in gettok: any sequence of digits, +,-, E, e, 155 and period was accepted as a valid number if it started with a period. 156 this would never have happened with the lex version. 157 158 other 1-character botches, now fixed, include a bare $ and a 159 bare " at the end of the input. 160 161Feb 7, 2001: 162 more (const char *) casts in b.c and tran.c to silence warnings. 163 164Nov 15, 2000: 165 fixed a bug introduced in august 1997 that caused expressions 166 like $f[1] to be syntax errors. thanks to arnold robbins for 167 noticing this and providing a fix. 168 169Oct 30, 2000: 170 fixed some nextfile bugs: not handling all cases. thanks to 171 arnold robbins for pointing this out. new regressions added. 172 173 close() is now a function. it returns whatever the library 174 fclose returns, and -1 for closing a file or pipe that wasn't 175 opened. 176 177Sep 24, 2000: 178 permit \n explicitly in character classes; won't work right 179 if comes in as "[\n]" but ok as /[\n]/, because of multiple 180 processing of \'s. thanks to arnold robbins. 181 182July 5, 2000: 183 minor fiddles in tran.c to keep compilers happy about uschar. 184 thanks to norman wilson. 185 186May 25, 2000: 187 yet another attempt at making 8-bit input work, with another 188 band-aid in b.c (member()), and some (uschar) casts to head 189 off potential errors in subscripts (like isdigit). also 190 changed HAT to NCHARS-2. thanks again to santiago vila. 191 192 changed maketab.c to ignore apparently out of range definitions 193 instead of halting; new freeBSD generates one. thanks to 194 jon snader <jsnader@ix.netcom.com> for pointing out the problem. 195 196May 2, 2000: 197 fixed an 8-bit problem in b.c by making several char*'s into 198 unsigned char*'s. not clear i have them all yet. thanks to 199 Santiago Vila <sanvila@unex.es> for the bug report. 200 201Apr 21, 2000: 202 finally found and fixed a memory leak in function call; it's 203 been there since functions were added ~1983. thanks to 204 jon bentley for the test case that found it. 205 206 added test in envinit to catch environment "variables" with 207 names begining with '='; thanks to Berend Hasselman. 208 209Jul 28, 1999: 210 added test in defn() to catch function foo(foo), which 211 otherwise recurses until core dump. thanks to arnold 212 robbins for noticing this. 213 214Jun 20, 1999: 215 added *bp in gettok in lex.c; appears possible to exit function 216 without terminating the string. thanks to russ cox. 217 218Jun 2, 1999: 219 added function stdinit() to run to initialize files[] array, 220 in case stdin, etc., are not constants; some compilers care. 221 222May 10, 1999: 223 replaced the ERROR ... FATAL, etc., macros with functions 224 based on vprintf, to avoid problems caused by overrunning 225 fixed-size errbuf array. thanks to ralph corderoy for the 226 impetus, and for pointing out a string termination bug in 227 qstring as well. 228 229Apr 21, 1999: 230 fixed bug that caused occasional core dumps with commandline 231 variable with value ending in \. (thanks to nelson beebe for 232 the test case.) 233 234Apr 16, 1999: 235 with code kindly provided by Bruce Lilly, awk now parses 236 /=/ and similar constructs more sensibly in more places. 237 Bruce also provided some helpful test cases. 238 239Apr 5, 1999: 240 changed true/false to True/False in run.c to make it 241 easier to compile with C++. Added some casts on malloc 242 and realloc to be honest about casts; ditto. changed 243 ltype int to long in struct rrow to reduce some 64-bit 244 complaints; other changes scattered throughout for the 245 same purpose. thanks to Nelson Beebe for these portability 246 improvements. 247 248 removed some horrible pointer-int casting in b.c and elsewhere 249 by adding ptoi and itonp to localize the casts, which are 250 all benign. fixed one incipient bug that showed up on sgi 251 in 64-bit mode. 252 253 reset lineno for new source file; include filename in error 254 message. also fixed line number error in continuation lines. 255 (thanks to Nelson Beebe for both of these.) 256 257Mar 24, 1999: 258 Nelson Beebe notes that irix 5.3 yacc dies with a bogus 259 error; use a newer version or switch to bison, since sgi 260 is unlikely to fix it. 261 262Mar 5, 1999: 263 changed isnumber to is_number to avoid the problem caused by 264 versions of ctype.h that include the name isnumber. 265 266 distribution now includes a script for building on a Mac, 267 thanks to Dan Allen. 268 269Feb 20, 1999: 270 fixed memory leaks in run.c (call) and tran.c (setfval). 271 thanks to Stephen Nutt for finding these and providing the fixes. 272 273Jan 13, 1999: 274 replaced srand argument by (unsigned int) in run.c; 275 avoids problem on Mac and potentially on Unix & Windows. 276 thanks to Dan Allen. 277 278 added a few (int) casts to silence useless compiler warnings. 279 e.g., errorflag= in run.c jump(). 280 281 added proctab.c to the bundle outout; one less thing 282 to have to compile out of the box. 283 284 added calls to _popen and _pclose to the win95 stub for 285 pipes (thanks to Steve Adams for this helpful suggestion). 286 seems to work, though properties are not well understood 287 by me, and it appears that under some circumstances the 288 pipe output is truncated. Be careful. 289 290Oct 19, 1998: 291 fixed a couple of bugs in getrec: could fail to update $0 292 after a getline var; because inputFS wasn't initialized, 293 could split $0 on every character, a misleading diversion. 294 295 fixed caching bug in makedfa: LRU was actually removing 296 least often used. 297 298 thanks to ross ridge for finding these, and for providing 299 great bug reports. 300 301May 12, 1998: 302 fixed potential bug in readrec: might fail to update record 303 pointer after growing. thanks to dan levy for spotting this 304 and suggesting the fix. 305 306Mar 12, 1998: 307 added -V to print version number and die. 308 309Feb 11, 1998: 310 subtle silent bug in lex.c: if the program ended with a number 311 longer than 1 digit, part of the input would be pushed back and 312 parsed again because token buffer wasn't terminated right. 313 example: awk 'length($0) > 10'. blush. at least i found it 314 myself. 315 316Aug 31, 1997: 317 s/adelete/awkdelete/: SGI uses this in malloc.h. 318 thanks to nelson beebe for pointing this one out. 319 320Aug 21, 1997: 321 fixed some bugs in sub and gsub when replacement includes \\. 322 this is a dark, horrible corner, but at least now i believe that 323 the behavior is the same as gawk and the intended posix standard. 324 thanks to arnold robbins for advice here. 325 326Aug 9, 1997: 327 somewhat regretfully, replaced the ancient lex-based lexical 328 analyzer with one written in C. it's longer, generates less code, 329 and more portable; the old one depended too much on mysterious 330 properties of lex that were not preserved in other environments. 331 in theory these recognize the same language. 332 333 now using strtod to test whether a string is a number, instead of 334 the convoluted original function. should be more portable and 335 reliable if strtod is implemented right. 336 337 removed now-pointless optimization in makefile that tries to avoid 338 recompilation when awkgram.y is changed but symbols are not. 339 340 removed most fixed-size arrays, though a handful remain, some 341 of which are unchecked. you have been warned. 342 343Aug 4, 1997: 344 with some trepidation, replaced the ancient code that managed 345 fields and $0 in fixed-size arrays with arrays that grow on 346 demand. there is still some tension between trying to make this 347 run fast and making it clean; not sure it's right yet. 348 349 the ill-conceived -mr and -mf arguments are now useful only 350 for debugging. previous dynamic string code removed. 351 352 numerous other minor cleanups along the way. 353 354Jul 30, 1997: 355 using code provided by dan levy (to whom profuse thanks), replaced 356 fixed-size arrays and awkward kludges by a fairly uniform mechanism 357 to grow arrays as needed for printf, sub, gsub, etc. 358 359Jul 23, 1997: 360 falling off the end of a function returns "" and 0, not 0. 361 thanks to arnold robbins. 362 363Jun 17, 1997: 364 replaced several fixed-size arrays by dynamically-created ones 365 in run.c; added overflow tests to some previously unchecked cases. 366 getline, toupper, tolower. 367 368 getline code is still broken in that recursive calls may wind 369 up using the same space. [fixed later] 370 371 increased RECSIZE to 8192 to push problems further over the horizon. 372 373 added \r to \n as input line separator for programs, not data. 374 damn CRLFs. 375 376 modified format() to permit explicit printf("%c", 0) to include 377 a null byte in output. thanks to ken stailey for the fix. 378 379 added a "-safe" argument that disables file output (print >, 380 print >>), process creation (cmd|getline, print |, system), and 381 access to the environment (ENVIRON). this is a first approximation 382 to a "safe" version of awk, but don't rely on it too much. thanks 383 to joan feigenbaum and matt blaze for the inspiration long ago. 384 385Jul 8, 1996: 386 fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to 387 ralph corderoy. 388 389Jun 29, 1996: 390 fixed awful bug in new field splitting; didn't get all the places 391 where input was done. 392 393Jun 28, 1996: 394 changed field-splitting to conform to posix definition: fields are 395 split using the value of FS at the time of input; it used to be 396 the value when the field or NF was first referred to, a much less 397 predictable definition. thanks to arnold robbins for encouragement 398 to do the right thing. 399 400May 28, 1996: 401 fixed appalling but apparently unimportant bug in parsing octal 402 numbers in reg exprs. 403 404 explicit hex in reg exprs now limited to 2 chars: \xa, \xaa. 405 406May 27, 1996: 407 cleaned up some declarations so gcc -Wall is now almost silent. 408 409 makefile now includes backup copies of ytab.c and lexyy.c in case 410 one makes before looking; it also avoids recreating lexyy.c unless 411 really needed. 412 413 s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes 414 with unwisely-written header files. 415 416 thanks to jeffrey friedl for several of these. 417 418May 26, 1996: 419 an attempt to rationalize the (unsigned) char issue. almost all 420 instances of unsigned char have been removed; the handful of places 421 in b.c where chars are used as table indices have been hand-crafted. 422 added some latin-1 tests to the regression, but i'm not confident; 423 none of my compilers seem to care much. thanks to nelson beebe for 424 pointing out some others that do care. 425 426May 2, 1996: 427 removed all register declarations. 428 429 enhanced split(), as in gawk, etc: split(s, a, "") splits s into 430 a[1]...a[length(s)] with each character a single element. 431 432 made the same changes for field-splitting if FS is "". 433 434 added nextfile, as in gawk: causes immediate advance to next 435 input file. (thanks to arnold robbins for inspiration and code). 436 437 small fixes to regexpr code: can now handle []], [[], and 438 variants; [] is now a syntax error, rather than matching 439 everything; [z-a] is now empty, not z. far from complete 440 or correct, however. (thanks to jeffrey friedl for pointing out 441 some awful behaviors.) 442 443Apr 29, 1996: 444 replaced uchar by uschar everwhere; apparently some compilers 445 usurp this name and this causes conflicts. 446 447 fixed call to time in run.c (bltin); arg is time_t *. 448 449 replaced horrible pointer/long punning in b.c by a legitimate 450 union. should be safer on 64-bit machines and cleaner everywhere. 451 (thanks to nelson beebe for pointing out some of these problems.) 452 453 replaced nested comments by #if 0...#endif in run.c, lib.c. 454 455 removed getsval, setsval, execute macros from run.c and lib.c. 456 machines are 100x faster than they were when these macros were 457 first used. 458 459 revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l, 460 y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of 461 portability to nameless systems. 462 463 "make bundle" now includes yacc and lex output files for recipients 464 who don't have yacc or lex. 465 466Aug 15, 1995: 467 initialized Cells in setsymtab more carefully; some fields 468 were not set. (thanks to purify, all of whose complaints i 469 think i now understand.) 470 471 fixed at least one error in gsub that looked at -1-th element 472 of an array when substituting for a null match (e.g., $). 473 474 delete arrayname is now legal; it clears the elements but leaves 475 the array, which may not be the right behavior. 476 477 modified makefile: my current make can't cope with the test used 478 to avoid unnecessary yacc invocations. 479 480Jul 17, 1995: 481 added dynamically growing strings to awk.lx.l and b.c 482 to permit regular expressions to be much bigger. 483 the state arrays can still overflow. 484 485Aug 24, 1994: 486 detect duplicate arguments in function definitions (mdm). 487 488May 11, 1994: 489 trivial fix to printf to limit string size in sub(). 490 491Apr 22, 1994: 492 fixed yet another subtle self-assignment problem: 493 $1 = $2; $1 = $1 clobbered $1. 494 495 Regression tests now use private echo, to avoid quoting problems. 496 497Feb 2, 1994: 498 changed error() to print line number as %d, not %g. 499 500Jul 23, 1993: 501 cosmetic changes: increased sizes of some arrays, 502 reworded some error messages. 503 504 added CONVFMT as in posix (just replaced OFMT in getsval) 505 506 FILENAME is now "" until the first thing that causes a file 507 to be opened. 508 509Nov 28, 1992: 510 deleted yyunput and yyoutput from proto.h; 511 different versions of lex give these different declarations. 512 513May 31, 1992: 514 added -mr N and -mf N options: more record and fields. 515 these really ought to adjust automatically. 516 517 cleaned up some error messages; "out of space" now means 518 malloc returned NULL in all cases. 519 520 changed rehash so that if it runs out, it just returns; 521 things will continue to run slow, but maybe a bit longer. 522 523Apr 24, 1992: 524 remove redundant close of stdin when using -f -. 525 526 got rid of core dump with -d; awk -d just prints date. 527 528Apr 12, 1992: 529 added explicit check for /dev/std(in,out,err) in redirection. 530 unlike gawk, no /dev/fd/n yet. 531 532 added (file/pipe) builtin. hard to test satisfactorily. 533 not posix. 534 535Feb 20, 1992: 536 recompile after abortive changes; should be unchanged. 537 538Dec 2, 1991: 539 die-casting time: converted to ansi C, installed that. 540 541Nov 30, 1991: 542 fixed storage leak in freefa, failing to recover [N]CCL. 543 thanks to Bill Jones (jones@cs.usask.ca) 544 545Nov 19, 1991: 546 use RAND_MAX instead of literal in builtin(). 547 548Nov 12, 1991: 549 cranked up some fixed-size arrays in b.c, and added a test for 550 overflow in penter. thanks to mark larsen. 551 552Sep 24, 1991: 553 increased buffer in gsub. a very crude fix to a general problem. 554 and again on Sep 26. 555 556Aug 18, 1991: 557 enforce variable name syntax for commandline variables: has to 558 start with letter or _. 559 560Jul 27, 1991: 561 allow newline after ; in for statements. 562 563Jul 21, 1991: 564 fixed so that in self-assignment like $1=$1, side effects 565 like recomputing $0 take place. (this is getting subtle.) 566 567Jun 30, 1991: 568 better test for detecting too-long output record. 569 570Jun 2, 1991: 571 better defense against very long printf strings. 572 made break and continue illegal outside of loops. 573 574May 13, 1991: 575 removed extra arg on gettemp, tempfree. minor error message rewording. 576 577May 6, 1991: 578 fixed silly bug in hex parsing in hexstr(). 579 removed an apparently unnecessary test in isnumber(). 580 warn about weird printf conversions. 581 fixed unchecked array overwrite in relex(). 582 583 changed for (i in array) to access elements in sorted order. 584 then unchanged it -- it really does run slower in too many cases. 585 left the code in place, commented out. 586 587Feb 10, 1991: 588 check error status on all writes, to avoid banging on full disks. 589 590Jan 28, 1991: 591 awk -f - reads the program from stdin. 592 593Jan 11, 1991: 594 failed to set numeric state on $0 in cmd|getline context in run.c. 595 596Nov 2, 1990: 597 fixed sleazy test for integrality in getsval; use modf. 598 599Oct 29, 1990: 600 fixed sleazy buggy code in lib.c that looked (incorrectly) for 601 too long input lines. 602 603Oct 14, 1990: 604 fixed the bug on p. 198 in which it couldn't deduce that an 605 argument was an array in some contexts. replaced the error 606 message in intest() by code that damn well makes it an array. 607 608Oct 8, 1990: 609 fixed horrible bug: types and values were not preserved in 610 some kinds of self-assignment. (in assign().) 611 612Aug 24, 1990: 613 changed NCHARS to 256 to handle 8-bit characters in strings 614 presented to match(), etc. 615 616Jun 26, 1990: 617 changed struct rrow (awk.h) to use long instead of int for lval, 618 since cfoll() stores a pointer in it. now works better when int's 619 are smaller than pointers! 620 621May 6, 1990: 622 AVA fixed the grammar so that ! is uniformly of the same precedence as 623 unary + and -. This renders illegal some constructs like !x=y, which 624 now has to be parenthesized as !(x=y), and makes others work properly: 625 !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements. 626 (These problems were pointed out by Bob Lenk of Posix.) 627 628 Added \x to regular expressions (already in strings). 629 Limited octal to octal digits; \8 and \9 are not octal. 630 Centralized the code for parsing escapes in regular expressions. 631 Added a bunch of tests to T.re and T.sub to verify some of this. 632 633Feb 9, 1990: 634 fixed null pointer dereference bug in main.c: -F[nothing]. sigh. 635 636 restored srand behavior: it returns the current seed. 637 638Jan 18, 1990: 639 srand now returns previous seed value (0 to start). 640 641Jan 5, 1990: 642 fix potential problem in tran.c -- something was freed, 643 then used in freesymtab. 644 645Oct 18, 1989: 646 another try to get the max number of open files set with 647 relatively machine-independent code. 648 649 small fix to input() in case of multiple reads after EOF. 650 651Oct 11, 1989: 652 FILENAME is now defined in the BEGIN block -- too many old 653 programs broke. 654 655 "-" means stdin in getline as well as on the commandline. 656 657 added a bunch of casts to the code to tell the truth about 658 char * vs. unsigned char *, a right royal pain. added a 659 setlocale call to the front of main, though probably no one 660 has it usefully implemented yet. 661 662Aug 24, 1989: 663 removed redundant relational tests against nullnode if parse 664 tree already had a relational at that point. 665 666Aug 11, 1989: 667 fixed bug: commandline variable assignment has to look like 668 var=something. (consider the man page for =, in file =.1) 669 670 changed number of arguments to functions to static arrays 671 to avoid repeated malloc calls. 672 673Aug 2, 1989: 674 restored -F (space) separator 675 676Jul 30, 1989: 677 added -v x=1 y=2 ... for immediate commandline variable assignment; 678 done before the BEGIN block for sure. they have to precede the 679 program if the program is on the commandline. 680 Modified Aug 2 to require a separate -v for each assignment. 681 682Jul 10, 1989: 683 fixed ref-thru-zero bug in environment code in tran.c 684 685Jun 23, 1989: 686 add newline to usage message. 687 688Jun 14, 1989: 689 added some missing ansi printf conversion letters: %i %X %E %G. 690 no sensible meaning for h or L, so they may not do what one expects. 691 692 made %* conversions work. 693 694 changed x^y so that if n is a positive integer, it's done 695 by explicit multiplication, thus achieving maximum accuracy. 696 (this should be done by pow() but it seems not to be locally.) 697 done to x ^= y as well. 698 699Jun 4, 1989: 700 ENVIRON array contains environment: if shell variable V=thing, 701 ENVIRON["V"] is "thing" 702 703 multiple -f arguments permitted. error reporting is naive. 704 (they were permitted before, but only the last was used.) 705 706 fixed a really stupid botch in the debugging macro dprintf 707 708 fixed order of evaluation of commandline assignments to match 709 what the book claims: an argument of the form x=e is evaluated 710 at the time it would have been opened if it were a filename (p 63). 711 this invalidates the suggested answer to ex 4-1 (p 195). 712 713 removed some code that permitted -F (space) fieldseparator, 714 since it didn't quite work right anyway. (restored aug 2) 715 716Apr 27, 1989: 717 Line number now accumulated correctly for comment lines. 718 719Apr 26, 1989: 720 Debugging output now includes a version date, 721 if one compiles it into the source each time. 722 723Apr 9, 1989: 724 Changed grammar to prohibit constants as 3rd arg of sub and gsub; 725 prevents class of overwriting-a-constant errors. (Last one?) 726 This invalidates the "banana" example on page 43 of the book. 727 728 Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal), 729 as in ANSI, for strings. Rescinded the sloppiness that permitted 730 non-octal digits in \ooo. Warning: not all compilers and libraries 731 will be able to deal with \x correctly. 732 733Jan 9, 1989: 734 Fixed bug that caused tempcell list to contain a duplicate. 735 The fix is kludgy. 736 737Dec 17, 1988: 738 Catches some more commandline errors in main. 739 Removed redundant decl of modf in run.c (confuses some compilers). 740 Warning: there's no single declaration of malloc, etc., in awk.h 741 that seems to satisfy all compilers. 742 743Dec 7, 1988: 744 Added a bit of code to error printing to avoid printing nulls. 745 (Not clear that it actually would.) 746 747Nov 27, 1988: 748 With fear and trembling, modified the grammar to permit 749 multiple pattern-action statements on one line without 750 an explicit separator. By definition, this capitulation 751 to the ghost of ancient implementations remains undefined 752 and thus subject to change without notice or apology. 753 DO NOT COUNT ON IT. 754 755Oct 30, 1988: 756 Fixed bug in call() that failed to recover storage. 757 758 A warning is now generated if there are more arguments 759 in the call than in the definition (in lieu of fixing 760 another storage leak). 761 762Oct 20, 1988: 763 Fixed %c: if expr is numeric, use numeric value; 764 otherwise print 1st char of string value. still 765 doesn't work if the value is 0 -- won't print \0. 766 767 Added a few more checks for running out of malloc. 768 769Oct 12, 1988: 770 Fixed bug in call() that freed local arrays twice. 771 772 Fixed to handle deletion of non-existent array right; 773 complains about attempt to delete non-array element. 774 775Sep 30, 1988: 776 Now guarantees to evaluate all arguments of built-in 777 functions, as in C; the appearance is that arguments 778 are evaluated before the function is called. Places 779 affected are sub (gsub was ok), substr, printf, and 780 all the built-in arithmetic functions in bltin(). 781 A warning is generated if a bltin() is called with 782 the wrong number of arguments. 783 784 This requires changing makeprof on p167 of the book. 785 786Aug 23, 1988: 787 setting FILENAME in BEGIN caused core dump, apparently 788 because it was freeing space not allocated by malloc. 789 790July 24, 1988: 791 fixed egregious error in toupper/tolower functions. 792 still subject to rescinding, however. 793 794July 2, 1988: 795 flush stdout before opening file or pipe 796 797July 2, 1988: 798 performance bug in b.c/cgoto(): not freeing some sets of states. 799 partial fix only right now, and the number of states increased 800 to make it less obvious. 801 802June 1, 1988: 803 check error status on close 804 805May 28, 1988: 806 srand returns seed value it's using. 807 see 1/18/90 808 809May 22, 1988: 810 Removed limit on depth of function calls. 811 812May 10, 1988: 813 Fixed lib.c to permit _ in commandline variable names. 814 815Mar 25, 1988: 816 main.c fixed to recognize -- as terminator of command- 817 line options. Illegal options flagged. 818 Error reporting slightly cleaned up. 819 820Dec 2, 1987: 821 Newer C compilers apply a strict scope rule to extern 822 declarations within functions. Two extern declarations in 823 lib.c and tran.c have been moved to obviate this problem. 824 825Oct xx, 1987: 826 Reluctantly added toupper and tolower functions. 827 Subject to rescinding without notice. 828 829Sep 17, 1987: 830 Error-message printer had printf(s) instead of 831 printf("%s",s); got core dumps when the message 832 included a %. 833 834Sep 12, 1987: 835 Very long printf strings caused core dump; 836 fixed aprintf, asprintf, format to catch them. 837 Can still get a core dump in printf itself. 838 839 840