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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 151 - (show annotations) (download)
Tue Apr 17 15:07:29 2007 UTC (7 years, 4 months ago) by ph10
File size: 126804 byte(s)
Tidies: added some casts and some missing #ifdefs.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.1 12-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 132html script that I use to generate
25 HTML has been updated to handle .nf/.fi and to complain if it encounters
26 .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, with the name
30 config.h.generic, for the benefit of those who have to compile without
31 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
32
33 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
34 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
35 makevp.bat for the current PCRE, using the additional files
36 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
37
38 6. A Windows user reported a minor discrepancy with test 2, which turned out
39 to be caused by a trailing space on an input line that had got lost in his
40 copy. The trailing space was an accident, so I've just removed it.
41
42 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
43 that is needed.
44
45 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
46 as "const" (a) because they are and (b) because it helps the PHP
47 maintainers who have recently made a script to detect big data structures
48 in the php code that should be moved to the .rodata section. I remembered
49 to update Builducptable as well, so it won't revert if ucptable.h is ever
50 re-created.
51
52 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
53 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
54 order to be able to cut out the UTF-8 tables in the latter when UTF-8
55 support is not required. This saves 1.5-2K of code, which is important in
56 some applications.
57
58 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
59 so as not to refer to the tables, even though these functions will never be
60 called when UTF-8 support is disabled. Otherwise there are problems with a
61 shared library.
62
63 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
64
65 (a) It was defining its arguments as char * instead of void *.
66
67 (b) It was assuming that all moves were upwards in memory; this was true
68 a long time ago when I wrote it, but is no longer the case.
69
70 The emulated memove() is provided for those environments that have neither
71 memmove() nor bcopy(). I didn't think anyone used it these days, but that
72 is clearly not the case, as these two bugs were recently reported.
73
74 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
75 and Detrail to create the HTML documentation, the .txt form of the man
76 pages, and it removes trailing spaces from listed files. It also creates
77 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
78 case, it wraps all the #defines with #ifndefs. This script should be run
79 before "make dist".
80
81 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
82 with Unicode property support.
83
84 (a) For a maximizing quantifier, if the two different cases of the
85 character were of different lengths in their UTF-8 codings (there are
86 some cases like this - I found 11), and the matching function had to
87 back up over a mixture of the two cases, it incorrectly assumed they
88 were both the same length.
89
90 (b) When PCRE was configured to use the heap rather than the stack for
91 recursion during matching, it was not correctly preserving the data for
92 the other case of a UTF-8 character when checking ahead for a match
93 while processing a minimizing repeat. If the check also involved
94 matching a wide character, but failed, corruption could cause an
95 erroneous result when trying to check for a repeat of the original
96 character.
97
98 13. Some tidying changes to the testing mechanism:
99
100 (a) The RunTest script now detects the internal link size and whether there
101 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
102 values substituted by "configure". (The RunGrepTest script already did
103 this for UTF-8.) The configure.ac script no longer substitutes the
104 relevant variables.
105
106 (b) The debugging options /B and /D in pcretest show the compiled bytecode
107 with length and offset values. This means that the output is different
108 for different internal link sizes. Test 2 is skipped for link sizes
109 other than 2 because of this, bypassing the problem. Unfortunately,
110 there was also a test in test 3 (the locale tests) that used /B and
111 failed for link sizes other than 2. Rather than cut the whole test out,
112 I have added a new /Z option to pcretest that replaces the length and
113 offset values with spaces. This is now used to make test 3 independent
114 of link size. (Test 2 will be tidied up later.)
115
116 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
117 segmentation fault instead of returning the appropriate error message.
118
119 15. In multiline mode when the newline sequence was set to "any", the pattern
120 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
121 This doesn't seem right; it now treats the CRLF combination as the line
122 ending, and so does not match in that case. It's only a pattern such as ^$
123 that would hit this one: something like ^ABC$ would have failed after \r
124 and then tried again after \r\n.
125
126 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
127 in an attempt to make files that differ only in their line terminators
128 compare equal. This works on Linux.
129
130 17. Under certain error circumstances pcregrep might try to free random memory
131 as it exited. This is now fixed, thanks to valgrind.
132
133 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
134 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
135 was because its rules for how to advance for /g after matching an empty
136 string at the end of a line did not allow for this case. They now check for
137 it specially.
138
139 20. pcretest is supposed to handle patterns and data of any length, by
140 extending its buffers when necessary. It was getting this wrong when the
141 buffer for a data line had to be extended.
142
143 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
144 CRLF as a newline sequence.
145
146 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
147 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
148 I have nevertheless tidied it up.
149
150 23. Added some casts to kill warnings from HP-UX ia64 compiler.
151
152
153 Version 7.0 19-Dec-06
154 ---------------------
155
156 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
157 moving to gcc 4.1.1.
158
159 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
160 sys/time.h, which is documented as needed for this function. It doesn't
161 seem to matter on Linux, but it showed up on some releases of OS X.
162
163 3. It seems that there are systems where bytes whose values are greater than
164 127 match isprint() in the "C" locale. The "C" locale should be the
165 default when a C program starts up. In most systems, only ASCII printing
166 characters match isprint(). This difference caused the output from pcretest
167 to vary, making some of the tests fail. I have changed pcretest so that:
168
169 (a) When it is outputting text in the compiled version of a pattern, bytes
170 other than 32-126 are always shown as hex escapes.
171
172 (b) When it is outputting text that is a matched part of a subject string,
173 it does the same, unless a different locale has been set for the match
174 (using the /L modifier). In this case, it uses isprint() to decide.
175
176 4. Fixed a major bug that caused incorrect computation of the amount of memory
177 required for a compiled pattern when options that changed within the
178 pattern affected the logic of the preliminary scan that determines the
179 length. The relevant options are -x, and -i in UTF-8 mode. The result was
180 that the computed length was too small. The symptoms of this bug were
181 either the PCRE error "internal error: code overflow" from pcre_compile(),
182 or a glibc crash with a message such as "pcretest: free(): invalid next
183 size (fast)". Examples of patterns that provoked this bug (shown in
184 pcretest format) are:
185
186 /(?-x: )/x
187 /(?x)(?-x: \s*#\s*)/
188 /((?i)[\x{c0}])/8
189 /(?i:[\x{c0}])/8
190
191 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
192 is now done differently.
193
194 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
195 wrapper classes; (b) implement a new function in the C++ scanner that is
196 more efficient than the old way of doing things because it avoids levels of
197 recursion in the regex matching; (c) add a paragraph to the documentation
198 for the FullMatch() function.
199
200 6. The escape sequence \n was being treated as whatever was defined as
201 "newline". Not only was this contrary to the documentation, which states
202 that \n is character 10 (hex 0A), but it also went horribly wrong when
203 "newline" was defined as CRLF. This has been fixed.
204
205 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
206 was being set to -1 for the "end of line" case (supposedly a value that no
207 character can have). Though this value is never used (the check for end of
208 line is "zero bytes in current character"), it caused compiler complaints.
209 I've changed it to 0xffffffff.
210
211 8. In pcre_version.c, the version string was being built by a sequence of
212 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
213 string (as it is for production releases) called a macro with an empty
214 argument. The C standard says the result of this is undefined. The gcc
215 compiler treats it as an empty string (which was what was wanted) but it is
216 reported that Visual C gives an error. The source has been hacked around to
217 avoid this problem.
218
219 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
220 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
221 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
222 of them did).
223
224 10. Originally, pcretest opened its input and output without "b"; then I was
225 told that "b" was needed in some environments, so it was added for release
226 5.0 to both the input and output. (It makes no difference on Unix-like
227 systems.) Later I was told that it is wrong for the input on Windows. I've
228 now abstracted the modes into two macros, to make it easier to fiddle with
229 them, and removed "b" from the input mode under Windows.
230
231 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
232
233 12. Added -help and --help to pcretest as an official way of being reminded
234 of the options.
235
236 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
237 and pcrecpp.cc because they annoy compilers at high warning levels.
238
239 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
240
241 15. Fixed an occurrence of == in configure.ac that should have been = (shell
242 scripts are not C programs :-) and which was not noticed because it works
243 on Linux.
244
245 16. pcretest is supposed to handle any length of pattern and data line (as one
246 line or as a continued sequence of lines) by extending its input buffer if
247 necessary. This feature was broken for very long pattern lines, leading to
248 a string of junk being passed to pcre_compile() if the pattern was longer
249 than about 50K.
250
251 17. I have done a major re-factoring of the way pcre_compile() computes the
252 amount of memory needed for a compiled pattern. Previously, there was code
253 that made a preliminary scan of the pattern in order to do this. That was
254 OK when PCRE was new, but as the facilities have expanded, it has become
255 harder and harder to keep it in step with the real compile phase, and there
256 have been a number of bugs (see for example, 4 above). I have now found a
257 cunning way of running the real compile function in a "fake" mode that
258 enables it to compute how much memory it would need, while actually only
259 ever using a few hundred bytes of working memory and without too many
260 tests of the mode. This should make future maintenance and development
261 easier. A side effect of this work is that the limit of 200 on the nesting
262 depth of parentheses has been removed (though this was never a serious
263 limitation, I suspect). However, there is a downside: pcre_compile() now
264 runs more slowly than before (30% or more, depending on the pattern). I
265 hope this isn't a big issue. There is no effect on runtime performance.
266
267 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
268 newline (only possible for the last line of a file) and it was a
269 pattern that set a locale (followed by /Lsomething), pcretest crashed.
270
271 19. Added additional timing features to pcretest. (1) The -tm option now times
272 matching only, not compiling. (2) Both -t and -tm can be followed, as a
273 separate command line item, by a number that specifies the number of
274 repeats to use when timing. The default is 50000; this gives better
275 precision, but takes uncomfortably long for very large patterns.
276
277 20. Extended pcre_study() to be more clever in cases where a branch of a
278 subpattern has no definite first character. For example, (a*|b*)[cd] would
279 previously give no result from pcre_study(). Now it recognizes that the
280 first character must be a, b, c, or d.
281
282 21. There was an incorrect error "recursive call could loop indefinitely" if
283 a subpattern (or the entire pattern) that was being tested for matching an
284 empty string contained only one non-empty item after a nested subpattern.
285 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
286 incorrectly, because the \d was being skipped in the check.
287
288 22. The pcretest program now has a new pattern option /B and a command line
289 option -b, which is equivalent to adding /B to every pattern. This causes
290 it to show the compiled bytecode, without the additional information that
291 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
292 is the same as /B/I).
293
294 23. A new optimization is now able automatically to treat some sequences such
295 as a*b as a*+b. More specifically, if something simple (such as a character
296 or a simple class like \d) has an unlimited quantifier, and is followed by
297 something that cannot possibly match the quantified thing, the quantifier
298 is automatically "possessified".
299
300 24. A recursive reference to a subpattern whose number was greater than 39
301 went wrong under certain circumstances in UTF-8 mode. This bug could also
302 have affected the operation of pcre_study().
303
304 25. Realized that a little bit of performance could be had by replacing
305 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
306
307 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
308
309 27. Possessive quantifiers such as a++ were previously implemented by turning
310 them into atomic groups such as ($>a+). Now they have their own opcodes,
311 which improves performance. This includes the automatically created ones
312 from 23 above.
313
314 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
315 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
316 the first matched character to be a colon. This applied both to named and
317 numbered groups.
318
319 29. The ucpinternal.h header file was missing its idempotency #ifdef.
320
321 30. I was sent a "project" file called libpcre.a.dev which I understand makes
322 building PCRE on Windows easier, so I have included it in the distribution.
323
324 31. There is now a check in pcretest against a ridiculously large number being
325 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
326 loop, the loop is abandoned.
327
328 32. Forward references to subpatterns in conditions such as (?(2)...) where
329 subpattern 2 is defined later cause pcre_compile() to search forwards in
330 the pattern for the relevant set of parentheses. This search went wrong
331 when there were unescaped parentheses in a character class, parentheses
332 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
333
334 33. "Subroutine" calls and backreferences were previously restricted to
335 referencing subpatterns earlier in the regex. This restriction has now
336 been removed.
337
338 34. Added a number of extra features that are going to be in Perl 5.10. On the
339 whole, these are just syntactic alternatives for features that PCRE had
340 previously implemented using the Python syntax or my own invention. The
341 other formats are all retained for compatibility.
342
343 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
344 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
345 also .NET compatible.
346
347 (b) A recursion or subroutine call to a named group can now be defined as
348 (?&name) as well as (?P>name).
349
350 (c) A backreference to a named group can now be defined as \k<name> or
351 \k'name' as well as (?P=name). The new forms, as well as being in Perl
352 5.10, are also .NET compatible.
353
354 (d) A conditional reference to a named group can now use the syntax
355 (?(<name>) or (?('name') as well as (?(name).
356
357 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
358 groups (named and numbered) that are never evaluated inline, but can be
359 called as "subroutines" from elsewhere. In effect, the DEFINE condition
360 is always false. There may be only one alternative in such a group.
361
362 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
363 as the simple (?(R). The condition is true only if the most recent
364 recursion is that of the given number or name. It does not search out
365 through the entire recursion stack.
366
367 (g) The escape \gN or \g{N} has been added, where N is a positive or
368 negative number, specifying an absolute or relative reference.
369
370 35. Tidied to get rid of some further signed/unsigned compiler warnings and
371 some "unreachable code" warnings.
372
373 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
374 things, this adds five new scripts.
375
376 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
377 There were also incompatibilities regarding the handling of \Q..\E inside
378 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
379 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
380
381 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
382 matches an empty string, and forcibly breaks the loop. There were bugs in
383 this code in non-simple cases. For a pattern such as ^(a()*)* matched
384 against aaaa the result was just "a" rather than "aaaa", for example. Two
385 separate and independent bugs (that affected different cases) have been
386 fixed.
387
388 39. Refactored the code to abolish the use of different opcodes for small
389 capturing bracket numbers. This is a tidy that I avoided doing when I
390 removed the limit on the number of capturing brackets for 3.5 back in 2001.
391 The new approach is not only tidier, it makes it possible to reduce the
392 memory needed to fix the previous bug (38).
393
394 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
395 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
396 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
397 mode.
398
399 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
400 report.
401
402 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
403 copy construction and assignment in the C++ wrapper.
404
405 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
406 couple of bugs that could have given wrong results in the "--newline=crlf"
407 case.
408
409 44. Added a number of casts and did some reorganization of signed/unsigned int
410 variables following suggestions from Dair Grant. Also renamed the variable
411 "this" as "item" because it is a C++ keyword.
412
413 45. Arranged for dftables to add
414
415 #include "pcre_internal.h"
416
417 to pcre_chartables.c because without it, gcc 4.x may remove the array
418 definition from the final binary if PCRE is built into a static library and
419 dead code stripping is activated.
420
421 46. For an unanchored pattern, if a match attempt fails at the start of a
422 newline sequence, and the newline setting is CRLF or ANY, and the next two
423 characters are CRLF, advance by two characters instead of one.
424
425
426 Version 6.7 04-Jul-06
427 ---------------------
428
429 1. In order to handle tests when input lines are enormously long, pcretest has
430 been re-factored so that it automatically extends its buffers when
431 necessary. The code is crude, but this _is_ just a test program. The
432 default size has been increased from 32K to 50K.
433
434 2. The code in pcre_study() was using the value of the re argument before
435 testing it for NULL. (Of course, in any sensible call of the function, it
436 won't be NULL.)
437
438 3. The memmove() emulation function in pcre_internal.h, which is used on
439 systems that lack both memmove() and bcopy() - that is, hardly ever -
440 was missing a "static" storage class specifier.
441
442 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
443 containing an extended class (one that cannot be represented by a bitmap
444 because it contains high-valued characters or Unicode property items, e.g.
445 [\pZ]). Almost always one would set UTF-8 mode when processing such a
446 pattern, but PCRE should not loop if you do not (it no longer does).
447 [Detail: two cases were found: (a) a repeated subpattern containing an
448 extended class; (b) a recursive reference to a subpattern that followed a
449 previous extended class. It wasn't skipping over the extended class
450 correctly when UTF-8 mode was not set.]
451
452 5. A negated single-character class was not being recognized as fixed-length
453 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
454 compile error "lookbehind assertion is not fixed length".
455
456 6. The RunPerlTest auxiliary script was showing an unexpected difference
457 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
458 write a Perl script that can interpret lines of an input file either as
459 byte characters or as UTF-8, which is what "perltest" was being required to
460 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
461 can't do is switch easily at run time between having the "use utf8;" pragma
462 or not. In the end, I fudged it by using the RunPerlTest script to insert
463 "use utf8;" explicitly for the UTF-8 tests.
464
465 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
466 the end of the subject string, contrary to the documentation and to what
467 Perl does. This was true of both matching functions. Now it matches only at
468 the start of the subject and immediately after *internal* newlines.
469
470 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
471 a pointer to an int instead of a pointer to an unsigned long int. This
472 caused problems on 64-bit systems.
473
474 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
475 instance of the 'standard' template library not being so standard".
476
477 10. There was no check on the number of named subpatterns nor the maximum
478 length of a subpattern name. The product of these values is used to compute
479 the size of the memory block for a compiled pattern. By supplying a very
480 long subpattern name and a large number of named subpatterns, the size
481 computation could be caused to overflow. This is now prevented by limiting
482 the length of names to 32 characters, and the number of named subpatterns
483 to 10,000.
484
485 11. Subpatterns that are repeated with specific counts have to be replicated in
486 the compiled pattern. The size of memory for this was computed from the
487 length of the subpattern and the repeat count. The latter is limited to
488 65535, but there was no limit on the former, meaning that integer overflow
489 could in principle occur. The compiled length of a repeated subpattern is
490 now limited to 30,000 bytes in order to prevent this.
491
492 12. Added the optional facility to have named substrings with the same name.
493
494 13. Added the ability to use a named substring as a condition, using the
495 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
496 are numbers (not recommended). Forward references are permitted.
497
498 14. Added forward references in named backreferences (if you see what I mean).
499
500 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
501 pattern could run off the end of the subject. For example, the pattern
502 "(?s)(.{1,5})"8 did this with the subject "ab".
503
504 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
505 PCRE_CASELESS was set when matching characters that were quantified with ?
506 or *.
507
508 17. A character class other than a single negated character that had a minimum
509 but no maximum quantifier - for example [ab]{6,} - was not handled
510 correctly by pce_dfa_exec(). It would match only one character.
511
512 18. A valid (though odd) pattern that looked like a POSIX character
513 class but used an invalid character after [ (for example [[,abc,]]) caused
514 pcre_compile() to give the error "Failed: internal error: code overflow" or
515 in some cases to crash with a glibc free() error. This could even happen if
516 the pattern terminated after [[ but there just happened to be a sequence of
517 letters, a binary zero, and a closing ] in the memory that followed.
518
519 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
520 over the years. Originally (before any Unicode support), just the bottom 8
521 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
522 output from "man perlunicode" includes this:
523
524 The regular expression compiler produces polymorphic opcodes. That
525 is, the pattern adapts to the data and automatically switches to
526 the Unicode character scheme when presented with Unicode data--or
527 instead uses a traditional byte scheme when presented with byte
528 data.
529
530 Sadly, a wide octal escape does not cause a switch, and in a string with
531 no other multibyte characters, these octal escapes are treated as before.
532 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
533 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
534 Unicode string.
535
536 I have not perpetrated such confusion in PCRE. Up till now, it took just
537 the bottom 8 bits, as in old Perl. I have now made octal escapes with
538 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
539 translate to the appropriate multibyte character.
540
541 29. Applied some refactoring to reduce the number of warnings from Microsoft
542 and Borland compilers. This has included removing the fudge introduced
543 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
544 a warning about an unused variable.
545
546 21. PCRE has not included VT (character 0x0b) in the set of whitespace
547 characters since release 4.0, because Perl (from release 5.004) does not.
548 [Or at least, is documented not to: some releases seem to be in conflict
549 with the documentation.] However, when a pattern was studied with
550 pcre_study() and all its branches started with \s, PCRE still included VT
551 as a possible starting character. Of course, this did no harm; it just
552 caused an unnecessary match attempt.
553
554 22. Removed a now-redundant internal flag bit that recorded the fact that case
555 dependency changed within the pattern. This was once needed for "required
556 byte" processing, but is no longer used. This recovers a now-scarce options
557 bit. Also moved the least significant internal flag bit to the most-
558 significant bit of the word, which was not previously used (hangover from
559 the days when it was an int rather than a uint) to free up another bit for
560 the future.
561
562 23. Added support for CRLF line endings as well as CR and LF. As well as the
563 default being selectable at build time, it can now be changed at runtime
564 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
565 specify that it is scanning data with non-default line endings.
566
567 24. Changed the definition of CXXLINK to make it agree with the definition of
568 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
569
570 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
571 recursions. This makes a big different to stack usage for some patterns.
572
573 26. If a subpattern containing a named recursion or subroutine reference such
574 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
575 the space required for the compiled pattern went wrong and gave too small a
576 value. Depending on the environment, this could lead to "Failed: internal
577 error: code overflow at offset 49" or "glibc detected double free or
578 corruption" errors.
579
580 27. Applied patches from Google (a) to support the new newline modes and (b) to
581 advance over multibyte UTF-8 characters in GlobalReplace.
582
583 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
584 difference for some implementation of PCRE in some Windows version.
585
586 29. Added some extra testing facilities to pcretest:
587
588 \q<number> in a data line sets the "match limit" value
589 \Q<number> in a data line sets the "match recursion limt" value
590 -S <number> sets the stack size, where <number> is in megabytes
591
592 The -S option isn't available for Windows.
593
594
595 Version 6.6 06-Feb-06
596 ---------------------
597
598 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
599 in pcreposix.h. I have copied the definition from pcre.h.
600
601 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
602 because pcre.h is no longer a built file.
603
604 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
605 not normally included in the compiled code.
606
607
608 Version 6.5 01-Feb-06
609 ---------------------
610
611 1. When using the partial match feature with pcre_dfa_exec(), it was not
612 anchoring the second and subsequent partial matches at the new starting
613 point. This could lead to incorrect results. For example, with the pattern
614 /1234/, partially matching against "123" and then "a4" gave a match.
615
616 2. Changes to pcregrep:
617
618 (a) All non-match returns from pcre_exec() were being treated as failures
619 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
620 error message is output. Some extra information is given for the
621 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
622 probably the only errors that are likely to be caused by users (by
623 specifying a regex that has nested indefinite repeats, for instance).
624 If there are more than 20 of these errors, pcregrep is abandoned.
625
626 (b) A binary zero was treated as data while matching, but terminated the
627 output line if it was written out. This has been fixed: binary zeroes
628 are now no different to any other data bytes.
629
630 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
631 used to set a locale for matching. The --locale=xxxx long option has
632 been added (no short equivalent) to specify a locale explicitly on the
633 pcregrep command, overriding the environment variables.
634
635 (d) When -B was used with -n, some line numbers in the output were one less
636 than they should have been.
637
638 (e) Added the -o (--only-matching) option.
639
640 (f) If -A or -C was used with -c (count only), some lines of context were
641 accidentally printed for the final match.
642
643 (g) Added the -H (--with-filename) option.
644
645 (h) The combination of options -rh failed to suppress file names for files
646 that were found from directory arguments.
647
648 (i) Added the -D (--devices) and -d (--directories) options.
649
650 (j) Added the -F (--fixed-strings) option.
651
652 (k) Allow "-" to be used as a file name for -f as well as for a data file.
653
654 (l) Added the --colo(u)r option.
655
656 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
657 is not present by default.
658
659 3. A nasty bug was discovered in the handling of recursive patterns, that is,
660 items such as (?R) or (?1), when the recursion could match a number of
661 alternatives. If it matched one of the alternatives, but subsequently,
662 outside the recursion, there was a failure, the code tried to back up into
663 the recursion. However, because of the way PCRE is implemented, this is not
664 possible, and the result was an incorrect result from the match.
665
666 In order to prevent this happening, the specification of recursion has
667 been changed so that all such subpatterns are automatically treated as
668 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
669
670 4. I had overlooked the fact that, in some locales, there are characters for
671 which isalpha() is true but neither isupper() nor islower() are true. In
672 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
673 and ordfeminine) are like this. This affected the treatment of \w and \W
674 when they appeared in character classes, but not when they appeared outside
675 a character class. The bit map for "word" characters is now created
676 separately from the results of isalnum() instead of just taking it from the
677 upper, lower, and digit maps. (Plus the underscore character, of course.)
678
679 5. The above bug also affected the handling of POSIX character classes such as
680 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
681 permanent tables. Instead, the bit maps for such a class were previously
682 created as the appropriate unions of the upper, lower, and digit bitmaps.
683 Now they are created by subtraction from the [[:word:]] class, which has
684 its own bitmap.
685
686 6. The [[:blank:]] character class matches horizontal, but not vertical space.
687 It is created by subtracting the vertical space characters (\x09, \x0a,
688 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
689 subtraction was done in the overall bitmap for a character class, meaning
690 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
691 be recognized. This bug has been fixed.
692
693 7. Patches from the folks at Google:
694
695 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
696 real life, but is still worth protecting against".
697
698 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
699 regular expressions".
700
701 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
702 have it.
703
704 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
705 "configure" and the latter not, in order to fix a problem somebody had
706 with compiling the Arg class on HP-UX.
707
708 (e) Improve the error-handling of the C++ wrapper a little bit.
709
710 (f) New tests for checking recursion limiting.
711
712 8. The pcre_memmove() function, which is used only if the environment does not
713 have a standard memmove() function (and is therefore rarely compiled),
714 contained two bugs: (a) use of int instead of size_t, and (b) it was not
715 returning a result (though PCRE never actually uses the result).
716
717 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
718 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
719 returned instead of calling malloc() with an overflowing number that would
720 most likely cause subsequent chaos.
721
722 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
723
724 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
725 with this option is matched, the nmatch and pmatch options of regexec() are
726 ignored.
727
728 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
729 provided in case anyone wants to the the POSIX interface with UTF-8
730 strings.
731
732 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
733 C++ linking (needed for some HP-UX environments).
734
735 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
736 (unused parameter) and in the pcre_printint() function (omitted "default"
737 switch label when the default is to do nothing).
738
739 15. Added some code to make it possible, when PCRE is compiled as a C++
740 library, to replace subject pointers for pcre_exec() with a smart pointer
741 class, thus making it possible to process discontinuous strings.
742
743 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
744 much the same function. They were added by different people who were trying
745 to make PCRE easy to compile on non-Unix systems. It has been suggested
746 that PCRE_EXPORT be abolished now that there is more automatic apparatus
747 for compiling on Windows systems. I have therefore replaced it with
748 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
749 defaults to "extern" for C or "extern C" for C++, which works fine on
750 Unix-like systems. It is now possible to override the value of PCRE_DATA_
751 SCOPE with something explicit in config.h. In addition:
752
753 (a) pcreposix.h still had just "extern" instead of either of these macros;
754 I have replaced it with PCRE_DATA_SCOPE.
755
756 (b) Functions such as _pcre_xclass(), which are internal to the library,
757 but external in the C sense, all had PCRE_EXPORT in their definitions.
758 This is apparently wrong for the Windows case, so I have removed it.
759 (It makes no difference on Unix-like systems.)
760
761 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
762 of recursive calls to match(). This is different to MATCH_LIMIT because
763 that limits the total number of calls to match(), not all of which increase
764 the depth of recursion. Limiting the recursion depth limits the amount of
765 stack (or heap if NO_RECURSE is set) that is used. The default can be set
766 when PCRE is compiled, and changed at run time. A patch from Google adds
767 this functionality to the C++ interface.
768
769 18. Changes to the handling of Unicode character properties:
770
771 (a) Updated the table to Unicode 4.1.0.
772
773 (b) Recognize characters that are not in the table as "Cn" (undefined).
774
775 (c) I revised the way the table is implemented to a much improved format
776 which includes recognition of ranges. It now supports the ranges that
777 are defined in UnicodeData.txt, and it also amalgamates other
778 characters into ranges. This has reduced the number of entries in the
779 table from around 16,000 to around 3,000, thus reducing its size
780 considerably. I realized I did not need to use a tree structure after
781 all - a binary chop search is just as efficient. Having reduced the
782 number of entries, I extended their size from 6 bytes to 8 bytes to
783 allow for more data.
784
785 (d) Added support for Unicode script names via properties such as \p{Han}.
786
787 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
788 matching that character.
789
790 20. When matching a repeated Unicode property with a minimum greater than zero,
791 (for example \pL{2,}), PCRE could look past the end of the subject if it
792 reached it while seeking the minimum number of characters. This could
793 happen only if some of the characters were more than one byte long, because
794 there is a check for at least the minimum number of bytes.
795
796 21. Refactored the implementation of \p and \P so as to be more general, to
797 allow for more different types of property in future. This has changed the
798 compiled form incompatibly. Anybody with saved compiled patterns that use
799 \p or \P will have to recompile them.
800
801 22. Added "Any" and "L&" to the supported property types.
802
803 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
804 but give a compile time error if the value is greater than 0xff.
805
806 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
807 accidentally not being installed or uninstalled.
808
809 25. The pcre.h file was built from pcre.h.in, but the only changes that were
810 made were to insert the current release number. This seemed silly, because
811 it made things harder for people building PCRE on systems that don't run
812 "configure". I have turned pcre.h into a distributed file, no longer built
813 by "configure", with the version identification directly included. There is
814 no longer a pcre.h.in file.
815
816 However, this change necessitated a change to the pcre-config script as
817 well. It is built from pcre-config.in, and one of the substitutions was the
818 release number. I have updated configure.ac so that ./configure now finds
819 the release number by grepping pcre.h.
820
821 26. Added the ability to run the tests under valgrind.
822
823
824 Version 6.4 05-Sep-05
825 ---------------------
826
827 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
828 "--" to be printed when multiple files were scanned, even when none of the
829 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
830 consider it to be a bug, and have restored the previous behaviour.
831
832 2. A couple of code tidies to get rid of compiler warnings.
833
834 3. The pcretest program used to cheat by referring to symbols in the library
835 whose names begin with _pcre_. These are internal symbols that are not
836 really supposed to be visible externally, and in some environments it is
837 possible to suppress them. The cheating is now confined to including
838 certain files from the library's source, which is a bit cleaner.
839
840 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
841 file's purpose clearer.
842
843 5. Reorganized pcre_ucp_findchar().
844
845
846 Version 6.3 15-Aug-05
847 ---------------------
848
849 1. The file libpcre.pc.in did not have general read permission in the tarball.
850
851 2. There were some problems when building without C++ support:
852
853 (a) If C++ support was not built, "make install" and "make test" still
854 tried to test it.
855
856 (b) There were problems when the value of CXX was explicitly set. Some
857 changes have been made to try to fix these, and ...
858
859 (c) --disable-cpp can now be used to explicitly disable C++ support.
860
861 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
862 backslash in a target when C++ was disabled. This confuses some
863 versions of "make", apparently. Using an intermediate variable solves
864 this. (Same for CPP_LOBJ.)
865
866 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
867 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
868 necessary on certain architectures.
869
870 4. Added a setting of -export-symbols-regex to the link command to remove
871 those symbols that are exported in the C sense, but actually are local
872 within the library, and not documented. Their names all begin with
873 "_pcre_". This is not a perfect job, because (a) we have to except some
874 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
875 available (and never for static libraries). I have made a note to try to
876 find a way round (a) in the future.
877
878
879 Version 6.2 01-Aug-05
880 ---------------------
881
882 1. There was no test for integer overflow of quantifier values. A construction
883 such as {1111111111111111} would give undefined results. What is worse, if
884 a minimum quantifier for a parenthesized subpattern overflowed and became
885 negative, the calculation of the memory size went wrong. This could have
886 led to memory overwriting.
887
888 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
889
890 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
891 operating environments where this matters.
892
893 4. Applied Giuseppe Maxia's patch to add additional features for controlling
894 PCRE options from within the C++ wrapper.
895
896 5. Named capturing subpatterns were not being correctly counted when a pattern
897 was compiled. This caused two problems: (a) If there were more than 100
898 such subpatterns, the calculation of the memory needed for the whole
899 compiled pattern went wrong, leading to an overflow error. (b) Numerical
900 back references of the form \12, where the number was greater than 9, were
901 not recognized as back references, even though there were sufficient
902 previous subpatterns.
903
904 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
905 versions of gcc, e.g. 2.95.4.
906
907
908 Version 6.1 21-Jun-05
909 ---------------------
910
911 1. There was one reference to the variable "posix" in pcretest.c that was not
912 surrounded by "#if !defined NOPOSIX".
913
914 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
915 the cross-check on the old pcre_info() function, for the benefit of the
916 cut-down version of PCRE that is currently imported into Exim.
917
918 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
919 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
920 patterns that start (?i) but I don't think that matters. The use of (?i) is
921 just an example; this all applies to the other options as well.
922
923 4. Since libtool seems to echo the compile commands it is issuing, the output
924 from "make" can be reduced a bit by putting "@" in front of each libtool
925 compile command.
926
927 5. Patch from the folks at Google for configure.in to be a bit more thorough
928 in checking for a suitable C++ installation before trying to compile the
929 C++ stuff. This should fix a reported problem when a compiler was present,
930 but no suitable headers.
931
932 6. The man pages all had just "PCRE" as their title. I have changed them to
933 be the relevant file name. I have also arranged that these names are
934 retained in the file doc/pcre.txt, which is a concatenation in text format
935 of all the man pages except the little individual ones for each function.
936
937 7. The NON-UNIX-USE file had not been updated for the different set of source
938 files that come with release 6. I also added a few comments about the C++
939 wrapper.
940
941
942 Version 6.0 07-Jun-05
943 ---------------------
944
945 1. Some minor internal re-organization to help with my DFA experiments.
946
947 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
948 didn't matter for the library itself when fully configured, but did matter
949 when compiling without UCP support, or within Exim, where the ucp files are
950 not imported.
951
952 3. Refactoring of the library code to split up the various functions into
953 different source modules. The addition of the new DFA matching code (see
954 below) to a single monolithic source would have made it really too
955 unwieldy, quite apart from causing all the code to be include in a
956 statically linked application, when only some functions are used. This is
957 relevant even without the DFA addition now that patterns can be compiled in
958 one application and matched in another.
959
960 The downside of splitting up is that there have to be some external
961 functions and data tables that are used internally in different modules of
962 the library but which are not part of the API. These have all had their
963 names changed to start with "_pcre_" so that they are unlikely to clash
964 with other external names.
965
966 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
967 a different (DFA) algorithm. Although it is slower than the original
968 function, it does have some advantages for certain types of matching
969 problem.
970
971 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
972 including restarting after a partial match.
973
974 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
975 defined when compiling for Windows was sent to me. I have put it into the
976 code, though I have no means of testing or verifying it.
977
978 7. Added the pcre_refcount() auxiliary function.
979
980 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
981 match before or at the first newline in the subject string. In pcretest,
982 the /f option on a pattern can be used to set this.
983
984 9. A repeated \w when used in UTF-8 mode with characters greater than 256
985 would behave wrongly. This has been present in PCRE since release 4.0.
986
987 10. A number of changes to the pcregrep command:
988
989 (a) Refactored how -x works; insert ^(...)$ instead of setting
990 PCRE_ANCHORED and checking the length, in preparation for adding
991 something similar for -w.
992
993 (b) Added the -w (match as a word) option.
994
995 (c) Refactored the way lines are read and buffered so as to have more
996 than one at a time available.
997
998 (d) Implemented a pcregrep test script.
999
1000 (e) Added the -M (multiline match) option. This allows patterns to match
1001 over several lines of the subject. The buffering ensures that at least
1002 8K, or the rest of the document (whichever is the shorter) is available
1003 for matching (and similarly the previous 8K for lookbehind assertions).
1004
1005 (f) Changed the --help output so that it now says
1006
1007 -w, --word-regex(p)
1008
1009 instead of two lines, one with "regex" and the other with "regexp"
1010 because that confused at least one person since the short forms are the
1011 same. (This required a bit of code, as the output is generated
1012 automatically from a table. It wasn't just a text change.)
1013
1014 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1015 option but starts with a hyphen. Could be a pattern or a path name
1016 starting with a hyphen, for instance.
1017
1018 (h) "-" can be given as a file name to represent stdin.
1019
1020 (i) When file names are being printed, "(standard input)" is used for
1021 the standard input, for compatibility with GNU grep. Previously
1022 "<stdin>" was used.
1023
1024 (j) The option --label=xxx can be used to supply a name to be used for
1025 stdin when file names are being printed. There is no short form.
1026
1027 (k) Re-factored the options decoding logic because we are going to add
1028 two more options that take data. Such options can now be given in four
1029 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1030
1031 (l) Added the -A, -B, and -C options for requesting that lines of context
1032 around matches be printed.
1033
1034 (m) Added the -L option to print the names of files that do not contain
1035 any matching lines, that is, the complement of -l.
1036
1037 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1038 continue to scan other files.
1039
1040 (o) The -s option was incorrectly implemented. For compatibility with other
1041 greps, it now suppresses the error message for a non-existent or non-
1042 accessible file (but not the return code). There is a new option called
1043 -q that suppresses the output of matching lines, which was what -s was
1044 previously doing.
1045
1046 (p) Added --include and --exclude options to specify files for inclusion
1047 and exclusion when recursing.
1048
1049 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1050 Hopefully, it now does.
1051
1052 12. Missing cast in pcre_study().
1053
1054 13. Added an "uninstall" target to the makefile.
1055
1056 14. Replaced "extern" in the function prototypes in Makefile.in with
1057 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1058 world, but is set differently for Windows.
1059
1060 15. Added a second compiling function called pcre_compile2(). The only
1061 difference is that it has an extra argument, which is a pointer to an
1062 integer error code. When there is a compile-time failure, this is set
1063 non-zero, in addition to the error test pointer being set to point to an
1064 error message. The new argument may be NULL if no error number is required
1065 (but then you may as well call pcre_compile(), which is now just a
1066 wrapper). This facility is provided because some applications need a
1067 numeric error indication, but it has also enabled me to tidy up the way
1068 compile-time errors are handled in the POSIX wrapper.
1069
1070 16. Added VPATH=.libs to the makefile; this should help when building with one
1071 prefix path and installing with another. (Or so I'm told by someone who
1072 knows more about this stuff than I do.)
1073
1074 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1075 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1076 match everything, including newlines. This is not POSIX-compatible, but
1077 somebody wanted the feature. From pcretest it can be activated by using
1078 both the P and the s flags.
1079
1080 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1081
1082 19. libpcre.pc was being incorrectly installed as executable.
1083
1084 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1085 it now also looks for '\r' so that it will work unmodified on Windows.
1086
1087 21. Added Google's contributed C++ wrapper to the distribution.
1088
1089 22. Added some untidy missing memory free() calls in pcretest, to keep
1090 Electric Fence happy when testing.
1091
1092
1093
1094 Version 5.0 13-Sep-04
1095 ---------------------
1096
1097 1. Internal change: literal characters are no longer packed up into items
1098 containing multiple characters in a single byte-string. Each character
1099 is now matched using a separate opcode. However, there may be more than one
1100 byte in the character in UTF-8 mode.
1101
1102 2. The pcre_callout_block structure has two new fields: pattern_position and
1103 next_item_length. These contain the offset in the pattern to the next match
1104 item, and its length, respectively.
1105
1106 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1107 insertion of callouts before each pattern item. Added the /C option to
1108 pcretest to make use of this.
1109
1110 4. On the advice of a Windows user, the lines
1111
1112 #if defined(_WIN32) || defined(WIN32)
1113 _setmode( _fileno( stdout ), 0x8000 );
1114 #endif /* defined(_WIN32) || defined(WIN32) */
1115
1116 have been added to the source of pcretest. This apparently does useful
1117 magic in relation to line terminators.
1118
1119 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1120 for the benefit of those environments where the "b" makes a difference.
1121
1122 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1123 to know about it. I have put a hack into configure.in that adds in code
1124 to set GCC=yes if CC=icc. This seems to end up at a point in the
1125 generated configure script that is early enough to affect the setting of
1126 compiler options, which is what is needed, but I have no means of testing
1127 whether it really works. (The user who reported this had patched the
1128 generated configure script, which of course I cannot do.)
1129
1130 LATER: After change 22 below (new libtool files), the configure script
1131 seems to know about icc (and also ecc). Therefore, I have commented out
1132 this hack in configure.in.
1133
1134 7. Added support for pkg-config (2 patches were sent in).
1135
1136 8. Negated POSIX character classes that used a combination of internal tables
1137 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1138 [[:^ascii]]. Typically, they would match almost any characters. The other
1139 POSIX classes were not broken in this way.
1140
1141 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1142 to find the match, as PCRE was deluded into thinking that the match had to
1143 start at the start point or following a newline. The same bug applied to
1144 patterns with negative forward assertions or any backward assertions
1145 preceding ".*" at the start, unless the pattern required a fixed first
1146 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1147
1148 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1149 starting at the last subject character, bytes beyond the end of the subject
1150 string were read.
1151
1152 11. Renamed the variable "class" as "classbits" to make life easier for C++
1153 users. (Previously there was a macro definition, but it apparently wasn't
1154 enough.)
1155
1156 12. Added the new field "tables" to the extra data so that tables can be passed
1157 in at exec time, or the internal tables can be re-selected. This allows
1158 a compiled regex to be saved and re-used at a later time by a different
1159 program that might have everything at different addresses.
1160
1161 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1162 -R library as well as a -L library.
1163
1164 14. The debugging options of pcretest (-d on the command line or D on a
1165 pattern) showed incorrect output for anything following an extended class
1166 that contained multibyte characters and which was followed by a quantifier.
1167
1168 15. Added optional support for general category Unicode character properties
1169 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1170 support. It adds about 90K to the size of the library. The meanings of the
1171 inbuilt class escapes such as \d and \s have NOT been changed.
1172
1173 16. Updated pcredemo.c to include calls to free() to release the memory for the
1174 compiled pattern.
1175
1176 17. The generated file chartables.c was being created in the source directory
1177 instead of in the building directory. This caused the build to fail if the
1178 source directory was different from the building directory, and was
1179 read-only.
1180
1181 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1182 file. No doubt somebody will tell me if they don't make sense... Also added
1183 Dan Mooney's comments about building on OpenVMS.
1184
1185 19. Added support for partial matching via the PCRE_PARTIAL option for
1186 pcre_exec() and the \P data escape in pcretest.
1187
1188 20. Extended pcretest with 3 new pattern features:
1189
1190 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1191 write the compiled pattern to the file whose name is "rest-of-line".
1192 This is a straight binary dump of the data, with the saved pointer to
1193 the character tables forced to be NULL. The study data, if any, is
1194 written too. After writing, pcretest reads a new pattern.
1195
1196 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1197 compiled pattern from the given file. There must not be any
1198 occurrences of "<" in the file name (pretty unlikely); if there are,
1199 pcretest will instead treat the initial "<" as a pattern delimiter.
1200 After reading in the pattern, pcretest goes on to read data lines as
1201 usual.
1202
1203 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1204 and 16-bit fields in a compiled pattern, to simulate a pattern that
1205 was compiled on a host of opposite endianness.
1206
1207 21. The pcre-exec() function can now cope with patterns that were compiled on
1208 hosts of opposite endianness, with this restriction:
1209
1210 As for any compiled expression that is saved and used later, the tables
1211 pointer field cannot be preserved; the extra_data field in the arguments
1212 to pcre_exec() should be used to pass in a tables address if a value
1213 other than the default internal tables were used at compile time.
1214
1215 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1216 now diagnosed as an error. Previously, most of the time, a negative number
1217 would have been treated as zero, but if in addition "ovector" was passed as
1218 NULL, a crash could occur.
1219
1220 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1221 new versions from the libtool 1.5 distribution (the last one is a copy of
1222 a file called libtool.m4). This seems to have fixed the need to patch
1223 "configure" to support Darwin 1.3 (which I used to do). However, I still
1224 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1225 workstation).
1226
1227 24. Changed the PCRE licence to be the more standard "BSD" licence.
1228
1229
1230 Version 4.5 01-Dec-03
1231 ---------------------
1232
1233 1. There has been some re-arrangement of the code for the match() function so
1234 that it can be compiled in a version that does not call itself recursively.
1235 Instead, it keeps those local variables that need separate instances for
1236 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1237 needs to "recurse". Keeping track of where control must go is done by means
1238 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1239 hide most of the details from the main code, and operates only if
1240 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1241 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1242 operating.
1243
1244 To make it easier for callers to provide specially tailored get/free
1245 functions for this usage, two new functions, pcre_stack_malloc, and
1246 pcre_stack_free, are used. They are always called in strict stacking order,
1247 and the size of block requested is always the same.
1248
1249 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1250 PCRE has been compiled to use the stack or the heap for recursion. The
1251 -C option of pcretest uses this to show which version is compiled.
1252
1253 A new data escape \S, is added to pcretest; it causes the amounts of store
1254 obtained and freed by both kinds of malloc/free at match time to be added
1255 to the output.
1256
1257 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1258 what's available on my current Linux desktop machine.
1259
1260 3. When matching a UTF-8 string, the test for a valid string at the start has
1261 been extended. If start_offset is not zero, PCRE now checks that it points
1262 to a byte that is the start of a UTF-8 character. If not, it returns
1263 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1264 this is necessary because there may be backward assertions in the pattern.
1265 When matching the same subject several times, it may save resources to use
1266 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1267
1268 4. The code for checking the validity of UTF-8 strings has been tightened so
1269 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1270 containing "overlong sequences".
1271
1272 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1273 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1274 should have been "&", but it just so happened that all the cases this let
1275 through by mistake were picked up later in the function.
1276
1277 6. I had used a variable called "isblank" - this is a C99 function, causing
1278 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1279
1280 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1281 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1282 so the version is shown; (c) stop "make test" echoing "./RunTest".
1283
1284 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1285
1286 9. The prototype for memmove() for systems that don't have it was using
1287 size_t, but the inclusion of the header that defines size_t was later. I've
1288 moved the #includes for the C headers earlier to avoid this.
1289
1290 10. Added some adjustments to the code to make it easier to compiler on certain
1291 special systems:
1292
1293 (a) Some "const" qualifiers were missing.
1294 (b) Added the macro EXPORT before all exported functions; by default this
1295 is defined to be empty.
1296 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1297 that it reads its output file name as an argument instead of writing
1298 to the standard output and assuming this can be redirected.
1299
1300 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1301 class containing characters with values greater than 255, PCRE compilation
1302 went into a loop.
1303
1304 12. A recursive reference to a subpattern that was within another subpattern
1305 that had a minimum quantifier of zero caused PCRE to crash. For example,
1306 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1307 recursion. If the recursively-called subpattern itself had a zero repeat,
1308 that was OK.
1309
1310 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1311 buffer into which it was copied (for escape processing) was still set at
1312 1024, so long lines caused crashes.
1313
1314 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1315 "internal error: code overflow...". This applied to any character class
1316 that was followed by a possessive quantifier.
1317
1318 15. Modified the Makefile to add libpcre.la as a prerequisite for
1319 libpcreposix.la because I was told this is needed for a parallel build to
1320 work.
1321
1322 16. If a pattern that contained .* following optional items at the start was
1323 studied, the wrong optimizing data was generated, leading to matching
1324 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1325 matching string must start with a or b or c. The correct conclusion for
1326 this pattern is that a match can start with any character.
1327
1328
1329 Version 4.4 13-Aug-03
1330 ---------------------
1331
1332 1. In UTF-8 mode, a character class containing characters with values between
1333 127 and 255 was not handled correctly if the compiled pattern was studied.
1334 In fixing this, I have also improved the studying algorithm for such
1335 classes (slightly).
1336
1337 2. Three internal functions had redundant arguments passed to them. Removal
1338 might give a very teeny performance improvement.
1339
1340 3. Documentation bug: the value of the capture_top field in a callout is *one
1341 more than* the number of the hightest numbered captured substring.
1342
1343 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1344 in incorrectly linking with a previously installed version. They now link
1345 explicitly with libpcre.la.
1346
1347 5. configure.in no longer needs to recognize Cygwin specially.
1348
1349 6. A problem in pcre.in for Windows platforms is fixed.
1350
1351 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1352 pcretest, it used to include the size of the study block as part of its
1353 output. Unfortunately, the structure contains a field that has a different
1354 size on different hardware architectures. This meant that the tests that
1355 showed this size failed. As the block is currently always of a fixed size,
1356 this information isn't actually particularly useful in pcretest output, so
1357 I have just removed it.
1358
1359 8. Three pre-processor statements accidentally did not start in column 1.
1360 Sadly, there are *still* compilers around that complain, even though
1361 standard C has not required this for well over a decade. Sigh.
1362
1363 9. In pcretest, the code for checking callouts passed small integers in the
1364 callout_data field, which is a void * field. However, some picky compilers
1365 complained about the casts involved for this on 64-bit systems. Now
1366 pcretest passes the address of the small integer instead, which should get
1367 rid of the warnings.
1368
1369 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1370 both compile and run time, and gives an error if an invalid UTF-8 sequence
1371 is found. There is a option for disabling this check in cases where the
1372 string is known to be correct and/or the maximum performance is wanted.
1373
1374 11. In response to a bug report, I changed one line in Makefile.in from
1375
1376 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1377 to
1378 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1379
1380 to look similar to other lines, but I have no way of telling whether this
1381 is the right thing to do, as I do not use Windows. No doubt I'll get told
1382 if it's wrong...
1383
1384
1385 Version 4.3 21-May-03
1386 ---------------------
1387
1388 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1389 Makefile.
1390
1391 2. Some refactoring to improve the quality of the code:
1392
1393 (i) The utf8_table... variables are now declared "const".
1394
1395 (ii) The code for \cx, which used the "case flipping" table to upper case
1396 lower case letters, now just substracts 32. This is ASCII-specific,
1397 but the whole concept of \cx is ASCII-specific, so it seems
1398 reasonable.
1399
1400 (iii) PCRE was using its character types table to recognize decimal and
1401 hexadecimal digits in the pattern. This is silly, because it handles
1402 only 0-9, a-f, and A-F, but the character types table is locale-
1403 specific, which means strange things might happen. A private
1404 table is now used for this - though it costs 256 bytes, a table is
1405 much faster than multiple explicit tests. Of course, the standard
1406 character types table is still used for matching digits in subject
1407 strings against \d.
1408
1409 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1410 ending in _t are). So I've renamed it as ESC_tee.
1411
1412 3. The first argument for regexec() in the POSIX wrapper should have been
1413 defined as "const".
1414
1415 4. Changed pcretest to use malloc() for its buffers so that they can be
1416 Electric Fenced for debugging.
1417
1418 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1419 to read one or more bytes before the start of the subject string. Often this
1420 had no effect on PCRE's behaviour, but in some circumstances it could
1421 provoke a segmentation fault.
1422
1423 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1424 to try to read one or more bytes before the start of the subject string.
1425
1426 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1427 UTF-8 support could misbehave in various ways if the subject string
1428 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1429 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1430 back over UTF-8 characters.)
1431
1432
1433 Version 4.2 14-Apr-03
1434 ---------------------
1435
1436 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1437
1438 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1439 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1440 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1441 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1442 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1443 and BUILD_EXEEXT
1444 Note: automatic setting of the BUILD variables is not yet working
1445 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1446 compile-time but not at link-time
1447 [LINK]: use for linking executables only
1448 make different versions for Windows and non-Windows
1449 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1450 libraries
1451 [LINK_FOR_BUILD]: new variable
1452 [OBJEXT]: use throughout
1453 [EXEEXT]: use throughout
1454 <winshared>: new target
1455 <wininstall>: new target
1456 <dftables.o>: use native compiler
1457 <dftables>: use native linker
1458 <install>: handle Windows platform correctly
1459 <clean>: ditto
1460 <check>: ditto
1461 copy DLL to top builddir before testing
1462
1463 As part of these changes, -no-undefined was removed again. This was reported
1464 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1465 in any case.
1466
1467 3. Some tidies to get rid of compiler warnings:
1468
1469 . In the match_data structure, match_limit was an unsigned long int, whereas
1470 match_call_count was an int. I've made them both unsigned long ints.
1471
1472 . In pcretest the fact that a const uschar * doesn't automatically cast to
1473 a void * provoked a warning.
1474
1475 . Turning on some more compiler warnings threw up some "shadow" variables
1476 and a few more missing casts.
1477
1478 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1479 option, a class that contained a single character with a value between 128
1480 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1481
1482 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1483 option, a class that contained several characters, but with at least one
1484 whose value was between 128 and 255 caused PCRE to crash.
1485
1486
1487 Version 4.1 12-Mar-03
1488 ---------------------
1489
1490 1. Compiling with gcc -pedantic found a couple of places where casts were
1491 needed, and a string in dftables.c that was longer than standard compilers are
1492 required to support.
1493
1494 2. Compiling with Sun's compiler found a few more places where the code could
1495 be tidied up in order to avoid warnings.
1496
1497 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1498 first of these names is deprecated in the latest Autoconf in favour of the name
1499 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1500 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1501 analogy I have changed it to CFLAGS_FOR_BUILD.
1502
1503 4. Added -no-undefined to the linking command in the Makefile, because this is
1504 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1505 linking step for the pcreposix library.
1506
1507 5. PCRE was failing to diagnose the case of two named groups with the same
1508 name.
1509
1510 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1511 literal character that is needed in the subject for a match, and scans along to
1512 ensure that it is present before embarking on the full matching process. This
1513 saves time in cases of nested unlimited repeats that are never going to match.
1514 Problem: the scan can take a lot of time if the subject is very long (e.g.
1515 megabytes), thus penalizing straightforward matches. It is now done only if the
1516 amount of subject to be scanned is less than 1000 bytes.
1517
1518 7. A lesser problem with the same optimization is that it was recording the
1519 first character of an anchored pattern as "needed", thus provoking a search
1520 right along the subject, even when the first match of the pattern was going to
1521 fail. The "needed" character is now not set for anchored patterns, unless it
1522 follows something in the pattern that is of non-fixed length. Thus, it still
1523 fulfils its original purpose of finding quick non-matches in cases of nested
1524 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1525
1526
1527 Version 4.0 17-Feb-03
1528 ---------------------
1529
1530 1. If a comment in an extended regex that started immediately after a meta-item
1531 extended to the end of string, PCRE compiled incorrect data. This could lead to
1532 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
1533
1534 2. Moved to autoconf 2.53 and libtool 1.4.2.
1535
1536 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
1537 the special perltest8 script is no longer needed - all the tests can be run
1538 from a single perltest script.
1539
1540 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
1541 by \s. It has now been removed in PCRE. This means it isn't recognized as
1542 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
1543 class [:space:] *does* include VT, thereby creating a mess.
1544
1545 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
1546 space and tab.
1547
1548 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
1549 its new features into the main test script, reducing the number of scripts.
1550
1551 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
1552 were backward compatible, and made the (?i) apply to the whole pattern, as if
1553 /i were given. Now it behaves more logically, and applies the option setting
1554 only to what follows. PCRE has been changed to follow suit. However, if it
1555 finds options settings right at the start of the pattern, it extracts them into
1556 the global options, as before. Thus, they show up in the info data.
1557
1558 8. Added support for the \Q...\E escape sequence. Characters in between are
1559 treated as literals. This is slightly different from Perl in that $ and @ are
1560 also handled as literals inside the quotes. In Perl, they will cause variable
1561 interpolation. Note the following examples:
1562
1563 Pattern PCRE matches Perl matches
1564
1565 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
1566 \Qabc\$xyz\E abc\$xyz abc\$xyz
1567 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
1568
1569 For compatibility with Perl, \Q...\E sequences are recognized inside character
1570 classes as well as outside them.
1571
1572 9. Re-organized 3 code statements in pcretest to avoid "overflow in
1573 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
1574 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
1575 signed/unsigned warnings.
1576
1577 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
1578 option for pcretest, so I've replaced it by a simple function that does just
1579 that job.
1580
1581 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
1582 "pcregrep -".
1583
1584 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
1585 Java package. This provides some syntactic sugar for simple cases of what my
1586 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
1587 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
1588 item, you can use this simplified notation. Note that only makes sense with
1589 greedy quantifiers. Consequently, the use of the possessive quantifier forces
1590 greediness, whatever the setting of the PCRE_UNGREEDY option.
1591
1592 13. A change of greediness default within a pattern was not taking effect at
1593 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
1594 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
1595 was abstracted outside.
1596
1597 14. PCRE now supports the \G assertion. It is true when the current matching
1598 position is at the start point of the match. This differs from \A when the
1599 starting offset is non-zero. Used with the /g option of pcretest (or similar
1600 code), it works in the same way as it does for Perl's /g option. If all
1601 alternatives of a regex begin with \G, the expression is anchored to the start
1602 match position, and the "anchored" flag is set in the compiled expression.
1603
1604 15. Some bugs concerning the handling of certain option changes within patterns
1605 have been fixed. These applied to options other than (?ims). For example,
1606 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
1607 been the other way round. Some of this was related to change 7 above.
1608
1609 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
1610 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
1611 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
1612 POSIX classes only within a class (e.g. /[[:alpha:]]/).
1613
1614 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
1615 mode. Unlike ".", it always matches newline, whatever the setting of
1616 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
1617 assertions. Perl allows it, but it doesn't (in general) work because it can't
1618 calculate the length of the lookbehind. At least, that's the case for Perl
1619 5.8.0 - I've been told they are going to document that it doesn't work in
1620 future.
1621
1622 18. Added an error diagnosis for escapes that PCRE does not support: these are
1623 \L, \l, \N, \P, \p, \U, \u, and \X.
1624
1625 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
1626 reading past the end of the pattern in cases such as /[abcd/.
1627
1628 20. PCRE was getting more memory than necessary for patterns with classes that
1629 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
1630
1631 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
1632 compiling PCRE for use with Virtual Pascal.
1633
1634 22. Small fix to the Makefile to make it work properly if the build is done
1635 outside the source tree.
1636
1637 23. Added a new extension: a condition to go with recursion. If a conditional
1638 subpattern starts with (?(R) the "true" branch is used if recursion has
1639 happened, whereas the "false" branch is used only at the top level.
1640
1641 24. When there was a very long string of literal characters (over 255 bytes
1642 without UTF support, over 250 bytes with UTF support), the computation of how
1643 much memory was required could be incorrect, leading to segfaults or other
1644 strange effects.
1645
1646 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
1647 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
1648 there was a subsequent back reference to those brackets. This meant that, for
1649 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
1650 possible to check for precisely this case. All we can do is abandon the
1651 optimization if .* occurs inside capturing brackets when there are any back
1652 references whatsoever. (See below for a better fix that came later.)
1653
1654 26. The handling of the optimization for finding the first character of a
1655 non-anchored pattern, and for finding a character that is required later in the
1656 match were failing in some cases. This didn't break the matching; it just
1657 failed to optimize when it could. The way this is done has been re-implemented.
1658
1659 27. Fixed typo in error message for invalid (?R item (it said "(?p").
1660
1661 28. Added a new feature that provides some of the functionality that Perl
1662 provides with (?{...}). The facility is termed a "callout". The way it is done
1663 in PCRE is for the caller to provide an optional function, by setting
1664 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
1665 global variable. By default it is unset, which disables all calling out. To get
1666 the function called, the regex must include (?C) at appropriate points. This
1667 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
1668 This provides a means of identifying different callout points. When PCRE
1669 reaches such a point in the regex, if pcre_callout has been set, the external
1670 function is called. It is provided with data in a structure called
1671 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
1672 matching continues; if it returns a non-zero value, the match at the current
1673 point fails. However, backtracking will occur if possible. [This was changed
1674 later and other features added - see item 49 below.]
1675
1676 29. pcretest is upgraded to test the callout functionality. It provides a
1677 callout function that displays information. By default, it shows the start of
1678 the match and the current position in the text. There are some new data escapes
1679 to vary what happens:
1680
1681 \C+ in addition, show current contents of captured substrings
1682 \C- do not supply a callout function
1683 \C!n return 1 when callout number n is reached
1684 \C!n!m return 1 when callout number n is reached for the mth time
1685
1686 30. If pcregrep was called with the -l option and just a single file name, it
1687 output "<stdin>" if a match was found, instead of the file name.
1688
1689 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
1690 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
1691 pcre_exec(). This saves a malloc/free per call. The default value of
1692 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
1693 when configuring.
1694
1695 32. The default maximum size of a compiled pattern is 64K. There have been a
1696 few cases of people hitting this limit. The code now uses macros to handle the
1697 storing of links as offsets within the compiled pattern. It defaults to 2-byte
1698 links, but this can be changed to 3 or 4 bytes by --with-link-size when
1699 configuring. Tests 2 and 5 work only with 2-byte links because they output
1700 debugging information about compiled patterns.
1701
1702 33. Internal code re-arrangements:
1703
1704 (a) Moved the debugging function for printing out a compiled regex into
1705 its own source file (printint.c) and used #include to pull it into
1706 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
1707 separate copies.
1708
1709 (b) Defined the list of op-code names for debugging as a macro in
1710 internal.h so that it is next to the definition of the opcodes.
1711
1712 (c) Defined a table of op-code lengths for simpler skipping along compiled
1713 code. This is again a macro in internal.h so that it is next to the
1714 definition of the opcodes.
1715
1716 34. Added support for recursive calls to individual subpatterns, along the
1717 lines of Robin Houston's patch (but implemented somewhat differently).
1718
1719 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
1720 allow it to read and process whole directories in Win32. This code was
1721 contributed by Lionel Fourquaux; it has not been tested by me.
1722
1723 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
1724 used to name a group. Names consist of alphanumerics and underscores, and must
1725 be unique. Back references use the syntax (?P=name) and recursive calls use
1726 (?P>name) which is a PCRE extension to the Python extension. Groups still have
1727 numbers. The function pcre_fullinfo() can be used after compilation to extract
1728 a name/number map. There are three relevant calls:
1729
1730 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
1731 PCRE_INFO_NAMECOUNT yields the number of entries
1732 PCRE_INFO_NAMETABLE yields a pointer to the map.
1733
1734 The map is a vector of fixed-size entries. The size of each entry depends on
1735 the length of the longest name used. The first two bytes of each entry are the
1736 group number, most significant byte first. There follows the corresponding
1737 name, zero terminated. The names are in alphabetical order.
1738
1739 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
1740 case instead of 255. Making it the same both with and without UTF-8 support
1741 means that the same test output works with both.
1742
1743 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
1744 calling malloc() with a zero argument.
1745
1746 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
1747 optimization. I've improved things by keeping a bitmap of backreferences with
1748 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
1749 fact referenced, the optimization can be applied. It is unlikely that a
1750 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
1751 the match to follow \n) will appear inside brackets with a number greater than
1752 31, but if it does, any back reference > 31 suppresses the optimization.
1753
1754 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
1755 of disabling numbered capturing parentheses. Any opening parenthesis that is
1756 not followed by ? behaves as if it were followed by ?: but named parentheses
1757 can still be used for capturing (and they will acquire numbers in the usual
1758 way).
1759
1760 41. Redesigned the return codes from the match() function into yes/no/error so
1761 that errors can be passed back from deep inside the nested calls. A malloc
1762 failure while inside a recursive subpattern call now causes the
1763 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
1764
1765 42. It is now possible to set a limit on the number of times the match()
1766 function is called in a call to pcre_exec(). This facility makes it possible to
1767 limit the amount of recursion and backtracking, though not in a directly
1768 obvious way, because the match() function is used in a number of different
1769 circumstances. The count starts from zero for each position in the subject
1770 string (for non-anchored patterns). The default limit is, for compatibility, a
1771 large number, namely 10 000 000. You can change this in two ways:
1772
1773 (a) When configuring PCRE before making, you can use --with-match-limit=n
1774 to set a default value for the compiled library.
1775
1776 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
1777 a different value is set. See 45 below.
1778
1779 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1780
1781 43. Added a new function pcre_config(int, void *) to enable run-time extraction
1782 of things that can be changed at compile time. The first argument specifies
1783 what is wanted and the second points to where the information is to be placed.
1784 The current list of available information is:
1785
1786 PCRE_CONFIG_UTF8
1787
1788 The output is an integer that is set to one if UTF-8 support is available;
1789 otherwise it is set to zero.
1790
1791 PCRE_CONFIG_NEWLINE
1792
1793 The output is an integer that it set to the value of the code that is used for
1794 newline. It is either LF (10) or CR (13).
1795
1796 PCRE_CONFIG_LINK_SIZE
1797
1798 The output is an integer that contains the number of bytes used for internal
1799 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
1800
1801 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
1802
1803 The output is an integer that contains the threshold above which the POSIX
1804 interface uses malloc() for output vectors. See item 31 above.
1805
1806 PCRE_CONFIG_MATCH_LIMIT
1807
1808 The output is an unsigned integer that contains the default limit of the number
1809 of match() calls in a pcre_exec() execution. See 42 above.
1810
1811 44. pcretest has been upgraded by the addition of the -C option. This causes it
1812 to extract all the available output from the new pcre_config() function, and to
1813 output it. The program then exits immediately.
1814
1815 45. A need has arisen to pass over additional data with calls to pcre_exec() in
1816 order to support additional features. One way would have been to define
1817 pcre_exec2() (for example) with extra arguments, but this would not have been
1818 extensible, and would also have required all calls to the original function to
1819 be mapped to the new one. Instead, I have chosen to extend the mechanism that
1820 is used for passing in "extra" data from pcre_study().
1821
1822 The pcre_extra structure is now exposed and defined in pcre.h. It currently
1823 contains the following fields:
1824
1825 flags a bitmap indicating which of the following fields are set
1826 study_data opaque data from pcre_study()
1827 match_limit a way of specifying a limit on match() calls for a specific
1828 call to pcre_exec()
1829 callout_data data for callouts (see 49 below)
1830
1831 The flag bits are also defined in pcre.h, and are
1832
1833 PCRE_EXTRA_STUDY_DATA
1834 PCRE_EXTRA_MATCH_LIMIT
1835 PCRE_EXTRA_CALLOUT_DATA
1836
1837 The pcre_study() function now returns one of these new pcre_extra blocks, with
1838 the actual study data pointed to by the study_data field, and the
1839 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
1840 before. That is, this change is entirely upwards-compatible and requires no
1841 change to existing code.
1842
1843 If you want to pass in additional data to pcre_exec(), you can either place it
1844 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
1845 block.
1846
1847 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
1848 data string contains the escape sequence \M, pcretest calls pcre_exec() several
1849 times with different match limits, until it finds the minimum value needed for
1850 pcre_exec() to complete. The value is then output. This can be instructive; for
1851 most simple matches the number is quite small, but for pathological cases it
1852 gets very large very quickly.
1853
1854 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
1855 returns the size of the data block pointed to by the study_data field in a
1856 pcre_extra block, that is, the value that was passed as the argument to
1857 pcre_malloc() when PCRE was getting memory in which to place the information
1858 created by pcre_study(). The fourth argument should point to a size_t variable.
1859 pcretest has been extended so that this information is shown after a successful
1860 pcre_study() call when information about the compiled regex is being displayed.
1861
1862 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
1863 because what follows is always an absolute path. (Later: it turns out that this
1864 is more than cosmetic for MinGW, because it doesn't like empty path
1865 components.)
1866
1867 49. Some changes have been made to the callout feature (see 28 above):
1868
1869 (i) A callout function now has three choices for what it returns:
1870
1871 0 => success, carry on matching
1872 > 0 => failure at this point, but backtrack if possible
1873 < 0 => serious error, return this value from pcre_exec()
1874
1875 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
1876 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
1877 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
1878 use by callout functions. It will never be used by PCRE itself.
1879
1880 (ii) The pcre_extra structure (see 45 above) has a void * field called
1881 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
1882 pcre_callout_block structure has a field of the same name. The contents of
1883 the field passed in the pcre_extra structure are passed to the callout
1884 function in the corresponding field in the callout block. This makes it
1885 easier to use the same callout-containing regex from multiple threads. For
1886 testing, the pcretest program has a new data escape
1887
1888 \C*n pass the number n (may be negative) as callout_data
1889
1890 If the callout function in pcretest receives a non-zero value as
1891 callout_data, it returns that value.
1892
1893 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
1894 there were some redundant $(CFLAGS) in commands that are now specified as
1895 $(LINK), which already includes $(CFLAGS).
1896
1897 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
1898 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
1899 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
1900 one-byte characters throughout. Note that case-insensitive matching applies
1901 only to characters whose values are less than 256. PCRE doesn't support the
1902 notion of cases for higher-valued characters.
1903
1904 (i) A character class whose characters are all within 0-255 is handled as
1905 a bit map, and the map is inverted for negative classes. Previously, a
1906 character > 255 always failed to match such a class; however it should
1907 match if the class was a negative one (e.g. [^ab]). This has been fixed.
1908
1909 (ii) A negated character class with a single character < 255 is coded as
1910 "not this character" (OP_NOT). This wasn't working properly when the test
1911 character was multibyte, either singly or repeated.
1912
1913 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
1914 mode, for example: \x{100}{2,3}.
1915
1916 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
1917 singly or repeated) now correctly test multibyte characters. However,
1918 PCRE doesn't recognize any characters with values greater than 255 as
1919 digits, spaces, or word characters. Such characters always match \D, \S,
1920 and \W, and never match \d, \s, or \w.
1921
1922 (v) Classes may now contain characters and character ranges with values
1923 greater than 255. For example: [ab\x{100}-\x{400}].
1924
1925 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
1926 PCRE in UTF-8 mode.
1927
1928 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
1929 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
1930 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
1931 value.)
1932
1933 53. The single man page has become too large. I have therefore split it up into
1934 a number of separate man pages. These also give rise to individual HTML pages;
1935 these are now put in a separate directory, and there is an index.html page that
1936 lists them all. Some hyperlinking between the pages has been installed.
1937
1938 54. Added convenience functions for handling named capturing parentheses.
1939
1940 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
1941 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
1942 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
1943 are faulted.
1944
1945 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
1946 calling configure. These values are used when compiling the dftables.c program
1947 which is run to generate the source of the default character tables. They
1948 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
1949 you will need to set these values.
1950
1951 57. Updated the building process for Windows DLL, as provided by Fred Cox.
1952
1953
1954 Version 3.9 02-Jan-02
1955 ---------------------
1956
1957 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
1958
1959 2. If --disable-static was given, the building process failed when trying to
1960 build pcretest and pcregrep. (For some reason it was using libtool to compile
1961 them, which is not right, as they aren't part of the library.)
1962
1963
1964 Version 3.8 18-Dec-01
1965 ---------------------
1966
1967 1. The experimental UTF-8 code was completely screwed up. It was packing the
1968 bytes in the wrong order. How dumb can you get?
1969
1970
1971 Version 3.7 29-Oct-01
1972 ---------------------
1973
1974 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
1975 This caused pcretest, when used on the test data, to segfault. Unfortunately,
1976 this didn't happen under Solaris 8, where I normally test things.
1977
1978 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
1979 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
1980 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
1981 this fix an hour or so after the initial 3.7 release.)
1982
1983
1984 Version 3.6 23-Oct-01
1985 ---------------------
1986
1987 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
1988 offsets passed as NULL with zero offset count.
1989
1990 2. The config.guess and config.sub files had not been updated when I moved to
1991 the latest autoconf.
1992
1993
1994 Version 3.5 15-Aug-01
1995 ---------------------
1996
1997 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
1998 had been forgotten.
1999
2000 2. By using declared but undefined structures, we can avoid using "void"
2001 definitions in pcre.h while keeping the internal definitions of the structures
2002 private.
2003
2004 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2005 user point of view, this means that both static and shared libraries are built
2006 by default, but this can be individually controlled. More of the work of
2007 handling this static/shared cases is now inside libtool instead of PCRE's make
2008 file.
2009
2010 4. The pcretest utility is now installed along with pcregrep because it is
2011 useful for users (to test regexs) and by doing this, it automatically gets
2012 relinked by libtool. The documentation has been turned into a man page, so
2013 there are now .1, .txt, and .html versions in /doc.
2014
2015 5. Upgrades to pcregrep:
2016 (i) Added long-form option names like gnu grep.
2017 (ii) Added --help to list all options with an explanatory phrase.
2018 (iii) Added -r, --recursive to recurse into sub-directories.
2019 (iv) Added -f, --file to read patterns from a file.
2020
2021 6. pcre_exec() was referring to its "code" argument before testing that
2022 argument for NULL (and giving an error if it was NULL).
2023
2024 7. Upgraded Makefile.in to allow for compiling in a different directory from
2025 the source directory.
2026
2027 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2028 options bits, the pointer it was passed was to an int instead of to an unsigned
2029 long int. This mattered only on 64-bit systems.
2030
2031 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2032 generated) instead of pcre.in, which it its source. Also made the same change
2033 in several of the .c files.
2034
2035 10. A new release of gcc defines printf() as a macro, which broke pcretest
2036 because it had an ifdef in the middle of a string argument for printf(). Fixed
2037 by using separate calls to printf().
2038
2039 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2040 script, to force use of CR or LF instead of \n in the source. On non-Unix
2041 systems, the value can be set in config.h.
2042
2043 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2044 absolute limit. Changed the text of the error message to make this clear, and
2045 likewise updated the man page.
2046
2047 13. The limit of 99 on the number of capturing subpatterns has been removed.
2048 The new limit is 65535, which I hope will not be a "real" limit.
2049
2050
2051 Version 3.4 22-Aug-00
2052 ---------------------
2053
2054 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2055
2056 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2057
2058
2059 Version 3.3 01-Aug-00
2060 ---------------------
2061
2062 1. If an octal character was given, but the value was greater than \377, it
2063 was not getting masked to the least significant bits, as documented. This could
2064 lead to crashes in some systems.
2065
2066 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2067 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2068
2069 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2070 These just pass their arguments on to (pcre_free)(), but they are provided
2071 because some uses of PCRE bind it to non-C systems that can call its functions,
2072 but cannot call free() or pcre_free() directly.
2073
2074 4. Add "make test" as a synonym for "make check". Corrected some comments in
2075 the Makefile.
2076
2077 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2078 Makefile.
2079
2080 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2081 command called pgrep for grepping around the active processes.
2082
2083 7. Added the beginnings of support for UTF-8 character strings.
2084
2085 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2086 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2087 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2088 out for the ar command.)
2089
2090
2091 Version 3.2 12-May-00
2092 ---------------------
2093
2094 This is purely a bug fixing release.
2095
2096 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2097 of ZA. This was just one example of several cases that could provoke this bug,
2098 which was introduced by change 9 of version 2.00. The code for breaking
2099 infinite loops after an iteration that matches an empty string was't working
2100 correctly.
2101
2102 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2103 when matched against abbab (for example). After matching an empty string, it
2104 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2105 caused it to match further down the string than it should.
2106
2107 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2108 was there because it doesn't seem to be needed, and it causes trouble on some
2109 systems, as it is not a Standard C header. It has been removed.
2110
2111 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2112 were reported on the Macintosh. The changes were from
2113
2114 while ((c = *(++ptr)) != 0 && c != '\n');
2115 to
2116 while ((c = *(++ptr)) != 0 && c != '\n') ;
2117
2118 Totally extraordinary, but if that's what it takes...
2119
2120 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2121 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2122 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2123 assumes the way PCRE uses memmove() (always moving upwards).
2124
2125 6. PCRE is being used in one environment where strchr() is not available. There
2126 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2127 faster code anyway.
2128
2129
2130 Version 3.1 09-Feb-00
2131 ---------------------
2132
2133 The only change in this release is the fixing of some bugs in Makefile.in for
2134 the "install" target:
2135
2136 (1) It was failing to install pcreposix.h.
2137
2138 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2139
2140
2141 Version 3.0 01-Feb-00
2142 ---------------------
2143
2144 1. Add support for the /+ modifier to perltest (to output $` like it does in
2145 pcretest).
2146
2147 2. Add support for the /g modifier to perltest.
2148
2149 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2150 matches null strings.
2151
2152 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2153 pattern. Perl treats empty patterns specially - it reuses the most recent
2154 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2155 effect.
2156
2157 5. The POSIX interface was broken in that it was just handing over the POSIX
2158 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2159 required a bigger vector, with some working space on the end. This means that
2160 the POSIX wrapper now has to get and free some memory, and copy the results.
2161
2162 6. Added some simple autoconf support, placing the test data and the
2163 documentation in separate directories, re-organizing some of the
2164 information files, and making it build pcre-config (a GNU standard). Also added
2165 libtool support for building PCRE as a shared library, which is now the
2166 default.
2167
2168 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2169 09 are not valid octal constants. Single digits will be used for minor values
2170 less than 10.
2171
2172 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2173 existing programs that set these in the POSIX interface can use PCRE without
2174 modification.
2175
2176 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2177 return all that pcre_info() returns, plus additional data. The pcre_info()
2178 function is retained for compatibility, but is considered to be obsolete.
2179
2180 10. Added experimental recursion feature (?R) to handle one common case that
2181 Perl 5.6 will be able to do with (?p{...}).
2182
2183 11. Added support for POSIX character classes like [:alpha:], which Perl is
2184 adopting.
2185
2186
2187 Version 2.08 31-Aug-99
2188 ----------------------
2189
2190 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2191 trying to match at the startoffset position, but instead was moving forward to
2192 the next newline as if a previous match had failed.
2193
2194 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2195 and could get into a loop if a null string was matched other than at the start
2196 of the subject.
2197
2198 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2199 be distinguished at compile time, and for completeness also added PCRE_DATE.
2200
2201 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2202 in GnuWin32 environments.
2203
2204
2205 Version 2.07 29-Jul-99
2206 ----------------------
2207
2208 1. The documentation is now supplied in plain text form and HTML as well as in
2209 the form of man page sources.
2210
2211 2. C++ compilers don't like assigning (void *) values to other pointer types.
2212 In particular this affects malloc(). Although there is no problem in Standard
2213 C, I've put in casts to keep C++ compilers happy.
2214
2215 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2216 should be (const char *).
2217
2218 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2219 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2220 However, I haven't made this a standard facility. The documentation doesn't
2221 mention it, and the Makefile doesn't support it.
2222
2223 5. The Makefile now contains an "install" target, with editable destinations at
2224 the top of the file. The pcretest program is not installed.
2225
2226 6. pgrep -V now gives the PCRE version number and date.
2227
2228 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2229 causing the entire string to be ignored, instead of just the last character.
2230
2231 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2232 non-matching string, it can take a very, very long time, even for strings of
2233 quite modest length, because of the nested recursion. PCRE now does better in
2234 some of these cases. It does this by remembering the last required literal
2235 character in the pattern, and pre-searching the subject to ensure it is present
2236 before running the real match. In other words, it applies a heuristic to detect
2237 some types of certain failure quickly, and in the above example, if presented
2238 with a string that has no trailing " it gives "no match" very quickly.
2239
2240 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2241 other alternatives are tried instead.
2242
2243
2244 Version 2.06 09-Jun-99
2245 ----------------------
2246
2247 1. Change pcretest's output for amount of store used to show just the code
2248 space, because the remainder (the data block) varies in size between 32-bit and
2249 64-bit systems.
2250
2251 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2252 start matching at. This allows lookbehinds to work when searching for multiple
2253 occurrences in a string.
2254
2255 3. Added additional options to pcretest for testing multiple occurrences:
2256
2257 /+ outputs the rest of the string that follows a match
2258 /g loops for multiple occurrences, using the new startoffset argument
2259 /G loops for multiple occurrences by passing an incremented pointer
2260
2261 4. PCRE wasn't doing the "first character" optimization for patterns starting
2262 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2263 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2264 the letter 'x'. On long subject strings, this gives a significant speed-up.
2265
2266
2267 Version 2.05 21-Apr-99
2268 ----------------------
2269
2270 1. Changed the type of magic_number from int to long int so that it works
2271 properly on 16-bit systems.
2272
2273 2. Fixed a bug which caused patterns starting with .* not to work correctly
2274 when the subject string contained newline characters. PCRE was assuming
2275 anchoring for such patterns in all cases, which is not correct because .* will
2276 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2277 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2278 must be retried after every newline in the subject.
2279
2280
2281 Version 2.04 18-Feb-99
2282 ----------------------
2283
2284 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2285 computation of the store needed to hold the pattern was incorrect (too large).
2286 If such patterns were nested a few deep, this could multiply and become a real
2287 problem.
2288
2289 2. Added /M option to pcretest to show the memory requirement of a specific
2290 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2291
2292 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2293 compiled in such a way that the backtracking after subsequent failure was
2294 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2295 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2296
2297
2298 Version 2.03 02-Feb-99
2299 ----------------------
2300
2301 1. Fixed typo and small mistake in man page.
2302
2303 2. Added 4th condition (GPL supersedes if conflict) and created separate
2304 LICENCE file containing the conditions.
2305
2306 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2307 Perl, that is the internal \ allows the delimiter to be included in the
2308 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2309 the final delimiter, add \ to the end of the pattern (to test the error).
2310
2311 4. Added the convenience functions for extracting substrings after a successful
2312 match. Updated pcretest to make it able to test these functions.
2313
2314
2315 Version 2.02 14-Jan-99
2316 ----------------------
2317
2318 1. Initialized the working variables associated with each extraction so that
2319 their saving and restoring doesn't refer to uninitialized store.
2320
2321 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2322 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2323 fix the problem.
2324
2325 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2326 calls, and wasn't printing the correct value for compiling calls. Increased the
2327 default value of LOOPREPEAT, and the number of significant figures in the
2328 times.
2329
2330 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2331
2332 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2333 a building problem on Windows NT with a FAT file system.
2334
2335
2336 Version 2.01 21-Oct-98
2337 ----------------------
2338
2339 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2340 to character tables built by pcre_maketables() in the current locale. If NULL
2341 is passed, the default tables are used.
2342
2343
2344 Version 2.00 24-Sep-98
2345 ----------------------
2346
2347 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2348 it any more.
2349
2350 2. Allow quantification of (?>) groups, and make it work correctly.
2351
2352 3. The first character computation wasn't working for (?>) groups.
2353
2354 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2355 end of the subject) and add 5.005's \z, which really does match only at the
2356 very end of the subject.
2357
2358 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2359
2360 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2361 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2362 localized options. All options to pcre_study() were also removed.
2363
2364 7. Add other new features from 5.005:
2365
2366 $(?<= positive lookbehind
2367 $(?<! negative lookbehind
2368 (?imsx-imsx) added the unsetting capability
2369 such a setting is global if at outer level; local otherwise
2370 (?imsx-imsx:) non-capturing groups with option setting
2371 (?(cond)re|re) conditional pattern matching
2372
2373 A backreference to itself in a repeated group matches the previous
2374 captured string.
2375
2376 8. General tidying up of studying (both automatic and via "study")
2377 consequential on the addition of new assertions.
2378
2379 9. As in 5.005, unlimited repeated groups that could match an empty substring
2380 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2381 runtime if any iteration does actually match an empty substring.
2382
2383 10. Include the RunTest script in the distribution.
2384
2385 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2386 discrepancies, some of which were old and were also with respect to 5.004. They
2387 have now been fixed.
2388
2389
2390 Version 1.09 28-Apr-98
2391 ----------------------
2392
2393 1. A negated single character class followed by a quantifier with a minimum
2394 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2395 program crashes, or just wrong answers. This did not apply to negated classes
2396 containing more than one character, or to minima other than one.
2397
2398
2399 Version 1.08 27-Mar-98
2400 ----------------------
2401
2402 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2403
2404 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2405 latter must appear before anything that relies on it in the pattern.
2406
2407
2408 Version 1.07 16-Feb-98
2409 ----------------------
2410
2411 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2412 repeat of a potentially empty string).
2413
2414
2415 Version 1.06 23-Jan-98
2416 ----------------------
2417
2418 1. Added Markus Oberhumer's little patches for C++.
2419
2420 2. Literal strings longer than 255 characters were broken.
2421
2422
2423 Version 1.05 23-Dec-97
2424 ----------------------
2425
2426 1. Negated character classes containing more than one character were failing if
2427 PCRE_CASELESS was set at run time.
2428
2429
2430 Version 1.04 19-Dec-97
2431 ----------------------
2432
2433 1. Corrected the man page, where some "const" qualifiers had been omitted.
2434
2435 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2436 input syntax.
2437
2438 3. Fixed memory leak which occurred when a regex with back references was
2439 matched with an offsets vector that wasn't big enough. The temporary memory
2440 that is used in this case wasn't being freed if the match failed.
2441
2442 4. Tidied pcretest to ensure it frees memory that it gets.
2443
2444 5. Temporary memory was being obtained in the case where the passed offsets
2445 vector was exactly big enough.
2446
2447 6. Corrected definition of offsetof() from change 5 below.
2448
2449 7. I had screwed up change 6 below and broken the rules for the use of
2450 setjmp(). Now fixed.
2451
2452
2453 Version 1.03 18-Dec-97
2454 ----------------------
2455
2456 1. A erroneous regex with a missing opening parenthesis was correctly
2457 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2458 on some systems.
2459
2460 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2461 it was reported that one broken compiler failed on the former because "code" is
2462 also an independent variable.
2463
2464 3. The erroneous regex a[]b caused an array overrun reference.
2465
2466 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2467 fail on data ending with that character. (It was going on too far, and checking
2468 the next character, typically a binary zero.) This was specific to the
2469 optimized code for single-character negative classes.
2470
2471 5. Added a contributed patch from the TIN world which does the following:
2472
2473 + Add an undef for memmove, in case the the system defines a macro for it.
2474
2475 + Add a definition of offsetof(), in case there isn't one. (I don't know
2476 the reason behind this - offsetof() is part of the ANSI standard - but
2477 it does no harm).
2478
2479 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2480 most of the places where whitespace preceded '#'. I have given up and
2481 allowed the remaining 2 cases to be at the margin.
2482
2483 + Rename some variables in pcre to eliminate shadowing. This seems very
2484 pedantic, but does no harm, of course.
2485
2486 6. Moved the call to setjmp() into its own function, to get rid of warnings
2487 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2488
2489 7. Constructs such as \d{8,} were compiling into the equivalent of
2490 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2491 outcome, but in this particular case used more store than had been allocated,
2492 which caused the bug to be discovered because it threw up an internal error.
2493
2494 8. The debugging code in both pcre and pcretest for outputting the compiled
2495 form of a regex was going wrong in the case of back references followed by
2496 curly-bracketed repeats.
2497
2498
2499 Version 1.02 12-Dec-97
2500 ----------------------
2501
2502 1. Typos in pcre.3 and comments in the source fixed.
2503
2504 2. Applied a contributed patch to get rid of places where it used to remove
2505 'const' from variables, and fixed some signed/unsigned and uninitialized
2506 variable warnings.
2507
2508 3. Added the "runtest" target to Makefile.
2509
2510 4. Set default compiler flag to -O2 rather than just -O.
2511
2512
2513 Version 1.01 19-Nov-97
2514 ----------------------
2515
2516 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2517 like /([ab]*)*/, that is, for classes with more than one character in them.
2518
2519 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2520 as /((?>a*))*/ (a PCRE_EXTRA facility).
2521
2522
2523 Version 1.00 18-Nov-97
2524 ----------------------
2525
2526 1. Added compile-time macros to support systems such as SunOS4 which don't have
2527 memmove() or strerror() but have other things that can be used instead.
2528
2529 2. Arranged that "make clean" removes the executables.
2530
2531
2532 Version 0.99 27-Oct-97
2533 ----------------------
2534
2535 1. Fixed bug in code for optimizing classes with only one character. It was
2536 initializing a 32-byte map regardless, which could cause it to run off the end
2537 of the memory it had got.
2538
2539 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
2540
2541
2542 Version 0.98 22-Oct-97
2543 ----------------------
2544
2545 1. Fixed bug in code for handling temporary memory usage when there are more
2546 back references than supplied space in the ovector. This could cause segfaults.
2547
2548
2549 Version 0.97 21-Oct-97
2550 ----------------------
2551
2552 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
2553
2554 2. Optimized negated single characters not to use a bit map.
2555
2556 3. Brought error texts together as macro definitions; clarified some of them;
2557 fixed one that was wrong - it said "range out of order" when it meant "invalid
2558 escape sequence".
2559
2560 4. Changed some char * arguments to const char *.
2561
2562 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
2563
2564 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
2565 pcretest.
2566
2567
2568 Version 0.96 16-Oct-97
2569 ----------------------
2570
2571 1. Added a simple "pgrep" utility to the distribution.
2572
2573 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
2574 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
2575 where "ddd" means "one or more decimal digits".
2576
2577 3. Fixed serious bug. If a pattern had a back reference, but the call to
2578 pcre_exec() didn't supply a large enough ovector to record the related
2579 identifying subpattern, the match always failed. PCRE now remembers the number
2580 of the largest back reference, and gets some temporary memory in which to save
2581 the offsets during matching if necessary, in order to ensure that
2582 backreferences always work.
2583
2584 4. Increased the compatibility with Perl in a number of ways:
2585
2586 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
2587 to request this handling. The option can be set at compile or exec time.
2588
2589 (b) $ matches before a terminating newline by default; an option
2590 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
2591 mode). The option can be set at compile or exec time.
2592
2593 (c) The handling of \ followed by a digit other than 0 is now supposed to be
2594 the same as Perl's. If the decimal number it represents is less than 10
2595 or there aren't that many previous left capturing parentheses, an octal
2596 escape is read. Inside a character class, it's always an octal escape,
2597 even if it is a single digit.
2598
2599 (d) An escaped but undefined alphabetic character is taken as a literal,
2600 unless PCRE_EXTRA is set. Currently this just reserves the remaining
2601 escapes.
2602
2603 (e) {0} is now permitted. (The previous item is removed from the compiled
2604 pattern).
2605
2606 5. Changed all the names of code files so that the basic parts are no longer
2607 than 10 characters, and abolished the teeny "globals.c" file.
2608
2609 6. Changed the handling of character classes; they are now done with a 32-byte
2610 bit map always.
2611
2612 7. Added the -d and /D options to pcretest to make it possible to look at the
2613 internals of compilation without having to recompile pcre.
2614
2615
2616 Version 0.95 23-Sep-97
2617 ----------------------
2618
2619 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
2620 \x20 at the start of a run of normal characters. These were being treated as
2621 real characters, instead of the source characters being re-checked.
2622
2623
2624 Version 0.94 18-Sep-97
2625 ----------------------
2626
2627 1. The functions are now thread-safe, with the caveat that the global variables
2628 containing pointers to malloc() and free() or alternative functions are the
2629 same for all threads.
2630
2631 2. Get pcre_study() to generate a bitmap of initial characters for non-
2632 anchored patterns when this is possible, and use it if passed to pcre_exec().
2633
2634
2635 Version 0.93 15-Sep-97
2636 ----------------------
2637
2638 1. /(b)|(:+)/ was computing an incorrect first character.
2639
2640 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
2641 but not actually doing anything yet.
2642
2643 3. Treat "-" characters in classes that cannot be part of ranges as literals,
2644 as Perl does (e.g. [-az] or [az-]).
2645
2646 4. Set the anchored flag if a branch starts with .* or .*? because that tests
2647 all possible positions.
2648
2649 5. Split up into different modules to avoid including unneeded functions in a
2650 compiled binary. However, compile and exec are still in one module. The "study"
2651 function is split off.
2652
2653 6. The character tables are now in a separate module whose source is generated
2654 by an auxiliary program - but can then be edited by hand if required. There are
2655 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
2656 toupper() in the code.
2657
2658 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
2659 make them global. Abolish the function for setting them, as the caller can now
2660 set them directly.
2661
2662
2663 Version 0.92 11-Sep-97
2664 ----------------------
2665
2666 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
2667 (e.g. /a{1,3}/) was broken (I mis-optimized it).
2668
2669 2. Caseless matching was not working in character classes if the characters in
2670 the pattern were in upper case.
2671
2672 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
2673
2674 4. Make PCRE_ANCHORED public and accept as a compile option.
2675
2676 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
2677 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
2678 pass them.
2679
2680 6. Give an error if bad option bits passed at compile or run time.
2681
2682 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
2683 pcretest to cause it to pass that flag.
2684
2685 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
2686 options, and the first character, if set.
2687
2688 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
2689
2690
2691 Version 0.91 10-Sep-97
2692 ----------------------
2693
2694 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
2695 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
2696
2697 2. PCRE was looping on encountering an indefinitely repeated back reference to
2698 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
2699 Perl does - treats the match as successful.
2700
2701 ****

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