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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 914 - (show annotations)
Mon Feb 13 06:04:50 2012 UTC (3 years, 5 months ago) by zherczeg
File size: 226342 byte(s)
Error occurred while calculating annotation data.
Partial matching support is added to the JIT compiler
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.31
5 -----------------------------
6
7 1. Fixing a wrong JIT test case and some compiler warnings.
8
9 2. Removed a bashism from the RunTest script.
10
11 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
12 to unsigned type, result still unsigned" that was given by an MS compiler
13 on encountering the code "-sizeof(xxx)".
14
15 4. Partial matching support is added to the JIT compiler.
16
17
18 Version 8.30 04-February-2012
19 -----------------------------
20
21 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
22 name is defined in ctype.h.
23
24 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
25 only in quite long subpatterns.
26
27 3. Removed the function pcre_info(), which has been obsolete and deprecated
28 since it was replaced by pcre_fullinfo() in February 2000.
29
30 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
31 match a (*MARK), and the match failed at the start of the subject, a
32 reference to memory before the start of the subject could occur. This bug
33 was introduced by fix 17 of release 8.21.
34
35 5. A reference to an unset group with zero minimum repetition was giving
36 totally wrong answers (in non-JavaScript-compatibility mode). For example,
37 /(another)?(\1?)test/ matched against "hello world test". This bug was
38 introduced in release 8.13.
39
40 6. Add support for 16-bit character strings (a large amount of work involving
41 many changes and refactorings).
42
43 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
44 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
45 from a file.
46
47 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
48 rounding is not applied in this particular case).
49
50 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
51 if they appear, or are escaped, in patterns.
52
53 10. Get rid of a number of -Wunused-but-set-variable warnings.
54
55 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
56 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
57 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
58 also returns the mark "x". This bug applied to capturing parentheses,
59 non-capturing parentheses, and atomic parentheses. It also applied to some
60 assertions.
61
62 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
63 information out of configure.ac instead of relying on pcre.h.generic, which
64 is not stored in the repository.
65
66 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
67 -lreadline.
68
69 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
70
71 15. Applied Graycode's patch to put the top-level frame on the stack rather
72 than the heap when not using the stack for recursion. This gives a
73 performance improvement in many cases when recursion is not deep.
74
75 16. Experimental code added to "pcretest -C" to output the stack frame size.
76
77
78 Version 8.21 12-Dec-2011
79 ------------------------
80
81 1. Updating the JIT compiler.
82
83 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
84 are added as well.
85
86 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
87 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
88 calling _pcre_jit_exec. Some extra comments are added.
89
90 4. (*MARK) settings inside atomic groups that do not contain any capturing
91 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
92 was introduced by change 18 for 8.20.
93
94 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
95 ECMA-262 standard.
96
97 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
98 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
99 This bug was probably introduced by change 9 of 8.13.
100
101 7. While fixing 6 above, I noticed that a number of other items were being
102 incorrectly rejected as "not fixed length". This arose partly because newer
103 opcodes had not been added to the fixed-length checking code. I have (a)
104 corrected the bug and added tests for these items, and (b) arranged for an
105 error to occur if an unknown opcode is encountered while checking for fixed
106 length instead of just assuming "not fixed length". The items that were
107 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
108 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
109 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
110
111 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
112 being incorrectly compiled and would have given unpredicatble results.
113
114 9. A possessively repeated subpattern with minimum repeat count greater than
115 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
116 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
117 the first (A) could occur when it should not.
118
119 10. Add a cast and remove a redundant test from the code.
120
121 11. JIT should use pcre_malloc/pcre_free for allocation.
122
123 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
124 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
125 is anything other than /usr, -L is still shown).
126
127 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
128
129 14. Perl does not support \N without a following name in a [] class; PCRE now
130 also gives an error.
131
132 15. If a forward reference was repeated with an upper limit of around 2000,
133 it caused the error "internal error: overran compiling workspace". The
134 maximum number of forward references (including repeats) was limited by the
135 internal workspace, and dependent on the LINK_SIZE. The code has been
136 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
137 the default depends on LINK_SIZE. There is a new upper limit (for safety)
138 of around 200,000 forward references. While doing this, I also speeded up
139 the filling in of repeated forward references.
140
141 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
142 incorrectly expecting the subject to contain another "a" after the start.
143
144 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
145 in the match, the SKIP should be ignored. This was not happening; instead
146 the SKIP was being treated as NOMATCH. For patterns such as
147 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
148 tested.
149
150 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
151 now much more compatible with Perl, in particular in cases where the result
152 is a non-match for a non-anchored pattern. For example, if
153 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
154 "m", where previously it did not return a name. A side effect of this
155 change is that for partial matches, the last encountered mark name is
156 returned, as for non matches. A number of tests that were previously not
157 Perl-compatible have been moved into the Perl-compatible test files. The
158 refactoring has had the pleasing side effect of removing one argument from
159 the match() function, thus reducing its stack requirements.
160
161 19. If the /S+ option was used in pcretest to study a pattern using JIT,
162 subsequent uses of /S (without +) incorrectly behaved like /S+.
163
164 21. Retrieve executable code size support for the JIT compiler and fixing
165 some warnings.
166
167 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
168 not work when the shorter character appeared right at the end of the
169 subject string.
170
171 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
172 systems.
173
174 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
175 output it when the /M option is used in pcretest.
176
177 25. The CheckMan script was not being included in the distribution. Also, added
178 an explicit "perl" to run Perl scripts from the PrepareRelease script
179 because this is reportedly needed in Windows.
180
181 26. If study data was being save in a file and studying had not found a set of
182 "starts with" bytes for the pattern, the data written to the file (though
183 never used) was taken from uninitialized memory and so caused valgrind to
184 complain.
185
186 27. Updated RunTest.bat as provided by Sheri Pierce.
187
188 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
189
190 29. Computation of memory usage for the table of capturing group names was
191 giving an unnecessarily large value.
192
193
194 Version 8.20 21-Oct-2011
195 ------------------------
196
197 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
198 a POSIX class. After further experiments with Perl, which convinced me that
199 Perl has bugs and confusions, a closing square bracket is no longer allowed
200 in a POSIX name. This bug also affected patterns with classes that started
201 with full stops.
202
203 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
204 captured substring, but while checking the failing first alternative,
205 substring 1 is temporarily captured. If the output vector supplied to
206 pcre_exec() was not big enough for this capture, the yield of the function
207 was still zero ("insufficient space for captured substrings"). This cannot
208 be totally fixed without adding another stack variable, which seems a lot
209 of expense for a edge case. However, I have improved the situation in cases
210 such as /(a)(b)x|abc/ matched against "abc", where the return code
211 indicates that fewer than the maximum number of slots in the ovector have
212 been set.
213
214 3. Related to (2) above: when there are more back references in a pattern than
215 slots in the output vector, pcre_exec() uses temporary memory during
216 matching, and copies in the captures as far as possible afterwards. It was
217 using the entire output vector, but this conflicts with the specification
218 that only 2/3 is used for passing back captured substrings. Now it uses
219 only the first 2/3, for compatibility. This is, of course, another edge
220 case.
221
222 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
223 main code base, and can be used by building with --enable-jit. When this is
224 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
225 runtime --no-jit option is given.
226
227 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
228 ovector, the return from the function was zero, implying that there were
229 other matches that did not fit. The correct "exactly full" value is now
230 returned.
231
232 6. If a subpattern that was called recursively or as a subroutine contained
233 (*PRUNE) or any other control that caused it to give a non-standard return,
234 invalid errors such as "Error -26 (nested recursion at the same subject
235 position)" or even infinite loops could occur.
236
237 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
238 computing the minimum length on reaching *ACCEPT, and so ended up with the
239 wrong value of 1 rather than 0. Further investigation indicates that
240 computing a minimum subject length in the presence of *ACCEPT is difficult
241 (think back references, subroutine calls), and so I have changed the code
242 so that no minimum is registered for a pattern that contains *ACCEPT.
243
244 8. If (*THEN) was present in the first (true) branch of a conditional group,
245 it was not handled as intended. [But see 16 below.]
246
247 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
248 Sheri Pierce.
249
250 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
251 the first byte in a match must be "a".
252
253 11. Change 17 for 8.13 increased the recursion depth for patterns like
254 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
255 pattern contains any instances of (*THEN). If it does not, the old
256 optimizations are restored. It would be nice to do this on a per-group
257 basis, but at the moment that is not feasible.
258
259 12. In some environments, the output of pcretest -C is CRLF terminated. This
260 broke RunTest's code that checks for the link size. A single white space
261 character after the value is now allowed for.
262
263 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
264 For "fr", it uses the Windows-specific input and output files.
265
266 14. If (*THEN) appeared in a group that was called recursively or as a
267 subroutine, it did not work as intended. [But see next item.]
268
269 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
270 pattern fragments (but not containing any | characters). If A and B are
271 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
272 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
273 D. In other words, Perl considers parentheses that do not contain any |
274 characters to be part of a surrounding alternative, whereas PCRE was
275 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
276 differently. PCRE now behaves in the same way as Perl, except in the case
277 of subroutine/recursion calls such as (?1) which have in any case always
278 been different (but PCRE had them first :-).
279
280 16. Related to 15 above: Perl does not treat the | in a conditional group as
281 creating alternatives. Such a group is treated in the same way as an
282 ordinary group without any | characters when processing (*THEN). PCRE has
283 been changed to match Perl's behaviour.
284
285 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
286 RunGrepTest script failed.
287
288 18. Change 22 for version 13 caused atomic groups to use more stack. This is
289 inevitable for groups that contain captures, but it can lead to a lot of
290 stack use in large patterns. The old behaviour has been restored for atomic
291 groups that do not contain any capturing parentheses.
292
293 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
294 suppress the check for a minimum subject length at run time. (If it was
295 given to pcre_exec() or pcre_dfa_exec() it did work.)
296
297 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
298 fail to work when decoding hex characters in data strings in EBCDIC
299 environments.
300
301 21. It appears that in at least one Mac OS environment, the isxdigit() function
302 is implemented as a macro that evaluates to its argument more than once,
303 contravening the C 90 Standard (I haven't checked a later standard). There
304 was an instance in pcretest which caused it to go wrong when processing
305 \x{...} escapes in subject strings. The has been rewritten to avoid using
306 things like p++ in the argument of isxdigit().
307
308
309 Version 8.13 16-Aug-2011
310 ------------------------
311
312 1. The Unicode data tables have been updated to Unicode 6.0.0.
313
314 2. Two minor typos in pcre_internal.h have been fixed.
315
316 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
317 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
318 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
319
320 4. There were a number of related bugs in the code for matching backrefences
321 caselessly in UTF-8 mode when codes for the characters concerned were
322 different numbers of bytes. For example, U+023A and U+2C65 are an upper
323 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
324 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
325 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
326 2-byte code at the end of the subject (it thought there wasn't enough data
327 left).
328
329 5. Comprehensive information about what went wrong is now returned by
330 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
331 as the output vector has at least 2 elements. The offset of the start of
332 the failing character and a reason code are placed in the vector.
333
334 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
335 now returned is for the first byte of the failing character, instead of the
336 last byte inspected. This is an incompatible change, but I hope it is small
337 enough not to be a problem. It makes the returned offset consistent with
338 pcre_exec() and pcre_dfa_exec().
339
340 7. pcretest now gives a text phrase as well as the error number when
341 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
342 failure, the offset and reason code are output.
343
344 8. When \R was used with a maximizing quantifier it failed to skip backwards
345 over a \r\n pair if the subsequent match failed. Instead, it just skipped
346 back over a single character (\n). This seems wrong (because it treated the
347 two characters as a single entity when going forwards), conflicts with the
348 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
349 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
350 has been changed.
351
352 9. Some internal refactoring has changed the processing so that the handling
353 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
354 time (the PCRE_DOTALL option was changed this way some time ago: version
355 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
356 which was always a bit of a fudge. It also means that there is one less
357 argument for the match() function, which reduces its stack requirements
358 slightly. This change also fixes an incompatibility with Perl: the pattern
359 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
360
361 10. More internal refactoring has drastically reduced the number of recursive
362 calls to match() for possessively repeated groups such as (abc)++ when
363 using pcre_exec().
364
365 11. While implementing 10, a number of bugs in the handling of groups were
366 discovered and fixed:
367
368 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
369 (a|)*(?1) gave a compile-time internal error.
370 ((a|)+)+ did not notice that the outer group could match an empty string.
371 (^a|^)+ was not marked as anchored.
372 (.*a|.*)+ was not marked as matching at start or after a newline.
373
374 12. Yet more internal refactoring has removed another argument from the match()
375 function. Special calls to this function are now indicated by setting a
376 value in a variable in the "match data" data block.
377
378 13. Be more explicit in pcre_study() instead of relying on "default" for
379 opcodes that mean there is no starting character; this means that when new
380 ones are added and accidentally left out of pcre_study(), testing should
381 pick them up.
382
383 14. The -s option of pcretest has been documented for ages as being an old
384 synonym of -m (show memory usage). I have changed it to mean "force study
385 for every regex", that is, assume /S for every regex. This is similar to -i
386 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
387 using it. It makes it easier to run collections of tests with and without
388 study enabled, and thereby test pcre_study() more easily. All the standard
389 tests are now run with and without -s (but some patterns can be marked as
390 "never study" - see 20 below).
391
392 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
393 restoration of the capturing data to the outer values was not happening
394 correctly.
395
396 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
397 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
398 pattern had matched an empty string, and so incorrectly returned a no
399 match.
400
401 17. There was optimizing code for the last branch of non-capturing parentheses,
402 and also for the obeyed branch of a conditional subexpression, which used
403 tail recursion to cut down on stack usage. Unfortunately, now that there is
404 the possibility of (*THEN) occurring in these branches, tail recursion is
405 no longer possible because the return has to be checked for (*THEN). These
406 two optimizations have therefore been removed. [But see 8.20/11 above.]
407
408 18. If a pattern containing \R was studied, it was assumed that \R always
409 matched two bytes, thus causing the minimum subject length to be
410 incorrectly computed because \R can also match just one byte.
411
412 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
413 was incorrectly computed.
414
415 20. If /S is present twice on a test pattern in pcretest input, it now
416 *disables* studying, thereby overriding the use of -s on the command line
417 (see 14 above). This is necessary for one or two tests to keep the output
418 identical in both cases.
419
420 21. When (*ACCEPT) was used in an assertion that matched an empty string and
421 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
422
423 22. When an atomic group that contained a capturing parenthesis was
424 successfully matched, but the branch in which it appeared failed, the
425 capturing was not being forgotten if a higher numbered group was later
426 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
427 group 1 to "a", when in fact it should be unset. This applied to multi-
428 branched capturing and non-capturing groups, repeated or not, and also to
429 positive assertions (capturing in negative assertions does not happen
430 in PCRE) and also to nested atomic groups.
431
432 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
433 subject after a captured substring, to make it easier to tell which of a
434 number of identical substrings has been captured.
435
436 24. The way atomic groups are processed by pcre_exec() has been changed so that
437 if they are repeated, backtracking one repetition now resets captured
438 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
439 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
440 "aaa". Previously, it would have been "a". As part of this code
441 refactoring, the way recursive calls are handled has also been changed.
442
443 25. If an assertion condition captured any substrings, they were not passed
444 back unless some other capturing happened later. For example, if
445 (?(?=(a))a) was matched against "a", no capturing was returned.
446
447 26. When studying a pattern that contained subroutine calls or assertions,
448 the code for finding the minimum length of a possible match was handling
449 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
450 group 1 called group 2 while simultaneously a separate group 2 called group
451 1). A stack overflow occurred in this case. I have fixed this by limiting
452 the recursion depth to 10.
453
454 27. Updated RunTest.bat in the distribution to the version supplied by Tom
455 Fortmann. This supports explicit test numbers on the command line, and has
456 argument validation and error reporting.
457
458 28. An instance of \X with an unlimited repeat could fail if at any point the
459 first character it looked at was a mark character.
460
461 29. Some minor code refactoring concerning Unicode properties and scripts
462 should reduce the stack requirement of match() slightly.
463
464 30. Added the '=' option to pcretest to check the setting of unused capturing
465 slots at the end of the pattern, which are documented as being -1, but are
466 not included in the return count.
467
468 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
469 compiled something random. Now it gives a compile-time error (as does
470 Perl).
471
472 32. A *MARK encountered during the processing of a positive assertion is now
473 recorded and passed back (compatible with Perl).
474
475 33. If --only-matching or --colour was set on a pcregrep call whose pattern
476 had alternative anchored branches, the search for a second match in a line
477 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
478 matched the line "0102" twice. The same bug affected patterns that started
479 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
480 twice.
481
482 34. Previously, PCRE did not allow quantification of assertions. However, Perl
483 does, and because of capturing effects, quantifying parenthesized
484 assertions may at times be useful. Quantifiers are now allowed for
485 parenthesized assertions.
486
487 35. A minor code tidy in pcre_compile() when checking options for \R usage.
488
489 36. \g was being checked for fancy things in a character class, when it should
490 just be a literal "g".
491
492 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
493 appearance of a nested POSIX class supersedes an apparent external class.
494 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
495 unescaped square brackets may also appear as part of class names. For
496 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
497 more like Perl. (But see 8.20/1 above.)
498
499 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
500 was because it thought it was \N{name}, which is not supported).
501
502 39. Add minix to OS list not supporting the -S option in pcretest.
503
504 40. PCRE tries to detect cases of infinite recursion at compile time, but it
505 cannot analyze patterns in sufficient detail to catch mutual recursions
506 such as ((?1))((?2)). There is now a runtime test that gives an error if a
507 subgroup is called recursively as a subpattern for a second time at the
508 same position in the subject string. In previous releases this might have
509 been caught by the recursion limit, or it might have run out of stack.
510
511 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
512 happen only once. PCRE was, however incorrectly giving a compile time error
513 "recursive call could loop indefinitely" because it cannot analyze the
514 pattern in sufficient detail. The compile time test no longer happens when
515 PCRE is compiling a conditional subpattern, but actual runaway loops are
516 now caught at runtime (see 40 above).
517
518 42. It seems that Perl allows any characters other than a closing parenthesis
519 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
520 has been changed to be the same.
521
522 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
523 as not to get warnings when autogen.sh is called. Also changed
524 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
525
526 44. To help people who use pcregrep to scan files containing exceedingly long
527 lines, the following changes have been made:
528
529 (a) The default value of the buffer size parameter has been increased from
530 8K to 20K. (The actual buffer used is three times this size.)
531
532 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
533 PCRE is built.
534
535 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
536 to be set at run time.
537
538 (d) Numerical values in pcregrep options can be followed by K or M, for
539 example --buffer-size=50K.
540
541 (e) If a line being scanned overflows pcregrep's buffer, an error is now
542 given and the return code is set to 2.
543
544 45. Add a pointer to the latest mark to the callout data block.
545
546 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
547 partial match of an empty string instead of no match. This was specific to
548 the use of ".".
549
550 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
551 complete match instead of a partial match. This bug was dependent on both
552 the PCRE_UTF8 and PCRE_DOTALL options being set.
553
554 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
555 starting byte set, because \b was not being ignored.
556
557
558 Version 8.12 15-Jan-2011
559 ------------------------
560
561 1. Fixed some typos in the markup of the man pages, and wrote a script that
562 checks for such things as part of the documentation building process.
563
564 2. On a big-endian 64-bit system, pcregrep did not correctly process the
565 --match-limit and --recursion-limit options (added for 8.11). In
566 particular, this made one of the standard tests fail. (The integer value
567 went into the wrong half of a long int.)
568
569 3. If the --colour option was given to pcregrep with -v (invert match), it
570 did strange things, either producing crazy output, or crashing. It should,
571 of course, ignore a request for colour when reporting lines that do not
572 match.
573
574 4. Another pcregrep bug caused similar problems if --colour was specified with
575 -M (multiline) and the pattern match finished with a line ending.
576
577 5. In pcregrep, when a pattern that ended with a literal newline sequence was
578 matched in multiline mode, the following line was shown as part of the
579 match. This seems wrong, so I have changed it.
580
581 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
582 the check for further matches in the same line (so they could be coloured)
583 to overrun the end of the current line. If another match was found, it was
584 incorrectly shown (and then shown again when found in the next line).
585
586 7. If pcregrep was compiled under Windows, there was a reference to the
587 function pcregrep_exit() before it was defined. I am assuming this was
588 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
589 reported by a user. I've moved the definition above the reference.
590
591
592 Version 8.11 10-Dec-2010
593 ------------------------
594
595 1. (*THEN) was not working properly if there were untried alternatives prior
596 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
597 backtracked to try for "b" instead of moving to the next alternative branch
598 at the same level (in this case, to look for "c"). The Perl documentation
599 is clear that when (*THEN) is backtracked onto, it goes to the "next
600 alternative in the innermost enclosing group".
601
602 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
603 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
604 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
605 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
606 (*THEN).
607
608 3. If \s appeared in a character class, it removed the VT character from
609 the class, even if it had been included by some previous item, for example
610 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
611 of \s, but is part of the POSIX "space" class.)
612
613 4. A partial match never returns an empty string (because you can always
614 match an empty string at the end of the subject); however the checking for
615 an empty string was starting at the "start of match" point. This has been
616 changed to the "earliest inspected character" point, because the returned
617 data for a partial match starts at this character. This means that, for
618 example, /(?<=abc)def/ gives a partial match for the subject "abc"
619 (previously it gave "no match").
620
621 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
622 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
623 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
624 has an implication that the given string is incomplete (because a partial
625 match is preferred over a full match). For this reason, these items now
626 give a partial match in this situation. [Aside: previously, the one case
627 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
628 match rather than a full match, which was wrong by the old rules, but is
629 now correct.]
630
631 6. There was a bug in the handling of #-introduced comments, recognized when
632 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
633 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
634 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
635 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
636 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
637 places in pcre_compile().
638
639 7. Related to (6) above, when pcre_compile() was skipping #-introduced
640 comments when looking ahead for named forward references to subpatterns,
641 the only newline sequence it recognized was NL. It now handles newlines
642 according to the set newline convention.
643
644 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
645 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
646 cater for a lack of strerror(). These oversights have been fixed.
647
648 9. Added --match-limit and --recursion-limit to pcregrep.
649
650 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
651
652 11. When the -o option was used, pcregrep was setting a return code of 1, even
653 when matches were found, and --line-buffered was not being honoured.
654
655 12. Added an optional parentheses number to the -o and --only-matching options
656 of pcregrep.
657
658 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
659 can match an empty string. The code to do it in pcretest and pcredemo
660 needed fixing:
661
662 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
663 only one byte after an empty string match just before CRLF (this case
664 just got forgotten; "any" and "anycrlf" were OK).
665
666 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
667 mode when an empty string match preceded an ASCII character followed by
668 a non-ASCII character. (The code for advancing by one character rather
669 than one byte was nonsense.)
670
671 (c) The pcredemo.c sample program did not have any code at all to handle
672 the cases when CRLF is a valid newline sequence.
673
674 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
675 as a starting offset was within the subject string. There is now a new
676 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
677 negative or greater than the length of the string. In order to test this,
678 pcretest is extended to allow the setting of negative starting offsets.
679
680 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
681 starting offset points to the beginning of a UTF-8 character was
682 unnecessarily clumsy. I tidied it up.
683
684 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
685 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
686
687 17. Nobody had reported that the --include_dir option, which was added in
688 release 7.7 should have been called --include-dir (hyphen, not underscore)
689 for compatibility with GNU grep. I have changed it to --include-dir, but
690 left --include_dir as an undocumented synonym, and the same for
691 --exclude-dir, though that is not available in GNU grep, at least as of
692 release 2.5.4.
693
694 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
695 characters from a string of bytes) have been redefined so as not to use
696 loops, in order to improve performance in some environments. At the same
697 time, I abstracted some of the common code into auxiliary macros to save
698 repetition (this should not affect the compiled code).
699
700 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
701 compile-time error is now given if \c is not followed by an ASCII
702 character, that is, a byte less than 128. (In EBCDIC mode, the code is
703 different, and any byte value is allowed.)
704
705 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
706 START_OPTIMIZE option, which is now allowed at compile time - but just
707 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
708 to pcregrep and other applications that have no direct access to PCRE
709 options. The new /Y option in pcretest sets this option when calling
710 pcre_compile().
711
712 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
713 back references. Groups containing recursive back references were forced to
714 be atomic by that change, but in the case of named groups, the amount of
715 memory required was incorrectly computed, leading to "Failed: internal
716 error: code overflow". This has been fixed.
717
718 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
719 pcretest.c, to avoid build problems in some Borland environments.
720
721
722 Version 8.10 25-Jun-2010
723 ------------------------
724
725 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
726 THEN.
727
728 2. (*ACCEPT) was not working when inside an atomic group.
729
730 3. Inside a character class, \B is treated as a literal by default, but
731 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
732 causes the error). The code is unchanged, but I tidied the documentation.
733
734 4. Inside a character class, PCRE always treated \R and \X as literals,
735 whereas Perl faults them if its -w option is set. I have changed PCRE so
736 that it faults them when PCRE_EXTRA is set.
737
738 5. Added support for \N, which always matches any character other than
739 newline. (It is the same as "." when PCRE_DOTALL is not set.)
740
741 6. When compiling pcregrep with newer versions of gcc which may have
742 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
743 declared with attribute warn_unused_result" were given. Just casting the
744 result to (void) does not stop the warnings; a more elaborate fudge is
745 needed. I've used a macro to implement this.
746
747 7. Minor change to pcretest.c to avoid a compiler warning.
748
749 8. Added four artifical Unicode properties to help with an option to make
750 \s etc use properties (see next item). The new properties are: Xan
751 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
752
753 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
754 use Unicode properties. (*UCP) at the start of a pattern can be used to set
755 this option. Modified pcretest to add /W to test this facility. Added
756 REG_UCP to make it available via the POSIX interface.
757
758 10. Added --line-buffered to pcregrep.
759
760 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
761 studied, and the match started with a letter with a code point greater than
762 127 whose first byte was different to the first byte of the other case of
763 the letter, the other case of this starting letter was not recognized
764 (#976).
765
766 12. If a pattern that was studied started with a repeated Unicode property
767 test, for example, \p{Nd}+, there was the theoretical possibility of
768 setting up an incorrect bitmap of starting bytes, but fortunately it could
769 not have actually happened in practice until change 8 above was made (it
770 added property types that matched character-matching opcodes).
771
772 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
773 possible starting bytes for non-anchored patterns.
774
775 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
776 \R, and also a number of cases that involve Unicode properties, both
777 explicit and implicit when PCRE_UCP is set.
778
779 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
780 input, it could crash or give wrong results if characters with values
781 greater than 0xc0 were present in the subject string. (Detail: it assumed
782 UTF-8 input when processing these items.)
783
784 16. Added a lot of (int) casts to avoid compiler warnings in systems where
785 size_t is 64-bit (#991).
786
787 17. Added a check for running out of memory when PCRE is compiled with
788 --disable-stack-for-recursion (#990).
789
790 18. If the last data line in a file for pcretest does not have a newline on
791 the end, a newline was missing in the output.
792
793 19. The default pcre_chartables.c file recognizes only ASCII characters (values
794 less than 128) in its various bitmaps. However, there is a facility for
795 generating tables according to the current locale when PCRE is compiled. It
796 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
797 characters, are recognized by isspace() and therefore were getting set in
798 these tables, and indeed these tables seem to approximate to ISO 8859. This
799 caused a problem in UTF-8 mode when pcre_study() was used to create a list
800 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
801 which of course cannot start UTF-8 characters. I have changed the code so
802 that only real ASCII characters (less than 128) and the correct starting
803 bytes for UTF-8 encodings are set for characters greater than 127 when in
804 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
805 altogether.)
806
807 20. Added the /T option to pcretest so as to be able to run tests with non-
808 standard character tables, thus making it possible to include the tests
809 used for 19 above in the standard set of tests.
810
811 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
812 reference to a subpattern the other side of a comment that contains an
813 opening parenthesis caused either an internal compiling error, or a
814 reference to the wrong subpattern.
815
816
817 Version 8.02 19-Mar-2010
818 ------------------------
819
820 1. The Unicode data tables have been updated to Unicode 5.2.0.
821
822 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
823 configured.
824
825 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
826 original author of that file, following a query about its status.
827
828 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
829 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
830
831 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
832 quantifier applied to a forward-referencing subroutine call, could compile
833 incorrect code or give the error "internal error: previously-checked
834 referenced subpattern not found".
835
836 6. Both MS Visual Studio and Symbian OS have problems with initializing
837 variables to point to external functions. For these systems, therefore,
838 pcre_malloc etc. are now initialized to local functions that call the
839 relevant global functions.
840
841 7. There were two entries missing in the vectors called coptable and poptable
842 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
843 I've fixed the data, and added a kludgy way of testing at compile time that
844 the lengths are correct (equal to the number of opcodes).
845
846 8. Following on from 7, I added a similar kludge to check the length of the
847 eint vector in pcreposix.c.
848
849 9. Error texts for pcre_compile() are held as one long string to avoid too
850 much relocation at load time. To find a text, the string is searched,
851 counting zeros. There was no check for running off the end of the string,
852 which could happen if a new error number was added without updating the
853 string.
854
855 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
856
857 11. \K was not working if it appeared in an atomic group or in a group that
858 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
859 \K is "not well defined" if used in an assertion. PCRE now accepts it if
860 the assertion is positive, but not if it is negative.
861
862 12. Change 11 fortuitously reduced the size of the stack frame used in the
863 "match()" function of pcre_exec.c by one pointer. Forthcoming
864 implementation of support for (*MARK) will need an extra pointer on the
865 stack; I have reserved it now, so that the stack frame size does not
866 decrease.
867
868 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
869 item in branch that calls a recursion is a subroutine call - as in the
870 second branch in the above example - was incorrectly given the compile-
871 time error "recursive call could loop indefinitely" because pcre_compile()
872 was not correctly checking the subroutine for matching a non-empty string.
873
874 14. The checks for overrunning compiling workspace could trigger after an
875 overrun had occurred. This is a "should never occur" error, but it can be
876 triggered by pathological patterns such as hundreds of nested parentheses.
877 The checks now trigger 100 bytes before the end of the workspace.
878
879 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
880
881
882 Version 8.01 19-Jan-2010
883 ------------------------
884
885 1. If a pattern contained a conditional subpattern with only one branch (in
886 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
887 computed the wrong minimum data length (which is of course zero for such
888 subpatterns). This could cause incorrect "no match" results.
889
890 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
891 the pattern is reset in the first branch, pcre_compile() failed with
892 "internal error: code overflow at offset...". This happened only when
893 the reset was to the original external option setting. (An optimization
894 abstracts leading options settings into an external setting, which was the
895 cause of this.)
896
897 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
898 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
899 assertion pattern did not match (meaning that the assertion was true), it
900 was incorrectly treated as false if the SKIP had been reached during the
901 matching. This also applied to assertions used as conditions.
902
903 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
904 assertion subpattern, including such a pattern used as a condition,
905 unpredictable results occurred, instead of the error return
906 PCRE_ERROR_DFA_UITEM.
907
908 5. The C++ GlobalReplace function was not working like Perl for the special
909 situation when an empty string is matched. It now does the fancy magic
910 stuff that is necessary.
911
912 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
913 removed. (These were left over from very, very early versions of PCRE.)
914
915 7. Some cosmetic changes to the code to make life easier when compiling it
916 as part of something else:
917
918 (a) Change DEBUG to PCRE_DEBUG.
919
920 (b) In pcre_compile(), rename the member of the "branch_chain" structure
921 called "current" as "current_branch", to prevent a collision with the
922 Linux macro when compiled as a kernel module.
923
924 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
925 prevent a collision with the Linux macro when compiled as a kernel
926 module.
927
928 8. In pcre_compile() there are some checks for integer overflows that used to
929 cast potentially large values to (double). This has been changed to that
930 when building, a check for int64_t is made, and if it is found, it is used
931 instead, thus avoiding the use of floating point arithmetic. (There is no
932 other use of FP in PCRE.) If int64_t is not found, the fallback is to
933 double.
934
935 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
936 2005 (difference between two addresses compared to an unsigned value).
937
938 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
939 custom one, because of the following reported problem in Windows:
940
941 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
942 under Win32.
943 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
944 therefore missing the function definition.
945 - The compiler thus generates a "C" signature for the test function.
946 - The linker fails to find the "C" function.
947 - PCRE fails to configure if asked to do so against libbz2.
948
949 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
950 messages were output:
951
952 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
953 rerunning libtoolize, to keep the correct libtool macros in-tree.
954 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
955
956 I have done both of these things.
957
958 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
959 most of the time, it *can* run out if it is given a pattern that contains a
960 runaway infinite recursion. I updated the discussion in the pcrestack man
961 page.
962
963 13. Now that we have gone to the x.xx style of version numbers, the minor
964 version may start with zero. Using 08 or 09 is a bad idea because users
965 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
966 interpreted as invalid octal numbers. I've updated the previous comment in
967 configure.ac, and also added a check that gives an error if 08 or 09 are
968 used.
969
970 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
971 causing partial matching to fail when the end of the subject matched \W
972 in a UTF-8 pattern where \W was quantified with a minimum of 3.
973
974 15. There were some discrepancies between the declarations in pcre_internal.h
975 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
976 their definitions. The declarations used "const uschar *" and the
977 definitions used USPTR. Even though USPTR is normally defined as "const
978 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
979 reported that: "This difference in casting confuses some C++ compilers, for
980 example, SunCC recognizes above declarations as different functions and
981 generates broken code for hbpcre." I have changed the declarations to use
982 USPTR.
983
984 16. GNU libtool is named differently on some systems. The autogen.sh script now
985 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
986 (FreeBSD).
987
988 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
989 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
990 comment: "Figure out how to create a longlong from a string: strtoll and
991 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
992 instance, but it only takes 2 args instead of 3!"
993
994 18. A subtle bug concerned with back references has been fixed by a change of
995 specification, with a corresponding code fix. A pattern such as
996 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
997 refers, was giving matches when it shouldn't. For example, xa=xaaa would
998 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
999 same bug. Such groups have to be quantified to be useful, or contained
1000 inside another quantified group. (If there's no repetition, the reference
1001 can never match.) The problem arises because, having left the group and
1002 moved on to the rest of the pattern, a later failure that backtracks into
1003 the group uses the captured value from the final iteration of the group
1004 rather than the correct earlier one. I have fixed this in PCRE by forcing
1005 any group that contains a reference to itself to be an atomic group; that
1006 is, there cannot be any backtracking into it once it has completed. This is
1007 similar to recursive and subroutine calls.
1008
1009
1010 Version 8.00 19-Oct-09
1011 ----------------------
1012
1013 1. The table for translating pcre_compile() error codes into POSIX error codes
1014 was out-of-date, and there was no check on the pcre_compile() error code
1015 being within the table. This could lead to an OK return being given in
1016 error.
1017
1018 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1019 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1020 in a Windows environment.
1021
1022 3. The pcregrep --count option prints the count for each file even when it is
1023 zero, as does GNU grep. However, pcregrep was also printing all files when
1024 --files-with-matches was added. Now, when both options are given, it prints
1025 counts only for those files that have at least one match. (GNU grep just
1026 prints the file name in this circumstance, but including the count seems
1027 more useful - otherwise, why use --count?) Also ensured that the
1028 combination -clh just lists non-zero counts, with no names.
1029
1030 4. The long form of the pcregrep -F option was incorrectly implemented as
1031 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1032 but it seems right to fix it, and I didn't think it was worth preserving
1033 the old behaviour.
1034
1035 5. The command line items --regex=pattern and --regexp=pattern were not
1036 recognized by pcregrep, which required --regex pattern or --regexp pattern
1037 (with a space rather than an '='). The man page documented the '=' forms,
1038 which are compatible with GNU grep; these now work.
1039
1040 6. No libpcreposix.pc file was created for pkg-config; there was just
1041 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1042
1043 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1044 when UCP support is not needed, by modifying the Python script that
1045 generates it from Unicode data files. This should not matter if the module
1046 is correctly used as a library, but I received one complaint about 50K of
1047 unwanted data. My guess is that the person linked everything into his
1048 program rather than using a library. Anyway, it does no harm.
1049
1050 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1051 was a minimum greater than 1 for a wide character in a possessive
1052 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1053 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1054 character. Chaos in the form of incorrect output or a compiling loop could
1055 result.
1056
1057 9. The restrictions on what a pattern can contain when partial matching is
1058 requested for pcre_exec() have been removed. All patterns can now be
1059 partially matched by this function. In addition, if there are at least two
1060 slots in the offset vector, the offset of the earliest inspected character
1061 for the match and the offset of the end of the subject are set in them when
1062 PCRE_ERROR_PARTIAL is returned.
1063
1064 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1065 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1066 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1067 and may be more useful for multi-segment matching.
1068
1069 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1070 used to be given if ever the end of the subject was reached; now it is
1071 given only if matching could not proceed because another character was
1072 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1073 string "Z", which now yields "no match" instead of "partial match". In the
1074 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1075 final character ended with (*FAIL).
1076
1077 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1078 if the pattern had a "must contain" character that was already found in the
1079 earlier partial match, unless partial matching was again requested. For
1080 example, with the pattern /dog.(body)?/, the "must contain" character is
1081 "g". If the first part-match was for the string "dog", restarting with
1082 "sbody" failed. This bug has been fixed.
1083
1084 13. The string returned by pcre_dfa_exec() after a partial match has been
1085 changed so that it starts at the first inspected character rather than the
1086 first character of the match. This makes a difference only if the pattern
1087 starts with a lookbehind assertion or \b or \B (\K is not supported by
1088 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1089 matching functions compatible, and I think it's the right thing to do.
1090
1091 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1092 so that the demonstration program is easily available in environments where
1093 PCRE has not been installed from source.
1094
1095 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1096 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1097 library.
1098
1099 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1100 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1101 is not the first non-POSIX option to be added. Clearly some people find
1102 these options useful.
1103
1104 17. If a caller to the POSIX matching function regexec() passes a non-zero
1105 value for nmatch with a NULL value for pmatch, the value of
1106 nmatch is forced to zero.
1107
1108 18. RunGrepTest did not have a test for the availability of the -u option of
1109 the diff command, as RunTest does. It now checks in the same way as
1110 RunTest, and also checks for the -b option.
1111
1112 19. If an odd number of negated classes containing just a single character
1113 interposed, within parentheses, between a forward reference to a named
1114 subpattern and the definition of the subpattern, compilation crashed with
1115 an internal error, complaining that it could not find the referenced
1116 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1117 [The bug was that it was starting one character too far in when skipping
1118 over the character class, thus treating the ] as data rather than
1119 terminating the class. This meant it could skip too much.]
1120
1121 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1122 /g option in pcretest when the pattern contains \K, which makes it possible
1123 to have an empty string match not at the start, even when the pattern is
1124 anchored. Updated pcretest and pcredemo to use this option.
1125
1126 21. If the maximum number of capturing subpatterns in a recursion was greater
1127 than the maximum at the outer level, the higher number was returned, but
1128 with unset values at the outer level. The correct (outer level) value is
1129 now given.
1130
1131 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1132 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1133 make it do so. The string so far is captured, making this feature
1134 compatible with Perl.
1135
1136 23. The tests have been re-organized, adding tests 11 and 12, to make it
1137 possible to check the Perl 5.10 features against Perl 5.10.
1138
1139 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1140 pattern matches a fixed length string. PCRE did not allow this; now it
1141 does. Neither allows recursion.
1142
1143 25. I finally figured out how to implement a request to provide the minimum
1144 length of subject string that was needed in order to match a given pattern.
1145 (It was back references and recursion that I had previously got hung up
1146 on.) This code has now been added to pcre_study(); it finds a lower bound
1147 to the length of subject needed. It is not necessarily the greatest lower
1148 bound, but using it to avoid searching strings that are too short does give
1149 some useful speed-ups. The value is available to calling programs via
1150 pcre_fullinfo().
1151
1152 26. While implementing 25, I discovered to my embarrassment that pcretest had
1153 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1154 study optimizations had never been tested with that matching function.
1155 Oops. What is worse, even when it was passed study data, there was a bug in
1156 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1157 were also very few tests of studied patterns with pcre_dfa_exec().
1158
1159 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1160 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1161 on the other side of the coin, they are no longer allowed to have different
1162 names, because these cannot be distinguished in PCRE, and this has caused
1163 confusion. (This is a difference from Perl.)
1164
1165 28. When duplicate subpattern names are present (necessarily with different
1166 numbers, as required by 27 above), and a test is made by name in a
1167 conditional pattern, either for a subpattern having been matched, or for
1168 recursion in such a pattern, all the associated numbered subpatterns are
1169 tested, and the overall condition is true if the condition is true for any
1170 one of them. This is the way Perl works, and is also more like the way
1171 testing by number works.
1172
1173
1174 Version 7.9 11-Apr-09
1175 ---------------------
1176
1177 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1178 (pcretest), all targets were linked against these libraries. This included
1179 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1180 libraries. This caused unwanted dependencies to be created. This problem
1181 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1182 pcretest is linked with readline.
1183
1184 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1185 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1186 moved outside it again, because FALSE and TRUE are already defined in AIX,
1187 but BOOL is not.
1188
1189 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1190 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1191
1192 4. The pcregrep documentation said spaces were inserted as well as colons (or
1193 hyphens) following file names and line numbers when outputting matching
1194 lines. This is not true; no spaces are inserted. I have also clarified the
1195 wording for the --colour (or --color) option.
1196
1197 5. In pcregrep, when --colour was used with -o, the list of matching strings
1198 was not coloured; this is different to GNU grep, so I have changed it to be
1199 the same.
1200
1201 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1202 each matching line was coloured. Now it goes on to look for further matches
1203 of any of the test patterns, which is the same behaviour as GNU grep.
1204
1205 7. A pattern that could match an empty string could cause pcregrep to loop; it
1206 doesn't make sense to accept an empty string match in pcregrep, so I have
1207 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1208 seems to be how GNU grep behaves.
1209
1210 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1211 start or after a newline", because the conditional assertion was not being
1212 correctly handled. The rule now is that both the assertion and what follows
1213 in the first alternative must satisfy the test.
1214
1215 9. If auto-callout was enabled in a pattern with a conditional group whose
1216 condition was an assertion, PCRE could crash during matching, both with
1217 pcre_exec() and pcre_dfa_exec().
1218
1219 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1220 used for matching.
1221
1222 11. Unicode property support in character classes was not working for
1223 characters (bytes) greater than 127 when not in UTF-8 mode.
1224
1225 12. Added the -M command line option to pcretest.
1226
1227 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1228
1229 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1230
1231 16. Added comments and documentation about mis-use of no_arg in the C++
1232 wrapper.
1233
1234 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1235 from Martin Jerabek that uses macro names for all relevant character and
1236 string constants.
1237
1238 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1239 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1240 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1241 these, but not everybody uses configure.
1242
1243 19. A conditional group that had only one branch was not being correctly
1244 recognized as an item that could match an empty string. This meant that an
1245 enclosing group might also not be so recognized, causing infinite looping
1246 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1247 with the subject "ab", where knowledge that the repeated group can match
1248 nothing is needed in order to break the loop.
1249
1250 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1251 exec(), but without supplying a callout function, matching went wrong.
1252
1253 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1254 leak if the size of the offset vector was greater than 30. When the vector
1255 is smaller, the saved offsets during recursion go onto a local stack
1256 vector, but for larger vectors malloc() is used. It was failing to free
1257 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1258 error, in fact).
1259
1260 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1261 heapframe that is used only when UTF-8 support is enabled. This caused no
1262 problem, but was untidy.
1263
1264 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1265 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1266 included within another project.
1267
1268 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1269 slightly modified by me:
1270
1271 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1272 not building pcregrep.
1273
1274 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1275 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1276
1277 25. Forward references, both numeric and by name, in patterns that made use of
1278 duplicate group numbers, could behave incorrectly or give incorrect errors,
1279 because when scanning forward to find the reference group, PCRE was not
1280 taking into account the duplicate group numbers. A pattern such as
1281 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1282
1283 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1284 the feature of a custom pointer more persuasive (as requested by a user).
1285
1286 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1287 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1288 pre-defined.
1289
1290 28. Added support for (*UTF8) at the start of a pattern.
1291
1292 29. Arrange for flags added by the "release type" setting in CMake to be shown
1293 in the configuration summary.
1294
1295
1296 Version 7.8 05-Sep-08
1297 ---------------------
1298
1299 1. Replaced UCP searching code with optimized version as implemented for Ad
1300 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1301 stage table and inline lookup instead of a function, giving speed ups of 2
1302 to 5 times on some simple patterns that I tested. Permission was given to
1303 distribute the MultiStage2.py script that generates the tables (it's not in
1304 the tarball, but is in the Subversion repository).
1305
1306 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1307 scripts.
1308
1309 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1310 a group with a zero qualifier. The result of the study could be incorrect,
1311 or the function might crash, depending on the pattern.
1312
1313 4. Caseless matching was not working for non-ASCII characters in back
1314 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1315 It now works when Unicode Property Support is available.
1316
1317 5. In pcretest, an escape such as \x{de} in the data was always generating
1318 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1319 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1320 truncation.
1321
1322 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1323
1324 7. Added two (int) casts to pcregrep when printing the difference of two
1325 pointers, in case they are 64-bit values.
1326
1327 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1328 test 2 if it fails.
1329
1330 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1331 and a #define of that name to empty if it is not externally set. This is to
1332 allow users of MSVC to set it if necessary.
1333
1334 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1335 the convenience functions in the pcre_get.c source file.
1336
1337 11. An option change at the start of a pattern that had top-level alternatives
1338 could cause overwriting and/or a crash. This command provoked a crash in
1339 some environments:
1340
1341 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1342
1343 This potential security problem was recorded as CVE-2008-2371.
1344
1345 12. For a pattern where the match had to start at the beginning or immediately
1346 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1347 pcre_dfa_exec() could read past the end of the passed subject if there was
1348 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1349 pcretest so that it places the subject at the end of its malloc-ed buffer.
1350
1351 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1352 exec() might read past the end of the data buffer in UTF-8 mode.
1353
1354 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1355 the data contained the byte 0x85 as part of a UTF-8 character within its
1356 first line. This applied both to normal and DFA matching.
1357
1358 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1359 /^[^d]*?$/8 failed to match "abc".
1360
1361 16. Added a missing copyright notice to pcrecpp_internal.h.
1362
1363 17. Make it more clear in the documentation that values returned from
1364 pcre_exec() in ovector are byte offsets, not character counts.
1365
1366 18. Tidied a few places to stop certain compilers from issuing warnings.
1367
1368 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1369 supplied by Stefan Weber. I made a further small update for 7.8 because
1370 there is a change of source arrangements: the pcre_searchfuncs.c module is
1371 replaced by pcre_ucd.c.
1372
1373
1374 Version 7.7 07-May-08
1375 ---------------------
1376
1377 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1378 a string to a long long, pretend we don't even have a long long." This is
1379 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1380
1381 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1382 pre-7.6 versions, which defined a global no_arg variable instead of putting
1383 it in the RE class. (See also #8 below.)
1384
1385 3. Remove a line of dead code, identified by coverity and reported by Nuno
1386 Lopes.
1387
1388 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1389
1390 (1) The include/exclude patterns were being applied to the whole pathnames
1391 of files, instead of just to the final components.
1392
1393 (2) If there was more than one level of directory, the subdirectories were
1394 skipped unless they satisfied the include/exclude conditions. This is
1395 inconsistent with GNU grep (and could even be seen as contrary to the
1396 pcregrep specification - which I improved to make it absolutely clear).
1397 The action now is always to scan all levels of directory, and just
1398 apply the include/exclude patterns to regular files.
1399
1400 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1401 --exclude_dir in the tests to avoid scanning .svn directories.
1402
1403 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1404 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1405 doesn't support NULs in patterns.
1406
1407 7. Added some missing "const"s to declarations of static tables in
1408 pcre_compile.c and pcre_dfa_exec.c.
1409
1410 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1411 caused by fix #2 above. (Subsequently also a second patch to fix the
1412 first patch. And a third patch - this was a messy problem.)
1413
1414 9. Applied Craig's patch to remove the use of push_back().
1415
1416 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1417 matching function regexec().
1418
1419 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1420 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1421 references. PCRE supports relative numbers with this syntax (I don't think
1422 Oniguruma does).
1423
1424 12. Previously, a group with a zero repeat such as (...){0} was completely
1425 omitted from the compiled regex. However, this means that if the group
1426 was called as a subroutine from elsewhere in the pattern, things went wrong
1427 (an internal error was given). Such groups are now left in the compiled
1428 pattern, with a new opcode that causes them to be skipped at execution
1429 time.
1430
1431 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1432 to the way PCRE behaves:
1433
1434 (a) A lone ] character is dis-allowed (Perl treats it as data).
1435
1436 (b) A back reference to an unmatched subpattern matches an empty string
1437 (Perl fails the current match path).
1438
1439 (c) A data ] in a character class must be notated as \] because if the
1440 first data character in a class is ], it defines an empty class. (In
1441 Perl it is not possible to have an empty class.) The empty class []
1442 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1443 The negative empty class [^] matches any one character, independently
1444 of the DOTALL setting.
1445
1446 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1447 non-existent subpattern following a character class starting with ']' and
1448 containing () gave an internal compiling error instead of "reference to
1449 non-existent subpattern". Fortunately, when the pattern did exist, the
1450 compiled code was correct. (When scanning forwards to check for the
1451 existencd of the subpattern, it was treating the data ']' as terminating
1452 the class, so got the count wrong. When actually compiling, the reference
1453 was subsequently set up correctly.)
1454
1455 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1456 it was being rejected as not supported by pcre_dfa_exec(), even though
1457 other assertions are supported. I have made pcre_dfa_exec() support
1458 (*FAIL).
1459
1460 16. The implementation of 13c above involved the invention of a new opcode,
1461 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1462 cannot be changed at match time, I realized I could make a small
1463 improvement to matching performance by compiling OP_ALLANY instead of
1464 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1465 on the OP_ANY path.
1466
1467 17. Compiling pcretest on Windows with readline support failed without the
1468 following two fixes: (1) Make the unistd.h include conditional on
1469 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1470
1471 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1472 ncurses library to be included for pcretest when ReadLine support is
1473 requested, but also to allow for it to be overridden. This patch came from
1474 Daniel Bergström.
1475
1476 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1477 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1478 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1479 spotting this.
1480
1481
1482 Version 7.6 28-Jan-08
1483 ---------------------
1484
1485 1. A character class containing a very large number of characters with
1486 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1487 overflow.
1488
1489 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1490 HAVE_LONG_LONG is not defined.
1491
1492 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1493 bring them up to date and include new features. This patch includes:
1494
1495 - Fixed PH's badly added libz and libbz2 support.
1496 - Fixed a problem with static linking.
1497 - Added pcredemo. [But later removed - see 7 below.]
1498 - Fixed dftables problem and added an option.
1499 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1500 HAVE_LONG_LONG.
1501 - Added readline support for pcretest.
1502 - Added an listing of the option settings after cmake has run.
1503
1504 4. A user submitted a patch to Makefile that makes it easy to create
1505 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1506 Makefile.am that cause it to include this special target, without
1507 affecting anything else. Note that the same mingw target plus all
1508 the other distribution libraries and programs are now supported
1509 when configuring with CMake (see 6 below) instead of with
1510 Configure/Make.
1511
1512 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1513 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1514 exported in the Windows port". It has not yet been confirmed that the patch
1515 solves the problem, but it does no harm.
1516
1517 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1518 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1519 with CMake, and also correct the comment about stack recursion.
1520
1521 7. Remove the automatic building of pcredemo from the ./configure system and
1522 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1523 of a program that users should build themselves after PCRE is installed, so
1524 building it automatically is not really right. What is more, it gave
1525 trouble in some build environments.
1526
1527 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1528
1529
1530 Version 7.5 10-Jan-08
1531 ---------------------
1532
1533 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1534 values in parens when parsing an RE using the C++ wrapper."
1535
1536 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1537 Characters greater than 255 were excluded from the class instead of being
1538 included.
1539
1540 3. The same bug as (2) above applied to negated POSIX classes such as
1541 [:^space:].
1542
1543 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1544 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1545 I have changed it.
1546
1547 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1548 first named subpattern) and a construct such as (?&a) would reference the
1549 first named subpattern whose name started with "a" (in other words, the
1550 length check was missing). Both these problems are fixed. "Subpattern name
1551 expected" is now given for (?&) (a zero-length name), and this patch also
1552 makes it give the same error for \k'' (previously it complained that that
1553 was a reference to a non-existent subpattern).
1554
1555 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1556 this is right because (?- can be followed by option settings as well as by
1557 digits. I have, however, made the messages clearer.
1558
1559 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1560 than the number used in the conditional) now cause a compile-time error.
1561 This is actually not compatible with Perl, which accepts such patterns, but
1562 treats the conditional as always being FALSE (as PCRE used to), but it
1563 seems to me that giving a diagnostic is better.
1564
1565 8. Change "alphameric" to the more common word "alphanumeric" in comments
1566 and messages.
1567
1568 9. Fix two occurrences of "backslash" in comments that should have been
1569 "backspace".
1570
1571 10. Remove two redundant lines of code that can never be obeyed (their function
1572 was moved elsewhere).
1573
1574 11. The program that makes PCRE's Unicode character property table had a bug
1575 which caused it to generate incorrect table entries for sequences of
1576 characters that have the same character type, but are in different scripts.
1577 It amalgamated them into a single range, with the script of the first of
1578 them. In other words, some characters were in the wrong script. There were
1579 thirteen such cases, affecting characters in the following ranges:
1580
1581 U+002b0 - U+002c1
1582 U+0060c - U+0060d
1583 U+0061e - U+00612
1584 U+0064b - U+0065e
1585 U+0074d - U+0076d
1586 U+01800 - U+01805
1587 U+01d00 - U+01d77
1588 U+01d9b - U+01dbf
1589 U+0200b - U+0200f
1590 U+030fc - U+030fe
1591 U+03260 - U+0327f
1592 U+0fb46 - U+0fbb1
1593 U+10450 - U+1049d
1594
1595 12. The -o option (show only the matching part of a line) for pcregrep was not
1596 compatible with GNU grep in that, if there was more than one match in a
1597 line, it showed only the first of them. It now behaves in the same way as
1598 GNU grep.
1599
1600 13. If the -o and -v options were combined for pcregrep, it printed a blank
1601 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1602 does the same. The return code can be used to tell if there were any
1603 non-matching lines.
1604
1605 14. Added --file-offsets and --line-offsets to pcregrep.
1606
1607 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1608 infinitely looping recursion. The bug was that positive lookaheads were not
1609 being skipped when checking for a possible empty match (negative lookaheads
1610 and both kinds of lookbehind were skipped).
1611
1612 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1613 inclusion of <windows.h> to before rather than after the definition of
1614 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1615
1616 17. Specifying a possessive quantifier with a specific limit for a Unicode
1617 character property caused pcre_compile() to compile bad code, which led at
1618 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1619 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1620 caused the error; without that there was no problem.
1621
1622 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1623
1624 19. Added --enable-pcretest-libreadline.
1625
1626 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1627 RE::GlobalReplace(). As a result, the number of replacements returned was
1628 double what it should be. I removed one of the increments, but Craig sent a
1629 later patch that removed the other one (the right fix) and added unit tests
1630 that check the return values (which was not done before).
1631
1632 21. Several CMake things:
1633
1634 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1635 the names libpcre and libpcreposix, not just pcre and pcreposix.
1636
1637 (2) The above change means that pcretest and pcregrep are now correctly
1638 linked with the newly-built libraries, not previously installed ones.
1639
1640 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1641
1642 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1643 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1644 UTF-8 newline character). The key issue is that the pattern starts .*;
1645 this means that the match must be either at the beginning, or after a
1646 newline. The bug was in the code for advancing after a failed match and
1647 checking that the new position followed a newline. It was not taking
1648 account of UTF-8 characters correctly.
1649
1650 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1651 character classes. PCRE was not treating the sequence [:...:] as a
1652 character class unless the ... were all letters. Perl, however, seems to
1653 allow any characters between [: and :], though of course it rejects as
1654 unknown any "names" that contain non-letters, because all the known class
1655 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1656 for example, whereas PCRE did not - it did not recognize a POSIX character
1657 class. This seemed a bit dangerous, so the code has been changed to be
1658 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1659 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1660 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1661 Perl does, and where it didn't before.
1662
1663 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1664 Windows environments %n is disabled by default.
1665
1666
1667 Version 7.4 21-Sep-07
1668 ---------------------
1669
1670 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1671 means that a class such as [\s] counted as "explicit reference to CR or
1672 LF". That isn't really right - the whole point of the change was to try to
1673 help when there was an actual mention of one of the two characters. So now
1674 the change happens only if \r or \n (or a literal CR or LF) character is
1675 encountered.
1676
1677 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1678 of both had grown to the point where there were only 3 bits left.
1679 Fortunately, there was spare space in the data structure, and so I have
1680 moved the internal flags into a new 16-bit field to free up more option
1681 bits.
1682
1683 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1684 but did not set the internal JCHANGED flag - either of these is enough to
1685 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1686 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1687 start sets both bits.
1688
1689 4. Added options (at build time, compile time, exec time) to change \R from
1690 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1691
1692 5. doc/pcresyntax.html was missing from the distribution.
1693
1694 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1695 compatibility, even though it is no longer used.
1696
1697 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1698 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1699 windows.h file is present (where different names are used). [This was
1700 reversed later after testing - see 16 below.]
1701
1702 8. Changed all #include <config.h> to #include "config.h". There were also
1703 some further <pcre.h> cases that I changed to "pcre.h".
1704
1705 9. When pcregrep was used with the --colour option, it missed the line ending
1706 sequence off the lines that it output.
1707
1708 10. It was pointed out to me that arrays of string pointers cause lots of
1709 relocations when a shared library is dynamically loaded. A technique of
1710 using a single long string with a table of offsets can drastically reduce
1711 these. I have refactored PCRE in four places to do this. The result is
1712 dramatic:
1713
1714 Originally: 290
1715 After changing UCP table: 187
1716 After changing error message table: 43
1717 After changing table of "verbs" 36
1718 After changing table of Posix names 22
1719
1720 Thanks to the folks working on Gregex for glib for this insight.
1721
1722 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1723 unicode-properties was also set.
1724
1725 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1726
1727 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1728 checked only for CRLF.
1729
1730 14. Added casts to pcretest.c to avoid compiler warnings.
1731
1732 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1733
1734 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1735 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1736 entirely. This removes changes made in 7 above.
1737
1738 17. The CMake files have been updated, and there is now more information about
1739 building with CMake in the NON-UNIX-USE document.
1740
1741
1742 Version 7.3 28-Aug-07
1743 ---------------------
1744
1745 1. In the rejigging of the build system that eventually resulted in 7.1, the
1746 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1747 brackets there is not right, since it causes compilers to look for an
1748 installed pcre.h, not the version that is in the source that is being
1749 compiled (which of course may be different). I have changed it back to:
1750
1751 #include "pcre.h"
1752
1753 I have a vague recollection that the change was concerned with compiling in
1754 different directories, but in the new build system, that is taken care of
1755 by the VPATH setting the Makefile.
1756
1757 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1758 when the subject happened to end in the byte 0x85 (e.g. if the last
1759 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1760 characters but of course it shouldn't be taken as a newline when it is part
1761 of another character. The bug was that, for an unlimited repeat of . in
1762 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1763 characters when looking for a newline.
1764
1765 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1766
1767 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1768 in debug output.
1769
1770 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1771 long printing in the pcrecpp unittest when running under MinGW.
1772
1773 6. ESC_K was left out of the EBCDIC table.
1774
1775 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1776 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1777 limit also applies to "virtual nesting" when a pattern is recursive, and in
1778 this case 1000 isn't so big. I have been able to remove this limit at the
1779 expense of backing off one optimization in certain circumstances. Normally,
1780 when pcre_exec() would call its internal match() function recursively and
1781 immediately return the result unconditionally, it uses a "tail recursion"
1782 feature to save stack. However, when a subpattern that can match an empty
1783 string has an unlimited repetition quantifier, it no longer makes this
1784 optimization. That gives it a stack frame in which to save the data for
1785 checking that an empty string has been matched. Previously this was taken
1786 from the 1000-entry workspace that had been reserved. So now there is no
1787 explicit limit, but more stack is used.
1788
1789 8. Applied Daniel's patches to solve problems with the import/export magic
1790 syntax that is required for Windows, and which was going wrong for the
1791 pcreposix and pcrecpp parts of the library. These were overlooked when this
1792 problem was solved for the main library.
1793
1794 9. There were some crude static tests to avoid integer overflow when computing
1795 the size of patterns that contain repeated groups with explicit upper
1796 limits. As the maximum quantifier is 65535, the maximum group length was
1797 set at 30,000 so that the product of these two numbers did not overflow a
1798 32-bit integer. However, it turns out that people want to use groups that
1799 are longer than 30,000 bytes (though not repeat them that many times).
1800 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1801 made it possible to implement the integer overflow checks in a much more
1802 dynamic way, which I have now done. The artificial limitation on group
1803 length has been removed - we now have only the limit on the total length of
1804 the compiled pattern, which depends on the LINK_SIZE setting.
1805
1806 10. Fixed a bug in the documentation for get/copy named substring when
1807 duplicate names are permitted. If none of the named substrings are set, the
1808 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1809 empty string.
1810
1811 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1812 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1813 because the ] is interpreted as the first data character and the
1814 terminating ] is not found. PCRE has been made compatible with Perl in this
1815 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1816 cause memory overwriting.
1817
1818 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1819 string has been matched (to stop an infinite loop). It was not recognizing
1820 a conditional subpattern that could match an empty string if that
1821 subpattern was within another subpattern. For example, it looped when
1822 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1823 condition was not nested. This bug has been fixed.
1824
1825 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1826 past the start of the subject in the presence of bytes with the top bit
1827 set, for example "\x8aBCD".
1828
1829 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1830 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1831
1832 14. Optimized (?!) to (*FAIL).
1833
1834 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1835 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1836 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1837 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1838 does: it's just the validity check that is more restrictive.
1839
1840 16. Inserted checks for integer overflows during escape sequence (backslash)
1841 processing, and also fixed erroneous offset values for syntax errors during
1842 backslash processing.
1843
1844 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1845 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1846
1847 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1848 caused an overrun.
1849
1850 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1851 something other than just ASCII characters) inside a group that had an
1852 unlimited repeat caused a loop at compile time (while checking to see
1853 whether the group could match an empty string).
1854
1855 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1856 [\P{Any}] did so. (Error in the code for printing property names.)
1857
1858 21. An orphan \E inside a character class could cause a crash.
1859
1860 22. A repeated capturing bracket such as (A)? could cause a wild memory
1861 reference during compilation.
1862
1863 23. There are several functions in pcre_compile() that scan along a compiled
1864 expression for various reasons (e.g. to see if it's fixed length for look
1865 behind). There were bugs in these functions when a repeated \p or \P was
1866 present in the pattern. These operators have additional parameters compared
1867 with \d, etc, and these were not being taken into account when moving along
1868 the compiled data. Specifically:
1869
1870 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1871 length.
1872
1873 (b) An item such as \pL+ within a repeated group could cause crashes or
1874 loops.
1875
1876 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1877 "reference to non-existent subpattern" error.
1878
1879 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1880
1881 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1882 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1883
1884 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1885
1886 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1887 character were causing crashes (broken optimization).
1888
1889 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1890 \p or \P) caused a compile-time loop.
1891
1892 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1893 break. For example, the unstudied pattern [\r\n]A does not match the string
1894 "\r\nA" because change 7.0/46 below moves the current point on by two
1895 characters after failing to match at the start. However, the pattern \nA
1896 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1897 the same is true. There doesn't seem any very clean way out of this, but
1898 what I have chosen to do makes the common cases work: PCRE now takes note
1899 of whether there can be an explicit match for \r or \n anywhere in the
1900 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1901 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1902 pattern has explicit CR or LF references.
1903
1904 29. Added (*CR) etc for changing newline setting at start of pattern.
1905
1906
1907 Version 7.2 19-Jun-07
1908 ---------------------
1909
1910 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1911 which is apparently normally available under Windows.
1912
1913 2. Re-jig the pcregrep tests with different newline settings in an attempt
1914 to make them independent of the local environment's newline setting.
1915
1916 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1917
1918 4. Some of the "internals" tests were previously cut out when the link size
1919 was not 2, because the output contained actual offsets. The recent new
1920 "Z" feature of pcretest means that these can be cut out, making the tests
1921 usable with all link sizes.
1922
1923 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1924 stack recursion. This gives a massive performance boost under BSD, but just
1925 a small improvement under Linux. However, it saves one field in the frame
1926 in all cases.
1927
1928 6. Added more features from the forthcoming Perl 5.10:
1929
1930 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1931 recursion call. It refers to the nth most recently opened parentheses.
1932
1933 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1934 to be opened parentheses.
1935
1936 (c) Conditions that refer to capturing parentheses can be specified
1937 relatively, for example, (?(-2)... or (?(+3)...
1938
1939 (d) \K resets the start of the current match so that everything before
1940 is not part of it.
1941
1942 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1943
1944 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1945 reference syntax.
1946
1947 (g) (?| introduces a group in which the numbering of parentheses in each
1948 alternative starts with the same number.
1949
1950 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1951
1952 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1953 PCRE_INFO_JCHANGED.
1954
1955 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1956 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1957 for detecting groups that can match an empty string.
1958
1959 9. A pattern with a very large number of alternatives (more than several
1960 hundred) was running out of internal workspace during the pre-compile
1961 phase, where pcre_compile() figures out how much memory will be needed. A
1962 bit of new cunning has reduced the workspace needed for groups with
1963 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1964 workspace instead of running out of the 4096 that are available.
1965
1966 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1967
1968 11. Applied patch from Google to remove an optimization that didn't quite work.
1969 The report of the bug said:
1970
1971 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1972 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1973 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1974
1975 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1976 it matched the wrong number of bytes.
1977
1978
1979 Version 7.1 24-Apr-07
1980 ---------------------
1981
1982 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1983 that is more "standard", making use of automake and other Autotools. There
1984 is some re-arrangement of the files and adjustment of comments consequent
1985 on this.
1986
1987 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1988 for recursive directory scanning broke on some systems because the files
1989 are not scanned in any specific order and on different systems the order
1990 was different. A call to "sort" has been inserted into RunGrepTest for the
1991 approprate test as a short-term fix. In the longer term there may be an
1992 alternative.
1993
1994 3. I had an email from Eric Raymond about problems translating some of PCRE's
1995 man pages to HTML (despite the fact that I distribute HTML pages, some
1996 people do their own conversions for various reasons). The problems
1997 concerned the use of low-level troff macros .br and .in. I have therefore
1998 removed all such uses from the man pages (some were redundant, some could
1999 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2000 HTML has been updated to handle .nf/.fi and to complain if it encounters
2001 .br or .in.
2002
2003 4. Updated comments in configure.ac that get placed in config.h.in and also
2004 arranged for config.h to be included in the distribution, with the name
2005 config.h.generic, for the benefit of those who have to compile without
2006 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2007
2008 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2009 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2010 makevp.bat for the current PCRE, using the additional files
2011 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2012
2013 6. A Windows user reported a minor discrepancy with test 2, which turned out
2014 to be caused by a trailing space on an input line that had got lost in his
2015 copy. The trailing space was an accident, so I've just removed it.
2016
2017 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2018 that is needed.
2019
2020 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2021 as "const" (a) because they are and (b) because it helps the PHP
2022 maintainers who have recently made a script to detect big data structures
2023 in the php code that should be moved to the .rodata section. I remembered
2024 to update Builducptable as well, so it won't revert if ucptable.h is ever
2025 re-created.
2026
2027 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2028 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2029 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2030 support is not required. This saves 1.5-2K of code, which is important in
2031 some applications.
2032
2033 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2034 so as not to refer to the tables, even though these functions will never be
2035 called when UTF-8 support is disabled. Otherwise there are problems with a
2036 shared library.
2037
2038 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2039
2040 (a) It was defining its arguments as char * instead of void *.
2041
2042 (b) It was assuming that all moves were upwards in memory; this was true
2043 a long time ago when I wrote it, but is no longer the case.
2044
2045 The emulated memove() is provided for those environments that have neither
2046 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2047 is clearly not the case, as these two bugs were recently reported.
2048
2049 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2050 and Detrail to create the HTML documentation, the .txt form of the man
2051 pages, and it removes trailing spaces from listed files. It also creates
2052 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2053 case, it wraps all the #defines with #ifndefs. This script should be run
2054 before "make dist".
2055
2056 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2057 with Unicode property support.
2058
2059 (a) For a maximizing quantifier, if the two different cases of the
2060 character were of different lengths in their UTF-8 codings (there are
2061 some cases like this - I found 11), and the matching function had to
2062 back up over a mixture of the two cases, it incorrectly assumed they
2063 were both the same length.
2064
2065 (b) When PCRE was configured to use the heap rather than the stack for
2066 recursion during matching, it was not correctly preserving the data for
2067 the other case of a UTF-8 character when checking ahead for a match
2068 while processing a minimizing repeat. If the check also involved
2069 matching a wide character, but failed, corruption could cause an
2070 erroneous result when trying to check for a repeat of the original
2071 character.
2072
2073 13. Some tidying changes to the testing mechanism:
2074
2075 (a) The RunTest script now detects the internal link size and whether there
2076 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2077 values substituted by "configure". (The RunGrepTest script already did
2078 this for UTF-8.) The configure.ac script no longer substitutes the
2079 relevant variables.
2080
2081 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2082 with length and offset values. This means that the output is different
2083 for different internal link sizes. Test 2 is skipped for link sizes
2084 other than 2 because of this, bypassing the problem. Unfortunately,
2085 there was also a test in test 3 (the locale tests) that used /B and
2086 failed for link sizes other than 2. Rather than cut the whole test out,
2087 I have added a new /Z option to pcretest that replaces the length and
2088 offset values with spaces. This is now used to make test 3 independent
2089 of link size. (Test 2 will be tidied up later.)
2090
2091 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2092 segmentation fault instead of returning the appropriate error message.
2093
2094 15. In multiline mode when the newline sequence was set to "any", the pattern
2095 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2096 This doesn't seem right; it now treats the CRLF combination as the line
2097 ending, and so does not match in that case. It's only a pattern such as ^$
2098 that would hit this one: something like ^ABC$ would have failed after \r
2099 and then tried again after \r\n.
2100
2101 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2102 in an attempt to make files that differ only in their line terminators
2103 compare equal. This works on Linux.
2104
2105 17. Under certain error circumstances pcregrep might try to free random memory
2106 as it exited. This is now fixed, thanks to valgrind.
2107
2108 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2109 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2110 was because its rules for how to advance for /g after matching an empty
2111 string at the end of a line did not allow for this case. They now check for
2112 it specially.
2113
2114 20. pcretest is supposed to handle patterns and data of any length, by
2115 extending its buffers when necessary. It was getting this wrong when the
2116 buffer for a data line had to be extended.
2117
2118 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2119 CRLF as a newline sequence.
2120
2121 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2122 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2123 I have nevertheless tidied it up.
2124
2125 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2126
2127 24. Added a man page for pcre-config.
2128
2129
2130 Version 7.0 19-Dec-06
2131 ---------------------
2132
2133 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2134 moving to gcc 4.1.1.
2135
2136 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2137 sys/time.h, which is documented as needed for this function. It doesn't
2138 seem to matter on Linux, but it showed up on some releases of OS X.
2139
2140 3. It seems that there are systems where bytes whose values are greater than
2141 127 match isprint() in the "C" locale. The "C" locale should be the
2142 default when a C program starts up. In most systems, only ASCII printing
2143 characters match isprint(). This difference caused the output from pcretest
2144 to vary, making some of the tests fail. I have changed pcretest so that:
2145
2146 (a) When it is outputting text in the compiled version of a pattern, bytes
2147 other than 32-126 are always shown as hex escapes.
2148
2149 (b) When it is outputting text that is a matched part of a subject string,
2150 it does the same, unless a different locale has been set for the match
2151 (using the /L modifier). In this case, it uses isprint() to decide.
2152
2153 4. Fixed a major bug that caused incorrect computation of the amount of memory
2154 required for a compiled pattern when options that changed within the
2155 pattern affected the logic of the preliminary scan that determines the
2156 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2157 that the computed length was too small. The symptoms of this bug were
2158 either the PCRE error "internal error: code overflow" from pcre_compile(),
2159 or a glibc crash with a message such as "pcretest: free(): invalid next
2160 size (fast)". Examples of patterns that provoked this bug (shown in
2161 pcretest format) are:
2162
2163 /(?-x: )/x
2164 /(?x)(?-x: \s*#\s*)/
2165 /((?i)[\x{c0}])/8
2166 /(?i:[\x{c0}])/8
2167
2168 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2169 is now done differently.
2170
2171 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2172 wrapper classes; (b) implement a new function in the C++ scanner that is
2173 more efficient than the old way of doing things because it avoids levels of
2174 recursion in the regex matching; (c) add a paragraph to the documentation
2175 for the FullMatch() function.
2176
2177 6. The escape sequence \n was being treated as whatever was defined as
2178 "newline". Not only was this contrary to the documentation, which states
2179 that \n is character 10 (hex 0A), but it also went horribly wrong when
2180 "newline" was defined as CRLF. This has been fixed.
2181
2182 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2183 was being set to -1 for the "end of line" case (supposedly a value that no
2184 character can have). Though this value is never used (the check for end of
2185 line is "zero bytes in current character"), it caused compiler complaints.
2186 I've changed it to 0xffffffff.
2187
2188 8. In pcre_version.c, the version string was being built by a sequence of
2189 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2190 string (as it is for production releases) called a macro with an empty
2191 argument. The C standard says the result of this is undefined. The gcc
2192 compiler treats it as an empty string (which was what was wanted) but it is
2193 reported that Visual C gives an error. The source has been hacked around to
2194 avoid this problem.
2195
2196 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2197 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2198 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2199 of them did).
2200
2201 10. Originally, pcretest opened its input and output without "b"; then I was
2202 told that "b" was needed in some environments, so it was added for release
2203 5.0 to both the input and output. (It makes no difference on Unix-like
2204 systems.) Later I was told that it is wrong for the input on Windows. I've
2205 now abstracted the modes into two macros, to make it easier to fiddle with
2206 them, and removed "b" from the input mode under Windows.
2207
2208 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2209
2210 12. Added -help and --help to pcretest as an official way of being reminded
2211 of the options.
2212
2213 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2214 and pcrecpp.cc because they annoy compilers at high warning levels.
2215
2216 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2217
2218 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2219 scripts are not C programs :-) and which was not noticed because it works
2220 on Linux.
2221
2222 16. pcretest is supposed to handle any length of pattern and data line (as one
2223 line or as a continued sequence of lines) by extending its input buffer if
2224 necessary. This feature was broken for very long pattern lines, leading to
2225 a string of junk being passed to pcre_compile() if the pattern was longer
2226 than about 50K.
2227
2228 17. I have done a major re-factoring of the way pcre_compile() computes the
2229 amount of memory needed for a compiled pattern. Previously, there was code
2230 that made a preliminary scan of the pattern in order to do this. That was
2231 OK when PCRE was new, but as the facilities have expanded, it has become
2232 harder and harder to keep it in step with the real compile phase, and there
2233 have been a number of bugs (see for example, 4 above). I have now found a
2234 cunning way of running the real compile function in a "fake" mode that
2235 enables it to compute how much memory it would need, while actually only
2236 ever using a few hundred bytes of working memory and without too many
2237 tests of the mode. This should make future maintenance and development
2238 easier. A side effect of this work is that the limit of 200 on the nesting
2239 depth of parentheses has been removed (though this was never a serious
2240 limitation, I suspect). However, there is a downside: pcre_compile() now
2241 runs more slowly than before (30% or more, depending on the pattern). I
2242 hope this isn't a big issue. There is no effect on runtime performance.
2243
2244 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2245 newline (only possible for the last line of a file) and it was a
2246 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2247
2248 19. Added additional timing features to pcretest. (1) The -tm option now times
2249 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2250 separate command line item, by a number that specifies the number of
2251 repeats to use when timing. The default is 50000; this gives better
2252 precision, but takes uncomfortably long for very large patterns.
2253
2254 20. Extended pcre_study() to be more clever in cases where a branch of a
2255 subpattern has no definite first character. For example, (a*|b*)[cd] would
2256 previously give no result from pcre_study(). Now it recognizes that the
2257 first character must be a, b, c, or d.
2258
2259 21. There was an incorrect error "recursive call could loop indefinitely" if
2260 a subpattern (or the entire pattern) that was being tested for matching an
2261 empty string contained only one non-empty item after a nested subpattern.
2262 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2263 incorrectly, because the \d was being skipped in the check.
2264
2265 22. The pcretest program now has a new pattern option /B and a command line
2266 option -b, which is equivalent to adding /B to every pattern. This causes
2267 it to show the compiled bytecode, without the additional information that
2268 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2269 is the same as /B/I).
2270
2271 23. A new optimization is now able automatically to treat some sequences such
2272 as a*b as a*+b. More specifically, if something simple (such as a character
2273 or a simple class like \d) has an unlimited quantifier, and is followed by
2274 something that cannot possibly match the quantified thing, the quantifier
2275 is automatically "possessified".
2276
2277 24. A recursive reference to a subpattern whose number was greater than 39
2278 went wrong under certain circumstances in UTF-8 mode. This bug could also
2279 have affected the operation of pcre_study().
2280
2281 25. Realized that a little bit of performance could be had by replacing
2282 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2283
2284 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2285
2286 27. Possessive quantifiers such as a++ were previously implemented by turning
2287 them into atomic groups such as ($>a+). Now they have their own opcodes,
2288 which improves performance. This includes the automatically created ones
2289 from 23 above.
2290
2291 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2292 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2293 the first matched character to be a colon. This applied both to named and
2294 numbered groups.
2295
2296 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2297
2298 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2299 building PCRE on Windows easier, so I have included it in the distribution.
2300
2301 31. There is now a check in pcretest against a ridiculously large number being
2302 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2303 loop, the loop is abandoned.
2304
2305 32. Forward references to subpatterns in conditions such as (?(2)...) where
2306 subpattern 2 is defined later cause pcre_compile() to search forwards in
2307 the pattern for the relevant set of parentheses. This search went wrong
2308 when there were unescaped parentheses in a character class, parentheses
2309 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2310
2311 33. "Subroutine" calls and backreferences were previously restricted to
2312 referencing subpatterns earlier in the regex. This restriction has now
2313 been removed.
2314
2315 34. Added a number of extra features that are going to be in Perl 5.10. On the
2316 whole, these are just syntactic alternatives for features that PCRE had
2317 previously implemented using the Python syntax or my own invention. The
2318 other formats are all retained for compatibility.
2319
2320 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2321 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2322 also .NET compatible.
2323
2324 (b) A recursion or subroutine call to a named group can now be defined as
2325 (?&name) as well as (?P>name).
2326
2327 (c) A backreference to a named group can now be defined as \k<name> or
2328 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2329 5.10, are also .NET compatible.
2330
2331 (d) A conditional reference to a named group can now use the syntax
2332 (?(<name>) or (?('name') as well as (?(name).
2333
2334 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2335 groups (named and numbered) that are never evaluated inline, but can be
2336 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2337 is always false. There may be only one alternative in such a group.
2338
2339 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2340 as the simple (?(R). The condition is true only if the most recent
2341 recursion is that of the given number or name. It does not search out
2342 through the entire recursion stack.
2343
2344 (g) The escape \gN or \g{N} has been added, where N is a positive or
2345 negative number, specifying an absolute or relative reference.
2346
2347 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2348 some "unreachable code" warnings.
2349
2350 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2351 things, this adds five new scripts.
2352
2353 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2354 There were also incompatibilities regarding the handling of \Q..\E inside
2355 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2356 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2357
2358 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2359 matches an empty string, and forcibly breaks the loop. There were bugs in
2360 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2361 against aaaa the result was just "a" rather than "aaaa", for example. Two
2362 separate and independent bugs (that affected different cases) have been
2363 fixed.
2364
2365 39. Refactored the code to abolish the use of different opcodes for small
2366 capturing bracket numbers. This is a tidy that I avoided doing when I
2367 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2368 The new approach is not only tidier, it makes it possible to reduce the
2369 memory needed to fix the previous bug (38).
2370
2371 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2372 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2373 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2374 mode.
2375
2376 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2377 report.
2378
2379 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2380 copy construction and assignment in the C++ wrapper.
2381
2382 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2383 couple of bugs that could have given wrong results in the "--newline=crlf"
2384 case.
2385
2386 44. Added a number of casts and did some reorganization of signed/unsigned int
2387 variables following suggestions from Dair Grant. Also renamed the variable
2388 "this" as "item" because it is a C++ keyword.
2389
2390 45. Arranged for dftables to add
2391
2392 #include "pcre_internal.h"
2393
2394 to pcre_chartables.c because without it, gcc 4.x may remove the array
2395 definition from the final binary if PCRE is built into a static library and
2396 dead code stripping is activated.
2397
2398 46. For an unanchored pattern, if a match attempt fails at the start of a
2399 newline sequence, and the newline setting is CRLF or ANY, and the next two
2400 characters are CRLF, advance by two characters instead of one.
2401
2402
2403 Version 6.7 04-Jul-06
2404 ---------------------
2405
2406 1. In order to handle tests when input lines are enormously long, pcretest has
2407 been re-factored so that it automatically extends its buffers when
2408 necessary. The code is crude, but this _is_ just a test program. The
2409 default size has been increased from 32K to 50K.
2410
2411 2. The code in pcre_study() was using the value of the re argument before
2412 testing it for NULL. (Of course, in any sensible call of the function, it
2413 won't be NULL.)
2414
2415 3. The memmove() emulation function in pcre_internal.h, which is used on
2416 systems that lack both memmove() and bcopy() - that is, hardly ever -
2417 was missing a "static" storage class specifier.
2418
2419 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2420 containing an extended class (one that cannot be represented by a bitmap
2421 because it contains high-valued characters or Unicode property items, e.g.
2422 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2423 pattern, but PCRE should not loop if you do not (it no longer does).
2424 [Detail: two cases were found: (a) a repeated subpattern containing an
2425 extended class; (b) a recursive reference to a subpattern that followed a
2426 previous extended class. It wasn't skipping over the extended class
2427 correctly when UTF-8 mode was not set.]
2428
2429 5. A negated single-character class was not being recognized as fixed-length
2430 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2431 compile error "lookbehind assertion is not fixed length".
2432
2433 6. The RunPerlTest auxiliary script was showing an unexpected difference
2434 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2435 write a Perl script that can interpret lines of an input file either as
2436 byte characters or as UTF-8, which is what "perltest" was being required to
2437 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2438 can't do is switch easily at run time between having the "use utf8;" pragma
2439 or not. In the end, I fudged it by using the RunPerlTest script to insert
2440 "use utf8;" explicitly for the UTF-8 tests.
2441
2442 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2443 the end of the subject string, contrary to the documentation and to what
2444 Perl does. This was true of both matching functions. Now it matches only at
2445 the start of the subject and immediately after *internal* newlines.
2446
2447 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2448 a pointer to an int instead of a pointer to an unsigned long int. This
2449 caused problems on 64-bit systems.
2450
2451 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2452 instance of the 'standard' template library not being so standard".
2453
2454 10. There was no check on the number of named subpatterns nor the maximum
2455 length of a subpattern name. The product of these values is used to compute
2456 the size of the memory block for a compiled pattern. By supplying a very
2457 long subpattern name and a large number of named subpatterns, the size
2458 computation could be caused to overflow. This is now prevented by limiting
2459 the length of names to 32 characters, and the number of named subpatterns
2460 to 10,000.
2461
2462 11. Subpatterns that are repeated with specific counts have to be replicated in
2463 the compiled pattern. The size of memory for this was computed from the
2464 length of the subpattern and the repeat count. The latter is limited to
2465 65535, but there was no limit on the former, meaning that integer overflow
2466 could in principle occur. The compiled length of a repeated subpattern is
2467 now limited to 30,000 bytes in order to prevent this.
2468
2469 12. Added the optional facility to have named substrings with the same name.
2470
2471 13. Added the ability to use a named substring as a condition, using the
2472 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2473 are numbers (not recommended). Forward references are permitted.
2474
2475 14. Added forward references in named backreferences (if you see what I mean).
2476
2477 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2478 pattern could run off the end of the subject. For example, the pattern
2479 "(?s)(.{1,5})"8 did this with the subject "ab".
2480
2481 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2482 PCRE_CASELESS was set when matching characters that were quantified with ?
2483 or *.
2484
2485 17. A character class other than a single negated character that had a minimum
2486 but no maximum quantifier - for example [ab]{6,} - was not handled
2487 correctly by pce_dfa_exec(). It would match only one character.
2488
2489 18. A valid (though odd) pattern that looked like a POSIX character
2490 class but used an invalid character after [ (for example [[,abc,]]) caused
2491 pcre_compile() to give the error "Failed: internal error: code overflow" or
2492 in some cases to crash with a glibc free() error. This could even happen if
2493 the pattern terminated after [[ but there just happened to be a sequence of
2494 letters, a binary zero, and a closing ] in the memory that followed.
2495
2496 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2497 over the years. Originally (before any Unicode support), just the bottom 8
2498 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2499 output from "man perlunicode" includes this:
2500
2501 The regular expression compiler produces polymorphic opcodes. That
2502 is, the pattern adapts to the data and automatically switches to
2503 the Unicode character scheme when presented with Unicode data--or
2504 instead uses a traditional byte scheme when presented with byte
2505 data.
2506
2507 Sadly, a wide octal escape does not cause a switch, and in a string with
2508 no other multibyte characters, these octal escapes are treated as before.
2509 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2510 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2511 Unicode string.
2512
2513 I have not perpetrated such confusion in PCRE. Up till now, it took just
2514 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2515 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2516 translate to the appropriate multibyte character.
2517
2518 29. Applied some refactoring to reduce the number of warnings from Microsoft
2519 and Borland compilers. This has included removing the fudge introduced
2520 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2521 a warning about an unused variable.
2522
2523 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2524 characters since release 4.0, because Perl (from release 5.004) does not.
2525 [Or at least, is documented not to: some releases seem to be in conflict
2526 with the documentation.] However, when a pattern was studied with
2527 pcre_study() and all its branches started with \s, PCRE still included VT
2528 as a possible starting character. Of course, this did no harm; it just
2529 caused an unnecessary match attempt.
2530
2531 22. Removed a now-redundant internal flag bit that recorded the fact that case
2532 dependency changed within the pattern. This was once needed for "required
2533 byte" processing, but is no longer used. This recovers a now-scarce options
2534 bit. Also moved the least significant internal flag bit to the most-
2535 significant bit of the word, which was not previously used (hangover from
2536 the days when it was an int rather than a uint) to free up another bit for
2537 the future.
2538
2539 23. Added support for CRLF line endings as well as CR and LF. As well as the
2540 default being selectable at build time, it can now be changed at runtime
2541 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2542 specify that it is scanning data with non-default line endings.
2543
2544 24. Changed the definition of CXXLINK to make it agree with the definition of
2545 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2546
2547 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2548 recursions. This makes a big different to stack usage for some patterns.
2549
2550 26. If a subpattern containing a named recursion or subroutine reference such
2551 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2552 the space required for the compiled pattern went wrong and gave too small a
2553 value. Depending on the environment, this could lead to "Failed: internal
2554 error: code overflow at offset 49" or "glibc detected double free or
2555 corruption" errors.
2556
2557 27. Applied patches from Google (a) to support the new newline modes and (b) to
2558 advance over multibyte UTF-8 characters in GlobalReplace.
2559
2560 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2561 difference for some implementation of PCRE in some Windows version.
2562
2563 29. Added some extra testing facilities to pcretest:
2564
2565 \q<number> in a data line sets the "match limit" value
2566 \Q<number> in a data line sets the "match recursion limt" value
2567 -S <number> sets the stack size, where <number> is in megabytes
2568
2569 The -S option isn't available for Windows.
2570
2571
2572 Version 6.6 06-Feb-06
2573 ---------------------
2574
2575 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2576 in pcreposix.h. I have copied the definition from pcre.h.
2577
2578 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2579 because pcre.h is no longer a built file.
2580
2581 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2582 not normally included in the compiled code.
2583
2584
2585 Version 6.5 01-Feb-06
2586 ---------------------
2587
2588 1. When using the partial match feature with pcre_dfa_exec(), it was not
2589 anchoring the second and subsequent partial matches at the new starting
2590 point. This could lead to incorrect results. For example, with the pattern
2591 /1234/, partially matching against "123" and then "a4" gave a match.
2592
2593 2. Changes to pcregrep:
2594
2595 (a) All non-match returns from pcre_exec() were being treated as failures
2596 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2597 error message is output. Some extra information is given for the
2598 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2599 probably the only errors that are likely to be caused by users (by
2600 specifying a regex that has nested indefinite repeats, for instance).
2601 If there are more than 20 of these errors, pcregrep is abandoned.
2602
2603 (b) A binary zero was treated as data while matching, but terminated the
2604 output line if it was written out. This has been fixed: binary zeroes
2605 are now no different to any other data bytes.
2606
2607 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2608 used to set a locale for matching. The --locale=xxxx long option has
2609 been added (no short equivalent) to specify a locale explicitly on the
2610 pcregrep command, overriding the environment variables.
2611
2612 (d) When -B was used with -n, some line numbers in the output were one less
2613 than they should have been.
2614
2615 (e) Added the -o (--only-matching) option.
2616
2617 (f) If -A or -C was used with -c (count only), some lines of context were
2618 accidentally printed for the final match.
2619
2620 (g) Added the -H (--with-filename) option.
2621
2622 (h) The combination of options -rh failed to suppress file names for files
2623 that were found from directory arguments.
2624
2625 (i) Added the -D (--devices) and -d (--directories) options.
2626
2627 (j) Added the -F (--fixed-strings) option.
2628
2629 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2630
2631 (l) Added the --colo(u)r option.
2632
2633 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2634 is not present by default.
2635
2636 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2637 items such as (?R) or (?1), when the recursion could match a number of
2638 alternatives. If it matched one of the alternatives, but subsequently,
2639 outside the recursion, there was a failure, the code tried to back up into
2640 the recursion. However, because of the way PCRE is implemented, this is not
2641 possible, and the result was an incorrect result from the match.
2642
2643 In order to prevent this happening, the specification of recursion has
2644 been changed so that all such subpatterns are automatically treated as
2645 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2646
2647 4. I had overlooked the fact that, in some locales, there are characters for
2648 which isalpha() is true but neither isupper() nor islower() are true. In
2649 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2650 and ordfeminine) are like this. This affected the treatment of \w and \W
2651 when they appeared in character classes, but not when they appeared outside
2652 a character class. The bit map for "word" characters is now created
2653 separately from the results of isalnum() instead of just taking it from the
2654 upper, lower, and digit maps. (Plus the underscore character, of course.)
2655
2656 5. The above bug also affected the handling of POSIX character classes such as
2657 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2658 permanent tables. Instead, the bit maps for such a class were previously
2659 created as the appropriate unions of the upper, lower, and digit bitmaps.
2660 Now they are created by subtraction from the [[:word:]] class, which has
2661 its own bitmap.
2662
2663 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2664 It is created by subtracting the vertical space characters (\x09, \x0a,
2665 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2666 subtraction was done in the overall bitmap for a character class, meaning
2667 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2668 be recognized. This bug has been fixed.
2669
2670 7. Patches from the folks at Google:
2671
2672 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2673 real life, but is still worth protecting against".
2674
2675 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2676 regular expressions".
2677
2678 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2679 have it.
2680
2681 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2682 "configure" and the latter not, in order to fix a problem somebody had
2683 with compiling the Arg class on HP-UX.
2684
2685 (e) Improve the error-handling of the C++ wrapper a little bit.
2686
2687 (f) New tests for checking recursion limiting.
2688
2689 8. The pcre_memmove() function, which is used only if the environment does not
2690 have a standard memmove() function (and is therefore rarely compiled),
2691 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2692 returning a result (though PCRE never actually uses the result).
2693
2694 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2695 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2696 returned instead of calling malloc() with an overflowing number that would
2697 most likely cause subsequent chaos.
2698
2699 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2700
2701 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2702 with this option is matched, the nmatch and pmatch options of regexec() are
2703 ignored.
2704
2705 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2706 provided in case anyone wants to the the POSIX interface with UTF-8
2707 strings.
2708
2709 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2710 C++ linking (needed for some HP-UX environments).
2711
2712 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2713 (unused parameter) and in the pcre_printint() function (omitted "default"
2714 switch label when the default is to do nothing).
2715
2716 15. Added some code to make it possible, when PCRE is compiled as a C++
2717 library, to replace subject pointers for pcre_exec() with a smart pointer
2718 class, thus making it possible to process discontinuous strings.
2719
2720 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2721 much the same function. They were added by different people who were trying
2722 to make PCRE easy to compile on non-Unix systems. It has been suggested
2723 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2724 for compiling on Windows systems. I have therefore replaced it with
2725 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2726 defaults to "extern" for C or "extern C" for C++, which works fine on
2727 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2728 SCOPE with something explicit in config.h. In addition:
2729
2730 (a) pcreposix.h still had just "extern" instead of either of these macros;
2731 I have replaced it with PCRE_DATA_SCOPE.
2732
2733 (b) Functions such as _pcre_xclass(), which are internal to the library,
2734 but external in the C sense, all had PCRE_EXPORT in their definitions.
2735 This is apparently wrong for the Windows case, so I have removed it.
2736 (It makes no difference on Unix-like systems.)
2737
2738 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2739 of recursive calls to match(). This is different to MATCH_LIMIT because
2740 that limits the total number of calls to match(), not all of which increase
2741 the depth of recursion. Limiting the recursion depth limits the amount of
2742 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2743 when PCRE is compiled, and changed at run time. A patch from Google adds
2744 this functionality to the C++ interface.
2745
2746 18. Changes to the handling of Unicode character properties:
2747
2748 (a) Updated the table to Unicode 4.1.0.
2749
2750 (b) Recognize characters that are not in the table as "Cn" (undefined).
2751
2752 (c) I revised the way the table is implemented to a much improved format
2753 which includes recognition of ranges. It now supports the ranges that
2754 are defined in UnicodeData.txt, and it also amalgamates other
2755 characters into ranges. This has reduced the number of entries in the
2756 table from around 16,000 to around 3,000, thus reducing its size
2757 considerably. I realized I did not need to use a tree structure after
2758 all - a binary chop search is just as efficient. Having reduced the
2759 number of entries, I extended their size from 6 bytes to 8 bytes to
2760 allow for more data.
2761
2762 (d) Added support for Unicode script names via properties such as \p{Han}.
2763
2764 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2765 matching that character.
2766
2767 20. When matching a repeated Unicode property with a minimum greater than zero,
2768 (for example \pL{2,}), PCRE could look past the end of the subject if it
2769 reached it while seeking the minimum number of characters. This could
2770 happen only if some of the characters were more than one byte long, because
2771 there is a check for at least the minimum number of bytes.
2772
2773 21. Refactored the implementation of \p and \P so as to be more general, to
2774 allow for more different types of property in future. This has changed the
2775 compiled form incompatibly. Anybody with saved compiled patterns that use
2776 \p or \P will have to recompile them.
2777
2778 22. Added "Any" and "L&" to the supported property types.
2779
2780 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2781 but give a compile time error if the value is greater than 0xff.
2782
2783 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2784 accidentally not being installed or uninstalled.
2785
2786 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2787 made were to insert the current release number. This seemed silly, because
2788 it made things harder for people building PCRE on systems that don't run
2789 "configure". I have turned pcre.h into a distributed file, no longer built
2790 by "configure", with the version identification directly included. There is
2791 no longer a pcre.h.in file.
2792
2793 However, this change necessitated a change to the pcre-config script as
2794 well. It is built from pcre-config.in, and one of the substitutions was the
2795 release number. I have updated configure.ac so that ./configure now finds
2796 the release number by grepping pcre.h.
2797
2798 26. Added the ability to run the tests under valgrind.
2799
2800
2801 Version 6.4 05-Sep-05
2802 ---------------------
2803
2804 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2805 "--" to be printed when multiple files were scanned, even when none of the
2806 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2807 consider it to be a bug, and have restored the previous behaviour.
2808
2809 2. A couple of code tidies to get rid of compiler warnings.
2810
2811 3. The pcretest program used to cheat by referring to symbols in the library
2812 whose names begin with _pcre_. These are internal symbols that are not
2813 really supposed to be visible externally, and in some environments it is
2814 possible to suppress them. The cheating is now confined to including
2815 certain files from the library's source, which is a bit cleaner.
2816
2817 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2818 file's purpose clearer.
2819
2820 5. Reorganized pcre_ucp_findchar().
2821
2822
2823 Version 6.3 15-Aug-05
2824 ---------------------
2825
2826 1. The file libpcre.pc.in did not have general read permission in the tarball.
2827
2828 2. There were some problems when building without C++ support:
2829
2830 (a) If C++ support was not built, "make install" and "make test" still
2831 tried to test it.
2832
2833 (b) There were problems when the value of CXX was explicitly set. Some
2834 changes have been made to try to fix these, and ...
2835
2836 (c) --disable-cpp can now be used to explicitly disable C++ support.
2837
2838 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2839 backslash in a target when C++ was disabled. This confuses some
2840 versions of "make", apparently. Using an intermediate variable solves
2841 this. (Same for CPP_LOBJ.)
2842
2843 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2844 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2845 necessary on certain architectures.
2846
2847 4. Added a setting of -export-symbols-regex to the link command to remove
2848 those symbols that are exported in the C sense, but actually are local
2849 within the library, and not documented. Their names all begin with
2850 "_pcre_". This is not a perfect job, because (a) we have to except some
2851 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2852 available (and never for static libraries). I have made a note to try to
2853 find a way round (a) in the future.
2854
2855
2856 Version 6.2 01-Aug-05
2857 ---------------------
2858
2859 1. There was no test for integer overflow of quantifier values. A construction
2860 such as {1111111111111111} would give undefined results. What is worse, if
2861 a minimum quantifier for a parenthesized subpattern overflowed and became
2862 negative, the calculation of the memory size went wrong. This could have
2863 led to memory overwriting.
2864
2865 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2866
2867 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2868 operating environments where this matters.
2869
2870 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2871 PCRE options from within the C++ wrapper.
2872
2873 5. Named capturing subpatterns were not being correctly counted when a pattern
2874 was compiled. This caused two problems: (a) If there were more than 100
2875 such subpatterns, the calculation of the memory needed for the whole
2876 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2877 back references of the form \12, where the number was greater than 9, were
2878 not recognized as back references, even though there were sufficient
2879 previous subpatterns.
2880
2881 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2882 versions of gcc, e.g. 2.95.4.
2883
2884
2885 Version 6.1 21-Jun-05
2886 ---------------------
2887
2888 1. There was one reference to the variable "posix" in pcretest.c that was not
2889 surrounded by "#if !defined NOPOSIX".
2890
2891 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2892 the cross-check on the old pcre_info() function, for the benefit of the
2893 cut-down version of PCRE that is currently imported into Exim.
2894
2895 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2896 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2897 patterns that start (?i) but I don't think that matters. The use of (?i) is
2898 just an example; this all applies to the other options as well.
2899
2900 4. Since libtool seems to echo the compile commands it is issuing, the output
2901 from "make" can be reduced a bit by putting "@" in front of each libtool
2902 compile command.
2903
2904 5. Patch from the folks at Google for configure.in to be a bit more thorough
2905 in checking for a suitable C++ installation before trying to compile the
2906 C++ stuff. This should fix a reported problem when a compiler was present,
2907 but no suitable headers.
2908
2909 6. The man pages all had just "PCRE" as their title. I have changed them to
2910 be the relevant file name. I have also arranged that these names are
2911 retained in the file doc/pcre.txt, which is a concatenation in text format
2912 of all the man pages except the little individual ones for each function.
2913
2914 7. The NON-UNIX-USE file had not been updated for the different set of source
2915 files that come with release 6. I also added a few comments about the C++
2916 wrapper.
2917
2918
2919 Version 6.0 07-Jun-05
2920 ---------------------
2921
2922 1. Some minor internal re-organization to help with my DFA experiments.
2923
2924 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2925 didn't matter for the library itself when fully configured, but did matter
2926 when compiling without UCP support, or within Exim, where the ucp files are
2927 not imported.
2928
2929 3. Refactoring of the library code to split up the various functions into
2930 different source modules. The addition of the new DFA matching code (see
2931 below) to a single monolithic source would have made it really too
2932 unwieldy, quite apart from causing all the code to be include in a
2933 statically linked application, when only some functions are used. This is
2934 relevant even without the DFA addition now that patterns can be compiled in
2935 one application and matched in another.
2936
2937 The downside of splitting up is that there have to be some external
2938 functions and data tables that are used internally in different modules of
2939 the library but which are not part of the API. These have all had their
2940 names changed to start with "_pcre_" so that they are unlikely to clash
2941 with other external names.
2942
2943 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2944 a different (DFA) algorithm. Although it is slower than the original
2945 function, it does have some advantages for certain types of matching
2946 problem.
2947
2948 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2949 including restarting after a partial match.
2950
2951 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2952 defined when compiling for Windows was sent to me. I have put it into the
2953 code, though I have no means of testing or verifying it.
2954
2955 7. Added the pcre_refcount() auxiliary function.
2956
2957 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2958 match before or at the first newline in the subject string. In pcretest,
2959 the /f option on a pattern can be used to set this.
2960
2961 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2962 would behave wrongly. This has been present in PCRE since release 4.0.
2963
2964 10. A number of changes to the pcregrep command:
2965
2966 (a) Refactored how -x works; insert ^(...)$ instead of setting
2967 PCRE_ANCHORED and checking the length, in preparation for adding
2968 something similar for -w.
2969
2970 (b) Added the -w (match as a word) option.
2971
2972 (c) Refactored the way lines are read and buffered so as to have more
2973 than one at a time available.
2974
2975 (d) Implemented a pcregrep test script.
2976
2977 (e) Added the -M (multiline match) option. This allows patterns to match
2978 over several lines of the subject. The buffering ensures that at least
2979 8K, or the rest of the document (whichever is the shorter) is available
2980 for matching (and similarly the previous 8K for lookbehind assertions).
2981
2982 (f) Changed the --help output so that it now says
2983
2984 -w, --word-regex(p)
2985
2986 instead of two lines, one with "regex" and the other with "regexp"
2987 because that confused at least one person since the short forms are the
2988 same. (This required a bit of code, as the output is generated
2989 automatically from a table. It wasn't just a text change.)
2990
2991 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2992 option but starts with a hyphen. Could be a pattern or a path name
2993 starting with a hyphen, for instance.
2994
2995 (h) "-" can be given as a file name to represent stdin.
2996
2997 (i) When file names are being printed, "(standard input)" is used for
2998 the standard input, for compatibility with GNU grep. Previously
2999 "<stdin>" was used.
3000
3001 (j) The option --label=xxx can be used to supply a name to be used for
3002 stdin when file names are being printed. There is no short form.
3003
3004 (k) Re-factored the options decoding logic because we are going to add
3005 two more options that take data. Such options can now be given in four
3006 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3007
3008 (l) Added the -A, -B, and -C options for requesting that lines of context
3009 around matches be printed.
3010
3011 (m) Added the -L option to print the names of files that do not contain
3012 any matching lines, that is, the complement of -l.
3013
3014 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3015 continue to scan other files.
3016
3017 (o) The -s option was incorrectly implemented. For compatibility with other
3018 greps, it now suppresses the error message for a non-existent or non-
3019 accessible file (but not the return code). There is a new option called
3020 -q that suppresses the output of matching lines, which was what -s was
3021 previously doing.
3022
3023 (p) Added --include and --exclude options to specify files for inclusion
3024 and exclusion when recursing.
3025
3026 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3027 Hopefully, it now does.
3028
3029 12. Missing cast in pcre_study().
3030
3031 13. Added an "uninstall" target to the makefile.
3032
3033 14. Replaced "extern" in the function prototypes in Makefile.in with
3034 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3035 world, but is set differently for Windows.
3036
3037 15. Added a second compiling function called pcre_compile2(). The only
3038 difference is that it has an extra argument, which is a pointer to an
3039 integer error code. When there is a compile-time failure, this is set
3040 non-zero, in addition to the error test pointer being set to point to an
3041 error message. The new argument may be NULL if no error number is required
3042 (but then you may as well call pcre_compile(), which is now just a
3043 wrapper). This facility is provided because some applications need a
3044 numeric error indication, but it has also enabled me to tidy up the way
3045 compile-time errors are handled in the POSIX wrapper.
3046
3047 16. Added VPATH=.libs to the makefile; this should help when building with one
3048 prefix path and installing with another. (Or so I'm told by someone who
3049 knows more about this stuff than I do.)
3050
3051 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3052 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3053 match everything, including newlines. This is not POSIX-compatible, but
3054 somebody wanted the feature. From pcretest it can be activated by using
3055 both the P and the s flags.
3056
3057 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3058
3059 19. libpcre.pc was being incorrectly installed as executable.
3060
3061 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3062 it now also looks for '\r' so that it will work unmodified on Windows.
3063
3064 21. Added Google's contributed C++ wrapper to the distribution.
3065
3066 22. Added some untidy missing memory free() calls in pcretest, to keep
3067 Electric Fence happy when testing.
3068
3069
3070
3071 Version 5.0 13-Sep-04
3072 ---------------------
3073
3074 1. Internal change: literal characters are no longer packed up into items
3075 containing multiple characters in a single byte-string. Each character
3076 is now matched using a separate opcode. However, there may be more than one
3077 byte in the character in UTF-8 mode.
3078
3079 2. The pcre_callout_block structure has two new fields: pattern_position and
3080 next_item_length. These contain the offset in the pattern to the next match
3081 item, and its length, respectively.
3082
3083 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3084 insertion of callouts before each pattern item. Added the /C option to
3085 pcretest to make use of this.
3086
3087 4. On the advice of a Windows user, the lines
3088
3089 #if defined(_WIN32) || defined(WIN32)
3090 _setmode( _fileno( stdout ), 0x8000 );
3091 #endif /* defined(_WIN32) || defined(WIN32) */
3092
3093 have been added to the source of pcretest. This apparently does useful
3094 magic in relation to line terminators.
3095
3096 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3097 for the benefit of those environments where the "b" makes a difference.
3098
3099 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3100 to know about it. I have put a hack into configure.in that adds in code
3101 to set GCC=yes if CC=icc. This seems to end up at a point in the
3102 generated configure script that is early enough to affect the setting of
3103 compiler options, which is what is needed, but I have no means of testing
3104 whether it really works. (The user who reported this had patched the
3105 generated configure script, which of course I cannot do.)
3106
3107 LATER: After change 22 below (new libtool files), the configure script
3108 seems to know about icc (and also ecc). Therefore, I have commented out
3109 this hack in configure.in.
3110
3111 7. Added support for pkg-config (2 patches were sent in).
3112
3113 8. Negated POSIX character classes that used a combination of internal tables
3114 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3115 [[:^ascii]]. Typically, they would match almost any characters. The other
3116 POSIX classes were not broken in this way.
3117
3118 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3119 to find the match, as PCRE was deluded into thinking that the match had to
3120 start at the start point or following a newline. The same bug applied to
3121 patterns with negative forward assertions or any backward assertions
3122 preceding ".*" at the start, unless the pattern required a fixed first
3123 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3124
3125 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3126 starting at the last subject character, bytes beyond the end of the subject
3127 string were read.
3128
3129 11. Renamed the variable "class" as "classbits" to make life easier for C++
3130 users. (Previously there was a macro definition, but it apparently wasn't
3131 enough.)
3132
3133 12. Added the new field "tables" to the extra data so that tables can be passed
3134 in at exec time, or the internal tables can be re-selected. This allows
3135 a compiled regex to be saved and re-used at a later time by a different
3136 program that might have everything at different addresses.
3137
3138 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3139 -R library as well as a -L library.
3140
3141 14. The debugging options of pcretest (-d on the command line or D on a
3142 pattern) showed incorrect output for anything following an extended class
3143 that contained multibyte characters and which was followed by a quantifier.
3144
3145 15. Added optional support for general category Unicode character properties
3146 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3147 support. It adds about 90K to the size of the library. The meanings of the
3148 inbuilt class escapes such as \d and \s have NOT been changed.
3149
3150 16. Updated pcredemo.c to include calls to free() to release the memory for the
3151 compiled pattern.
3152
3153 17. The generated file chartables.c was being created in the source directory
3154 instead of in the building directory. This caused the build to fail if the
3155 source directory was different from the building directory, and was
3156 read-only.
3157
3158 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3159 file. No doubt somebody will tell me if they don't make sense... Also added
3160 Dan Mooney's comments about building on OpenVMS.
3161
3162 19. Added support for partial matching via the PCRE_PARTIAL option for
3163 pcre_exec() and the \P data escape in pcretest.
3164
3165 20. Extended pcretest with 3 new pattern features:
3166
3167 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3168 write the compiled pattern to the file whose name is "rest-of-line".
3169 This is a straight binary dump of the data, with the saved pointer to
3170 the character tables forced to be NULL. The study data, if any, is
3171 written too. After writing, pcretest reads a new pattern.
3172
3173 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3174 compiled pattern from the given file. There must not be any
3175 occurrences of "<" in the file name (pretty unlikely); if there are,
3176 pcretest will instead treat the initial "<" as a pattern delimiter.
3177 After reading in the pattern, pcretest goes on to read data lines as
3178 usual.
3179
3180 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3181 and 16-bit fields in a compiled pattern, to simulate a pattern that
3182 was compiled on a host of opposite endianness.
3183
3184 21. The pcre-exec() function can now cope with patterns that were compiled on
3185 hosts of opposite endianness, with this restriction:
3186
3187 As for any compiled expression that is saved and used later, the tables
3188 pointer field cannot be preserved; the extra_data field in the arguments
3189 to pcre_exec() should be used to pass in a tables address if a value
3190 other than the default internal tables were used at compile time.
3191
3192 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3193 now diagnosed as an error. Previously, most of the time, a negative number
3194 would have been treated as zero, but if in addition "ovector" was passed as
3195 NULL, a crash could occur.
3196
3197 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3198 new versions from the libtool 1.5 distribution (the last one is a copy of
3199 a file called libtool.m4). This seems to have fixed the need to patch
3200 "configure" to support Darwin 1.3 (which I used to do). However, I still
3201 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3202 workstation).
3203
3204 24. Changed the PCRE licence to be the more standard "BSD" licence.
3205
3206
3207 Version 4.5 01-Dec-03
3208 ---------------------
3209
3210 1. There has been some re-arrangement of the code for the match() function so
3211 that it can be compiled in a version that does not call itself recursively.
3212 Instead, it keeps those local variables that need separate instances for
3213 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3214 needs to "recurse". Keeping track of where control must go is done by means
3215 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3216 hide most of the details from the main code, and operates only if
3217 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3218 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3219 operating.
3220
3221 To make it easier for callers to provide specially tailored get/free
3222 functions for this usage, two new functions, pcre_stack_malloc, and
3223 pcre_stack_free, are used. They are always called in strict stacking order,
3224 and the size of block requested is always the same.
3225
3226 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3227 PCRE has been compiled to use the stack or the heap for recursion. The
3228 -C option of pcretest uses this to show which version is compiled.
3229
3230 A new data escape \S, is added to pcretest; it causes the amounts of store
3231 obtained and freed by both kinds of malloc/free at match time to be added
3232 to the output.
3233
3234 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3235 what's available on my current Linux desktop machine.
3236
3237 3. When matching a UTF-8 string, the test for a valid string at the start has
3238 been extended. If start_offset is not zero, PCRE now checks that it points
3239 to a byte that is the start of a UTF-8 character. If not, it returns
3240 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3241 this is necessary because there may be backward assertions in the pattern.
3242 When matching the same subject several times, it may save resources to use
3243 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3244
3245 4. The code for checking the validity of UTF-8 strings has been tightened so
3246 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3247 containing "overlong sequences".
3248
3249 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3250 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3251 should have been "&", but it just so happened that all the cases this let
3252 through by mistake were picked up later in the function.
3253
3254 6. I had used a variable called "isblank" - this is a C99 function, causing
3255 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3256
3257 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3258 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3259 so the version is shown; (c) stop "make test" echoing "./RunTest".
3260
3261 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3262
3263 9. The prototype for memmove() for systems that don't have it was using
3264 size_t, but the inclusion of the header that defines size_t was later. I've
3265 moved the #includes for the C headers earlier to avoid this.
3266
3267 10. Added some adjustments to the code to make it easier to compiler on certain
3268 special systems:
3269
3270 (a) Some "const" qualifiers were missing.
3271 (b) Added the macro EXPORT before all exported functions; by default this
3272 is defined to be empty.
3273 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3274 that it reads its output file name as an argument instead of writing
3275 to the standard output and assuming this can be redirected.
3276
3277 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3278 class containing characters with values greater than 255, PCRE compilation
3279 went into a loop.
3280
3281 12. A recursive reference to a subpattern that was within another subpattern
3282 that had a minimum quantifier of zero caused PCRE to crash. For example,
3283 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3284 recursion. If the recursively-called subpattern itself had a zero repeat,
3285 that was OK.
3286
3287 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3288 buffer into which it was copied (for escape processing) was still set at
3289 1024, so long lines caused crashes.
3290
3291 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3292 "internal error: code overflow...". This applied to any character class
3293 that was followed by a possessive quantifier.
3294
3295 15. Modified the Makefile to add libpcre.la as a prerequisite for
3296 libpcreposix.la because I was told this is needed for a parallel build to
3297 work.
3298
3299 16. If a pattern that contained .* following optional items at the start was
3300 studied, the wrong optimizing data was generated, leading to matching
3301 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3302 matching string must start with a or b or c. The correct conclusion for
3303 this pattern is that a match can start with any character.
3304
3305
3306 Version 4.4 13-Aug-03
3307 ---------------------
3308
3309 1. In UTF-8 mode, a character class containing characters with values between
3310 127 and 255 was not handled correctly if the compiled pattern was studied.
3311 In fixing this, I have also improved the studying algorithm for such
3312 classes (slightly).
3313
3314 2. Three internal functions had redundant arguments passed to them. Removal
3315 might give a very teeny performance improvement.
3316
3317 3. Documentation bug: the value of the capture_top field in a callout is *one
3318 more than* the number of the hightest numbered captured substring.
3319
3320 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3321 in incorrectly linking with a previously installed version. They now link
3322 explicitly with libpcre.la.
3323
3324 5. configure.in no longer needs to recognize Cygwin specially.
3325
3326 6. A problem in pcre.in for Windows platforms is fixed.
3327
3328 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3329 pcretest, it used to include the size of the study block as part of its
3330 output. Unfortunately, the structure contains a field that has a different
3331 size on different hardware architectures. This meant that the tests that
3332 showed this size failed. As the block is currently always of a fixed size,
3333 this information isn't actually particularly useful in pcretest output, so
3334 I have just removed it.
3335
3336 8. Three pre-processor statements accidentally did not start in column 1.
3337 Sadly, there are *still* compilers around that complain, even though
3338 standard C has not required this for well over a decade. Sigh.
3339
3340 9. In pcretest, the code for checking callouts passed small integers in the
3341 callout_data field, which is a void * field. However, some picky compilers
3342 complained about the casts involved for this on 64-bit systems. Now
3343 pcretest passes the address of the small integer instead, which should get
3344 rid of the warnings.
3345
3346 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3347 both compile and run time, and gives an error if an invalid UTF-8 sequence
3348 is found. There is a option for disabling this check in cases where the
3349 string is known to be correct and/or the maximum performance is wanted.
3350
3351 11. In response to a bug report, I changed one line in Makefile.in from
3352
3353 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3354 to
3355 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3356
3357 to look similar to other lines, but I have no way of telling whether this
3358 is the right thing to do, as I do not use Windows. No doubt I'll get told
3359 if it's wrong...
3360
3361
3362 Version 4.3 21-May-03
3363 ---------------------
3364
3365 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3366 Makefile.
3367
3368 2. Some refactoring to improve the quality of the code:
3369
3370 (i) The utf8_table... variables are now declared "const".
3371
3372 (ii) The code for \cx, which used the "case flipping" table to upper case
3373 lower case letters, now just substracts 32. This is ASCII-specific,
3374 but the whole concept of \cx is ASCII-specific, so it seems
3375 reasonable.
3376
3377 (iii) PCRE was using its character types table to recognize decimal and
3378 hexadecimal digits in the pattern. This is silly, because it handles
3379 only 0-9, a-f, and A-F, but the character types table is locale-
3380 specific, which means strange things might happen. A private
3381 table is now used for this - though it costs 256 bytes, a table is
3382 much faster than multiple explicit tests. Of course, the standard
3383 character types table is still used for matching digits in subject
3384 strings against \d.
3385
3386 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3387 ending in _t are). So I've renamed it as ESC_tee.
3388
3389 3. The first argument for regexec() in the POSIX wrapper should have been
3390 defined as "const".
3391
3392 4. Changed pcretest to use malloc() for its buffers so that they can be
3393 Electric Fenced for debugging.
3394
3395 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3396 to read one or more bytes before the start of the subject string. Often this
3397 had no effect on PCRE's behaviour, but in some circumstances it could
3398 provoke a segmentation fault.
3399
3400 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3401 to try to read one or more bytes before the start of the subject string.
3402
3403 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3404 UTF-8 support could misbehave in various ways if the subject string
3405 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3406 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3407 back over UTF-8 characters.)
3408
3409
3410 Version 4.2 14-Apr-03
3411 ---------------------
3412
3413 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3414
3415 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3416 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3417 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3418 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3419 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3420 and BUILD_EXEEXT
3421 Note: automatic setting of the BUILD variables is not yet working
3422 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3423 compile-time but not at link-time
3424 [LINK]: use for linking executables only
3425 make different versions for Windows and non-Windows
3426 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3427 libraries
3428 [LINK_FOR_BUILD]: new variable
3429 [OBJEXT]: use throughout
3430 [EXEEXT]: use throughout
3431 <winshared>: new target
3432 <wininstall>: new target
3433 <dftables.o>: use native compiler
3434 <dftables>: use native linker
3435 <install>: handle Windows platform correctly
3436 <clean>: ditto
3437 <check>: ditto
3438 copy DLL to top builddir before testing
3439
3440 As part of these changes, -no-undefined was removed again. This was reported
3441 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3442 in any case.
3443
3444 3. Some tidies to get rid of compiler warnings:
3445
3446 . In the match_data structure, match_limit was an unsigned long int, whereas
3447 match_call_count was an int. I've made them both unsigned long ints.
3448
3449 . In pcretest the fact that a const uschar * doesn't automatically cast to
3450 a void * provoked a warning.
3451
3452 . Turning on some more compiler warnings threw up some "shadow" variables
3453 and a few more missing casts.
3454
3455 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3456 option, a class that contained a single character with a value between 128
3457 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3458
3459 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3460 option, a class that contained several characters, but with at least one
3461 whose value was between 128 and 255 caused PCRE to crash.
3462
3463
3464 Version 4.1 12-Mar-03
3465 ---------------------
3466
3467 1. Compiling with gcc -pedantic found a couple of places where casts were
3468 needed, and a string in dftables.c that was longer than standard compilers are
3469 required to support.
3470
3471 2. Compiling with Sun's compiler found a few more places where the code could
3472 be tidied up in order to avoid warnings.
3473
3474 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3475 first of these names is deprecated in the latest Autoconf in favour of the name
3476 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3477 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3478 analogy I have changed it to CFLAGS_FOR_BUILD.
3479
3480 4. Added -no-undefined to the linking command in the Makefile, because this is
3481 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3482 linking step for the pcreposix library.
3483
3484 5. PCRE was failing to diagnose the case of two named groups with the same
3485 name.
3486
3487 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3488 literal character that is needed in the subject for a match, and scans along to
3489 ensure that it is present before embarking on the full matching process. This
3490 saves time in cases of nested unlimited repeats that are never going to match.
3491 Problem: the scan can take a lot of time if the subject is very long (e.g.
3492 megabytes), thus penalizing straightforward matches. It is now done only if the
3493 amount of subject to be scanned is less than 1000 bytes.
3494
3495 7. A lesser problem with the same optimization is that it was recording the
3496 first character of an anchored pattern as "needed", thus provoking a search
3497 right along the subject, even when the first match of the pattern was going to
3498 fail. The "needed" character is now not set for anchored patterns, unless it
3499 follows something in the pattern that is of non-fixed length. Thus, it still
3500 fulfils its original purpose of finding quick non-matches in cases of nested
3501 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3502
3503
3504 Version 4.0 17-Feb-03
3505 ---------------------
3506
3507 1. If a comment in an extended regex that started immediately after a meta-item
3508 extended to the end of string, PCRE compiled incorrect data. This could lead to
3509 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3510
3511 2. Moved to autoconf 2.53 and libtool 1.4.2.
3512
3513 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3514 the special perltest8 script is no longer needed - all the tests can be run
3515 from a single perltest script.
3516
3517 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3518 by \s. It has now been removed in PCRE. This means it isn't recognized as
3519 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3520 class [:space:] *does* include VT, thereby creating a mess.
3521
3522 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3523 space and tab.
3524
3525 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3526 its new features into the main test script, reducing the number of scripts.
3527
3528 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3529 were backward compatible, and made the (?i) apply to the whole pattern, as if
3530 /i were given. Now it behaves more logically, and applies the option setting
3531 only to what follows. PCRE has been changed to follow suit. However, if it
3532 finds options settings right at the start of the pattern, it extracts them into
3533 the global options, as before. Thus, they show up in the info data.
3534
3535 8. Added support for the \Q...\E escape sequence. Characters in between are
3536 treated as literals. This is slightly different from Perl in that $ and @ are
3537 also handled as literals inside the quotes. In Perl, they will cause variable
3538 interpolation. Note the following examples:
3539
3540 Pattern PCRE matches Perl matches
3541
3542 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3543 \Qabc\$xyz\E abc\$xyz abc\$xyz
3544 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3545
3546 For compatibility with Perl, \Q...\E sequences are recognized inside character
3547 classes as well as outside them.
3548
3549 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3550 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3551 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3552 signed/unsigned warnings.
3553
3554 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3555 option for pcretest, so I've replaced it by a simple function that does just
3556 that job.
3557
3558 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3559 "pcregrep -".
3560
3561 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3562 Java package. This provides some syntactic sugar for simple cases of what my
3563 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3564 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3565 item, you can use this simplified notation. Note that only makes sense with
3566 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3567 greediness, whatever the setting of the PCRE_UNGREEDY option.
3568
3569 13. A change of greediness default within a pattern was not taking effect at
3570 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3571 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3572 was abstracted outside.
3573
3574 14. PCRE now supports the \G assertion. It is true when the current matching
3575 position is at the start point of the match. This differs from \A when the
3576 starting offset is non-zero. Used with the /g option of pcretest (or similar
3577 code), it works in the same way as it does for Perl's /g option. If all
3578 alternatives of a regex begin with \G, the expression is anchored to the start
3579 match position, and the "anchored" flag is set in the compiled expression.
3580
3581 15. Some bugs concerning the handling of certain option changes within patterns
3582 have been fixed. These applied to options other than (?ims). For example,
3583 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3584 been the other way round. Some of this was related to change 7 above.
3585
3586 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3587 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3588 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3589 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3590
3591 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3592 mode. Unlike ".", it always matches newline, whatever the setting of
3593 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3594 assertions. Perl allows it, but it doesn't (in general) work because it can't
3595 calculate the length of the lookbehind. At least, that's the case for Perl
3596 5.8.0 - I've been told they are going to document that it doesn't work in
3597 future.
3598
3599 18. Added an error diagnosis for escapes that PCRE does not support: these are
3600 \L, \l, \N, \P, \p, \U, \u, and \X.
3601
3602 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3603 reading past the end of the pattern in cases such as /[abcd/.
3604
3605 20. PCRE was getting more memory than necessary for patterns with classes that
3606 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3607
3608 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3609 compiling PCRE for use with Virtual Pascal.
3610
3611 22. Small fix to the Makefile to make it work properly if the build is done
3612 outside the source tree.
3613
3614 23. Added a new extension: a condition to go with recursion. If a conditional
3615 subpattern starts with (?(R) the "true" branch is used if recursion has
3616 happened, whereas the "false" branch is used only at the top level.
3617
3618 24. When there was a very long string of literal characters (over 255 bytes
3619 without UTF support, over 250 bytes with UTF support), the computation of how
3620 much memory was required could be incorrect, leading to segfaults or other
3621 strange effects.
3622
3623 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3624 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3625 there was a subsequent back reference to those brackets. This meant that, for
3626 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3627 possible to check for precisely this case. All we can do is abandon the
3628 optimization if .* occurs inside capturing brackets when there are any back
3629 references whatsoever. (See below for a better fix that came later.)
3630
3631 26. The handling of the optimization for finding the first character of a
3632 non-anchored pattern, and for finding a character that is required later in the
3633 match were failing in some cases. This didn't break the matching; it just
3634 failed to optimize when it could. The way this is done has been re-implemented.
3635
3636 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3637
3638 28. Added a new feature that provides some of the functionality that Perl
3639 provides with (?{...}). The facility is termed a "callout". The way it is done
3640 in PCRE is for the caller to provide an optional function, by setting
3641 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3642 global variable. By default it is unset, which disables all calling out. To get
3643 the function called, the regex must include (?C) at appropriate points. This
3644 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3645 This provides a means of identifying different callout points. When PCRE
3646 reaches such a point in the regex, if pcre_callout has been set, the external
3647 function is called. It is provided with data in a structure called
3648 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3649 matching continues; if it returns a non-zero value, the match at the current
3650 point fails. However, backtracking will occur if possible. [This was changed
3651 later and other features added - see item 49 below.]
3652
3653 29. pcretest is upgraded to test the callout functionality. It provides a
3654 callout function that displays information. By default, it shows the start of
3655 the match and the current position in the text. There are some new data escapes
3656 to vary what happens:
3657
3658 \C+ in addition, show current contents of captured substrings
3659 \C- do not supply a callout function
3660 \C!n return 1 when callout number n is reached
3661 \C!n!m return 1 when callout number n is reached for the mth time
3662
3663 30. If pcregrep was called with the -l option and just a single file name, it
3664 output "<stdin>" if a match was found, instead of the file name.
3665
3666 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3667 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3668 pcre_exec(). This saves a malloc/free per call. The default value of
3669 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3670 when configuring.
3671
3672 32. The default maximum size of a compiled pattern is 64K. There have been a
3673 few cases of people hitting this limit. The code now uses macros to handle the
3674 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3675 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3676 configuring. Tests 2 and 5 work only with 2-byte links because they output
3677 debugging information about compiled patterns.
3678
3679 33. Internal code re-arrangements:
3680
3681 (a) Moved the debugging function for printing out a compiled regex into
3682 its own source file (printint.c) and used #include to pull it into
3683 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3684 separate copies.
3685
3686 (b) Defined the list of op-code names for debugging as a macro in
3687 internal.h so that it is next to the definition of the opcodes.
3688
3689 (c) Defined a table of op-code lengths for simpler skipping along compiled
3690 code. This is again a macro in internal.h so that it is next to the
3691 definition of the opcodes.
3692
3693 34. Added support for recursive calls to individual subpatterns, along the
3694 lines of Robin Houston's patch (but implemented somewhat differently).
3695
3696 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3697 allow it to read and process whole directories in Win32. This code was
3698 contributed by Lionel Fourquaux; it has not been tested by me.
3699
3700 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3701 used to name a group. Names consist of alphanumerics and underscores, and must
3702 be unique. Back references use the syntax (?P=name) and recursive calls use
3703 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3704 numbers. The function pcre_fullinfo() can be used after compilation to extract
3705 a name/number map. There are three relevant calls:
3706
3707 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3708 PCRE_INFO_NAMECOUNT yields the number of entries
3709 PCRE_INFO_NAMETABLE yields a pointer to the map.
3710
3711 The map is a vector of fixed-size entries. The size of each entry depends on
3712 the length of the longest name used. The first two bytes of each entry are the
3713 group number, most significant byte first. There follows the corresponding
3714 name, zero terminated. The names are in alphabetical order.
3715
3716 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3717 case instead of 255. Making it the same both with and without UTF-8 support
3718 means that the same test output works with both.
3719
3720 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3721 calling malloc() with a zero argument.
3722
3723 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3724 optimization. I've improved things by keeping a bitmap of backreferences with
3725 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3726 fact referenced, the optimization can be applied. It is unlikely that a
3727 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3728 the match to follow \n) will appear inside brackets with a number greater than
3729 31, but if it does, any back reference > 31 suppresses the optimization.
3730
3731 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3732 of disabling numbered capturing parentheses. Any opening parenthesis that is
3733 not followed by ? behaves as if it were followed by ?: but named parentheses
3734 can still be used for capturing (and they will acquire numbers in the usual
3735 way).
3736
3737 41. Redesigned the return codes from the match() function into yes/no/error so
3738 that errors can be passed back from deep inside the nested calls. A malloc
3739 failure while inside a recursive subpattern call now causes the
3740 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3741
3742 42. It is now possible to set a limit on the number of times the match()
3743 function is called in a call to pcre_exec(). This facility makes it possible to
3744 limit the amount of recursion and backtracking, though not in a directly
3745 obvious way, because the match() function is used in a number of different
3746 circumstances. The count starts from zero for each position in the subject
3747 string (for non-anchored patterns). The default limit is, for compatibility, a
3748 large number, namely 10 000 000. You can change this in two ways:
3749
3750 (a) When configuring PCRE before making, you can use --with-match-limit=n
3751 to set a default value for the compiled library.
3752
3753 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3754 a different value is set. See 45 below.
3755
3756 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3757
3758 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3759 of things that can be changed at compile time. The first argument specifies
3760 what is wanted and the second points to where the information is to be placed.
3761 The current list of available information is:
3762
3763 PCRE_CONFIG_UTF8
3764
3765 The output is an integer that is set to one if UTF-8 support is available;
3766 otherwise it is set to zero.
3767
3768 PCRE_CONFIG_NEWLINE
3769
3770 The output is an integer that it set to the value of the code that is used for
3771 newline. It is either LF (10) or CR (13).
3772
3773 PCRE_CONFIG_LINK_SIZE
3774
3775 The output is an integer that contains the number of bytes used for internal
3776 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3777
3778 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3779
3780 The output is an integer that contains the threshold above which the POSIX
3781 interface uses malloc() for output vectors. See item 31 above.
3782
3783 PCRE_CONFIG_MATCH_LIMIT
3784
3785 The output is an unsigned integer that contains the default limit of the number
3786 of match() calls in a pcre_exec() execution. See 42 above.
3787
3788 44. pcretest has been upgraded by the addition of the -C option. This causes it
3789 to extract all the available output from the new pcre_config() function, and to
3790 output it. The program then exits immediately.
3791
3792 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3793 order to support additional features. One way would have been to define
3794 pcre_exec2() (for example) with extra arguments, but this would not have been
3795 extensible, and would also have required all calls to the original function to
3796 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3797 is used for passing in "extra" data from pcre_study().
3798
3799 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3800 contains the following fields:
3801
3802 flags a bitmap indicating which of the following fields are set
3803 study_data opaque data from pcre_study()
3804 match_limit a way of specifying a limit on match() calls for a specific
3805 call to pcre_exec()
3806 callout_data data for callouts (see 49 below)
3807
3808 The flag bits are also defined in pcre.h, and are
3809
3810 PCRE_EXTRA_STUDY_DATA
3811 PCRE_EXTRA_MATCH_LIMIT
3812 PCRE_EXTRA_CALLOUT_DATA
3813
3814 The pcre_study() function now returns one of these new pcre_extra blocks, with
3815 the actual study data pointed to by the study_data field, and the
3816 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3817 before. That is, this change is entirely upwards-compatible and requires no
3818 change to existing code.
3819
3820 If you want to pass in additional data to pcre_exec(), you can either place it
3821 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3822 block.
3823
3824 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3825 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3826 times with different match limits, until it finds the minimum value needed for
3827 pcre_exec() to complete. The value is then output. This can be instructive; for
3828 most simple matches the number is quite small, but for pathological cases it
3829 gets very large very quickly.
3830
3831 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3832 returns the size of the data block pointed to by the study_data field in a
3833 pcre_extra block, that is, the value that was passed as the argument to
3834 pcre_malloc() when PCRE was getting memory in which to place the information
3835 created by pcre_study(). The fourth argument should point to a size_t variable.
3836 pcretest has been extended so that this information is shown after a successful
3837 pcre_study() call when information about the compiled regex is being displayed.
3838
3839 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3840 because what follows is always an absolute path. (Later: it turns out that this
3841 is more than cosmetic for MinGW, because it doesn't like empty path
3842 components.)
3843
3844 49. Some changes have been made to the callout feature (see 28 above):
3845
3846 (i) A callout function now has three choices for what it returns:
3847
3848 0 => success, carry on matching
3849 > 0 => failure at this point, but backtrack if possible
3850 < 0 => serious error, return this value from pcre_exec()
3851
3852 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3853 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3854 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3855 use by callout functions. It will never be used by PCRE itself.
3856
3857 (ii) The pcre_extra structure (see 45 above) has a void * field called
3858 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3859 pcre_callout_block structure has a field of the same name. The contents of
3860 the field passed in the pcre_extra structure are passed to the callout
3861 function in the corresponding field in the callout block. This makes it
3862 easier to use the same callout-containing regex from multiple threads. For
3863 testing, the pcretest program has a new data escape
3864
3865 \C*n pass the number n (may be negative) as callout_data
3866
3867 If the callout function in pcretest receives a non-zero value as
3868 callout_data, it returns that value.
3869
3870 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3871 there were some redundant $(CFLAGS) in commands that are now specified as
3872 $(LINK), which already includes $(CFLAGS).
3873
3874 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3875 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3876 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3877 one-byte characters throughout. Note that case-insensitive matching applies
3878 only to characters whose values are less than 256. PCRE doesn't support the
3879 notion of cases for higher-valued characters.
3880
3881 (i) A character class whose characters are all within 0-255 is handled as
3882 a bit map, and the map is inverted for negative classes. Previously, a
3883 character > 255 always failed to match such a class; however it should
3884 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3885
3886 (ii) A negated character class with a single character < 255 is coded as
3887 "not this character" (OP_NOT). This wasn't working properly when the test
3888 character was multibyte, either singly or repeated.
3889
3890 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3891 mode, for example: \x{100}{2,3}.
3892
3893 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3894 singly or repeated) now correctly test multibyte characters. However,
3895 PCRE doesn't recognize any characters with values greater than 255 as
3896 digits, spaces, or word characters. Such characters always match \D, \S,
3897 and \W, and never match \d, \s, or \w.
3898
3899 (v) Classes may now contain characters and character ranges with values
3900 greater than 255. For example: [ab\x{100}-\x{400}].
3901
3902 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3903 PCRE in UTF-8 mode.
3904
3905 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3906 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3907 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3908 value.)
3909
3910 53. The single man page has become too large. I have therefore split it up into
3911 a number of separate man pages. These also give rise to individual HTML pages;
3912 these are now put in a separate directory, and there is an index.html page that
3913 lists them all. Some hyperlinking between the pages has been installed.
3914
3915 54. Added convenience functions for handling named capturing parentheses.
3916
3917 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3918 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3919 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3920 are faulted.
3921
3922 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3923 calling configure. These values are used when compiling the dftables.c program
3924 which is run to generate the source of the default character tables. They
3925 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3926 you will need to set these values.
3927
3928 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3929
3930
3931 Version 3.9 02-Jan-02
3932 ---------------------
3933
3934 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3935
3936 2. If --disable-static was given, the building process failed when trying to
3937 build pcretest and pcregrep. (For some reason it was using libtool to compile
3938 them, which is not right, as they aren't part of the library.)
3939
3940
3941 Version 3.8 18-Dec-01
3942 ---------------------
3943
3944 1. The experimental UTF-8 code was completely screwed up. It was packing the
3945 bytes in the wrong order. How dumb can you get?
3946
3947
3948 Version 3.7 29-Oct-01
3949 ---------------------
3950
3951 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3952 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3953 this didn't happen under Solaris 8, where I normally test things.
3954
3955 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3956 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3957 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3958 this fix an hour or so after the initial 3.7 release.)
3959
3960
3961 Version 3.6 23-Oct-01
3962 ---------------------
3963
3964 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3965 offsets passed as NULL with zero offset count.
3966
3967 2. The config.guess and config.sub files had not been updated when I moved to
3968 the latest autoconf.
3969
3970
3971 Version 3.5 15-Aug-01
3972 ---------------------
3973
3974 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3975 had been forgotten.
3976
3977 2. By using declared but undefined structures, we can avoid using "void"
3978 definitions in pcre.h while keeping the internal definitions of the structures
3979 private.
3980
3981 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3982 user point of view, this means that both static and shared libraries are built
3983 by default, but this can be individually controlled. More of the work of
3984 handling this static/shared cases is now inside libtool instead of PCRE's make
3985 file.
3986
3987 4. The pcretest utility is now installed along with pcregrep because it is
3988 useful for users (to test regexs) and by doing this, it automatically gets
3989 relinked by libtool. The documentation has been turned into a man page, so
3990 there are now .1, .txt, and .html versions in /doc.
3991
3992 5. Upgrades to pcregrep:
3993 (i) Added long-form option names like gnu grep.
3994 (ii) Added --help to list all options with an explanatory phrase.
3995 (iii) Added -r, --recursive to recurse into sub-directories.
3996 (iv) Added -f, --file to read patterns from a file.
3997
3998 6. pcre_exec() was referring to its "code" argument before testing that
3999 argument for NULL (and giving an error if it was NULL).
4000
4001 7. Upgraded Makefile.in to allow for compiling in a different directory from
4002 the source directory.
4003
4004 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4005 options bits, the pointer it was passed was to an int instead of to an unsigned
4006 long int. This mattered only on 64-bit systems.
4007
4008 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4009 generated) instead of pcre.in, which it its source. Also made the same change
4010 in several of the .c files.
4011
4012 10. A new release of gcc defines printf() as a macro, which broke pcretest
4013 because it had an ifdef in the middle of a string argument for printf(). Fixed
4014 by using separate calls to printf().
4015
4016 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4017 script, to force use of CR or LF instead of \n in the source. On non-Unix
4018 systems, the value can be set in config.h.
4019
4020 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4021 absolute limit. Changed the text of the error message to make this clear, and
4022 likewise updated the man page.
4023
4024 13. The limit of 99 on the number of capturing subpatterns has been removed.
4025 The new limit is 65535, which I hope will not be a "real" limit.
4026
4027
4028 Version 3.4 22-Aug-00
4029 ---------------------
4030
4031 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4032
4033 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4034
4035
4036 Version 3.3 01-Aug-00
4037 ---------------------
4038
4039 1. If an octal character was given, but the value was greater than \377, it
4040 was not getting masked to the least significant bits, as documented. This could
4041 lead to crashes in some systems.
4042
4043 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4044 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4045
4046 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4047 These just pass their arguments on to (pcre_free)(), but they are provided
4048 because some uses of PCRE bind it to non-C systems that can call its functions,
4049 but cannot call free() or pcre_free() directly.
4050
4051 4. Add "make test" as a synonym for "make check". Corrected some comments in
4052 the Makefile.
4053
4054 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4055 Makefile.
4056
4057 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4058 command called pgrep for grepping around the active processes.
4059
4060 7. Added the beginnings of support for UTF-8 character strings.
4061
4062 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4063 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4064 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4065 out for the ar command.)
4066
4067
4068 Version 3.2 12-May-00
4069 ---------------------
4070
4071 This is purely a bug fixing release.
4072
4073 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4074 of ZA. This was just one example of several cases that could provoke this bug,
4075 which was introduced by change 9 of version 2.00. The code for breaking
4076 infinite loops after an iteration that matches an empty string was't working
4077 correctly.
4078
4079 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4080 when matched against abbab (for example). After matching an empty string, it
4081 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4082 caused it to match further down the string than it should.
4083
4084 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4085 was there because it doesn't seem to be needed, and it causes trouble on some
4086 systems, as it is not a Standard C header. It has been removed.
4087
4088 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4089 were reported on the Macintosh. The changes were from
4090
4091 while ((c = *(++ptr)) != 0 && c != '\n');
4092 to
4093 while ((c = *(++ptr)) != 0 && c != '\n') ;
4094
4095 Totally extraordinary, but if that's what it takes...
4096
4097 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4098 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4099 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4100 assumes the way PCRE uses memmove() (always moving upwards).
4101
4102 6. PCRE is being used in one environment where strchr() is not available. There
4103 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4104 faster code anyway.
4105
4106
4107 Version 3.1 09-Feb-00
4108 ---------------------
4109
4110 The only change in this release is the fixing of some bugs in Makefile.in for
4111 the "install" target:
4112
4113 (1) It was failing to install pcreposix.h.
4114
4115 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4116
4117
4118 Version 3.0 01-Feb-00
4119 ---------------------
4120
4121 1. Add support for the /+ modifier to perltest (to output $` like it does in
4122 pcretest).
4123
4124 2. Add support for the /g modifier to perltest.
4125
4126 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4127 matches null strings.
4128
4129 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4130 pattern. Perl treats empty patterns specially - it reuses the most recent
4131 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4132 effect.
4133
4134 5. The POSIX interface was broken in that it was just handing over the POSIX
4135 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4136 required a bigger vector, with some working space on the end. This means that
4137 the POSIX wrapper now has to get and free some memory, and copy the results.
4138
4139 6. Added some simple autoconf support, placing the test data and the
4140 documentation in separate directories, re-organizing some of the
4141 information files, and making it build pcre-config (a GNU standard). Also added
4142 libtool support for building PCRE as a shared library, which is now the
4143 default.
4144
4145 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4146 09 are not valid octal constants. Single digits will be used for minor values
4147 less than 10.
4148
4149 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4150 existing programs that set these in the POSIX interface can use PCRE without
4151 modification.
4152
4153 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4154 return all that pcre_info() returns, plus additional data. The pcre_info()
4155 function is retained for compatibility, but is considered to be obsolete.
4156
4157 10. Added experimental recursion feature (?R) to handle one common case that
4158 Perl 5.6 will be able to do with (?p{...}).
4159
4160 11. Added support for POSIX character classes like [:alpha:], which Perl is
4161 adopting.
4162
4163
4164 Version 2.08 31-Aug-99
4165 ----------------------
4166
4167 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4168 trying to match at the startoffset position, but instead was moving forward to
4169 the next newline as if a previous match had failed.
4170
4171 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4172 and could get into a loop if a null string was matched other than at the start
4173 of the subject.
4174
4175 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4176 be distinguished at compile time, and for completeness also added PCRE_DATE.
4177
4178 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4179 in GnuWin32 environments.
4180
4181
4182 Version 2.07 29-Jul-99
4183 ----------------------
4184
4185 1. The documentation is now supplied in plain text form and HTML as well as in
4186 the form of man page sources.
4187
4188 2. C++ compilers don't like assigning (void *) values to other pointer types.
4189 In particular this affects malloc(). Although there is no problem in Standard
4190 C, I've put in casts to keep C++ compilers happy.
4191
4192 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4193 should be (const char *).
4194
4195 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4196 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4197 However, I haven't made this a standard facility. The documentation doesn't
4198 mention it, and the Makefile doesn't support it.
4199
4200 5. The Makefile now contains an "install" target, with editable destinations at
4201 the top of the file. The pcretest program is not installed.
4202
4203 6. pgrep -V now gives the PCRE version number and date.
4204
4205 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4206 causing the entire string to be ignored, instead of just the last character.
4207
4208 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4209 non-matching string, it can take a very, very long time, even for strings of
4210 quite modest length, because of the nested recursion. PCRE now does better in
4211 some of these cases. It does this by remembering the last required literal
4212 character in the pattern, and pre-searching the subject to ensure it is present
4213 before running the real match. In other words, it applies a heuristic to detect
4214 some types of certain failure quickly, and in the above example, if presented
4215 with a string that has no trailing " it gives "no match" very quickly.
4216
4217 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4218 other alternatives are tried instead.
4219
4220
4221 Version 2.06 09-Jun-99
4222 ----------------------
4223
4224 1. Change pcretest's output for amount of store used to show just the code
4225 space, because the remainder (the data block) varies in size between 32-bit and
4226 64-bit systems.
4227
4228 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4229 start matching at. This allows lookbehinds to work when searching for multiple
4230 occurrences in a string.
4231
4232 3. Added additional options to pcretest for testing multiple occurrences:
4233
4234 /+ outputs the rest of the string that follows a match
4235 /g loops for multiple occurrences, using the new startoffset argument
4236 /G loops for multiple occurrences by passing an incremented pointer
4237
4238 4. PCRE wasn't doing the "first character" optimization for patterns starting
4239 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4240 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4241 the letter 'x'. On long subject strings, this gives a significant speed-up.
4242
4243
4244 Version 2.05 21-Apr-99
4245 ----------------------
4246
4247 1. Changed the type of magic_number from int to long int so that it works
4248 properly on 16-bit systems.
4249
4250 2. Fixed a bug which caused patterns starting with .* not to work correctly
4251 when the subject string contained newline characters. PCRE was assuming
4252 anchoring for such patterns in all cases, which is not correct because .* will
4253 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4254 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4255 must be retried after every newline in the subject.
4256
4257
4258 Version 2.04 18-Feb-99
4259 ----------------------
4260
4261 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4262 computation of the store needed to hold the pattern was incorrect (too large).
4263 If such patterns were nested a few deep, this could multiply and become a real
4264 problem.
4265
4266 2. Added /M option to pcretest to show the memory requirement of a specific
4267 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4268
4269 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4270 compiled in such a way that the backtracking after subsequent failure was
4271 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4272 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4273
4274
4275 Version 2.03 02-Feb-99
4276 ----------------------
4277
4278 1. Fixed typo and small mistake in man page.
4279
4280 2. Added 4th condition (GPL supersedes if conflict) and created separate
4281 LICENCE file containing the conditions.
4282
4283 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4284 Perl, that is the internal \ allows the delimiter to be included in the
4285 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4286 the final delimiter, add \ to the end of the pattern (to test the error).
4287
4288 4. Added the convenience functions for extracting substrings after a successful
4289 match. Updated pcretest to make it able to test these functions.
4290
4291
4292 Version 2.02 14-Jan-99
4293 ----------------------
4294
4295 1. Initialized the working variables associated with each extraction so that
4296 their saving and restoring doesn't refer to uninitialized store.
4297
4298 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4299 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4300 fix the problem.
4301
4302 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4303 calls, and wasn't printing the correct value for compiling calls. Increased the
4304 default value of LOOPREPEAT, and the number of significant figures in the
4305 times.
4306
4307 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4308
4309 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4310 a building problem on Windows NT with a FAT file system.
4311
4312
4313 Version 2.01 21-Oct-98
4314 ----------------------
4315
4316 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4317 to character tables built by pcre_maketables() in the current locale. If NULL
4318 is passed, the default tables are used.
4319
4320
4321 Version 2.00 24-Sep-98
4322 ----------------------
4323
4324 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4325 it any more.
4326
4327 2. Allow quantification of (?>) groups, and make it work correctly.
4328
4329 3. The first character computation wasn't working for (?>) groups.
4330
4331 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4332 end of the subject) and add 5.005's \z, which really does match only at the
4333 very end of the subject.
4334
4335 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4336
4337 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4338 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4339 localized options. All options to pcre_study() were also removed.
4340
4341 7. Add other new features from 5.005:
4342
4343 $(?<= positive lookbehind
4344 $(?<! negative lookbehind
4345 (?imsx-imsx) added the unsetting capability
4346 such a setting is global if at outer level; local otherwise
4347 (?imsx-imsx:) non-capturing groups with option setting
4348 (?(cond)re|re) conditional pattern matching
4349
4350 A backreference to itself in a repeated group matches the previous
4351 captured string.
4352
4353 8. General tidying up of studying (both automatic and via "study")
4354 consequential on the addition of new assertions.
4355
4356 9. As in 5.005, unlimited repeated groups that could match an empty substring
4357 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4358 runtime if any iteration does actually match an empty substring.
4359
4360 10. Include the RunTest script in the distribution.
4361
4362 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4363 discrepancies, some of which were old and were also with respect to 5.004. They
4364 have now been fixed.
4365
4366
4367 Version 1.09 28-Apr-98
4368 ----------------------
4369
4370 1. A negated single character class followed by a quantifier with a minimum
4371 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4372 program crashes, or just wrong answers. This did not apply to negated classes
4373 containing more than one character, or to minima other than one.
4374
4375
4376 Version 1.08 27-Mar-98
4377 ----------------------
4378
4379 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4380
4381 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4382 latter must appear before anything that relies on it in the pattern.
4383
4384
4385 Version 1.07 16-Feb-98
4386 ----------------------
4387
4388 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4389 repeat of a potentially empty string).
4390
4391
4392 Version 1.06 23-Jan-98
4393 ----------------------
4394
4395 1. Added Markus Oberhumer's little patches for C++.
4396
4397 2. Literal strings longer than 255 characters were broken.
4398
4399
4400 Version 1.05 23-Dec-97
4401 ----------------------
4402
4403 1. Negated character classes containing more than one character were failing if
4404 PCRE_CASELESS was set at run time.
4405
4406
4407 Version 1.04 19-Dec-97
4408 ----------------------
4409
4410 1. Corrected the man page, where some "const" qualifiers had been omitted.
4411
4412 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4413 input syntax.
4414
4415 3. Fixed memory leak which occurred when a regex with back references was
4416 matched with an offsets vector that wasn't big enough. The temporary memory
4417 that is used in this case wasn't being freed if the match failed.
4418
4419 4. Tidied pcretest to ensure it frees memory that it gets.
4420
4421 5. Temporary memory was being obtained in the case where the passed offsets
4422 vector was exactly big enough.
4423
4424 6. Corrected definition of offsetof() from change 5 below.
4425
4426 7. I had screwed up change 6 below and broken the rules for the use of
4427 setjmp(). Now fixed.
4428
4429
4430 Version 1.03 18-Dec-97
4431 ----------------------
4432
4433 1. A erroneous regex with a missing opening parenthesis was correctly
4434 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4435 on some systems.
4436
4437 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4438 it was reported that one broken compiler failed on the former because "code" is
4439 also an independent variable.
4440
4441 3. The erroneous regex a[]b caused an array overrun reference.
4442
4443 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4444 fail on data ending with that character. (It was going on too far, and checking
4445 the next character, typically a binary zero.) This was specific to the
4446 optimized code for single-character negative classes.
4447
4448 5. Added a contributed patch from the TIN world which does the following:
4449
4450 + Add an undef for memmove, in case the the system defines a macro for it.
4451
4452 + Add a definition of offsetof(), in case there isn't one. (I don't know
4453 the reason behind this - offsetof() is part of the ANSI standard - but
4454 it does no harm).
4455
4456 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4457 most of the places where whitespace preceded '#'. I have given up and
4458 allowed the remaining 2 cases to be at the margin.
4459
4460 + Rename some variables in pcre to eliminate shadowing. This seems very
4461 pedantic, but does no harm, of course.
4462
4463 6. Moved the call to setjmp() into its own function, to get rid of warnings
4464 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4465
4466 7. Constructs such as \d{8,} were compiling into the equivalent of
4467 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4468 outcome, but in this particular case used more store than had been allocated,
4469 which caused the bug to be discovered because it threw up an internal error.
4470
4471 8. The debugging code in both pcre and pcretest for outputting the compiled
4472 form of a regex was going wrong in the case of back references followed by
4473 curly-bracketed repeats.
4474
4475
4476 Version 1.02 12-Dec-97
4477 ----------------------
4478
4479 1. Typos in pcre.3 and comments in the source fixed.
4480
4481 2. Applied a contributed patch to get rid of places where it used to remove
4482 'const' from variables, and fixed some signed/unsigned and uninitialized
4483 variable warnings.
4484
4485 3. Added the "runtest" target to Makefile.
4486
4487 4. Set default compiler flag to -O2 rather than just -O.
4488
4489
4490 Version 1.01 19-Nov-97
4491 ----------------------
4492
4493 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4494 like /([ab]*)*/, that is, for classes with more than one character in them.
4495
4496 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4497 as /((?>a*))*/ (a PCRE_EXTRA facility).
4498
4499
4500 Version 1.00 18-Nov-97
4501 ----------------------
4502
4503 1. Added compile-time macros to support systems such as SunOS4 which don't have
4504 memmove() or strerror() but have other things that can be used instead.
4505
4506 2. Arranged that "make clean" removes the executables.
4507
4508
4509 Version 0.99 27-Oct-97
4510 ----------------------
4511
4512 1. Fixed bug in code for optimizing classes with only one character. It was
4513 initializing a 32-byte map regardless, which could cause it to run off the end
4514 of the memory it had got.
4515
4516 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4517
4518
4519 Version 0.98 22-Oct-97
4520 ----------------------
4521
4522 1. Fixed bug in code for handling temporary memory usage when there are more
4523 back references than supplied space in the ovector. This could cause segfaults.
4524
4525
4526 Version 0.97 21-Oct-97
4527 ----------------------
4528
4529 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4530
4531 2. Optimized negated single characters not to use a bit map.
4532
4533 3. Brought error texts together as macro definitions; clarified some of them;
4534 fixed one that was wrong - it said "range out of order" when it meant "invalid
4535 escape sequence".
4536
4537 4. Changed some char * arguments to const char *.
4538
4539 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4540
4541 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4542 pcretest.
4543
4544
4545 Version 0.96 16-Oct-97
4546 ----------------------
4547
4548 1. Added a simple "pgrep" utility to the distribution.
4549
4550 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4551 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4552 where "ddd" means "one or more decimal digits".
4553
4554 3. Fixed serious bug. If a pattern had a back reference, but the call to
4555 pcre_exec() didn't supply a large enough ovector to record the related
4556 identifying subpattern, the match always failed. PCRE now remembers the number
4557 of the largest back reference, and gets some temporary memory in which to save
4558 the offsets during matching if necessary, in order to ensure that
4559 backreferences always work.
4560
4561 4. Increased the compatibility with Perl in a number of ways:
4562
4563 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4564 to request this handling. The option can be set at compile or exec time.
4565
4566 (b) $ matches before a terminating newline by default; an option
4567 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4568 mode). The option can be set at compile or exec time.
4569
4570 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4571 the same as Perl's. If the decimal number it represents is less than 10
4572 or there aren't that many previous left capturing parentheses, an octal
4573 escape is read. Inside a character class, it's always an octal escape,
4574 even if it is a single digit.
4575
4576 (d) An escaped but undefined alphabetic character is taken as a literal,
4577 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4578 escapes.
4579
4580 (e) {0} is now permitted. (The previous item is removed from the compiled
4581 pattern).
4582
4583 5. Changed all the names of code files so that the basic parts are no longer
4584 than 10 characters, and abolished the teeny "globals.c" file.
4585
4586 6. Changed the handling of character classes; they are now done with a 32-byte
4587 bit map always.
4588
4589 7. Added the -d and /D options to pcretest to make it possible to look at the
4590 internals of compilation without having to recompile pcre.
4591
4592
4593 Version 0.95 23-Sep-97
4594 ----------------------
4595
4596 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4597 \x20 at the start of a run of normal characters. These were being treated as
4598 real characters, instead of the source characters being re-checked.
4599
4600
4601 Version 0.94 18-Sep-97
4602 ----------------------
4603
4604 1. The functions are now thread-safe, with the caveat that the global variables
4605 containing pointers to malloc() and free() or alternative functions are the
4606 same for all threads.
4607
4608 2. Get pcre_study() to generate a bitmap of initial characters for non-
4609 anchored patterns when this is possible, and use it if passed to pcre_exec().
4610
4611
4612 Version 0.93 15-Sep-97
4613 ----------------------
4614
4615 1. /(b)|(:+)/ was computing an incorrect first character.
4616
4617 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4618 but not actually doing anything yet.
4619
4620 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4621 as Perl does (e.g. [-az] or [az-]).
4622
4623 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4624 all possible positions.
4625
4626 5. Split up into different modules to avoid including unneeded functions in a
4627 compiled binary. However, compile and exec are still in one module. The "study"
4628 function is split off.
4629
4630 6. The character tables are now in a separate module whose source is generated
4631 by an auxiliary program - but can then be edited by hand if required. There are
4632 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4633 toupper() in the code.
4634
4635 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4636 make them global. Abolish the function for setting them, as the caller can now
4637 set them directly.
4638
4639
4640 Version 0.92 11-Sep-97
4641 ----------------------
4642
4643 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4644 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4645
4646 2. Caseless matching was not working in character classes if the characters in
4647 the pattern were in upper case.
4648
4649 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4650
4651 4. Make PCRE_ANCHORED public and accept as a compile option.
4652
4653 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4654 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4655 pass them.
4656
4657 6. Give an error if bad option bits passed at compile or run time.
4658
4659 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4660 pcretest to cause it to pass that flag.
4661
4662 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4663 options, and the first character, if set.
4664
4665 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4666
4667
4668 Version 0.91 10-Sep-97
4669 ----------------------
4670
4671 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4672 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4673
4674 2. PCRE was looping on encountering an indefinitely repeated back reference to
4675 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4676 Perl does - treats the match as successful.
4677
4678 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5