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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 107 - (show annotations) (download)
Wed Mar 7 11:02:28 2007 UTC (7 years, 1 month ago) by ph10
File size: 121794 byte(s)
Added some additional #ifdef SUPPORT_UTF8 to minimize the code when UTF-8 
support is not compiled.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.1 05-Mar-07
5 ---------------------
6
7 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
8 that is more "standard", making use of automake and other autotools. There
9 is some re-arrangement of the files and adjustment of comments consequent
10 on this.
11
12 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
13 for recursive directory scanning broke on some systems because the files
14 are not scanned in any specific order and on different systems the order
15 was different. A call to "sort" has been inserted into RunGrepTest for the
16 approprate test as a short-term fix. In the longer term there may be an
17 alternative.
18
19 3. I had an email from Eric Raymond about problems translating some of PCRE's
20 man pages to HTML (despite the fact that I distribute HTML pages, some
21 people do their own conversions for various reasons). The problems
22 concerned the use of low-level troff macros .br and .in. I have therefore
23 removed all such uses from the man pages (some were redundant, some could
24 be replaced by .nf/.fi pairs). The maintain/132html script that I use to
25 generate HTML has been updated to handle .nf/.fi and to complain if it
26 encounters .br or .in.
27
28 4. Updated comments in configure.ac that get placed in config.h.in and also
29 arranged for config.h to be included in the distribution, for the benefit
30 of those who have to compile without Autotools (compare pcre.h).
31
32 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
33 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
34 makevp.bat for the current PCRE, using the additional files !compile.txt,
35 !linklib.txt, and pcregexp.pas.
36
37 6. A Windows user reported a minor discrepancy with test 2, which turned out
38 to be caused by a trailing space on an input line that had got lost in his
39 copy. The trailing space was an accident, so I've just removed it.
40
41 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
42 that is needed.
43
44 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
45 as "const" (a) because they are and (b) because it helps the PHP
46 maintainers who have recently made a script to detect big data structures
47 in the php code that should be moved to the .rodata section. I remembered
48 to update Builducptable as well, so it won't revert if ucptable.h is ever
49 re-created.
50
51 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
52 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
53 order to be able to cut out the UTF-8 tables in the latter when UTF-8
54 support is not required. This saves 1.5-2K of code, which is important in
55 some applications.
56
57
58 Version 7.0 19-Dec-06
59 ---------------------
60
61 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
62 moving to gcc 4.1.1.
63
64 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
65 sys/time.h, which is documented as needed for this function. It doesn't
66 seem to matter on Linux, but it showed up on some releases of OS X.
67
68 3. It seems that there are systems where bytes whose values are greater than
69 127 match isprint() in the "C" locale. The "C" locale should be the
70 default when a C program starts up. In most systems, only ASCII printing
71 characters match isprint(). This difference caused the output from pcretest
72 to vary, making some of the tests fail. I have changed pcretest so that:
73
74 (a) When it is outputting text in the compiled version of a pattern, bytes
75 other than 32-126 are always shown as hex escapes.
76
77 (b) When it is outputting text that is a matched part of a subject string,
78 it does the same, unless a different locale has been set for the match
79 (using the /L modifier). In this case, it uses isprint() to decide.
80
81 4. Fixed a major bug that caused incorrect computation of the amount of memory
82 required for a compiled pattern when options that changed within the
83 pattern affected the logic of the preliminary scan that determines the
84 length. The relevant options are -x, and -i in UTF-8 mode. The result was
85 that the computed length was too small. The symptoms of this bug were
86 either the PCRE error "internal error: code overflow" from pcre_compile(),
87 or a glibc crash with a message such as "pcretest: free(): invalid next
88 size (fast)". Examples of patterns that provoked this bug (shown in
89 pcretest format) are:
90
91 /(?-x: )/x
92 /(?x)(?-x: \s*#\s*)/
93 /((?i)[\x{c0}])/8
94 /(?i:[\x{c0}])/8
95
96 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
97 is now done differently.
98
99 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
100 wrapper classes; (b) implement a new function in the C++ scanner that is
101 more efficient than the old way of doing things because it avoids levels of
102 recursion in the regex matching; (c) add a paragraph to the documentation
103 for the FullMatch() function.
104
105 6. The escape sequence \n was being treated as whatever was defined as
106 "newline". Not only was this contrary to the documentation, which states
107 that \n is character 10 (hex 0A), but it also went horribly wrong when
108 "newline" was defined as CRLF. This has been fixed.
109
110 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
111 was being set to -1 for the "end of line" case (supposedly a value that no
112 character can have). Though this value is never used (the check for end of
113 line is "zero bytes in current character"), it caused compiler complaints.
114 I've changed it to 0xffffffff.
115
116 8. In pcre_version.c, the version string was being built by a sequence of
117 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
118 string (as it is for production releases) called a macro with an empty
119 argument. The C standard says the result of this is undefined. The gcc
120 compiler treats it as an empty string (which was what was wanted) but it is
121 reported that Visual C gives an error. The source has been hacked around to
122 avoid this problem.
123
124 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
125 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
126 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
127 of them did).
128
129 10. Originally, pcretest opened its input and output without "b"; then I was
130 told that "b" was needed in some environments, so it was added for release
131 5.0 to both the input and output. (It makes no difference on Unix-like
132 systems.) Later I was told that it is wrong for the input on Windows. I've
133 now abstracted the modes into two macros, to make it easier to fiddle with
134 them, and removed "b" from the input mode under Windows.
135
136 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
137
138 12. Added -help and --help to pcretest as an official way of being reminded
139 of the options.
140
141 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
142 and pcrecpp.cc because they annoy compilers at high warning levels.
143
144 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
145
146 15. Fixed an occurrence of == in configure.ac that should have been = (shell
147 scripts are not C programs :-) and which was not noticed because it works
148 on Linux.
149
150 16. pcretest is supposed to handle any length of pattern and data line (as one
151 line or as a continued sequence of lines) by extending its input buffer if
152 necessary. This feature was broken for very long pattern lines, leading to
153 a string of junk being passed to pcre_compile() if the pattern was longer
154 than about 50K.
155
156 17. I have done a major re-factoring of the way pcre_compile() computes the
157 amount of memory needed for a compiled pattern. Previously, there was code
158 that made a preliminary scan of the pattern in order to do this. That was
159 OK when PCRE was new, but as the facilities have expanded, it has become
160 harder and harder to keep it in step with the real compile phase, and there
161 have been a number of bugs (see for example, 4 above). I have now found a
162 cunning way of running the real compile function in a "fake" mode that
163 enables it to compute how much memory it would need, while actually only
164 ever using a few hundred bytes of working memory and without too many
165 tests of the mode. This should make future maintenance and development
166 easier. A side effect of this work is that the limit of 200 on the nesting
167 depth of parentheses has been removed (though this was never a serious
168 limitation, I suspect). However, there is a downside: pcre_compile() now
169 runs more slowly than before (30% or more, depending on the pattern). I
170 hope this isn't a big issue. There is no effect on runtime performance.
171
172 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
173 newline (only possible for the last line of a file) and it was a
174 pattern that set a locale (followed by /Lsomething), pcretest crashed.
175
176 19. Added additional timing features to pcretest. (1) The -tm option now times
177 matching only, not compiling. (2) Both -t and -tm can be followed, as a
178 separate command line item, by a number that specifies the number of
179 repeats to use when timing. The default is 50000; this gives better
180 precision, but takes uncomfortably long for very large patterns.
181
182 20. Extended pcre_study() to be more clever in cases where a branch of a
183 subpattern has no definite first character. For example, (a*|b*)[cd] would
184 previously give no result from pcre_study(). Now it recognizes that the
185 first character must be a, b, c, or d.
186
187 21. There was an incorrect error "recursive call could loop indefinitely" if
188 a subpattern (or the entire pattern) that was being tested for matching an
189 empty string contained only one non-empty item after a nested subpattern.
190 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
191 incorrectly, because the \d was being skipped in the check.
192
193 22. The pcretest program now has a new pattern option /B and a command line
194 option -b, which is equivalent to adding /B to every pattern. This causes
195 it to show the compiled bytecode, without the additional information that
196 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
197 is the same as /B/I).
198
199 23. A new optimization is now able automatically to treat some sequences such
200 as a*b as a*+b. More specifically, if something simple (such as a character
201 or a simple class like \d) has an unlimited quantifier, and is followed by
202 something that cannot possibly match the quantified thing, the quantifier
203 is automatically "possessified".
204
205 24. A recursive reference to a subpattern whose number was greater than 39
206 went wrong under certain circumstances in UTF-8 mode. This bug could also
207 have affected the operation of pcre_study().
208
209 25. Realized that a little bit of performance could be had by replacing
210 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
211
212 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
213
214 27. Possessive quantifiers such as a++ were previously implemented by turning
215 them into atomic groups such as ($>a+). Now they have their own opcodes,
216 which improves performance. This includes the automatically created ones
217 from 23 above.
218
219 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
220 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
221 the first matched character to be a colon. This applied both to named and
222 numbered groups.
223
224 29. The ucpinternal.h header file was missing its idempotency #ifdef.
225
226 30. I was sent a "project" file called libpcre.a.dev which I understand makes
227 building PCRE on Windows easier, so I have included it in the distribution.
228
229 31. There is now a check in pcretest against a ridiculously large number being
230 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
231 loop, the loop is abandoned.
232
233 32. Forward references to subpatterns in conditions such as (?(2)...) where
234 subpattern 2 is defined later cause pcre_compile() to search forwards in
235 the pattern for the relevant set of parentheses. This search went wrong
236 when there were unescaped parentheses in a character class, parentheses
237 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
238
239 33. "Subroutine" calls and backreferences were previously restricted to
240 referencing subpatterns earlier in the regex. This restriction has now
241 been removed.
242
243 34. Added a number of extra features that are going to be in Perl 5.10. On the
244 whole, these are just syntactic alternatives for features that PCRE had
245 previously implemented using the Python syntax or my own invention. The
246 other formats are all retained for compatibility.
247
248 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
249 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
250 also .NET compatible.
251
252 (b) A recursion or subroutine call to a named group can now be defined as
253 (?&name) as well as (?P>name).
254
255 (c) A backreference to a named group can now be defined as \k<name> or
256 \k'name' as well as (?P=name). The new forms, as well as being in Perl
257 5.10, are also .NET compatible.
258
259 (d) A conditional reference to a named group can now use the syntax
260 (?(<name>) or (?('name') as well as (?(name).
261
262 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
263 groups (named and numbered) that are never evaluated inline, but can be
264 called as "subroutines" from elsewhere. In effect, the DEFINE condition
265 is always false. There may be only one alternative in such a group.
266
267 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
268 as the simple (?(R). The condition is true only if the most recent
269 recursion is that of the given number or name. It does not search out
270 through the entire recursion stack.
271
272 (g) The escape \gN or \g{N} has been added, where N is a positive or
273 negative number, specifying an absolute or relative reference.
274
275 35. Tidied to get rid of some further signed/unsigned compiler warnings and
276 some "unreachable code" warnings.
277
278 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
279 things, this adds five new scripts.
280
281 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
282 There were also incompatibilities regarding the handling of \Q..\E inside
283 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
284 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
285
286 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
287 matches an empty string, and forcibly breaks the loop. There were bugs in
288 this code in non-simple cases. For a pattern such as ^(a()*)* matched
289 against aaaa the result was just "a" rather than "aaaa", for example. Two
290 separate and independent bugs (that affected different cases) have been
291 fixed.
292
293 39. Refactored the code to abolish the use of different opcodes for small
294 capturing bracket numbers. This is a tidy that I avoided doing when I
295 removed the limit on the number of capturing brackets for 3.5 back in 2001.
296 The new approach is not only tidier, it makes it possible to reduce the
297 memory needed to fix the previous bug (38).
298
299 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
300 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
301 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
302 mode.
303
304 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
305 report.
306
307 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
308 copy construction and assignment in the C++ wrapper.
309
310 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
311 couple of bugs that could have given wrong results in the "--newline=crlf"
312 case.
313
314 44. Added a number of casts and did some reorganization of signed/unsigned int
315 variables following suggestions from Dair Grant. Also renamed the variable
316 "this" as "item" because it is a C++ keyword.
317
318 45. Arranged for dftables to add
319
320 #include "pcre_internal.h"
321
322 to pcre_chartables.c because without it, gcc 4.x may remove the array
323 definition from the final binary if PCRE is built into a static library and
324 dead code stripping is activated.
325
326 46. For an unanchored pattern, if a match attempt fails at the start of a
327 newline sequence, and the newline setting is CRLF or ANY, and the next two
328 characters are CRLF, advance by two characters instead of one.
329
330
331 Version 6.7 04-Jul-06
332 ---------------------
333
334 1. In order to handle tests when input lines are enormously long, pcretest has
335 been re-factored so that it automatically extends its buffers when
336 necessary. The code is crude, but this _is_ just a test program. The
337 default size has been increased from 32K to 50K.
338
339 2. The code in pcre_study() was using the value of the re argument before
340 testing it for NULL. (Of course, in any sensible call of the function, it
341 won't be NULL.)
342
343 3. The memmove() emulation function in pcre_internal.h, which is used on
344 systems that lack both memmove() and bcopy() - that is, hardly ever -
345 was missing a "static" storage class specifier.
346
347 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
348 containing an extended class (one that cannot be represented by a bitmap
349 because it contains high-valued characters or Unicode property items, e.g.
350 [\pZ]). Almost always one would set UTF-8 mode when processing such a
351 pattern, but PCRE should not loop if you do not (it no longer does).
352 [Detail: two cases were found: (a) a repeated subpattern containing an
353 extended class; (b) a recursive reference to a subpattern that followed a
354 previous extended class. It wasn't skipping over the extended class
355 correctly when UTF-8 mode was not set.]
356
357 5. A negated single-character class was not being recognized as fixed-length
358 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
359 compile error "lookbehind assertion is not fixed length".
360
361 6. The RunPerlTest auxiliary script was showing an unexpected difference
362 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
363 write a Perl script that can interpret lines of an input file either as
364 byte characters or as UTF-8, which is what "perltest" was being required to
365 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
366 can't do is switch easily at run time between having the "use utf8;" pragma
367 or not. In the end, I fudged it by using the RunPerlTest script to insert
368 "use utf8;" explicitly for the UTF-8 tests.
369
370 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
371 the end of the subject string, contrary to the documentation and to what
372 Perl does. This was true of both matching functions. Now it matches only at
373 the start of the subject and immediately after *internal* newlines.
374
375 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
376 a pointer to an int instead of a pointer to an unsigned long int. This
377 caused problems on 64-bit systems.
378
379 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
380 instance of the 'standard' template library not being so standard".
381
382 10. There was no check on the number of named subpatterns nor the maximum
383 length of a subpattern name. The product of these values is used to compute
384 the size of the memory block for a compiled pattern. By supplying a very
385 long subpattern name and a large number of named subpatterns, the size
386 computation could be caused to overflow. This is now prevented by limiting
387 the length of names to 32 characters, and the number of named subpatterns
388 to 10,000.
389
390 11. Subpatterns that are repeated with specific counts have to be replicated in
391 the compiled pattern. The size of memory for this was computed from the
392 length of the subpattern and the repeat count. The latter is limited to
393 65535, but there was no limit on the former, meaning that integer overflow
394 could in principle occur. The compiled length of a repeated subpattern is
395 now limited to 30,000 bytes in order to prevent this.
396
397 12. Added the optional facility to have named substrings with the same name.
398
399 13. Added the ability to use a named substring as a condition, using the
400 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
401 are numbers (not recommended). Forward references are permitted.
402
403 14. Added forward references in named backreferences (if you see what I mean).
404
405 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
406 pattern could run off the end of the subject. For example, the pattern
407 "(?s)(.{1,5})"8 did this with the subject "ab".
408
409 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
410 PCRE_CASELESS was set when matching characters that were quantified with ?
411 or *.
412
413 17. A character class other than a single negated character that had a minimum
414 but no maximum quantifier - for example [ab]{6,} - was not handled
415 correctly by pce_dfa_exec(). It would match only one character.
416
417 18. A valid (though odd) pattern that looked like a POSIX character
418 class but used an invalid character after [ (for example [[,abc,]]) caused
419 pcre_compile() to give the error "Failed: internal error: code overflow" or
420 in some cases to crash with a glibc free() error. This could even happen if
421 the pattern terminated after [[ but there just happened to be a sequence of
422 letters, a binary zero, and a closing ] in the memory that followed.
423
424 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
425 over the years. Originally (before any Unicode support), just the bottom 8
426 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
427 output from "man perlunicode" includes this:
428
429 The regular expression compiler produces polymorphic opcodes. That
430 is, the pattern adapts to the data and automatically switches to
431 the Unicode character scheme when presented with Unicode data--or
432 instead uses a traditional byte scheme when presented with byte
433 data.
434
435 Sadly, a wide octal escape does not cause a switch, and in a string with
436 no other multibyte characters, these octal escapes are treated as before.
437 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
438 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
439 Unicode string.
440
441 I have not perpetrated such confusion in PCRE. Up till now, it took just
442 the bottom 8 bits, as in old Perl. I have now made octal escapes with
443 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
444 translate to the appropriate multibyte character.
445
446 29. Applied some refactoring to reduce the number of warnings from Microsoft
447 and Borland compilers. This has included removing the fudge introduced
448 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
449 a warning about an unused variable.
450
451 21. PCRE has not included VT (character 0x0b) in the set of whitespace
452 characters since release 4.0, because Perl (from release 5.004) does not.
453 [Or at least, is documented not to: some releases seem to be in conflict
454 with the documentation.] However, when a pattern was studied with
455 pcre_study() and all its branches started with \s, PCRE still included VT
456 as a possible starting character. Of course, this did no harm; it just
457 caused an unnecessary match attempt.
458
459 22. Removed a now-redundant internal flag bit that recorded the fact that case
460 dependency changed within the pattern. This was once needed for "required
461 byte" processing, but is no longer used. This recovers a now-scarce options
462 bit. Also moved the least significant internal flag bit to the most-
463 significant bit of the word, which was not previously used (hangover from
464 the days when it was an int rather than a uint) to free up another bit for
465 the future.
466
467 23. Added support for CRLF line endings as well as CR and LF. As well as the
468 default being selectable at build time, it can now be changed at runtime
469 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
470 specify that it is scanning data with non-default line endings.
471
472 24. Changed the definition of CXXLINK to make it agree with the definition of
473 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
474
475 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
476 recursions. This makes a big different to stack usage for some patterns.
477
478 26. If a subpattern containing a named recursion or subroutine reference such
479 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
480 the space required for the compiled pattern went wrong and gave too small a
481 value. Depending on the environment, this could lead to "Failed: internal
482 error: code overflow at offset 49" or "glibc detected double free or
483 corruption" errors.
484
485 27. Applied patches from Google (a) to support the new newline modes and (b) to
486 advance over multibyte UTF-8 characters in GlobalReplace.
487
488 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
489 difference for some implementation of PCRE in some Windows version.
490
491 29. Added some extra testing facilities to pcretest:
492
493 \q<number> in a data line sets the "match limit" value
494 \Q<number> in a data line sets the "match recursion limt" value
495 -S <number> sets the stack size, where <number> is in megabytes
496
497 The -S option isn't available for Windows.
498
499
500 Version 6.6 06-Feb-06
501 ---------------------
502
503 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
504 in pcreposix.h. I have copied the definition from pcre.h.
505
506 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
507 because pcre.h is no longer a built file.
508
509 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
510 not normally included in the compiled code.
511
512
513 Version 6.5 01-Feb-06
514 ---------------------
515
516 1. When using the partial match feature with pcre_dfa_exec(), it was not
517 anchoring the second and subsequent partial matches at the new starting
518 point. This could lead to incorrect results. For example, with the pattern
519 /1234/, partially matching against "123" and then "a4" gave a match.
520
521 2. Changes to pcregrep:
522
523 (a) All non-match returns from pcre_exec() were being treated as failures
524 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
525 error message is output. Some extra information is given for the
526 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
527 probably the only errors that are likely to be caused by users (by
528 specifying a regex that has nested indefinite repeats, for instance).
529 If there are more than 20 of these errors, pcregrep is abandoned.
530
531 (b) A binary zero was treated as data while matching, but terminated the
532 output line if it was written out. This has been fixed: binary zeroes
533 are now no different to any other data bytes.
534
535 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
536 used to set a locale for matching. The --locale=xxxx long option has
537 been added (no short equivalent) to specify a locale explicitly on the
538 pcregrep command, overriding the environment variables.
539
540 (d) When -B was used with -n, some line numbers in the output were one less
541 than they should have been.
542
543 (e) Added the -o (--only-matching) option.
544
545 (f) If -A or -C was used with -c (count only), some lines of context were
546 accidentally printed for the final match.
547
548 (g) Added the -H (--with-filename) option.
549
550 (h) The combination of options -rh failed to suppress file names for files
551 that were found from directory arguments.
552
553 (i) Added the -D (--devices) and -d (--directories) options.
554
555 (j) Added the -F (--fixed-strings) option.
556
557 (k) Allow "-" to be used as a file name for -f as well as for a data file.
558
559 (l) Added the --colo(u)r option.
560
561 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
562 is not present by default.
563
564 3. A nasty bug was discovered in the handling of recursive patterns, that is,
565 items such as (?R) or (?1), when the recursion could match a number of
566 alternatives. If it matched one of the alternatives, but subsequently,
567 outside the recursion, there was a failure, the code tried to back up into
568 the recursion. However, because of the way PCRE is implemented, this is not
569 possible, and the result was an incorrect result from the match.
570
571 In order to prevent this happening, the specification of recursion has
572 been changed so that all such subpatterns are automatically treated as
573 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
574
575 4. I had overlooked the fact that, in some locales, there are characters for
576 which isalpha() is true but neither isupper() nor islower() are true. In
577 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
578 and ordfeminine) are like this. This affected the treatment of \w and \W
579 when they appeared in character classes, but not when they appeared outside
580 a character class. The bit map for "word" characters is now created
581 separately from the results of isalnum() instead of just taking it from the
582 upper, lower, and digit maps. (Plus the underscore character, of course.)
583
584 5. The above bug also affected the handling of POSIX character classes such as
585 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
586 permanent tables. Instead, the bit maps for such a class were previously
587 created as the appropriate unions of the upper, lower, and digit bitmaps.
588 Now they are created by subtraction from the [[:word:]] class, which has
589 its own bitmap.
590
591 6. The [[:blank:]] character class matches horizontal, but not vertical space.
592 It is created by subtracting the vertical space characters (\x09, \x0a,
593 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
594 subtraction was done in the overall bitmap for a character class, meaning
595 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
596 be recognized. This bug has been fixed.
597
598 7. Patches from the folks at Google:
599
600 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
601 real life, but is still worth protecting against".
602
603 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
604 regular expressions".
605
606 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
607 have it.
608
609 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
610 "configure" and the latter not, in order to fix a problem somebody had
611 with compiling the Arg class on HP-UX.
612
613 (e) Improve the error-handling of the C++ wrapper a little bit.
614
615 (f) New tests for checking recursion limiting.
616
617 8. The pcre_memmove() function, which is used only if the environment does not
618 have a standard memmove() function (and is therefore rarely compiled),
619 contained two bugs: (a) use of int instead of size_t, and (b) it was not
620 returning a result (though PCRE never actually uses the result).
621
622 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
623 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
624 returned instead of calling malloc() with an overflowing number that would
625 most likely cause subsequent chaos.
626
627 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
628
629 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
630 with this option is matched, the nmatch and pmatch options of regexec() are
631 ignored.
632
633 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
634 provided in case anyone wants to the the POSIX interface with UTF-8
635 strings.
636
637 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
638 C++ linking (needed for some HP-UX environments).
639
640 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
641 (unused parameter) and in the pcre_printint() function (omitted "default"
642 switch label when the default is to do nothing).
643
644 15. Added some code to make it possible, when PCRE is compiled as a C++
645 library, to replace subject pointers for pcre_exec() with a smart pointer
646 class, thus making it possible to process discontinuous strings.
647
648 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
649 much the same function. They were added by different people who were trying
650 to make PCRE easy to compile on non-Unix systems. It has been suggested
651 that PCRE_EXPORT be abolished now that there is more automatic apparatus
652 for compiling on Windows systems. I have therefore replaced it with
653 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
654 defaults to "extern" for C or "extern C" for C++, which works fine on
655 Unix-like systems. It is now possible to override the value of PCRE_DATA_
656 SCOPE with something explicit in config.h. In addition:
657
658 (a) pcreposix.h still had just "extern" instead of either of these macros;
659 I have replaced it with PCRE_DATA_SCOPE.
660
661 (b) Functions such as _pcre_xclass(), which are internal to the library,
662 but external in the C sense, all had PCRE_EXPORT in their definitions.
663 This is apparently wrong for the Windows case, so I have removed it.
664 (It makes no difference on Unix-like systems.)
665
666 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
667 of recursive calls to match(). This is different to MATCH_LIMIT because
668 that limits the total number of calls to match(), not all of which increase
669 the depth of recursion. Limiting the recursion depth limits the amount of
670 stack (or heap if NO_RECURSE is set) that is used. The default can be set
671 when PCRE is compiled, and changed at run time. A patch from Google adds
672 this functionality to the C++ interface.
673
674 18. Changes to the handling of Unicode character properties:
675
676 (a) Updated the table to Unicode 4.1.0.
677
678 (b) Recognize characters that are not in the table as "Cn" (undefined).
679
680 (c) I revised the way the table is implemented to a much improved format
681 which includes recognition of ranges. It now supports the ranges that
682 are defined in UnicodeData.txt, and it also amalgamates other
683 characters into ranges. This has reduced the number of entries in the
684 table from around 16,000 to around 3,000, thus reducing its size
685 considerably. I realized I did not need to use a tree structure after
686 all - a binary chop search is just as efficient. Having reduced the
687 number of entries, I extended their size from 6 bytes to 8 bytes to
688 allow for more data.
689
690 (d) Added support for Unicode script names via properties such as \p{Han}.
691
692 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
693 matching that character.
694
695 20. When matching a repeated Unicode property with a minimum greater than zero,
696 (for example \pL{2,}), PCRE could look past the end of the subject if it
697 reached it while seeking the minimum number of characters. This could
698 happen only if some of the characters were more than one byte long, because
699 there is a check for at least the minimum number of bytes.
700
701 21. Refactored the implementation of \p and \P so as to be more general, to
702 allow for more different types of property in future. This has changed the
703 compiled form incompatibly. Anybody with saved compiled patterns that use
704 \p or \P will have to recompile them.
705
706 22. Added "Any" and "L&" to the supported property types.
707
708 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
709 but give a compile time error if the value is greater than 0xff.
710
711 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
712 accidentally not being installed or uninstalled.
713
714 25. The pcre.h file was built from pcre.h.in, but the only changes that were
715 made were to insert the current release number. This seemed silly, because
716 it made things harder for people building PCRE on systems that don't run
717 "configure". I have turned pcre.h into a distributed file, no longer built
718 by "configure", with the version identification directly included. There is
719 no longer a pcre.h.in file.
720
721 However, this change necessitated a change to the pcre-config script as
722 well. It is built from pcre-config.in, and one of the substitutions was the
723 release number. I have updated configure.ac so that ./configure now finds
724 the release number by grepping pcre.h.
725
726 26. Added the ability to run the tests under valgrind.
727
728
729 Version 6.4 05-Sep-05
730 ---------------------
731
732 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
733 "--" to be printed when multiple files were scanned, even when none of the
734 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
735 consider it to be a bug, and have restored the previous behaviour.
736
737 2. A couple of code tidies to get rid of compiler warnings.
738
739 3. The pcretest program used to cheat by referring to symbols in the library
740 whose names begin with _pcre_. These are internal symbols that are not
741 really supposed to be visible externally, and in some environments it is
742 possible to suppress them. The cheating is now confined to including
743 certain files from the library's source, which is a bit cleaner.
744
745 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
746 file's purpose clearer.
747
748 5. Reorganized pcre_ucp_findchar().
749
750
751 Version 6.3 15-Aug-05
752 ---------------------
753
754 1. The file libpcre.pc.in did not have general read permission in the tarball.
755
756 2. There were some problems when building without C++ support:
757
758 (a) If C++ support was not built, "make install" and "make test" still
759 tried to test it.
760
761 (b) There were problems when the value of CXX was explicitly set. Some
762 changes have been made to try to fix these, and ...
763
764 (c) --disable-cpp can now be used to explicitly disable C++ support.
765
766 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
767 backslash in a target when C++ was disabled. This confuses some
768 versions of "make", apparently. Using an intermediate variable solves
769 this. (Same for CPP_LOBJ.)
770
771 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
772 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
773 necessary on certain architectures.
774
775 4. Added a setting of -export-symbols-regex to the link command to remove
776 those symbols that are exported in the C sense, but actually are local
777 within the library, and not documented. Their names all begin with
778 "_pcre_". This is not a perfect job, because (a) we have to except some
779 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
780 available (and never for static libraries). I have made a note to try to
781 find a way round (a) in the future.
782
783
784 Version 6.2 01-Aug-05
785 ---------------------
786
787 1. There was no test for integer overflow of quantifier values. A construction
788 such as {1111111111111111} would give undefined results. What is worse, if
789 a minimum quantifier for a parenthesized subpattern overflowed and became
790 negative, the calculation of the memory size went wrong. This could have
791 led to memory overwriting.
792
793 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
794
795 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
796 operating environments where this matters.
797
798 4. Applied Giuseppe Maxia's patch to add additional features for controlling
799 PCRE options from within the C++ wrapper.
800
801 5. Named capturing subpatterns were not being correctly counted when a pattern
802 was compiled. This caused two problems: (a) If there were more than 100
803 such subpatterns, the calculation of the memory needed for the whole
804 compiled pattern went wrong, leading to an overflow error. (b) Numerical
805 back references of the form \12, where the number was greater than 9, were
806 not recognized as back references, even though there were sufficient
807 previous subpatterns.
808
809 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
810 versions of gcc, e.g. 2.95.4.
811
812
813 Version 6.1 21-Jun-05
814 ---------------------
815
816 1. There was one reference to the variable "posix" in pcretest.c that was not
817 surrounded by "#if !defined NOPOSIX".
818
819 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
820 the cross-check on the old pcre_info() function, for the benefit of the
821 cut-down version of PCRE that is currently imported into Exim.
822
823 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
824 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
825 patterns that start (?i) but I don't think that matters. The use of (?i) is
826 just an example; this all applies to the other options as well.
827
828 4. Since libtool seems to echo the compile commands it is issuing, the output
829 from "make" can be reduced a bit by putting "@" in front of each libtool
830 compile command.
831
832 5. Patch from the folks at Google for configure.in to be a bit more thorough
833 in checking for a suitable C++ installation before trying to compile the
834 C++ stuff. This should fix a reported problem when a compiler was present,
835 but no suitable headers.
836
837 6. The man pages all had just "PCRE" as their title. I have changed them to
838 be the relevant file name. I have also arranged that these names are
839 retained in the file doc/pcre.txt, which is a concatenation in text format
840 of all the man pages except the little individual ones for each function.
841
842 7. The NON-UNIX-USE file had not been updated for the different set of source
843 files that come with release 6. I also added a few comments about the C++
844 wrapper.
845
846
847 Version 6.0 07-Jun-05
848 ---------------------
849
850 1. Some minor internal re-organization to help with my DFA experiments.
851
852 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
853 didn't matter for the library itself when fully configured, but did matter
854 when compiling without UCP support, or within Exim, where the ucp files are
855 not imported.
856
857 3. Refactoring of the library code to split up the various functions into
858 different source modules. The addition of the new DFA matching code (see
859 below) to a single monolithic source would have made it really too
860 unwieldy, quite apart from causing all the code to be include in a
861 statically linked application, when only some functions are used. This is
862 relevant even without the DFA addition now that patterns can be compiled in
863 one application and matched in another.
864
865 The downside of splitting up is that there have to be some external
866 functions and data tables that are used internally in different modules of
867 the library but which are not part of the API. These have all had their
868 names changed to start with "_pcre_" so that they are unlikely to clash
869 with other external names.
870
871 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
872 a different (DFA) algorithm. Although it is slower than the original
873 function, it does have some advantages for certain types of matching
874 problem.
875
876 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
877 including restarting after a partial match.
878
879 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
880 defined when compiling for Windows was sent to me. I have put it into the
881 code, though I have no means of testing or verifying it.
882
883 7. Added the pcre_refcount() auxiliary function.
884
885 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
886 match before or at the first newline in the subject string. In pcretest,
887 the /f option on a pattern can be used to set this.
888
889 9. A repeated \w when used in UTF-8 mode with characters greater than 256
890 would behave wrongly. This has been present in PCRE since release 4.0.
891
892 10. A number of changes to the pcregrep command:
893
894 (a) Refactored how -x works; insert ^(...)$ instead of setting
895 PCRE_ANCHORED and checking the length, in preparation for adding
896 something similar for -w.
897
898 (b) Added the -w (match as a word) option.
899
900 (c) Refactored the way lines are read and buffered so as to have more
901 than one at a time available.
902
903 (d) Implemented a pcregrep test script.
904
905 (e) Added the -M (multiline match) option. This allows patterns to match
906 over several lines of the subject. The buffering ensures that at least
907 8K, or the rest of the document (whichever is the shorter) is available
908 for matching (and similarly the previous 8K for lookbehind assertions).
909
910 (f) Changed the --help output so that it now says
911
912 -w, --word-regex(p)
913
914 instead of two lines, one with "regex" and the other with "regexp"
915 because that confused at least one person since the short forms are the
916 same. (This required a bit of code, as the output is generated
917 automatically from a table. It wasn't just a text change.)
918
919 (g) -- can be used to terminate pcregrep options if the next thing isn't an
920 option but starts with a hyphen. Could be a pattern or a path name
921 starting with a hyphen, for instance.
922
923 (h) "-" can be given as a file name to represent stdin.
924
925 (i) When file names are being printed, "(standard input)" is used for
926 the standard input, for compatibility with GNU grep. Previously
927 "<stdin>" was used.
928
929 (j) The option --label=xxx can be used to supply a name to be used for
930 stdin when file names are being printed. There is no short form.
931
932 (k) Re-factored the options decoding logic because we are going to add
933 two more options that take data. Such options can now be given in four
934 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
935
936 (l) Added the -A, -B, and -C options for requesting that lines of context
937 around matches be printed.
938
939 (m) Added the -L option to print the names of files that do not contain
940 any matching lines, that is, the complement of -l.
941
942 (n) The return code is 2 if any file cannot be opened, but pcregrep does
943 continue to scan other files.
944
945 (o) The -s option was incorrectly implemented. For compatibility with other
946 greps, it now suppresses the error message for a non-existent or non-
947 accessible file (but not the return code). There is a new option called
948 -q that suppresses the output of matching lines, which was what -s was
949 previously doing.
950
951 (p) Added --include and --exclude options to specify files for inclusion
952 and exclusion when recursing.
953
954 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
955 Hopefully, it now does.
956
957 12. Missing cast in pcre_study().
958
959 13. Added an "uninstall" target to the makefile.
960
961 14. Replaced "extern" in the function prototypes in Makefile.in with
962 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
963 world, but is set differently for Windows.
964
965 15. Added a second compiling function called pcre_compile2(). The only
966 difference is that it has an extra argument, which is a pointer to an
967 integer error code. When there is a compile-time failure, this is set
968 non-zero, in addition to the error test pointer being set to point to an
969 error message. The new argument may be NULL if no error number is required
970 (but then you may as well call pcre_compile(), which is now just a
971 wrapper). This facility is provided because some applications need a
972 numeric error indication, but it has also enabled me to tidy up the way
973 compile-time errors are handled in the POSIX wrapper.
974
975 16. Added VPATH=.libs to the makefile; this should help when building with one
976 prefix path and installing with another. (Or so I'm told by someone who
977 knows more about this stuff than I do.)
978
979 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
980 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
981 match everything, including newlines. This is not POSIX-compatible, but
982 somebody wanted the feature. From pcretest it can be activated by using
983 both the P and the s flags.
984
985 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
986
987 19. libpcre.pc was being incorrectly installed as executable.
988
989 20. A couple of places in pcretest check for end-of-line by looking for '\n';
990 it now also looks for '\r' so that it will work unmodified on Windows.
991
992 21. Added Google's contributed C++ wrapper to the distribution.
993
994 22. Added some untidy missing memory free() calls in pcretest, to keep
995 Electric Fence happy when testing.
996
997
998
999 Version 5.0 13-Sep-04
1000 ---------------------
1001
1002 1. Internal change: literal characters are no longer packed up into items
1003 containing multiple characters in a single byte-string. Each character
1004 is now matched using a separate opcode. However, there may be more than one
1005 byte in the character in UTF-8 mode.
1006
1007 2. The pcre_callout_block structure has two new fields: pattern_position and
1008 next_item_length. These contain the offset in the pattern to the next match
1009 item, and its length, respectively.
1010
1011 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1012 insertion of callouts before each pattern item. Added the /C option to
1013 pcretest to make use of this.
1014
1015 4. On the advice of a Windows user, the lines
1016
1017 #if defined(_WIN32) || defined(WIN32)
1018 _setmode( _fileno( stdout ), 0x8000 );
1019 #endif /* defined(_WIN32) || defined(WIN32) */
1020
1021 have been added to the source of pcretest. This apparently does useful
1022 magic in relation to line terminators.
1023
1024 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1025 for the benefit of those environments where the "b" makes a difference.
1026
1027 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1028 to know about it. I have put a hack into configure.in that adds in code
1029 to set GCC=yes if CC=icc. This seems to end up at a point in the
1030 generated configure script that is early enough to affect the setting of
1031 compiler options, which is what is needed, but I have no means of testing
1032 whether it really works. (The user who reported this had patched the
1033 generated configure script, which of course I cannot do.)
1034
1035 LATER: After change 22 below (new libtool files), the configure script
1036 seems to know about icc (and also ecc). Therefore, I have commented out
1037 this hack in configure.in.
1038
1039 7. Added support for pkg-config (2 patches were sent in).
1040
1041 8. Negated POSIX character classes that used a combination of internal tables
1042 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1043 [[:^ascii]]. Typically, they would match almost any characters. The other
1044 POSIX classes were not broken in this way.
1045
1046 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1047 to find the match, as PCRE was deluded into thinking that the match had to
1048 start at the start point or following a newline. The same bug applied to
1049 patterns with negative forward assertions or any backward assertions
1050 preceding ".*" at the start, unless the pattern required a fixed first
1051 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1052
1053 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1054 starting at the last subject character, bytes beyond the end of the subject
1055 string were read.
1056
1057 11. Renamed the variable "class" as "classbits" to make life easier for C++
1058 users. (Previously there was a macro definition, but it apparently wasn't
1059 enough.)
1060
1061 12. Added the new field "tables" to the extra data so that tables can be passed
1062 in at exec time, or the internal tables can be re-selected. This allows
1063 a compiled regex to be saved and re-used at a later time by a different
1064 program that might have everything at different addresses.
1065
1066 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1067 -R library as well as a -L library.
1068
1069 14. The debugging options of pcretest (-d on the command line or D on a
1070 pattern) showed incorrect output for anything following an extended class
1071 that contained multibyte characters and which was followed by a quantifier.
1072
1073 15. Added optional support for general category Unicode character properties
1074 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1075 support. It adds about 90K to the size of the library. The meanings of the
1076 inbuilt class escapes such as \d and \s have NOT been changed.
1077
1078 16. Updated pcredemo.c to include calls to free() to release the memory for the
1079 compiled pattern.
1080
1081 17. The generated file chartables.c was being created in the source directory
1082 instead of in the building directory. This caused the build to fail if the
1083 source directory was different from the building directory, and was
1084 read-only.
1085
1086 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1087 file. No doubt somebody will tell me if they don't make sense... Also added
1088 Dan Mooney's comments about building on OpenVMS.
1089
1090 19. Added support for partial matching via the PCRE_PARTIAL option for
1091 pcre_exec() and the \P data escape in pcretest.
1092
1093 20. Extended pcretest with 3 new pattern features:
1094
1095 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1096 write the compiled pattern to the file whose name is "rest-of-line".
1097 This is a straight binary dump of the data, with the saved pointer to
1098 the character tables forced to be NULL. The study data, if any, is
1099 written too. After writing, pcretest reads a new pattern.
1100
1101 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1102 compiled pattern from the given file. There must not be any
1103 occurrences of "<" in the file name (pretty unlikely); if there are,
1104 pcretest will instead treat the initial "<" as a pattern delimiter.
1105 After reading in the pattern, pcretest goes on to read data lines as
1106 usual.
1107
1108 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1109 and 16-bit fields in a compiled pattern, to simulate a pattern that
1110 was compiled on a host of opposite endianness.
1111
1112 21. The pcre-exec() function can now cope with patterns that were compiled on
1113 hosts of opposite endianness, with this restriction:
1114
1115 As for any compiled expression that is saved and used later, the tables
1116 pointer field cannot be preserved; the extra_data field in the arguments
1117 to pcre_exec() should be used to pass in a tables address if a value
1118 other than the default internal tables were used at compile time.
1119
1120 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1121 now diagnosed as an error. Previously, most of the time, a negative number
1122 would have been treated as zero, but if in addition "ovector" was passed as
1123 NULL, a crash could occur.
1124
1125 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1126 new versions from the libtool 1.5 distribution (the last one is a copy of
1127 a file called libtool.m4). This seems to have fixed the need to patch
1128 "configure" to support Darwin 1.3 (which I used to do). However, I still
1129 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1130 workstation).
1131
1132 24. Changed the PCRE licence to be the more standard "BSD" licence.
1133
1134
1135 Version 4.5 01-Dec-03
1136 ---------------------
1137
1138 1. There has been some re-arrangement of the code for the match() function so
1139 that it can be compiled in a version that does not call itself recursively.
1140 Instead, it keeps those local variables that need separate instances for
1141 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1142 needs to "recurse". Keeping track of where control must go is done by means
1143 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1144 hide most of the details from the main code, and operates only if
1145 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1146 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1147 operating.
1148
1149 To make it easier for callers to provide specially tailored get/free
1150 functions for this usage, two new functions, pcre_stack_malloc, and
1151 pcre_stack_free, are used. They are always called in strict stacking order,
1152 and the size of block requested is always the same.
1153
1154 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1155 PCRE has been compiled to use the stack or the heap for recursion. The
1156 -C option of pcretest uses this to show which version is compiled.
1157
1158 A new data escape \S, is added to pcretest; it causes the amounts of store
1159 obtained and freed by both kinds of malloc/free at match time to be added
1160 to the output.
1161
1162 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1163 what's available on my current Linux desktop machine.
1164
1165 3. When matching a UTF-8 string, the test for a valid string at the start has
1166 been extended. If start_offset is not zero, PCRE now checks that it points
1167 to a byte that is the start of a UTF-8 character. If not, it returns
1168 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1169 this is necessary because there may be backward assertions in the pattern.
1170 When matching the same subject several times, it may save resources to use
1171 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1172
1173 4. The code for checking the validity of UTF-8 strings has been tightened so
1174 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1175 containing "overlong sequences".
1176
1177 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1178 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1179 should have been "&", but it just so happened that all the cases this let
1180 through by mistake were picked up later in the function.
1181
1182 6. I had used a variable called "isblank" - this is a C99 function, causing
1183 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1184
1185 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1186 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1187 so the version is shown; (c) stop "make test" echoing "./RunTest".
1188
1189 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1190
1191 9. The prototype for memmove() for systems that don't have it was using
1192 size_t, but the inclusion of the header that defines size_t was later. I've
1193 moved the #includes for the C headers earlier to avoid this.
1194
1195 10. Added some adjustments to the code to make it easier to compiler on certain
1196 special systems:
1197
1198 (a) Some "const" qualifiers were missing.
1199 (b) Added the macro EXPORT before all exported functions; by default this
1200 is defined to be empty.
1201 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1202 that it reads its output file name as an argument instead of writing
1203 to the standard output and assuming this can be redirected.
1204
1205 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1206 class containing characters with values greater than 255, PCRE compilation
1207 went into a loop.
1208
1209 12. A recursive reference to a subpattern that was within another subpattern
1210 that had a minimum quantifier of zero caused PCRE to crash. For example,
1211 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1212 recursion. If the recursively-called subpattern itself had a zero repeat,
1213 that was OK.
1214
1215 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1216 buffer into which it was copied (for escape processing) was still set at
1217 1024, so long lines caused crashes.
1218
1219 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1220 "internal error: code overflow...". This applied to any character class
1221 that was followed by a possessive quantifier.
1222
1223 15. Modified the Makefile to add libpcre.la as a prerequisite for
1224 libpcreposix.la because I was told this is needed for a parallel build to
1225 work.
1226
1227 16. If a pattern that contained .* following optional items at the start was
1228 studied, the wrong optimizing data was generated, leading to matching
1229 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1230 matching string must start with a or b or c. The correct conclusion for
1231 this pattern is that a match can start with any character.
1232
1233
1234 Version 4.4 13-Aug-03
1235 ---------------------
1236
1237 1. In UTF-8 mode, a character class containing characters with values between
1238 127 and 255 was not handled correctly if the compiled pattern was studied.
1239 In fixing this, I have also improved the studying algorithm for such
1240 classes (slightly).
1241
1242 2. Three internal functions had redundant arguments passed to them. Removal
1243 might give a very teeny performance improvement.
1244
1245 3. Documentation bug: the value of the capture_top field in a callout is *one
1246 more than* the number of the hightest numbered captured substring.
1247
1248 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1249 in incorrectly linking with a previously installed version. They now link
1250 explicitly with libpcre.la.
1251
1252 5. configure.in no longer needs to recognize Cygwin specially.
1253
1254 6. A problem in pcre.in for Windows platforms is fixed.
1255
1256 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1257 pcretest, it used to include the size of the study block as part of its
1258 output. Unfortunately, the structure contains a field that has a different
1259 size on different hardware architectures. This meant that the tests that
1260 showed this size failed. As the block is currently always of a fixed size,
1261 this information isn't actually particularly useful in pcretest output, so
1262 I have just removed it.
1263
1264 8. Three pre-processor statements accidentally did not start in column 1.
1265 Sadly, there are *still* compilers around that complain, even though
1266 standard C has not required this for well over a decade. Sigh.
1267
1268 9. In pcretest, the code for checking callouts passed small integers in the
1269 callout_data field, which is a void * field. However, some picky compilers
1270 complained about the casts involved for this on 64-bit systems. Now
1271 pcretest passes the address of the small integer instead, which should get
1272 rid of the warnings.
1273
1274 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1275 both compile and run time, and gives an error if an invalid UTF-8 sequence
1276 is found. There is a option for disabling this check in cases where the
1277 string is known to be correct and/or the maximum performance is wanted.
1278
1279 11. In response to a bug report, I changed one line in Makefile.in from
1280
1281 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1282 to
1283 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1284
1285 to look similar to other lines, but I have no way of telling whether this
1286 is the right thing to do, as I do not use Windows. No doubt I'll get told
1287 if it's wrong...
1288
1289
1290 Version 4.3 21-May-03
1291 ---------------------
1292
1293 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1294 Makefile.
1295
1296 2. Some refactoring to improve the quality of the code:
1297
1298 (i) The utf8_table... variables are now declared "const".
1299
1300 (ii) The code for \cx, which used the "case flipping" table to upper case
1301 lower case letters, now just substracts 32. This is ASCII-specific,
1302 but the whole concept of \cx is ASCII-specific, so it seems
1303 reasonable.
1304
1305 (iii) PCRE was using its character types table to recognize decimal and
1306 hexadecimal digits in the pattern. This is silly, because it handles
1307 only 0-9, a-f, and A-F, but the character types table is locale-
1308 specific, which means strange things might happen. A private
1309 table is now used for this - though it costs 256 bytes, a table is
1310 much faster than multiple explicit tests. Of course, the standard
1311 character types table is still used for matching digits in subject
1312 strings against \d.
1313
1314 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1315 ending in _t are). So I've renamed it as ESC_tee.
1316
1317 3. The first argument for regexec() in the POSIX wrapper should have been
1318 defined as "const".
1319
1320 4. Changed pcretest to use malloc() for its buffers so that they can be
1321 Electric Fenced for debugging.
1322
1323 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1324 to read one or more bytes before the start of the subject string. Often this
1325 had no effect on PCRE's behaviour, but in some circumstances it could
1326 provoke a segmentation fault.
1327
1328 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1329 to try to read one or more bytes before the start of the subject string.
1330
1331 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1332 UTF-8 support could misbehave in various ways if the subject string
1333 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1334 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1335 back over UTF-8 characters.)
1336
1337
1338 Version 4.2 14-Apr-03
1339 ---------------------
1340
1341 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1342
1343 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1344 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1345 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1346 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1347 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1348 and BUILD_EXEEXT
1349 Note: automatic setting of the BUILD variables is not yet working
1350 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1351 compile-time but not at link-time
1352 [LINK]: use for linking executables only
1353 make different versions for Windows and non-Windows
1354 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1355 libraries
1356 [LINK_FOR_BUILD]: new variable
1357 [OBJEXT]: use throughout
1358 [EXEEXT]: use throughout
1359 <winshared>: new target
1360 <wininstall>: new target
1361 <dftables.o>: use native compiler
1362 <dftables>: use native linker
1363 <install>: handle Windows platform correctly
1364 <clean>: ditto
1365 <check>: ditto
1366 copy DLL to top builddir before testing
1367
1368 As part of these changes, -no-undefined was removed again. This was reported
1369 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1370 in any case.
1371
1372 3. Some tidies to get rid of compiler warnings:
1373
1374 . In the match_data structure, match_limit was an unsigned long int, whereas
1375 match_call_count was an int. I've made them both unsigned long ints.
1376
1377 . In pcretest the fact that a const uschar * doesn't automatically cast to
1378 a void * provoked a warning.
1379
1380 . Turning on some more compiler warnings threw up some "shadow" variables
1381 and a few more missing casts.
1382
1383 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1384 option, a class that contained a single character with a value between 128
1385 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1386
1387 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1388 option, a class that contained several characters, but with at least one
1389 whose value was between 128 and 255 caused PCRE to crash.
1390
1391
1392 Version 4.1 12-Mar-03
1393 ---------------------
1394
1395 1. Compiling with gcc -pedantic found a couple of places where casts were
1396 needed, and a string in dftables.c that was longer than standard compilers are
1397 required to support.
1398
1399 2. Compiling with Sun's compiler found a few more places where the code could
1400 be tidied up in order to avoid warnings.
1401
1402 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1403 first of these names is deprecated in the latest Autoconf in favour of the name
1404 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1405 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1406 analogy I have changed it to CFLAGS_FOR_BUILD.
1407
1408 4. Added -no-undefined to the linking command in the Makefile, because this is
1409 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1410 linking step for the pcreposix library.
1411
1412 5. PCRE was failing to diagnose the case of two named groups with the same
1413 name.
1414
1415 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1416 literal character that is needed in the subject for a match, and scans along to
1417 ensure that it is present before embarking on the full matching process. This
1418 saves time in cases of nested unlimited repeats that are never going to match.
1419 Problem: the scan can take a lot of time if the subject is very long (e.g.
1420 megabytes), thus penalizing straightforward matches. It is now done only if the
1421 amount of subject to be scanned is less than 1000 bytes.
1422
1423 7. A lesser problem with the same optimization is that it was recording the
1424 first character of an anchored pattern as "needed", thus provoking a search
1425 right along the subject, even when the first match of the pattern was going to
1426 fail. The "needed" character is now not set for anchored patterns, unless it
1427 follows something in the pattern that is of non-fixed length. Thus, it still
1428 fulfils its original purpose of finding quick non-matches in cases of nested
1429 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1430
1431
1432 Version 4.0 17-Feb-03
1433 ---------------------
1434
1435 1. If a comment in an extended regex that started immediately after a meta-item
1436 extended to the end of string, PCRE compiled incorrect data. This could lead to
1437 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
1438
1439 2. Moved to autoconf 2.53 and libtool 1.4.2.
1440
1441 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
1442 the special perltest8 script is no longer needed - all the tests can be run
1443 from a single perltest script.
1444
1445 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
1446 by \s. It has now been removed in PCRE. This means it isn't recognized as
1447 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
1448 class [:space:] *does* include VT, thereby creating a mess.
1449
1450 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
1451 space and tab.
1452
1453 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
1454 its new features into the main test script, reducing the number of scripts.
1455
1456 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
1457 were backward compatible, and made the (?i) apply to the whole pattern, as if
1458 /i were given. Now it behaves more logically, and applies the option setting
1459 only to what follows. PCRE has been changed to follow suit. However, if it
1460 finds options settings right at the start of the pattern, it extracts them into
1461 the global options, as before. Thus, they show up in the info data.
1462
1463 8. Added support for the \Q...\E escape sequence. Characters in between are
1464 treated as literals. This is slightly different from Perl in that $ and @ are
1465 also handled as literals inside the quotes. In Perl, they will cause variable
1466 interpolation. Note the following examples:
1467
1468 Pattern PCRE matches Perl matches
1469
1470 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
1471 \Qabc\$xyz\E abc\$xyz abc\$xyz
1472 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
1473
1474 For compatibility with Perl, \Q...\E sequences are recognized inside character
1475 classes as well as outside them.
1476
1477 9. Re-organized 3 code statements in pcretest to avoid "overflow in
1478 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
1479 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
1480 signed/unsigned warnings.
1481
1482 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
1483 option for pcretest, so I've replaced it by a simple function that does just
1484 that job.
1485
1486 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
1487 "pcregrep -".
1488
1489 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
1490 Java package. This provides some syntactic sugar for simple cases of what my
1491 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
1492 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
1493 item, you can use this simplified notation. Note that only makes sense with
1494 greedy quantifiers. Consequently, the use of the possessive quantifier forces
1495 greediness, whatever the setting of the PCRE_UNGREEDY option.
1496
1497 13. A change of greediness default within a pattern was not taking effect at
1498 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
1499 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
1500 was abstracted outside.
1501
1502 14. PCRE now supports the \G assertion. It is true when the current matching
1503 position is at the start point of the match. This differs from \A when the
1504 starting offset is non-zero. Used with the /g option of pcretest (or similar
1505 code), it works in the same way as it does for Perl's /g option. If all
1506 alternatives of a regex begin with \G, the expression is anchored to the start
1507 match position, and the "anchored" flag is set in the compiled expression.
1508
1509 15. Some bugs concerning the handling of certain option changes within patterns
1510 have been fixed. These applied to options other than (?ims). For example,
1511 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
1512 been the other way round. Some of this was related to change 7 above.
1513
1514 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
1515 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
1516 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
1517 POSIX classes only within a class (e.g. /[[:alpha:]]/).
1518
1519 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
1520 mode. Unlike ".", it always matches newline, whatever the setting of
1521 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
1522 assertions. Perl allows it, but it doesn't (in general) work because it can't
1523 calculate the length of the lookbehind. At least, that's the case for Perl
1524 5.8.0 - I've been told they are going to document that it doesn't work in
1525 future.
1526
1527 18. Added an error diagnosis for escapes that PCRE does not support: these are
1528 \L, \l, \N, \P, \p, \U, \u, and \X.
1529
1530 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
1531 reading past the end of the pattern in cases such as /[abcd/.
1532
1533 20. PCRE was getting more memory than necessary for patterns with classes that
1534 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
1535
1536 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
1537 compiling PCRE for use with Virtual Pascal.
1538
1539 22. Small fix to the Makefile to make it work properly if the build is done
1540 outside the source tree.
1541
1542 23. Added a new extension: a condition to go with recursion. If a conditional
1543 subpattern starts with (?(R) the "true" branch is used if recursion has
1544 happened, whereas the "false" branch is used only at the top level.
1545
1546 24. When there was a very long string of literal characters (over 255 bytes
1547 without UTF support, over 250 bytes with UTF support), the computation of how
1548 much memory was required could be incorrect, leading to segfaults or other
1549 strange effects.
1550
1551 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
1552 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
1553 there was a subsequent back reference to those brackets. This meant that, for
1554 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
1555 possible to check for precisely this case. All we can do is abandon the
1556 optimization if .* occurs inside capturing brackets when there are any back
1557 references whatsoever. (See below for a better fix that came later.)
1558
1559 26. The handling of the optimization for finding the first character of a
1560 non-anchored pattern, and for finding a character that is required later in the
1561 match were failing in some cases. This didn't break the matching; it just
1562 failed to optimize when it could. The way this is done has been re-implemented.
1563
1564 27. Fixed typo in error message for invalid (?R item (it said "(?p").
1565
1566 28. Added a new feature that provides some of the functionality that Perl
1567 provides with (?{...}). The facility is termed a "callout". The way it is done
1568 in PCRE is for the caller to provide an optional function, by setting
1569 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
1570 global variable. By default it is unset, which disables all calling out. To get
1571 the function called, the regex must include (?C) at appropriate points. This
1572 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
1573 This provides a means of identifying different callout points. When PCRE
1574 reaches such a point in the regex, if pcre_callout has been set, the external
1575 function is called. It is provided with data in a structure called
1576 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
1577 matching continues; if it returns a non-zero value, the match at the current
1578 point fails. However, backtracking will occur if possible. [This was changed
1579 later and other features added - see item 49 below.]
1580
1581 29. pcretest is upgraded to test the callout functionality. It provides a
1582 callout function that displays information. By default, it shows the start of
1583 the match and the current position in the text. There are some new data escapes
1584 to vary what happens:
1585
1586 \C+ in addition, show current contents of captured substrings
1587 \C- do not supply a callout function
1588 \C!n return 1 when callout number n is reached
1589 \C!n!m return 1 when callout number n is reached for the mth time
1590
1591 30. If pcregrep was called with the -l option and just a single file name, it
1592 output "<stdin>" if a match was found, instead of the file name.
1593
1594 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
1595 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
1596 pcre_exec(). This saves a malloc/free per call. The default value of
1597 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
1598 when configuring.
1599
1600 32. The default maximum size of a compiled pattern is 64K. There have been a
1601 few cases of people hitting this limit. The code now uses macros to handle the
1602 storing of links as offsets within the compiled pattern. It defaults to 2-byte
1603 links, but this can be changed to 3 or 4 bytes by --with-link-size when
1604 configuring. Tests 2 and 5 work only with 2-byte links because they output
1605 debugging information about compiled patterns.
1606
1607 33. Internal code re-arrangements:
1608
1609 (a) Moved the debugging function for printing out a compiled regex into
1610 its own source file (printint.c) and used #include to pull it into
1611 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
1612 separate copies.
1613
1614 (b) Defined the list of op-code names for debugging as a macro in
1615 internal.h so that it is next to the definition of the opcodes.
1616
1617 (c) Defined a table of op-code lengths for simpler skipping along compiled
1618 code. This is again a macro in internal.h so that it is next to the
1619 definition of the opcodes.
1620
1621 34. Added support for recursive calls to individual subpatterns, along the
1622 lines of Robin Houston's patch (but implemented somewhat differently).
1623
1624 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
1625 allow it to read and process whole directories in Win32. This code was
1626 contributed by Lionel Fourquaux; it has not been tested by me.
1627
1628 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
1629 used to name a group. Names consist of alphanumerics and underscores, and must
1630 be unique. Back references use the syntax (?P=name) and recursive calls use
1631 (?P>name) which is a PCRE extension to the Python extension. Groups still have
1632 numbers. The function pcre_fullinfo() can be used after compilation to extract
1633 a name/number map. There are three relevant calls:
1634
1635 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
1636 PCRE_INFO_NAMECOUNT yields the number of entries
1637 PCRE_INFO_NAMETABLE yields a pointer to the map.
1638
1639 The map is a vector of fixed-size entries. The size of each entry depends on
1640 the length of the longest name used. The first two bytes of each entry are the
1641 group number, most significant byte first. There follows the corresponding
1642 name, zero terminated. The names are in alphabetical order.
1643
1644 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
1645 case instead of 255. Making it the same both with and without UTF-8 support
1646 means that the same test output works with both.
1647
1648 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
1649 calling malloc() with a zero argument.
1650
1651 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
1652 optimization. I've improved things by keeping a bitmap of backreferences with
1653 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
1654 fact referenced, the optimization can be applied. It is unlikely that a
1655 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
1656 the match to follow \n) will appear inside brackets with a number greater than
1657 31, but if it does, any back reference > 31 suppresses the optimization.
1658
1659 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
1660 of disabling numbered capturing parentheses. Any opening parenthesis that is
1661 not followed by ? behaves as if it were followed by ?: but named parentheses
1662 can still be used for capturing (and they will acquire numbers in the usual
1663 way).
1664
1665 41. Redesigned the return codes from the match() function into yes/no/error so
1666 that errors can be passed back from deep inside the nested calls. A malloc
1667 failure while inside a recursive subpattern call now causes the
1668 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
1669
1670 42. It is now possible to set a limit on the number of times the match()
1671 function is called in a call to pcre_exec(). This facility makes it possible to
1672 limit the amount of recursion and backtracking, though not in a directly
1673 obvious way, because the match() function is used in a number of different
1674 circumstances. The count starts from zero for each position in the subject
1675 string (for non-anchored patterns). The default limit is, for compatibility, a
1676 large number, namely 10 000 000. You can change this in two ways:
1677
1678 (a) When configuring PCRE before making, you can use --with-match-limit=n
1679 to set a default value for the compiled library.
1680
1681 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
1682 a different value is set. See 45 below.
1683
1684 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1685
1686 43. Added a new function pcre_config(int, void *) to enable run-time extraction
1687 of things that can be changed at compile time. The first argument specifies
1688 what is wanted and the second points to where the information is to be placed.
1689 The current list of available information is:
1690
1691 PCRE_CONFIG_UTF8
1692
1693 The output is an integer that is set to one if UTF-8 support is available;
1694 otherwise it is set to zero.
1695
1696 PCRE_CONFIG_NEWLINE
1697
1698 The output is an integer that it set to the value of the code that is used for
1699 newline. It is either LF (10) or CR (13).
1700
1701 PCRE_CONFIG_LINK_SIZE
1702
1703 The output is an integer that contains the number of bytes used for internal
1704 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
1705
1706 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
1707
1708 The output is an integer that contains the threshold above which the POSIX
1709 interface uses malloc() for output vectors. See item 31 above.
1710
1711 PCRE_CONFIG_MATCH_LIMIT
1712
1713 The output is an unsigned integer that contains the default limit of the number
1714 of match() calls in a pcre_exec() execution. See 42 above.
1715
1716 44. pcretest has been upgraded by the addition of the -C option. This causes it
1717 to extract all the available output from the new pcre_config() function, and to
1718 output it. The program then exits immediately.
1719
1720 45. A need has arisen to pass over additional data with calls to pcre_exec() in
1721 order to support additional features. One way would have been to define
1722 pcre_exec2() (for example) with extra arguments, but this would not have been
1723 extensible, and would also have required all calls to the original function to
1724 be mapped to the new one. Instead, I have chosen to extend the mechanism that
1725 is used for passing in "extra" data from pcre_study().
1726
1727 The pcre_extra structure is now exposed and defined in pcre.h. It currently
1728 contains the following fields:
1729
1730 flags a bitmap indicating which of the following fields are set
1731 study_data opaque data from pcre_study()
1732 match_limit a way of specifying a limit on match() calls for a specific
1733 call to pcre_exec()
1734 callout_data data for callouts (see 49 below)
1735
1736 The flag bits are also defined in pcre.h, and are
1737
1738 PCRE_EXTRA_STUDY_DATA
1739 PCRE_EXTRA_MATCH_LIMIT
1740 PCRE_EXTRA_CALLOUT_DATA
1741
1742 The pcre_study() function now returns one of these new pcre_extra blocks, with
1743 the actual study data pointed to by the study_data field, and the
1744 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
1745 before. That is, this change is entirely upwards-compatible and requires no
1746 change to existing code.
1747
1748 If you want to pass in additional data to pcre_exec(), you can either place it
1749 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
1750 block.
1751
1752 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
1753 data string contains the escape sequence \M, pcretest calls pcre_exec() several
1754 times with different match limits, until it finds the minimum value needed for
1755 pcre_exec() to complete. The value is then output. This can be instructive; for
1756 most simple matches the number is quite small, but for pathological cases it
1757 gets very large very quickly.
1758
1759 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
1760 returns the size of the data block pointed to by the study_data field in a
1761 pcre_extra block, that is, the value that was passed as the argument to
1762 pcre_malloc() when PCRE was getting memory in which to place the information
1763 created by pcre_study(). The fourth argument should point to a size_t variable.
1764 pcretest has been extended so that this information is shown after a successful
1765 pcre_study() call when information about the compiled regex is being displayed.
1766
1767 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
1768 because what follows is always an absolute path. (Later: it turns out that this
1769 is more than cosmetic for MinGW, because it doesn't like empty path
1770 components.)
1771
1772 49. Some changes have been made to the callout feature (see 28 above):
1773
1774 (i) A callout function now has three choices for what it returns:
1775
1776 0 => success, carry on matching
1777 > 0 => failure at this point, but backtrack if possible
1778 < 0 => serious error, return this value from pcre_exec()
1779
1780 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
1781 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
1782 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
1783 use by callout functions. It will never be used by PCRE itself.
1784
1785 (ii) The pcre_extra structure (see 45 above) has a void * field called
1786 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
1787 pcre_callout_block structure has a field of the same name. The contents of
1788 the field passed in the pcre_extra structure are passed to the callout
1789 function in the corresponding field in the callout block. This makes it
1790 easier to use the same callout-containing regex from multiple threads. For
1791 testing, the pcretest program has a new data escape
1792
1793 \C*n pass the number n (may be negative) as callout_data
1794
1795 If the callout function in pcretest receives a non-zero value as
1796 callout_data, it returns that value.
1797
1798 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
1799 there were some redundant $(CFLAGS) in commands that are now specified as
1800 $(LINK), which already includes $(CFLAGS).
1801
1802 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
1803 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
1804 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
1805 one-byte characters throughout. Note that case-insensitive matching applies
1806 only to characters whose values are less than 256. PCRE doesn't support the
1807 notion of cases for higher-valued characters.
1808
1809 (i) A character class whose characters are all within 0-255 is handled as
1810 a bit map, and the map is inverted for negative classes. Previously, a
1811 character > 255 always failed to match such a class; however it should
1812 match if the class was a negative one (e.g. [^ab]). This has been fixed.
1813
1814 (ii) A negated character class with a single character < 255 is coded as
1815 "not this character" (OP_NOT). This wasn't working properly when the test
1816 character was multibyte, either singly or repeated.
1817
1818 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
1819 mode, for example: \x{100}{2,3}.
1820
1821 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
1822 singly or repeated) now correctly test multibyte characters. However,
1823 PCRE doesn't recognize any characters with values greater than 255 as
1824 digits, spaces, or word characters. Such characters always match \D, \S,
1825 and \W, and never match \d, \s, or \w.
1826
1827 (v) Classes may now contain characters and character ranges with values
1828 greater than 255. For example: [ab\x{100}-\x{400}].
1829
1830 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
1831 PCRE in UTF-8 mode.
1832
1833 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
1834 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
1835 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
1836 value.)
1837
1838 53. The single man page has become too large. I have therefore split it up into
1839 a number of separate man pages. These also give rise to individual HTML pages;
1840 these are now put in a separate directory, and there is an index.html page that
1841 lists them all. Some hyperlinking between the pages has been installed.
1842
1843 54. Added convenience functions for handling named capturing parentheses.
1844
1845 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
1846 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
1847 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
1848 are faulted.
1849
1850 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
1851 calling configure. These values are used when compiling the dftables.c program
1852 which is run to generate the source of the default character tables. They
1853 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
1854 you will need to set these values.
1855
1856 57. Updated the building process for Windows DLL, as provided by Fred Cox.
1857
1858
1859 Version 3.9 02-Jan-02
1860 ---------------------
1861
1862 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
1863
1864 2. If --disable-static was given, the building process failed when trying to
1865 build pcretest and pcregrep. (For some reason it was using libtool to compile
1866 them, which is not right, as they aren't part of the library.)
1867
1868
1869 Version 3.8 18-Dec-01
1870 ---------------------
1871
1872 1. The experimental UTF-8 code was completely screwed up. It was packing the
1873 bytes in the wrong order. How dumb can you get?
1874
1875
1876 Version 3.7 29-Oct-01
1877 ---------------------
1878
1879 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
1880 This caused pcretest, when used on the test data, to segfault. Unfortunately,
1881 this didn't happen under Solaris 8, where I normally test things.
1882
1883 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
1884 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
1885 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
1886 this fix an hour or so after the initial 3.7 release.)
1887
1888
1889 Version 3.6 23-Oct-01
1890 ---------------------
1891
1892 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
1893 offsets passed as NULL with zero offset count.
1894
1895 2. The config.guess and config.sub files had not been updated when I moved to
1896 the latest autoconf.
1897
1898
1899 Version 3.5 15-Aug-01
1900 ---------------------
1901
1902 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
1903 had been forgotten.
1904
1905 2. By using declared but undefined structures, we can avoid using "void"
1906 definitions in pcre.h while keeping the internal definitions of the structures
1907 private.
1908
1909 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
1910 user point of view, this means that both static and shared libraries are built
1911 by default, but this can be individually controlled. More of the work of
1912 handling this static/shared cases is now inside libtool instead of PCRE's make
1913 file.
1914
1915 4. The pcretest utility is now installed along with pcregrep because it is
1916 useful for users (to test regexs) and by doing this, it automatically gets
1917 relinked by libtool. The documentation has been turned into a man page, so
1918 there are now .1, .txt, and .html versions in /doc.
1919
1920 5. Upgrades to pcregrep:
1921 (i) Added long-form option names like gnu grep.
1922 (ii) Added --help to list all options with an explanatory phrase.
1923 (iii) Added -r, --recursive to recurse into sub-directories.
1924 (iv) Added -f, --file to read patterns from a file.
1925
1926 6. pcre_exec() was referring to its "code" argument before testing that
1927 argument for NULL (and giving an error if it was NULL).
1928
1929 7. Upgraded Makefile.in to allow for compiling in a different directory from
1930 the source directory.
1931
1932 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
1933 options bits, the pointer it was passed was to an int instead of to an unsigned
1934 long int. This mattered only on 64-bit systems.
1935
1936 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
1937 generated) instead of pcre.in, which it its source. Also made the same change
1938 in several of the .c files.
1939
1940 10. A new release of gcc defines printf() as a macro, which broke pcretest
1941 because it had an ifdef in the middle of a string argument for printf(). Fixed
1942 by using separate calls to printf().
1943
1944 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
1945 script, to force use of CR or LF instead of \n in the source. On non-Unix
1946 systems, the value can be set in config.h.
1947
1948 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
1949 absolute limit. Changed the text of the error message to make this clear, and
1950 likewise updated the man page.
1951
1952 13. The limit of 99 on the number of capturing subpatterns has been removed.
1953 The new limit is 65535, which I hope will not be a "real" limit.
1954
1955
1956 Version 3.4 22-Aug-00
1957 ---------------------
1958
1959 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
1960
1961 2. Diagnose condition (?(0) as an error instead of crashing on matching.
1962
1963
1964 Version 3.3 01-Aug-00
1965 ---------------------
1966
1967 1. If an octal character was given, but the value was greater than \377, it
1968 was not getting masked to the least significant bits, as documented. This could
1969 lead to crashes in some systems.
1970
1971 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
1972 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
1973
1974 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
1975 These just pass their arguments on to (pcre_free)(), but they are provided
1976 because some uses of PCRE bind it to non-C systems that can call its functions,
1977 but cannot call free() or pcre_free() directly.
1978
1979 4. Add "make test" as a synonym for "make check". Corrected some comments in
1980 the Makefile.
1981
1982 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
1983 Makefile.
1984
1985 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
1986 command called pgrep for grepping around the active processes.
1987
1988 7. Added the beginnings of support for UTF-8 character strings.
1989
1990 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
1991 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
1992 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
1993 out for the ar command.)
1994
1995
1996 Version 3.2 12-May-00
1997 ---------------------
1998
1999 This is purely a bug fixing release.
2000
2001 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2002 of ZA. This was just one example of several cases that could provoke this bug,
2003 which was introduced by change 9 of version 2.00. The code for breaking
2004 infinite loops after an iteration that matches an empty string was't working
2005 correctly.
2006
2007 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2008 when matched against abbab (for example). After matching an empty string, it
2009 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2010 caused it to match further down the string than it should.
2011
2012 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2013 was there because it doesn't seem to be needed, and it causes trouble on some
2014 systems, as it is not a Standard C header. It has been removed.
2015
2016 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2017 were reported on the Macintosh. The changes were from
2018
2019 while ((c = *(++ptr)) != 0 && c != '\n');
2020 to
2021 while ((c = *(++ptr)) != 0 && c != '\n') ;
2022
2023 Totally extraordinary, but if that's what it takes...
2024
2025 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2026 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2027 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2028 assumes the way PCRE uses memmove() (always moving upwards).
2029
2030 6. PCRE is being used in one environment where strchr() is not available. There
2031 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2032 faster code anyway.
2033
2034
2035 Version 3.1 09-Feb-00
2036 ---------------------
2037
2038 The only change in this release is the fixing of some bugs in Makefile.in for
2039 the "install" target:
2040
2041 (1) It was failing to install pcreposix.h.
2042
2043 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2044
2045
2046 Version 3.0 01-Feb-00
2047 ---------------------
2048
2049 1. Add support for the /+ modifier to perltest (to output $` like it does in
2050 pcretest).
2051
2052 2. Add support for the /g modifier to perltest.
2053
2054 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2055 matches null strings.
2056
2057 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2058 pattern. Perl treats empty patterns specially - it reuses the most recent
2059 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2060 effect.
2061
2062 5. The POSIX interface was broken in that it was just handing over the POSIX
2063 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2064 required a bigger vector, with some working space on the end. This means that
2065 the POSIX wrapper now has to get and free some memory, and copy the results.
2066
2067 6. Added some simple autoconf support, placing the test data and the
2068 documentation in separate directories, re-organizing some of the
2069 information files, and making it build pcre-config (a GNU standard). Also added
2070 libtool support for building PCRE as a shared library, which is now the
2071 default.
2072
2073 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2074 09 are not valid octal constants. Single digits will be used for minor values
2075 less than 10.
2076
2077 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2078 existing programs that set these in the POSIX interface can use PCRE without
2079 modification.
2080
2081 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2082 return all that pcre_info() returns, plus additional data. The pcre_info()
2083 function is retained for compatibility, but is considered to be obsolete.
2084
2085 10. Added experimental recursion feature (?R) to handle one common case that
2086 Perl 5.6 will be able to do with (?p{...}).
2087
2088 11. Added support for POSIX character classes like [:alpha:], which Perl is
2089 adopting.
2090
2091
2092 Version 2.08 31-Aug-99
2093 ----------------------
2094
2095 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2096 trying to match at the startoffset position, but instead was moving forward to
2097 the next newline as if a previous match had failed.
2098
2099 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2100 and could get into a loop if a null string was matched other than at the start
2101 of the subject.
2102
2103 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2104 be distinguished at compile time, and for completeness also added PCRE_DATE.
2105
2106 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2107 in GnuWin32 environments.
2108
2109
2110 Version 2.07 29-Jul-99
2111 ----------------------
2112
2113 1. The documentation is now supplied in plain text form and HTML as well as in
2114 the form of man page sources.
2115
2116 2. C++ compilers don't like assigning (void *) values to other pointer types.
2117 In particular this affects malloc(). Although there is no problem in Standard
2118 C, I've put in casts to keep C++ compilers happy.
2119
2120 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2121 should be (const char *).
2122
2123 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2124 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2125 However, I haven't made this a standard facility. The documentation doesn't
2126 mention it, and the Makefile doesn't support it.
2127
2128 5. The Makefile now contains an "install" target, with editable destinations at
2129 the top of the file. The pcretest program is not installed.
2130
2131 6. pgrep -V now gives the PCRE version number and date.
2132
2133 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2134 causing the entire string to be ignored, instead of just the last character.
2135
2136 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2137 non-matching string, it can take a very, very long time, even for strings of
2138 quite modest length, because of the nested recursion. PCRE now does better in
2139 some of these cases. It does this by remembering the last required literal
2140 character in the pattern, and pre-searching the subject to ensure it is present
2141 before running the real match. In other words, it applies a heuristic to detect
2142 some types of certain failure quickly, and in the above example, if presented
2143 with a string that has no trailing " it gives "no match" very quickly.
2144
2145 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2146 other alternatives are tried instead.
2147
2148
2149 Version 2.06 09-Jun-99
2150 ----------------------
2151
2152 1. Change pcretest's output for amount of store used to show just the code
2153 space, because the remainder (the data block) varies in size between 32-bit and
2154 64-bit systems.
2155
2156 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2157 start matching at. This allows lookbehinds to work when searching for multiple
2158 occurrences in a string.
2159
2160 3. Added additional options to pcretest for testing multiple occurrences:
2161
2162 /+ outputs the rest of the string that follows a match
2163 /g loops for multiple occurrences, using the new startoffset argument
2164 /G loops for multiple occurrences by passing an incremented pointer
2165
2166 4. PCRE wasn't doing the "first character" optimization for patterns starting
2167 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2168 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2169 the letter 'x'. On long subject strings, this gives a significant speed-up.
2170
2171
2172 Version 2.05 21-Apr-99
2173 ----------------------
2174
2175 1. Changed the type of magic_number from int to long int so that it works
2176 properly on 16-bit systems.
2177
2178 2. Fixed a bug which caused patterns starting with .* not to work correctly
2179 when the subject string contained newline characters. PCRE was assuming
2180 anchoring for such patterns in all cases, which is not correct because .* will
2181 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2182 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2183 must be retried after every newline in the subject.
2184
2185
2186 Version 2.04 18-Feb-99
2187 ----------------------
2188
2189 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2190 computation of the store needed to hold the pattern was incorrect (too large).
2191 If such patterns were nested a few deep, this could multiply and become a real
2192 problem.
2193
2194 2. Added /M option to pcretest to show the memory requirement of a specific
2195 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2196
2197 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2198 compiled in such a way that the backtracking after subsequent failure was
2199 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2200 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2201
2202
2203 Version 2.03 02-Feb-99
2204 ----------------------
2205
2206 1. Fixed typo and small mistake in man page.
2207
2208 2. Added 4th condition (GPL supersedes if conflict) and created separate
2209 LICENCE file containing the conditions.
2210
2211 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2212 Perl, that is the internal \ allows the delimiter to be included in the
2213 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2214 the final delimiter, add \ to the end of the pattern (to test the error).
2215
2216 4. Added the convenience functions for extracting substrings after a successful
2217 match. Updated pcretest to make it able to test these functions.
2218
2219
2220 Version 2.02 14-Jan-99
2221 ----------------------
2222
2223 1. Initialized the working variables associated with each extraction so that
2224 their saving and restoring doesn't refer to uninitialized store.
2225
2226 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2227 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2228 fix the problem.
2229
2230 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2231 calls, and wasn't printing the correct value for compiling calls. Increased the
2232 default value of LOOPREPEAT, and the number of significant figures in the
2233 times.
2234
2235 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2236
2237 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2238 a building problem on Windows NT with a FAT file system.
2239
2240
2241 Version 2.01 21-Oct-98
2242 ----------------------
2243
2244 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2245 to character tables built by pcre_maketables() in the current locale. If NULL
2246 is passed, the default tables are used.
2247
2248
2249 Version 2.00 24-Sep-98
2250 ----------------------
2251
2252 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2253 it any more.
2254
2255 2. Allow quantification of (?>) groups, and make it work correctly.
2256
2257 3. The first character computation wasn't working for (?>) groups.
2258
2259 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2260 end of the subject) and add 5.005's \z, which really does match only at the
2261 very end of the subject.
2262
2263 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2264
2265 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2266 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2267 localized options. All options to pcre_study() were also removed.
2268
2269 7. Add other new features from 5.005:
2270
2271 $(?<= positive lookbehind
2272 $(?<! negative lookbehind
2273 (?imsx-imsx) added the unsetting capability
2274 such a setting is global if at outer level; local otherwise
2275 (?imsx-imsx:) non-capturing groups with option setting
2276 (?(cond)re|re) conditional pattern matching
2277
2278 A backreference to itself in a repeated group matches the previous
2279 captured string.
2280
2281 8. General tidying up of studying (both automatic and via "study")
2282 consequential on the addition of new assertions.
2283
2284 9. As in 5.005, unlimited repeated groups that could match an empty substring
2285 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2286 runtime if any iteration does actually match an empty substring.
2287
2288 10. Include the RunTest script in the distribution.
2289
2290 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2291 discrepancies, some of which were old and were also with respect to 5.004. They
2292 have now been fixed.
2293
2294
2295 Version 1.09 28-Apr-98
2296 ----------------------
2297
2298 1. A negated single character class followed by a quantifier with a minimum
2299 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2300 program crashes, or just wrong answers. This did not apply to negated classes
2301 containing more than one character, or to minima other than one.
2302
2303
2304 Version 1.08 27-Mar-98
2305 ----------------------
2306
2307 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2308
2309 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2310 latter must appear before anything that relies on it in the pattern.
2311
2312
2313 Version 1.07 16-Feb-98
2314 ----------------------
2315
2316 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2317 repeat of a potentially empty string).
2318
2319
2320 Version 1.06 23-Jan-98
2321 ----------------------
2322
2323 1. Added Markus Oberhumer's little patches for C++.
2324
2325 2. Literal strings longer than 255 characters were broken.
2326
2327
2328 Version 1.05 23-Dec-97
2329 ----------------------
2330
2331 1. Negated character classes containing more than one character were failing if
2332 PCRE_CASELESS was set at run time.
2333
2334
2335 Version 1.04 19-Dec-97
2336 ----------------------
2337
2338 1. Corrected the man page, where some "const" qualifiers had been omitted.
2339
2340 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2341 input syntax.
2342
2343 3. Fixed memory leak which occurred when a regex with back references was
2344 matched with an offsets vector that wasn't big enough. The temporary memory
2345 that is used in this case wasn't being freed if the match failed.
2346
2347 4. Tidied pcretest to ensure it frees memory that it gets.
2348
2349 5. Temporary memory was being obtained in the case where the passed offsets
2350 vector was exactly big enough.
2351
2352 6. Corrected definition of offsetof() from change 5 below.
2353
2354 7. I had screwed up change 6 below and broken the rules for the use of
2355 setjmp(). Now fixed.
2356
2357
2358 Version 1.03 18-Dec-97
2359 ----------------------
2360
2361 1. A erroneous regex with a missing opening parenthesis was correctly
2362 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2363 on some systems.
2364
2365 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2366 it was reported that one broken compiler failed on the former because "code" is
2367 also an independent variable.
2368
2369 3. The erroneous regex a[]b caused an array overrun reference.
2370
2371 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2372 fail on data ending with that character. (It was going on too far, and checking
2373 the next character, typically a binary zero.) This was specific to the
2374 optimized code for single-character negative classes.
2375
2376 5. Added a contributed patch from the TIN world which does the following:
2377
2378 + Add an undef for memmove, in case the the system defines a macro for it.
2379
2380 + Add a definition of offsetof(), in case there isn't one. (I don't know
2381 the reason behind this - offsetof() is part of the ANSI standard - but
2382 it does no harm).
2383
2384 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2385 most of the places where whitespace preceded '#'. I have given up and
2386 allowed the remaining 2 cases to be at the margin.
2387
2388 + Rename some variables in pcre to eliminate shadowing. This seems very
2389 pedantic, but does no harm, of course.
2390
2391 6. Moved the call to setjmp() into its own function, to get rid of warnings
2392 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2393
2394 7. Constructs such as \d{8,} were compiling into the equivalent of
2395 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2396 outcome, but in this particular case used more store than had been allocated,
2397 which caused the bug to be discovered because it threw up an internal error.
2398
2399 8. The debugging code in both pcre and pcretest for outputting the compiled
2400 form of a regex was going wrong in the case of back references followed by
2401 curly-bracketed repeats.
2402
2403
2404 Version 1.02 12-Dec-97
2405 ----------------------
2406
2407 1. Typos in pcre.3 and comments in the source fixed.
2408
2409 2. Applied a contributed patch to get rid of places where it used to remove
2410 'const' from variables, and fixed some signed/unsigned and uninitialized
2411 variable warnings.
2412
2413 3. Added the "runtest" target to Makefile.
2414
2415 4. Set default compiler flag to -O2 rather than just -O.
2416
2417
2418 Version 1.01 19-Nov-97
2419 ----------------------
2420
2421 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2422 like /([ab]*)*/, that is, for classes with more than one character in them.
2423
2424 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2425 as /((?>a*))*/ (a PCRE_EXTRA facility).
2426
2427
2428 Version 1.00 18-Nov-97
2429 ----------------------
2430
2431 1. Added compile-time macros to support systems such as SunOS4 which don't have
2432 memmove() or strerror() but have other things that can be used instead.
2433
2434 2. Arranged that "make clean" removes the executables.
2435
2436
2437 Version 0.99 27-Oct-97
2438 ----------------------
2439
2440 1. Fixed bug in code for optimizing classes with only one character. It was
2441 initializing a 32-byte map regardless, which could cause it to run off the end
2442 of the memory it had got.
2443
2444 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
2445
2446
2447 Version 0.98 22-Oct-97
2448 ----------------------
2449
2450 1. Fixed bug in code for handling temporary memory usage when there are more
2451 back references than supplied space in the ovector. This could cause segfaults.
2452
2453
2454 Version 0.97 21-Oct-97
2455 ----------------------
2456
2457 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
2458
2459 2. Optimized negated single characters not to use a bit map.
2460
2461 3. Brought error texts together as macro definitions; clarified some of them;
2462 fixed one that was wrong - it said "range out of order" when it meant "invalid
2463 escape sequence".
2464
2465 4. Changed some char * arguments to const char *.
2466
2467 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
2468
2469 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
2470 pcretest.
2471
2472
2473 Version 0.96 16-Oct-97
2474 ----------------------
2475
2476 1. Added a simple "pgrep" utility to the distribution.
2477
2478 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
2479 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
2480 where "ddd" means "one or more decimal digits".
2481
2482 3. Fixed serious bug. If a pattern had a back reference, but the call to
2483 pcre_exec() didn't supply a large enough ovector to record the related
2484 identifying subpattern, the match always failed. PCRE now remembers the number
2485 of the largest back reference, and gets some temporary memory in which to save
2486 the offsets during matching if necessary, in order to ensure that
2487 backreferences always work.
2488
2489 4. Increased the compatibility with Perl in a number of ways:
2490
2491 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
2492 to request this handling. The option can be set at compile or exec time.
2493
2494 (b) $ matches before a terminating newline by default; an option
2495 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
2496 mode). The option can be set at compile or exec time.
2497
2498 (c) The handling of \ followed by a digit other than 0 is now supposed to be
2499 the same as Perl's. If the decimal number it represents is less than 10
2500 or there aren't that many previous left capturing parentheses, an octal
2501 escape is read. Inside a character class, it's always an octal escape,
2502 even if it is a single digit.
2503
2504 (d) An escaped but undefined alphabetic character is taken as a literal,
2505 unless PCRE_EXTRA is set. Currently this just reserves the remaining
2506 escapes.
2507
2508 (e) {0} is now permitted. (The previous item is removed from the compiled
2509 pattern).
2510
2511 5. Changed all the names of code files so that the basic parts are no longer
2512 than 10 characters, and abolished the teeny "globals.c" file.
2513
2514 6. Changed the handling of character classes; they are now done with a 32-byte
2515 bit map always.
2516
2517 7. Added the -d and /D options to pcretest to make it possible to look at the
2518 internals of compilation without having to recompile pcre.
2519
2520
2521 Version 0.95 23-Sep-97
2522 ----------------------
2523
2524 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
2525 \x20 at the start of a run of normal characters. These were being treated as
2526 real characters, instead of the source characters being re-checked.
2527
2528
2529 Version 0.94 18-Sep-97
2530 ----------------------
2531
2532 1. The functions are now thread-safe, with the caveat that the global variables
2533 containing pointers to malloc() and free() or alternative functions are the
2534 same for all threads.
2535
2536 2. Get pcre_study() to generate a bitmap of initial characters for non-
2537 anchored patterns when this is possible, and use it if passed to pcre_exec().
2538
2539
2540 Version 0.93 15-Sep-97
2541 ----------------------
2542
2543 1. /(b)|(:+)/ was computing an incorrect first character.
2544
2545 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
2546 but not actually doing anything yet.
2547
2548 3. Treat "-" characters in classes that cannot be part of ranges as literals,
2549 as Perl does (e.g. [-az] or [az-]).
2550
2551 4. Set the anchored flag if a branch starts with .* or .*? because that tests
2552 all possible positions.
2553
2554 5. Split up into different modules to avoid including unneeded functions in a
2555 compiled binary. However, compile and exec are still in one module. The "study"
2556 function is split off.
2557
2558 6. The character tables are now in a separate module whose source is generated
2559 by an auxiliary program - but can then be edited by hand if required. There are
2560 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
2561 toupper() in the code.
2562
2563 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
2564 make them global. Abolish the function for setting them, as the caller can now
2565 set them directly.
2566
2567
2568 Version 0.92 11-Sep-97
2569 ----------------------
2570
2571 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
2572 (e.g. /a{1,3}/) was broken (I mis-optimized it).
2573
2574 2. Caseless matching was not working in character classes if the characters in
2575 the pattern were in upper case.
2576
2577 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
2578
2579 4. Make PCRE_ANCHORED public and accept as a compile option.
2580
2581 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
2582 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
2583 pass them.
2584
2585 6. Give an error if bad option bits passed at compile or run time.
2586
2587 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
2588 pcretest to cause it to pass that flag.
2589
2590 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
2591 options, and the first character, if set.
2592
2593 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
2594
2595
2596 Version 0.91 10-Sep-97
2597 ----------------------
2598
2599 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
2600 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
2601
2602 2. PCRE was looping on encountering an indefinitely repeated back reference to
2603 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
2604 Perl does - treats the match as successful.
2605
2606 ****

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12