FIXES revision 146299
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 28146299SruApr 24, 2005: 29146299Sru modified lib.c so that values of $0 et al are preserved in the END 30146299Sru block, apparently as required by posix. thanks to havard eidnes 31146299Sru for the report and code. 32146299Sru 33146299SruJan 14, 2005: 34146299Sru fixed infinite loop in parsing, originally found by brian tsang. 35146299Sru thanks to arnold robbins for a suggestion that started me 36146299Sru rethinking it. 37146299Sru 38146299SruDec 31, 2004: 39146299Sru prevent overflow of -f array in main, head off potential error in 40146299Sru call of SYNTAX(), test malloc return in lib.c, all with thanks to 41146299Sru todd miller. 42146299Sru 43146299SruDec 22, 2004: 44146299Sru cranked up size of NCHARS; coverity thinks it can be overrun with 45146299Sru smaller size, and i think that's right. added some assertions to b.c 46146299Sru to catch places where it might overrun. the RE code is still fragile. 47146299Sru 48146299SruDec 5, 2004: 49146299Sru fixed a couple of overflow problems with ridiculous field numbers: 50146299Sru e.g., print $(2^32-1). thanks to ruslan ermilov, giorgos keramidas 51146299Sru and david o'brien at freebsd.org for patches. this really should 52146299Sru be re-done from scratch. 53146299Sru 54146299SruNov 21, 2004: 55146299Sru fixed another 25-year-old RE bug, in split. it's another failure 56146299Sru to (re-)initialize. thanks to steve fisher for spotting this and 57146299Sru providing a good test case. 58146299Sru 59125601SruNov 22, 2003: 60125601Sru fixed a bug in regular expressions that dates (so help me) from 1977; 61125601Sru it's been there from the beginning. an anchored longest match that 62125601Sru was longer than the number of states triggered a failure to initialize 63125601Sru the machine properly. many thanks to moinak ghosh for not only finding 64125601Sru this one but for providing a fix, in some of the most mysterious 65125601Sru code known to man. 66125601Sru 67125601Sru fixed a storage leak in call() that appears to have been there since 68125601Sru 1983 or so -- a function without an explicit return that assigns a 69125601Sru string to a parameter leaked a Cell. thanks to moinak ghosh for 70125601Sru spotting this very subtle one. 71125601Sru 72125601SruJul 31, 2003: 73125601Sru fixed, thanks to andrey chernov and ruslan ermilov, a bug in lex.c 74125601Sru that mis-handled the character 255 in input. (it was being compared 75125601Sru to EOF with a signed comparison.) 76125601Sru 77118194SruJul 29, 2003: 78118194Sru fixed (i think) the long-standing botch that included the beginning of 79118194Sru line state ^ for RE's in the set of valid characters; this led to a 80118194Sru variety of odd problems, including failure to properly match certain 81118194Sru regular expressions in non-US locales. thanks to ruslan for keeping 82118194Sru at this one. 83118194Sru 84118194SruJul 28, 2003: 85118194Sru n-th try at getting internationalization right, with thanks to volker 86118194Sru kiefel, arnold robbins and ruslan ermilov for advice, though they 87118194Sru should not be blamed for the outcome. according to posix, "." is the 88118194Sru radix character in programs and command line arguments regardless of 89118194Sru the locale; otherwise, the locale should prevail for input and output 90118194Sru of numbers. so it's intended to work that way. 91118194Sru 92118194Sru i have rescinded the attempt to use strcoll in expanding shorthands in 93118194Sru regular expressions (cclenter). its properties are much too 94118194Sru surprising; for example [a-c] matches aAbBc in locale en_US but abBcC 95118194Sru in locale fr_CA. i can see how this might arise by implementation 96118194Sru but i cannot explain it to a human user. (this behavior can be seen 97118194Sru in gawk as well; we're leaning on the same library.) 98118194Sru 99118194Sru the issue appears to be that strcoll is meant for sorting, where 100118194Sru merging upper and lower case may make sense (though note that unix 101118194Sru sort does not do this by default either). it is not appropriate 102118194Sru for regular expressions, where the goal is to match specific 103118194Sru patterns of characters. in any case, the notations [:lower:], etc., 104118194Sru are available in awk, and they are more likely to work correctly in 105118194Sru most locales. 106118194Sru 107118194Sru a moratorium is hereby declared on internationalization changes. 108118194Sru i apologize to friends and colleagues in other parts of the world. 109118194Sru i would truly like to get this "right", but i don't know what 110118194Sru that is, and i do not want to keep making changes until it's clear. 111118194Sru 112118194SruJul 4, 2003: 113118194Sru fixed bug that permitted non-terminated RE, as in "awk /x". 114118194Sru 115118194SruJun 1, 2003: 116118194Sru subtle change to split: if source is empty, number of elems 117118194Sru is always 0 and the array is not set. 118118194Sru 119118194SruMar 21, 2003: 120118194Sru added some parens to isblank, in another attempt to make things 121118194Sru internationally portable. 122118194Sru 123112336SobrienMar 14, 2003: 124112336Sobrien the internationalization changes, somewhat modified, are now 125112336Sobrien reinstated. in theory awk will now do character comparisons 126112336Sobrien and case conversions in national language, but "." will always 127112336Sobrien be the decimal point separator on input and output regardless 128112336Sobrien of national language. isblank(){} has an #ifndef. 129112336Sobrien 130112336Sobrien this no longer compiles on windows: LC_MESSAGES isn't defined 131112336Sobrien in vc6++. 132112336Sobrien 133112336Sobrien fixed subtle behavior in field and record splitting: if FS is 134112336Sobrien a single character and RS is not empty, \n is NOT a separator. 135112336Sobrien this tortuous reading is found in the awk book; behavior now 136112336Sobrien matches gawk and mawk. 137112336Sobrien 138108072SobrienDec 13, 2002: 139108072Sobrien for the moment, the internationalization changes of nov 29 are 140108072Sobrien rolled back -- programs like x = 1.2 don't work in some locales, 141108072Sobrien because the parser is expecting x = 1,2. until i understand this 142108072Sobrien better, this will have to wait. 143108072Sobrien 144107806SobrienNov 29, 2002: 145107806Sobrien modified b.c (with tiny changes in main and run) to support 146107806Sobrien locales, using strcoll and iswhatever tests for posix character 147107806Sobrien classes. thanks to ruslan ermilov (ru@freebsd.org) for code. 148107806Sobrien the function isblank doesn't seem to have propagated to any 149107806Sobrien header file near me, so it's there explicitly. not properly 150107806Sobrien tested on non-ascii character sets by me. 151107806Sobrien 152107806SobrienJun 28, 2002: 153107806Sobrien modified run/format() and tran/getsval() to do a slightly better 154107806Sobrien job on using OFMT for output from print and CONVFMT for other 155107806Sobrien number->string conversions, as promised by posix and done by 156107806Sobrien gawk and mawk. there are still places where it doesn't work 157107806Sobrien right if CONVFMT is changed; by then the STR attribute of the 158107806Sobrien variable has been irrevocably set. thanks to arnold robbins for 159107806Sobrien code and examples. 160107806Sobrien 161107806Sobrien fixed subtle bug in format that could get core dump. thanks to 162107806Sobrien Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing. 163107806Sobrien minor cleanup in run.c / format() at the same time. 164107806Sobrien 165107806Sobrien added some tests for null pointers to debugging printf's, which 166107806Sobrien were never intended for external consumption. thanks to dave 167107806Sobrien kerns (dkerns@lucent.com) for pointing this out. 168107806Sobrien 169107806Sobrien GNU compatibility: an empty regexp matches anything (thanks to 170107806Sobrien dag-erling smorgrav, des@ofug.org). subject to reversion if 171107806Sobrien this does more harm than good. 172107806Sobrien 173107806Sobrien pervasive small changes to make things more const-correct, as 174107806Sobrien reported by gcc's -Wwrite-strings. as it says in the gcc manual, 175107806Sobrien this may be more nuisance than useful. provoked by a suggestion 176107806Sobrien and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk 177107806Sobrien 178107806Sobrien minor documentation changes to note that this now compiles out 179107806Sobrien of the box on Mac OS X. 180107806Sobrien 18190902SdesFeb 10, 2002: 18290902Sdes changed types in posix chars structure to quiet solaris cc. 18390902Sdes 18490902SdesJan 1, 2002: 18590902Sdes fflush() or fflush("") flushes all files and pipes. 18690902Sdes 18790902Sdes length(arrayname) returns number of elements; thanks to 18890902Sdes arnold robbins for suggestion. 18990902Sdes 19090902Sdes added a makefile.win to make it easier to build on windows. 19190902Sdes based on dan allen's buildwin.bat. 19290902Sdes 19390902SdesNov 16, 2001: 19490902Sdes added support for posix character class names like [:digit:], 19590902Sdes which are not exactly shorter than [0-9] and perhaps no more 19690902Sdes portable. thanks to dag-erling smorgrav for code. 19790902Sdes 19890902SdesFeb 16, 2001: 19990902Sdes removed -m option; no longer needed, and it was actually 20090902Sdes broken (noted thanks to volker kiefel). 20190902Sdes 20290902SdesFeb 10, 2001: 20390902Sdes fixed an appalling bug in gettok: any sequence of digits, +,-, E, e, 20490902Sdes and period was accepted as a valid number if it started with a period. 20590902Sdes this would never have happened with the lex version. 20690902Sdes 20790902Sdes other 1-character botches, now fixed, include a bare $ and a 20890902Sdes bare " at the end of the input. 20990902Sdes 21090902SdesFeb 7, 2001: 21190902Sdes more (const char *) casts in b.c and tran.c to silence warnings. 21290902Sdes 21385587SobrienNov 15, 2000: 21485587Sobrien fixed a bug introduced in august 1997 that caused expressions 21585587Sobrien like $f[1] to be syntax errors. thanks to arnold robbins for 21685587Sobrien noticing this and providing a fix. 21785587Sobrien 21885587SobrienOct 30, 2000: 21985587Sobrien fixed some nextfile bugs: not handling all cases. thanks to 22085587Sobrien arnold robbins for pointing this out. new regressions added. 22185587Sobrien 22285587Sobrien close() is now a function. it returns whatever the library 22385587Sobrien fclose returns, and -1 for closing a file or pipe that wasn't 22485587Sobrien opened. 22585587Sobrien 22685587SobrienSep 24, 2000: 22785587Sobrien permit \n explicitly in character classes; won't work right 22885587Sobrien if comes in as "[\n]" but ok as /[\n]/, because of multiple 22985587Sobrien processing of \'s. thanks to arnold robbins. 23085587Sobrien 23185587SobrienJuly 5, 2000: 23285587Sobrien minor fiddles in tran.c to keep compilers happy about uschar. 23385587Sobrien thanks to norman wilson. 23485587Sobrien 23585587SobrienMay 25, 2000: 23685587Sobrien yet another attempt at making 8-bit input work, with another 23785587Sobrien band-aid in b.c (member()), and some (uschar) casts to head 23885587Sobrien off potential errors in subscripts (like isdigit). also 23985587Sobrien changed HAT to NCHARS-2. thanks again to santiago vila. 24085587Sobrien 24185587Sobrien changed maketab.c to ignore apparently out of range definitions 24285587Sobrien instead of halting; new freeBSD generates one. thanks to 24385587Sobrien jon snader <jsnader@ix.netcom.com> for pointing out the problem. 24485587Sobrien 24585587SobrienMay 2, 2000: 24685587Sobrien fixed an 8-bit problem in b.c by making several char*'s into 24785587Sobrien unsigned char*'s. not clear i have them all yet. thanks to 24885587Sobrien Santiago Vila <sanvila@unex.es> for the bug report. 24985587Sobrien 25085587SobrienApr 21, 2000: 25185587Sobrien finally found and fixed a memory leak in function call; it's 25285587Sobrien been there since functions were added ~1983. thanks to 25385587Sobrien jon bentley for the test case that found it. 25485587Sobrien 25585587Sobrien added test in envinit to catch environment "variables" with 256146299Sru names beginning with '='; thanks to Berend Hasselman. 25785587Sobrien 25885587SobrienJul 28, 1999: 25985587Sobrien added test in defn() to catch function foo(foo), which 26085587Sobrien otherwise recurses until core dump. thanks to arnold 26185587Sobrien robbins for noticing this. 26285587Sobrien 26385587SobrienJun 20, 1999: 26485587Sobrien added *bp in gettok in lex.c; appears possible to exit function 26585587Sobrien without terminating the string. thanks to russ cox. 26685587Sobrien 26785587SobrienJun 2, 1999: 26885587Sobrien added function stdinit() to run to initialize files[] array, 26985587Sobrien in case stdin, etc., are not constants; some compilers care. 27085587Sobrien 27185587SobrienMay 10, 1999: 27285587Sobrien replaced the ERROR ... FATAL, etc., macros with functions 27385587Sobrien based on vprintf, to avoid problems caused by overrunning 27485587Sobrien fixed-size errbuf array. thanks to ralph corderoy for the 27585587Sobrien impetus, and for pointing out a string termination bug in 27685587Sobrien qstring as well. 27785587Sobrien 27885587SobrienApr 21, 1999: 27985587Sobrien fixed bug that caused occasional core dumps with commandline 28085587Sobrien variable with value ending in \. (thanks to nelson beebe for 28185587Sobrien the test case.) 28285587Sobrien 28385587SobrienApr 16, 1999: 28485587Sobrien with code kindly provided by Bruce Lilly, awk now parses 28585587Sobrien /=/ and similar constructs more sensibly in more places. 28685587Sobrien Bruce also provided some helpful test cases. 28785587Sobrien 28885587SobrienApr 5, 1999: 28985587Sobrien changed true/false to True/False in run.c to make it 29085587Sobrien easier to compile with C++. Added some casts on malloc 29185587Sobrien and realloc to be honest about casts; ditto. changed 29285587Sobrien ltype int to long in struct rrow to reduce some 64-bit 29385587Sobrien complaints; other changes scattered throughout for the 29485587Sobrien same purpose. thanks to Nelson Beebe for these portability 29585587Sobrien improvements. 29685587Sobrien 29785587Sobrien removed some horrible pointer-int casting in b.c and elsewhere 29885587Sobrien by adding ptoi and itonp to localize the casts, which are 29985587Sobrien all benign. fixed one incipient bug that showed up on sgi 30085587Sobrien in 64-bit mode. 30185587Sobrien 30285587Sobrien reset lineno for new source file; include filename in error 30385587Sobrien message. also fixed line number error in continuation lines. 30485587Sobrien (thanks to Nelson Beebe for both of these.) 30585587Sobrien 30685587SobrienMar 24, 1999: 30785587Sobrien Nelson Beebe notes that irix 5.3 yacc dies with a bogus 30885587Sobrien error; use a newer version or switch to bison, since sgi 30985587Sobrien is unlikely to fix it. 31085587Sobrien 31185587SobrienMar 5, 1999: 31285587Sobrien changed isnumber to is_number to avoid the problem caused by 31385587Sobrien versions of ctype.h that include the name isnumber. 31485587Sobrien 31585587Sobrien distribution now includes a script for building on a Mac, 31685587Sobrien thanks to Dan Allen. 31785587Sobrien 31885587SobrienFeb 20, 1999: 31985587Sobrien fixed memory leaks in run.c (call) and tran.c (setfval). 32085587Sobrien thanks to Stephen Nutt for finding these and providing the fixes. 32185587Sobrien 32285587SobrienJan 13, 1999: 32385587Sobrien replaced srand argument by (unsigned int) in run.c; 32485587Sobrien avoids problem on Mac and potentially on Unix & Windows. 32585587Sobrien thanks to Dan Allen. 32685587Sobrien 32785587Sobrien added a few (int) casts to silence useless compiler warnings. 32885587Sobrien e.g., errorflag= in run.c jump(). 32985587Sobrien 33085587Sobrien added proctab.c to the bundle outout; one less thing 33185587Sobrien to have to compile out of the box. 33285587Sobrien 33385587Sobrien added calls to _popen and _pclose to the win95 stub for 33485587Sobrien pipes (thanks to Steve Adams for this helpful suggestion). 33585587Sobrien seems to work, though properties are not well understood 33685587Sobrien by me, and it appears that under some circumstances the 33785587Sobrien pipe output is truncated. Be careful. 33885587Sobrien 33985587SobrienOct 19, 1998: 34085587Sobrien fixed a couple of bugs in getrec: could fail to update $0 34185587Sobrien after a getline var; because inputFS wasn't initialized, 34285587Sobrien could split $0 on every character, a misleading diversion. 34385587Sobrien 34485587Sobrien fixed caching bug in makedfa: LRU was actually removing 34585587Sobrien least often used. 34685587Sobrien 34785587Sobrien thanks to ross ridge for finding these, and for providing 34885587Sobrien great bug reports. 34985587Sobrien 35085587SobrienMay 12, 1998: 35185587Sobrien fixed potential bug in readrec: might fail to update record 35285587Sobrien pointer after growing. thanks to dan levy for spotting this 35385587Sobrien and suggesting the fix. 35485587Sobrien 35585587SobrienMar 12, 1998: 35685587Sobrien added -V to print version number and die. 35785587Sobrien 35885587SobrienFeb 11, 1998: 35985587Sobrien subtle silent bug in lex.c: if the program ended with a number 36085587Sobrien longer than 1 digit, part of the input would be pushed back and 36185587Sobrien parsed again because token buffer wasn't terminated right. 36285587Sobrien example: awk 'length($0) > 10'. blush. at least i found it 36385587Sobrien myself. 36485587Sobrien 36585587SobrienAug 31, 1997: 36685587Sobrien s/adelete/awkdelete/: SGI uses this in malloc.h. 36785587Sobrien thanks to nelson beebe for pointing this one out. 36885587Sobrien 36985587SobrienAug 21, 1997: 37085587Sobrien fixed some bugs in sub and gsub when replacement includes \\. 37185587Sobrien this is a dark, horrible corner, but at least now i believe that 37285587Sobrien the behavior is the same as gawk and the intended posix standard. 37385587Sobrien thanks to arnold robbins for advice here. 37485587Sobrien 37585587SobrienAug 9, 1997: 37685587Sobrien somewhat regretfully, replaced the ancient lex-based lexical 37785587Sobrien analyzer with one written in C. it's longer, generates less code, 37885587Sobrien and more portable; the old one depended too much on mysterious 37985587Sobrien properties of lex that were not preserved in other environments. 38085587Sobrien in theory these recognize the same language. 38185587Sobrien 38285587Sobrien now using strtod to test whether a string is a number, instead of 38385587Sobrien the convoluted original function. should be more portable and 38485587Sobrien reliable if strtod is implemented right. 38585587Sobrien 38685587Sobrien removed now-pointless optimization in makefile that tries to avoid 38785587Sobrien recompilation when awkgram.y is changed but symbols are not. 38885587Sobrien 38985587Sobrien removed most fixed-size arrays, though a handful remain, some 39085587Sobrien of which are unchecked. you have been warned. 39185587Sobrien 39285587SobrienAug 4, 1997: 39385587Sobrien with some trepidation, replaced the ancient code that managed 39485587Sobrien fields and $0 in fixed-size arrays with arrays that grow on 39585587Sobrien demand. there is still some tension between trying to make this 39685587Sobrien run fast and making it clean; not sure it's right yet. 39785587Sobrien 39885587Sobrien the ill-conceived -mr and -mf arguments are now useful only 39985587Sobrien for debugging. previous dynamic string code removed. 40085587Sobrien 40185587Sobrien numerous other minor cleanups along the way. 40285587Sobrien 40385587SobrienJul 30, 1997: 40485587Sobrien using code provided by dan levy (to whom profuse thanks), replaced 40585587Sobrien fixed-size arrays and awkward kludges by a fairly uniform mechanism 40685587Sobrien to grow arrays as needed for printf, sub, gsub, etc. 40785587Sobrien 40885587SobrienJul 23, 1997: 40985587Sobrien falling off the end of a function returns "" and 0, not 0. 41085587Sobrien thanks to arnold robbins. 41185587Sobrien 41285587SobrienJun 17, 1997: 41385587Sobrien replaced several fixed-size arrays by dynamically-created ones 41485587Sobrien in run.c; added overflow tests to some previously unchecked cases. 41585587Sobrien getline, toupper, tolower. 41685587Sobrien 41785587Sobrien getline code is still broken in that recursive calls may wind 41885587Sobrien up using the same space. [fixed later] 41985587Sobrien 42085587Sobrien increased RECSIZE to 8192 to push problems further over the horizon. 42185587Sobrien 42285587Sobrien added \r to \n as input line separator for programs, not data. 42385587Sobrien damn CRLFs. 42485587Sobrien 42585587Sobrien modified format() to permit explicit printf("%c", 0) to include 42685587Sobrien a null byte in output. thanks to ken stailey for the fix. 42785587Sobrien 42885587Sobrien added a "-safe" argument that disables file output (print >, 42985587Sobrien print >>), process creation (cmd|getline, print |, system), and 43085587Sobrien access to the environment (ENVIRON). this is a first approximation 43185587Sobrien to a "safe" version of awk, but don't rely on it too much. thanks 43285587Sobrien to joan feigenbaum and matt blaze for the inspiration long ago. 43385587Sobrien 43485587SobrienJul 8, 1996: 43585587Sobrien fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to 43685587Sobrien ralph corderoy. 43785587Sobrien 43885587SobrienJun 29, 1996: 43985587Sobrien fixed awful bug in new field splitting; didn't get all the places 44085587Sobrien where input was done. 44185587Sobrien 44285587SobrienJun 28, 1996: 44385587Sobrien changed field-splitting to conform to posix definition: fields are 44485587Sobrien split using the value of FS at the time of input; it used to be 44585587Sobrien the value when the field or NF was first referred to, a much less 44685587Sobrien predictable definition. thanks to arnold robbins for encouragement 44785587Sobrien to do the right thing. 44885587Sobrien 44985587SobrienMay 28, 1996: 45085587Sobrien fixed appalling but apparently unimportant bug in parsing octal 45185587Sobrien numbers in reg exprs. 45285587Sobrien 45385587Sobrien explicit hex in reg exprs now limited to 2 chars: \xa, \xaa. 45485587Sobrien 45585587SobrienMay 27, 1996: 45685587Sobrien cleaned up some declarations so gcc -Wall is now almost silent. 45785587Sobrien 45885587Sobrien makefile now includes backup copies of ytab.c and lexyy.c in case 45985587Sobrien one makes before looking; it also avoids recreating lexyy.c unless 46085587Sobrien really needed. 46185587Sobrien 46285587Sobrien s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes 46385587Sobrien with unwisely-written header files. 46485587Sobrien 46585587Sobrien thanks to jeffrey friedl for several of these. 46685587Sobrien 46785587SobrienMay 26, 1996: 46885587Sobrien an attempt to rationalize the (unsigned) char issue. almost all 46985587Sobrien instances of unsigned char have been removed; the handful of places 47085587Sobrien in b.c where chars are used as table indices have been hand-crafted. 47185587Sobrien added some latin-1 tests to the regression, but i'm not confident; 47285587Sobrien none of my compilers seem to care much. thanks to nelson beebe for 47385587Sobrien pointing out some others that do care. 47485587Sobrien 47585587SobrienMay 2, 1996: 47685587Sobrien removed all register declarations. 47785587Sobrien 47885587Sobrien enhanced split(), as in gawk, etc: split(s, a, "") splits s into 47985587Sobrien a[1]...a[length(s)] with each character a single element. 48085587Sobrien 48185587Sobrien made the same changes for field-splitting if FS is "". 48285587Sobrien 48385587Sobrien added nextfile, as in gawk: causes immediate advance to next 48485587Sobrien input file. (thanks to arnold robbins for inspiration and code). 48585587Sobrien 48685587Sobrien small fixes to regexpr code: can now handle []], [[], and 48785587Sobrien variants; [] is now a syntax error, rather than matching 48885587Sobrien everything; [z-a] is now empty, not z. far from complete 48985587Sobrien or correct, however. (thanks to jeffrey friedl for pointing out 49085587Sobrien some awful behaviors.) 49185587Sobrien 49285587SobrienApr 29, 1996: 493146299Sru replaced uchar by uschar everywhere; apparently some compilers 49485587Sobrien usurp this name and this causes conflicts. 49585587Sobrien 49685587Sobrien fixed call to time in run.c (bltin); arg is time_t *. 49785587Sobrien 49885587Sobrien replaced horrible pointer/long punning in b.c by a legitimate 49985587Sobrien union. should be safer on 64-bit machines and cleaner everywhere. 50085587Sobrien (thanks to nelson beebe for pointing out some of these problems.) 50185587Sobrien 50285587Sobrien replaced nested comments by #if 0...#endif in run.c, lib.c. 50385587Sobrien 50485587Sobrien removed getsval, setsval, execute macros from run.c and lib.c. 50585587Sobrien machines are 100x faster than they were when these macros were 50685587Sobrien first used. 50785587Sobrien 50885587Sobrien revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l, 50985587Sobrien y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of 51085587Sobrien portability to nameless systems. 51185587Sobrien 51285587Sobrien "make bundle" now includes yacc and lex output files for recipients 51385587Sobrien who don't have yacc or lex. 51485587Sobrien 51585587SobrienAug 15, 1995: 51685587Sobrien initialized Cells in setsymtab more carefully; some fields 51785587Sobrien were not set. (thanks to purify, all of whose complaints i 51885587Sobrien think i now understand.) 51985587Sobrien 52085587Sobrien fixed at least one error in gsub that looked at -1-th element 52185587Sobrien of an array when substituting for a null match (e.g., $). 52285587Sobrien 52385587Sobrien delete arrayname is now legal; it clears the elements but leaves 52485587Sobrien the array, which may not be the right behavior. 52585587Sobrien 52685587Sobrien modified makefile: my current make can't cope with the test used 52785587Sobrien to avoid unnecessary yacc invocations. 52885587Sobrien 52985587SobrienJul 17, 1995: 53085587Sobrien added dynamically growing strings to awk.lx.l and b.c 53185587Sobrien to permit regular expressions to be much bigger. 53285587Sobrien the state arrays can still overflow. 53385587Sobrien 53485587SobrienAug 24, 1994: 53585587Sobrien detect duplicate arguments in function definitions (mdm). 53685587Sobrien 53785587SobrienMay 11, 1994: 53885587Sobrien trivial fix to printf to limit string size in sub(). 53985587Sobrien 54085587SobrienApr 22, 1994: 54185587Sobrien fixed yet another subtle self-assignment problem: 54285587Sobrien $1 = $2; $1 = $1 clobbered $1. 54385587Sobrien 54485587Sobrien Regression tests now use private echo, to avoid quoting problems. 54585587Sobrien 54685587SobrienFeb 2, 1994: 54785587Sobrien changed error() to print line number as %d, not %g. 54885587Sobrien 54985587SobrienJul 23, 1993: 55085587Sobrien cosmetic changes: increased sizes of some arrays, 55185587Sobrien reworded some error messages. 55285587Sobrien 55385587Sobrien added CONVFMT as in posix (just replaced OFMT in getsval) 55485587Sobrien 55585587Sobrien FILENAME is now "" until the first thing that causes a file 55685587Sobrien to be opened. 55785587Sobrien 55885587SobrienNov 28, 1992: 55985587Sobrien deleted yyunput and yyoutput from proto.h; 56085587Sobrien different versions of lex give these different declarations. 56185587Sobrien 56285587SobrienMay 31, 1992: 56385587Sobrien added -mr N and -mf N options: more record and fields. 56485587Sobrien these really ought to adjust automatically. 56585587Sobrien 56685587Sobrien cleaned up some error messages; "out of space" now means 56785587Sobrien malloc returned NULL in all cases. 56885587Sobrien 56985587Sobrien changed rehash so that if it runs out, it just returns; 57085587Sobrien things will continue to run slow, but maybe a bit longer. 57185587Sobrien 57285587SobrienApr 24, 1992: 57385587Sobrien remove redundant close of stdin when using -f -. 57485587Sobrien 57585587Sobrien got rid of core dump with -d; awk -d just prints date. 57685587Sobrien 57785587SobrienApr 12, 1992: 57885587Sobrien added explicit check for /dev/std(in,out,err) in redirection. 57985587Sobrien unlike gawk, no /dev/fd/n yet. 58085587Sobrien 58185587Sobrien added (file/pipe) builtin. hard to test satisfactorily. 58285587Sobrien not posix. 58385587Sobrien 58485587SobrienFeb 20, 1992: 58585587Sobrien recompile after abortive changes; should be unchanged. 58685587Sobrien 58785587SobrienDec 2, 1991: 58885587Sobrien die-casting time: converted to ansi C, installed that. 58985587Sobrien 59085587SobrienNov 30, 1991: 59185587Sobrien fixed storage leak in freefa, failing to recover [N]CCL. 59285587Sobrien thanks to Bill Jones (jones@cs.usask.ca) 59385587Sobrien 59485587SobrienNov 19, 1991: 59585587Sobrien use RAND_MAX instead of literal in builtin(). 59685587Sobrien 59785587SobrienNov 12, 1991: 59885587Sobrien cranked up some fixed-size arrays in b.c, and added a test for 59985587Sobrien overflow in penter. thanks to mark larsen. 60085587Sobrien 60185587SobrienSep 24, 1991: 60285587Sobrien increased buffer in gsub. a very crude fix to a general problem. 60385587Sobrien and again on Sep 26. 60485587Sobrien 60585587SobrienAug 18, 1991: 60685587Sobrien enforce variable name syntax for commandline variables: has to 60785587Sobrien start with letter or _. 60885587Sobrien 60985587SobrienJul 27, 1991: 61085587Sobrien allow newline after ; in for statements. 61185587Sobrien 61285587SobrienJul 21, 1991: 61385587Sobrien fixed so that in self-assignment like $1=$1, side effects 61485587Sobrien like recomputing $0 take place. (this is getting subtle.) 61585587Sobrien 61685587SobrienJun 30, 1991: 61785587Sobrien better test for detecting too-long output record. 61885587Sobrien 61985587SobrienJun 2, 1991: 62085587Sobrien better defense against very long printf strings. 62185587Sobrien made break and continue illegal outside of loops. 62285587Sobrien 62385587SobrienMay 13, 1991: 62485587Sobrien removed extra arg on gettemp, tempfree. minor error message rewording. 62585587Sobrien 62685587SobrienMay 6, 1991: 62785587Sobrien fixed silly bug in hex parsing in hexstr(). 62885587Sobrien removed an apparently unnecessary test in isnumber(). 62985587Sobrien warn about weird printf conversions. 63085587Sobrien fixed unchecked array overwrite in relex(). 63185587Sobrien 63285587Sobrien changed for (i in array) to access elements in sorted order. 63385587Sobrien then unchanged it -- it really does run slower in too many cases. 63485587Sobrien left the code in place, commented out. 63585587Sobrien 63685587SobrienFeb 10, 1991: 63785587Sobrien check error status on all writes, to avoid banging on full disks. 63885587Sobrien 63985587SobrienJan 28, 1991: 64085587Sobrien awk -f - reads the program from stdin. 64185587Sobrien 64285587SobrienJan 11, 1991: 64385587Sobrien failed to set numeric state on $0 in cmd|getline context in run.c. 64485587Sobrien 64585587SobrienNov 2, 1990: 64685587Sobrien fixed sleazy test for integrality in getsval; use modf. 64785587Sobrien 64885587SobrienOct 29, 1990: 64985587Sobrien fixed sleazy buggy code in lib.c that looked (incorrectly) for 65085587Sobrien too long input lines. 65185587Sobrien 65285587SobrienOct 14, 1990: 65385587Sobrien fixed the bug on p. 198 in which it couldn't deduce that an 65485587Sobrien argument was an array in some contexts. replaced the error 65585587Sobrien message in intest() by code that damn well makes it an array. 65685587Sobrien 65785587SobrienOct 8, 1990: 65885587Sobrien fixed horrible bug: types and values were not preserved in 65985587Sobrien some kinds of self-assignment. (in assign().) 66085587Sobrien 66185587SobrienAug 24, 1990: 66285587Sobrien changed NCHARS to 256 to handle 8-bit characters in strings 66385587Sobrien presented to match(), etc. 66485587Sobrien 66585587SobrienJun 26, 1990: 66685587Sobrien changed struct rrow (awk.h) to use long instead of int for lval, 66785587Sobrien since cfoll() stores a pointer in it. now works better when int's 66885587Sobrien are smaller than pointers! 66985587Sobrien 67085587SobrienMay 6, 1990: 67185587Sobrien AVA fixed the grammar so that ! is uniformly of the same precedence as 67285587Sobrien unary + and -. This renders illegal some constructs like !x=y, which 67385587Sobrien now has to be parenthesized as !(x=y), and makes others work properly: 67485587Sobrien !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements. 67585587Sobrien (These problems were pointed out by Bob Lenk of Posix.) 67685587Sobrien 67785587Sobrien Added \x to regular expressions (already in strings). 67885587Sobrien Limited octal to octal digits; \8 and \9 are not octal. 67985587Sobrien Centralized the code for parsing escapes in regular expressions. 68085587Sobrien Added a bunch of tests to T.re and T.sub to verify some of this. 68185587Sobrien 68285587SobrienFeb 9, 1990: 68385587Sobrien fixed null pointer dereference bug in main.c: -F[nothing]. sigh. 68485587Sobrien 68585587Sobrien restored srand behavior: it returns the current seed. 68685587Sobrien 68785587SobrienJan 18, 1990: 68885587Sobrien srand now returns previous seed value (0 to start). 68985587Sobrien 69085587SobrienJan 5, 1990: 69185587Sobrien fix potential problem in tran.c -- something was freed, 69285587Sobrien then used in freesymtab. 69385587Sobrien 69485587SobrienOct 18, 1989: 69585587Sobrien another try to get the max number of open files set with 69685587Sobrien relatively machine-independent code. 69785587Sobrien 69885587Sobrien small fix to input() in case of multiple reads after EOF. 69985587Sobrien 70085587SobrienOct 11, 1989: 70185587Sobrien FILENAME is now defined in the BEGIN block -- too many old 70285587Sobrien programs broke. 70385587Sobrien 70485587Sobrien "-" means stdin in getline as well as on the commandline. 70585587Sobrien 70685587Sobrien added a bunch of casts to the code to tell the truth about 70785587Sobrien char * vs. unsigned char *, a right royal pain. added a 70885587Sobrien setlocale call to the front of main, though probably no one 70985587Sobrien has it usefully implemented yet. 71085587Sobrien 71185587SobrienAug 24, 1989: 71285587Sobrien removed redundant relational tests against nullnode if parse 71385587Sobrien tree already had a relational at that point. 71485587Sobrien 71585587SobrienAug 11, 1989: 71685587Sobrien fixed bug: commandline variable assignment has to look like 71785587Sobrien var=something. (consider the man page for =, in file =.1) 71885587Sobrien 71985587Sobrien changed number of arguments to functions to static arrays 72085587Sobrien to avoid repeated malloc calls. 72185587Sobrien 72285587SobrienAug 2, 1989: 72385587Sobrien restored -F (space) separator 72485587Sobrien 72585587SobrienJul 30, 1989: 72685587Sobrien added -v x=1 y=2 ... for immediate commandline variable assignment; 72785587Sobrien done before the BEGIN block for sure. they have to precede the 72885587Sobrien program if the program is on the commandline. 72985587Sobrien Modified Aug 2 to require a separate -v for each assignment. 73085587Sobrien 73185587SobrienJul 10, 1989: 73285587Sobrien fixed ref-thru-zero bug in environment code in tran.c 73385587Sobrien 73485587SobrienJun 23, 1989: 73585587Sobrien add newline to usage message. 73685587Sobrien 73785587SobrienJun 14, 1989: 73885587Sobrien added some missing ansi printf conversion letters: %i %X %E %G. 73985587Sobrien no sensible meaning for h or L, so they may not do what one expects. 74085587Sobrien 74185587Sobrien made %* conversions work. 74285587Sobrien 74385587Sobrien changed x^y so that if n is a positive integer, it's done 74485587Sobrien by explicit multiplication, thus achieving maximum accuracy. 74585587Sobrien (this should be done by pow() but it seems not to be locally.) 74685587Sobrien done to x ^= y as well. 74785587Sobrien 74885587SobrienJun 4, 1989: 74985587Sobrien ENVIRON array contains environment: if shell variable V=thing, 75085587Sobrien ENVIRON["V"] is "thing" 75185587Sobrien 75285587Sobrien multiple -f arguments permitted. error reporting is naive. 75385587Sobrien (they were permitted before, but only the last was used.) 75485587Sobrien 75585587Sobrien fixed a really stupid botch in the debugging macro dprintf 75685587Sobrien 75785587Sobrien fixed order of evaluation of commandline assignments to match 75885587Sobrien what the book claims: an argument of the form x=e is evaluated 75985587Sobrien at the time it would have been opened if it were a filename (p 63). 76085587Sobrien this invalidates the suggested answer to ex 4-1 (p 195). 76185587Sobrien 76285587Sobrien removed some code that permitted -F (space) fieldseparator, 76385587Sobrien since it didn't quite work right anyway. (restored aug 2) 76485587Sobrien 76585587SobrienApr 27, 1989: 76685587Sobrien Line number now accumulated correctly for comment lines. 76785587Sobrien 76885587SobrienApr 26, 1989: 76985587Sobrien Debugging output now includes a version date, 77085587Sobrien if one compiles it into the source each time. 77185587Sobrien 77285587SobrienApr 9, 1989: 77385587Sobrien Changed grammar to prohibit constants as 3rd arg of sub and gsub; 77485587Sobrien prevents class of overwriting-a-constant errors. (Last one?) 77585587Sobrien This invalidates the "banana" example on page 43 of the book. 77685587Sobrien 77785587Sobrien Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal), 77885587Sobrien as in ANSI, for strings. Rescinded the sloppiness that permitted 77985587Sobrien non-octal digits in \ooo. Warning: not all compilers and libraries 78085587Sobrien will be able to deal with \x correctly. 78185587Sobrien 78285587SobrienJan 9, 1989: 78385587Sobrien Fixed bug that caused tempcell list to contain a duplicate. 78485587Sobrien The fix is kludgy. 78585587Sobrien 78685587SobrienDec 17, 1988: 78785587Sobrien Catches some more commandline errors in main. 78885587Sobrien Removed redundant decl of modf in run.c (confuses some compilers). 78985587Sobrien Warning: there's no single declaration of malloc, etc., in awk.h 79085587Sobrien that seems to satisfy all compilers. 79185587Sobrien 79285587SobrienDec 7, 1988: 79385587Sobrien Added a bit of code to error printing to avoid printing nulls. 79485587Sobrien (Not clear that it actually would.) 79585587Sobrien 79685587SobrienNov 27, 1988: 79785587Sobrien With fear and trembling, modified the grammar to permit 79885587Sobrien multiple pattern-action statements on one line without 79985587Sobrien an explicit separator. By definition, this capitulation 80085587Sobrien to the ghost of ancient implementations remains undefined 80185587Sobrien and thus subject to change without notice or apology. 80285587Sobrien DO NOT COUNT ON IT. 80385587Sobrien 80485587SobrienOct 30, 1988: 80585587Sobrien Fixed bug in call() that failed to recover storage. 80685587Sobrien 80785587Sobrien A warning is now generated if there are more arguments 80885587Sobrien in the call than in the definition (in lieu of fixing 80985587Sobrien another storage leak). 81085587Sobrien 81185587SobrienOct 20, 1988: 81285587Sobrien Fixed %c: if expr is numeric, use numeric value; 81385587Sobrien otherwise print 1st char of string value. still 81485587Sobrien doesn't work if the value is 0 -- won't print \0. 81585587Sobrien 81685587Sobrien Added a few more checks for running out of malloc. 81785587Sobrien 81885587SobrienOct 12, 1988: 81985587Sobrien Fixed bug in call() that freed local arrays twice. 82085587Sobrien 82185587Sobrien Fixed to handle deletion of non-existent array right; 82285587Sobrien complains about attempt to delete non-array element. 82385587Sobrien 82485587SobrienSep 30, 1988: 82585587Sobrien Now guarantees to evaluate all arguments of built-in 82685587Sobrien functions, as in C; the appearance is that arguments 82785587Sobrien are evaluated before the function is called. Places 82885587Sobrien affected are sub (gsub was ok), substr, printf, and 82985587Sobrien all the built-in arithmetic functions in bltin(). 83085587Sobrien A warning is generated if a bltin() is called with 83185587Sobrien the wrong number of arguments. 83285587Sobrien 83385587Sobrien This requires changing makeprof on p167 of the book. 83485587Sobrien 83585587SobrienAug 23, 1988: 83685587Sobrien setting FILENAME in BEGIN caused core dump, apparently 83785587Sobrien because it was freeing space not allocated by malloc. 83885587Sobrien 83985587SobrienJuly 24, 1988: 84085587Sobrien fixed egregious error in toupper/tolower functions. 84185587Sobrien still subject to rescinding, however. 84285587Sobrien 84385587SobrienJuly 2, 1988: 84485587Sobrien flush stdout before opening file or pipe 84585587Sobrien 84685587SobrienJuly 2, 1988: 84785587Sobrien performance bug in b.c/cgoto(): not freeing some sets of states. 84885587Sobrien partial fix only right now, and the number of states increased 84985587Sobrien to make it less obvious. 85085587Sobrien 85185587SobrienJune 1, 1988: 85285587Sobrien check error status on close 85385587Sobrien 85485587SobrienMay 28, 1988: 85585587Sobrien srand returns seed value it's using. 85685587Sobrien see 1/18/90 85785587Sobrien 85885587SobrienMay 22, 1988: 85985587Sobrien Removed limit on depth of function calls. 86085587Sobrien 86185587SobrienMay 10, 1988: 86285587Sobrien Fixed lib.c to permit _ in commandline variable names. 86385587Sobrien 86485587SobrienMar 25, 1988: 86585587Sobrien main.c fixed to recognize -- as terminator of command- 86685587Sobrien line options. Illegal options flagged. 86785587Sobrien Error reporting slightly cleaned up. 86885587Sobrien 86985587SobrienDec 2, 1987: 87085587Sobrien Newer C compilers apply a strict scope rule to extern 87185587Sobrien declarations within functions. Two extern declarations in 87285587Sobrien lib.c and tran.c have been moved to obviate this problem. 87385587Sobrien 87485587SobrienOct xx, 1987: 87585587Sobrien Reluctantly added toupper and tolower functions. 87685587Sobrien Subject to rescinding without notice. 87785587Sobrien 87885587SobrienSep 17, 1987: 87985587Sobrien Error-message printer had printf(s) instead of 88085587Sobrien printf("%s",s); got core dumps when the message 88185587Sobrien included a %. 88285587Sobrien 88385587SobrienSep 12, 1987: 88485587Sobrien Very long printf strings caused core dump; 88585587Sobrien fixed aprintf, asprintf, format to catch them. 88685587Sobrien Can still get a core dump in printf itself. 88785587Sobrien 88885587Sobrien 889