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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 915 - (show annotations)
Tue Feb 14 13:05:39 2012 UTC (7 years, 7 months ago) by zherczeg
File size: 226396 byte(s)
Improved \X and back reference partial matching
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 5. Improved \X and back reference partial matching.
18
19
20 Version 8.30 04-February-2012
21 -----------------------------
22
23 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
24 name is defined in ctype.h.
25
26 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
27 only in quite long subpatterns.
28
29 3. Removed the function pcre_info(), which has been obsolete and deprecated
30 since it was replaced by pcre_fullinfo() in February 2000.
31
32 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
33 match a (*MARK), and the match failed at the start of the subject, a
34 reference to memory before the start of the subject could occur. This bug
35 was introduced by fix 17 of release 8.21.
36
37 5. A reference to an unset group with zero minimum repetition was giving
38 totally wrong answers (in non-JavaScript-compatibility mode). For example,
39 /(another)?(\1?)test/ matched against "hello world test". This bug was
40 introduced in release 8.13.
41
42 6. Add support for 16-bit character strings (a large amount of work involving
43 many changes and refactorings).
44
45 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
46 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
47 from a file.
48
49 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
50 rounding is not applied in this particular case).
51
52 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
53 if they appear, or are escaped, in patterns.
54
55 10. Get rid of a number of -Wunused-but-set-variable warnings.
56
57 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
58 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
59 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
60 also returns the mark "x". This bug applied to capturing parentheses,
61 non-capturing parentheses, and atomic parentheses. It also applied to some
62 assertions.
63
64 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
65 information out of configure.ac instead of relying on pcre.h.generic, which
66 is not stored in the repository.
67
68 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
69 -lreadline.
70
71 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
72
73 15. Applied Graycode's patch to put the top-level frame on the stack rather
74 than the heap when not using the stack for recursion. This gives a
75 performance improvement in many cases when recursion is not deep.
76
77 16. Experimental code added to "pcretest -C" to output the stack frame size.
78
79
80 Version 8.21 12-Dec-2011
81 ------------------------
82
83 1. Updating the JIT compiler.
84
85 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
86 are added as well.
87
88 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
89 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
90 calling _pcre_jit_exec. Some extra comments are added.
91
92 4. (*MARK) settings inside atomic groups that do not contain any capturing
93 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
94 was introduced by change 18 for 8.20.
95
96 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
97 ECMA-262 standard.
98
99 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
100 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
101 This bug was probably introduced by change 9 of 8.13.
102
103 7. While fixing 6 above, I noticed that a number of other items were being
104 incorrectly rejected as "not fixed length". This arose partly because newer
105 opcodes had not been added to the fixed-length checking code. I have (a)
106 corrected the bug and added tests for these items, and (b) arranged for an
107 error to occur if an unknown opcode is encountered while checking for fixed
108 length instead of just assuming "not fixed length". The items that were
109 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
110 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
111 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
112
113 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
114 being incorrectly compiled and would have given unpredicatble results.
115
116 9. A possessively repeated subpattern with minimum repeat count greater than
117 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
118 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
119 the first (A) could occur when it should not.
120
121 10. Add a cast and remove a redundant test from the code.
122
123 11. JIT should use pcre_malloc/pcre_free for allocation.
124
125 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
126 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
127 is anything other than /usr, -L is still shown).
128
129 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
130
131 14. Perl does not support \N without a following name in a [] class; PCRE now
132 also gives an error.
133
134 15. If a forward reference was repeated with an upper limit of around 2000,
135 it caused the error "internal error: overran compiling workspace". The
136 maximum number of forward references (including repeats) was limited by the
137 internal workspace, and dependent on the LINK_SIZE. The code has been
138 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
139 the default depends on LINK_SIZE. There is a new upper limit (for safety)
140 of around 200,000 forward references. While doing this, I also speeded up
141 the filling in of repeated forward references.
142
143 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
144 incorrectly expecting the subject to contain another "a" after the start.
145
146 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
147 in the match, the SKIP should be ignored. This was not happening; instead
148 the SKIP was being treated as NOMATCH. For patterns such as
149 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
150 tested.
151
152 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
153 now much more compatible with Perl, in particular in cases where the result
154 is a non-match for a non-anchored pattern. For example, if
155 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
156 "m", where previously it did not return a name. A side effect of this
157 change is that for partial matches, the last encountered mark name is
158 returned, as for non matches. A number of tests that were previously not
159 Perl-compatible have been moved into the Perl-compatible test files. The
160 refactoring has had the pleasing side effect of removing one argument from
161 the match() function, thus reducing its stack requirements.
162
163 19. If the /S+ option was used in pcretest to study a pattern using JIT,
164 subsequent uses of /S (without +) incorrectly behaved like /S+.
165
166 21. Retrieve executable code size support for the JIT compiler and fixing
167 some warnings.
168
169 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
170 not work when the shorter character appeared right at the end of the
171 subject string.
172
173 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
174 systems.
175
176 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
177 output it when the /M option is used in pcretest.
178
179 25. The CheckMan script was not being included in the distribution. Also, added
180 an explicit "perl" to run Perl scripts from the PrepareRelease script
181 because this is reportedly needed in Windows.
182
183 26. If study data was being save in a file and studying had not found a set of
184 "starts with" bytes for the pattern, the data written to the file (though
185 never used) was taken from uninitialized memory and so caused valgrind to
186 complain.
187
188 27. Updated RunTest.bat as provided by Sheri Pierce.
189
190 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
191
192 29. Computation of memory usage for the table of capturing group names was
193 giving an unnecessarily large value.
194
195
196 Version 8.20 21-Oct-2011
197 ------------------------
198
199 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
200 a POSIX class. After further experiments with Perl, which convinced me that
201 Perl has bugs and confusions, a closing square bracket is no longer allowed
202 in a POSIX name. This bug also affected patterns with classes that started
203 with full stops.
204
205 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
206 captured substring, but while checking the failing first alternative,
207 substring 1 is temporarily captured. If the output vector supplied to
208 pcre_exec() was not big enough for this capture, the yield of the function
209 was still zero ("insufficient space for captured substrings"). This cannot
210 be totally fixed without adding another stack variable, which seems a lot
211 of expense for a edge case. However, I have improved the situation in cases
212 such as /(a)(b)x|abc/ matched against "abc", where the return code
213 indicates that fewer than the maximum number of slots in the ovector have
214 been set.
215
216 3. Related to (2) above: when there are more back references in a pattern than
217 slots in the output vector, pcre_exec() uses temporary memory during
218 matching, and copies in the captures as far as possible afterwards. It was
219 using the entire output vector, but this conflicts with the specification
220 that only 2/3 is used for passing back captured substrings. Now it uses
221 only the first 2/3, for compatibility. This is, of course, another edge
222 case.
223
224 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
225 main code base, and can be used by building with --enable-jit. When this is
226 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
227 runtime --no-jit option is given.
228
229 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
230 ovector, the return from the function was zero, implying that there were
231 other matches that did not fit. The correct "exactly full" value is now
232 returned.
233
234 6. If a subpattern that was called recursively or as a subroutine contained
235 (*PRUNE) or any other control that caused it to give a non-standard return,
236 invalid errors such as "Error -26 (nested recursion at the same subject
237 position)" or even infinite loops could occur.
238
239 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
240 computing the minimum length on reaching *ACCEPT, and so ended up with the
241 wrong value of 1 rather than 0. Further investigation indicates that
242 computing a minimum subject length in the presence of *ACCEPT is difficult
243 (think back references, subroutine calls), and so I have changed the code
244 so that no minimum is registered for a pattern that contains *ACCEPT.
245
246 8. If (*THEN) was present in the first (true) branch of a conditional group,
247 it was not handled as intended. [But see 16 below.]
248
249 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
250 Sheri Pierce.
251
252 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
253 the first byte in a match must be "a".
254
255 11. Change 17 for 8.13 increased the recursion depth for patterns like
256 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
257 pattern contains any instances of (*THEN). If it does not, the old
258 optimizations are restored. It would be nice to do this on a per-group
259 basis, but at the moment that is not feasible.
260
261 12. In some environments, the output of pcretest -C is CRLF terminated. This
262 broke RunTest's code that checks for the link size. A single white space
263 character after the value is now allowed for.
264
265 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
266 For "fr", it uses the Windows-specific input and output files.
267
268 14. If (*THEN) appeared in a group that was called recursively or as a
269 subroutine, it did not work as intended. [But see next item.]
270
271 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
272 pattern fragments (but not containing any | characters). If A and B are
273 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
274 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
275 D. In other words, Perl considers parentheses that do not contain any |
276 characters to be part of a surrounding alternative, whereas PCRE was
277 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
278 differently. PCRE now behaves in the same way as Perl, except in the case
279 of subroutine/recursion calls such as (?1) which have in any case always
280 been different (but PCRE had them first :-).
281
282 16. Related to 15 above: Perl does not treat the | in a conditional group as
283 creating alternatives. Such a group is treated in the same way as an
284 ordinary group without any | characters when processing (*THEN). PCRE has
285 been changed to match Perl's behaviour.
286
287 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
288 RunGrepTest script failed.
289
290 18. Change 22 for version 13 caused atomic groups to use more stack. This is
291 inevitable for groups that contain captures, but it can lead to a lot of
292 stack use in large patterns. The old behaviour has been restored for atomic
293 groups that do not contain any capturing parentheses.
294
295 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
296 suppress the check for a minimum subject length at run time. (If it was
297 given to pcre_exec() or pcre_dfa_exec() it did work.)
298
299 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
300 fail to work when decoding hex characters in data strings in EBCDIC
301 environments.
302
303 21. It appears that in at least one Mac OS environment, the isxdigit() function
304 is implemented as a macro that evaluates to its argument more than once,
305 contravening the C 90 Standard (I haven't checked a later standard). There
306 was an instance in pcretest which caused it to go wrong when processing
307 \x{...} escapes in subject strings. The has been rewritten to avoid using
308 things like p++ in the argument of isxdigit().
309
310
311 Version 8.13 16-Aug-2011
312 ------------------------
313
314 1. The Unicode data tables have been updated to Unicode 6.0.0.
315
316 2. Two minor typos in pcre_internal.h have been fixed.
317
318 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
319 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
320 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
321
322 4. There were a number of related bugs in the code for matching backrefences
323 caselessly in UTF-8 mode when codes for the characters concerned were
324 different numbers of bytes. For example, U+023A and U+2C65 are an upper
325 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
326 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
327 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
328 2-byte code at the end of the subject (it thought there wasn't enough data
329 left).
330
331 5. Comprehensive information about what went wrong is now returned by
332 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
333 as the output vector has at least 2 elements. The offset of the start of
334 the failing character and a reason code are placed in the vector.
335
336 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
337 now returned is for the first byte of the failing character, instead of the
338 last byte inspected. This is an incompatible change, but I hope it is small
339 enough not to be a problem. It makes the returned offset consistent with
340 pcre_exec() and pcre_dfa_exec().
341
342 7. pcretest now gives a text phrase as well as the error number when
343 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
344 failure, the offset and reason code are output.
345
346 8. When \R was used with a maximizing quantifier it failed to skip backwards
347 over a \r\n pair if the subsequent match failed. Instead, it just skipped
348 back over a single character (\n). This seems wrong (because it treated the
349 two characters as a single entity when going forwards), conflicts with the
350 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
351 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
352 has been changed.
353
354 9. Some internal refactoring has changed the processing so that the handling
355 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
356 time (the PCRE_DOTALL option was changed this way some time ago: version
357 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
358 which was always a bit of a fudge. It also means that there is one less
359 argument for the match() function, which reduces its stack requirements
360 slightly. This change also fixes an incompatibility with Perl: the pattern
361 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
362
363 10. More internal refactoring has drastically reduced the number of recursive
364 calls to match() for possessively repeated groups such as (abc)++ when
365 using pcre_exec().
366
367 11. While implementing 10, a number of bugs in the handling of groups were
368 discovered and fixed:
369
370 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
371 (a|)*(?1) gave a compile-time internal error.
372 ((a|)+)+ did not notice that the outer group could match an empty string.
373 (^a|^)+ was not marked as anchored.
374 (.*a|.*)+ was not marked as matching at start or after a newline.
375
376 12. Yet more internal refactoring has removed another argument from the match()
377 function. Special calls to this function are now indicated by setting a
378 value in a variable in the "match data" data block.
379
380 13. Be more explicit in pcre_study() instead of relying on "default" for
381 opcodes that mean there is no starting character; this means that when new
382 ones are added and accidentally left out of pcre_study(), testing should
383 pick them up.
384
385 14. The -s option of pcretest has been documented for ages as being an old
386 synonym of -m (show memory usage). I have changed it to mean "force study
387 for every regex", that is, assume /S for every regex. This is similar to -i
388 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
389 using it. It makes it easier to run collections of tests with and without
390 study enabled, and thereby test pcre_study() more easily. All the standard
391 tests are now run with and without -s (but some patterns can be marked as
392 "never study" - see 20 below).
393
394 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
395 restoration of the capturing data to the outer values was not happening
396 correctly.
397
398 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
399 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
400 pattern had matched an empty string, and so incorrectly returned a no
401 match.
402
403 17. There was optimizing code for the last branch of non-capturing parentheses,
404 and also for the obeyed branch of a conditional subexpression, which used
405 tail recursion to cut down on stack usage. Unfortunately, now that there is
406 the possibility of (*THEN) occurring in these branches, tail recursion is
407 no longer possible because the return has to be checked for (*THEN). These
408 two optimizations have therefore been removed. [But see 8.20/11 above.]
409
410 18. If a pattern containing \R was studied, it was assumed that \R always
411 matched two bytes, thus causing the minimum subject length to be
412 incorrectly computed because \R can also match just one byte.
413
414 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
415 was incorrectly computed.
416
417 20. If /S is present twice on a test pattern in pcretest input, it now
418 *disables* studying, thereby overriding the use of -s on the command line
419 (see 14 above). This is necessary for one or two tests to keep the output
420 identical in both cases.
421
422 21. When (*ACCEPT) was used in an assertion that matched an empty string and
423 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
424
425 22. When an atomic group that contained a capturing parenthesis was
426 successfully matched, but the branch in which it appeared failed, the
427 capturing was not being forgotten if a higher numbered group was later
428 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
429 group 1 to "a", when in fact it should be unset. This applied to multi-
430 branched capturing and non-capturing groups, repeated or not, and also to
431 positive assertions (capturing in negative assertions does not happen
432 in PCRE) and also to nested atomic groups.
433
434 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
435 subject after a captured substring, to make it easier to tell which of a
436 number of identical substrings has been captured.
437
438 24. The way atomic groups are processed by pcre_exec() has been changed so that
439 if they are repeated, backtracking one repetition now resets captured
440 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
441 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
442 "aaa". Previously, it would have been "a". As part of this code
443 refactoring, the way recursive calls are handled has also been changed.
444
445 25. If an assertion condition captured any substrings, they were not passed
446 back unless some other capturing happened later. For example, if
447 (?(?=(a))a) was matched against "a", no capturing was returned.
448
449 26. When studying a pattern that contained subroutine calls or assertions,
450 the code for finding the minimum length of a possible match was handling
451 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
452 group 1 called group 2 while simultaneously a separate group 2 called group
453 1). A stack overflow occurred in this case. I have fixed this by limiting
454 the recursion depth to 10.
455
456 27. Updated RunTest.bat in the distribution to the version supplied by Tom
457 Fortmann. This supports explicit test numbers on the command line, and has
458 argument validation and error reporting.
459
460 28. An instance of \X with an unlimited repeat could fail if at any point the
461 first character it looked at was a mark character.
462
463 29. Some minor code refactoring concerning Unicode properties and scripts
464 should reduce the stack requirement of match() slightly.
465
466 30. Added the '=' option to pcretest to check the setting of unused capturing
467 slots at the end of the pattern, which are documented as being -1, but are
468 not included in the return count.
469
470 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
471 compiled something random. Now it gives a compile-time error (as does
472 Perl).
473
474 32. A *MARK encountered during the processing of a positive assertion is now
475 recorded and passed back (compatible with Perl).
476
477 33. If --only-matching or --colour was set on a pcregrep call whose pattern
478 had alternative anchored branches, the search for a second match in a line
479 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
480 matched the line "0102" twice. The same bug affected patterns that started
481 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
482 twice.
483
484 34. Previously, PCRE did not allow quantification of assertions. However, Perl
485 does, and because of capturing effects, quantifying parenthesized
486 assertions may at times be useful. Quantifiers are now allowed for
487 parenthesized assertions.
488
489 35. A minor code tidy in pcre_compile() when checking options for \R usage.
490
491 36. \g was being checked for fancy things in a character class, when it should
492 just be a literal "g".
493
494 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
495 appearance of a nested POSIX class supersedes an apparent external class.
496 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
497 unescaped square brackets may also appear as part of class names. For
498 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
499 more like Perl. (But see 8.20/1 above.)
500
501 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
502 was because it thought it was \N{name}, which is not supported).
503
504 39. Add minix to OS list not supporting the -S option in pcretest.
505
506 40. PCRE tries to detect cases of infinite recursion at compile time, but it
507 cannot analyze patterns in sufficient detail to catch mutual recursions
508 such as ((?1))((?2)). There is now a runtime test that gives an error if a
509 subgroup is called recursively as a subpattern for a second time at the
510 same position in the subject string. In previous releases this might have
511 been caught by the recursion limit, or it might have run out of stack.
512
513 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
514 happen only once. PCRE was, however incorrectly giving a compile time error
515 "recursive call could loop indefinitely" because it cannot analyze the
516 pattern in sufficient detail. The compile time test no longer happens when
517 PCRE is compiling a conditional subpattern, but actual runaway loops are
518 now caught at runtime (see 40 above).
519
520 42. It seems that Perl allows any characters other than a closing parenthesis
521 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
522 has been changed to be the same.
523
524 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
525 as not to get warnings when autogen.sh is called. Also changed
526 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
527
528 44. To help people who use pcregrep to scan files containing exceedingly long
529 lines, the following changes have been made:
530
531 (a) The default value of the buffer size parameter has been increased from
532 8K to 20K. (The actual buffer used is three times this size.)
533
534 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
535 PCRE is built.
536
537 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
538 to be set at run time.
539
540 (d) Numerical values in pcregrep options can be followed by K or M, for
541 example --buffer-size=50K.
542
543 (e) If a line being scanned overflows pcregrep's buffer, an error is now
544 given and the return code is set to 2.
545
546 45. Add a pointer to the latest mark to the callout data block.
547
548 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
549 partial match of an empty string instead of no match. This was specific to
550 the use of ".".
551
552 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
553 complete match instead of a partial match. This bug was dependent on both
554 the PCRE_UTF8 and PCRE_DOTALL options being set.
555
556 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
557 starting byte set, because \b was not being ignored.
558
559
560 Version 8.12 15-Jan-2011
561 ------------------------
562
563 1. Fixed some typos in the markup of the man pages, and wrote a script that
564 checks for such things as part of the documentation building process.
565
566 2. On a big-endian 64-bit system, pcregrep did not correctly process the
567 --match-limit and --recursion-limit options (added for 8.11). In
568 particular, this made one of the standard tests fail. (The integer value
569 went into the wrong half of a long int.)
570
571 3. If the --colour option was given to pcregrep with -v (invert match), it
572 did strange things, either producing crazy output, or crashing. It should,
573 of course, ignore a request for colour when reporting lines that do not
574 match.
575
576 4. Another pcregrep bug caused similar problems if --colour was specified with
577 -M (multiline) and the pattern match finished with a line ending.
578
579 5. In pcregrep, when a pattern that ended with a literal newline sequence was
580 matched in multiline mode, the following line was shown as part of the
581 match. This seems wrong, so I have changed it.
582
583 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
584 the check for further matches in the same line (so they could be coloured)
585 to overrun the end of the current line. If another match was found, it was
586 incorrectly shown (and then shown again when found in the next line).
587
588 7. If pcregrep was compiled under Windows, there was a reference to the
589 function pcregrep_exit() before it was defined. I am assuming this was
590 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
591 reported by a user. I've moved the definition above the reference.
592
593
594 Version 8.11 10-Dec-2010
595 ------------------------
596
597 1. (*THEN) was not working properly if there were untried alternatives prior
598 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
599 backtracked to try for "b" instead of moving to the next alternative branch
600 at the same level (in this case, to look for "c"). The Perl documentation
601 is clear that when (*THEN) is backtracked onto, it goes to the "next
602 alternative in the innermost enclosing group".
603
604 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
605 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
606 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
607 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
608 (*THEN).
609
610 3. If \s appeared in a character class, it removed the VT character from
611 the class, even if it had been included by some previous item, for example
612 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
613 of \s, but is part of the POSIX "space" class.)
614
615 4. A partial match never returns an empty string (because you can always
616 match an empty string at the end of the subject); however the checking for
617 an empty string was starting at the "start of match" point. This has been
618 changed to the "earliest inspected character" point, because the returned
619 data for a partial match starts at this character. This means that, for
620 example, /(?<=abc)def/ gives a partial match for the subject "abc"
621 (previously it gave "no match").
622
623 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
624 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
625 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
626 has an implication that the given string is incomplete (because a partial
627 match is preferred over a full match). For this reason, these items now
628 give a partial match in this situation. [Aside: previously, the one case
629 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
630 match rather than a full match, which was wrong by the old rules, but is
631 now correct.]
632
633 6. There was a bug in the handling of #-introduced comments, recognized when
634 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
635 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
636 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
637 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
638 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
639 places in pcre_compile().
640
641 7. Related to (6) above, when pcre_compile() was skipping #-introduced
642 comments when looking ahead for named forward references to subpatterns,
643 the only newline sequence it recognized was NL. It now handles newlines
644 according to the set newline convention.
645
646 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
647 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
648 cater for a lack of strerror(). These oversights have been fixed.
649
650 9. Added --match-limit and --recursion-limit to pcregrep.
651
652 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
653
654 11. When the -o option was used, pcregrep was setting a return code of 1, even
655 when matches were found, and --line-buffered was not being honoured.
656
657 12. Added an optional parentheses number to the -o and --only-matching options
658 of pcregrep.
659
660 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
661 can match an empty string. The code to do it in pcretest and pcredemo
662 needed fixing:
663
664 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
665 only one byte after an empty string match just before CRLF (this case
666 just got forgotten; "any" and "anycrlf" were OK).
667
668 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
669 mode when an empty string match preceded an ASCII character followed by
670 a non-ASCII character. (The code for advancing by one character rather
671 than one byte was nonsense.)
672
673 (c) The pcredemo.c sample program did not have any code at all to handle
674 the cases when CRLF is a valid newline sequence.
675
676 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
677 as a starting offset was within the subject string. There is now a new
678 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
679 negative or greater than the length of the string. In order to test this,
680 pcretest is extended to allow the setting of negative starting offsets.
681
682 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
683 starting offset points to the beginning of a UTF-8 character was
684 unnecessarily clumsy. I tidied it up.
685
686 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
687 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
688
689 17. Nobody had reported that the --include_dir option, which was added in
690 release 7.7 should have been called --include-dir (hyphen, not underscore)
691 for compatibility with GNU grep. I have changed it to --include-dir, but
692 left --include_dir as an undocumented synonym, and the same for
693 --exclude-dir, though that is not available in GNU grep, at least as of
694 release 2.5.4.
695
696 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
697 characters from a string of bytes) have been redefined so as not to use
698 loops, in order to improve performance in some environments. At the same
699 time, I abstracted some of the common code into auxiliary macros to save
700 repetition (this should not affect the compiled code).
701
702 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
703 compile-time error is now given if \c is not followed by an ASCII
704 character, that is, a byte less than 128. (In EBCDIC mode, the code is
705 different, and any byte value is allowed.)
706
707 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
708 START_OPTIMIZE option, which is now allowed at compile time - but just
709 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
710 to pcregrep and other applications that have no direct access to PCRE
711 options. The new /Y option in pcretest sets this option when calling
712 pcre_compile().
713
714 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
715 back references. Groups containing recursive back references were forced to
716 be atomic by that change, but in the case of named groups, the amount of
717 memory required was incorrectly computed, leading to "Failed: internal
718 error: code overflow". This has been fixed.
719
720 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
721 pcretest.c, to avoid build problems in some Borland environments.
722
723
724 Version 8.10 25-Jun-2010
725 ------------------------
726
727 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
728 THEN.
729
730 2. (*ACCEPT) was not working when inside an atomic group.
731
732 3. Inside a character class, \B is treated as a literal by default, but
733 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
734 causes the error). The code is unchanged, but I tidied the documentation.
735
736 4. Inside a character class, PCRE always treated \R and \X as literals,
737 whereas Perl faults them if its -w option is set. I have changed PCRE so
738 that it faults them when PCRE_EXTRA is set.
739
740 5. Added support for \N, which always matches any character other than
741 newline. (It is the same as "." when PCRE_DOTALL is not set.)
742
743 6. When compiling pcregrep with newer versions of gcc which may have
744 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
745 declared with attribute warn_unused_result" were given. Just casting the
746 result to (void) does not stop the warnings; a more elaborate fudge is
747 needed. I've used a macro to implement this.
748
749 7. Minor change to pcretest.c to avoid a compiler warning.
750
751 8. Added four artifical Unicode properties to help with an option to make
752 \s etc use properties (see next item). The new properties are: Xan
753 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
754
755 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
756 use Unicode properties. (*UCP) at the start of a pattern can be used to set
757 this option. Modified pcretest to add /W to test this facility. Added
758 REG_UCP to make it available via the POSIX interface.
759
760 10. Added --line-buffered to pcregrep.
761
762 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
763 studied, and the match started with a letter with a code point greater than
764 127 whose first byte was different to the first byte of the other case of
765 the letter, the other case of this starting letter was not recognized
766 (#976).
767
768 12. If a pattern that was studied started with a repeated Unicode property
769 test, for example, \p{Nd}+, there was the theoretical possibility of
770 setting up an incorrect bitmap of starting bytes, but fortunately it could
771 not have actually happened in practice until change 8 above was made (it
772 added property types that matched character-matching opcodes).
773
774 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
775 possible starting bytes for non-anchored patterns.
776
777 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
778 \R, and also a number of cases that involve Unicode properties, both
779 explicit and implicit when PCRE_UCP is set.
780
781 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
782 input, it could crash or give wrong results if characters with values
783 greater than 0xc0 were present in the subject string. (Detail: it assumed
784 UTF-8 input when processing these items.)
785
786 16. Added a lot of (int) casts to avoid compiler warnings in systems where
787 size_t is 64-bit (#991).
788
789 17. Added a check for running out of memory when PCRE is compiled with
790 --disable-stack-for-recursion (#990).
791
792 18. If the last data line in a file for pcretest does not have a newline on
793 the end, a newline was missing in the output.
794
795 19. The default pcre_chartables.c file recognizes only ASCII characters (values
796 less than 128) in its various bitmaps. However, there is a facility for
797 generating tables according to the current locale when PCRE is compiled. It
798 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
799 characters, are recognized by isspace() and therefore were getting set in
800 these tables, and indeed these tables seem to approximate to ISO 8859. This
801 caused a problem in UTF-8 mode when pcre_study() was used to create a list
802 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
803 which of course cannot start UTF-8 characters. I have changed the code so
804 that only real ASCII characters (less than 128) and the correct starting
805 bytes for UTF-8 encodings are set for characters greater than 127 when in
806 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
807 altogether.)
808
809 20. Added the /T option to pcretest so as to be able to run tests with non-
810 standard character tables, thus making it possible to include the tests
811 used for 19 above in the standard set of tests.
812
813 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
814 reference to a subpattern the other side of a comment that contains an
815 opening parenthesis caused either an internal compiling error, or a
816 reference to the wrong subpattern.
817
818
819 Version 8.02 19-Mar-2010
820 ------------------------
821
822 1. The Unicode data tables have been updated to Unicode 5.2.0.
823
824 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
825 configured.
826
827 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
828 original author of that file, following a query about its status.
829
830 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
831 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
832
833 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
834 quantifier applied to a forward-referencing subroutine call, could compile
835 incorrect code or give the error "internal error: previously-checked
836 referenced subpattern not found".
837
838 6. Both MS Visual Studio and Symbian OS have problems with initializing
839 variables to point to external functions. For these systems, therefore,
840 pcre_malloc etc. are now initialized to local functions that call the
841 relevant global functions.
842
843 7. There were two entries missing in the vectors called coptable and poptable
844 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
845 I've fixed the data, and added a kludgy way of testing at compile time that
846 the lengths are correct (equal to the number of opcodes).
847
848 8. Following on from 7, I added a similar kludge to check the length of the
849 eint vector in pcreposix.c.
850
851 9. Error texts for pcre_compile() are held as one long string to avoid too
852 much relocation at load time. To find a text, the string is searched,
853 counting zeros. There was no check for running off the end of the string,
854 which could happen if a new error number was added without updating the
855 string.
856
857 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
858
859 11. \K was not working if it appeared in an atomic group or in a group that
860 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
861 \K is "not well defined" if used in an assertion. PCRE now accepts it if
862 the assertion is positive, but not if it is negative.
863
864 12. Change 11 fortuitously reduced the size of the stack frame used in the
865 "match()" function of pcre_exec.c by one pointer. Forthcoming
866 implementation of support for (*MARK) will need an extra pointer on the
867 stack; I have reserved it now, so that the stack frame size does not
868 decrease.
869
870 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
871 item in branch that calls a recursion is a subroutine call - as in the
872 second branch in the above example - was incorrectly given the compile-
873 time error "recursive call could loop indefinitely" because pcre_compile()
874 was not correctly checking the subroutine for matching a non-empty string.
875
876 14. The checks for overrunning compiling workspace could trigger after an
877 overrun had occurred. This is a "should never occur" error, but it can be
878 triggered by pathological patterns such as hundreds of nested parentheses.
879 The checks now trigger 100 bytes before the end of the workspace.
880
881 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
882
883
884 Version 8.01 19-Jan-2010
885 ------------------------
886
887 1. If a pattern contained a conditional subpattern with only one branch (in
888 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
889 computed the wrong minimum data length (which is of course zero for such
890 subpatterns). This could cause incorrect "no match" results.
891
892 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
893 the pattern is reset in the first branch, pcre_compile() failed with
894 "internal error: code overflow at offset...". This happened only when
895 the reset was to the original external option setting. (An optimization
896 abstracts leading options settings into an external setting, which was the
897 cause of this.)
898
899 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
900 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
901 assertion pattern did not match (meaning that the assertion was true), it
902 was incorrectly treated as false if the SKIP had been reached during the
903 matching. This also applied to assertions used as conditions.
904
905 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
906 assertion subpattern, including such a pattern used as a condition,
907 unpredictable results occurred, instead of the error return
908 PCRE_ERROR_DFA_UITEM.
909
910 5. The C++ GlobalReplace function was not working like Perl for the special
911 situation when an empty string is matched. It now does the fancy magic
912 stuff that is necessary.
913
914 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
915 removed. (These were left over from very, very early versions of PCRE.)
916
917 7. Some cosmetic changes to the code to make life easier when compiling it
918 as part of something else:
919
920 (a) Change DEBUG to PCRE_DEBUG.
921
922 (b) In pcre_compile(), rename the member of the "branch_chain" structure
923 called "current" as "current_branch", to prevent a collision with the
924 Linux macro when compiled as a kernel module.
925
926 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
927 prevent a collision with the Linux macro when compiled as a kernel
928 module.
929
930 8. In pcre_compile() there are some checks for integer overflows that used to
931 cast potentially large values to (double). This has been changed to that
932 when building, a check for int64_t is made, and if it is found, it is used
933 instead, thus avoiding the use of floating point arithmetic. (There is no
934 other use of FP in PCRE.) If int64_t is not found, the fallback is to
935 double.
936
937 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
938 2005 (difference between two addresses compared to an unsigned value).
939
940 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
941 custom one, because of the following reported problem in Windows:
942
943 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
944 under Win32.
945 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
946 therefore missing the function definition.
947 - The compiler thus generates a "C" signature for the test function.
948 - The linker fails to find the "C" function.
949 - PCRE fails to configure if asked to do so against libbz2.
950
951 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
952 messages were output:
953
954 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
955 rerunning libtoolize, to keep the correct libtool macros in-tree.
956 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
957
958 I have done both of these things.
959
960 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
961 most of the time, it *can* run out if it is given a pattern that contains a
962 runaway infinite recursion. I updated the discussion in the pcrestack man
963 page.
964
965 13. Now that we have gone to the x.xx style of version numbers, the minor
966 version may start with zero. Using 08 or 09 is a bad idea because users
967 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
968 interpreted as invalid octal numbers. I've updated the previous comment in
969 configure.ac, and also added a check that gives an error if 08 or 09 are
970 used.
971
972 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
973 causing partial matching to fail when the end of the subject matched \W
974 in a UTF-8 pattern where \W was quantified with a minimum of 3.
975
976 15. There were some discrepancies between the declarations in pcre_internal.h
977 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
978 their definitions. The declarations used "const uschar *" and the
979 definitions used USPTR. Even though USPTR is normally defined as "const
980 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
981 reported that: "This difference in casting confuses some C++ compilers, for
982 example, SunCC recognizes above declarations as different functions and
983 generates broken code for hbpcre." I have changed the declarations to use
984 USPTR.
985
986 16. GNU libtool is named differently on some systems. The autogen.sh script now
987 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
988 (FreeBSD).
989
990 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
991 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
992 comment: "Figure out how to create a longlong from a string: strtoll and
993 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
994 instance, but it only takes 2 args instead of 3!"
995
996 18. A subtle bug concerned with back references has been fixed by a change of
997 specification, with a corresponding code fix. A pattern such as
998 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
999 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1000 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1001 same bug. Such groups have to be quantified to be useful, or contained
1002 inside another quantified group. (If there's no repetition, the reference
1003 can never match.) The problem arises because, having left the group and
1004 moved on to the rest of the pattern, a later failure that backtracks into
1005 the group uses the captured value from the final iteration of the group
1006 rather than the correct earlier one. I have fixed this in PCRE by forcing
1007 any group that contains a reference to itself to be an atomic group; that
1008 is, there cannot be any backtracking into it once it has completed. This is
1009 similar to recursive and subroutine calls.
1010
1011
1012 Version 8.00 19-Oct-09
1013 ----------------------
1014
1015 1. The table for translating pcre_compile() error codes into POSIX error codes
1016 was out-of-date, and there was no check on the pcre_compile() error code
1017 being within the table. This could lead to an OK return being given in
1018 error.
1019
1020 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1021 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1022 in a Windows environment.
1023
1024 3. The pcregrep --count option prints the count for each file even when it is
1025 zero, as does GNU grep. However, pcregrep was also printing all files when
1026 --files-with-matches was added. Now, when both options are given, it prints
1027 counts only for those files that have at least one match. (GNU grep just
1028 prints the file name in this circumstance, but including the count seems
1029 more useful - otherwise, why use --count?) Also ensured that the
1030 combination -clh just lists non-zero counts, with no names.
1031
1032 4. The long form of the pcregrep -F option was incorrectly implemented as
1033 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1034 but it seems right to fix it, and I didn't think it was worth preserving
1035 the old behaviour.
1036
1037 5. The command line items --regex=pattern and --regexp=pattern were not
1038 recognized by pcregrep, which required --regex pattern or --regexp pattern
1039 (with a space rather than an '='). The man page documented the '=' forms,
1040 which are compatible with GNU grep; these now work.
1041
1042 6. No libpcreposix.pc file was created for pkg-config; there was just
1043 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1044
1045 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1046 when UCP support is not needed, by modifying the Python script that
1047 generates it from Unicode data files. This should not matter if the module
1048 is correctly used as a library, but I received one complaint about 50K of
1049 unwanted data. My guess is that the person linked everything into his
1050 program rather than using a library. Anyway, it does no harm.
1051
1052 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1053 was a minimum greater than 1 for a wide character in a possessive
1054 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1055 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1056 character. Chaos in the form of incorrect output or a compiling loop could
1057 result.
1058
1059 9. The restrictions on what a pattern can contain when partial matching is
1060 requested for pcre_exec() have been removed. All patterns can now be
1061 partially matched by this function. In addition, if there are at least two
1062 slots in the offset vector, the offset of the earliest inspected character
1063 for the match and the offset of the end of the subject are set in them when
1064 PCRE_ERROR_PARTIAL is returned.
1065
1066 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1067 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1068 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1069 and may be more useful for multi-segment matching.
1070
1071 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1072 used to be given if ever the end of the subject was reached; now it is
1073 given only if matching could not proceed because another character was
1074 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1075 string "Z", which now yields "no match" instead of "partial match". In the
1076 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1077 final character ended with (*FAIL).
1078
1079 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1080 if the pattern had a "must contain" character that was already found in the
1081 earlier partial match, unless partial matching was again requested. For
1082 example, with the pattern /dog.(body)?/, the "must contain" character is
1083 "g". If the first part-match was for the string "dog", restarting with
1084 "sbody" failed. This bug has been fixed.
1085
1086 13. The string returned by pcre_dfa_exec() after a partial match has been
1087 changed so that it starts at the first inspected character rather than the
1088 first character of the match. This makes a difference only if the pattern
1089 starts with a lookbehind assertion or \b or \B (\K is not supported by
1090 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1091 matching functions compatible, and I think it's the right thing to do.
1092
1093 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1094 so that the demonstration program is easily available in environments where
1095 PCRE has not been installed from source.
1096
1097 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1098 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1099 library.
1100
1101 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1102 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1103 is not the first non-POSIX option to be added. Clearly some people find
1104 these options useful.
1105
1106 17. If a caller to the POSIX matching function regexec() passes a non-zero
1107 value for nmatch with a NULL value for pmatch, the value of
1108 nmatch is forced to zero.
1109
1110 18. RunGrepTest did not have a test for the availability of the -u option of
1111 the diff command, as RunTest does. It now checks in the same way as
1112 RunTest, and also checks for the -b option.
1113
1114 19. If an odd number of negated classes containing just a single character
1115 interposed, within parentheses, between a forward reference to a named
1116 subpattern and the definition of the subpattern, compilation crashed with
1117 an internal error, complaining that it could not find the referenced
1118 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1119 [The bug was that it was starting one character too far in when skipping
1120 over the character class, thus treating the ] as data rather than
1121 terminating the class. This meant it could skip too much.]
1122
1123 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1124 /g option in pcretest when the pattern contains \K, which makes it possible
1125 to have an empty string match not at the start, even when the pattern is
1126 anchored. Updated pcretest and pcredemo to use this option.
1127
1128 21. If the maximum number of capturing subpatterns in a recursion was greater
1129 than the maximum at the outer level, the higher number was returned, but
1130 with unset values at the outer level. The correct (outer level) value is
1131 now given.
1132
1133 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1134 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1135 make it do so. The string so far is captured, making this feature
1136 compatible with Perl.
1137
1138 23. The tests have been re-organized, adding tests 11 and 12, to make it
1139 possible to check the Perl 5.10 features against Perl 5.10.
1140
1141 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1142 pattern matches a fixed length string. PCRE did not allow this; now it
1143 does. Neither allows recursion.
1144
1145 25. I finally figured out how to implement a request to provide the minimum
1146 length of subject string that was needed in order to match a given pattern.
1147 (It was back references and recursion that I had previously got hung up
1148 on.) This code has now been added to pcre_study(); it finds a lower bound
1149 to the length of subject needed. It is not necessarily the greatest lower
1150 bound, but using it to avoid searching strings that are too short does give
1151 some useful speed-ups. The value is available to calling programs via
1152 pcre_fullinfo().
1153
1154 26. While implementing 25, I discovered to my embarrassment that pcretest had
1155 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1156 study optimizations had never been tested with that matching function.
1157 Oops. What is worse, even when it was passed study data, there was a bug in
1158 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1159 were also very few tests of studied patterns with pcre_dfa_exec().
1160
1161 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1162 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1163 on the other side of the coin, they are no longer allowed to have different
1164 names, because these cannot be distinguished in PCRE, and this has caused
1165 confusion. (This is a difference from Perl.)
1166
1167 28. When duplicate subpattern names are present (necessarily with different
1168 numbers, as required by 27 above), and a test is made by name in a
1169 conditional pattern, either for a subpattern having been matched, or for
1170 recursion in such a pattern, all the associated numbered subpatterns are
1171 tested, and the overall condition is true if the condition is true for any
1172 one of them. This is the way Perl works, and is also more like the way
1173 testing by number works.
1174
1175
1176 Version 7.9 11-Apr-09
1177 ---------------------
1178
1179 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1180 (pcretest), all targets were linked against these libraries. This included
1181 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1182 libraries. This caused unwanted dependencies to be created. This problem
1183 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1184 pcretest is linked with readline.
1185
1186 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1187 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1188 moved outside it again, because FALSE and TRUE are already defined in AIX,
1189 but BOOL is not.
1190
1191 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1192 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1193
1194 4. The pcregrep documentation said spaces were inserted as well as colons (or
1195 hyphens) following file names and line numbers when outputting matching
1196 lines. This is not true; no spaces are inserted. I have also clarified the
1197 wording for the --colour (or --color) option.
1198
1199 5. In pcregrep, when --colour was used with -o, the list of matching strings
1200 was not coloured; this is different to GNU grep, so I have changed it to be
1201 the same.
1202
1203 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1204 each matching line was coloured. Now it goes on to look for further matches
1205 of any of the test patterns, which is the same behaviour as GNU grep.
1206
1207 7. A pattern that could match an empty string could cause pcregrep to loop; it
1208 doesn't make sense to accept an empty string match in pcregrep, so I have
1209 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1210 seems to be how GNU grep behaves.
1211
1212 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1213 start or after a newline", because the conditional assertion was not being
1214 correctly handled. The rule now is that both the assertion and what follows
1215 in the first alternative must satisfy the test.
1216
1217 9. If auto-callout was enabled in a pattern with a conditional group whose
1218 condition was an assertion, PCRE could crash during matching, both with
1219 pcre_exec() and pcre_dfa_exec().
1220
1221 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1222 used for matching.
1223
1224 11. Unicode property support in character classes was not working for
1225 characters (bytes) greater than 127 when not in UTF-8 mode.
1226
1227 12. Added the -M command line option to pcretest.
1228
1229 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1230
1231 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1232
1233 16. Added comments and documentation about mis-use of no_arg in the C++
1234 wrapper.
1235
1236 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1237 from Martin Jerabek that uses macro names for all relevant character and
1238 string constants.
1239
1240 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1241 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1242 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1243 these, but not everybody uses configure.
1244
1245 19. A conditional group that had only one branch was not being correctly
1246 recognized as an item that could match an empty string. This meant that an
1247 enclosing group might also not be so recognized, causing infinite looping
1248 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1249 with the subject "ab", where knowledge that the repeated group can match
1250 nothing is needed in order to break the loop.
1251
1252 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1253 exec(), but without supplying a callout function, matching went wrong.
1254
1255 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1256 leak if the size of the offset vector was greater than 30. When the vector
1257 is smaller, the saved offsets during recursion go onto a local stack
1258 vector, but for larger vectors malloc() is used. It was failing to free
1259 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1260 error, in fact).
1261
1262 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1263 heapframe that is used only when UTF-8 support is enabled. This caused no
1264 problem, but was untidy.
1265
1266 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1267 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1268 included within another project.
1269
1270 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1271 slightly modified by me:
1272
1273 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1274 not building pcregrep.
1275
1276 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1277 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1278
1279 25. Forward references, both numeric and by name, in patterns that made use of
1280 duplicate group numbers, could behave incorrectly or give incorrect errors,
1281 because when scanning forward to find the reference group, PCRE was not
1282 taking into account the duplicate group numbers. A pattern such as
1283 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1284
1285 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1286 the feature of a custom pointer more persuasive (as requested by a user).
1287
1288 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1289 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1290 pre-defined.
1291
1292 28. Added support for (*UTF8) at the start of a pattern.
1293
1294 29. Arrange for flags added by the "release type" setting in CMake to be shown
1295 in the configuration summary.
1296
1297
1298 Version 7.8 05-Sep-08
1299 ---------------------
1300
1301 1. Replaced UCP searching code with optimized version as implemented for Ad
1302 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1303 stage table and inline lookup instead of a function, giving speed ups of 2
1304 to 5 times on some simple patterns that I tested. Permission was given to
1305 distribute the MultiStage2.py script that generates the tables (it's not in
1306 the tarball, but is in the Subversion repository).
1307
1308 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1309 scripts.
1310
1311 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1312 a group with a zero qualifier. The result of the study could be incorrect,
1313 or the function might crash, depending on the pattern.
1314
1315 4. Caseless matching was not working for non-ASCII characters in back
1316 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1317 It now works when Unicode Property Support is available.
1318
1319 5. In pcretest, an escape such as \x{de} in the data was always generating
1320 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1321 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1322 truncation.
1323
1324 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1325
1326 7. Added two (int) casts to pcregrep when printing the difference of two
1327 pointers, in case they are 64-bit values.
1328
1329 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1330 test 2 if it fails.
1331
1332 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1333 and a #define of that name to empty if it is not externally set. This is to
1334 allow users of MSVC to set it if necessary.
1335
1336 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1337 the convenience functions in the pcre_get.c source file.
1338
1339 11. An option change at the start of a pattern that had top-level alternatives
1340 could cause overwriting and/or a crash. This command provoked a crash in
1341 some environments:
1342
1343 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1344
1345 This potential security problem was recorded as CVE-2008-2371.
1346
1347 12. For a pattern where the match had to start at the beginning or immediately
1348 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1349 pcre_dfa_exec() could read past the end of the passed subject if there was
1350 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1351 pcretest so that it places the subject at the end of its malloc-ed buffer.
1352
1353 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1354 exec() might read past the end of the data buffer in UTF-8 mode.
1355
1356 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1357 the data contained the byte 0x85 as part of a UTF-8 character within its
1358 first line. This applied both to normal and DFA matching.
1359
1360 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1361 /^[^d]*?$/8 failed to match "abc".
1362
1363 16. Added a missing copyright notice to pcrecpp_internal.h.
1364
1365 17. Make it more clear in the documentation that values returned from
1366 pcre_exec() in ovector are byte offsets, not character counts.
1367
1368 18. Tidied a few places to stop certain compilers from issuing warnings.
1369
1370 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1371 supplied by Stefan Weber. I made a further small update for 7.8 because
1372 there is a change of source arrangements: the pcre_searchfuncs.c module is
1373 replaced by pcre_ucd.c.
1374
1375
1376 Version 7.7 07-May-08
1377 ---------------------
1378
1379 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1380 a string to a long long, pretend we don't even have a long long." This is
1381 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1382
1383 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1384 pre-7.6 versions, which defined a global no_arg variable instead of putting
1385 it in the RE class. (See also #8 below.)
1386
1387 3. Remove a line of dead code, identified by coverity and reported by Nuno
1388 Lopes.
1389
1390 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1391
1392 (1) The include/exclude patterns were being applied to the whole pathnames
1393 of files, instead of just to the final components.
1394
1395 (2) If there was more than one level of directory, the subdirectories were
1396 skipped unless they satisfied the include/exclude conditions. This is
1397 inconsistent with GNU grep (and could even be seen as contrary to the
1398 pcregrep specification - which I improved to make it absolutely clear).
1399 The action now is always to scan all levels of directory, and just
1400 apply the include/exclude patterns to regular files.
1401
1402 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1403 --exclude_dir in the tests to avoid scanning .svn directories.
1404
1405 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1406 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1407 doesn't support NULs in patterns.
1408
1409 7. Added some missing "const"s to declarations of static tables in
1410 pcre_compile.c and pcre_dfa_exec.c.
1411
1412 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1413 caused by fix #2 above. (Subsequently also a second patch to fix the
1414 first patch. And a third patch - this was a messy problem.)
1415
1416 9. Applied Craig's patch to remove the use of push_back().
1417
1418 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1419 matching function regexec().
1420
1421 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1422 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1423 references. PCRE supports relative numbers with this syntax (I don't think
1424 Oniguruma does).
1425
1426 12. Previously, a group with a zero repeat such as (...){0} was completely
1427 omitted from the compiled regex. However, this means that if the group
1428 was called as a subroutine from elsewhere in the pattern, things went wrong
1429 (an internal error was given). Such groups are now left in the compiled
1430 pattern, with a new opcode that causes them to be skipped at execution
1431 time.
1432
1433 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1434 to the way PCRE behaves:
1435
1436 (a) A lone ] character is dis-allowed (Perl treats it as data).
1437
1438 (b) A back reference to an unmatched subpattern matches an empty string
1439 (Perl fails the current match path).
1440
1441 (c) A data ] in a character class must be notated as \] because if the
1442 first data character in a class is ], it defines an empty class. (In
1443 Perl it is not possible to have an empty class.) The empty class []
1444 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1445 The negative empty class [^] matches any one character, independently
1446 of the DOTALL setting.
1447
1448 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1449 non-existent subpattern following a character class starting with ']' and
1450 containing () gave an internal compiling error instead of "reference to
1451 non-existent subpattern". Fortunately, when the pattern did exist, the
1452 compiled code was correct. (When scanning forwards to check for the
1453 existencd of the subpattern, it was treating the data ']' as terminating
1454 the class, so got the count wrong. When actually compiling, the reference
1455 was subsequently set up correctly.)
1456
1457 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1458 it was being rejected as not supported by pcre_dfa_exec(), even though
1459 other assertions are supported. I have made pcre_dfa_exec() support
1460 (*FAIL).
1461
1462 16. The implementation of 13c above involved the invention of a new opcode,
1463 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1464 cannot be changed at match time, I realized I could make a small
1465 improvement to matching performance by compiling OP_ALLANY instead of
1466 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1467 on the OP_ANY path.
1468
1469 17. Compiling pcretest on Windows with readline support failed without the
1470 following two fixes: (1) Make the unistd.h include conditional on
1471 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1472
1473 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1474 ncurses library to be included for pcretest when ReadLine support is
1475 requested, but also to allow for it to be overridden. This patch came from
1476 Daniel Bergström.
1477
1478 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1479 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1480 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1481 spotting this.
1482
1483
1484 Version 7.6 28-Jan-08
1485 ---------------------
1486
1487 1. A character class containing a very large number of characters with
1488 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1489 overflow.
1490
1491 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1492 HAVE_LONG_LONG is not defined.
1493
1494 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1495 bring them up to date and include new features. This patch includes:
1496
1497 - Fixed PH's badly added libz and libbz2 support.
1498 - Fixed a problem with static linking.
1499 - Added pcredemo. [But later removed - see 7 below.]
1500 - Fixed dftables problem and added an option.
1501 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1502 HAVE_LONG_LONG.
1503 - Added readline support for pcretest.
1504 - Added an listing of the option settings after cmake has run.
1505
1506 4. A user submitted a patch to Makefile that makes it easy to create
1507 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1508 Makefile.am that cause it to include this special target, without
1509 affecting anything else. Note that the same mingw target plus all
1510 the other distribution libraries and programs are now supported
1511 when configuring with CMake (see 6 below) instead of with
1512 Configure/Make.
1513
1514 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1515 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1516 exported in the Windows port". It has not yet been confirmed that the patch
1517 solves the problem, but it does no harm.
1518
1519 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1520 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1521 with CMake, and also correct the comment about stack recursion.
1522
1523 7. Remove the automatic building of pcredemo from the ./configure system and
1524 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1525 of a program that users should build themselves after PCRE is installed, so
1526 building it automatically is not really right. What is more, it gave
1527 trouble in some build environments.
1528
1529 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1530
1531
1532 Version 7.5 10-Jan-08
1533 ---------------------
1534
1535 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1536 values in parens when parsing an RE using the C++ wrapper."
1537
1538 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1539 Characters greater than 255 were excluded from the class instead of being
1540 included.
1541
1542 3. The same bug as (2) above applied to negated POSIX classes such as
1543 [:^space:].
1544
1545 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1546 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1547 I have changed it.
1548
1549 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1550 first named subpattern) and a construct such as (?&a) would reference the
1551 first named subpattern whose name started with "a" (in other words, the
1552 length check was missing). Both these problems are fixed. "Subpattern name
1553 expected" is now given for (?&) (a zero-length name), and this patch also
1554 makes it give the same error for \k'' (previously it complained that that
1555 was a reference to a non-existent subpattern).
1556
1557 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1558 this is right because (?- can be followed by option settings as well as by
1559 digits. I have, however, made the messages clearer.
1560
1561 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1562 than the number used in the conditional) now cause a compile-time error.
1563 This is actually not compatible with Perl, which accepts such patterns, but
1564 treats the conditional as always being FALSE (as PCRE used to), but it
1565 seems to me that giving a diagnostic is better.
1566
1567 8. Change "alphameric" to the more common word "alphanumeric" in comments
1568 and messages.
1569
1570 9. Fix two occurrences of "backslash" in comments that should have been
1571 "backspace".
1572
1573 10. Remove two redundant lines of code that can never be obeyed (their function
1574 was moved elsewhere).
1575
1576 11. The program that makes PCRE's Unicode character property table had a bug
1577 which caused it to generate incorrect table entries for sequences of
1578 characters that have the same character type, but are in different scripts.
1579 It amalgamated them into a single range, with the script of the first of
1580 them. In other words, some characters were in the wrong script. There were
1581 thirteen such cases, affecting characters in the following ranges:
1582
1583 U+002b0 - U+002c1
1584 U+0060c - U+0060d
1585 U+0061e - U+00612
1586 U+0064b - U+0065e
1587 U+0074d - U+0076d
1588 U+01800 - U+01805
1589 U+01d00 - U+01d77
1590 U+01d9b - U+01dbf
1591 U+0200b - U+0200f
1592 U+030fc - U+030fe
1593 U+03260 - U+0327f
1594 U+0fb46 - U+0fbb1
1595 U+10450 - U+1049d
1596
1597 12. The -o option (show only the matching part of a line) for pcregrep was not
1598 compatible with GNU grep in that, if there was more than one match in a
1599 line, it showed only the first of them. It now behaves in the same way as
1600 GNU grep.
1601
1602 13. If the -o and -v options were combined for pcregrep, it printed a blank
1603 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1604 does the same. The return code can be used to tell if there were any
1605 non-matching lines.
1606
1607 14. Added --file-offsets and --line-offsets to pcregrep.
1608
1609 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1610 infinitely looping recursion. The bug was that positive lookaheads were not
1611 being skipped when checking for a possible empty match (negative lookaheads
1612 and both kinds of lookbehind were skipped).
1613
1614 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1615 inclusion of <windows.h> to before rather than after the definition of
1616 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1617
1618 17. Specifying a possessive quantifier with a specific limit for a Unicode
1619 character property caused pcre_compile() to compile bad code, which led at
1620 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1621 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1622 caused the error; without that there was no problem.
1623
1624 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1625
1626 19. Added --enable-pcretest-libreadline.
1627
1628 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1629 RE::GlobalReplace(). As a result, the number of replacements returned was
1630 double what it should be. I removed one of the increments, but Craig sent a
1631 later patch that removed the other one (the right fix) and added unit tests
1632 that check the return values (which was not done before).
1633
1634 21. Several CMake things:
1635
1636 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1637 the names libpcre and libpcreposix, not just pcre and pcreposix.
1638
1639 (2) The above change means that pcretest and pcregrep are now correctly
1640 linked with the newly-built libraries, not previously installed ones.
1641
1642 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1643
1644 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1645 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1646 UTF-8 newline character). The key issue is that the pattern starts .*;
1647 this means that the match must be either at the beginning, or after a
1648 newline. The bug was in the code for advancing after a failed match and
1649 checking that the new position followed a newline. It was not taking
1650 account of UTF-8 characters correctly.
1651
1652 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1653 character classes. PCRE was not treating the sequence [:...:] as a
1654 character class unless the ... were all letters. Perl, however, seems to
1655 allow any characters between [: and :], though of course it rejects as
1656 unknown any "names" that contain non-letters, because all the known class
1657 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1658 for example, whereas PCRE did not - it did not recognize a POSIX character
1659 class. This seemed a bit dangerous, so the code has been changed to be
1660 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1661 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1662 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1663 Perl does, and where it didn't before.
1664
1665 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1666 Windows environments %n is disabled by default.
1667
1668
1669 Version 7.4 21-Sep-07
1670 ---------------------
1671
1672 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1673 means that a class such as [\s] counted as "explicit reference to CR or
1674 LF". That isn't really right - the whole point of the change was to try to
1675 help when there was an actual mention of one of the two characters. So now
1676 the change happens only if \r or \n (or a literal CR or LF) character is
1677 encountered.
1678
1679 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1680 of both had grown to the point where there were only 3 bits left.
1681 Fortunately, there was spare space in the data structure, and so I have
1682 moved the internal flags into a new 16-bit field to free up more option
1683 bits.
1684
1685 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1686 but did not set the internal JCHANGED flag - either of these is enough to
1687 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1688 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1689 start sets both bits.
1690
1691 4. Added options (at build time, compile time, exec time) to change \R from
1692 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1693
1694 5. doc/pcresyntax.html was missing from the distribution.
1695
1696 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1697 compatibility, even though it is no longer used.
1698
1699 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1700 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1701 windows.h file is present (where different names are used). [This was
1702 reversed later after testing - see 16 below.]
1703
1704 8. Changed all #include <config.h> to #include "config.h". There were also
1705 some further <pcre.h> cases that I changed to "pcre.h".
1706
1707 9. When pcregrep was used with the --colour option, it missed the line ending
1708 sequence off the lines that it output.
1709
1710 10. It was pointed out to me that arrays of string pointers cause lots of
1711 relocations when a shared library is dynamically loaded. A technique of
1712 using a single long string with a table of offsets can drastically reduce
1713 these. I have refactored PCRE in four places to do this. The result is
1714 dramatic:
1715
1716 Originally: 290
1717 After changing UCP table: 187
1718 After changing error message table: 43
1719 After changing table of "verbs" 36
1720 After changing table of Posix names 22
1721
1722 Thanks to the folks working on Gregex for glib for this insight.
1723
1724 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1725 unicode-properties was also set.
1726
1727 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1728
1729 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1730 checked only for CRLF.
1731
1732 14. Added casts to pcretest.c to avoid compiler warnings.
1733
1734 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1735
1736 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1737 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1738 entirely. This removes changes made in 7 above.
1739
1740 17. The CMake files have been updated, and there is now more information about
1741 building with CMake in the NON-UNIX-USE document.
1742
1743
1744 Version 7.3 28-Aug-07
1745 ---------------------
1746
1747 1. In the rejigging of the build system that eventually resulted in 7.1, the
1748 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1749 brackets there is not right, since it causes compilers to look for an
1750 installed pcre.h, not the version that is in the source that is being
1751 compiled (which of course may be different). I have changed it back to:
1752
1753 #include "pcre.h"
1754
1755 I have a vague recollection that the change was concerned with compiling in
1756 different directories, but in the new build system, that is taken care of
1757 by the VPATH setting the Makefile.
1758
1759 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1760 when the subject happened to end in the byte 0x85 (e.g. if the last
1761 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1762 characters but of course it shouldn't be taken as a newline when it is part
1763 of another character. The bug was that, for an unlimited repeat of . in
1764 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1765 characters when looking for a newline.
1766
1767 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1768
1769 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1770 in debug output.
1771
1772 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1773 long printing in the pcrecpp unittest when running under MinGW.
1774
1775 6. ESC_K was left out of the EBCDIC table.
1776
1777 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1778 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1779 limit also applies to "virtual nesting" when a pattern is recursive, and in
1780 this case 1000 isn't so big. I have been able to remove this limit at the
1781 expense of backing off one optimization in certain circumstances. Normally,
1782 when pcre_exec() would call its internal match() function recursively and
1783 immediately return the result unconditionally, it uses a "tail recursion"
1784 feature to save stack. However, when a subpattern that can match an empty
1785 string has an unlimited repetition quantifier, it no longer makes this
1786 optimization. That gives it a stack frame in which to save the data for
1787 checking that an empty string has been matched. Previously this was taken
1788 from the 1000-entry workspace that had been reserved. So now there is no
1789 explicit limit, but more stack is used.
1790
1791 8. Applied Daniel's patches to solve problems with the import/export magic
1792 syntax that is required for Windows, and which was going wrong for the
1793 pcreposix and pcrecpp parts of the library. These were overlooked when this
1794 problem was solved for the main library.
1795
1796 9. There were some crude static tests to avoid integer overflow when computing
1797 the size of patterns that contain repeated groups with explicit upper
1798 limits. As the maximum quantifier is 65535, the maximum group length was
1799 set at 30,000 so that the product of these two numbers did not overflow a
1800 32-bit integer. However, it turns out that people want to use groups that
1801 are longer than 30,000 bytes (though not repeat them that many times).
1802 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1803 made it possible to implement the integer overflow checks in a much more
1804 dynamic way, which I have now done. The artificial limitation on group
1805 length has been removed - we now have only the limit on the total length of
1806 the compiled pattern, which depends on the LINK_SIZE setting.
1807
1808 10. Fixed a bug in the documentation for get/copy named substring when
1809 duplicate names are permitted. If none of the named substrings are set, the
1810 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1811 empty string.
1812
1813 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1814 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1815 because the ] is interpreted as the first data character and the
1816 terminating ] is not found. PCRE has been made compatible with Perl in this
1817 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1818 cause memory overwriting.
1819
1820 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1821 string has been matched (to stop an infinite loop). It was not recognizing
1822 a conditional subpattern that could match an empty string if that
1823 subpattern was within another subpattern. For example, it looped when
1824 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1825 condition was not nested. This bug has been fixed.
1826
1827 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1828 past the start of the subject in the presence of bytes with the top bit
1829 set, for example "\x8aBCD".
1830
1831 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1832 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1833
1834 14. Optimized (?!) to (*FAIL).
1835
1836 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1837 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1838 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1839 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1840 does: it's just the validity check that is more restrictive.
1841
1842 16. Inserted checks for integer overflows during escape sequence (backslash)
1843 processing, and also fixed erroneous offset values for syntax errors during
1844 backslash processing.
1845
1846 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1847 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1848
1849 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1850 caused an overrun.
1851
1852 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1853 something other than just ASCII characters) inside a group that had an
1854 unlimited repeat caused a loop at compile time (while checking to see
1855 whether the group could match an empty string).
1856
1857 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1858 [\P{Any}] did so. (Error in the code for printing property names.)
1859
1860 21. An orphan \E inside a character class could cause a crash.
1861
1862 22. A repeated capturing bracket such as (A)? could cause a wild memory
1863 reference during compilation.
1864
1865 23. There are several functions in pcre_compile() that scan along a compiled
1866 expression for various reasons (e.g. to see if it's fixed length for look
1867 behind). There were bugs in these functions when a repeated \p or \P was
1868 present in the pattern. These operators have additional parameters compared
1869 with \d, etc, and these were not being taken into account when moving along
1870 the compiled data. Specifically:
1871
1872 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1873 length.
1874
1875 (b) An item such as \pL+ within a repeated group could cause crashes or
1876 loops.
1877
1878 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1879 "reference to non-existent subpattern" error.
1880
1881 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1882
1883 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1884 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1885
1886 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1887
1888 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1889 character were causing crashes (broken optimization).
1890
1891 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1892 \p or \P) caused a compile-time loop.
1893
1894 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1895 break. For example, the unstudied pattern [\r\n]A does not match the string
1896 "\r\nA" because change 7.0/46 below moves the current point on by two
1897 characters after failing to match at the start. However, the pattern \nA
1898 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1899 the same is true. There doesn't seem any very clean way out of this, but
1900 what I have chosen to do makes the common cases work: PCRE now takes note
1901 of whether there can be an explicit match for \r or \n anywhere in the
1902 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1903 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1904 pattern has explicit CR or LF references.
1905
1906 29. Added (*CR) etc for changing newline setting at start of pattern.
1907
1908
1909 Version 7.2 19-Jun-07
1910 ---------------------
1911
1912 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1913 which is apparently normally available under Windows.
1914
1915 2. Re-jig the pcregrep tests with different newline settings in an attempt
1916 to make them independent of the local environment's newline setting.
1917
1918 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1919
1920 4. Some of the "internals" tests were previously cut out when the link size
1921 was not 2, because the output contained actual offsets. The recent new
1922 "Z" feature of pcretest means that these can be cut out, making the tests
1923 usable with all link sizes.
1924
1925 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1926 stack recursion. This gives a massive performance boost under BSD, but just
1927 a small improvement under Linux. However, it saves one field in the frame
1928 in all cases.
1929
1930 6. Added more features from the forthcoming Perl 5.10:
1931
1932 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1933 recursion call. It refers to the nth most recently opened parentheses.
1934
1935 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1936 to be opened parentheses.
1937
1938 (c) Conditions that refer to capturing parentheses can be specified
1939 relatively, for example, (?(-2)... or (?(+3)...
1940
1941 (d) \K resets the start of the current match so that everything before
1942 is not part of it.
1943
1944 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1945
1946 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1947 reference syntax.
1948
1949 (g) (?| introduces a group in which the numbering of parentheses in each
1950 alternative starts with the same number.
1951
1952 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1953
1954 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1955 PCRE_INFO_JCHANGED.
1956
1957 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1958 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1959 for detecting groups that can match an empty string.
1960
1961 9. A pattern with a very large number of alternatives (more than several
1962 hundred) was running out of internal workspace during the pre-compile
1963 phase, where pcre_compile() figures out how much memory will be needed. A
1964 bit of new cunning has reduced the workspace needed for groups with
1965 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1966 workspace instead of running out of the 4096 that are available.
1967
1968 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1969
1970 11. Applied patch from Google to remove an optimization that didn't quite work.
1971 The report of the bug said:
1972
1973 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1974 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1975 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1976
1977 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1978 it matched the wrong number of bytes.
1979
1980
1981 Version 7.1 24-Apr-07
1982 ---------------------
1983
1984 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1985 that is more "standard", making use of automake and other Autotools. There
1986 is some re-arrangement of the files and adjustment of comments consequent
1987 on this.
1988
1989 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1990 for recursive directory scanning broke on some systems because the files
1991 are not scanned in any specific order and on different systems the order
1992 was different. A call to "sort" has been inserted into RunGrepTest for the
1993 approprate test as a short-term fix. In the longer term there may be an
1994 alternative.
1995
1996 3. I had an email from Eric Raymond about problems translating some of PCRE's
1997 man pages to HTML (despite the fact that I distribute HTML pages, some
1998 people do their own conversions for various reasons). The problems
1999 concerned the use of low-level troff macros .br and .in. I have therefore
2000 removed all such uses from the man pages (some were redundant, some could
2001 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2002 HTML has been updated to handle .nf/.fi and to complain if it encounters
2003 .br or .in.
2004
2005 4. Updated comments in configure.ac that get placed in config.h.in and also
2006 arranged for config.h to be included in the distribution, with the name
2007 config.h.generic, for the benefit of those who have to compile without
2008 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2009
2010 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2011 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2012 makevp.bat for the current PCRE, using the additional files
2013 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2014
2015 6. A Windows user reported a minor discrepancy with test 2, which turned out
2016 to be caused by a trailing space on an input line that had got lost in his
2017 copy. The trailing space was an accident, so I've just removed it.
2018
2019 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2020 that is needed.
2021
2022 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2023 as "const" (a) because they are and (b) because it helps the PHP
2024 maintainers who have recently made a script to detect big data structures
2025 in the php code that should be moved to the .rodata section. I remembered
2026 to update Builducptable as well, so it won't revert if ucptable.h is ever
2027 re-created.
2028
2029 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2030 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2031 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2032 support is not required. This saves 1.5-2K of code, which is important in
2033 some applications.
2034
2035 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2036 so as not to refer to the tables, even though these functions will never be
2037 called when UTF-8 support is disabled. Otherwise there are problems with a
2038 shared library.
2039
2040 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2041
2042 (a) It was defining its arguments as char * instead of void *.
2043
2044 (b) It was assuming that all moves were upwards in memory; this was true
2045 a long time ago when I wrote it, but is no longer the case.
2046
2047 The emulated memove() is provided for those environments that have neither
2048 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2049 is clearly not the case, as these two bugs were recently reported.
2050
2051 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2052 and Detrail to create the HTML documentation, the .txt form of the man
2053 pages, and it removes trailing spaces from listed files. It also creates
2054 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2055 case, it wraps all the #defines with #ifndefs. This script should be run
2056 before "make dist".
2057
2058 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2059 with Unicode property support.
2060
2061 (a) For a maximizing quantifier, if the two different cases of the
2062 character were of different lengths in their UTF-8 codings (there are
2063 some cases like this - I found 11), and the matching function had to
2064 back up over a mixture of the two cases, it incorrectly assumed they
2065 were both the same length.
2066
2067 (b) When PCRE was configured to use the heap rather than the stack for
2068 recursion during matching, it was not correctly preserving the data for
2069 the other case of a UTF-8 character when checking ahead for a match
2070 while processing a minimizing repeat. If the check also involved
2071 matching a wide character, but failed, corruption could cause an
2072 erroneous result when trying to check for a repeat of the original
2073 character.
2074
2075 13. Some tidying changes to the testing mechanism:
2076
2077 (a) The RunTest script now detects the internal link size and whether there
2078 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2079 values substituted by "configure". (The RunGrepTest script already did
2080 this for UTF-8.) The configure.ac script no longer substitutes the
2081 relevant variables.
2082
2083 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2084 with length and offset values. This means that the output is different
2085 for different internal link sizes. Test 2 is skipped for link sizes
2086 other than 2 because of this, bypassing the problem. Unfortunately,
2087 there was also a test in test 3 (the locale tests) that used /B and
2088 failed for link sizes other than 2. Rather than cut the whole test out,
2089 I have added a new /Z option to pcretest that replaces the length and
2090 offset values with spaces. This is now used to make test 3 independent
2091 of link size. (Test 2 will be tidied up later.)
2092
2093 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2094 segmentation fault instead of returning the appropriate error message.
2095
2096 15. In multiline mode when the newline sequence was set to "any", the pattern
2097 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2098 This doesn't seem right; it now treats the CRLF combination as the line
2099 ending, and so does not match in that case. It's only a pattern such as ^$
2100 that would hit this one: something like ^ABC$ would have failed after \r
2101 and then tried again after \r\n.
2102
2103 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2104 in an attempt to make files that differ only in their line terminators
2105 compare equal. This works on Linux.
2106
2107 17. Under certain error circumstances pcregrep might try to free random memory
2108 as it exited. This is now fixed, thanks to valgrind.
2109
2110 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2111 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2112 was because its rules for how to advance for /g after matching an empty
2113 string at the end of a line did not allow for this case. They now check for
2114 it specially.
2115
2116 20. pcretest is supposed to handle patterns and data of any length, by
2117 extending its buffers when necessary. It was getting this wrong when the
2118 buffer for a data line had to be extended.
2119
2120 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2121 CRLF as a newline sequence.
2122
2123 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2124 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2125 I have nevertheless tidied it up.
2126
2127 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2128
2129 24. Added a man page for pcre-config.
2130
2131
2132 Version 7.0 19-Dec-06
2133 ---------------------
2134
2135 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2136 moving to gcc 4.1.1.
2137
2138 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2139 sys/time.h, which is documented as needed for this function. It doesn't
2140 seem to matter on Linux, but it showed up on some releases of OS X.
2141
2142 3. It seems that there are systems where bytes whose values are greater than
2143 127 match isprint() in the "C" locale. The "C" locale should be the
2144 default when a C program starts up. In most systems, only ASCII printing
2145 characters match isprint(). This difference caused the output from pcretest
2146 to vary, making some of the tests fail. I have changed pcretest so that:
2147
2148 (a) When it is outputting text in the compiled version of a pattern, bytes
2149 other than 32-126 are always shown as hex escapes.
2150
2151 (b) When it is outputting text that is a matched part of a subject string,
2152 it does the same, unless a different locale has been set for the match
2153 (using the /L modifier). In this case, it uses isprint() to decide.
2154
2155 4. Fixed a major bug that caused incorrect computation of the amount of memory
2156 required for a compiled pattern when options that changed within the
2157 pattern affected the logic of the preliminary scan that determines the
2158 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2159 that the computed length was too small. The symptoms of this bug were
2160 either the PCRE error "internal error: code overflow" from pcre_compile(),
2161 or a glibc crash with a message such as "pcretest: free(): invalid next
2162 size (fast)". Examples of patterns that provoked this bug (shown in
2163 pcretest format) are:
2164
2165 /(?-x: )/x
2166 /(?x)(?-x: \s*#\s*)/
2167 /((?i)[\x{c0}])/8
2168 /(?i:[\x{c0}])/8
2169
2170 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2171 is now done differently.
2172
2173 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2174 wrapper classes; (b) implement a new function in the C++ scanner that is
2175 more efficient than the old way of doing things because it avoids levels of
2176 recursion in the regex matching; (c) add a paragraph to the documentation
2177 for the FullMatch() function.
2178
2179 6. The escape sequence \n was being treated as whatever was defined as
2180 "newline". Not only was this contrary to the documentation, which states
2181 that \n is character 10 (hex 0A), but it also went horribly wrong when
2182 "newline" was defined as CRLF. This has been fixed.
2183
2184 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2185 was being set to -1 for the "end of line" case (supposedly a value that no
2186 character can have). Though this value is never used (the check for end of
2187 line is "zero bytes in current character"), it caused compiler complaints.
2188 I've changed it to 0xffffffff.
2189
2190 8. In pcre_version.c, the version string was being built by a sequence of
2191 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2192 string (as it is for production releases) called a macro with an empty
2193 argument. The C standard says the result of this is undefined. The gcc
2194 compiler treats it as an empty string (which was what was wanted) but it is
2195 reported that Visual C gives an error. The source has been hacked around to
2196 avoid this problem.
2197
2198 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2199 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2200 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2201 of them did).
2202
2203 10. Originally, pcretest opened its input and output without "b"; then I was
2204 told that "b" was needed in some environments, so it was added for release
2205 5.0 to both the input and output. (It makes no difference on Unix-like
2206 systems.) Later I was told that it is wrong for the input on Windows. I've
2207 now abstracted the modes into two macros, to make it easier to fiddle with
2208 them, and removed "b" from the input mode under Windows.
2209
2210 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2211
2212 12. Added -help and --help to pcretest as an official way of being reminded
2213 of the options.
2214
2215 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2216 and pcrecpp.cc because they annoy compilers at high warning levels.
2217
2218 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2219
2220 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2221 scripts are not C programs :-) and which was not noticed because it works
2222 on Linux.
2223
2224 16. pcretest is supposed to handle any length of pattern and data line (as one
2225 line or as a continued sequence of lines) by extending its input buffer if
2226 necessary. This feature was broken for very long pattern lines, leading to
2227 a string of junk being passed to pcre_compile() if the pattern was longer
2228 than about 50K.
2229
2230 17. I have done a major re-factoring of the way pcre_compile() computes the
2231 amount of memory needed for a compiled pattern. Previously, there was code
2232 that made a preliminary scan of the pattern in order to do this. That was
2233 OK when PCRE was new, but as the facilities have expanded, it has become
2234 harder and harder to keep it in step with the real compile phase, and there
2235 have been a number of bugs (see for example, 4 above). I have now found a
2236 cunning way of running the real compile function in a "fake" mode that
2237 enables it to compute how much memory it would need, while actually only
2238 ever using a few hundred bytes of working memory and without too many
2239 tests of the mode. This should make future maintenance and development
2240 easier. A side effect of this work is that the limit of 200 on the nesting
2241 depth of parentheses has been removed (though this was never a serious
2242 limitation, I suspect). However, there is a downside: pcre_compile() now
2243 runs more slowly than before (30% or more, depending on the pattern). I
2244 hope this isn't a big issue. There is no effect on runtime performance.
2245
2246 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2247 newline (only possible for the last line of a file) and it was a
2248 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2249
2250 19. Added additional timing features to pcretest. (1) The -tm option now times
2251 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2252 separate command line item, by a number that specifies the number of
2253 repeats to use when timing. The default is 50000; this gives better
2254 precision, but takes uncomfortably long for very large patterns.
2255
2256 20. Extended pcre_study() to be more clever in cases where a branch of a
2257 subpattern has no definite first character. For example, (a*|b*)[cd] would
2258 previously give no result from pcre_study(). Now it recognizes that the
2259 first character must be a, b, c, or d.
2260
2261 21. There was an incorrect error "recursive call could loop indefinitely" if
2262 a subpattern (or the entire pattern) that was being tested for matching an
2263 empty string contained only one non-empty item after a nested subpattern.
2264 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2265 incorrectly, because the \d was being skipped in the check.
2266
2267 22. The pcretest program now has a new pattern option /B and a command line
2268 option -b, which is equivalent to adding /B to every pattern. This causes
2269 it to show the compiled bytecode, without the additional information that
2270 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2271 is the same as /B/I).
2272
2273 23. A new optimization is now able automatically to treat some sequences such
2274 as a*b as a*+b. More specifically, if something simple (such as a character
2275 or a simple class like \d) has an unlimited quantifier, and is followed by
2276 something that cannot possibly match the quantified thing, the quantifier
2277 is automatically "possessified".
2278
2279 24. A recursive reference to a subpattern whose number was greater than 39
2280 went wrong under certain circumstances in UTF-8 mode. This bug could also
2281 have affected the operation of pcre_study().
2282
2283 25. Realized that a little bit of performance could be had by replacing
2284 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2285
2286 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2287
2288 27. Possessive quantifiers such as a++ were previously implemented by turning
2289 them into atomic groups such as ($>a+). Now they have their own opcodes,
2290 which improves performance. This includes the automatically created ones
2291 from 23 above.
2292
2293 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2294 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2295 the first matched character to be a colon. This applied both to named and
2296 numbered groups.
2297
2298 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2299
2300 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2301 building PCRE on Windows easier, so I have included it in the distribution.
2302
2303 31. There is now a check in pcretest against a ridiculously large number being
2304 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2305 loop, the loop is abandoned.
2306
2307 32. Forward references to subpatterns in conditions such as (?(2)...) where
2308 subpattern 2 is defined later cause pcre_compile() to search forwards in
2309 the pattern for the relevant set of parentheses. This search went wrong
2310 when there were unescaped parentheses in a character class, parentheses
2311 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2312
2313 33. "Subroutine" calls and backreferences were previously restricted to
2314 referencing subpatterns earlier in the regex. This restriction has now
2315 been removed.
2316
2317 34. Added a number of extra features that are going to be in Perl 5.10. On the
2318 whole, these are just syntactic alternatives for features that PCRE had
2319 previously implemented using the Python syntax or my own invention. The
2320 other formats are all retained for compatibility.
2321
2322 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2323 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2324 also .NET compatible.
2325
2326 (b) A recursion or subroutine call to a named group can now be defined as
2327 (?&name) as well as (?P>name).
2328
2329 (c) A backreference to a named group can now be defined as \k<name> or
2330 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2331 5.10, are also .NET compatible.
2332
2333 (d) A conditional reference to a named group can now use the syntax
2334 (?(<name>) or (?('name') as well as (?(name).
2335
2336 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2337 groups (named and numbered) that are never evaluated inline, but can be
2338 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2339 is always false. There may be only one alternative in such a group.
2340
2341 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2342 as the simple (?(R). The condition is true only if the most recent
2343 recursion is that of the given number or name. It does not search out
2344 through the entire recursion stack.
2345
2346 (g) The escape \gN or \g{N} has been added, where N is a positive or
2347 negative number, specifying an absolute or relative reference.
2348
2349 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2350 some "unreachable code" warnings.
2351
2352 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2353 things, this adds five new scripts.
2354
2355 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2356 There were also incompatibilities regarding the handling of \Q..\E inside
2357 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2358 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2359
2360 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2361 matches an empty string, and forcibly breaks the loop. There were bugs in
2362 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2363 against aaaa the result was just "a" rather than "aaaa", for example. Two
2364 separate and independent bugs (that affected different cases) have been
2365 fixed.
2366
2367 39. Refactored the code to abolish the use of different opcodes for small
2368 capturing bracket numbers. This is a tidy that I avoided doing when I
2369 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2370 The new approach is not only tidier, it makes it possible to reduce the
2371 memory needed to fix the previous bug (38).
2372
2373 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2374 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2375 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2376 mode.
2377
2378 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2379 report.
2380
2381 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2382 copy construction and assignment in the C++ wrapper.
2383
2384 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2385 couple of bugs that could have given wrong results in the "--newline=crlf"
2386 case.
2387
2388 44. Added a number of casts and did some reorganization of signed/unsigned int
2389 variables following suggestions from Dair Grant. Also renamed the variable
2390 "this" as "item" because it is a C++ keyword.
2391
2392 45. Arranged for dftables to add
2393
2394 #include "pcre_internal.h"
2395
2396 to pcre_chartables.c because without it, gcc 4.x may remove the array
2397 definition from the final binary if PCRE is built into a static library and
2398 dead code stripping is activated.
2399
2400 46. For an unanchored pattern, if a match attempt fails at the start of a
2401 newline sequence, and the newline setting is CRLF or ANY, and the next two
2402 characters are CRLF, advance by two characters instead of one.
2403
2404
2405 Version 6.7 04-Jul-06
2406 ---------------------
2407
2408 1. In order to handle tests when input lines are enormously long, pcretest has
2409 been re-factored so that it automatically extends its buffers when
2410 necessary. The code is crude, but this _is_ just a test program. The
2411 default size has been increased from 32K to 50K.
2412
2413 2. The code in pcre_study() was using the value of the re argument before
2414 testing it for NULL. (Of course, in any sensible call of the function, it
2415 won't be NULL.)
2416
2417 3. The memmove() emulation function in pcre_internal.h, which is used on
2418 systems that lack both memmove() and bcopy() - that is, hardly ever -
2419 was missing a "static" storage class specifier.
2420
2421 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2422 containing an extended class (one that cannot be represented by a bitmap
2423 because it contains high-valued characters or Unicode property items, e.g.
2424 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2425 pattern, but PCRE should not loop if you do not (it no longer does).
2426 [Detail: two cases were found: (a) a repeated subpattern containing an
2427 extended class; (b) a recursive reference to a subpattern that followed a
2428 previous extended class. It wasn't skipping over the extended class
2429 correctly when UTF-8 mode was not set.]
2430
2431 5. A negated single-character class was not being recognized as fixed-length
2432 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2433 compile error "lookbehind assertion is not fixed length".
2434
2435 6. The RunPerlTest auxiliary script was showing an unexpected difference
2436 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2437 write a Perl script that can interpret lines of an input file either as
2438 byte characters or as UTF-8, which is what "perltest" was being required to
2439 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2440 can't do is switch easily at run time between having the "use utf8;" pragma
2441 or not. In the end, I fudged it by using the RunPerlTest script to insert
2442 "use utf8;" explicitly for the UTF-8 tests.
2443
2444 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2445 the end of the subject string, contrary to the documentation and to what
2446 Perl does. This was true of both matching functions. Now it matches only at
2447 the start of the subject and immediately after *internal* newlines.
2448
2449 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2450 a pointer to an int instead of a pointer to an unsigned long int. This
2451 caused problems on 64-bit systems.
2452
2453 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2454 instance of the 'standard' template library not being so standard".
2455
2456 10. There was no check on the number of named subpatterns nor the maximum
2457 length of a subpattern name. The product of these values is used to compute
2458 the size of the memory block for a compiled pattern. By supplying a very
2459 long subpattern name and a large number of named subpatterns, the size
2460 computation could be caused to overflow. This is now prevented by limiting
2461 the length of names to 32 characters, and the number of named subpatterns
2462 to 10,000.
2463
2464 11. Subpatterns that are repeated with specific counts have to be replicated in
2465 the compiled pattern. The size of memory for this was computed from the
2466 length of the subpattern and the repeat count. The latter is limited to
2467 65535, but there was no limit on the former, meaning that integer overflow
2468 could in principle occur. The compiled length of a repeated subpattern is
2469 now limited to 30,000 bytes in order to prevent this.
2470
2471 12. Added the optional facility to have named substrings with the same name.
2472
2473 13. Added the ability to use a named substring as a condition, using the
2474 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2475 are numbers (not recommended). Forward references are permitted.
2476
2477 14. Added forward references in named backreferences (if you see what I mean).
2478
2479 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2480 pattern could run off the end of the subject. For example, the pattern
2481 "(?s)(.{1,5})"8 did this with the subject "ab".
2482
2483 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2484 PCRE_CASELESS was set when matching characters that were quantified with ?
2485 or *.
2486
2487 17. A character class other than a single negated character that had a minimum
2488 but no maximum quantifier - for example [ab]{6,} - was not handled
2489 correctly by pce_dfa_exec(). It would match only one character.
2490
2491 18. A valid (though odd) pattern that looked like a POSIX character
2492 class but used an invalid character after [ (for example [[,abc,]]) caused
2493 pcre_compile() to give the error "Failed: internal error: code overflow" or
2494 in some cases to crash with a glibc free() error. This could even happen if
2495 the pattern terminated after [[ but there just happened to be a sequence of
2496 letters, a binary zero, and a closing ] in the memory that followed.
2497
2498 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2499 over the years. Originally (before any Unicode support), just the bottom 8
2500 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2501 output from "man perlunicode" includes this:
2502
2503 The regular expression compiler produces polymorphic opcodes. That
2504 is, the pattern adapts to the data and automatically switches to
2505 the Unicode character scheme when presented with Unicode data--or
2506 instead uses a traditional byte scheme when presented with byte
2507 data.
2508
2509 Sadly, a wide octal escape does not cause a switch, and in a string with
2510 no other multibyte characters, these octal escapes are treated as before.
2511 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2512 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2513 Unicode string.
2514
2515 I have not perpetrated such confusion in PCRE. Up till now, it took just
2516 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2517 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2518 translate to the appropriate multibyte character.
2519
2520 29. Applied some refactoring to reduce the number of warnings from Microsoft
2521 and Borland compilers. This has included removing the fudge introduced
2522 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2523 a warning about an unused variable.
2524
2525 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2526 characters since release 4.0, because Perl (from release 5.004) does not.
2527 [Or at least, is documented not to: some releases seem to be in conflict
2528 with the documentation.] However, when a pattern was studied with
2529 pcre_study() and all its branches started with \s, PCRE still included VT
2530 as a possible starting character. Of course, this did no harm; it just
2531 caused an unnecessary match attempt.
2532
2533 22. Removed a now-redundant internal flag bit that recorded the fact that case
2534 dependency changed within the pattern. This was once needed for "required
2535 byte" processing, but is no longer used. This recovers a now-scarce options
2536 bit. Also moved the least significant internal flag bit to the most-
2537 significant bit of the word, which was not previously used (hangover from
2538 the days when it was an int rather than a uint) to free up another bit for
2539 the future.
2540
2541 23. Added support for CRLF line endings as well as CR and LF. As well as the
2542 default being selectable at build time, it can now be changed at runtime
2543 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2544 specify that it is scanning data with non-default line endings.
2545
2546 24. Changed the definition of CXXLINK to make it agree with the definition of
2547 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2548
2549 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2550 recursions. This makes a big different to stack usage for some patterns.
2551
2552 26. If a subpattern containing a named recursion or subroutine reference such
2553 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2554 the space required for the compiled pattern went wrong and gave too small a
2555 value. Depending on the environment, this could lead to "Failed: internal
2556 error: code overflow at offset 49" or "glibc detected double free or
2557 corruption" errors.
2558
2559 27. Applied patches from Google (a) to support the new newline modes and (b) to
2560 advance over multibyte UTF-8 characters in GlobalReplace.
2561
2562 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2563 difference for some implementation of PCRE in some Windows version.
2564
2565 29. Added some extra testing facilities to pcretest:
2566
2567 \q<number> in a data line sets the "match limit" value
2568 \Q<number> in a data line sets the "match recursion limt" value
2569 -S <number> sets the stack size, where <number> is in megabytes
2570
2571 The -S option isn't available for Windows.
2572
2573
2574 Version 6.6 06-Feb-06
2575 ---------------------
2576
2577 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2578 in pcreposix.h. I have copied the definition from pcre.h.
2579
2580 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2581 because pcre.h is no longer a built file.
2582
2583 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2584 not normally included in the compiled code.
2585
2586
2587 Version 6.5 01-Feb-06
2588 ---------------------
2589
2590 1. When using the partial match feature with pcre_dfa_exec(), it was not
2591 anchoring the second and subsequent partial matches at the new starting
2592 point. This could lead to incorrect results. For example, with the pattern
2593 /1234/, partially matching against "123" and then "a4" gave a match.
2594
2595 2. Changes to pcregrep:
2596
2597 (a) All non-match returns from pcre_exec() were being treated as failures
2598 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2599 error message is output. Some extra information is given for the
2600 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2601 probably the only errors that are likely to be caused by users (by
2602 specifying a regex that has nested indefinite repeats, for instance).
2603 If there are more than 20 of these errors, pcregrep is abandoned.
2604
2605 (b) A binary zero was treated as data while matching, but terminated the
2606 output line if it was written out. This has been fixed: binary zeroes
2607 are now no different to any other data bytes.
2608
2609 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2610 used to set a locale for matching. The --locale=xxxx long option has
2611 been added (no short equivalent) to specify a locale explicitly on the
2612 pcregrep command, overriding the environment variables.
2613
2614 (d) When -B was used with -n, some line numbers in the output were one less
2615 than they should have been.
2616
2617 (e) Added the -o (--only-matching) option.
2618
2619 (f) If -A or -C was used with -c (count only), some lines of context were
2620 accidentally printed for the final match.
2621
2622 (g) Added the -H (--with-filename) option.
2623
2624 (h) The combination of options -rh failed to suppress file names for files
2625 that were found from directory arguments.
2626
2627 (i) Added the -D (--devices) and -d (--directories) options.
2628
2629 (j) Added the -F (--fixed-strings) option.
2630
2631 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2632
2633 (l) Added the --colo(u)r option.
2634
2635 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2636 is not present by default.
2637
2638 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2639 items such as (?R) or (?1), when the recursion could match a number of
2640 alternatives. If it matched one of the alternatives, but subsequently,
2641 outside the recursion, there was a failure, the code tried to back up into
2642 the recursion. However, because of the way PCRE is implemented, this is not
2643 possible, and the result was an incorrect result from the match.
2644
2645 In order to prevent this happening, the specification of recursion has
2646 been changed so that all such subpatterns are automatically treated as
2647 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2648
2649 4. I had overlooked the fact that, in some locales, there are characters for
2650 which isalpha() is true but neither isupper() nor islower() are true. In
2651 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2652 and ordfeminine) are like this. This affected the treatment of \w and \W
2653 when they appeared in character classes, but not when they appeared outside
2654 a character class. The bit map for "word" characters is now created
2655 separately from the results of isalnum() instead of just taking it from the
2656 upper, lower, and digit maps. (Plus the underscore character, of course.)
2657
2658 5. The above bug also affected the handling of POSIX character classes such as
2659 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2660 permanent tables. Instead, the bit maps for such a class were previously
2661 created as the appropriate unions of the upper, lower, and digit bitmaps.
2662 Now they are created by subtraction from the [[:word:]] class, which has
2663 its own bitmap.
2664
2665 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2666 It is created by subtracting the vertical space characters (\x09, \x0a,
2667 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2668 subtraction was done in the overall bitmap for a character class, meaning
2669 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2670 be recognized. This bug has been fixed.
2671
2672 7. Patches from the folks at Google:
2673
2674 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2675 real life, but is still worth protecting against".
2676
2677 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2678 regular expressions".
2679
2680 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2681 have it.
2682
2683 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2684 "configure" and the latter not, in order to fix a problem somebody had
2685 with compiling the Arg class on HP-UX.
2686
2687 (e) Improve the error-handling of the C++ wrapper a little bit.
2688
2689 (f) New tests for checking recursion limiting.
2690
2691 8. The pcre_memmove() function, which is used only if the environment does not
2692 have a standard memmove() function (and is therefore rarely compiled),
2693 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2694 returning a result (though PCRE never actually uses the result).
2695
2696 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2697 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2698 returned instead of calling malloc() with an overflowing number that would
2699 most likely cause subsequent chaos.
2700
2701 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2702
2703 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2704 with this option is matched, the nmatch and pmatch options of regexec() are
2705 ignored.
2706
2707 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2708 provided in case anyone wants to the the POSIX interface with UTF-8
2709 strings.
2710
2711 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2712 C++ linking (needed for some HP-UX environments).
2713
2714 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2715 (unused parameter) and in the pcre_printint() function (omitted "default"
2716 switch label when the default is to do nothing).
2717
2718 15. Added some code to make it possible, when PCRE is compiled as a C++
2719 library, to replace subject pointers for pcre_exec() with a smart pointer
2720 class, thus making it possible to process discontinuous strings.
2721
2722 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2723 much the same function. They were added by different people who were trying
2724 to make PCRE easy to compile on non-Unix systems. It has been suggested
2725 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2726 for compiling on Windows systems. I have therefore replaced it with
2727 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2728 defaults to "extern" for C or "extern C" for C++, which works fine on
2729 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2730 SCOPE with something explicit in config.h. In addition:
2731
2732 (a) pcreposix.h still had just "extern" instead of either of these macros;
2733 I have replaced it with PCRE_DATA_SCOPE.
2734
2735 (b) Functions such as _pcre_xclass(), which are internal to the library,
2736 but external in the C sense, all had PCRE_EXPORT in their definitions.
2737 This is apparently wrong for the Windows case, so I have removed it.
2738 (It makes no difference on Unix-like systems.)
2739
2740 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2741 of recursive calls to match(). This is different to MATCH_LIMIT because
2742 that limits the total number of calls to match(), not all of which increase
2743 the depth of recursion. Limiting the recursion depth limits the amount of
2744 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2745 when PCRE is compiled, and changed at run time. A patch from Google adds
2746 this functionality to the C++ interface.
2747
2748 18. Changes to the handling of Unicode character properties:
2749
2750 (a) Updated the table to Unicode 4.1.0.
2751
2752 (b) Recognize characters that are not in the table as "Cn" (undefined).
2753
2754 (c) I revised the way the table is implemented to a much improved format
2755 which includes recognition of ranges. It now supports the ranges that
2756 are defined in UnicodeData.txt, and it also amalgamates other
2757 characters into ranges. This has reduced the number of entries in the
2758 table from around 16,000 to around 3,000, thus reducing its size
2759 considerably. I realized I did not need to use a tree structure after
2760 all - a binary chop search is just as efficient. Having reduced the
2761 number of entries, I extended their size from 6 bytes to 8 bytes to
2762 allow for more data.
2763
2764 (d) Added support for Unicode script names via properties such as \p{Han}.
2765
2766 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2767 matching that character.
2768
2769 20. When matching a repeated Unicode property with a minimum greater than zero,
2770 (for example \pL{2,}), PCRE could look past the end of the subject if it
2771 reached it while seeking the minimum number of characters. This could
2772 happen only if some of the characters were more than one byte long, because
2773 there is a check for at least the minimum number of bytes.
2774
2775 21. Refactored the implementation of \p and \P so as to be more general, to
2776 allow for more different types of property in future. This has changed the
2777 compiled form incompatibly. Anybody with saved compiled patterns that use
2778 \p or \P will have to recompile them.
2779
2780 22. Added "Any" and "L&" to the supported property types.
2781
2782 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2783 but give a compile time error if the value is greater than 0xff.
2784
2785 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2786 accidentally not being installed or uninstalled.
2787
2788 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2789 made were to insert the current release number. This seemed silly, because
2790 it made things harder for people building PCRE on systems that don't run
2791 "configure". I have turned pcre.h into a distributed file, no longer built
2792 by "configure", with the version identification directly included. There is
2793 no longer a pcre.h.in file.
2794
2795 However, this change necessitated a change to the pcre-config script as
2796 well. It is built from pcre-config.in, and one of the substitutions was the
2797 release number. I have updated configure.ac so that ./configure now finds
2798 the release number by grepping pcre.h.
2799
2800 26. Added the ability to run the tests under valgrind.
2801
2802
2803 Version 6.4 05-Sep-05
2804 ---------------------
2805
2806 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2807 "--" to be printed when multiple files were scanned, even when none of the
2808 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2809 consider it to be a bug, and have restored the previous behaviour.
2810
2811 2. A couple of code tidies to get rid of compiler warnings.
2812
2813 3. The pcretest program used to cheat by referring to symbols in the library
2814 whose names begin with _pcre_. These are internal symbols that are not
2815 really supposed to be visible externally, and in some environments it is
2816 possible to suppress them. The cheating is now confined to including
2817 certain files from the library's source, which is a bit cleaner.
2818
2819 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2820 file's purpose clearer.
2821
2822 5. Reorganized pcre_ucp_findchar().
2823
2824
2825 Version 6.3 15-Aug-05
2826 ---------------------
2827
2828 1. The file libpcre.pc.in did not have general read permission in the tarball.
2829
2830 2. There were some problems when building without C++ support:
2831
2832 (a) If C++ support was not built, "make install" and "make test" still
2833 tried to test it.
2834
2835 (b) There were problems when the value of CXX was explicitly set. Some
2836 changes have been made to try to fix these, and ...
2837
2838 (c) --disable-cpp can now be used to explicitly disable C++ support.
2839
2840 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2841 backslash in a target when C++ was disabled. This confuses some
2842 versions of "make", apparently. Using an intermediate variable solves
2843 this. (Same for CPP_LOBJ.)
2844
2845 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2846 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2847 necessary on certain architectures.
2848
2849 4. Added a setting of -export-symbols-regex to the link command to remove
2850 those symbols that are exported in the C sense, but actually are local
2851 within the library, and not documented. Their names all begin with
2852 "_pcre_". This is not a perfect job, because (a) we have to except some
2853 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2854 available (and never for static libraries). I have made a note to try to
2855 find a way round (a) in the future.
2856
2857
2858 Version 6.2 01-Aug-05
2859 ---------------------
2860
2861 1. There was no test for integer overflow of quantifier values. A construction
2862 such as {1111111111111111} would give undefined results. What is worse, if
2863 a minimum quantifier for a parenthesized subpattern overflowed and became
2864 negative, the calculation of the memory size went wrong. This could have
2865 led to memory overwriting.
2866
2867 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2868
2869 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2870 operating environments where this matters.
2871
2872 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2873 PCRE options from within the C++ wrapper.
2874
2875 5. Named capturing subpatterns were not being correctly counted when a pattern
2876 was compiled. This caused two problems: (a) If there were more than 100
2877 such subpatterns, the calculation of the memory needed for the whole
2878 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2879 back references of the form \12, where the number was greater than 9, were
2880 not recognized as back references, even though there were sufficient
2881 previous subpatterns.
2882
2883 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2884 versions of gcc, e.g. 2.95.4.
2885
2886
2887 Version 6.1 21-Jun-05
2888 ---------------------
2889
2890 1. There was one reference to the variable "posix" in pcretest.c that was not
2891 surrounded by "#if !defined NOPOSIX".
2892
2893 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2894 the cross-check on the old pcre_info() function, for the benefit of the
2895 cut-down version of PCRE that is currently imported into Exim.
2896
2897 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2898 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2899 patterns that start (?i) but I don't think that matters. The use of (?i) is
2900 just an example; this all applies to the other options as well.
2901
2902 4. Since libtool seems to echo the compile commands it is issuing, the output
2903 from "make" can be reduced a bit by putting "@" in front of each libtool
2904 compile command.
2905
2906 5. Patch from the folks at Google for configure.in to be a bit more thorough
2907 in checking for a suitable C++ installation before trying to compile the
2908 C++ stuff. This should fix a reported problem when a compiler was present,
2909 but no suitable headers.
2910
2911 6. The man pages all had just "PCRE" as their title. I have changed them to
2912 be the relevant file name. I have also arranged that these names are
2913 retained in the file doc/pcre.txt, which is a concatenation in text format
2914 of all the man pages except the little individual ones for each function.
2915
2916 7. The NON-UNIX-USE file had not been updated for the different set of source
2917 files that come with release 6. I also added a few comments about the C++
2918 wrapper.
2919
2920
2921 Version 6.0 07-Jun-05
2922 ---------------------
2923
2924 1. Some minor internal re-organization to help with my DFA experiments.
2925
2926 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2927 didn't matter for the library itself when fully configured, but did matter
2928 when compiling without UCP support, or within Exim, where the ucp files are
2929 not imported.
2930
2931 3. Refactoring of the library code to split up the various functions into
2932 different source modules. The addition of the new DFA matching code (see
2933 below) to a single monolithic source would have made it really too
2934 unwieldy, quite apart from causing all the code to be include in a
2935 statically linked application, when only some functions are used. This is
2936 relevant even without the DFA addition now that patterns can be compiled in
2937 one application and matched in another.
2938
2939 The downside of splitting up is that there have to be some external
2940 functions and data tables that are used internally in different modules of
2941 the library but which are not part of the API. These have all had their
2942 names changed to start with "_pcre_" so that they are unlikely to clash
2943 with other external names.
2944
2945 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2946 a different (DFA) algorithm. Although it is slower than the original
2947 function, it does have some advantages for certain types of matching
2948 problem.
2949
2950 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2951 including restarting after a partial match.
2952
2953 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2954 defined when compiling for Windows was sent to me. I have put it into the
2955 code, though I have no means of testing or verifying it.
2956
2957 7. Added the pcre_refcount() auxiliary function.
2958
2959 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2960 match before or at the first newline in the subject string. In pcretest,
2961 the /f option on a pattern can be used to set this.
2962
2963 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2964 would behave wrongly. This has been present in PCRE since release 4.0.
2965
2966 10. A number of changes to the pcregrep command:
2967
2968 (a) Refactored how -x works; insert ^(...)$ instead of setting
2969 PCRE_ANCHORED and checking the length, in preparation for adding
2970 something similar for -w.
2971
2972 (b) Added the -w (match as a word) option.
2973
2974 (c) Refactored the way lines are read and buffered so as to have more
2975 than one at a time available.
2976
2977 (d) Implemented a pcregrep test script.
2978
2979 (e) Added the -M (multiline match) option. This allows patterns to match
2980 over several lines of the subject. The buffering ensures that at least
2981 8K, or the rest of the document (whichever is the shorter) is available
2982 for matching (and similarly the previous 8K for lookbehind assertions).
2983
2984 (f) Changed the --help output so that it now says
2985
2986 -w, --word-regex(p)
2987
2988 instead of two lines, one with "regex" and the other with "regexp"
2989 because that confused at least one person since the short forms are the
2990 same. (This required a bit of code, as the output is generated
2991 automatically from a table. It wasn't just a text change.)
2992
2993 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2994 option but starts with a hyphen. Could be a pattern or a path name
2995 starting with a hyphen, for instance.
2996
2997 (h) "-" can be given as a file name to represent stdin.
2998
2999 (i) When file names are being printed, "(standard input)" is used for
3000 the standard input, for compatibility with GNU grep. Previously
3001 "<stdin>" was used.
3002
3003 (j) The option --label=xxx can be used to supply a name to be used for
3004 stdin when file names are being printed. There is no short form.
3005
3006 (k) Re-factored the options decoding logic because we are going to add
3007 two more options that take data. Such options can now be given in four
3008 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3009
3010 (l) Added the -A, -B, and -C options for requesting that lines of context
3011 around matches be printed.
3012
3013 (m) Added the -L option to print the names of files that do not contain
3014 any matching lines, that is, the complement of -l.
3015
3016 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3017 continue to scan other files.
3018
3019 (o) The -s option was incorrectly implemented. For compatibility with other
3020 greps, it now suppresses the error message for a non-existent or non-
3021 accessible file (but not the return code). There is a new option called
3022 -q that suppresses the output of matching lines, which was what -s was
3023 previously doing.
3024
3025 (p) Added --include and --exclude options to specify files for inclusion
3026 and exclusion when recursing.
3027
3028 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3029 Hopefully, it now does.
3030
3031 12. Missing cast in pcre_study().
3032
3033 13. Added an "uninstall" target to the makefile.
3034
3035 14. Replaced "extern" in the function prototypes in Makefile.in with
3036 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3037 world, but is set differently for Windows.
3038
3039 15. Added a second compiling function called pcre_compile2(). The only
3040 difference is that it has an extra argument, which is a pointer to an
3041 integer error code. When there is a compile-time failure, this is set
3042 non-zero, in addition to the error test pointer being set to point to an
3043 error message. The new argument may be NULL if no error number is required
3044 (but then you may as well call pcre_compile(), which is now just a
3045 wrapper). This facility is provided because some applications need a
3046 numeric error indication, but it has also enabled me to tidy up the way
3047 compile-time errors are handled in the POSIX wrapper.
3048
3049 16. Added VPATH=.libs to the makefile; this should help when building with one
3050 prefix path and installing with another. (Or so I'm told by someone who
3051 knows more about this stuff than I do.)
3052
3053 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3054 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3055 match everything, including newlines. This is not POSIX-compatible, but
3056 somebody wanted the feature. From pcretest it can be activated by using
3057 both the P and the s flags.
3058
3059 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3060
3061 19. libpcre.pc was being incorrectly installed as executable.
3062
3063 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3064 it now also looks for '\r' so that it will work unmodified on Windows.
3065
3066 21. Added Google's contributed C++ wrapper to the distribution.
3067
3068 22. Added some untidy missing memory free() calls in pcretest, to keep
3069 Electric Fence happy when testing.
3070
3071
3072
3073 Version 5.0 13-Sep-04
3074 ---------------------
3075
3076 1. Internal change: literal characters are no longer packed up into items
3077 containing multiple characters in a single byte-string. Each character
3078 is now matched using a separate opcode. However, there may be more than one
3079 byte in the character in UTF-8 mode.
3080
3081 2. The pcre_callout_block structure has two new fields: pattern_position and
3082 next_item_length. These contain the offset in the pattern to the next match
3083 item, and its length, respectively.
3084
3085 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3086 insertion of callouts before each pattern item. Added the /C option to
3087 pcretest to make use of this.
3088
3089 4. On the advice of a Windows user, the lines
3090
3091 #if defined(_WIN32) || defined(WIN32)
3092 _setmode( _fileno( stdout ), 0x8000 );
3093 #endif /* defined(_WIN32) || defined(WIN32) */
3094
3095 have been added to the source of pcretest. This apparently does useful
3096 magic in relation to line terminators.
3097
3098 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3099 for the benefit of those environments where the "b" makes a difference.
3100
3101 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3102 to know about it. I have put a hack into configure.in that adds in code
3103 to set GCC=yes if CC=icc. This seems to end up at a point in the
3104 generated configure script that is early enough to affect the setting of
3105 compiler options, which is what is needed, but I have no means of testing
3106 whether it really works. (The user who reported this had patched the
3107 generated configure script, which of course I cannot do.)
3108
3109 LATER: After change 22 below (new libtool files), the configure script
3110 seems to know about icc (and also ecc). Therefore, I have commented out
3111 this hack in configure.in.
3112
3113 7. Added support for pkg-config (2 patches were sent in).
3114
3115 8. Negated POSIX character classes that used a combination of internal tables
3116 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3117 [[:^ascii]]. Typically, they would match almost any characters. The other
3118 POSIX classes were not broken in this way.
3119
3120 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3121 to find the match, as PCRE was deluded into thinking that the match had to
3122 start at the start point or following a newline. The same bug applied to
3123 patterns with negative forward assertions or any backward assertions
3124 preceding ".*" at the start, unless the pattern required a fixed first
3125 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3126
3127 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3128 starting at the last subject character, bytes beyond the end of the subject
3129 string were read.
3130
3131 11. Renamed the variable "class" as "classbits" to make life easier for C++
3132 users. (Previously there was a macro definition, but it apparently wasn't
3133 enough.)
3134
3135 12. Added the new field "tables" to the extra data so that tables can be passed
3136 in at exec time, or the internal tables can be re-selected. This allows
3137 a compiled regex to be saved and re-used at a later time by a different
3138 program that might have everything at different addresses.
3139
3140 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3141 -R library as well as a -L library.
3142
3143 14. The debugging options of pcretest (-d on the command line or D on a
3144 pattern) showed incorrect output for anything following an extended class
3145 that contained multibyte characters and which was followed by a quantifier.
3146
3147 15. Added optional support for general category Unicode character properties
3148 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3149 support. It adds about 90K to the size of the library. The meanings of the
3150 inbuilt class escapes such as \d and \s have NOT been changed.
3151
3152 16. Updated pcredemo.c to include calls to free() to release the memory for the
3153 compiled pattern.
3154
3155 17. The generated file chartables.c was being created in the source directory
3156 instead of in the building directory. This caused the build to fail if the
3157 source directory was different from the building directory, and was
3158 read-only.
3159
3160 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3161 file. No doubt somebody will tell me if they don't make sense... Also added
3162 Dan Mooney's comments about building on OpenVMS.
3163
3164 19. Added support for partial matching via the PCRE_PARTIAL option for
3165 pcre_exec() and the \P data escape in pcretest.
3166
3167 20. Extended pcretest with 3 new pattern features:
3168
3169 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3170 write the compiled pattern to the file whose name is "rest-of-line".
3171 This is a straight binary dump of the data, with the saved pointer to
3172 the character tables forced to be NULL. The study data, if any, is
3173 written too. After writing, pcretest reads a new pattern.
3174
3175 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3176 compiled pattern from the given file. There must not be any
3177 occurrences of "<" in the file name (pretty unlikely); if there are,
3178 pcretest will instead treat the initial "<" as a pattern delimiter.
3179 After reading in the pattern, pcretest goes on to read data lines as
3180 usual.
3181
3182 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3183 and 16-bit fields in a compiled pattern, to simulate a pattern that
3184 was compiled on a host of opposite endianness.
3185
3186 21. The pcre-exec() function can now cope with patterns that were compiled on
3187 hosts of opposite endianness, with this restriction:
3188
3189 As for any compiled expression that is saved and used later, the tables
3190 pointer field cannot be preserved; the extra_data field in the arguments
3191 to pcre_exec() should be used to pass in a tables address if a value
3192 other than the default internal tables were used at compile time.
3193
3194 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3195 now diagnosed as an error. Previously, most of the time, a negative number
3196 would have been treated as zero, but if in addition "ovector" was passed as
3197 NULL, a crash could occur.
3198
3199 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3200 new versions from the libtool 1.5 distribution (the last one is a copy of
3201 a file called libtool.m4). This seems to have fixed the need to patch
3202 "configure" to support Darwin 1.3 (which I used to do). However, I still
3203 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3204 workstation).
3205
3206 24. Changed the PCRE licence to be the more standard "BSD" licence.
3207
3208
3209 Version 4.5 01-Dec-03
3210 ---------------------
3211
3212 1. There has been some re-arrangement of the code for the match() function so
3213 that it can be compiled in a version that does not call itself recursively.
3214 Instead, it keeps those local variables that need separate instances for
3215 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3216 needs to "recurse". Keeping track of where control must go is done by means
3217 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3218 hide most of the details from the main code, and operates only if
3219 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3220 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3221 operating.
3222
3223 To make it easier for callers to provide specially tailored get/free
3224 functions for this usage, two new functions, pcre_stack_malloc, and
3225 pcre_stack_free, are used. They are always called in strict stacking order,
3226 and the size of block requested is always the same.
3227
3228 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3229 PCRE has been compiled to use the stack or the heap for recursion. The
3230 -C option of pcretest uses this to show which version is compiled.
3231
3232 A new data escape \S, is added to pcretest; it causes the amounts of store
3233 obtained and freed by both kinds of malloc/free at match time to be added
3234 to the output.
3235
3236 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3237 what's available on my current Linux desktop machine.
3238
3239 3. When matching a UTF-8 string, the test for a valid string at the start has
3240 been extended. If start_offset is not zero, PCRE now checks that it points
3241 to a byte that is the start of a UTF-8 character. If not, it returns
3242 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3243 this is necessary because there may be backward assertions in the pattern.
3244 When matching the same subject several times, it may save resources to use
3245 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3246
3247 4. The code for checking the validity of UTF-8 strings has been tightened so
3248 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3249 containing "overlong sequences".
3250
3251 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3252 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3253 should have been "&", but it just so happened that all the cases this let
3254 through by mistake were picked up later in the function.
3255
3256 6. I had used a variable called "isblank" - this is a C99 function, causing
3257 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3258
3259 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3260 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3261 so the version is shown; (c) stop "make test" echoing "./RunTest".
3262
3263 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3264
3265 9. The prototype for memmove() for systems that don't have it was using
3266 size_t, but the inclusion of the header that defines size_t was later. I've
3267 moved the #includes for the C headers earlier to avoid this.
3268
3269 10. Added some adjustments to the code to make it easier to compiler on certain
3270 special systems:
3271
3272 (a) Some "const" qualifiers were missing.
3273 (b) Added the macro EXPORT before all exported functions; by default this
3274 is defined to be empty.
3275 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3276 that it reads its output file name as an argument instead of writing
3277 to the standard output and assuming this can be redirected.
3278
3279 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3280 class containing characters with values greater than 255, PCRE compilation
3281 went into a loop.
3282
3283 12. A recursive reference to a subpattern that was within another subpattern
3284 that had a minimum quantifier of zero caused PCRE to crash. For example,
3285 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3286 recursion. If the recursively-called subpattern itself had a zero repeat,
3287 that was OK.
3288
3289 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3290 buffer into which it was copied (for escape processing) was still set at
3291 1024, so long lines caused crashes.
3292
3293 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3294 "internal error: code overflow...". This applied to any character class
3295 that was followed by a possessive quantifier.
3296
3297 15. Modified the Makefile to add libpcre.la as a prerequisite for
3298 libpcreposix.la because I was told this is needed for a parallel build to
3299 work.
3300
3301 16. If a pattern that contained .* following optional items at the start was
3302 studied, the wrong optimizing data was generated, leading to matching
3303 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3304 matching string must start with a or b or c. The correct conclusion for
3305 this pattern is that a match can start with any character.
3306
3307
3308 Version 4.4 13-Aug-03
3309 ---------------------
3310
3311 1. In UTF-8 mode, a character class containing characters with values between
3312 127 and 255 was not handled correctly if the compiled pattern was studied.
3313 In fixing this, I have also improved the studying algorithm for such
3314 classes (slightly).
3315
3316 2. Three internal functions had redundant arguments passed to them. Removal
3317 might give a very teeny performance improvement.
3318
3319 3. Documentation bug: the value of the capture_top field in a callout is *one
3320 more than* the number of the hightest numbered captured substring.
3321
3322 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3323 in incorrectly linking with a previously installed version. They now link
3324 explicitly with libpcre.la.
3325
3326 5. configure.in no longer needs to recognize Cygwin specially.
3327
3328 6. A problem in pcre.in for Windows platforms is fixed.
3329
3330 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3331 pcretest, it used to include the size of the study block as part of its
3332 output. Unfortunately, the structure contains a field that has a different
3333 size on different hardware architectures. This meant that the tests that
3334 showed this size failed. As the block is currently always of a fixed size,
3335 this information isn't actually particularly useful in pcretest output, so
3336 I have just removed it.
3337
3338 8. Three pre-processor statements accidentally did not start in column 1.
3339 Sadly, there are *still* compilers around that complain, even though
3340 standard C has not required this for well over a decade. Sigh.
3341
3342 9. In pcretest, the code for checking callouts passed small integers in the
3343 callout_data field, which is a void * field. However, some picky compilers
3344 complained about the casts involved for this on 64-bit systems. Now
3345 pcretest passes the address of the small integer instead, which should get
3346 rid of the warnings.
3347
3348 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3349 both compile and run time, and gives an error if an invalid UTF-8 sequence
3350 is found. There is a option for disabling this check in cases where the
3351 string is known to be correct and/or the maximum performance is wanted.
3352
3353 11. In response to a bug report, I changed one line in Makefile.in from
3354
3355 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3356 to
3357 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3358
3359 to look similar to other lines, but I have no way of telling whether this
3360 is the right thing to do, as I do not use Windows. No doubt I'll get told
3361 if it's wrong...
3362
3363
3364 Version 4.3 21-May-03
3365 ---------------------
3366
3367 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3368 Makefile.
3369
3370 2. Some refactoring to improve the quality of the code:
3371
3372 (i) The utf8_table... variables are now declared "const".
3373
3374 (ii) The code for \cx, which used the "case flipping" table to upper case
3375 lower case letters, now just substracts 32. This is ASCII-specific,
3376 but the whole concept of \cx is ASCII-specific, so it seems
3377 reasonable.
3378
3379 (iii) PCRE was using its character types table to recognize decimal and
3380 hexadecimal digits in the pattern. This is silly, because it handles
3381 only 0-9, a-f, and A-F, but the character types table is locale-
3382 specific, which means strange things might happen. A private
3383 table is now used for this - though it costs 256 bytes, a table is
3384 much faster than multiple explicit tests. Of course, the standard
3385 character types table is still used for matching digits in subject
3386 strings against \d.
3387
3388 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3389 ending in _t are). So I've renamed it as ESC_tee.
3390
3391 3. The first argument for regexec() in the POSIX wrapper should have been
3392 defined as "const".
3393
3394 4. Changed pcretest to use malloc() for its buffers so that they can be
3395 Electric Fenced for debugging.
3396
3397 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3398 to read one or more bytes before the start of the subject string. Often this
3399 had no effect on PCRE's behaviour, but in some circumstances it could
3400 provoke a segmentation fault.
3401
3402 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3403 to try to read one or more bytes before the start of the subject string.
3404
3405 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3406 UTF-8 support could misbehave in various ways if the subject string
3407 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3408 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3409 back over UTF-8 characters.)
3410
3411
3412 Version 4.2 14-Apr-03
3413 ---------------------
3414
3415 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3416
3417 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3418 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3419 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3420 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3421 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3422 and BUILD_EXEEXT
3423 Note: automatic setting of the BUILD variables is not yet working
3424 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3425 compile-time but not at link-time
3426 [LINK]: use for linking executables only
3427 make different versions for Windows and non-Windows
3428 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3429 libraries
3430 [LINK_FOR_BUILD]: new variable
3431 [OBJEXT]: use throughout
3432 [EXEEXT]: use throughout
3433 <winshared>: new target
3434 <wininstall>: new target
3435 <dftables.o>: use native compiler
3436 <dftables>: use native linker
3437 <install>: handle Windows platform correctly
3438 <clean>: ditto
3439 <check>: ditto
3440 copy DLL to top builddir before testing
3441
3442 As part of these changes, -no-undefined was removed again. This was reported
3443 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3444 in any case.
3445
3446 3. Some tidies to get rid of compiler warnings:
3447
3448 . In the match_data structure, match_limit was an unsigned long int, whereas
3449 match_call_count was an int. I've made them both unsigned long ints.
3450
3451 . In pcretest the fact that a const uschar * doesn't automatically cast to
3452 a void * provoked a warning.
3453
3454 . Turning on some more compiler warnings threw up some "shadow" variables
3455 and a few more missing casts.
3456
3457 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3458 option, a class that contained a single character with a value between 128
3459 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3460
3461 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3462 option, a class that contained several characters, but with at least one
3463 whose value was between 128 and 255 caused PCRE to crash.
3464
3465
3466 Version 4.1 12-Mar-03
3467 ---------------------
3468
3469 1. Compiling with gcc -pedantic found a couple of places where casts were
3470 needed, and a string in dftables.c that was longer than standard compilers are
3471 required to support.
3472
3473 2. Compiling with Sun's compiler found a few more places where the code could
3474 be tidied up in order to avoid warnings.
3475
3476 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3477 first of these names is deprecated in the latest Autoconf in favour of the name
3478 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3479 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3480 analogy I have changed it to CFLAGS_FOR_BUILD.
3481
3482 4. Added -no-undefined to the linking command in the Makefile, because this is
3483 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3484 linking step for the pcreposix library.
3485
3486 5. PCRE was failing to diagnose the case of two named groups with the same
3487 name.
3488
3489 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3490 literal character that is needed in the subject for a match, and scans along to
3491 ensure that it is present before embarking on the full matching process. This
3492 saves time in cases of nested unlimited repeats that are never going to match.
3493 Problem: the scan can take a lot of time if the subject is very long (e.g.
3494 megabytes), thus penalizing straightforward matches. It is now done only if the
3495 amount of subject to be scanned is less than 1000 bytes.
3496
3497 7. A lesser problem with the same optimization is that it was recording the
3498 first character of an anchored pattern as "needed", thus provoking a search
3499 right along the subject, even when the first match of the pattern was going to
3500 fail. The "needed" character is now not set for anchored patterns, unless it
3501 follows something in the pattern that is of non-fixed length. Thus, it still
3502 fulfils its original purpose of finding quick non-matches in cases of nested
3503 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3504
3505
3506 Version 4.0 17-Feb-03
3507 ---------------------
3508
3509 1. If a comment in an extended regex that started immediately after a meta-item
3510 extended to the end of string, PCRE compiled incorrect data. This could lead to
3511 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3512
3513 2. Moved to autoconf 2.53 and libtool 1.4.2.
3514
3515 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3516 the special perltest8 script is no longer needed - all the tests can be run
3517 from a single perltest script.
3518
3519 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3520 by \s. It has now been removed in PCRE. This means it isn't recognized as
3521 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3522 class [:space:] *does* include VT, thereby creating a mess.
3523
3524 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3525 space and tab.
3526
3527 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3528 its new features into the main test script, reducing the number of scripts.
3529
3530 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3531 were backward compatible, and made the (?i) apply to the whole pattern, as if
3532 /i were given. Now it behaves more logically, and applies the option setting
3533 only to what follows. PCRE has been changed to follow suit. However, if it
3534 finds options settings right at the start of the pattern, it extracts them into
3535 the global options, as before. Thus, they show up in the info data.
3536
3537 8. Added support for the \Q...\E escape sequence. Characters in between are
3538 treated as literals. This is slightly different from Perl in that $ and @ are
3539 also handled as literals inside the quotes. In Perl, they will cause variable
3540 interpolation. Note the following examples:
3541
3542 Pattern PCRE matches Perl matches
3543
3544 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3545 \Qabc\$xyz\E abc\$xyz abc\$xyz
3546 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3547
3548 For compatibility with Perl, \Q...\E sequences are recognized inside character
3549 classes as well as outside them.
3550
3551 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3552 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3553 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3554 signed/unsigned warnings.
3555
3556 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3557 option for pcretest, so I've replaced it by a simple function that does just
3558 that job.
3559
3560 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3561 "pcregrep -".
3562
3563 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3564 Java package. This provides some syntactic sugar for simple cases of what my
3565 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3566 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3567 item, you can use this simplified notation. Note that only makes sense with
3568 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3569 greediness, whatever the setting of the PCRE_UNGREEDY option.
3570
3571 13. A change of greediness default within a pattern was not taking effect at
3572 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3573 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3574 was abstracted outside.
3575
3576 14. PCRE now supports the \G assertion. It is true when the current matching
3577 position is at the start point of the match. This differs from \A when the
3578 starting offset is non-zero. Used with the /g option of pcretest (or similar
3579 code), it works in the same way as it does for Perl's /g option. If all
3580 alternatives of a regex begin with \G, the expression is anchored to the start
3581 match position, and the "anchored" flag is set in the compiled expression.
3582
3583 15. Some bugs concerning the handling of certain option changes within patterns
3584 have been fixed. These applied to options other than (?ims). For example,
3585 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3586 been the other way round. Some of this was related to change 7 above.
3587
3588 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3589 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3590 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3591 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3592
3593 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3594 mode. Unlike ".", it always matches newline, whatever the setting of
3595 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3596 assertions. Perl allows it, but it doesn't (in general) work because it can't
3597 calculate the length of the lookbehind. At least, that's the case for Perl
3598 5.8.0 - I've been told they are going to document that it doesn't work in
3599 future.
3600
3601 18. Added an error diagnosis for escapes that PCRE does not support: these are
3602 \L, \l, \N, \P, \p, \U, \u, and \X.
3603
3604 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3605 reading past the end of the pattern in cases such as /[abcd/.
3606
3607 20. PCRE was getting more memory than necessary for patterns with classes that
3608 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3609
3610 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3611 compiling PCRE for use with Virtual Pascal.
3612
3613 22. Small fix to the Makefile to make it work properly if the build is done
3614 outside the source tree.
3615
3616 23. Added a new extension: a condition to go with recursion. If a conditional
3617 subpattern starts with (?(R) the "true" branch is used if recursion has
3618 happened, whereas the "false" branch is used only at the top level.
3619
3620 24. When there was a very long string of literal characters (over 255 bytes
3621 without UTF support, over 250 bytes with UTF support), the computation of how
3622 much memory was required could be incorrect, leading to segfaults or other
3623 strange effects.
3624
3625 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3626 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3627 there was a subsequent back reference to those brackets. This meant that, for
3628 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3629 possible to check for precisely this case. All we can do is abandon the
3630 optimization if .* occurs inside capturing brackets when there are any back
3631 references whatsoever. (See below for a better fix that came later.)
3632
3633 26. The handling of the optimization for finding the first character of a
3634 non-anchored pattern, and for finding a character that is required later in the
3635 match were failing in some cases. This didn't break the matching; it just
3636 failed to optimize when it could. The way this is done has been re-implemented.
3637
3638 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3639
3640 28. Added a new feature that provides some of the functionality that Perl
3641 provides with (?{...}). The facility is termed a "callout". The way it is done
3642 in PCRE is for the caller to provide an optional function, by setting
3643 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3644 global variable. By default it is unset, which disables all calling out. To get
3645 the function called, the regex must include (?C) at appropriate points. This
3646 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3647 This provides a means of identifying different callout points. When PCRE
3648 reaches such a point in the regex, if pcre_callout has been set, the external
3649 function is called. It is provided with data in a structure called
3650 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3651 matching continues; if it returns a non-zero value, the match at the current
3652 point fails. However, backtracking will occur if possible. [This was changed
3653 later and other features added - see item 49 below.]
3654
3655 29. pcretest is upgraded to test the callout functionality. It provides a
3656 callout function that displays information. By default, it shows the start of
3657 the match and the current position in the text. There are some new data escapes
3658 to vary what happens:
3659
3660 \C+ in addition, show current contents of captured substrings
3661 \C- do not supply a callout function
3662 \C!n return 1 when callout number n is reached
3663 \C!n!m return 1 when callout number n is reached for the mth time
3664
3665 30. If pcregrep was called with the -l option and just a single file name, it
3666 output "<stdin>" if a match was found, instead of the file name.
3667
3668 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3669 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3670 pcre_exec(). This saves a malloc/free per call. The default value of
3671 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3672 when configuring.
3673
3674 32. The default maximum size of a compiled pattern is 64K. There have been a
3675 few cases of people hitting this limit. The code now uses macros to handle the
3676 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3677 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3678 configuring. Tests 2 and 5 work only with 2-byte links because they output
3679 debugging information about compiled patterns.
3680
3681 33. Internal code re-arrangements:
3682
3683 (a) Moved the debugging function for printing out a compiled regex into
3684 its own source file (printint.c) and used #include to pull it into
3685 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3686 separate copies.
3687
3688 (b) Defined the list of op-code names for debugging as a macro in
3689 internal.h so that it is next to the definition of the opcodes.
3690
3691 (c) Defined a table of op-code lengths for simpler skipping along compiled
3692 code. This is again a macro in internal.h so that it is next to the
3693 definition of the opcodes.
3694
3695 34. Added support for recursive calls to individual subpatterns, along the
3696 lines of Robin Houston's patch (but implemented somewhat differently).
3697
3698 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3699 allow it to read and process whole directories in Win32. This code was
3700 contributed by Lionel Fourquaux; it has not been tested by me.
3701
3702 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3703 used to name a group. Names consist of alphanumerics and underscores, and must
3704 be unique. Back references use the syntax (?P=name) and recursive calls use
3705 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3706 numbers. The function pcre_fullinfo() can be used after compilation to extract
3707 a name/number map. There are three relevant calls:
3708
3709 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3710 PCRE_INFO_NAMECOUNT yields the number of entries
3711 PCRE_INFO_NAMETABLE yields a pointer to the map.
3712
3713 The map is a vector of fixed-size entries. The size of each entry depends on
3714 the length of the longest name used. The first two bytes of each entry are the
3715 group number, most significant byte first. There follows the corresponding
3716 name, zero terminated. The names are in alphabetical order.
3717
3718 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3719 case instead of 255. Making it the same both with and without UTF-8 support
3720 means that the same test output works with both.
3721
3722 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3723 calling malloc() with a zero argument.
3724
3725 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3726 optimization. I've improved things by keeping a bitmap of backreferences with
3727 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3728 fact referenced, the optimization can be applied. It is unlikely that a
3729 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3730 the match to follow \n) will appear inside brackets with a number greater than
3731 31, but if it does, any back reference > 31 suppresses the optimization.
3732
3733 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3734 of disabling numbered capturing parentheses. Any opening parenthesis that is
3735 not followed by ? behaves as if it were followed by ?: but named parentheses
3736 can still be used for capturing (and they will acquire numbers in the usual
3737 way).
3738
3739 41. Redesigned the return codes from the match() function into yes/no/error so
3740 that errors can be passed back from deep inside the nested calls. A malloc
3741 failure while inside a recursive subpattern call now causes the
3742 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3743
3744 42. It is now possible to set a limit on the number of times the match()
3745 function is called in a call to pcre_exec(). This facility makes it possible to
3746 limit the amount of recursion and backtracking, though not in a directly
3747 obvious way, because the match() function is used in a number of different
3748 circumstances. The count starts from zero for each position in the subject
3749 string (for non-anchored patterns). The default limit is, for compatibility, a
3750 large number, namely 10 000 000. You can change this in two ways:
3751
3752 (a) When configuring PCRE before making, you can use --with-match-limit=n
3753 to set a default value for the compiled library.
3754
3755 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3756 a different value is set. See 45 below.
3757
3758 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3759
3760 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3761 of things that can be changed at compile time. The first argument specifies
3762 what is wanted and the second points to where the information is to be placed.
3763 The current list of available information is:
3764
3765 PCRE_CONFIG_UTF8
3766
3767 The output is an integer that is set to one if UTF-8 support is available;
3768 otherwise it is set to zero.
3769
3770 PCRE_CONFIG_NEWLINE
3771
3772 The output is an integer that it set to the value of the code that is used for
3773 newline. It is either LF (10) or CR (13).
3774
3775 PCRE_CONFIG_LINK_SIZE
3776
3777 The output is an integer that contains the number of bytes used for internal
3778 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3779
3780 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3781
3782 The output is an integer that contains the threshold above which the POSIX
3783 interface uses malloc() for output vectors. See item 31 above.
3784
3785 PCRE_CONFIG_MATCH_LIMIT
3786
3787 The output is an unsigned integer that contains the default limit of the number
3788 of match() calls in a pcre_exec() execution. See 42 above.
3789
3790 44. pcretest has been upgraded by the addition of the -C option. This causes it
3791 to extract all the available output from the new pcre_config() function, and to
3792 output it. The program then exits immediately.
3793
3794 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3795 order to support additional features. One way would have been to define
3796 pcre_exec2() (for example) with extra arguments, but this would not have been
3797 extensible, and would also have required all calls to the original function to
3798 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3799 is used for passing in "extra" data from pcre_study().
3800
3801 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3802 contains the following fields:
3803
3804 flags a bitmap indicating which of the following fields are set
3805 study_data opaque data from pcre_study()
3806 match_limit a way of specifying a limit on match() calls for a specific
3807 call to pcre_exec()
3808 callout_data data for callouts (see 49 below)
3809
3810 The flag bits are also defined in pcre.h, and are
3811
3812 PCRE_EXTRA_STUDY_DATA
3813 PCRE_EXTRA_MATCH_LIMIT
3814 PCRE_EXTRA_CALLOUT_DATA
3815
3816 The pcre_study() function now returns one of these new pcre_extra blocks, with
3817 the actual study data pointed to by the study_data field, and the
3818 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3819 before. That is, this change is entirely upwards-compatible and requires no
3820 change to existing code.
3821
3822 If you want to pass in additional data to pcre_exec(), you can either place it
3823 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3824 block.
3825
3826 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3827 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3828 times with different match limits, until it finds the minimum value needed for
3829 pcre_exec() to complete. The value is then output. This can be instructive; for
3830 most simple matches the number is quite small, but for pathological cases it
3831 gets very large very quickly.
3832
3833 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3834 returns the size of the data block pointed to by the study_data field in a
3835 pcre_extra block, that is, the value that was passed as the argument to
3836 pcre_malloc() when PCRE was getting memory in which to place the information
3837 created by pcre_study(). The fourth argument should point to a size_t variable.
3838 pcretest has been extended so that this information is shown after a successful
3839 pcre_study() call when information about the compiled regex is being displayed.
3840
3841 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3842 because what follows is always an absolute path. (Later: it turns out that this
3843 is more than cosmetic for MinGW, because it doesn't like empty path
3844 components.)
3845
3846 49. Some changes have been made to the callout feature (see 28 above):
3847
3848 (i) A callout function now has three choices for what it returns:
3849
3850 0 => success, carry on matching
3851 > 0 => failure at this point, but backtrack if possible
3852 < 0 => serious error, return this value from pcre_exec()
3853
3854 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3855 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3856 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3857 use by callout functions. It will never be used by PCRE itself.
3858
3859 (ii) The pcre_extra structure (see 45 above) has a void * field called
3860 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3861 pcre_callout_block structure has a field of the same name. The contents of
3862 the field passed in the pcre_extra structure are passed to the callout
3863 function in the corresponding field in the callout block. This makes it
3864 easier to use the same callout-containing regex from multiple threads. For
3865 testing, the pcretest program has a new data escape
3866
3867 \C*n pass the number n (may be negative) as callout_data
3868
3869 If the callout function in pcretest receives a non-zero value as
3870 callout_data, it returns that value.
3871
3872 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3873 there were some redundant $(CFLAGS) in commands that are now specified as
3874 $(LINK), which already includes $(CFLAGS).
3875
3876 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3877 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3878 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3879 one-byte characters throughout. Note that case-insensitive matching applies
3880 only to characters whose values are less than 256. PCRE doesn't support the
3881 notion of cases for higher-valued characters.
3882
3883 (i) A character class whose characters are all within 0-255 is handled as
3884 a bit map, and the map is inverted for negative classes. Previously, a
3885 character > 255 always failed to match such a class; however it should
3886 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3887
3888 (ii) A negated character class with a single character < 255 is coded as
3889 "not this character" (OP_NOT). This wasn't working properly when the test
3890 character was multibyte, either singly or repeated.
3891
3892 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3893 mode, for example: \x{100}{2,3}.
3894
3895 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3896 singly or repeated) now correctly test multibyte characters. However,
3897 PCRE doesn't recognize any characters with values greater than 255 as
3898 digits, spaces, or word characters. Such characters always match \D, \S,
3899 and \W, and never match \d, \s, or \w.
3900
3901 (v) Classes may now contain characters and character ranges with values
3902 greater than 255. For example: [ab\x{100}-\x{400}].
3903
3904 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3905 PCRE in UTF-8 mode.
3906
3907 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3908 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3909 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3910 value.)
3911
3912 53. The single man page has become too large. I have therefore split it up into
3913 a number of separate man pages. These also give rise to individual HTML pages;
3914 these are now put in a separate directory, and there is an index.html page that
3915 lists them all. Some hyperlinking between the pages has been installed.
3916
3917 54. Added convenience functions for handling named capturing parentheses.
3918
3919 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3920 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3921 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3922 are faulted.
3923
3924 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3925 calling configure. These values are used when compiling the dftables.c program
3926 which is run to generate the source of the default character tables. They
3927 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3928 you will need to set these values.
3929
3930 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3931
3932
3933 Version 3.9 02-Jan-02
3934 ---------------------
3935
3936 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3937
3938 2. If --disable-static was given, the building process failed when trying to
3939 build pcretest and pcregrep. (For some reason it was using libtool to compile
3940 them, which is not right, as they aren't part of the library.)
3941
3942
3943 Version 3.8 18-Dec-01
3944 ---------------------
3945
3946 1. The experimental UTF-8 code was completely screwed up. It was packing the
3947 bytes in the wrong order. How dumb can you get?
3948
3949
3950 Version 3.7 29-Oct-01
3951 ---------------------
3952
3953 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3954 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3955 this didn't happen under Solaris 8, where I normally test things.
3956
3957 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3958 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3959 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3960 this fix an hour or so after the initial 3.7 release.)
3961
3962
3963 Version 3.6 23-Oct-01
3964 ---------------------
3965
3966 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3967 offsets passed as NULL with zero offset count.
3968
3969 2. The config.guess and config.sub files had not been updated when I moved to
3970 the latest autoconf.
3971
3972
3973 Version 3.5 15-Aug-01
3974 ---------------------
3975
3976 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3977 had been forgotten.
3978
3979 2. By using declared but undefined structures, we can avoid using "void"
3980 definitions in pcre.h while keeping the internal definitions of the structures
3981 private.
3982
3983 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3984 user point of view, this means that both static and shared libraries are built
3985 by default, but this can be individually controlled. More of the work of
3986 handling this static/shared cases is now inside libtool instead of PCRE's make
3987 file.
3988
3989 4. The pcretest utility is now installed along with pcregrep because it is
3990 useful for users (to test regexs) and by doing this, it automatically gets
3991 relinked by libtool. The documentation has been turned into a man page, so
3992 there are now .1, .txt, and .html versions in /doc.
3993
3994 5. Upgrades to pcregrep:
3995 (i) Added long-form option names like gnu grep.
3996 (ii) Added --help to list all options with an explanatory phrase.
3997 (iii) Added -r, --recursive to recurse into sub-directories.
3998 (iv) Added -f, --file to read patterns from a file.
3999
4000 6. pcre_exec() was referring to its "code" argument before testing that
4001 argument for NULL (and giving an error if it was NULL).
4002
4003 7. Upgraded Makefile.in to allow for compiling in a different directory from
4004 the source directory.
4005
4006 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4007 options bits, the pointer it was passed was to an int instead of to an unsigned
4008 long int. This mattered only on 64-bit systems.
4009
4010 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4011 generated) instead of pcre.in, which it its source. Also made the same change
4012 in several of the .c files.
4013
4014 10. A new release of gcc defines printf() as a macro, which broke pcretest
4015 because it had an ifdef in the middle of a string argument for printf(). Fixed
4016 by using separate calls to printf().
4017
4018 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4019 script, to force use of CR or LF instead of \n in the source. On non-Unix
4020 systems, the value can be set in config.h.
4021
4022 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4023 absolute limit. Changed the text of the error message to make this clear, and
4024 likewise updated the man page.
4025
4026 13. The limit of 99 on the number of capturing subpatterns has been removed.
4027 The new limit is 65535, which I hope will not be a "real" limit.
4028
4029
4030 Version 3.4 22-Aug-00
4031 ---------------------
4032
4033 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4034
4035 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4036
4037
4038 Version 3.3 01-Aug-00
4039 ---------------------
4040
4041 1. If an octal character was given, but the value was greater than \377, it
4042 was not getting masked to the least significant bits, as documented. This could
4043 lead to crashes in some systems.
4044
4045 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4046 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4047
4048 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4049 These just pass their arguments on to (pcre_free)(), but they are provided
4050 because some uses of PCRE bind it to non-C systems that can call its functions,
4051 but cannot call free() or pcre_free() directly.
4052
4053 4. Add "make test" as a synonym for "make check". Corrected some comments in
4054 the Makefile.
4055
4056 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4057 Makefile.
4058
4059 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4060 command called pgrep for grepping around the active processes.
4061
4062 7. Added the beginnings of support for UTF-8 character strings.
4063
4064 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4065 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4066 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4067 out for the ar command.)
4068
4069
4070 Version 3.2 12-May-00
4071 ---------------------
4072
4073 This is purely a bug fixing release.
4074
4075 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4076 of ZA. This was just one example of several cases that could provoke this bug,
4077 which was introduced by change 9 of version 2.00. The code for breaking
4078 infinite loops after an iteration that matches an empty string was't working
4079 correctly.
4080
4081 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4082 when matched against abbab (for example). After matching an empty string, it
4083 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4084 caused it to match further down the string than it should.
4085
4086 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4087 was there because it doesn't seem to be needed, and it causes trouble on some
4088 systems, as it is not a Standard C header. It has been removed.
4089
4090 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4091 were reported on the Macintosh. The changes were from
4092
4093 while ((c = *(++ptr)) != 0 && c != '\n');
4094 to
4095 while ((c = *(++ptr)) != 0 && c != '\n') ;
4096
4097 Totally extraordinary, but if that's what it takes...
4098
4099 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4100 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4101 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4102 assumes the way PCRE uses memmove() (always moving upwards).
4103
4104 6. PCRE is being used in one environment where strchr() is not available. There
4105 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4106 faster code anyway.
4107
4108
4109 Version 3.1 09-Feb-00
4110 ---------------------
4111
4112 The only change in this release is the fixing of some bugs in Makefile.in for
4113 the "install" target:
4114
4115 (1) It was failing to install pcreposix.h.
4116
4117 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4118
4119
4120 Version 3.0 01-Feb-00
4121 ---------------------
4122
4123 1. Add support for the /+ modifier to perltest (to output $` like it does in
4124 pcretest).
4125
4126 2. Add support for the /g modifier to perltest.
4127
4128 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4129 matches null strings.
4130
4131 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4132 pattern. Perl treats empty patterns specially - it reuses the most recent
4133 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4134 effect.
4135
4136 5. The POSIX interface was broken in that it was just handing over the POSIX
4137 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4138 required a bigger vector, with some working space on the end. This means that
4139 the POSIX wrapper now has to get and free some memory, and copy the results.
4140
4141 6. Added some simple autoconf support, placing the test data and the
4142 documentation in separate directories, re-organizing some of the
4143 information files, and making it build pcre-config (a GNU standard). Also added
4144 libtool support for building PCRE as a shared library, which is now the
4145 default.
4146
4147 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4148 09 are not valid octal constants. Single digits will be used for minor values
4149 less than 10.
4150
4151 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4152 existing programs that set these in the POSIX interface can use PCRE without
4153 modification.
4154
4155 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4156 return all that pcre_info() returns, plus additional data. The pcre_info()
4157 function is retained for compatibility, but is considered to be obsolete.
4158
4159 10. Added experimental recursion feature (?R) to handle one common case that
4160 Perl 5.6 will be able to do with (?p{...}).
4161
4162 11. Added support for POSIX character classes like [:alpha:], which Perl is
4163 adopting.
4164
4165
4166 Version 2.08 31-Aug-99
4167 ----------------------
4168
4169 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4170 trying to match at the startoffset position, but instead was moving forward to
4171 the next newline as if a previous match had failed.
4172
4173 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4174 and could get into a loop if a null string was matched other than at the start
4175 of the subject.
4176
4177 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4178 be distinguished at compile time, and for completeness also added PCRE_DATE.
4179
4180 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4181 in GnuWin32 environments.
4182
4183
4184 Version 2.07 29-Jul-99
4185 ----------------------
4186
4187 1. The documentation is now supplied in plain text form and HTML as well as in
4188 the form of man page sources.
4189
4190 2. C++ compilers don't like assigning (void *) values to other pointer types.
4191 In particular this affects malloc(). Although there is no problem in Standard
4192 C, I've put in casts to keep C++ compilers happy.
4193
4194 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4195 should be (const char *).
4196
4197 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4198 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4199 However, I haven't made this a standard facility. The documentation doesn't
4200 mention it, and the Makefile doesn't support it.
4201
4202 5. The Makefile now contains an "install" target, with editable destinations at
4203 the top of the file. The pcretest program is not installed.
4204
4205 6. pgrep -V now gives the PCRE version number and date.
4206
4207 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4208 causing the entire string to be ignored, instead of just the last character.
4209
4210 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4211 non-matching string, it can take a very, very long time, even for strings of
4212 quite modest length, because of the nested recursion. PCRE now does better in
4213 some of these cases. It does this by remembering the last required literal
4214 character in the pattern, and pre-searching the subject to ensure it is present
4215 before running the real match. In other words, it applies a heuristic to detect
4216 some types of certain failure quickly, and in the above example, if presented
4217 with a string that has no trailing " it gives "no match" very quickly.
4218
4219 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4220 other alternatives are tried instead.
4221
4222
4223 Version 2.06 09-Jun-99
4224 ----------------------
4225
4226 1. Change pcretest's output for amount of store used to show just the code
4227 space, because the remainder (the data block) varies in size between 32-bit and
4228 64-bit systems.
4229
4230 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4231 start matching at. This allows lookbehinds to work when searching for multiple
4232 occurrences in a string.
4233
4234 3. Added additional options to pcretest for testing multiple occurrences:
4235
4236 /+ outputs the rest of the string that follows a match
4237 /g loops for multiple occurrences, using the new startoffset argument
4238 /G loops for multiple occurrences by passing an incremented pointer
4239
4240 4. PCRE wasn't doing the "first character" optimization for patterns starting
4241 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4242 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4243 the letter 'x'. On long subject strings, this gives a significant speed-up.
4244
4245
4246 Version 2.05 21-Apr-99
4247 ----------------------
4248
4249 1. Changed the type of magic_number from int to long int so that it works
4250 properly on 16-bit systems.
4251
4252 2. Fixed a bug which caused patterns starting with .* not to work correctly
4253 when the subject string contained newline characters. PCRE was assuming
4254 anchoring for such patterns in all cases, which is not correct because .* will
4255 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4256 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4257 must be retried after every newline in the subject.
4258
4259
4260 Version 2.04 18-Feb-99
4261 ----------------------
4262
4263 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4264 computation of the store needed to hold the pattern was incorrect (too large).
4265 If such patterns were nested a few deep, this could multiply and become a real
4266 problem.
4267
4268 2. Added /M option to pcretest to show the memory requirement of a specific
4269 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4270
4271 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4272 compiled in such a way that the backtracking after subsequent failure was
4273 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4274 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4275
4276
4277 Version 2.03 02-Feb-99
4278 ----------------------
4279
4280 1. Fixed typo and small mistake in man page.
4281
4282 2. Added 4th condition (GPL supersedes if conflict) and created separate
4283 LICENCE file containing the conditions.
4284
4285 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4286 Perl, that is the internal \ allows the delimiter to be included in the
4287 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4288 the final delimiter, add \ to the end of the pattern (to test the error).
4289
4290 4. Added the convenience functions for extracting substrings after a successful
4291 match. Updated pcretest to make it able to test these functions.
4292
4293
4294 Version 2.02 14-Jan-99
4295 ----------------------
4296
4297 1. Initialized the working variables associated with each extraction so that
4298 their saving and restoring doesn't refer to uninitialized store.
4299
4300 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4301 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4302 fix the problem.
4303
4304 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4305 calls, and wasn't printing the correct value for compiling calls. Increased the
4306 default value of LOOPREPEAT, and the number of significant figures in the
4307 times.
4308
4309 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4310
4311 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4312 a building problem on Windows NT with a FAT file system.
4313
4314
4315 Version 2.01 21-Oct-98
4316 ----------------------
4317
4318 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4319 to character tables built by pcre_maketables() in the current locale. If NULL
4320 is passed, the default tables are used.
4321
4322
4323 Version 2.00 24-Sep-98
4324 ----------------------
4325
4326 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4327 it any more.
4328
4329 2. Allow quantification of (?>) groups, and make it work correctly.
4330
4331 3. The first character computation wasn't working for (?>) groups.
4332
4333 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4334 end of the subject) and add 5.005's \z, which really does match only at the
4335 very end of the subject.
4336
4337 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4338
4339 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4340 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4341 localized options. All options to pcre_study() were also removed.
4342
4343 7. Add other new features from 5.005:
4344
4345 $(?<= positive lookbehind
4346 $(?<! negative lookbehind
4347 (?imsx-imsx) added the unsetting capability
4348 such a setting is global if at outer level; local otherwise
4349 (?imsx-imsx:) non-capturing groups with option setting
4350 (?(cond)re|re) conditional pattern matching
4351
4352 A backreference to itself in a repeated group matches the previous
4353 captured string.
4354
4355 8. General tidying up of studying (both automatic and via "study")
4356 consequential on the addition of new assertions.
4357
4358 9. As in 5.005, unlimited repeated groups that could match an empty substring
4359 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4360 runtime if any iteration does actually match an empty substring.
4361
4362 10. Include the RunTest script in the distribution.
4363
4364 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4365 discrepancies, some of which were old and were also with respect to 5.004. They
4366 have now been fixed.
4367
4368
4369 Version 1.09 28-Apr-98
4370 ----------------------
4371
4372 1. A negated single character class followed by a quantifier with a minimum
4373 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4374 program crashes, or just wrong answers. This did not apply to negated classes
4375 containing more than one character, or to minima other than one.
4376
4377
4378 Version 1.08 27-Mar-98
4379 ----------------------
4380
4381 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4382
4383 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4384 latter must appear before anything that relies on it in the pattern.
4385
4386
4387 Version 1.07 16-Feb-98
4388 ----------------------
4389
4390 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4391 repeat of a potentially empty string).
4392
4393
4394 Version 1.06 23-Jan-98
4395 ----------------------
4396
4397 1. Added Markus Oberhumer's little patches for C++.
4398
4399 2. Literal strings longer than 255 characters were broken.
4400
4401
4402 Version 1.05 23-Dec-97
4403 ----------------------
4404
4405 1. Negated character classes containing more than one character were failing if
4406 PCRE_CASELESS was set at run time.
4407
4408
4409 Version 1.04 19-Dec-97
4410 ----------------------
4411
4412 1. Corrected the man page, where some "const" qualifiers had been omitted.
4413
4414 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4415 input syntax.
4416
4417 3. Fixed memory leak which occurred when a regex with back references was
4418 matched with an offsets vector that wasn't big enough. The temporary memory
4419 that is used in this case wasn't being freed if the match failed.
4420
4421 4. Tidied pcretest to ensure it frees memory that it gets.
4422
4423 5. Temporary memory was being obtained in the case where the passed offsets
4424 vector was exactly big enough.
4425
4426 6. Corrected definition of offsetof() from change 5 below.
4427
4428 7. I had screwed up change 6 below and broken the rules for the use of
4429 setjmp(). Now fixed.
4430
4431
4432 Version 1.03 18-Dec-97
4433 ----------------------
4434
4435 1. A erroneous regex with a missing opening parenthesis was correctly
4436 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4437 on some systems.
4438
4439 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4440 it was reported that one broken compiler failed on the former because "code" is
4441 also an independent variable.
4442
4443 3. The erroneous regex a[]b caused an array overrun reference.
4444
4445 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4446 fail on data ending with that character. (It was going on too far, and checking
4447 the next character, typically a binary zero.) This was specific to the
4448 optimized code for single-character negative classes.
4449
4450 5. Added a contributed patch from the TIN world which does the following:
4451
4452 + Add an undef for memmove, in case the the system defines a macro for it.
4453
4454 + Add a definition of offsetof(), in case there isn't one. (I don't know
4455 the reason behind this - offsetof() is part of the ANSI standard - but
4456 it does no harm).
4457
4458 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4459 most of the places where whitespace preceded '#'. I have given up and
4460 allowed the remaining 2 cases to be at the margin.
4461
4462 + Rename some variables in pcre to eliminate shadowing. This seems very
4463 pedantic, but does no harm, of course.
4464
4465 6. Moved the call to setjmp() into its own function, to get rid of warnings
4466 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4467
4468 7. Constructs such as \d{8,} were compiling into the equivalent of
4469 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4470 outcome, but in this particular case used more store than had been allocated,
4471 which caused the bug to be discovered because it threw up an internal error.
4472
4473 8. The debugging code in both pcre and pcretest for outputting the compiled
4474 form of a regex was going wrong in the case of back references followed by
4475 curly-bracketed repeats.
4476
4477
4478 Version 1.02 12-Dec-97
4479 ----------------------
4480
4481 1. Typos in pcre.3 and comments in the source fixed.
4482
4483 2. Applied a contributed patch to get rid of places where it used to remove
4484 'const' from variables, and fixed some signed/unsigned and uninitialized
4485 variable warnings.
4486
4487 3. Added the "runtest" target to Makefile.
4488
4489 4. Set default compiler flag to -O2 rather than just -O.
4490
4491
4492 Version 1.01 19-Nov-97
4493 ----------------------
4494
4495 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4496 like /([ab]*)*/, that is, for classes with more than one character in them.
4497
4498 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4499 as /((?>a*))*/ (a PCRE_EXTRA facility).
4500
4501
4502 Version 1.00 18-Nov-97
4503 ----------------------
4504
4505 1. Added compile-time macros to support systems such as SunOS4 which don't have
4506 memmove() or strerror() but have other things that can be used instead.
4507
4508 2. Arranged that "make clean" removes the executables.
4509
4510
4511 Version 0.99 27-Oct-97
4512 ----------------------
4513
4514 1. Fixed bug in code for optimizing classes with only one character. It was
4515 initializing a 32-byte map regardless, which could cause it to run off the end
4516 of the memory it had got.
4517
4518 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4519
4520
4521 Version 0.98 22-Oct-97
4522 ----------------------
4523
4524 1. Fixed bug in code for handling temporary memory usage when there are more
4525 back references than supplied space in the ovector. This could cause segfaults.
4526
4527
4528 Version 0.97 21-Oct-97
4529 ----------------------
4530
4531 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4532
4533 2. Optimized negated single characters not to use a bit map.
4534
4535 3. Brought error texts together as macro definitions; clarified some of them;
4536 fixed one that was wrong - it said "range out of order" when it meant "invalid
4537 escape sequence".
4538
4539 4. Changed some char * arguments to const char *.
4540
4541 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4542
4543 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4544 pcretest.
4545
4546
4547 Version 0.96 16-Oct-97
4548 ----------------------
4549
4550 1. Added a simple "pgrep" utility to the distribution.
4551
4552 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4553 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4554 where "ddd" means "one or more decimal digits".
4555
4556 3. Fixed serious bug. If a pattern had a back reference, but the call to
4557 pcre_exec() didn't supply a large enough ovector to record the related
4558 identifying subpattern, the match always failed. PCRE now remembers the number
4559 of the largest back reference, and gets some temporary memory in which to save
4560 the offsets during matching if necessary, in order to ensure that
4561 backreferences always work.
4562
4563 4. Increased the compatibility with Perl in a number of ways:
4564
4565 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4566 to request this handling. The option can be set at compile or exec time.
4567
4568 (b) $ matches before a terminating newline by default; an option
4569 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4570 mode). The option can be set at compile or exec time.
4571
4572 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4573 the same as Perl's. If the decimal number it represents is less than 10
4574 or there aren't that many previous left capturing parentheses, an octal
4575 escape is read. Inside a character class, it's always an octal escape,
4576 even if it is a single digit.
4577
4578 (d) An escaped but undefined alphabetic character is taken as a literal,
4579 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4580 escapes.
4581
4582 (e) {0} is now permitted. (The previous item is removed from the compiled
4583 pattern).
4584
4585 5. Changed all the names of code files so that the basic parts are no longer
4586 than 10 characters, and abolished the teeny "globals.c" file.
4587
4588 6. Changed the handling of character classes; they are now done with a 32-byte
4589 bit map always.
4590
4591 7. Added the -d and /D options to pcretest to make it possible to look at the
4592 internals of compilation without having to recompile pcre.
4593
4594
4595 Version 0.95 23-Sep-97
4596 ----------------------
4597
4598 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4599 \x20 at the start of a run of normal characters. These were being treated as
4600 real characters, instead of the source characters being re-checked.
4601
4602
4603 Version 0.94 18-Sep-97
4604 ----------------------
4605
4606 1. The functions are now thread-safe, with the caveat that the global variables
4607 containing pointers to malloc() and free() or alternative functions are the
4608 same for all threads.
4609
4610 2. Get pcre_study() to generate a bitmap of initial characters for non-
4611 anchored patterns when this is possible, and use it if passed to pcre_exec().
4612
4613
4614 Version 0.93 15-Sep-97
4615 ----------------------
4616
4617 1. /(b)|(:+)/ was computing an incorrect first character.
4618
4619 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4620 but not actually doing anything yet.
4621
4622 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4623 as Perl does (e.g. [-az] or [az-]).
4624
4625 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4626 all possible positions.
4627
4628 5. Split up into different modules to avoid including unneeded functions in a
4629 compiled binary. However, compile and exec are still in one module. The "study"
4630 function is split off.
4631
4632 6. The character tables are now in a separate module whose source is generated
4633 by an auxiliary program - but can then be edited by hand if required. There are
4634 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4635 toupper() in the code.
4636
4637 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4638 make them global. Abolish the function for setting them, as the caller can now
4639 set them directly.
4640
4641
4642 Version 0.92 11-Sep-97
4643 ----------------------
4644
4645 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4646 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4647
4648 2. Caseless matching was not working in character classes if the characters in
4649 the pattern were in upper case.
4650
4651 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4652
4653 4. Make PCRE_ANCHORED public and accept as a compile option.
4654
4655 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4656 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4657 pass them.
4658
4659 6. Give an error if bad option bits passed at compile or run time.
4660
4661 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4662 pcretest to cause it to pass that flag.
4663
4664 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4665 options, and the first character, if set.
4666
4667 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4668
4669
4670 Version 0.91 10-Sep-97
4671 ----------------------
4672
4673 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4674 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4675
4676 2. PCRE was looping on encountering an indefinitely repeated back reference to
4677 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4678 Perl does - treats the match as successful.
4679
4680 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5