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 long-standing overflow bug in run.c; many thanks to
30 nelson beebe for spotting it and providing the fix.
32 fixed bug that didn't parse -vd=1 properly; thanks to santiago
36 i give up. replaced isblank with isspace in b.c; there are
37 no consistent header files.
40 fixed a long-standing issue with when FS takes effect. a
41 change to FS is now noticed immediately for subsequent splits.
43 changed the name getline() to awkgetline() to avoid yet another
44 name conflict somewhere.
47 temporarily for now defined HAS_ISBLANK, since that seems to
48 be the best way through the thicket. isblank arrived in C99,
49 but seems to be arriving at different systems at different
53 fixed typo in b.c that set tmpvec wrongly. no one had ever
54 run into the problem, apparently. thanks to alistair crooks.
57 minor fix in lib.c: increase inputFS to 100, change malloc
60 fixed memory fault caused by out of order test in setsval.
62 thanks to david o'brien, freebsd, for both fixes.
65 fiddle in makefile to fix for BSD make; thanks to igor sobrado.
68 fixed some null pointer refs calling adjbuf.
71 fixed a bug in matching the null RE in sub and gsub. thanks to al aho
72 who actually did the fix (in b.c), and to wolfgang seeberg for finding
73 it and providing a very compact test case.
75 fixed quotation in b.c; thanks to Hal Pratt and the Princeton Dante
78 removed some no-effect asserts in run.c.
80 fiddled maketab.c to not complain about bison-generated values.
82 removed the obsolete -V argument; fixed --version to print the
85 fixed wording and an outright error in the usage message; thanks to igor
86 sobrado and jason mcintyre.
88 fixed a bug in -d that caused core dump if no program followed.
91 dropped mac.code from makefile; there are few non-MacOSX
95 system() not flagged as unsafe in the unadvertised -safe option.
96 found it while enhancing tests before shipping the ;login: article.
97 practice what you preach.
99 removed the 9-years-obsolete -mr and -mf flags.
101 added -version and --version options.
103 core dump on linux with BEGIN {nextfile}, now fixed.
105 removed some #ifdef's in run.c and lex.c that appear to no
109 modified lib.c so that values of $0 et al are preserved in the END
110 block, apparently as required by posix. thanks to havard eidnes
111 for the report and code.
114 fixed infinite loop in parsing, originally found by brian tsang.
115 thanks to arnold robbins for a suggestion that started me
119 prevent overflow of -f array in main, head off potential error in
120 call of SYNTAX(), test malloc return in lib.c, all with thanks to
124 cranked up size of NCHARS; coverity thinks it can be overrun with
125 smaller size, and i think that's right. added some assertions to b.c
126 to catch places where it might overrun. the RE code is still fragile.
129 fixed a couple of overflow problems with ridiculous field numbers:
130 e.g., print $(2^32-1). thanks to ruslan ermilov, giorgos keramidas
131 and david o'brien at freebsd.org for patches. this really should
132 be re-done from scratch.
135 fixed another 25-year-old RE bug, in split. it's another failure
136 to (re-)initialize. thanks to steve fisher for spotting this and
137 providing a good test case.
140 fixed a bug in regular expressions that dates (so help me) from 1977;
141 it's been there from the beginning. an anchored longest match that
142 was longer than the number of states triggered a failure to initialize
143 the machine properly. many thanks to moinak ghosh for not only finding
144 this one but for providing a fix, in some of the most mysterious
147 fixed a storage leak in call() that appears to have been there since
148 1983 or so -- a function without an explicit return that assigns a
149 string to a parameter leaked a Cell. thanks to moinak ghosh for
150 spotting this very subtle one.
153 fixed, thanks to andrey chernov and ruslan ermilov, a bug in lex.c
154 that mis-handled the character 255 in input. (it was being compared
155 to EOF with a signed comparison.)
158 fixed (i think) the long-standing botch that included the beginning of
159 line state ^ for RE's in the set of valid characters; this led to a
160 variety of odd problems, including failure to properly match certain
161 regular expressions in non-US locales. thanks to ruslan for keeping
165 n-th try at getting internationalization right, with thanks to volker
166 kiefel, arnold robbins and ruslan ermilov for advice, though they
167 should not be blamed for the outcome. according to posix, "." is the
168 radix character in programs and command line arguments regardless of
169 the locale; otherwise, the locale should prevail for input and output
170 of numbers. so it's intended to work that way.
172 i have rescinded the attempt to use strcoll in expanding shorthands in
173 regular expressions (cclenter). its properties are much too
174 surprising; for example [a-c] matches aAbBc in locale en_US but abBcC
175 in locale fr_CA. i can see how this might arise by implementation
176 but i cannot explain it to a human user. (this behavior can be seen
177 in gawk as well; we're leaning on the same library.)
179 the issue appears to be that strcoll is meant for sorting, where
180 merging upper and lower case may make sense (though note that unix
181 sort does not do this by default either). it is not appropriate
182 for regular expressions, where the goal is to match specific
183 patterns of characters. in any case, the notations [:lower:], etc.,
184 are available in awk, and they are more likely to work correctly in
187 a moratorium is hereby declared on internationalization changes.
188 i apologize to friends and colleagues in other parts of the world.
189 i would truly like to get this "right", but i don't know what
190 that is, and i do not want to keep making changes until it's clear.
193 fixed bug that permitted non-terminated RE, as in "awk /x".
196 subtle change to split: if source is empty, number of elems
197 is always 0 and the array is not set.
200 added some parens to isblank, in another attempt to make things
201 internationally portable.
204 the internationalization changes, somewhat modified, are now
205 reinstated. in theory awk will now do character comparisons
206 and case conversions in national language, but "." will always
207 be the decimal point separator on input and output regardless
208 of national language. isblank(){} has an #ifndef.
210 this no longer compiles on windows: LC_MESSAGES isn't defined
213 fixed subtle behavior in field and record splitting: if FS is
214 a single character and RS is not empty, \n is NOT a separator.
215 this tortuous reading is found in the awk book; behavior now
216 matches gawk and mawk.
219 for the moment, the internationalization changes of nov 29 are
220 rolled back -- programs like x = 1.2 don't work in some locales,
221 because the parser is expecting x = 1,2. until i understand this
222 better, this will have to wait.
225 modified b.c (with tiny changes in main and run) to support
226 locales, using strcoll and iswhatever tests for posix character
227 classes. thanks to ruslan ermilov (ru@freebsd.org) for code.
228 the function isblank doesn't seem to have propagated to any
229 header file near me, so it's there explicitly. not properly
230 tested on non-ascii character sets by me.
233 modified run/format() and tran/getsval() to do a slightly better
234 job on using OFMT for output from print and CONVFMT for other
235 number->string conversions, as promised by posix and done by
236 gawk and mawk. there are still places where it doesn't work
237 right if CONVFMT is changed; by then the STR attribute of the
238 variable has been irrevocably set. thanks to arnold robbins for
241 fixed subtle bug in format that could get core dump. thanks to
242 Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing.
243 minor cleanup in run.c / format() at the same time.
245 added some tests for null pointers to debugging printf's, which
246 were never intended for external consumption. thanks to dave
247 kerns (dkerns@lucent.com) for pointing this out.
249 GNU compatibility: an empty regexp matches anything (thanks to
250 dag-erling smorgrav, des@ofug.org). subject to reversion if
251 this does more harm than good.
253 pervasive small changes to make things more const-correct, as
254 reported by gcc's -Wwrite-strings. as it says in the gcc manual,
255 this may be more nuisance than useful. provoked by a suggestion
256 and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk
258 minor documentation changes to note that this now compiles out
259 of the box on Mac OS X.
262 changed types in posix chars structure to quiet solaris cc.
265 fflush() or fflush("") flushes all files and pipes.
267 length(arrayname) returns number of elements; thanks to
268 arnold robbins for suggestion.
270 added a makefile.win to make it easier to build on windows.
271 based on dan allen's buildwin.bat.
274 added support for posix character class names like [:digit:],
275 which are not exactly shorter than [0-9] and perhaps no more
276 portable. thanks to dag-erling smorgrav for code.
279 removed -m option; no longer needed, and it was actually
280 broken (noted thanks to volker kiefel).
283 fixed an appalling bug in gettok: any sequence of digits, +,-, E, e,
284 and period was accepted as a valid number if it started with a period.
285 this would never have happened with the lex version.
287 other 1-character botches, now fixed, include a bare $ and a
288 bare " at the end of the input.
291 more (const char *) casts in b.c and tran.c to silence warnings.
294 fixed a bug introduced in august 1997 that caused expressions
295 like $f[1] to be syntax errors. thanks to arnold robbins for
296 noticing this and providing a fix.
299 fixed some nextfile bugs: not handling all cases. thanks to
300 arnold robbins for pointing this out. new regressions added.
302 close() is now a function. it returns whatever the library
303 fclose returns, and -1 for closing a file or pipe that wasn't
307 permit \n explicitly in character classes; won't work right
308 if comes in as "[\n]" but ok as /[\n]/, because of multiple
309 processing of \'s. thanks to arnold robbins.
312 minor fiddles in tran.c to keep compilers happy about uschar.
313 thanks to norman wilson.
316 yet another attempt at making 8-bit input work, with another
317 band-aid in b.c (member()), and some (uschar) casts to head
318 off potential errors in subscripts (like isdigit). also
319 changed HAT to NCHARS-2. thanks again to santiago vila.
321 changed maketab.c to ignore apparently out of range definitions
322 instead of halting; new freeBSD generates one. thanks to
323 jon snader <jsnader@ix.netcom.com> for pointing out the problem.
326 fixed an 8-bit problem in b.c by making several char*'s into
327 unsigned char*'s. not clear i have them all yet. thanks to
328 Santiago Vila <sanvila@unex.es> for the bug report.
331 finally found and fixed a memory leak in function call; it's
332 been there since functions were added ~1983. thanks to
333 jon bentley for the test case that found it.
335 added test in envinit to catch environment "variables" with
336 names beginning with '='; thanks to Berend Hasselman.
339 added test in defn() to catch function foo(foo), which
340 otherwise recurses until core dump. thanks to arnold
341 robbins for noticing this.
344 added *bp in gettok in lex.c; appears possible to exit function
345 without terminating the string. thanks to russ cox.
348 added function stdinit() to run to initialize files[] array,
349 in case stdin, etc., are not constants; some compilers care.
352 replaced the ERROR ... FATAL, etc., macros with functions
353 based on vprintf, to avoid problems caused by overrunning
354 fixed-size errbuf array. thanks to ralph corderoy for the
355 impetus, and for pointing out a string termination bug in
359 fixed bug that caused occasional core dumps with commandline
360 variable with value ending in \. (thanks to nelson beebe for
364 with code kindly provided by Bruce Lilly, awk now parses
365 /=/ and similar constructs more sensibly in more places.
366 Bruce also provided some helpful test cases.
369 changed true/false to True/False in run.c to make it
370 easier to compile with C++. Added some casts on malloc
371 and realloc to be honest about casts; ditto. changed
372 ltype int to long in struct rrow to reduce some 64-bit
373 complaints; other changes scattered throughout for the
374 same purpose. thanks to Nelson Beebe for these portability
377 removed some horrible pointer-int casting in b.c and elsewhere
378 by adding ptoi and itonp to localize the casts, which are
379 all benign. fixed one incipient bug that showed up on sgi
382 reset lineno for new source file; include filename in error
383 message. also fixed line number error in continuation lines.
384 (thanks to Nelson Beebe for both of these.)
387 Nelson Beebe notes that irix 5.3 yacc dies with a bogus
388 error; use a newer version or switch to bison, since sgi
389 is unlikely to fix it.
392 changed isnumber to is_number to avoid the problem caused by
393 versions of ctype.h that include the name isnumber.
395 distribution now includes a script for building on a Mac,
399 fixed memory leaks in run.c (call) and tran.c (setfval).
400 thanks to Stephen Nutt for finding these and providing the fixes.
403 replaced srand argument by (unsigned int) in run.c;
404 avoids problem on Mac and potentially on Unix & Windows.
407 added a few (int) casts to silence useless compiler warnings.
408 e.g., errorflag= in run.c jump().
410 added proctab.c to the bundle outout; one less thing
411 to have to compile out of the box.
413 added calls to _popen and _pclose to the win95 stub for
414 pipes (thanks to Steve Adams for this helpful suggestion).
415 seems to work, though properties are not well understood
416 by me, and it appears that under some circumstances the
417 pipe output is truncated. Be careful.
420 fixed a couple of bugs in getrec: could fail to update $0
421 after a getline var; because inputFS wasn't initialized,
422 could split $0 on every character, a misleading diversion.
424 fixed caching bug in makedfa: LRU was actually removing
427 thanks to ross ridge for finding these, and for providing
431 fixed potential bug in readrec: might fail to update record
432 pointer after growing. thanks to dan levy for spotting this
433 and suggesting the fix.
436 added -V to print version number and die.
439 subtle silent bug in lex.c: if the program ended with a number
440 longer than 1 digit, part of the input would be pushed back and
441 parsed again because token buffer wasn't terminated right.
442 example: awk 'length($0) > 10'. blush. at least i found it
446 s/adelete/awkdelete/: SGI uses this in malloc.h.
447 thanks to nelson beebe for pointing this one out.
450 fixed some bugs in sub and gsub when replacement includes \\.
451 this is a dark, horrible corner, but at least now i believe that
452 the behavior is the same as gawk and the intended posix standard.
453 thanks to arnold robbins for advice here.
456 somewhat regretfully, replaced the ancient lex-based lexical
457 analyzer with one written in C. it's longer, generates less code,
458 and more portable; the old one depended too much on mysterious
459 properties of lex that were not preserved in other environments.
460 in theory these recognize the same language.
462 now using strtod to test whether a string is a number, instead of
463 the convoluted original function. should be more portable and
464 reliable if strtod is implemented right.
466 removed now-pointless optimization in makefile that tries to avoid
467 recompilation when awkgram.y is changed but symbols are not.
469 removed most fixed-size arrays, though a handful remain, some
470 of which are unchecked. you have been warned.
473 with some trepidation, replaced the ancient code that managed
474 fields and $0 in fixed-size arrays with arrays that grow on
475 demand. there is still some tension between trying to make this
476 run fast and making it clean; not sure it's right yet.
478 the ill-conceived -mr and -mf arguments are now useful only
479 for debugging. previous dynamic string code removed.
481 numerous other minor cleanups along the way.
484 using code provided by dan levy (to whom profuse thanks), replaced
485 fixed-size arrays and awkward kludges by a fairly uniform mechanism
486 to grow arrays as needed for printf, sub, gsub, etc.
489 falling off the end of a function returns "" and 0, not 0.
490 thanks to arnold robbins.
493 replaced several fixed-size arrays by dynamically-created ones
494 in run.c; added overflow tests to some previously unchecked cases.
495 getline, toupper, tolower.
497 getline code is still broken in that recursive calls may wind
498 up using the same space. [fixed later]
500 increased RECSIZE to 8192 to push problems further over the horizon.
502 added \r to \n as input line separator for programs, not data.
505 modified format() to permit explicit printf("%c", 0) to include
506 a null byte in output. thanks to ken stailey for the fix.
508 added a "-safe" argument that disables file output (print >,
509 print >>), process creation (cmd|getline, print |, system), and
510 access to the environment (ENVIRON). this is a first approximation
511 to a "safe" version of awk, but don't rely on it too much. thanks
512 to joan feigenbaum and matt blaze for the inspiration long ago.
515 fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to
519 fixed awful bug in new field splitting; didn't get all the places
520 where input was done.
523 changed field-splitting to conform to posix definition: fields are
524 split using the value of FS at the time of input; it used to be
525 the value when the field or NF was first referred to, a much less
526 predictable definition. thanks to arnold robbins for encouragement
527 to do the right thing.
530 fixed appalling but apparently unimportant bug in parsing octal
531 numbers in reg exprs.
533 explicit hex in reg exprs now limited to 2 chars: \xa, \xaa.
536 cleaned up some declarations so gcc -Wall is now almost silent.
538 makefile now includes backup copies of ytab.c and lexyy.c in case
539 one makes before looking; it also avoids recreating lexyy.c unless
542 s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes
543 with unwisely-written header files.
545 thanks to jeffrey friedl for several of these.
548 an attempt to rationalize the (unsigned) char issue. almost all
549 instances of unsigned char have been removed; the handful of places
550 in b.c where chars are used as table indices have been hand-crafted.
551 added some latin-1 tests to the regression, but i'm not confident;
552 none of my compilers seem to care much. thanks to nelson beebe for
553 pointing out some others that do care.
556 removed all register declarations.
558 enhanced split(), as in gawk, etc: split(s, a, "") splits s into
559 a[1]...a[length(s)] with each character a single element.
561 made the same changes for field-splitting if FS is "".
563 added nextfile, as in gawk: causes immediate advance to next
564 input file. (thanks to arnold robbins for inspiration and code).
566 small fixes to regexpr code: can now handle []], [[], and
567 variants; [] is now a syntax error, rather than matching
568 everything; [z-a] is now empty, not z. far from complete
569 or correct, however. (thanks to jeffrey friedl for pointing out
570 some awful behaviors.)
573 replaced uchar by uschar everywhere; apparently some compilers
574 usurp this name and this causes conflicts.
576 fixed call to time in run.c (bltin); arg is time_t *.
578 replaced horrible pointer/long punning in b.c by a legitimate
579 union. should be safer on 64-bit machines and cleaner everywhere.
580 (thanks to nelson beebe for pointing out some of these problems.)
582 replaced nested comments by #if 0...#endif in run.c, lib.c.
584 removed getsval, setsval, execute macros from run.c and lib.c.
585 machines are 100x faster than they were when these macros were
588 revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l,
589 y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of
590 portability to nameless systems.
592 "make bundle" now includes yacc and lex output files for recipients
593 who don't have yacc or lex.
596 initialized Cells in setsymtab more carefully; some fields
597 were not set. (thanks to purify, all of whose complaints i
598 think i now understand.)
600 fixed at least one error in gsub that looked at -1-th element
601 of an array when substituting for a null match (e.g., $).
603 delete arrayname is now legal; it clears the elements but leaves
604 the array, which may not be the right behavior.
606 modified makefile: my current make can't cope with the test used
607 to avoid unnecessary yacc invocations.
610 added dynamically growing strings to awk.lx.l and b.c
611 to permit regular expressions to be much bigger.
612 the state arrays can still overflow.
615 detect duplicate arguments in function definitions (mdm).
618 trivial fix to printf to limit string size in sub().
621 fixed yet another subtle self-assignment problem:
622 $1 = $2; $1 = $1 clobbered $1.
624 Regression tests now use private echo, to avoid quoting problems.
627 changed error() to print line number as %d, not %g.
630 cosmetic changes: increased sizes of some arrays,
631 reworded some error messages.
633 added CONVFMT as in posix (just replaced OFMT in getsval)
635 FILENAME is now "" until the first thing that causes a file
639 deleted yyunput and yyoutput from proto.h;
640 different versions of lex give these different declarations.
643 added -mr N and -mf N options: more record and fields.
644 these really ought to adjust automatically.
646 cleaned up some error messages; "out of space" now means
647 malloc returned NULL in all cases.
649 changed rehash so that if it runs out, it just returns;
650 things will continue to run slow, but maybe a bit longer.
653 remove redundant close of stdin when using -f -.
655 got rid of core dump with -d; awk -d just prints date.
658 added explicit check for /dev/std(in,out,err) in redirection.
659 unlike gawk, no /dev/fd/n yet.
661 added (file/pipe) builtin. hard to test satisfactorily.
665 recompile after abortive changes; should be unchanged.
668 die-casting time: converted to ansi C, installed that.
671 fixed storage leak in freefa, failing to recover [N]CCL.
672 thanks to Bill Jones (jones@cs.usask.ca)
675 use RAND_MAX instead of literal in builtin().
678 cranked up some fixed-size arrays in b.c, and added a test for
679 overflow in penter. thanks to mark larsen.
682 increased buffer in gsub. a very crude fix to a general problem.
686 enforce variable name syntax for commandline variables: has to
687 start with letter or _.
690 allow newline after ; in for statements.
693 fixed so that in self-assignment like $1=$1, side effects
694 like recomputing $0 take place. (this is getting subtle.)
697 better test for detecting too-long output record.
700 better defense against very long printf strings.
701 made break and continue illegal outside of loops.
704 removed extra arg on gettemp, tempfree. minor error message rewording.
707 fixed silly bug in hex parsing in hexstr().
708 removed an apparently unnecessary test in isnumber().
709 warn about weird printf conversions.
710 fixed unchecked array overwrite in relex().
712 changed for (i in array) to access elements in sorted order.
713 then unchanged it -- it really does run slower in too many cases.
714 left the code in place, commented out.
717 check error status on all writes, to avoid banging on full disks.
720 awk -f - reads the program from stdin.
723 failed to set numeric state on $0 in cmd|getline context in run.c.
726 fixed sleazy test for integrality in getsval; use modf.
729 fixed sleazy buggy code in lib.c that looked (incorrectly) for
730 too long input lines.
733 fixed the bug on p. 198 in which it couldn't deduce that an
734 argument was an array in some contexts. replaced the error
735 message in intest() by code that damn well makes it an array.
738 fixed horrible bug: types and values were not preserved in
739 some kinds of self-assignment. (in assign().)
742 changed NCHARS to 256 to handle 8-bit characters in strings
743 presented to match(), etc.
746 changed struct rrow (awk.h) to use long instead of int for lval,
747 since cfoll() stores a pointer in it. now works better when int's
748 are smaller than pointers!
751 AVA fixed the grammar so that ! is uniformly of the same precedence as
752 unary + and -. This renders illegal some constructs like !x=y, which
753 now has to be parenthesized as !(x=y), and makes others work properly:
754 !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements.
755 (These problems were pointed out by Bob Lenk of Posix.)
757 Added \x to regular expressions (already in strings).
758 Limited octal to octal digits; \8 and \9 are not octal.
759 Centralized the code for parsing escapes in regular expressions.
760 Added a bunch of tests to T.re and T.sub to verify some of this.
763 fixed null pointer dereference bug in main.c: -F[nothing]. sigh.
765 restored srand behavior: it returns the current seed.
768 srand now returns previous seed value (0 to start).
771 fix potential problem in tran.c -- something was freed,
772 then used in freesymtab.
775 another try to get the max number of open files set with
776 relatively machine-independent code.
778 small fix to input() in case of multiple reads after EOF.
781 FILENAME is now defined in the BEGIN block -- too many old
784 "-" means stdin in getline as well as on the commandline.
786 added a bunch of casts to the code to tell the truth about
787 char * vs. unsigned char *, a right royal pain. added a
788 setlocale call to the front of main, though probably no one
789 has it usefully implemented yet.
792 removed redundant relational tests against nullnode if parse
793 tree already had a relational at that point.
796 fixed bug: commandline variable assignment has to look like
797 var=something. (consider the man page for =, in file =.1)
799 changed number of arguments to functions to static arrays
800 to avoid repeated malloc calls.
803 restored -F (space) separator
806 added -v x=1 y=2 ... for immediate commandline variable assignment;
807 done before the BEGIN block for sure. they have to precede the
808 program if the program is on the commandline.
809 Modified Aug 2 to require a separate -v for each assignment.
812 fixed ref-thru-zero bug in environment code in tran.c
815 add newline to usage message.
818 added some missing ansi printf conversion letters: %i %X %E %G.
819 no sensible meaning for h or L, so they may not do what one expects.
821 made %* conversions work.
823 changed x^y so that if n is a positive integer, it's done
824 by explicit multiplication, thus achieving maximum accuracy.
825 (this should be done by pow() but it seems not to be locally.)
826 done to x ^= y as well.
829 ENVIRON array contains environment: if shell variable V=thing,
830 ENVIRON["V"] is "thing"
832 multiple -f arguments permitted. error reporting is naive.
833 (they were permitted before, but only the last was used.)
835 fixed a really stupid botch in the debugging macro dprintf
837 fixed order of evaluation of commandline assignments to match
838 what the book claims: an argument of the form x=e is evaluated
839 at the time it would have been opened if it were a filename (p 63).
840 this invalidates the suggested answer to ex 4-1 (p 195).
842 removed some code that permitted -F (space) fieldseparator,
843 since it didn't quite work right anyway. (restored aug 2)
846 Line number now accumulated correctly for comment lines.
849 Debugging output now includes a version date,
850 if one compiles it into the source each time.
853 Changed grammar to prohibit constants as 3rd arg of sub and gsub;
854 prevents class of overwriting-a-constant errors. (Last one?)
855 This invalidates the "banana" example on page 43 of the book.
857 Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal),
858 as in ANSI, for strings. Rescinded the sloppiness that permitted
859 non-octal digits in \ooo. Warning: not all compilers and libraries
860 will be able to deal with \x correctly.
863 Fixed bug that caused tempcell list to contain a duplicate.
867 Catches some more commandline errors in main.
868 Removed redundant decl of modf in run.c (confuses some compilers).
869 Warning: there's no single declaration of malloc, etc., in awk.h
870 that seems to satisfy all compilers.
873 Added a bit of code to error printing to avoid printing nulls.
874 (Not clear that it actually would.)
877 With fear and trembling, modified the grammar to permit
878 multiple pattern-action statements on one line without
879 an explicit separator. By definition, this capitulation
880 to the ghost of ancient implementations remains undefined
881 and thus subject to change without notice or apology.
885 Fixed bug in call() that failed to recover storage.
887 A warning is now generated if there are more arguments
888 in the call than in the definition (in lieu of fixing
889 another storage leak).
892 Fixed %c: if expr is numeric, use numeric value;
893 otherwise print 1st char of string value. still
894 doesn't work if the value is 0 -- won't print \0.
896 Added a few more checks for running out of malloc.
899 Fixed bug in call() that freed local arrays twice.
901 Fixed to handle deletion of non-existent array right;
902 complains about attempt to delete non-array element.
905 Now guarantees to evaluate all arguments of built-in
906 functions, as in C; the appearance is that arguments
907 are evaluated before the function is called. Places
908 affected are sub (gsub was ok), substr, printf, and
909 all the built-in arithmetic functions in bltin().
910 A warning is generated if a bltin() is called with
911 the wrong number of arguments.
913 This requires changing makeprof on p167 of the book.
916 setting FILENAME in BEGIN caused core dump, apparently
917 because it was freeing space not allocated by malloc.
920 fixed egregious error in toupper/tolower functions.
921 still subject to rescinding, however.
924 flush stdout before opening file or pipe
927 performance bug in b.c/cgoto(): not freeing some sets of states.
928 partial fix only right now, and the number of states increased
929 to make it less obvious.
932 check error status on close
935 srand returns seed value it's using.
939 Removed limit on depth of function calls.
942 Fixed lib.c to permit _ in commandline variable names.
945 main.c fixed to recognize -- as terminator of command-
946 line options. Illegal options flagged.
947 Error reporting slightly cleaned up.
950 Newer C compilers apply a strict scope rule to extern
951 declarations within functions. Two extern declarations in
952 lib.c and tran.c have been moved to obviate this problem.
955 Reluctantly added toupper and tolower functions.
956 Subject to rescinding without notice.
959 Error-message printer had printf(s) instead of
960 printf("%s",s); got core dumps when the message
964 Very long printf strings caused core dump;
965 fixed aprintf, asprintf, format to catch them.
966 Can still get a core dump in printf itself.