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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 35 - (show annotations) (download)
Sat Feb 24 21:39:05 2007 UTC (7 years, 2 months ago) by nigel
File size: 16928 byte(s)
Load pcre-2.06 into code/trunk.

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12