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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 141 - (show annotations) (download)
Fri Mar 30 15:46:27 2007 UTC (7 years ago) by ph10
File size: 126210 byte(s)
Fixed obscure bug when matching a null string in multiline mode with 
newline=any.

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

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