FIXES revision 125601
185587Sobrien/**************************************************************** 285587SobrienCopyright (C) Lucent Technologies 1997 385587SobrienAll Rights Reserved 485587Sobrien 585587SobrienPermission to use, copy, modify, and distribute this software and 685587Sobrienits documentation for any purpose and without fee is hereby 785587Sobriengranted, provided that the above copyright notice appear in all 885587Sobriencopies and that both that the copyright notice and this 985587Sobrienpermission notice and warranty disclaimer appear in supporting 1085587Sobriendocumentation, and that the name Lucent Technologies or any of 1185587Sobrienits entities not be used in advertising or publicity pertaining 1285587Sobriento distribution of the software without specific, written prior 1385587Sobrienpermission. 1485587Sobrien 1585587SobrienLUCENT DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, 1685587SobrienINCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. 1785587SobrienIN NO EVENT SHALL LUCENT OR ANY OF ITS ENTITIES BE LIABLE FOR ANY 1885587SobrienSPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES 1985587SobrienWHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER 2085587SobrienIN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, 2185587SobrienARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF 2285587SobrienTHIS SOFTWARE. 2385587Sobrien****************************************************************/ 2485587Sobrien 2585587SobrienThis file lists all bug fixes, changes, etc., made since the AWK book 2685587Sobrienwas sent to the printers in August, 1987. 2785587Sobrien 28125601SruNov 22, 2003: 29125601Sru fixed a bug in regular expressions that dates (so help me) from 1977; 30125601Sru it's been there from the beginning. an anchored longest match that 31125601Sru was longer than the number of states triggered a failure to initialize 32125601Sru the machine properly. many thanks to moinak ghosh for not only finding 33125601Sru this one but for providing a fix, in some of the most mysterious 34125601Sru code known to man. 35125601Sru 36125601Sru fixed a storage leak in call() that appears to have been there since 37125601Sru 1983 or so -- a function without an explicit return that assigns a 38125601Sru string to a parameter leaked a Cell. thanks to moinak ghosh for 39125601Sru spotting this very subtle one. 40125601Sru 41125601SruJul 31, 2003: 42125601Sru fixed, thanks to andrey chernov and ruslan ermilov, a bug in lex.c 43125601Sru that mis-handled the character 255 in input. (it was being compared 44125601Sru to EOF with a signed comparison.) 45125601Sru 46118194SruJul 29, 2003: 47118194Sru fixed (i think) the long-standing botch that included the beginning of 48118194Sru line state ^ for RE's in the set of valid characters; this led to a 49118194Sru variety of odd problems, including failure to properly match certain 50118194Sru regular expressions in non-US locales. thanks to ruslan for keeping 51118194Sru at this one. 52118194Sru 53118194SruJul 28, 2003: 54118194Sru n-th try at getting internationalization right, with thanks to volker 55118194Sru kiefel, arnold robbins and ruslan ermilov for advice, though they 56118194Sru should not be blamed for the outcome. according to posix, "." is the 57118194Sru radix character in programs and command line arguments regardless of 58118194Sru the locale; otherwise, the locale should prevail for input and output 59118194Sru of numbers. so it's intended to work that way. 60118194Sru 61118194Sru i have rescinded the attempt to use strcoll in expanding shorthands in 62118194Sru regular expressions (cclenter). its properties are much too 63118194Sru surprising; for example [a-c] matches aAbBc in locale en_US but abBcC 64118194Sru in locale fr_CA. i can see how this might arise by implementation 65118194Sru but i cannot explain it to a human user. (this behavior can be seen 66118194Sru in gawk as well; we're leaning on the same library.) 67118194Sru 68118194Sru the issue appears to be that strcoll is meant for sorting, where 69118194Sru merging upper and lower case may make sense (though note that unix 70118194Sru sort does not do this by default either). it is not appropriate 71118194Sru for regular expressions, where the goal is to match specific 72118194Sru patterns of characters. in any case, the notations [:lower:], etc., 73118194Sru are available in awk, and they are more likely to work correctly in 74118194Sru most locales. 75118194Sru 76118194Sru a moratorium is hereby declared on internationalization changes. 77118194Sru i apologize to friends and colleagues in other parts of the world. 78118194Sru i would truly like to get this "right", but i don't know what 79118194Sru that is, and i do not want to keep making changes until it's clear. 80118194Sru 81118194SruJul 4, 2003: 82118194Sru fixed bug that permitted non-terminated RE, as in "awk /x". 83118194Sru 84118194SruJun 1, 2003: 85118194Sru subtle change to split: if source is empty, number of elems 86118194Sru is always 0 and the array is not set. 87118194Sru 88118194SruMar 21, 2003: 89118194Sru added some parens to isblank, in another attempt to make things 90118194Sru internationally portable. 91118194Sru 92112336SobrienMar 14, 2003: 93112336Sobrien the internationalization changes, somewhat modified, are now 94112336Sobrien reinstated. in theory awk will now do character comparisons 95112336Sobrien and case conversions in national language, but "." will always 96112336Sobrien be the decimal point separator on input and output regardless 97112336Sobrien of national language. isblank(){} has an #ifndef. 98112336Sobrien 99112336Sobrien this no longer compiles on windows: LC_MESSAGES isn't defined 100112336Sobrien in vc6++. 101112336Sobrien 102112336Sobrien fixed subtle behavior in field and record splitting: if FS is 103112336Sobrien a single character and RS is not empty, \n is NOT a separator. 104112336Sobrien this tortuous reading is found in the awk book; behavior now 105112336Sobrien matches gawk and mawk. 106112336Sobrien 107108072SobrienDec 13, 2002: 108108072Sobrien for the moment, the internationalization changes of nov 29 are 109108072Sobrien rolled back -- programs like x = 1.2 don't work in some locales, 110108072Sobrien because the parser is expecting x = 1,2. until i understand this 111108072Sobrien better, this will have to wait. 112108072Sobrien 113107806SobrienNov 29, 2002: 114107806Sobrien modified b.c (with tiny changes in main and run) to support 115107806Sobrien locales, using strcoll and iswhatever tests for posix character 116107806Sobrien classes. thanks to ruslan ermilov (ru@freebsd.org) for code. 117107806Sobrien the function isblank doesn't seem to have propagated to any 118107806Sobrien header file near me, so it's there explicitly. not properly 119107806Sobrien tested on non-ascii character sets by me. 120107806Sobrien 121107806SobrienJun 28, 2002: 122107806Sobrien modified run/format() and tran/getsval() to do a slightly better 123107806Sobrien job on using OFMT for output from print and CONVFMT for other 124107806Sobrien number->string conversions, as promised by posix and done by 125107806Sobrien gawk and mawk. there are still places where it doesn't work 126107806Sobrien right if CONVFMT is changed; by then the STR attribute of the 127107806Sobrien variable has been irrevocably set. thanks to arnold robbins for 128107806Sobrien code and examples. 129107806Sobrien 130107806Sobrien fixed subtle bug in format that could get core dump. thanks to 131107806Sobrien Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing. 132107806Sobrien minor cleanup in run.c / format() at the same time. 133107806Sobrien 134107806Sobrien added some tests for null pointers to debugging printf's, which 135107806Sobrien were never intended for external consumption. thanks to dave 136107806Sobrien kerns (dkerns@lucent.com) for pointing this out. 137107806Sobrien 138107806Sobrien GNU compatibility: an empty regexp matches anything (thanks to 139107806Sobrien dag-erling smorgrav, des@ofug.org). subject to reversion if 140107806Sobrien this does more harm than good. 141107806Sobrien 142107806Sobrien pervasive small changes to make things more const-correct, as 143107806Sobrien reported by gcc's -Wwrite-strings. as it says in the gcc manual, 144107806Sobrien this may be more nuisance than useful. provoked by a suggestion 145107806Sobrien and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk 146107806Sobrien 147107806Sobrien minor documentation changes to note that this now compiles out 148107806Sobrien of the box on Mac OS X. 149107806Sobrien 15090902SdesFeb 10, 2002: 15190902Sdes changed types in posix chars structure to quiet solaris cc. 15290902Sdes 15390902SdesJan 1, 2002: 15490902Sdes fflush() or fflush("") flushes all files and pipes. 15590902Sdes 15690902Sdes length(arrayname) returns number of elements; thanks to 15790902Sdes arnold robbins for suggestion. 15890902Sdes 15990902Sdes added a makefile.win to make it easier to build on windows. 16090902Sdes based on dan allen's buildwin.bat. 16190902Sdes 16290902SdesNov 16, 2001: 16390902Sdes added support for posix character class names like [:digit:], 16490902Sdes which are not exactly shorter than [0-9] and perhaps no more 16590902Sdes portable. thanks to dag-erling smorgrav for code. 16690902Sdes 16790902SdesFeb 16, 2001: 16890902Sdes removed -m option; no longer needed, and it was actually 16990902Sdes broken (noted thanks to volker kiefel). 17090902Sdes 17190902SdesFeb 10, 2001: 17290902Sdes fixed an appalling bug in gettok: any sequence of digits, +,-, E, e, 17390902Sdes and period was accepted as a valid number if it started with a period. 17490902Sdes this would never have happened with the lex version. 17590902Sdes 17690902Sdes other 1-character botches, now fixed, include a bare $ and a 17790902Sdes bare " at the end of the input. 17890902Sdes 17990902SdesFeb 7, 2001: 18090902Sdes more (const char *) casts in b.c and tran.c to silence warnings. 18190902Sdes 18285587SobrienNov 15, 2000: 18385587Sobrien fixed a bug introduced in august 1997 that caused expressions 18485587Sobrien like $f[1] to be syntax errors. thanks to arnold robbins for 18585587Sobrien noticing this and providing a fix. 18685587Sobrien 18785587SobrienOct 30, 2000: 18885587Sobrien fixed some nextfile bugs: not handling all cases. thanks to 18985587Sobrien arnold robbins for pointing this out. new regressions added. 19085587Sobrien 19185587Sobrien close() is now a function. it returns whatever the library 19285587Sobrien fclose returns, and -1 for closing a file or pipe that wasn't 19385587Sobrien opened. 19485587Sobrien 19585587SobrienSep 24, 2000: 19685587Sobrien permit \n explicitly in character classes; won't work right 19785587Sobrien if comes in as "[\n]" but ok as /[\n]/, because of multiple 19885587Sobrien processing of \'s. thanks to arnold robbins. 19985587Sobrien 20085587SobrienJuly 5, 2000: 20185587Sobrien minor fiddles in tran.c to keep compilers happy about uschar. 20285587Sobrien thanks to norman wilson. 20385587Sobrien 20485587SobrienMay 25, 2000: 20585587Sobrien yet another attempt at making 8-bit input work, with another 20685587Sobrien band-aid in b.c (member()), and some (uschar) casts to head 20785587Sobrien off potential errors in subscripts (like isdigit). also 20885587Sobrien changed HAT to NCHARS-2. thanks again to santiago vila. 20985587Sobrien 21085587Sobrien changed maketab.c to ignore apparently out of range definitions 21185587Sobrien instead of halting; new freeBSD generates one. thanks to 21285587Sobrien jon snader <jsnader@ix.netcom.com> for pointing out the problem. 21385587Sobrien 21485587SobrienMay 2, 2000: 21585587Sobrien fixed an 8-bit problem in b.c by making several char*'s into 21685587Sobrien unsigned char*'s. not clear i have them all yet. thanks to 21785587Sobrien Santiago Vila <sanvila@unex.es> for the bug report. 21885587Sobrien 21985587SobrienApr 21, 2000: 22085587Sobrien finally found and fixed a memory leak in function call; it's 22185587Sobrien been there since functions were added ~1983. thanks to 22285587Sobrien jon bentley for the test case that found it. 22385587Sobrien 22485587Sobrien added test in envinit to catch environment "variables" with 22585587Sobrien names begining with '='; thanks to Berend Hasselman. 22685587Sobrien 22785587SobrienJul 28, 1999: 22885587Sobrien added test in defn() to catch function foo(foo), which 22985587Sobrien otherwise recurses until core dump. thanks to arnold 23085587Sobrien robbins for noticing this. 23185587Sobrien 23285587SobrienJun 20, 1999: 23385587Sobrien added *bp in gettok in lex.c; appears possible to exit function 23485587Sobrien without terminating the string. thanks to russ cox. 23585587Sobrien 23685587SobrienJun 2, 1999: 23785587Sobrien added function stdinit() to run to initialize files[] array, 23885587Sobrien in case stdin, etc., are not constants; some compilers care. 23985587Sobrien 24085587SobrienMay 10, 1999: 24185587Sobrien replaced the ERROR ... FATAL, etc., macros with functions 24285587Sobrien based on vprintf, to avoid problems caused by overrunning 24385587Sobrien fixed-size errbuf array. thanks to ralph corderoy for the 24485587Sobrien impetus, and for pointing out a string termination bug in 24585587Sobrien qstring as well. 24685587Sobrien 24785587SobrienApr 21, 1999: 24885587Sobrien fixed bug that caused occasional core dumps with commandline 24985587Sobrien variable with value ending in \. (thanks to nelson beebe for 25085587Sobrien the test case.) 25185587Sobrien 25285587SobrienApr 16, 1999: 25385587Sobrien with code kindly provided by Bruce Lilly, awk now parses 25485587Sobrien /=/ and similar constructs more sensibly in more places. 25585587Sobrien Bruce also provided some helpful test cases. 25685587Sobrien 25785587SobrienApr 5, 1999: 25885587Sobrien changed true/false to True/False in run.c to make it 25985587Sobrien easier to compile with C++. Added some casts on malloc 26085587Sobrien and realloc to be honest about casts; ditto. changed 26185587Sobrien ltype int to long in struct rrow to reduce some 64-bit 26285587Sobrien complaints; other changes scattered throughout for the 26385587Sobrien same purpose. thanks to Nelson Beebe for these portability 26485587Sobrien improvements. 26585587Sobrien 26685587Sobrien removed some horrible pointer-int casting in b.c and elsewhere 26785587Sobrien by adding ptoi and itonp to localize the casts, which are 26885587Sobrien all benign. fixed one incipient bug that showed up on sgi 26985587Sobrien in 64-bit mode. 27085587Sobrien 27185587Sobrien reset lineno for new source file; include filename in error 27285587Sobrien message. also fixed line number error in continuation lines. 27385587Sobrien (thanks to Nelson Beebe for both of these.) 27485587Sobrien 27585587SobrienMar 24, 1999: 27685587Sobrien Nelson Beebe notes that irix 5.3 yacc dies with a bogus 27785587Sobrien error; use a newer version or switch to bison, since sgi 27885587Sobrien is unlikely to fix it. 27985587Sobrien 28085587SobrienMar 5, 1999: 28185587Sobrien changed isnumber to is_number to avoid the problem caused by 28285587Sobrien versions of ctype.h that include the name isnumber. 28385587Sobrien 28485587Sobrien distribution now includes a script for building on a Mac, 28585587Sobrien thanks to Dan Allen. 28685587Sobrien 28785587SobrienFeb 20, 1999: 28885587Sobrien fixed memory leaks in run.c (call) and tran.c (setfval). 28985587Sobrien thanks to Stephen Nutt for finding these and providing the fixes. 29085587Sobrien 29185587SobrienJan 13, 1999: 29285587Sobrien replaced srand argument by (unsigned int) in run.c; 29385587Sobrien avoids problem on Mac and potentially on Unix & Windows. 29485587Sobrien thanks to Dan Allen. 29585587Sobrien 29685587Sobrien added a few (int) casts to silence useless compiler warnings. 29785587Sobrien e.g., errorflag= in run.c jump(). 29885587Sobrien 29985587Sobrien added proctab.c to the bundle outout; one less thing 30085587Sobrien to have to compile out of the box. 30185587Sobrien 30285587Sobrien added calls to _popen and _pclose to the win95 stub for 30385587Sobrien pipes (thanks to Steve Adams for this helpful suggestion). 30485587Sobrien seems to work, though properties are not well understood 30585587Sobrien by me, and it appears that under some circumstances the 30685587Sobrien pipe output is truncated. Be careful. 30785587Sobrien 30885587SobrienOct 19, 1998: 30985587Sobrien fixed a couple of bugs in getrec: could fail to update $0 31085587Sobrien after a getline var; because inputFS wasn't initialized, 31185587Sobrien could split $0 on every character, a misleading diversion. 31285587Sobrien 31385587Sobrien fixed caching bug in makedfa: LRU was actually removing 31485587Sobrien least often used. 31585587Sobrien 31685587Sobrien thanks to ross ridge for finding these, and for providing 31785587Sobrien great bug reports. 31885587Sobrien 31985587SobrienMay 12, 1998: 32085587Sobrien fixed potential bug in readrec: might fail to update record 32185587Sobrien pointer after growing. thanks to dan levy for spotting this 32285587Sobrien and suggesting the fix. 32385587Sobrien 32485587SobrienMar 12, 1998: 32585587Sobrien added -V to print version number and die. 32685587Sobrien 32785587SobrienFeb 11, 1998: 32885587Sobrien subtle silent bug in lex.c: if the program ended with a number 32985587Sobrien longer than 1 digit, part of the input would be pushed back and 33085587Sobrien parsed again because token buffer wasn't terminated right. 33185587Sobrien example: awk 'length($0) > 10'. blush. at least i found it 33285587Sobrien myself. 33385587Sobrien 33485587SobrienAug 31, 1997: 33585587Sobrien s/adelete/awkdelete/: SGI uses this in malloc.h. 33685587Sobrien thanks to nelson beebe for pointing this one out. 33785587Sobrien 33885587SobrienAug 21, 1997: 33985587Sobrien fixed some bugs in sub and gsub when replacement includes \\. 34085587Sobrien this is a dark, horrible corner, but at least now i believe that 34185587Sobrien the behavior is the same as gawk and the intended posix standard. 34285587Sobrien thanks to arnold robbins for advice here. 34385587Sobrien 34485587SobrienAug 9, 1997: 34585587Sobrien somewhat regretfully, replaced the ancient lex-based lexical 34685587Sobrien analyzer with one written in C. it's longer, generates less code, 34785587Sobrien and more portable; the old one depended too much on mysterious 34885587Sobrien properties of lex that were not preserved in other environments. 34985587Sobrien in theory these recognize the same language. 35085587Sobrien 35185587Sobrien now using strtod to test whether a string is a number, instead of 35285587Sobrien the convoluted original function. should be more portable and 35385587Sobrien reliable if strtod is implemented right. 35485587Sobrien 35585587Sobrien removed now-pointless optimization in makefile that tries to avoid 35685587Sobrien recompilation when awkgram.y is changed but symbols are not. 35785587Sobrien 35885587Sobrien removed most fixed-size arrays, though a handful remain, some 35985587Sobrien of which are unchecked. you have been warned. 36085587Sobrien 36185587SobrienAug 4, 1997: 36285587Sobrien with some trepidation, replaced the ancient code that managed 36385587Sobrien fields and $0 in fixed-size arrays with arrays that grow on 36485587Sobrien demand. there is still some tension between trying to make this 36585587Sobrien run fast and making it clean; not sure it's right yet. 36685587Sobrien 36785587Sobrien the ill-conceived -mr and -mf arguments are now useful only 36885587Sobrien for debugging. previous dynamic string code removed. 36985587Sobrien 37085587Sobrien numerous other minor cleanups along the way. 37185587Sobrien 37285587SobrienJul 30, 1997: 37385587Sobrien using code provided by dan levy (to whom profuse thanks), replaced 37485587Sobrien fixed-size arrays and awkward kludges by a fairly uniform mechanism 37585587Sobrien to grow arrays as needed for printf, sub, gsub, etc. 37685587Sobrien 37785587SobrienJul 23, 1997: 37885587Sobrien falling off the end of a function returns "" and 0, not 0. 37985587Sobrien thanks to arnold robbins. 38085587Sobrien 38185587SobrienJun 17, 1997: 38285587Sobrien replaced several fixed-size arrays by dynamically-created ones 38385587Sobrien in run.c; added overflow tests to some previously unchecked cases. 38485587Sobrien getline, toupper, tolower. 38585587Sobrien 38685587Sobrien getline code is still broken in that recursive calls may wind 38785587Sobrien up using the same space. [fixed later] 38885587Sobrien 38985587Sobrien increased RECSIZE to 8192 to push problems further over the horizon. 39085587Sobrien 39185587Sobrien added \r to \n as input line separator for programs, not data. 39285587Sobrien damn CRLFs. 39385587Sobrien 39485587Sobrien modified format() to permit explicit printf("%c", 0) to include 39585587Sobrien a null byte in output. thanks to ken stailey for the fix. 39685587Sobrien 39785587Sobrien added a "-safe" argument that disables file output (print >, 39885587Sobrien print >>), process creation (cmd|getline, print |, system), and 39985587Sobrien access to the environment (ENVIRON). this is a first approximation 40085587Sobrien to a "safe" version of awk, but don't rely on it too much. thanks 40185587Sobrien to joan feigenbaum and matt blaze for the inspiration long ago. 40285587Sobrien 40385587SobrienJul 8, 1996: 40485587Sobrien fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to 40585587Sobrien ralph corderoy. 40685587Sobrien 40785587SobrienJun 29, 1996: 40885587Sobrien fixed awful bug in new field splitting; didn't get all the places 40985587Sobrien where input was done. 41085587Sobrien 41185587SobrienJun 28, 1996: 41285587Sobrien changed field-splitting to conform to posix definition: fields are 41385587Sobrien split using the value of FS at the time of input; it used to be 41485587Sobrien the value when the field or NF was first referred to, a much less 41585587Sobrien predictable definition. thanks to arnold robbins for encouragement 41685587Sobrien to do the right thing. 41785587Sobrien 41885587SobrienMay 28, 1996: 41985587Sobrien fixed appalling but apparently unimportant bug in parsing octal 42085587Sobrien numbers in reg exprs. 42185587Sobrien 42285587Sobrien explicit hex in reg exprs now limited to 2 chars: \xa, \xaa. 42385587Sobrien 42485587SobrienMay 27, 1996: 42585587Sobrien cleaned up some declarations so gcc -Wall is now almost silent. 42685587Sobrien 42785587Sobrien makefile now includes backup copies of ytab.c and lexyy.c in case 42885587Sobrien one makes before looking; it also avoids recreating lexyy.c unless 42985587Sobrien really needed. 43085587Sobrien 43185587Sobrien s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes 43285587Sobrien with unwisely-written header files. 43385587Sobrien 43485587Sobrien thanks to jeffrey friedl for several of these. 43585587Sobrien 43685587SobrienMay 26, 1996: 43785587Sobrien an attempt to rationalize the (unsigned) char issue. almost all 43885587Sobrien instances of unsigned char have been removed; the handful of places 43985587Sobrien in b.c where chars are used as table indices have been hand-crafted. 44085587Sobrien added some latin-1 tests to the regression, but i'm not confident; 44185587Sobrien none of my compilers seem to care much. thanks to nelson beebe for 44285587Sobrien pointing out some others that do care. 44385587Sobrien 44485587SobrienMay 2, 1996: 44585587Sobrien removed all register declarations. 44685587Sobrien 44785587Sobrien enhanced split(), as in gawk, etc: split(s, a, "") splits s into 44885587Sobrien a[1]...a[length(s)] with each character a single element. 44985587Sobrien 45085587Sobrien made the same changes for field-splitting if FS is "". 45185587Sobrien 45285587Sobrien added nextfile, as in gawk: causes immediate advance to next 45385587Sobrien input file. (thanks to arnold robbins for inspiration and code). 45485587Sobrien 45585587Sobrien small fixes to regexpr code: can now handle []], [[], and 45685587Sobrien variants; [] is now a syntax error, rather than matching 45785587Sobrien everything; [z-a] is now empty, not z. far from complete 45885587Sobrien or correct, however. (thanks to jeffrey friedl for pointing out 45985587Sobrien some awful behaviors.) 46085587Sobrien 46185587SobrienApr 29, 1996: 46285587Sobrien replaced uchar by uschar everwhere; apparently some compilers 46385587Sobrien usurp this name and this causes conflicts. 46485587Sobrien 46585587Sobrien fixed call to time in run.c (bltin); arg is time_t *. 46685587Sobrien 46785587Sobrien replaced horrible pointer/long punning in b.c by a legitimate 46885587Sobrien union. should be safer on 64-bit machines and cleaner everywhere. 46985587Sobrien (thanks to nelson beebe for pointing out some of these problems.) 47085587Sobrien 47185587Sobrien replaced nested comments by #if 0...#endif in run.c, lib.c. 47285587Sobrien 47385587Sobrien removed getsval, setsval, execute macros from run.c and lib.c. 47485587Sobrien machines are 100x faster than they were when these macros were 47585587Sobrien first used. 47685587Sobrien 47785587Sobrien revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l, 47885587Sobrien y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of 47985587Sobrien portability to nameless systems. 48085587Sobrien 48185587Sobrien "make bundle" now includes yacc and lex output files for recipients 48285587Sobrien who don't have yacc or lex. 48385587Sobrien 48485587SobrienAug 15, 1995: 48585587Sobrien initialized Cells in setsymtab more carefully; some fields 48685587Sobrien were not set. (thanks to purify, all of whose complaints i 48785587Sobrien think i now understand.) 48885587Sobrien 48985587Sobrien fixed at least one error in gsub that looked at -1-th element 49085587Sobrien of an array when substituting for a null match (e.g., $). 49185587Sobrien 49285587Sobrien delete arrayname is now legal; it clears the elements but leaves 49385587Sobrien the array, which may not be the right behavior. 49485587Sobrien 49585587Sobrien modified makefile: my current make can't cope with the test used 49685587Sobrien to avoid unnecessary yacc invocations. 49785587Sobrien 49885587SobrienJul 17, 1995: 49985587Sobrien added dynamically growing strings to awk.lx.l and b.c 50085587Sobrien to permit regular expressions to be much bigger. 50185587Sobrien the state arrays can still overflow. 50285587Sobrien 50385587SobrienAug 24, 1994: 50485587Sobrien detect duplicate arguments in function definitions (mdm). 50585587Sobrien 50685587SobrienMay 11, 1994: 50785587Sobrien trivial fix to printf to limit string size in sub(). 50885587Sobrien 50985587SobrienApr 22, 1994: 51085587Sobrien fixed yet another subtle self-assignment problem: 51185587Sobrien $1 = $2; $1 = $1 clobbered $1. 51285587Sobrien 51385587Sobrien Regression tests now use private echo, to avoid quoting problems. 51485587Sobrien 51585587SobrienFeb 2, 1994: 51685587Sobrien changed error() to print line number as %d, not %g. 51785587Sobrien 51885587SobrienJul 23, 1993: 51985587Sobrien cosmetic changes: increased sizes of some arrays, 52085587Sobrien reworded some error messages. 52185587Sobrien 52285587Sobrien added CONVFMT as in posix (just replaced OFMT in getsval) 52385587Sobrien 52485587Sobrien FILENAME is now "" until the first thing that causes a file 52585587Sobrien to be opened. 52685587Sobrien 52785587SobrienNov 28, 1992: 52885587Sobrien deleted yyunput and yyoutput from proto.h; 52985587Sobrien different versions of lex give these different declarations. 53085587Sobrien 53185587SobrienMay 31, 1992: 53285587Sobrien added -mr N and -mf N options: more record and fields. 53385587Sobrien these really ought to adjust automatically. 53485587Sobrien 53585587Sobrien cleaned up some error messages; "out of space" now means 53685587Sobrien malloc returned NULL in all cases. 53785587Sobrien 53885587Sobrien changed rehash so that if it runs out, it just returns; 53985587Sobrien things will continue to run slow, but maybe a bit longer. 54085587Sobrien 54185587SobrienApr 24, 1992: 54285587Sobrien remove redundant close of stdin when using -f -. 54385587Sobrien 54485587Sobrien got rid of core dump with -d; awk -d just prints date. 54585587Sobrien 54685587SobrienApr 12, 1992: 54785587Sobrien added explicit check for /dev/std(in,out,err) in redirection. 54885587Sobrien unlike gawk, no /dev/fd/n yet. 54985587Sobrien 55085587Sobrien added (file/pipe) builtin. hard to test satisfactorily. 55185587Sobrien not posix. 55285587Sobrien 55385587SobrienFeb 20, 1992: 55485587Sobrien recompile after abortive changes; should be unchanged. 55585587Sobrien 55685587SobrienDec 2, 1991: 55785587Sobrien die-casting time: converted to ansi C, installed that. 55885587Sobrien 55985587SobrienNov 30, 1991: 56085587Sobrien fixed storage leak in freefa, failing to recover [N]CCL. 56185587Sobrien thanks to Bill Jones (jones@cs.usask.ca) 56285587Sobrien 56385587SobrienNov 19, 1991: 56485587Sobrien use RAND_MAX instead of literal in builtin(). 56585587Sobrien 56685587SobrienNov 12, 1991: 56785587Sobrien cranked up some fixed-size arrays in b.c, and added a test for 56885587Sobrien overflow in penter. thanks to mark larsen. 56985587Sobrien 57085587SobrienSep 24, 1991: 57185587Sobrien increased buffer in gsub. a very crude fix to a general problem. 57285587Sobrien and again on Sep 26. 57385587Sobrien 57485587SobrienAug 18, 1991: 57585587Sobrien enforce variable name syntax for commandline variables: has to 57685587Sobrien start with letter or _. 57785587Sobrien 57885587SobrienJul 27, 1991: 57985587Sobrien allow newline after ; in for statements. 58085587Sobrien 58185587SobrienJul 21, 1991: 58285587Sobrien fixed so that in self-assignment like $1=$1, side effects 58385587Sobrien like recomputing $0 take place. (this is getting subtle.) 58485587Sobrien 58585587SobrienJun 30, 1991: 58685587Sobrien better test for detecting too-long output record. 58785587Sobrien 58885587SobrienJun 2, 1991: 58985587Sobrien better defense against very long printf strings. 59085587Sobrien made break and continue illegal outside of loops. 59185587Sobrien 59285587SobrienMay 13, 1991: 59385587Sobrien removed extra arg on gettemp, tempfree. minor error message rewording. 59485587Sobrien 59585587SobrienMay 6, 1991: 59685587Sobrien fixed silly bug in hex parsing in hexstr(). 59785587Sobrien removed an apparently unnecessary test in isnumber(). 59885587Sobrien warn about weird printf conversions. 59985587Sobrien fixed unchecked array overwrite in relex(). 60085587Sobrien 60185587Sobrien changed for (i in array) to access elements in sorted order. 60285587Sobrien then unchanged it -- it really does run slower in too many cases. 60385587Sobrien left the code in place, commented out. 60485587Sobrien 60585587SobrienFeb 10, 1991: 60685587Sobrien check error status on all writes, to avoid banging on full disks. 60785587Sobrien 60885587SobrienJan 28, 1991: 60985587Sobrien awk -f - reads the program from stdin. 61085587Sobrien 61185587SobrienJan 11, 1991: 61285587Sobrien failed to set numeric state on $0 in cmd|getline context in run.c. 61385587Sobrien 61485587SobrienNov 2, 1990: 61585587Sobrien fixed sleazy test for integrality in getsval; use modf. 61685587Sobrien 61785587SobrienOct 29, 1990: 61885587Sobrien fixed sleazy buggy code in lib.c that looked (incorrectly) for 61985587Sobrien too long input lines. 62085587Sobrien 62185587SobrienOct 14, 1990: 62285587Sobrien fixed the bug on p. 198 in which it couldn't deduce that an 62385587Sobrien argument was an array in some contexts. replaced the error 62485587Sobrien message in intest() by code that damn well makes it an array. 62585587Sobrien 62685587SobrienOct 8, 1990: 62785587Sobrien fixed horrible bug: types and values were not preserved in 62885587Sobrien some kinds of self-assignment. (in assign().) 62985587Sobrien 63085587SobrienAug 24, 1990: 63185587Sobrien changed NCHARS to 256 to handle 8-bit characters in strings 63285587Sobrien presented to match(), etc. 63385587Sobrien 63485587SobrienJun 26, 1990: 63585587Sobrien changed struct rrow (awk.h) to use long instead of int for lval, 63685587Sobrien since cfoll() stores a pointer in it. now works better when int's 63785587Sobrien are smaller than pointers! 63885587Sobrien 63985587SobrienMay 6, 1990: 64085587Sobrien AVA fixed the grammar so that ! is uniformly of the same precedence as 64185587Sobrien unary + and -. This renders illegal some constructs like !x=y, which 64285587Sobrien now has to be parenthesized as !(x=y), and makes others work properly: 64385587Sobrien !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements. 64485587Sobrien (These problems were pointed out by Bob Lenk of Posix.) 64585587Sobrien 64685587Sobrien Added \x to regular expressions (already in strings). 64785587Sobrien Limited octal to octal digits; \8 and \9 are not octal. 64885587Sobrien Centralized the code for parsing escapes in regular expressions. 64985587Sobrien Added a bunch of tests to T.re and T.sub to verify some of this. 65085587Sobrien 65185587SobrienFeb 9, 1990: 65285587Sobrien fixed null pointer dereference bug in main.c: -F[nothing]. sigh. 65385587Sobrien 65485587Sobrien restored srand behavior: it returns the current seed. 65585587Sobrien 65685587SobrienJan 18, 1990: 65785587Sobrien srand now returns previous seed value (0 to start). 65885587Sobrien 65985587SobrienJan 5, 1990: 66085587Sobrien fix potential problem in tran.c -- something was freed, 66185587Sobrien then used in freesymtab. 66285587Sobrien 66385587SobrienOct 18, 1989: 66485587Sobrien another try to get the max number of open files set with 66585587Sobrien relatively machine-independent code. 66685587Sobrien 66785587Sobrien small fix to input() in case of multiple reads after EOF. 66885587Sobrien 66985587SobrienOct 11, 1989: 67085587Sobrien FILENAME is now defined in the BEGIN block -- too many old 67185587Sobrien programs broke. 67285587Sobrien 67385587Sobrien "-" means stdin in getline as well as on the commandline. 67485587Sobrien 67585587Sobrien added a bunch of casts to the code to tell the truth about 67685587Sobrien char * vs. unsigned char *, a right royal pain. added a 67785587Sobrien setlocale call to the front of main, though probably no one 67885587Sobrien has it usefully implemented yet. 67985587Sobrien 68085587SobrienAug 24, 1989: 68185587Sobrien removed redundant relational tests against nullnode if parse 68285587Sobrien tree already had a relational at that point. 68385587Sobrien 68485587SobrienAug 11, 1989: 68585587Sobrien fixed bug: commandline variable assignment has to look like 68685587Sobrien var=something. (consider the man page for =, in file =.1) 68785587Sobrien 68885587Sobrien changed number of arguments to functions to static arrays 68985587Sobrien to avoid repeated malloc calls. 69085587Sobrien 69185587SobrienAug 2, 1989: 69285587Sobrien restored -F (space) separator 69385587Sobrien 69485587SobrienJul 30, 1989: 69585587Sobrien added -v x=1 y=2 ... for immediate commandline variable assignment; 69685587Sobrien done before the BEGIN block for sure. they have to precede the 69785587Sobrien program if the program is on the commandline. 69885587Sobrien Modified Aug 2 to require a separate -v for each assignment. 69985587Sobrien 70085587SobrienJul 10, 1989: 70185587Sobrien fixed ref-thru-zero bug in environment code in tran.c 70285587Sobrien 70385587SobrienJun 23, 1989: 70485587Sobrien add newline to usage message. 70585587Sobrien 70685587SobrienJun 14, 1989: 70785587Sobrien added some missing ansi printf conversion letters: %i %X %E %G. 70885587Sobrien no sensible meaning for h or L, so they may not do what one expects. 70985587Sobrien 71085587Sobrien made %* conversions work. 71185587Sobrien 71285587Sobrien changed x^y so that if n is a positive integer, it's done 71385587Sobrien by explicit multiplication, thus achieving maximum accuracy. 71485587Sobrien (this should be done by pow() but it seems not to be locally.) 71585587Sobrien done to x ^= y as well. 71685587Sobrien 71785587SobrienJun 4, 1989: 71885587Sobrien ENVIRON array contains environment: if shell variable V=thing, 71985587Sobrien ENVIRON["V"] is "thing" 72085587Sobrien 72185587Sobrien multiple -f arguments permitted. error reporting is naive. 72285587Sobrien (they were permitted before, but only the last was used.) 72385587Sobrien 72485587Sobrien fixed a really stupid botch in the debugging macro dprintf 72585587Sobrien 72685587Sobrien fixed order of evaluation of commandline assignments to match 72785587Sobrien what the book claims: an argument of the form x=e is evaluated 72885587Sobrien at the time it would have been opened if it were a filename (p 63). 72985587Sobrien this invalidates the suggested answer to ex 4-1 (p 195). 73085587Sobrien 73185587Sobrien removed some code that permitted -F (space) fieldseparator, 73285587Sobrien since it didn't quite work right anyway. (restored aug 2) 73385587Sobrien 73485587SobrienApr 27, 1989: 73585587Sobrien Line number now accumulated correctly for comment lines. 73685587Sobrien 73785587SobrienApr 26, 1989: 73885587Sobrien Debugging output now includes a version date, 73985587Sobrien if one compiles it into the source each time. 74085587Sobrien 74185587SobrienApr 9, 1989: 74285587Sobrien Changed grammar to prohibit constants as 3rd arg of sub and gsub; 74385587Sobrien prevents class of overwriting-a-constant errors. (Last one?) 74485587Sobrien This invalidates the "banana" example on page 43 of the book. 74585587Sobrien 74685587Sobrien Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal), 74785587Sobrien as in ANSI, for strings. Rescinded the sloppiness that permitted 74885587Sobrien non-octal digits in \ooo. Warning: not all compilers and libraries 74985587Sobrien will be able to deal with \x correctly. 75085587Sobrien 75185587SobrienJan 9, 1989: 75285587Sobrien Fixed bug that caused tempcell list to contain a duplicate. 75385587Sobrien The fix is kludgy. 75485587Sobrien 75585587SobrienDec 17, 1988: 75685587Sobrien Catches some more commandline errors in main. 75785587Sobrien Removed redundant decl of modf in run.c (confuses some compilers). 75885587Sobrien Warning: there's no single declaration of malloc, etc., in awk.h 75985587Sobrien that seems to satisfy all compilers. 76085587Sobrien 76185587SobrienDec 7, 1988: 76285587Sobrien Added a bit of code to error printing to avoid printing nulls. 76385587Sobrien (Not clear that it actually would.) 76485587Sobrien 76585587SobrienNov 27, 1988: 76685587Sobrien With fear and trembling, modified the grammar to permit 76785587Sobrien multiple pattern-action statements on one line without 76885587Sobrien an explicit separator. By definition, this capitulation 76985587Sobrien to the ghost of ancient implementations remains undefined 77085587Sobrien and thus subject to change without notice or apology. 77185587Sobrien DO NOT COUNT ON IT. 77285587Sobrien 77385587SobrienOct 30, 1988: 77485587Sobrien Fixed bug in call() that failed to recover storage. 77585587Sobrien 77685587Sobrien A warning is now generated if there are more arguments 77785587Sobrien in the call than in the definition (in lieu of fixing 77885587Sobrien another storage leak). 77985587Sobrien 78085587SobrienOct 20, 1988: 78185587Sobrien Fixed %c: if expr is numeric, use numeric value; 78285587Sobrien otherwise print 1st char of string value. still 78385587Sobrien doesn't work if the value is 0 -- won't print \0. 78485587Sobrien 78585587Sobrien Added a few more checks for running out of malloc. 78685587Sobrien 78785587SobrienOct 12, 1988: 78885587Sobrien Fixed bug in call() that freed local arrays twice. 78985587Sobrien 79085587Sobrien Fixed to handle deletion of non-existent array right; 79185587Sobrien complains about attempt to delete non-array element. 79285587Sobrien 79385587SobrienSep 30, 1988: 79485587Sobrien Now guarantees to evaluate all arguments of built-in 79585587Sobrien functions, as in C; the appearance is that arguments 79685587Sobrien are evaluated before the function is called. Places 79785587Sobrien affected are sub (gsub was ok), substr, printf, and 79885587Sobrien all the built-in arithmetic functions in bltin(). 79985587Sobrien A warning is generated if a bltin() is called with 80085587Sobrien the wrong number of arguments. 80185587Sobrien 80285587Sobrien This requires changing makeprof on p167 of the book. 80385587Sobrien 80485587SobrienAug 23, 1988: 80585587Sobrien setting FILENAME in BEGIN caused core dump, apparently 80685587Sobrien because it was freeing space not allocated by malloc. 80785587Sobrien 80885587SobrienJuly 24, 1988: 80985587Sobrien fixed egregious error in toupper/tolower functions. 81085587Sobrien still subject to rescinding, however. 81185587Sobrien 81285587SobrienJuly 2, 1988: 81385587Sobrien flush stdout before opening file or pipe 81485587Sobrien 81585587SobrienJuly 2, 1988: 81685587Sobrien performance bug in b.c/cgoto(): not freeing some sets of states. 81785587Sobrien partial fix only right now, and the number of states increased 81885587Sobrien to make it less obvious. 81985587Sobrien 82085587SobrienJune 1, 1988: 82185587Sobrien check error status on close 82285587Sobrien 82385587SobrienMay 28, 1988: 82485587Sobrien srand returns seed value it's using. 82585587Sobrien see 1/18/90 82685587Sobrien 82785587SobrienMay 22, 1988: 82885587Sobrien Removed limit on depth of function calls. 82985587Sobrien 83085587SobrienMay 10, 1988: 83185587Sobrien Fixed lib.c to permit _ in commandline variable names. 83285587Sobrien 83385587SobrienMar 25, 1988: 83485587Sobrien main.c fixed to recognize -- as terminator of command- 83585587Sobrien line options. Illegal options flagged. 83685587Sobrien Error reporting slightly cleaned up. 83785587Sobrien 83885587SobrienDec 2, 1987: 83985587Sobrien Newer C compilers apply a strict scope rule to extern 84085587Sobrien declarations within functions. Two extern declarations in 84185587Sobrien lib.c and tran.c have been moved to obviate this problem. 84285587Sobrien 84385587SobrienOct xx, 1987: 84485587Sobrien Reluctantly added toupper and tolower functions. 84585587Sobrien Subject to rescinding without notice. 84685587Sobrien 84785587SobrienSep 17, 1987: 84885587Sobrien Error-message printer had printf(s) instead of 84985587Sobrien printf("%s",s); got core dumps when the message 85085587Sobrien included a %. 85185587Sobrien 85285587SobrienSep 12, 1987: 85385587Sobrien Very long printf strings caused core dump; 85485587Sobrien fixed aprintf, asprintf, format to catch them. 85585587Sobrien Can still get a core dump in printf itself. 85685587Sobrien 85785587Sobrien 858