1 /****************************************************************
2 Copyright (C) Lucent Technologies 1997
5 Permission to use, copy, modify, and distribute this software and
6 its documentation for any purpose and without fee is hereby
7 granted, provided that the above copyright notice appear in all
8 copies and that both that the copyright notice and this
9 permission notice and warranty disclaimer appear in supporting
10 documentation, and that the name Lucent Technologies or any of
11 its entities not be used in advertising or publicity pertaining
12 to distribution of the software without specific, written prior
15 LUCENT DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
16 INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS.
17 IN NO EVENT SHALL LUCENT OR ANY OF ITS ENTITIES BE LIABLE FOR ANY
18 SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
19 WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER
20 IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
21 ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
23 ****************************************************************/
25 This file lists all bug fixes, changes, etc., made since the AWK book
26 was sent to the printers in August, 1987.
29 modified lib.c so that values of $0 et al are preserved in the END
30 block, apparently as required by posix. thanks to havard eidnes
31 for the report and code.
34 fixed infinite loop in parsing, originally found by brian tsang.
35 thanks to arnold robbins for a suggestion that started me
39 prevent overflow of -f array in main, head off potential error in
40 call of SYNTAX(), test malloc return in lib.c, all with thanks to
44 cranked up size of NCHARS; coverity thinks it can be overrun with
45 smaller size, and i think that's right. added some assertions to b.c
46 to catch places where it might overrun. the RE code is still fragile.
49 fixed a couple of overflow problems with ridiculous field numbers:
50 e.g., print $(2^32-1). thanks to ruslan ermilov, giorgos keramidas
51 and david o'brien at freebsd.org for patches. this really should
52 be re-done from scratch.
55 fixed another 25-year-old RE bug, in split. it's another failure
56 to (re-)initialize. thanks to steve fisher for spotting this and
57 providing a good test case.
60 fixed a bug in regular expressions that dates (so help me) from 1977;
61 it's been there from the beginning. an anchored longest match that
62 was longer than the number of states triggered a failure to initialize
63 the machine properly. many thanks to moinak ghosh for not only finding
64 this one but for providing a fix, in some of the most mysterious
67 fixed a storage leak in call() that appears to have been there since
68 1983 or so -- a function without an explicit return that assigns a
69 string to a parameter leaked a Cell. thanks to moinak ghosh for
70 spotting this very subtle one.
73 fixed, thanks to andrey chernov and ruslan ermilov, a bug in lex.c
74 that mis-handled the character 255 in input. (it was being compared
75 to EOF with a signed comparison.)
78 fixed (i think) the long-standing botch that included the beginning of
79 line state ^ for RE's in the set of valid characters; this led to a
80 variety of odd problems, including failure to properly match certain
81 regular expressions in non-US locales. thanks to ruslan for keeping
85 n-th try at getting internationalization right, with thanks to volker
86 kiefel, arnold robbins and ruslan ermilov for advice, though they
87 should not be blamed for the outcome. according to posix, "." is the
88 radix character in programs and command line arguments regardless of
89 the locale; otherwise, the locale should prevail for input and output
90 of numbers. so it's intended to work that way.
92 i have rescinded the attempt to use strcoll in expanding shorthands in
93 regular expressions (cclenter). its properties are much too
94 surprising; for example [a-c] matches aAbBc in locale en_US but abBcC
95 in locale fr_CA. i can see how this might arise by implementation
96 but i cannot explain it to a human user. (this behavior can be seen
97 in gawk as well; we're leaning on the same library.)
99 the issue appears to be that strcoll is meant for sorting, where
100 merging upper and lower case may make sense (though note that unix
101 sort does not do this by default either). it is not appropriate
102 for regular expressions, where the goal is to match specific
103 patterns of characters. in any case, the notations [:lower:], etc.,
104 are available in awk, and they are more likely to work correctly in
107 a moratorium is hereby declared on internationalization changes.
108 i apologize to friends and colleagues in other parts of the world.
109 i would truly like to get this "right", but i don't know what
110 that is, and i do not want to keep making changes until it's clear.
113 fixed bug that permitted non-terminated RE, as in "awk /x".
116 subtle change to split: if source is empty, number of elems
117 is always 0 and the array is not set.
120 added some parens to isblank, in another attempt to make things
121 internationally portable.
124 the internationalization changes, somewhat modified, are now
125 reinstated. in theory awk will now do character comparisons
126 and case conversions in national language, but "." will always
127 be the decimal point separator on input and output regardless
128 of national language. isblank(){} has an #ifndef.
130 this no longer compiles on windows: LC_MESSAGES isn't defined
133 fixed subtle behavior in field and record splitting: if FS is
134 a single character and RS is not empty, \n is NOT a separator.
135 this tortuous reading is found in the awk book; behavior now
136 matches gawk and mawk.
139 for the moment, the internationalization changes of nov 29 are
140 rolled back -- programs like x = 1.2 don't work in some locales,
141 because the parser is expecting x = 1,2. until i understand this
142 better, this will have to wait.
145 modified b.c (with tiny changes in main and run) to support
146 locales, using strcoll and iswhatever tests for posix character
147 classes. thanks to ruslan ermilov (ru@freebsd.org) for code.
148 the function isblank doesn't seem to have propagated to any
149 header file near me, so it's there explicitly. not properly
150 tested on non-ascii character sets by me.
153 modified run/format() and tran/getsval() to do a slightly better
154 job on using OFMT for output from print and CONVFMT for other
155 number->string conversions, as promised by posix and done by
156 gawk and mawk. there are still places where it doesn't work
157 right if CONVFMT is changed; by then the STR attribute of the
158 variable has been irrevocably set. thanks to arnold robbins for
161 fixed subtle bug in format that could get core dump. thanks to
162 Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing.
163 minor cleanup in run.c / format() at the same time.
165 added some tests for null pointers to debugging printf's, which
166 were never intended for external consumption. thanks to dave
167 kerns (dkerns@lucent.com) for pointing this out.
169 GNU compatibility: an empty regexp matches anything (thanks to
170 dag-erling smorgrav, des@ofug.org). subject to reversion if
171 this does more harm than good.
173 pervasive small changes to make things more const-correct, as
174 reported by gcc's -Wwrite-strings. as it says in the gcc manual,
175 this may be more nuisance than useful. provoked by a suggestion
176 and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk
178 minor documentation changes to note that this now compiles out
179 of the box on Mac OS X.
182 changed types in posix chars structure to quiet solaris cc.
185 fflush() or fflush("") flushes all files and pipes.
187 length(arrayname) returns number of elements; thanks to
188 arnold robbins for suggestion.
190 added a makefile.win to make it easier to build on windows.
191 based on dan allen's buildwin.bat.
194 added support for posix character class names like [:digit:],
195 which are not exactly shorter than [0-9] and perhaps no more
196 portable. thanks to dag-erling smorgrav for code.
199 removed -m option; no longer needed, and it was actually
200 broken (noted thanks to volker kiefel).
203 fixed an appalling bug in gettok: any sequence of digits, +,-, E, e,
204 and period was accepted as a valid number if it started with a period.
205 this would never have happened with the lex version.
207 other 1-character botches, now fixed, include a bare $ and a
208 bare " at the end of the input.
211 more (const char *) casts in b.c and tran.c to silence warnings.
214 fixed a bug introduced in august 1997 that caused expressions
215 like $f[1] to be syntax errors. thanks to arnold robbins for
216 noticing this and providing a fix.
219 fixed some nextfile bugs: not handling all cases. thanks to
220 arnold robbins for pointing this out. new regressions added.
222 close() is now a function. it returns whatever the library
223 fclose returns, and -1 for closing a file or pipe that wasn't
227 permit \n explicitly in character classes; won't work right
228 if comes in as "[\n]" but ok as /[\n]/, because of multiple
229 processing of \'s. thanks to arnold robbins.
232 minor fiddles in tran.c to keep compilers happy about uschar.
233 thanks to norman wilson.
236 yet another attempt at making 8-bit input work, with another
237 band-aid in b.c (member()), and some (uschar) casts to head
238 off potential errors in subscripts (like isdigit). also
239 changed HAT to NCHARS-2. thanks again to santiago vila.
241 changed maketab.c to ignore apparently out of range definitions
242 instead of halting; new freeBSD generates one. thanks to
243 jon snader <jsnader@ix.netcom.com> for pointing out the problem.
246 fixed an 8-bit problem in b.c by making several char*'s into
247 unsigned char*'s. not clear i have them all yet. thanks to
248 Santiago Vila <sanvila@unex.es> for the bug report.
251 finally found and fixed a memory leak in function call; it's
252 been there since functions were added ~1983. thanks to
253 jon bentley for the test case that found it.
255 added test in envinit to catch environment "variables" with
256 names beginning with '='; thanks to Berend Hasselman.
259 added test in defn() to catch function foo(foo), which
260 otherwise recurses until core dump. thanks to arnold
261 robbins for noticing this.
264 added *bp in gettok in lex.c; appears possible to exit function
265 without terminating the string. thanks to russ cox.
268 added function stdinit() to run to initialize files[] array,
269 in case stdin, etc., are not constants; some compilers care.
272 replaced the ERROR ... FATAL, etc., macros with functions
273 based on vprintf, to avoid problems caused by overrunning
274 fixed-size errbuf array. thanks to ralph corderoy for the
275 impetus, and for pointing out a string termination bug in
279 fixed bug that caused occasional core dumps with commandline
280 variable with value ending in \. (thanks to nelson beebe for
284 with code kindly provided by Bruce Lilly, awk now parses
285 /=/ and similar constructs more sensibly in more places.
286 Bruce also provided some helpful test cases.
289 changed true/false to True/False in run.c to make it
290 easier to compile with C++. Added some casts on malloc
291 and realloc to be honest about casts; ditto. changed
292 ltype int to long in struct rrow to reduce some 64-bit
293 complaints; other changes scattered throughout for the
294 same purpose. thanks to Nelson Beebe for these portability
297 removed some horrible pointer-int casting in b.c and elsewhere
298 by adding ptoi and itonp to localize the casts, which are
299 all benign. fixed one incipient bug that showed up on sgi
302 reset lineno for new source file; include filename in error
303 message. also fixed line number error in continuation lines.
304 (thanks to Nelson Beebe for both of these.)
307 Nelson Beebe notes that irix 5.3 yacc dies with a bogus
308 error; use a newer version or switch to bison, since sgi
309 is unlikely to fix it.
312 changed isnumber to is_number to avoid the problem caused by
313 versions of ctype.h that include the name isnumber.
315 distribution now includes a script for building on a Mac,
319 fixed memory leaks in run.c (call) and tran.c (setfval).
320 thanks to Stephen Nutt for finding these and providing the fixes.
323 replaced srand argument by (unsigned int) in run.c;
324 avoids problem on Mac and potentially on Unix & Windows.
327 added a few (int) casts to silence useless compiler warnings.
328 e.g., errorflag= in run.c jump().
330 added proctab.c to the bundle outout; one less thing
331 to have to compile out of the box.
333 added calls to _popen and _pclose to the win95 stub for
334 pipes (thanks to Steve Adams for this helpful suggestion).
335 seems to work, though properties are not well understood
336 by me, and it appears that under some circumstances the
337 pipe output is truncated. Be careful.
340 fixed a couple of bugs in getrec: could fail to update $0
341 after a getline var; because inputFS wasn't initialized,
342 could split $0 on every character, a misleading diversion.
344 fixed caching bug in makedfa: LRU was actually removing
347 thanks to ross ridge for finding these, and for providing
351 fixed potential bug in readrec: might fail to update record
352 pointer after growing. thanks to dan levy for spotting this
353 and suggesting the fix.
356 added -V to print version number and die.
359 subtle silent bug in lex.c: if the program ended with a number
360 longer than 1 digit, part of the input would be pushed back and
361 parsed again because token buffer wasn't terminated right.
362 example: awk 'length($0) > 10'. blush. at least i found it
366 s/adelete/awkdelete/: SGI uses this in malloc.h.
367 thanks to nelson beebe for pointing this one out.
370 fixed some bugs in sub and gsub when replacement includes \\.
371 this is a dark, horrible corner, but at least now i believe that
372 the behavior is the same as gawk and the intended posix standard.
373 thanks to arnold robbins for advice here.
376 somewhat regretfully, replaced the ancient lex-based lexical
377 analyzer with one written in C. it's longer, generates less code,
378 and more portable; the old one depended too much on mysterious
379 properties of lex that were not preserved in other environments.
380 in theory these recognize the same language.
382 now using strtod to test whether a string is a number, instead of
383 the convoluted original function. should be more portable and
384 reliable if strtod is implemented right.
386 removed now-pointless optimization in makefile that tries to avoid
387 recompilation when awkgram.y is changed but symbols are not.
389 removed most fixed-size arrays, though a handful remain, some
390 of which are unchecked. you have been warned.
393 with some trepidation, replaced the ancient code that managed
394 fields and $0 in fixed-size arrays with arrays that grow on
395 demand. there is still some tension between trying to make this
396 run fast and making it clean; not sure it's right yet.
398 the ill-conceived -mr and -mf arguments are now useful only
399 for debugging. previous dynamic string code removed.
401 numerous other minor cleanups along the way.
404 using code provided by dan levy (to whom profuse thanks), replaced
405 fixed-size arrays and awkward kludges by a fairly uniform mechanism
406 to grow arrays as needed for printf, sub, gsub, etc.
409 falling off the end of a function returns "" and 0, not 0.
410 thanks to arnold robbins.
413 replaced several fixed-size arrays by dynamically-created ones
414 in run.c; added overflow tests to some previously unchecked cases.
415 getline, toupper, tolower.
417 getline code is still broken in that recursive calls may wind
418 up using the same space. [fixed later]
420 increased RECSIZE to 8192 to push problems further over the horizon.
422 added \r to \n as input line separator for programs, not data.
425 modified format() to permit explicit printf("%c", 0) to include
426 a null byte in output. thanks to ken stailey for the fix.
428 added a "-safe" argument that disables file output (print >,
429 print >>), process creation (cmd|getline, print |, system), and
430 access to the environment (ENVIRON). this is a first approximation
431 to a "safe" version of awk, but don't rely on it too much. thanks
432 to joan feigenbaum and matt blaze for the inspiration long ago.
435 fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to
439 fixed awful bug in new field splitting; didn't get all the places
440 where input was done.
443 changed field-splitting to conform to posix definition: fields are
444 split using the value of FS at the time of input; it used to be
445 the value when the field or NF was first referred to, a much less
446 predictable definition. thanks to arnold robbins for encouragement
447 to do the right thing.
450 fixed appalling but apparently unimportant bug in parsing octal
451 numbers in reg exprs.
453 explicit hex in reg exprs now limited to 2 chars: \xa, \xaa.
456 cleaned up some declarations so gcc -Wall is now almost silent.
458 makefile now includes backup copies of ytab.c and lexyy.c in case
459 one makes before looking; it also avoids recreating lexyy.c unless
462 s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes
463 with unwisely-written header files.
465 thanks to jeffrey friedl for several of these.
468 an attempt to rationalize the (unsigned) char issue. almost all
469 instances of unsigned char have been removed; the handful of places
470 in b.c where chars are used as table indices have been hand-crafted.
471 added some latin-1 tests to the regression, but i'm not confident;
472 none of my compilers seem to care much. thanks to nelson beebe for
473 pointing out some others that do care.
476 removed all register declarations.
478 enhanced split(), as in gawk, etc: split(s, a, "") splits s into
479 a[1]...a[length(s)] with each character a single element.
481 made the same changes for field-splitting if FS is "".
483 added nextfile, as in gawk: causes immediate advance to next
484 input file. (thanks to arnold robbins for inspiration and code).
486 small fixes to regexpr code: can now handle []], [[], and
487 variants; [] is now a syntax error, rather than matching
488 everything; [z-a] is now empty, not z. far from complete
489 or correct, however. (thanks to jeffrey friedl for pointing out
490 some awful behaviors.)
493 replaced uchar by uschar everywhere; apparently some compilers
494 usurp this name and this causes conflicts.
496 fixed call to time in run.c (bltin); arg is time_t *.
498 replaced horrible pointer/long punning in b.c by a legitimate
499 union. should be safer on 64-bit machines and cleaner everywhere.
500 (thanks to nelson beebe for pointing out some of these problems.)
502 replaced nested comments by #if 0...#endif in run.c, lib.c.
504 removed getsval, setsval, execute macros from run.c and lib.c.
505 machines are 100x faster than they were when these macros were
508 revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l,
509 y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of
510 portability to nameless systems.
512 "make bundle" now includes yacc and lex output files for recipients
513 who don't have yacc or lex.
516 initialized Cells in setsymtab more carefully; some fields
517 were not set. (thanks to purify, all of whose complaints i
518 think i now understand.)
520 fixed at least one error in gsub that looked at -1-th element
521 of an array when substituting for a null match (e.g., $).
523 delete arrayname is now legal; it clears the elements but leaves
524 the array, which may not be the right behavior.
526 modified makefile: my current make can't cope with the test used
527 to avoid unnecessary yacc invocations.
530 added dynamically growing strings to awk.lx.l and b.c
531 to permit regular expressions to be much bigger.
532 the state arrays can still overflow.
535 detect duplicate arguments in function definitions (mdm).
538 trivial fix to printf to limit string size in sub().
541 fixed yet another subtle self-assignment problem:
542 $1 = $2; $1 = $1 clobbered $1.
544 Regression tests now use private echo, to avoid quoting problems.
547 changed error() to print line number as %d, not %g.
550 cosmetic changes: increased sizes of some arrays,
551 reworded some error messages.
553 added CONVFMT as in posix (just replaced OFMT in getsval)
555 FILENAME is now "" until the first thing that causes a file
559 deleted yyunput and yyoutput from proto.h;
560 different versions of lex give these different declarations.
563 added -mr N and -mf N options: more record and fields.
564 these really ought to adjust automatically.
566 cleaned up some error messages; "out of space" now means
567 malloc returned NULL in all cases.
569 changed rehash so that if it runs out, it just returns;
570 things will continue to run slow, but maybe a bit longer.
573 remove redundant close of stdin when using -f -.
575 got rid of core dump with -d; awk -d just prints date.
578 added explicit check for /dev/std(in,out,err) in redirection.
579 unlike gawk, no /dev/fd/n yet.
581 added (file/pipe) builtin. hard to test satisfactorily.
585 recompile after abortive changes; should be unchanged.
588 die-casting time: converted to ansi C, installed that.
591 fixed storage leak in freefa, failing to recover [N]CCL.
592 thanks to Bill Jones (jones@cs.usask.ca)
595 use RAND_MAX instead of literal in builtin().
598 cranked up some fixed-size arrays in b.c, and added a test for
599 overflow in penter. thanks to mark larsen.
602 increased buffer in gsub. a very crude fix to a general problem.
606 enforce variable name syntax for commandline variables: has to
607 start with letter or _.
610 allow newline after ; in for statements.
613 fixed so that in self-assignment like $1=$1, side effects
614 like recomputing $0 take place. (this is getting subtle.)
617 better test for detecting too-long output record.
620 better defense against very long printf strings.
621 made break and continue illegal outside of loops.
624 removed extra arg on gettemp, tempfree. minor error message rewording.
627 fixed silly bug in hex parsing in hexstr().
628 removed an apparently unnecessary test in isnumber().
629 warn about weird printf conversions.
630 fixed unchecked array overwrite in relex().
632 changed for (i in array) to access elements in sorted order.
633 then unchanged it -- it really does run slower in too many cases.
634 left the code in place, commented out.
637 check error status on all writes, to avoid banging on full disks.
640 awk -f - reads the program from stdin.
643 failed to set numeric state on $0 in cmd|getline context in run.c.
646 fixed sleazy test for integrality in getsval; use modf.
649 fixed sleazy buggy code in lib.c that looked (incorrectly) for
650 too long input lines.
653 fixed the bug on p. 198 in which it couldn't deduce that an
654 argument was an array in some contexts. replaced the error
655 message in intest() by code that damn well makes it an array.
658 fixed horrible bug: types and values were not preserved in
659 some kinds of self-assignment. (in assign().)
662 changed NCHARS to 256 to handle 8-bit characters in strings
663 presented to match(), etc.
666 changed struct rrow (awk.h) to use long instead of int for lval,
667 since cfoll() stores a pointer in it. now works better when int's
668 are smaller than pointers!
671 AVA fixed the grammar so that ! is uniformly of the same precedence as
672 unary + and -. This renders illegal some constructs like !x=y, which
673 now has to be parenthesized as !(x=y), and makes others work properly:
674 !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements.
675 (These problems were pointed out by Bob Lenk of Posix.)
677 Added \x to regular expressions (already in strings).
678 Limited octal to octal digits; \8 and \9 are not octal.
679 Centralized the code for parsing escapes in regular expressions.
680 Added a bunch of tests to T.re and T.sub to verify some of this.
683 fixed null pointer dereference bug in main.c: -F[nothing]. sigh.
685 restored srand behavior: it returns the current seed.
688 srand now returns previous seed value (0 to start).
691 fix potential problem in tran.c -- something was freed,
692 then used in freesymtab.
695 another try to get the max number of open files set with
696 relatively machine-independent code.
698 small fix to input() in case of multiple reads after EOF.
701 FILENAME is now defined in the BEGIN block -- too many old
704 "-" means stdin in getline as well as on the commandline.
706 added a bunch of casts to the code to tell the truth about
707 char * vs. unsigned char *, a right royal pain. added a
708 setlocale call to the front of main, though probably no one
709 has it usefully implemented yet.
712 removed redundant relational tests against nullnode if parse
713 tree already had a relational at that point.
716 fixed bug: commandline variable assignment has to look like
717 var=something. (consider the man page for =, in file =.1)
719 changed number of arguments to functions to static arrays
720 to avoid repeated malloc calls.
723 restored -F (space) separator
726 added -v x=1 y=2 ... for immediate commandline variable assignment;
727 done before the BEGIN block for sure. they have to precede the
728 program if the program is on the commandline.
729 Modified Aug 2 to require a separate -v for each assignment.
732 fixed ref-thru-zero bug in environment code in tran.c
735 add newline to usage message.
738 added some missing ansi printf conversion letters: %i %X %E %G.
739 no sensible meaning for h or L, so they may not do what one expects.
741 made %* conversions work.
743 changed x^y so that if n is a positive integer, it's done
744 by explicit multiplication, thus achieving maximum accuracy.
745 (this should be done by pow() but it seems not to be locally.)
746 done to x ^= y as well.
749 ENVIRON array contains environment: if shell variable V=thing,
750 ENVIRON["V"] is "thing"
752 multiple -f arguments permitted. error reporting is naive.
753 (they were permitted before, but only the last was used.)
755 fixed a really stupid botch in the debugging macro dprintf
757 fixed order of evaluation of commandline assignments to match
758 what the book claims: an argument of the form x=e is evaluated
759 at the time it would have been opened if it were a filename (p 63).
760 this invalidates the suggested answer to ex 4-1 (p 195).
762 removed some code that permitted -F (space) fieldseparator,
763 since it didn't quite work right anyway. (restored aug 2)
766 Line number now accumulated correctly for comment lines.
769 Debugging output now includes a version date,
770 if one compiles it into the source each time.
773 Changed grammar to prohibit constants as 3rd arg of sub and gsub;
774 prevents class of overwriting-a-constant errors. (Last one?)
775 This invalidates the "banana" example on page 43 of the book.
777 Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal),
778 as in ANSI, for strings. Rescinded the sloppiness that permitted
779 non-octal digits in \ooo. Warning: not all compilers and libraries
780 will be able to deal with \x correctly.
783 Fixed bug that caused tempcell list to contain a duplicate.
787 Catches some more commandline errors in main.
788 Removed redundant decl of modf in run.c (confuses some compilers).
789 Warning: there's no single declaration of malloc, etc., in awk.h
790 that seems to satisfy all compilers.
793 Added a bit of code to error printing to avoid printing nulls.
794 (Not clear that it actually would.)
797 With fear and trembling, modified the grammar to permit
798 multiple pattern-action statements on one line without
799 an explicit separator. By definition, this capitulation
800 to the ghost of ancient implementations remains undefined
801 and thus subject to change without notice or apology.
805 Fixed bug in call() that failed to recover storage.
807 A warning is now generated if there are more arguments
808 in the call than in the definition (in lieu of fixing
809 another storage leak).
812 Fixed %c: if expr is numeric, use numeric value;
813 otherwise print 1st char of string value. still
814 doesn't work if the value is 0 -- won't print \0.
816 Added a few more checks for running out of malloc.
819 Fixed bug in call() that freed local arrays twice.
821 Fixed to handle deletion of non-existent array right;
822 complains about attempt to delete non-array element.
825 Now guarantees to evaluate all arguments of built-in
826 functions, as in C; the appearance is that arguments
827 are evaluated before the function is called. Places
828 affected are sub (gsub was ok), substr, printf, and
829 all the built-in arithmetic functions in bltin().
830 A warning is generated if a bltin() is called with
831 the wrong number of arguments.
833 This requires changing makeprof on p167 of the book.
836 setting FILENAME in BEGIN caused core dump, apparently
837 because it was freeing space not allocated by malloc.
840 fixed egregious error in toupper/tolower functions.
841 still subject to rescinding, however.
844 flush stdout before opening file or pipe
847 performance bug in b.c/cgoto(): not freeing some sets of states.
848 partial fix only right now, and the number of states increased
849 to make it less obvious.
852 check error status on close
855 srand returns seed value it's using.
859 Removed limit on depth of function calls.
862 Fixed lib.c to permit _ in commandline variable names.
865 main.c fixed to recognize -- as terminator of command-
866 line options. Illegal options flagged.
867 Error reporting slightly cleaned up.
870 Newer C compilers apply a strict scope rule to extern
871 declarations within functions. Two extern declarations in
872 lib.c and tran.c have been moved to obviate this problem.
875 Reluctantly added toupper and tolower functions.
876 Subject to rescinding without notice.
879 Error-message printer had printf(s) instead of
880 printf("%s",s); got core dumps when the message
884 Very long printf strings caused core dump;
885 fixed aprintf, asprintf, format to catch them.
886 Can still get a core dump in printf itself.