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 fixed a long-standing issue with when FS takes effect. a
30 change to FS is now noticed immediately for subsequent splits.
32 changed the name getline() to awkgetline() to avoid yet another
33 name conflict somewhere.
36 temporarily for now defined HAS_ISBLANK, since that seems to
37 be the best way through the thicket. isblank arrived in C99,
38 but seems to be arriving at different systems at different
42 fixed typo in b.c that set tmpvec wrongly. no one had ever
43 run into the problem, apparently. thanks to alistair crooks.
46 minor fix in lib.c: increase inputFS to 100, change malloc
49 fixed memory fault caused by out of order test in setsval.
51 thanks to david o'brien, freebsd, for both fixes.
54 fiddle in makefile to fix for BSD make; thanks to igor sobrado.
57 fixed some null pointer refs calling adjbuf.
60 fixed a bug in matching the null RE in sub and gsub. thanks to al aho
61 who actually did the fix (in b.c), and to wolfgang seeberg for finding
62 it and providing a very compact test case.
64 fixed quotation in b.c; thanks to Hal Pratt and the Princeton Dante
67 removed some no-effect asserts in run.c.
69 fiddled maketab.c to not complain about bison-generated values.
71 removed the obsolete -V argument; fixed --version to print the
74 fixed wording and an outright error in the usage message; thanks to igor
75 sobrado and jason mcintyre.
77 fixed a bug in -d that caused core dump if no program followed.
80 dropped mac.code from makefile; there are few non-MacOSX
84 system() not flagged as unsafe in the unadvertised -safe option.
85 found it while enhancing tests before shipping the ;login: article.
86 practice what you preach.
88 removed the 9-years-obsolete -mr and -mf flags.
90 added -version and --version options.
92 core dump on linux with BEGIN {nextfile}, now fixed.
94 removed some #ifdef's in run.c and lex.c that appear to no
98 modified lib.c so that values of $0 et al are preserved in the END
99 block, apparently as required by posix. thanks to havard eidnes
100 for the report and code.
103 fixed infinite loop in parsing, originally found by brian tsang.
104 thanks to arnold robbins for a suggestion that started me
108 prevent overflow of -f array in main, head off potential error in
109 call of SYNTAX(), test malloc return in lib.c, all with thanks to
113 cranked up size of NCHARS; coverity thinks it can be overrun with
114 smaller size, and i think that's right. added some assertions to b.c
115 to catch places where it might overrun. the RE code is still fragile.
118 fixed a couple of overflow problems with ridiculous field numbers:
119 e.g., print $(2^32-1). thanks to ruslan ermilov, giorgos keramidas
120 and david o'brien at freebsd.org for patches. this really should
121 be re-done from scratch.
124 fixed another 25-year-old RE bug, in split. it's another failure
125 to (re-)initialize. thanks to steve fisher for spotting this and
126 providing a good test case.
129 fixed a bug in regular expressions that dates (so help me) from 1977;
130 it's been there from the beginning. an anchored longest match that
131 was longer than the number of states triggered a failure to initialize
132 the machine properly. many thanks to moinak ghosh for not only finding
133 this one but for providing a fix, in some of the most mysterious
136 fixed a storage leak in call() that appears to have been there since
137 1983 or so -- a function without an explicit return that assigns a
138 string to a parameter leaked a Cell. thanks to moinak ghosh for
139 spotting this very subtle one.
142 fixed, thanks to andrey chernov and ruslan ermilov, a bug in lex.c
143 that mis-handled the character 255 in input. (it was being compared
144 to EOF with a signed comparison.)
147 fixed (i think) the long-standing botch that included the beginning of
148 line state ^ for RE's in the set of valid characters; this led to a
149 variety of odd problems, including failure to properly match certain
150 regular expressions in non-US locales. thanks to ruslan for keeping
154 n-th try at getting internationalization right, with thanks to volker
155 kiefel, arnold robbins and ruslan ermilov for advice, though they
156 should not be blamed for the outcome. according to posix, "." is the
157 radix character in programs and command line arguments regardless of
158 the locale; otherwise, the locale should prevail for input and output
159 of numbers. so it's intended to work that way.
161 i have rescinded the attempt to use strcoll in expanding shorthands in
162 regular expressions (cclenter). its properties are much too
163 surprising; for example [a-c] matches aAbBc in locale en_US but abBcC
164 in locale fr_CA. i can see how this might arise by implementation
165 but i cannot explain it to a human user. (this behavior can be seen
166 in gawk as well; we're leaning on the same library.)
168 the issue appears to be that strcoll is meant for sorting, where
169 merging upper and lower case may make sense (though note that unix
170 sort does not do this by default either). it is not appropriate
171 for regular expressions, where the goal is to match specific
172 patterns of characters. in any case, the notations [:lower:], etc.,
173 are available in awk, and they are more likely to work correctly in
176 a moratorium is hereby declared on internationalization changes.
177 i apologize to friends and colleagues in other parts of the world.
178 i would truly like to get this "right", but i don't know what
179 that is, and i do not want to keep making changes until it's clear.
182 fixed bug that permitted non-terminated RE, as in "awk /x".
185 subtle change to split: if source is empty, number of elems
186 is always 0 and the array is not set.
189 added some parens to isblank, in another attempt to make things
190 internationally portable.
193 the internationalization changes, somewhat modified, are now
194 reinstated. in theory awk will now do character comparisons
195 and case conversions in national language, but "." will always
196 be the decimal point separator on input and output regardless
197 of national language. isblank(){} has an #ifndef.
199 this no longer compiles on windows: LC_MESSAGES isn't defined
202 fixed subtle behavior in field and record splitting: if FS is
203 a single character and RS is not empty, \n is NOT a separator.
204 this tortuous reading is found in the awk book; behavior now
205 matches gawk and mawk.
208 for the moment, the internationalization changes of nov 29 are
209 rolled back -- programs like x = 1.2 don't work in some locales,
210 because the parser is expecting x = 1,2. until i understand this
211 better, this will have to wait.
214 modified b.c (with tiny changes in main and run) to support
215 locales, using strcoll and iswhatever tests for posix character
216 classes. thanks to ruslan ermilov (ru@freebsd.org) for code.
217 the function isblank doesn't seem to have propagated to any
218 header file near me, so it's there explicitly. not properly
219 tested on non-ascii character sets by me.
222 modified run/format() and tran/getsval() to do a slightly better
223 job on using OFMT for output from print and CONVFMT for other
224 number->string conversions, as promised by posix and done by
225 gawk and mawk. there are still places where it doesn't work
226 right if CONVFMT is changed; by then the STR attribute of the
227 variable has been irrevocably set. thanks to arnold robbins for
230 fixed subtle bug in format that could get core dump. thanks to
231 Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing.
232 minor cleanup in run.c / format() at the same time.
234 added some tests for null pointers to debugging printf's, which
235 were never intended for external consumption. thanks to dave
236 kerns (dkerns@lucent.com) for pointing this out.
238 GNU compatibility: an empty regexp matches anything (thanks to
239 dag-erling smorgrav, des@ofug.org). subject to reversion if
240 this does more harm than good.
242 pervasive small changes to make things more const-correct, as
243 reported by gcc's -Wwrite-strings. as it says in the gcc manual,
244 this may be more nuisance than useful. provoked by a suggestion
245 and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk
247 minor documentation changes to note that this now compiles out
248 of the box on Mac OS X.
251 changed types in posix chars structure to quiet solaris cc.
254 fflush() or fflush("") flushes all files and pipes.
256 length(arrayname) returns number of elements; thanks to
257 arnold robbins for suggestion.
259 added a makefile.win to make it easier to build on windows.
260 based on dan allen's buildwin.bat.
263 added support for posix character class names like [:digit:],
264 which are not exactly shorter than [0-9] and perhaps no more
265 portable. thanks to dag-erling smorgrav for code.
268 removed -m option; no longer needed, and it was actually
269 broken (noted thanks to volker kiefel).
272 fixed an appalling bug in gettok: any sequence of digits, +,-, E, e,
273 and period was accepted as a valid number if it started with a period.
274 this would never have happened with the lex version.
276 other 1-character botches, now fixed, include a bare $ and a
277 bare " at the end of the input.
280 more (const char *) casts in b.c and tran.c to silence warnings.
283 fixed a bug introduced in august 1997 that caused expressions
284 like $f[1] to be syntax errors. thanks to arnold robbins for
285 noticing this and providing a fix.
288 fixed some nextfile bugs: not handling all cases. thanks to
289 arnold robbins for pointing this out. new regressions added.
291 close() is now a function. it returns whatever the library
292 fclose returns, and -1 for closing a file or pipe that wasn't
296 permit \n explicitly in character classes; won't work right
297 if comes in as "[\n]" but ok as /[\n]/, because of multiple
298 processing of \'s. thanks to arnold robbins.
301 minor fiddles in tran.c to keep compilers happy about uschar.
302 thanks to norman wilson.
305 yet another attempt at making 8-bit input work, with another
306 band-aid in b.c (member()), and some (uschar) casts to head
307 off potential errors in subscripts (like isdigit). also
308 changed HAT to NCHARS-2. thanks again to santiago vila.
310 changed maketab.c to ignore apparently out of range definitions
311 instead of halting; new freeBSD generates one. thanks to
312 jon snader <jsnader@ix.netcom.com> for pointing out the problem.
315 fixed an 8-bit problem in b.c by making several char*'s into
316 unsigned char*'s. not clear i have them all yet. thanks to
317 Santiago Vila <sanvila@unex.es> for the bug report.
320 finally found and fixed a memory leak in function call; it's
321 been there since functions were added ~1983. thanks to
322 jon bentley for the test case that found it.
324 added test in envinit to catch environment "variables" with
325 names beginning with '='; thanks to Berend Hasselman.
328 added test in defn() to catch function foo(foo), which
329 otherwise recurses until core dump. thanks to arnold
330 robbins for noticing this.
333 added *bp in gettok in lex.c; appears possible to exit function
334 without terminating the string. thanks to russ cox.
337 added function stdinit() to run to initialize files[] array,
338 in case stdin, etc., are not constants; some compilers care.
341 replaced the ERROR ... FATAL, etc., macros with functions
342 based on vprintf, to avoid problems caused by overrunning
343 fixed-size errbuf array. thanks to ralph corderoy for the
344 impetus, and for pointing out a string termination bug in
348 fixed bug that caused occasional core dumps with commandline
349 variable with value ending in \. (thanks to nelson beebe for
353 with code kindly provided by Bruce Lilly, awk now parses
354 /=/ and similar constructs more sensibly in more places.
355 Bruce also provided some helpful test cases.
358 changed true/false to True/False in run.c to make it
359 easier to compile with C++. Added some casts on malloc
360 and realloc to be honest about casts; ditto. changed
361 ltype int to long in struct rrow to reduce some 64-bit
362 complaints; other changes scattered throughout for the
363 same purpose. thanks to Nelson Beebe for these portability
366 removed some horrible pointer-int casting in b.c and elsewhere
367 by adding ptoi and itonp to localize the casts, which are
368 all benign. fixed one incipient bug that showed up on sgi
371 reset lineno for new source file; include filename in error
372 message. also fixed line number error in continuation lines.
373 (thanks to Nelson Beebe for both of these.)
376 Nelson Beebe notes that irix 5.3 yacc dies with a bogus
377 error; use a newer version or switch to bison, since sgi
378 is unlikely to fix it.
381 changed isnumber to is_number to avoid the problem caused by
382 versions of ctype.h that include the name isnumber.
384 distribution now includes a script for building on a Mac,
388 fixed memory leaks in run.c (call) and tran.c (setfval).
389 thanks to Stephen Nutt for finding these and providing the fixes.
392 replaced srand argument by (unsigned int) in run.c;
393 avoids problem on Mac and potentially on Unix & Windows.
396 added a few (int) casts to silence useless compiler warnings.
397 e.g., errorflag= in run.c jump().
399 added proctab.c to the bundle outout; one less thing
400 to have to compile out of the box.
402 added calls to _popen and _pclose to the win95 stub for
403 pipes (thanks to Steve Adams for this helpful suggestion).
404 seems to work, though properties are not well understood
405 by me, and it appears that under some circumstances the
406 pipe output is truncated. Be careful.
409 fixed a couple of bugs in getrec: could fail to update $0
410 after a getline var; because inputFS wasn't initialized,
411 could split $0 on every character, a misleading diversion.
413 fixed caching bug in makedfa: LRU was actually removing
416 thanks to ross ridge for finding these, and for providing
420 fixed potential bug in readrec: might fail to update record
421 pointer after growing. thanks to dan levy for spotting this
422 and suggesting the fix.
425 added -V to print version number and die.
428 subtle silent bug in lex.c: if the program ended with a number
429 longer than 1 digit, part of the input would be pushed back and
430 parsed again because token buffer wasn't terminated right.
431 example: awk 'length($0) > 10'. blush. at least i found it
435 s/adelete/awkdelete/: SGI uses this in malloc.h.
436 thanks to nelson beebe for pointing this one out.
439 fixed some bugs in sub and gsub when replacement includes \\.
440 this is a dark, horrible corner, but at least now i believe that
441 the behavior is the same as gawk and the intended posix standard.
442 thanks to arnold robbins for advice here.
445 somewhat regretfully, replaced the ancient lex-based lexical
446 analyzer with one written in C. it's longer, generates less code,
447 and more portable; the old one depended too much on mysterious
448 properties of lex that were not preserved in other environments.
449 in theory these recognize the same language.
451 now using strtod to test whether a string is a number, instead of
452 the convoluted original function. should be more portable and
453 reliable if strtod is implemented right.
455 removed now-pointless optimization in makefile that tries to avoid
456 recompilation when awkgram.y is changed but symbols are not.
458 removed most fixed-size arrays, though a handful remain, some
459 of which are unchecked. you have been warned.
462 with some trepidation, replaced the ancient code that managed
463 fields and $0 in fixed-size arrays with arrays that grow on
464 demand. there is still some tension between trying to make this
465 run fast and making it clean; not sure it's right yet.
467 the ill-conceived -mr and -mf arguments are now useful only
468 for debugging. previous dynamic string code removed.
470 numerous other minor cleanups along the way.
473 using code provided by dan levy (to whom profuse thanks), replaced
474 fixed-size arrays and awkward kludges by a fairly uniform mechanism
475 to grow arrays as needed for printf, sub, gsub, etc.
478 falling off the end of a function returns "" and 0, not 0.
479 thanks to arnold robbins.
482 replaced several fixed-size arrays by dynamically-created ones
483 in run.c; added overflow tests to some previously unchecked cases.
484 getline, toupper, tolower.
486 getline code is still broken in that recursive calls may wind
487 up using the same space. [fixed later]
489 increased RECSIZE to 8192 to push problems further over the horizon.
491 added \r to \n as input line separator for programs, not data.
494 modified format() to permit explicit printf("%c", 0) to include
495 a null byte in output. thanks to ken stailey for the fix.
497 added a "-safe" argument that disables file output (print >,
498 print >>), process creation (cmd|getline, print |, system), and
499 access to the environment (ENVIRON). this is a first approximation
500 to a "safe" version of awk, but don't rely on it too much. thanks
501 to joan feigenbaum and matt blaze for the inspiration long ago.
504 fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to
508 fixed awful bug in new field splitting; didn't get all the places
509 where input was done.
512 changed field-splitting to conform to posix definition: fields are
513 split using the value of FS at the time of input; it used to be
514 the value when the field or NF was first referred to, a much less
515 predictable definition. thanks to arnold robbins for encouragement
516 to do the right thing.
519 fixed appalling but apparently unimportant bug in parsing octal
520 numbers in reg exprs.
522 explicit hex in reg exprs now limited to 2 chars: \xa, \xaa.
525 cleaned up some declarations so gcc -Wall is now almost silent.
527 makefile now includes backup copies of ytab.c and lexyy.c in case
528 one makes before looking; it also avoids recreating lexyy.c unless
531 s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes
532 with unwisely-written header files.
534 thanks to jeffrey friedl for several of these.
537 an attempt to rationalize the (unsigned) char issue. almost all
538 instances of unsigned char have been removed; the handful of places
539 in b.c where chars are used as table indices have been hand-crafted.
540 added some latin-1 tests to the regression, but i'm not confident;
541 none of my compilers seem to care much. thanks to nelson beebe for
542 pointing out some others that do care.
545 removed all register declarations.
547 enhanced split(), as in gawk, etc: split(s, a, "") splits s into
548 a[1]...a[length(s)] with each character a single element.
550 made the same changes for field-splitting if FS is "".
552 added nextfile, as in gawk: causes immediate advance to next
553 input file. (thanks to arnold robbins for inspiration and code).
555 small fixes to regexpr code: can now handle []], [[], and
556 variants; [] is now a syntax error, rather than matching
557 everything; [z-a] is now empty, not z. far from complete
558 or correct, however. (thanks to jeffrey friedl for pointing out
559 some awful behaviors.)
562 replaced uchar by uschar everywhere; apparently some compilers
563 usurp this name and this causes conflicts.
565 fixed call to time in run.c (bltin); arg is time_t *.
567 replaced horrible pointer/long punning in b.c by a legitimate
568 union. should be safer on 64-bit machines and cleaner everywhere.
569 (thanks to nelson beebe for pointing out some of these problems.)
571 replaced nested comments by #if 0...#endif in run.c, lib.c.
573 removed getsval, setsval, execute macros from run.c and lib.c.
574 machines are 100x faster than they were when these macros were
577 revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l,
578 y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of
579 portability to nameless systems.
581 "make bundle" now includes yacc and lex output files for recipients
582 who don't have yacc or lex.
585 initialized Cells in setsymtab more carefully; some fields
586 were not set. (thanks to purify, all of whose complaints i
587 think i now understand.)
589 fixed at least one error in gsub that looked at -1-th element
590 of an array when substituting for a null match (e.g., $).
592 delete arrayname is now legal; it clears the elements but leaves
593 the array, which may not be the right behavior.
595 modified makefile: my current make can't cope with the test used
596 to avoid unnecessary yacc invocations.
599 added dynamically growing strings to awk.lx.l and b.c
600 to permit regular expressions to be much bigger.
601 the state arrays can still overflow.
604 detect duplicate arguments in function definitions (mdm).
607 trivial fix to printf to limit string size in sub().
610 fixed yet another subtle self-assignment problem:
611 $1 = $2; $1 = $1 clobbered $1.
613 Regression tests now use private echo, to avoid quoting problems.
616 changed error() to print line number as %d, not %g.
619 cosmetic changes: increased sizes of some arrays,
620 reworded some error messages.
622 added CONVFMT as in posix (just replaced OFMT in getsval)
624 FILENAME is now "" until the first thing that causes a file
628 deleted yyunput and yyoutput from proto.h;
629 different versions of lex give these different declarations.
632 added -mr N and -mf N options: more record and fields.
633 these really ought to adjust automatically.
635 cleaned up some error messages; "out of space" now means
636 malloc returned NULL in all cases.
638 changed rehash so that if it runs out, it just returns;
639 things will continue to run slow, but maybe a bit longer.
642 remove redundant close of stdin when using -f -.
644 got rid of core dump with -d; awk -d just prints date.
647 added explicit check for /dev/std(in,out,err) in redirection.
648 unlike gawk, no /dev/fd/n yet.
650 added (file/pipe) builtin. hard to test satisfactorily.
654 recompile after abortive changes; should be unchanged.
657 die-casting time: converted to ansi C, installed that.
660 fixed storage leak in freefa, failing to recover [N]CCL.
661 thanks to Bill Jones (jones@cs.usask.ca)
664 use RAND_MAX instead of literal in builtin().
667 cranked up some fixed-size arrays in b.c, and added a test for
668 overflow in penter. thanks to mark larsen.
671 increased buffer in gsub. a very crude fix to a general problem.
675 enforce variable name syntax for commandline variables: has to
676 start with letter or _.
679 allow newline after ; in for statements.
682 fixed so that in self-assignment like $1=$1, side effects
683 like recomputing $0 take place. (this is getting subtle.)
686 better test for detecting too-long output record.
689 better defense against very long printf strings.
690 made break and continue illegal outside of loops.
693 removed extra arg on gettemp, tempfree. minor error message rewording.
696 fixed silly bug in hex parsing in hexstr().
697 removed an apparently unnecessary test in isnumber().
698 warn about weird printf conversions.
699 fixed unchecked array overwrite in relex().
701 changed for (i in array) to access elements in sorted order.
702 then unchanged it -- it really does run slower in too many cases.
703 left the code in place, commented out.
706 check error status on all writes, to avoid banging on full disks.
709 awk -f - reads the program from stdin.
712 failed to set numeric state on $0 in cmd|getline context in run.c.
715 fixed sleazy test for integrality in getsval; use modf.
718 fixed sleazy buggy code in lib.c that looked (incorrectly) for
719 too long input lines.
722 fixed the bug on p. 198 in which it couldn't deduce that an
723 argument was an array in some contexts. replaced the error
724 message in intest() by code that damn well makes it an array.
727 fixed horrible bug: types and values were not preserved in
728 some kinds of self-assignment. (in assign().)
731 changed NCHARS to 256 to handle 8-bit characters in strings
732 presented to match(), etc.
735 changed struct rrow (awk.h) to use long instead of int for lval,
736 since cfoll() stores a pointer in it. now works better when int's
737 are smaller than pointers!
740 AVA fixed the grammar so that ! is uniformly of the same precedence as
741 unary + and -. This renders illegal some constructs like !x=y, which
742 now has to be parenthesized as !(x=y), and makes others work properly:
743 !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements.
744 (These problems were pointed out by Bob Lenk of Posix.)
746 Added \x to regular expressions (already in strings).
747 Limited octal to octal digits; \8 and \9 are not octal.
748 Centralized the code for parsing escapes in regular expressions.
749 Added a bunch of tests to T.re and T.sub to verify some of this.
752 fixed null pointer dereference bug in main.c: -F[nothing]. sigh.
754 restored srand behavior: it returns the current seed.
757 srand now returns previous seed value (0 to start).
760 fix potential problem in tran.c -- something was freed,
761 then used in freesymtab.
764 another try to get the max number of open files set with
765 relatively machine-independent code.
767 small fix to input() in case of multiple reads after EOF.
770 FILENAME is now defined in the BEGIN block -- too many old
773 "-" means stdin in getline as well as on the commandline.
775 added a bunch of casts to the code to tell the truth about
776 char * vs. unsigned char *, a right royal pain. added a
777 setlocale call to the front of main, though probably no one
778 has it usefully implemented yet.
781 removed redundant relational tests against nullnode if parse
782 tree already had a relational at that point.
785 fixed bug: commandline variable assignment has to look like
786 var=something. (consider the man page for =, in file =.1)
788 changed number of arguments to functions to static arrays
789 to avoid repeated malloc calls.
792 restored -F (space) separator
795 added -v x=1 y=2 ... for immediate commandline variable assignment;
796 done before the BEGIN block for sure. they have to precede the
797 program if the program is on the commandline.
798 Modified Aug 2 to require a separate -v for each assignment.
801 fixed ref-thru-zero bug in environment code in tran.c
804 add newline to usage message.
807 added some missing ansi printf conversion letters: %i %X %E %G.
808 no sensible meaning for h or L, so they may not do what one expects.
810 made %* conversions work.
812 changed x^y so that if n is a positive integer, it's done
813 by explicit multiplication, thus achieving maximum accuracy.
814 (this should be done by pow() but it seems not to be locally.)
815 done to x ^= y as well.
818 ENVIRON array contains environment: if shell variable V=thing,
819 ENVIRON["V"] is "thing"
821 multiple -f arguments permitted. error reporting is naive.
822 (they were permitted before, but only the last was used.)
824 fixed a really stupid botch in the debugging macro dprintf
826 fixed order of evaluation of commandline assignments to match
827 what the book claims: an argument of the form x=e is evaluated
828 at the time it would have been opened if it were a filename (p 63).
829 this invalidates the suggested answer to ex 4-1 (p 195).
831 removed some code that permitted -F (space) fieldseparator,
832 since it didn't quite work right anyway. (restored aug 2)
835 Line number now accumulated correctly for comment lines.
838 Debugging output now includes a version date,
839 if one compiles it into the source each time.
842 Changed grammar to prohibit constants as 3rd arg of sub and gsub;
843 prevents class of overwriting-a-constant errors. (Last one?)
844 This invalidates the "banana" example on page 43 of the book.
846 Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal),
847 as in ANSI, for strings. Rescinded the sloppiness that permitted
848 non-octal digits in \ooo. Warning: not all compilers and libraries
849 will be able to deal with \x correctly.
852 Fixed bug that caused tempcell list to contain a duplicate.
856 Catches some more commandline errors in main.
857 Removed redundant decl of modf in run.c (confuses some compilers).
858 Warning: there's no single declaration of malloc, etc., in awk.h
859 that seems to satisfy all compilers.
862 Added a bit of code to error printing to avoid printing nulls.
863 (Not clear that it actually would.)
866 With fear and trembling, modified the grammar to permit
867 multiple pattern-action statements on one line without
868 an explicit separator. By definition, this capitulation
869 to the ghost of ancient implementations remains undefined
870 and thus subject to change without notice or apology.
874 Fixed bug in call() that failed to recover storage.
876 A warning is now generated if there are more arguments
877 in the call than in the definition (in lieu of fixing
878 another storage leak).
881 Fixed %c: if expr is numeric, use numeric value;
882 otherwise print 1st char of string value. still
883 doesn't work if the value is 0 -- won't print \0.
885 Added a few more checks for running out of malloc.
888 Fixed bug in call() that freed local arrays twice.
890 Fixed to handle deletion of non-existent array right;
891 complains about attempt to delete non-array element.
894 Now guarantees to evaluate all arguments of built-in
895 functions, as in C; the appearance is that arguments
896 are evaluated before the function is called. Places
897 affected are sub (gsub was ok), substr, printf, and
898 all the built-in arithmetic functions in bltin().
899 A warning is generated if a bltin() is called with
900 the wrong number of arguments.
902 This requires changing makeprof on p167 of the book.
905 setting FILENAME in BEGIN caused core dump, apparently
906 because it was freeing space not allocated by malloc.
909 fixed egregious error in toupper/tolower functions.
910 still subject to rescinding, however.
913 flush stdout before opening file or pipe
916 performance bug in b.c/cgoto(): not freeing some sets of states.
917 partial fix only right now, and the number of states increased
918 to make it less obvious.
921 check error status on close
924 srand returns seed value it's using.
928 Removed limit on depth of function calls.
931 Fixed lib.c to permit _ in commandline variable names.
934 main.c fixed to recognize -- as terminator of command-
935 line options. Illegal options flagged.
936 Error reporting slightly cleaned up.
939 Newer C compilers apply a strict scope rule to extern
940 declarations within functions. Two extern declarations in
941 lib.c and tran.c have been moved to obviate this problem.
944 Reluctantly added toupper and tolower functions.
945 Subject to rescinding without notice.
948 Error-message printer had printf(s) instead of
949 printf("%s",s); got core dumps when the message
953 Very long printf strings caused core dump;
954 fixed aprintf, asprintf, format to catch them.
955 Can still get a core dump in printf itself.