/[pcre]/code/trunk/ChangeLog
ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 37 - (show annotations) (download)
Sat Feb 24 21:39:09 2007 UTC (7 years, 9 months ago) by nigel
File size: 18767 byte(s)
Load pcre-2.07 into code/trunk.

1 ChangeLog for PCRE
2 ------------------
3
4
5 Version 2.07 29-Jul-99
6 ----------------------
7
8 1. The documentation is now supplied in plain text form and HTML as well as in
9 the form of man page sources.
10
11 2. C++ compilers don't like assigning (void *) values to other pointer types.
12 In particular this affects malloc(). Although there is no problem in Standard
13 C, I've put in casts to keep C++ compilers happy.
14
15 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
16 should be (const char *).
17
18 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
19 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
20 However, I haven't made this a standard facility. The documentation doesn't
21 mention it, and the Makefile doesn't support it.
22
23 5. The Makefile now contains an "install" target, with editable destinations at
24 the top of the file. The pcretest program is not installed.
25
26 6. pgrep -V now gives the PCRE version number and date.
27
28 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
29 causing the entire string to be ignored, instead of just the last character.
30
31 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
32 non-matching string, it can take a very, very long time, even for strings of
33 quite modest length, because of the nested recursion. PCRE now does better in
34 some of these cases. It does this by remembering the last required literal
35 character in the pattern, and pre-searching the subject to ensure it is present
36 before running the real match. In other words, it applies a heuristic to detect
37 some types of certain failure quickly, and in the above example, if presented
38 with a string that has no trailing " it gives "no match" very quickly.
39
40 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
41 other alternatives are tried instead.
42
43
44 Version 2.06 09-Jun-99
45 ----------------------
46
47 1. Change pcretest's output for amount of store used to show just the code
48 space, because the remainder (the data block) varies in size between 32-bit and
49 64-bit systems.
50
51 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
52 start matching at. This allows lookbehinds to work when searching for multiple
53 occurrences in a string.
54
55 3. Added additional options to pcretest for testing multiple occurrences:
56
57 /+ outputs the rest of the string that follows a match
58 /g loops for multiple occurrences, using the new startoffset argument
59 /G loops for multiple occurrences by passing an incremented pointer
60
61 4. PCRE wasn't doing the "first character" optimization for patterns starting
62 with \b or \B, though it was doing it for other lookbehind assertions. That is,
63 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
64 the letter 'x'. On long subject strings, this gives a significant speed-up.
65
66
67 Version 2.05 21-Apr-99
68 ----------------------
69
70 1. Changed the type of magic_number from int to long int so that it works
71 properly on 16-bit systems.
72
73 2. Fixed a bug which caused patterns starting with .* not to work correctly
74 when the subject string contained newline characters. PCRE was assuming
75 anchoring for such patterns in all cases, which is not correct because .* will
76 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
77 DOTALL is set at top level; otherwise it knows that patterns starting with .*
78 must be retried after every newline in the subject.
79
80
81 Version 2.04 18-Feb-99
82 ----------------------
83
84 1. For parenthesized subpatterns with repeats whose minimum was zero, the
85 computation of the store needed to hold the pattern was incorrect (too large).
86 If such patterns were nested a few deep, this could multiply and become a real
87 problem.
88
89 2. Added /M option to pcretest to show the memory requirement of a specific
90 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
91
92 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
93 compiled in such a way that the backtracking after subsequent failure was
94 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
95 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
96
97
98 Version 2.03 02-Feb-99
99 ----------------------
100
101 1. Fixed typo and small mistake in man page.
102
103 2. Added 4th condition (GPL supersedes if conflict) and created separate
104 LICENCE file containing the conditions.
105
106 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
107 Perl, that is the internal \ allows the delimiter to be included in the
108 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
109 the final delimiter, add \ to the end of the pattern (to test the error).
110
111 4. Added the convenience functions for extracting substrings after a successful
112 match. Updated pcretest to make it able to test these functions.
113
114
115 Version 2.02 14-Jan-99
116 ----------------------
117
118 1. Initialized the working variables associated with each extraction so that
119 their saving and restoring doesn't refer to uninitialized store.
120
121 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
122 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
123 fix the problem.
124
125 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
126 calls, and wasn't printing the correct value for compiling calls. Increased the
127 default value of LOOPREPEAT, and the number of significant figures in the
128 times.
129
130 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
131
132 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
133 a building problem on Windows NT with a FAT file system.
134
135
136 Version 2.01 21-Oct-98
137 ----------------------
138
139 1. Changed the API for pcre_compile() to allow for the provision of a pointer
140 to character tables built by pcre_maketables() in the current locale. If NULL
141 is passed, the default tables are used.
142
143
144 Version 2.00 24-Sep-98
145 ----------------------
146
147 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
148 it any more.
149
150 2. Allow quantification of (?>) groups, and make it work correctly.
151
152 3. The first character computation wasn't working for (?>) groups.
153
154 4. Correct the implementation of \Z (it is permitted to match on the \n at the
155 end of the subject) and add 5.005's \z, which really does match only at the
156 very end of the subject.
157
158 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
159
160 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
161 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
162 localized options. All options to pcre_study() were also removed.
163
164 7. Add other new features from 5.005:
165
166 $(?<= positive lookbehind
167 $(?<! negative lookbehind
168 (?imsx-imsx) added the unsetting capability
169 such a setting is global if at outer level; local otherwise
170 (?imsx-imsx:) non-capturing groups with option setting
171 (?(cond)re|re) conditional pattern matching
172
173 A backreference to itself in a repeated group matches the previous
174 captured string.
175
176 8. General tidying up of studying (both automatic and via "study")
177 consequential on the addition of new assertions.
178
179 9. As in 5.005, unlimited repeated groups that could match an empty substring
180 are no longer faulted at compile time. Instead, the loop is forcibly broken at
181 runtime if any iteration does actually match an empty substring.
182
183 10. Include the RunTest script in the distribution.
184
185 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
186 discrepancies, some of which were old and were also with respect to 5.004. They
187 have now been fixed.
188
189
190 Version 1.09 28-Apr-98
191 ----------------------
192
193 1. A negated single character class followed by a quantifier with a minimum
194 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
195 program crashes, or just wrong answers. This did not apply to negated classes
196 containing more than one character, or to minima other than one.
197
198
199 Version 1.08 27-Mar-98
200 ----------------------
201
202 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
203
204 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
205 latter must appear before anything that relies on it in the pattern.
206
207
208 Version 1.07 16-Feb-98
209 ----------------------
210
211 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
212 repeat of a potentially empty string).
213
214
215 Version 1.06 23-Jan-98
216 ----------------------
217
218 1. Added Markus Oberhumer's little patches for C++.
219
220 2. Literal strings longer than 255 characters were broken.
221
222
223 Version 1.05 23-Dec-97
224 ----------------------
225
226 1. Negated character classes containing more than one character were failing if
227 PCRE_CASELESS was set at run time.
228
229
230 Version 1.04 19-Dec-97
231 ----------------------
232
233 1. Corrected the man page, where some "const" qualifiers had been omitted.
234
235 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
236 input syntax.
237
238 3. Fixed memory leak which occurred when a regex with back references was
239 matched with an offsets vector that wasn't big enough. The temporary memory
240 that is used in this case wasn't being freed if the match failed.
241
242 4. Tidied pcretest to ensure it frees memory that it gets.
243
244 5. Temporary memory was being obtained in the case where the passed offsets
245 vector was exactly big enough.
246
247 6. Corrected definition of offsetof() from change 5 below.
248
249 7. I had screwed up change 6 below and broken the rules for the use of
250 setjmp(). Now fixed.
251
252
253 Version 1.03 18-Dec-97
254 ----------------------
255
256 1. A erroneous regex with a missing opening parenthesis was correctly
257 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
258 on some systems.
259
260 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
261 it was reported that one broken compiler failed on the former because "code" is
262 also an independent variable.
263
264 3. The erroneous regex a[]b caused an array overrun reference.
265
266 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
267 fail on data ending with that character. (It was going on too far, and checking
268 the next character, typically a binary zero.) This was specific to the
269 optimized code for single-character negative classes.
270
271 5. Added a contributed patch from the TIN world which does the following:
272
273 + Add an undef for memmove, in case the the system defines a macro for it.
274
275 + Add a definition of offsetof(), in case there isn't one. (I don't know
276 the reason behind this - offsetof() is part of the ANSI standard - but
277 it does no harm).
278
279 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
280 most of the places where whitespace preceded '#'. I have given up and
281 allowed the remaining 2 cases to be at the margin.
282
283 + Rename some variables in pcre to eliminate shadowing. This seems very
284 pedantic, but does no harm, of course.
285
286 6. Moved the call to setjmp() into its own function, to get rid of warnings
287 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
288
289 7. Constructs such as \d{8,} were compiling into the equivalent of
290 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
291 outcome, but in this particular case used more store than had been allocated,
292 which caused the bug to be discovered because it threw up an internal error.
293
294 8. The debugging code in both pcre and pcretest for outputting the compiled
295 form of a regex was going wrong in the case of back references followed by
296 curly-bracketed repeats.
297
298
299 Version 1.02 12-Dec-97
300 ----------------------
301
302 1. Typos in pcre.3 and comments in the source fixed.
303
304 2. Applied a contributed patch to get rid of places where it used to remove
305 'const' from variables, and fixed some signed/unsigned and uninitialized
306 variable warnings.
307
308 3. Added the "runtest" target to Makefile.
309
310 4. Set default compiler flag to -O2 rather than just -O.
311
312
313 Version 1.01 19-Nov-97
314 ----------------------
315
316 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
317 like /([ab]*)*/, that is, for classes with more than one character in them.
318
319 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
320 as /((?>a*))*/ (a PCRE_EXTRA facility).
321
322
323 Version 1.00 18-Nov-97
324 ----------------------
325
326 1. Added compile-time macros to support systems such as SunOS4 which don't have
327 memmove() or strerror() but have other things that can be used instead.
328
329 2. Arranged that "make clean" removes the executables.
330
331
332 Version 0.99 27-Oct-97
333 ----------------------
334
335 1. Fixed bug in code for optimizing classes with only one character. It was
336 initializing a 32-byte map regardless, which could cause it to run off the end
337 of the memory it had got.
338
339 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
340
341
342 Version 0.98 22-Oct-97
343 ----------------------
344
345 1. Fixed bug in code for handling temporary memory usage when there are more
346 back references than supplied space in the ovector. This could cause segfaults.
347
348
349 Version 0.97 21-Oct-97
350 ----------------------
351
352 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
353
354 2. Optimized negated single characters not to use a bit map.
355
356 3. Brought error texts together as macro definitions; clarified some of them;
357 fixed one that was wrong - it said "range out of order" when it meant "invalid
358 escape sequence".
359
360 4. Changed some char * arguments to const char *.
361
362 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
363
364 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
365 pcretest.
366
367
368 Version 0.96 16-Oct-97
369 ----------------------
370
371 1. Added a simple "pgrep" utility to the distribution.
372
373 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
374 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
375 where "ddd" means "one or more decimal digits".
376
377 3. Fixed serious bug. If a pattern had a back reference, but the call to
378 pcre_exec() didn't supply a large enough ovector to record the related
379 identifying subpattern, the match always failed. PCRE now remembers the number
380 of the largest back reference, and gets some temporary memory in which to save
381 the offsets during matching if necessary, in order to ensure that
382 backreferences always work.
383
384 4. Increased the compatibility with Perl in a number of ways:
385
386 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
387 to request this handling. The option can be set at compile or exec time.
388
389 (b) $ matches before a terminating newline by default; an option
390 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
391 mode). The option can be set at compile or exec time.
392
393 (c) The handling of \ followed by a digit other than 0 is now supposed to be
394 the same as Perl's. If the decimal number it represents is less than 10
395 or there aren't that many previous left capturing parentheses, an octal
396 escape is read. Inside a character class, it's always an octal escape,
397 even if it is a single digit.
398
399 (d) An escaped but undefined alphabetic character is taken as a literal,
400 unless PCRE_EXTRA is set. Currently this just reserves the remaining
401 escapes.
402
403 (e) {0} is now permitted. (The previous item is removed from the compiled
404 pattern).
405
406 5. Changed all the names of code files so that the basic parts are no longer
407 than 10 characters, and abolished the teeny "globals.c" file.
408
409 6. Changed the handling of character classes; they are now done with a 32-byte
410 bit map always.
411
412 7. Added the -d and /D options to pcretest to make it possible to look at the
413 internals of compilation without having to recompile pcre.
414
415
416 Version 0.95 23-Sep-97
417 ----------------------
418
419 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
420 \x20 at the start of a run of normal characters. These were being treated as
421 real characters, instead of the source characters being re-checked.
422
423
424 Version 0.94 18-Sep-97
425 ----------------------
426
427 1. The functions are now thread-safe, with the caveat that the global variables
428 containing pointers to malloc() and free() or alternative functions are the
429 same for all threads.
430
431 2. Get pcre_study() to generate a bitmap of initial characters for non-
432 anchored patterns when this is possible, and use it if passed to pcre_exec().
433
434
435 Version 0.93 15-Sep-97
436 ----------------------
437
438 1. /(b)|(:+)/ was computing an incorrect first character.
439
440 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
441 but not actually doing anything yet.
442
443 3. Treat "-" characters in classes that cannot be part of ranges as literals,
444 as Perl does (e.g. [-az] or [az-]).
445
446 4. Set the anchored flag if a branch starts with .* or .*? because that tests
447 all possible positions.
448
449 5. Split up into different modules to avoid including unneeded functions in a
450 compiled binary. However, compile and exec are still in one module. The "study"
451 function is split off.
452
453 6. The character tables are now in a separate module whose source is generated
454 by an auxiliary program - but can then be edited by hand if required. There are
455 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
456 toupper() in the code.
457
458 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
459 make them global. Abolish the function for setting them, as the caller can now
460 set them directly.
461
462
463 Version 0.92 11-Sep-97
464 ----------------------
465
466 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
467 (e.g. /a{1,3}/) was broken (I mis-optimized it).
468
469 2. Caseless matching was not working in character classes if the characters in
470 the pattern were in upper case.
471
472 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
473
474 4. Make PCRE_ANCHORED public and accept as a compile option.
475
476 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
477 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
478 pass them.
479
480 6. Give an error if bad option bits passed at compile or run time.
481
482 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
483 pcretest to cause it to pass that flag.
484
485 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
486 options, and the first character, if set.
487
488 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
489
490
491 Version 0.91 10-Sep-97
492 ----------------------
493
494 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
495 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
496
497 2. PCRE was looping on encountering an indefinitely repeated back reference to
498 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
499 Perl does - treats the match as successful.
500
501 ****

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12