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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 913 - (show annotations)
Sun Feb 12 17:06:59 2012 UTC (3 years, 3 months ago) by ph10
File size: 226282 byte(s)
Error occurred while calculating annotation data.
Add a cast to fix a compiler warning.
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
16 Version 8.30 04-February-2012
17 -----------------------------
18
19 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
20 name is defined in ctype.h.
21
22 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
23 only in quite long subpatterns.
24
25 3. Removed the function pcre_info(), which has been obsolete and deprecated
26 since it was replaced by pcre_fullinfo() in February 2000.
27
28 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
29 match a (*MARK), and the match failed at the start of the subject, a
30 reference to memory before the start of the subject could occur. This bug
31 was introduced by fix 17 of release 8.21.
32
33 5. A reference to an unset group with zero minimum repetition was giving
34 totally wrong answers (in non-JavaScript-compatibility mode). For example,
35 /(another)?(\1?)test/ matched against "hello world test". This bug was
36 introduced in release 8.13.
37
38 6. Add support for 16-bit character strings (a large amount of work involving
39 many changes and refactorings).
40
41 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
42 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
43 from a file.
44
45 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
46 rounding is not applied in this particular case).
47
48 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
49 if they appear, or are escaped, in patterns.
50
51 10. Get rid of a number of -Wunused-but-set-variable warnings.
52
53 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
54 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
55 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
56 also returns the mark "x". This bug applied to capturing parentheses,
57 non-capturing parentheses, and atomic parentheses. It also applied to some
58 assertions.
59
60 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
61 information out of configure.ac instead of relying on pcre.h.generic, which
62 is not stored in the repository.
63
64 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
65 -lreadline.
66
67 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
68
69 15. Applied Graycode's patch to put the top-level frame on the stack rather
70 than the heap when not using the stack for recursion. This gives a
71 performance improvement in many cases when recursion is not deep.
72
73 16. Experimental code added to "pcretest -C" to output the stack frame size.
74
75
76 Version 8.21 12-Dec-2011
77 ------------------------
78
79 1. Updating the JIT compiler.
80
81 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
82 are added as well.
83
84 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
85 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
86 calling _pcre_jit_exec. Some extra comments are added.
87
88 4. (*MARK) settings inside atomic groups that do not contain any capturing
89 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
90 was introduced by change 18 for 8.20.
91
92 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
93 ECMA-262 standard.
94
95 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
96 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
97 This bug was probably introduced by change 9 of 8.13.
98
99 7. While fixing 6 above, I noticed that a number of other items were being
100 incorrectly rejected as "not fixed length". This arose partly because newer
101 opcodes had not been added to the fixed-length checking code. I have (a)
102 corrected the bug and added tests for these items, and (b) arranged for an
103 error to occur if an unknown opcode is encountered while checking for fixed
104 length instead of just assuming "not fixed length". The items that were
105 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
106 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
107 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
108
109 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
110 being incorrectly compiled and would have given unpredicatble results.
111
112 9. A possessively repeated subpattern with minimum repeat count greater than
113 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
114 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
115 the first (A) could occur when it should not.
116
117 10. Add a cast and remove a redundant test from the code.
118
119 11. JIT should use pcre_malloc/pcre_free for allocation.
120
121 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
122 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
123 is anything other than /usr, -L is still shown).
124
125 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
126
127 14. Perl does not support \N without a following name in a [] class; PCRE now
128 also gives an error.
129
130 15. If a forward reference was repeated with an upper limit of around 2000,
131 it caused the error "internal error: overran compiling workspace". The
132 maximum number of forward references (including repeats) was limited by the
133 internal workspace, and dependent on the LINK_SIZE. The code has been
134 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
135 the default depends on LINK_SIZE. There is a new upper limit (for safety)
136 of around 200,000 forward references. While doing this, I also speeded up
137 the filling in of repeated forward references.
138
139 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
140 incorrectly expecting the subject to contain another "a" after the start.
141
142 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
143 in the match, the SKIP should be ignored. This was not happening; instead
144 the SKIP was being treated as NOMATCH. For patterns such as
145 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
146 tested.
147
148 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
149 now much more compatible with Perl, in particular in cases where the result
150 is a non-match for a non-anchored pattern. For example, if
151 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
152 "m", where previously it did not return a name. A side effect of this
153 change is that for partial matches, the last encountered mark name is
154 returned, as for non matches. A number of tests that were previously not
155 Perl-compatible have been moved into the Perl-compatible test files. The
156 refactoring has had the pleasing side effect of removing one argument from
157 the match() function, thus reducing its stack requirements.
158
159 19. If the /S+ option was used in pcretest to study a pattern using JIT,
160 subsequent uses of /S (without +) incorrectly behaved like /S+.
161
162 21. Retrieve executable code size support for the JIT compiler and fixing
163 some warnings.
164
165 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
166 not work when the shorter character appeared right at the end of the
167 subject string.
168
169 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
170 systems.
171
172 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
173 output it when the /M option is used in pcretest.
174
175 25. The CheckMan script was not being included in the distribution. Also, added
176 an explicit "perl" to run Perl scripts from the PrepareRelease script
177 because this is reportedly needed in Windows.
178
179 26. If study data was being save in a file and studying had not found a set of
180 "starts with" bytes for the pattern, the data written to the file (though
181 never used) was taken from uninitialized memory and so caused valgrind to
182 complain.
183
184 27. Updated RunTest.bat as provided by Sheri Pierce.
185
186 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
187
188 29. Computation of memory usage for the table of capturing group names was
189 giving an unnecessarily large value.
190
191
192 Version 8.20 21-Oct-2011
193 ------------------------
194
195 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
196 a POSIX class. After further experiments with Perl, which convinced me that
197 Perl has bugs and confusions, a closing square bracket is no longer allowed
198 in a POSIX name. This bug also affected patterns with classes that started
199 with full stops.
200
201 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
202 captured substring, but while checking the failing first alternative,
203 substring 1 is temporarily captured. If the output vector supplied to
204 pcre_exec() was not big enough for this capture, the yield of the function
205 was still zero ("insufficient space for captured substrings"). This cannot
206 be totally fixed without adding another stack variable, which seems a lot
207 of expense for a edge case. However, I have improved the situation in cases
208 such as /(a)(b)x|abc/ matched against "abc", where the return code
209 indicates that fewer than the maximum number of slots in the ovector have
210 been set.
211
212 3. Related to (2) above: when there are more back references in a pattern than
213 slots in the output vector, pcre_exec() uses temporary memory during
214 matching, and copies in the captures as far as possible afterwards. It was
215 using the entire output vector, but this conflicts with the specification
216 that only 2/3 is used for passing back captured substrings. Now it uses
217 only the first 2/3, for compatibility. This is, of course, another edge
218 case.
219
220 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
221 main code base, and can be used by building with --enable-jit. When this is
222 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
223 runtime --no-jit option is given.
224
225 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
226 ovector, the return from the function was zero, implying that there were
227 other matches that did not fit. The correct "exactly full" value is now
228 returned.
229
230 6. If a subpattern that was called recursively or as a subroutine contained
231 (*PRUNE) or any other control that caused it to give a non-standard return,
232 invalid errors such as "Error -26 (nested recursion at the same subject
233 position)" or even infinite loops could occur.
234
235 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
236 computing the minimum length on reaching *ACCEPT, and so ended up with the
237 wrong value of 1 rather than 0. Further investigation indicates that
238 computing a minimum subject length in the presence of *ACCEPT is difficult
239 (think back references, subroutine calls), and so I have changed the code
240 so that no minimum is registered for a pattern that contains *ACCEPT.
241
242 8. If (*THEN) was present in the first (true) branch of a conditional group,
243 it was not handled as intended. [But see 16 below.]
244
245 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
246 Sheri Pierce.
247
248 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
249 the first byte in a match must be "a".
250
251 11. Change 17 for 8.13 increased the recursion depth for patterns like
252 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
253 pattern contains any instances of (*THEN). If it does not, the old
254 optimizations are restored. It would be nice to do this on a per-group
255 basis, but at the moment that is not feasible.
256
257 12. In some environments, the output of pcretest -C is CRLF terminated. This
258 broke RunTest's code that checks for the link size. A single white space
259 character after the value is now allowed for.
260
261 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
262 For "fr", it uses the Windows-specific input and output files.
263
264 14. If (*THEN) appeared in a group that was called recursively or as a
265 subroutine, it did not work as intended. [But see next item.]
266
267 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
268 pattern fragments (but not containing any | characters). If A and B are
269 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
270 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
271 D. In other words, Perl considers parentheses that do not contain any |
272 characters to be part of a surrounding alternative, whereas PCRE was
273 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
274 differently. PCRE now behaves in the same way as Perl, except in the case
275 of subroutine/recursion calls such as (?1) which have in any case always
276 been different (but PCRE had them first :-).
277
278 16. Related to 15 above: Perl does not treat the | in a conditional group as
279 creating alternatives. Such a group is treated in the same way as an
280 ordinary group without any | characters when processing (*THEN). PCRE has
281 been changed to match Perl's behaviour.
282
283 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
284 RunGrepTest script failed.
285
286 18. Change 22 for version 13 caused atomic groups to use more stack. This is
287 inevitable for groups that contain captures, but it can lead to a lot of
288 stack use in large patterns. The old behaviour has been restored for atomic
289 groups that do not contain any capturing parentheses.
290
291 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
292 suppress the check for a minimum subject length at run time. (If it was
293 given to pcre_exec() or pcre_dfa_exec() it did work.)
294
295 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
296 fail to work when decoding hex characters in data strings in EBCDIC
297 environments.
298
299 21. It appears that in at least one Mac OS environment, the isxdigit() function
300 is implemented as a macro that evaluates to its argument more than once,
301 contravening the C 90 Standard (I haven't checked a later standard). There
302 was an instance in pcretest which caused it to go wrong when processing
303 \x{...} escapes in subject strings. The has been rewritten to avoid using
304 things like p++ in the argument of isxdigit().
305
306
307 Version 8.13 16-Aug-2011
308 ------------------------
309
310 1. The Unicode data tables have been updated to Unicode 6.0.0.
311
312 2. Two minor typos in pcre_internal.h have been fixed.
313
314 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
315 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
316 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
317
318 4. There were a number of related bugs in the code for matching backrefences
319 caselessly in UTF-8 mode when codes for the characters concerned were
320 different numbers of bytes. For example, U+023A and U+2C65 are an upper
321 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
322 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
323 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
324 2-byte code at the end of the subject (it thought there wasn't enough data
325 left).
326
327 5. Comprehensive information about what went wrong is now returned by
328 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
329 as the output vector has at least 2 elements. The offset of the start of
330 the failing character and a reason code are placed in the vector.
331
332 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
333 now returned is for the first byte of the failing character, instead of the
334 last byte inspected. This is an incompatible change, but I hope it is small
335 enough not to be a problem. It makes the returned offset consistent with
336 pcre_exec() and pcre_dfa_exec().
337
338 7. pcretest now gives a text phrase as well as the error number when
339 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
340 failure, the offset and reason code are output.
341
342 8. When \R was used with a maximizing quantifier it failed to skip backwards
343 over a \r\n pair if the subsequent match failed. Instead, it just skipped
344 back over a single character (\n). This seems wrong (because it treated the
345 two characters as a single entity when going forwards), conflicts with the
346 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
347 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
348 has been changed.
349
350 9. Some internal refactoring has changed the processing so that the handling
351 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
352 time (the PCRE_DOTALL option was changed this way some time ago: version
353 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
354 which was always a bit of a fudge. It also means that there is one less
355 argument for the match() function, which reduces its stack requirements
356 slightly. This change also fixes an incompatibility with Perl: the pattern
357 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
358
359 10. More internal refactoring has drastically reduced the number of recursive
360 calls to match() for possessively repeated groups such as (abc)++ when
361 using pcre_exec().
362
363 11. While implementing 10, a number of bugs in the handling of groups were
364 discovered and fixed:
365
366 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
367 (a|)*(?1) gave a compile-time internal error.
368 ((a|)+)+ did not notice that the outer group could match an empty string.
369 (^a|^)+ was not marked as anchored.
370 (.*a|.*)+ was not marked as matching at start or after a newline.
371
372 12. Yet more internal refactoring has removed another argument from the match()
373 function. Special calls to this function are now indicated by setting a
374 value in a variable in the "match data" data block.
375
376 13. Be more explicit in pcre_study() instead of relying on "default" for
377 opcodes that mean there is no starting character; this means that when new
378 ones are added and accidentally left out of pcre_study(), testing should
379 pick them up.
380
381 14. The -s option of pcretest has been documented for ages as being an old
382 synonym of -m (show memory usage). I have changed it to mean "force study
383 for every regex", that is, assume /S for every regex. This is similar to -i
384 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
385 using it. It makes it easier to run collections of tests with and without
386 study enabled, and thereby test pcre_study() more easily. All the standard
387 tests are now run with and without -s (but some patterns can be marked as
388 "never study" - see 20 below).
389
390 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
391 restoration of the capturing data to the outer values was not happening
392 correctly.
393
394 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
395 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
396 pattern had matched an empty string, and so incorrectly returned a no
397 match.
398
399 17. There was optimizing code for the last branch of non-capturing parentheses,
400 and also for the obeyed branch of a conditional subexpression, which used
401 tail recursion to cut down on stack usage. Unfortunately, now that there is
402 the possibility of (*THEN) occurring in these branches, tail recursion is
403 no longer possible because the return has to be checked for (*THEN). These
404 two optimizations have therefore been removed. [But see 8.20/11 above.]
405
406 18. If a pattern containing \R was studied, it was assumed that \R always
407 matched two bytes, thus causing the minimum subject length to be
408 incorrectly computed because \R can also match just one byte.
409
410 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
411 was incorrectly computed.
412
413 20. If /S is present twice on a test pattern in pcretest input, it now
414 *disables* studying, thereby overriding the use of -s on the command line
415 (see 14 above). This is necessary for one or two tests to keep the output
416 identical in both cases.
417
418 21. When (*ACCEPT) was used in an assertion that matched an empty string and
419 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
420
421 22. When an atomic group that contained a capturing parenthesis was
422 successfully matched, but the branch in which it appeared failed, the
423 capturing was not being forgotten if a higher numbered group was later
424 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
425 group 1 to "a", when in fact it should be unset. This applied to multi-
426 branched capturing and non-capturing groups, repeated or not, and also to
427 positive assertions (capturing in negative assertions does not happen
428 in PCRE) and also to nested atomic groups.
429
430 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
431 subject after a captured substring, to make it easier to tell which of a
432 number of identical substrings has been captured.
433
434 24. The way atomic groups are processed by pcre_exec() has been changed so that
435 if they are repeated, backtracking one repetition now resets captured
436 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
437 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
438 "aaa". Previously, it would have been "a". As part of this code
439 refactoring, the way recursive calls are handled has also been changed.
440
441 25. If an assertion condition captured any substrings, they were not passed
442 back unless some other capturing happened later. For example, if
443 (?(?=(a))a) was matched against "a", no capturing was returned.
444
445 26. When studying a pattern that contained subroutine calls or assertions,
446 the code for finding the minimum length of a possible match was handling
447 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
448 group 1 called group 2 while simultaneously a separate group 2 called group
449 1). A stack overflow occurred in this case. I have fixed this by limiting
450 the recursion depth to 10.
451
452 27. Updated RunTest.bat in the distribution to the version supplied by Tom
453 Fortmann. This supports explicit test numbers on the command line, and has
454 argument validation and error reporting.
455
456 28. An instance of \X with an unlimited repeat could fail if at any point the
457 first character it looked at was a mark character.
458
459 29. Some minor code refactoring concerning Unicode properties and scripts
460 should reduce the stack requirement of match() slightly.
461
462 30. Added the '=' option to pcretest to check the setting of unused capturing
463 slots at the end of the pattern, which are documented as being -1, but are
464 not included in the return count.
465
466 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
467 compiled something random. Now it gives a compile-time error (as does
468 Perl).
469
470 32. A *MARK encountered during the processing of a positive assertion is now
471 recorded and passed back (compatible with Perl).
472
473 33. If --only-matching or --colour was set on a pcregrep call whose pattern
474 had alternative anchored branches, the search for a second match in a line
475 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
476 matched the line "0102" twice. The same bug affected patterns that started
477 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
478 twice.
479
480 34. Previously, PCRE did not allow quantification of assertions. However, Perl
481 does, and because of capturing effects, quantifying parenthesized
482 assertions may at times be useful. Quantifiers are now allowed for
483 parenthesized assertions.
484
485 35. A minor code tidy in pcre_compile() when checking options for \R usage.
486
487 36. \g was being checked for fancy things in a character class, when it should
488 just be a literal "g".
489
490 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
491 appearance of a nested POSIX class supersedes an apparent external class.
492 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
493 unescaped square brackets may also appear as part of class names. For
494 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
495 more like Perl. (But see 8.20/1 above.)
496
497 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
498 was because it thought it was \N{name}, which is not supported).
499
500 39. Add minix to OS list not supporting the -S option in pcretest.
501
502 40. PCRE tries to detect cases of infinite recursion at compile time, but it
503 cannot analyze patterns in sufficient detail to catch mutual recursions
504 such as ((?1))((?2)). There is now a runtime test that gives an error if a
505 subgroup is called recursively as a subpattern for a second time at the
506 same position in the subject string. In previous releases this might have
507 been caught by the recursion limit, or it might have run out of stack.
508
509 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
510 happen only once. PCRE was, however incorrectly giving a compile time error
511 "recursive call could loop indefinitely" because it cannot analyze the
512 pattern in sufficient detail. The compile time test no longer happens when
513 PCRE is compiling a conditional subpattern, but actual runaway loops are
514 now caught at runtime (see 40 above).
515
516 42. It seems that Perl allows any characters other than a closing parenthesis
517 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
518 has been changed to be the same.
519
520 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
521 as not to get warnings when autogen.sh is called. Also changed
522 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
523
524 44. To help people who use pcregrep to scan files containing exceedingly long
525 lines, the following changes have been made:
526
527 (a) The default value of the buffer size parameter has been increased from
528 8K to 20K. (The actual buffer used is three times this size.)
529
530 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
531 PCRE is built.
532
533 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
534 to be set at run time.
535
536 (d) Numerical values in pcregrep options can be followed by K or M, for
537 example --buffer-size=50K.
538
539 (e) If a line being scanned overflows pcregrep's buffer, an error is now
540 given and the return code is set to 2.
541
542 45. Add a pointer to the latest mark to the callout data block.
543
544 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
545 partial match of an empty string instead of no match. This was specific to
546 the use of ".".
547
548 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
549 complete match instead of a partial match. This bug was dependent on both
550 the PCRE_UTF8 and PCRE_DOTALL options being set.
551
552 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
553 starting byte set, because \b was not being ignored.
554
555
556 Version 8.12 15-Jan-2011
557 ------------------------
558
559 1. Fixed some typos in the markup of the man pages, and wrote a script that
560 checks for such things as part of the documentation building process.
561
562 2. On a big-endian 64-bit system, pcregrep did not correctly process the
563 --match-limit and --recursion-limit options (added for 8.11). In
564 particular, this made one of the standard tests fail. (The integer value
565 went into the wrong half of a long int.)
566
567 3. If the --colour option was given to pcregrep with -v (invert match), it
568 did strange things, either producing crazy output, or crashing. It should,
569 of course, ignore a request for colour when reporting lines that do not
570 match.
571
572 4. Another pcregrep bug caused similar problems if --colour was specified with
573 -M (multiline) and the pattern match finished with a line ending.
574
575 5. In pcregrep, when a pattern that ended with a literal newline sequence was
576 matched in multiline mode, the following line was shown as part of the
577 match. This seems wrong, so I have changed it.
578
579 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
580 the check for further matches in the same line (so they could be coloured)
581 to overrun the end of the current line. If another match was found, it was
582 incorrectly shown (and then shown again when found in the next line).
583
584 7. If pcregrep was compiled under Windows, there was a reference to the
585 function pcregrep_exit() before it was defined. I am assuming this was
586 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
587 reported by a user. I've moved the definition above the reference.
588
589
590 Version 8.11 10-Dec-2010
591 ------------------------
592
593 1. (*THEN) was not working properly if there were untried alternatives prior
594 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
595 backtracked to try for "b" instead of moving to the next alternative branch
596 at the same level (in this case, to look for "c"). The Perl documentation
597 is clear that when (*THEN) is backtracked onto, it goes to the "next
598 alternative in the innermost enclosing group".
599
600 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
601 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
602 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
603 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
604 (*THEN).
605
606 3. If \s appeared in a character class, it removed the VT character from
607 the class, even if it had been included by some previous item, for example
608 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
609 of \s, but is part of the POSIX "space" class.)
610
611 4. A partial match never returns an empty string (because you can always
612 match an empty string at the end of the subject); however the checking for
613 an empty string was starting at the "start of match" point. This has been
614 changed to the "earliest inspected character" point, because the returned
615 data for a partial match starts at this character. This means that, for
616 example, /(?<=abc)def/ gives a partial match for the subject "abc"
617 (previously it gave "no match").
618
619 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
620 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
621 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
622 has an implication that the given string is incomplete (because a partial
623 match is preferred over a full match). For this reason, these items now
624 give a partial match in this situation. [Aside: previously, the one case
625 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
626 match rather than a full match, which was wrong by the old rules, but is
627 now correct.]
628
629 6. There was a bug in the handling of #-introduced comments, recognized when
630 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
631 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
632 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
633 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
634 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
635 places in pcre_compile().
636
637 7. Related to (6) above, when pcre_compile() was skipping #-introduced
638 comments when looking ahead for named forward references to subpatterns,
639 the only newline sequence it recognized was NL. It now handles newlines
640 according to the set newline convention.
641
642 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
643 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
644 cater for a lack of strerror(). These oversights have been fixed.
645
646 9. Added --match-limit and --recursion-limit to pcregrep.
647
648 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
649
650 11. When the -o option was used, pcregrep was setting a return code of 1, even
651 when matches were found, and --line-buffered was not being honoured.
652
653 12. Added an optional parentheses number to the -o and --only-matching options
654 of pcregrep.
655
656 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
657 can match an empty string. The code to do it in pcretest and pcredemo
658 needed fixing:
659
660 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
661 only one byte after an empty string match just before CRLF (this case
662 just got forgotten; "any" and "anycrlf" were OK).
663
664 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
665 mode when an empty string match preceded an ASCII character followed by
666 a non-ASCII character. (The code for advancing by one character rather
667 than one byte was nonsense.)
668
669 (c) The pcredemo.c sample program did not have any code at all to handle
670 the cases when CRLF is a valid newline sequence.
671
672 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
673 as a starting offset was within the subject string. There is now a new
674 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
675 negative or greater than the length of the string. In order to test this,
676 pcretest is extended to allow the setting of negative starting offsets.
677
678 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
679 starting offset points to the beginning of a UTF-8 character was
680 unnecessarily clumsy. I tidied it up.
681
682 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
683 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
684
685 17. Nobody had reported that the --include_dir option, which was added in
686 release 7.7 should have been called --include-dir (hyphen, not underscore)
687 for compatibility with GNU grep. I have changed it to --include-dir, but
688 left --include_dir as an undocumented synonym, and the same for
689 --exclude-dir, though that is not available in GNU grep, at least as of
690 release 2.5.4.
691
692 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
693 characters from a string of bytes) have been redefined so as not to use
694 loops, in order to improve performance in some environments. At the same
695 time, I abstracted some of the common code into auxiliary macros to save
696 repetition (this should not affect the compiled code).
697
698 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
699 compile-time error is now given if \c is not followed by an ASCII
700 character, that is, a byte less than 128. (In EBCDIC mode, the code is
701 different, and any byte value is allowed.)
702
703 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
704 START_OPTIMIZE option, which is now allowed at compile time - but just
705 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
706 to pcregrep and other applications that have no direct access to PCRE
707 options. The new /Y option in pcretest sets this option when calling
708 pcre_compile().
709
710 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
711 back references. Groups containing recursive back references were forced to
712 be atomic by that change, but in the case of named groups, the amount of
713 memory required was incorrectly computed, leading to "Failed: internal
714 error: code overflow". This has been fixed.
715
716 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
717 pcretest.c, to avoid build problems in some Borland environments.
718
719
720 Version 8.10 25-Jun-2010
721 ------------------------
722
723 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
724 THEN.
725
726 2. (*ACCEPT) was not working when inside an atomic group.
727
728 3. Inside a character class, \B is treated as a literal by default, but
729 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
730 causes the error). The code is unchanged, but I tidied the documentation.
731
732 4. Inside a character class, PCRE always treated \R and \X as literals,
733 whereas Perl faults them if its -w option is set. I have changed PCRE so
734 that it faults them when PCRE_EXTRA is set.
735
736 5. Added support for \N, which always matches any character other than
737 newline. (It is the same as "." when PCRE_DOTALL is not set.)
738
739 6. When compiling pcregrep with newer versions of gcc which may have
740 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
741 declared with attribute warn_unused_result" were given. Just casting the
742 result to (void) does not stop the warnings; a more elaborate fudge is
743 needed. I've used a macro to implement this.
744
745 7. Minor change to pcretest.c to avoid a compiler warning.
746
747 8. Added four artifical Unicode properties to help with an option to make
748 \s etc use properties (see next item). The new properties are: Xan
749 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
750
751 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
752 use Unicode properties. (*UCP) at the start of a pattern can be used to set
753 this option. Modified pcretest to add /W to test this facility. Added
754 REG_UCP to make it available via the POSIX interface.
755
756 10. Added --line-buffered to pcregrep.
757
758 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
759 studied, and the match started with a letter with a code point greater than
760 127 whose first byte was different to the first byte of the other case of
761 the letter, the other case of this starting letter was not recognized
762 (#976).
763
764 12. If a pattern that was studied started with a repeated Unicode property
765 test, for example, \p{Nd}+, there was the theoretical possibility of
766 setting up an incorrect bitmap of starting bytes, but fortunately it could
767 not have actually happened in practice until change 8 above was made (it
768 added property types that matched character-matching opcodes).
769
770 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
771 possible starting bytes for non-anchored patterns.
772
773 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
774 \R, and also a number of cases that involve Unicode properties, both
775 explicit and implicit when PCRE_UCP is set.
776
777 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
778 input, it could crash or give wrong results if characters with values
779 greater than 0xc0 were present in the subject string. (Detail: it assumed
780 UTF-8 input when processing these items.)
781
782 16. Added a lot of (int) casts to avoid compiler warnings in systems where
783 size_t is 64-bit (#991).
784
785 17. Added a check for running out of memory when PCRE is compiled with
786 --disable-stack-for-recursion (#990).
787
788 18. If the last data line in a file for pcretest does not have a newline on
789 the end, a newline was missing in the output.
790
791 19. The default pcre_chartables.c file recognizes only ASCII characters (values
792 less than 128) in its various bitmaps. However, there is a facility for
793 generating tables according to the current locale when PCRE is compiled. It
794 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
795 characters, are recognized by isspace() and therefore were getting set in
796 these tables, and indeed these tables seem to approximate to ISO 8859. This
797 caused a problem in UTF-8 mode when pcre_study() was used to create a list
798 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
799 which of course cannot start UTF-8 characters. I have changed the code so
800 that only real ASCII characters (less than 128) and the correct starting
801 bytes for UTF-8 encodings are set for characters greater than 127 when in
802 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
803 altogether.)
804
805 20. Added the /T option to pcretest so as to be able to run tests with non-
806 standard character tables, thus making it possible to include the tests
807 used for 19 above in the standard set of tests.
808
809 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
810 reference to a subpattern the other side of a comment that contains an
811 opening parenthesis caused either an internal compiling error, or a
812 reference to the wrong subpattern.
813
814
815 Version 8.02 19-Mar-2010
816 ------------------------
817
818 1. The Unicode data tables have been updated to Unicode 5.2.0.
819
820 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
821 configured.
822
823 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
824 original author of that file, following a query about its status.
825
826 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
827 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
828
829 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
830 quantifier applied to a forward-referencing subroutine call, could compile
831 incorrect code or give the error "internal error: previously-checked
832 referenced subpattern not found".
833
834 6. Both MS Visual Studio and Symbian OS have problems with initializing
835 variables to point to external functions. For these systems, therefore,
836 pcre_malloc etc. are now initialized to local functions that call the
837 relevant global functions.
838
839 7. There were two entries missing in the vectors called coptable and poptable
840 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
841 I've fixed the data, and added a kludgy way of testing at compile time that
842 the lengths are correct (equal to the number of opcodes).
843
844 8. Following on from 7, I added a similar kludge to check the length of the
845 eint vector in pcreposix.c.
846
847 9. Error texts for pcre_compile() are held as one long string to avoid too
848 much relocation at load time. To find a text, the string is searched,
849 counting zeros. There was no check for running off the end of the string,
850 which could happen if a new error number was added without updating the
851 string.
852
853 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
854
855 11. \K was not working if it appeared in an atomic group or in a group that
856 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
857 \K is "not well defined" if used in an assertion. PCRE now accepts it if
858 the assertion is positive, but not if it is negative.
859
860 12. Change 11 fortuitously reduced the size of the stack frame used in the
861 "match()" function of pcre_exec.c by one pointer. Forthcoming
862 implementation of support for (*MARK) will need an extra pointer on the
863 stack; I have reserved it now, so that the stack frame size does not
864 decrease.
865
866 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
867 item in branch that calls a recursion is a subroutine call - as in the
868 second branch in the above example - was incorrectly given the compile-
869 time error "recursive call could loop indefinitely" because pcre_compile()
870 was not correctly checking the subroutine for matching a non-empty string.
871
872 14. The checks for overrunning compiling workspace could trigger after an
873 overrun had occurred. This is a "should never occur" error, but it can be
874 triggered by pathological patterns such as hundreds of nested parentheses.
875 The checks now trigger 100 bytes before the end of the workspace.
876
877 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
878
879
880 Version 8.01 19-Jan-2010
881 ------------------------
882
883 1. If a pattern contained a conditional subpattern with only one branch (in
884 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
885 computed the wrong minimum data length (which is of course zero for such
886 subpatterns). This could cause incorrect "no match" results.
887
888 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
889 the pattern is reset in the first branch, pcre_compile() failed with
890 "internal error: code overflow at offset...". This happened only when
891 the reset was to the original external option setting. (An optimization
892 abstracts leading options settings into an external setting, which was the
893 cause of this.)
894
895 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
896 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
897 assertion pattern did not match (meaning that the assertion was true), it
898 was incorrectly treated as false if the SKIP had been reached during the
899 matching. This also applied to assertions used as conditions.
900
901 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
902 assertion subpattern, including such a pattern used as a condition,
903 unpredictable results occurred, instead of the error return
904 PCRE_ERROR_DFA_UITEM.
905
906 5. The C++ GlobalReplace function was not working like Perl for the special
907 situation when an empty string is matched. It now does the fancy magic
908 stuff that is necessary.
909
910 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
911 removed. (These were left over from very, very early versions of PCRE.)
912
913 7. Some cosmetic changes to the code to make life easier when compiling it
914 as part of something else:
915
916 (a) Change DEBUG to PCRE_DEBUG.
917
918 (b) In pcre_compile(), rename the member of the "branch_chain" structure
919 called "current" as "current_branch", to prevent a collision with the
920 Linux macro when compiled as a kernel module.
921
922 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
923 prevent a collision with the Linux macro when compiled as a kernel
924 module.
925
926 8. In pcre_compile() there are some checks for integer overflows that used to
927 cast potentially large values to (double). This has been changed to that
928 when building, a check for int64_t is made, and if it is found, it is used
929 instead, thus avoiding the use of floating point arithmetic. (There is no
930 other use of FP in PCRE.) If int64_t is not found, the fallback is to
931 double.
932
933 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
934 2005 (difference between two addresses compared to an unsigned value).
935
936 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
937 custom one, because of the following reported problem in Windows:
938
939 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
940 under Win32.
941 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
942 therefore missing the function definition.
943 - The compiler thus generates a "C" signature for the test function.
944 - The linker fails to find the "C" function.
945 - PCRE fails to configure if asked to do so against libbz2.
946
947 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
948 messages were output:
949
950 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
951 rerunning libtoolize, to keep the correct libtool macros in-tree.
952 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
953
954 I have done both of these things.
955
956 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
957 most of the time, it *can* run out if it is given a pattern that contains a
958 runaway infinite recursion. I updated the discussion in the pcrestack man
959 page.
960
961 13. Now that we have gone to the x.xx style of version numbers, the minor
962 version may start with zero. Using 08 or 09 is a bad idea because users
963 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
964 interpreted as invalid octal numbers. I've updated the previous comment in
965 configure.ac, and also added a check that gives an error if 08 or 09 are
966 used.
967
968 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
969 causing partial matching to fail when the end of the subject matched \W
970 in a UTF-8 pattern where \W was quantified with a minimum of 3.
971
972 15. There were some discrepancies between the declarations in pcre_internal.h
973 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
974 their definitions. The declarations used "const uschar *" and the
975 definitions used USPTR. Even though USPTR is normally defined as "const
976 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
977 reported that: "This difference in casting confuses some C++ compilers, for
978 example, SunCC recognizes above declarations as different functions and
979 generates broken code for hbpcre." I have changed the declarations to use
980 USPTR.
981
982 16. GNU libtool is named differently on some systems. The autogen.sh script now
983 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
984 (FreeBSD).
985
986 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
987 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
988 comment: "Figure out how to create a longlong from a string: strtoll and
989 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
990 instance, but it only takes 2 args instead of 3!"
991
992 18. A subtle bug concerned with back references has been fixed by a change of
993 specification, with a corresponding code fix. A pattern such as
994 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
995 refers, was giving matches when it shouldn't. For example, xa=xaaa would
996 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
997 same bug. Such groups have to be quantified to be useful, or contained
998 inside another quantified group. (If there's no repetition, the reference
999 can never match.) The problem arises because, having left the group and
1000 moved on to the rest of the pattern, a later failure that backtracks into
1001 the group uses the captured value from the final iteration of the group
1002 rather than the correct earlier one. I have fixed this in PCRE by forcing
1003 any group that contains a reference to itself to be an atomic group; that
1004 is, there cannot be any backtracking into it once it has completed. This is
1005 similar to recursive and subroutine calls.
1006
1007
1008 Version 8.00 19-Oct-09
1009 ----------------------
1010
1011 1. The table for translating pcre_compile() error codes into POSIX error codes
1012 was out-of-date, and there was no check on the pcre_compile() error code
1013 being within the table. This could lead to an OK return being given in
1014 error.
1015
1016 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1017 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1018 in a Windows environment.
1019
1020 3. The pcregrep --count option prints the count for each file even when it is
1021 zero, as does GNU grep. However, pcregrep was also printing all files when
1022 --files-with-matches was added. Now, when both options are given, it prints
1023 counts only for those files that have at least one match. (GNU grep just
1024 prints the file name in this circumstance, but including the count seems
1025 more useful - otherwise, why use --count?) Also ensured that the
1026 combination -clh just lists non-zero counts, with no names.
1027
1028 4. The long form of the pcregrep -F option was incorrectly implemented as
1029 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1030 but it seems right to fix it, and I didn't think it was worth preserving
1031 the old behaviour.
1032
1033 5. The command line items --regex=pattern and --regexp=pattern were not
1034 recognized by pcregrep, which required --regex pattern or --regexp pattern
1035 (with a space rather than an '='). The man page documented the '=' forms,
1036 which are compatible with GNU grep; these now work.
1037
1038 6. No libpcreposix.pc file was created for pkg-config; there was just
1039 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1040
1041 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1042 when UCP support is not needed, by modifying the Python script that
1043 generates it from Unicode data files. This should not matter if the module
1044 is correctly used as a library, but I received one complaint about 50K of
1045 unwanted data. My guess is that the person linked everything into his
1046 program rather than using a library. Anyway, it does no harm.
1047
1048 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1049 was a minimum greater than 1 for a wide character in a possessive
1050 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1051 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1052 character. Chaos in the form of incorrect output or a compiling loop could
1053 result.
1054
1055 9. The restrictions on what a pattern can contain when partial matching is
1056 requested for pcre_exec() have been removed. All patterns can now be
1057 partially matched by this function. In addition, if there are at least two
1058 slots in the offset vector, the offset of the earliest inspected character
1059 for the match and the offset of the end of the subject are set in them when
1060 PCRE_ERROR_PARTIAL is returned.
1061
1062 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1063 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1064 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1065 and may be more useful for multi-segment matching.
1066
1067 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1068 used to be given if ever the end of the subject was reached; now it is
1069 given only if matching could not proceed because another character was
1070 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1071 string "Z", which now yields "no match" instead of "partial match". In the
1072 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1073 final character ended with (*FAIL).
1074
1075 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1076 if the pattern had a "must contain" character that was already found in the
1077 earlier partial match, unless partial matching was again requested. For
1078 example, with the pattern /dog.(body)?/, the "must contain" character is
1079 "g". If the first part-match was for the string "dog", restarting with
1080 "sbody" failed. This bug has been fixed.
1081
1082 13. The string returned by pcre_dfa_exec() after a partial match has been
1083 changed so that it starts at the first inspected character rather than the
1084 first character of the match. This makes a difference only if the pattern
1085 starts with a lookbehind assertion or \b or \B (\K is not supported by
1086 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1087 matching functions compatible, and I think it's the right thing to do.
1088
1089 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1090 so that the demonstration program is easily available in environments where
1091 PCRE has not been installed from source.
1092
1093 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1094 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1095 library.
1096
1097 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1098 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1099 is not the first non-POSIX option to be added. Clearly some people find
1100 these options useful.
1101
1102 17. If a caller to the POSIX matching function regexec() passes a non-zero
1103 value for nmatch with a NULL value for pmatch, the value of
1104 nmatch is forced to zero.
1105
1106 18. RunGrepTest did not have a test for the availability of the -u option of
1107 the diff command, as RunTest does. It now checks in the same way as
1108 RunTest, and also checks for the -b option.
1109
1110 19. If an odd number of negated classes containing just a single character
1111 interposed, within parentheses, between a forward reference to a named
1112 subpattern and the definition of the subpattern, compilation crashed with
1113 an internal error, complaining that it could not find the referenced
1114 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1115 [The bug was that it was starting one character too far in when skipping
1116 over the character class, thus treating the ] as data rather than
1117 terminating the class. This meant it could skip too much.]
1118
1119 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1120 /g option in pcretest when the pattern contains \K, which makes it possible
1121 to have an empty string match not at the start, even when the pattern is
1122 anchored. Updated pcretest and pcredemo to use this option.
1123
1124 21. If the maximum number of capturing subpatterns in a recursion was greater
1125 than the maximum at the outer level, the higher number was returned, but
1126 with unset values at the outer level. The correct (outer level) value is
1127 now given.
1128
1129 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1130 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1131 make it do so. The string so far is captured, making this feature
1132 compatible with Perl.
1133
1134 23. The tests have been re-organized, adding tests 11 and 12, to make it
1135 possible to check the Perl 5.10 features against Perl 5.10.
1136
1137 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1138 pattern matches a fixed length string. PCRE did not allow this; now it
1139 does. Neither allows recursion.
1140
1141 25. I finally figured out how to implement a request to provide the minimum
1142 length of subject string that was needed in order to match a given pattern.
1143 (It was back references and recursion that I had previously got hung up
1144 on.) This code has now been added to pcre_study(); it finds a lower bound
1145 to the length of subject needed. It is not necessarily the greatest lower
1146 bound, but using it to avoid searching strings that are too short does give
1147 some useful speed-ups. The value is available to calling programs via
1148 pcre_fullinfo().
1149
1150 26. While implementing 25, I discovered to my embarrassment that pcretest had
1151 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1152 study optimizations had never been tested with that matching function.
1153 Oops. What is worse, even when it was passed study data, there was a bug in
1154 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1155 were also very few tests of studied patterns with pcre_dfa_exec().
1156
1157 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1158 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1159 on the other side of the coin, they are no longer allowed to have different
1160 names, because these cannot be distinguished in PCRE, and this has caused
1161 confusion. (This is a difference from Perl.)
1162
1163 28. When duplicate subpattern names are present (necessarily with different
1164 numbers, as required by 27 above), and a test is made by name in a
1165 conditional pattern, either for a subpattern having been matched, or for
1166 recursion in such a pattern, all the associated numbered subpatterns are
1167 tested, and the overall condition is true if the condition is true for any
1168 one of them. This is the way Perl works, and is also more like the way
1169 testing by number works.
1170
1171
1172 Version 7.9 11-Apr-09
1173 ---------------------
1174
1175 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1176 (pcretest), all targets were linked against these libraries. This included
1177 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1178 libraries. This caused unwanted dependencies to be created. This problem
1179 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1180 pcretest is linked with readline.
1181
1182 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1183 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1184 moved outside it again, because FALSE and TRUE are already defined in AIX,
1185 but BOOL is not.
1186
1187 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1188 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1189
1190 4. The pcregrep documentation said spaces were inserted as well as colons (or
1191 hyphens) following file names and line numbers when outputting matching
1192 lines. This is not true; no spaces are inserted. I have also clarified the
1193 wording for the --colour (or --color) option.
1194
1195 5. In pcregrep, when --colour was used with -o, the list of matching strings
1196 was not coloured; this is different to GNU grep, so I have changed it to be
1197 the same.
1198
1199 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1200 each matching line was coloured. Now it goes on to look for further matches
1201 of any of the test patterns, which is the same behaviour as GNU grep.
1202
1203 7. A pattern that could match an empty string could cause pcregrep to loop; it
1204 doesn't make sense to accept an empty string match in pcregrep, so I have
1205 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1206 seems to be how GNU grep behaves.
1207
1208 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1209 start or after a newline", because the conditional assertion was not being
1210 correctly handled. The rule now is that both the assertion and what follows
1211 in the first alternative must satisfy the test.
1212
1213 9. If auto-callout was enabled in a pattern with a conditional group whose
1214 condition was an assertion, PCRE could crash during matching, both with
1215 pcre_exec() and pcre_dfa_exec().
1216
1217 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1218 used for matching.
1219
1220 11. Unicode property support in character classes was not working for
1221 characters (bytes) greater than 127 when not in UTF-8 mode.
1222
1223 12. Added the -M command line option to pcretest.
1224
1225 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1226
1227 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1228
1229 16. Added comments and documentation about mis-use of no_arg in the C++
1230 wrapper.
1231
1232 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1233 from Martin Jerabek that uses macro names for all relevant character and
1234 string constants.
1235
1236 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1237 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1238 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1239 these, but not everybody uses configure.
1240
1241 19. A conditional group that had only one branch was not being correctly
1242 recognized as an item that could match an empty string. This meant that an
1243 enclosing group might also not be so recognized, causing infinite looping
1244 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1245 with the subject "ab", where knowledge that the repeated group can match
1246 nothing is needed in order to break the loop.
1247
1248 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1249 exec(), but without supplying a callout function, matching went wrong.
1250
1251 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1252 leak if the size of the offset vector was greater than 30. When the vector
1253 is smaller, the saved offsets during recursion go onto a local stack
1254 vector, but for larger vectors malloc() is used. It was failing to free
1255 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1256 error, in fact).
1257
1258 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1259 heapframe that is used only when UTF-8 support is enabled. This caused no
1260 problem, but was untidy.
1261
1262 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1263 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1264 included within another project.
1265
1266 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1267 slightly modified by me:
1268
1269 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1270 not building pcregrep.
1271
1272 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1273 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1274
1275 25. Forward references, both numeric and by name, in patterns that made use of
1276 duplicate group numbers, could behave incorrectly or give incorrect errors,
1277 because when scanning forward to find the reference group, PCRE was not
1278 taking into account the duplicate group numbers. A pattern such as
1279 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1280
1281 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1282 the feature of a custom pointer more persuasive (as requested by a user).
1283
1284 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1285 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1286 pre-defined.
1287
1288 28. Added support for (*UTF8) at the start of a pattern.
1289
1290 29. Arrange for flags added by the "release type" setting in CMake to be shown
1291 in the configuration summary.
1292
1293
1294 Version 7.8 05-Sep-08
1295 ---------------------
1296
1297 1. Replaced UCP searching code with optimized version as implemented for Ad
1298 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1299 stage table and inline lookup instead of a function, giving speed ups of 2
1300 to 5 times on some simple patterns that I tested. Permission was given to
1301 distribute the MultiStage2.py script that generates the tables (it's not in
1302 the tarball, but is in the Subversion repository).
1303
1304 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1305 scripts.
1306
1307 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1308 a group with a zero qualifier. The result of the study could be incorrect,
1309 or the function might crash, depending on the pattern.
1310
1311 4. Caseless matching was not working for non-ASCII characters in back
1312 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1313 It now works when Unicode Property Support is available.
1314
1315 5. In pcretest, an escape such as \x{de} in the data was always generating
1316 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1317 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1318 truncation.
1319
1320 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1321
1322 7. Added two (int) casts to pcregrep when printing the difference of two
1323 pointers, in case they are 64-bit values.
1324
1325 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1326 test 2 if it fails.
1327
1328 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1329 and a #define of that name to empty if it is not externally set. This is to
1330 allow users of MSVC to set it if necessary.
1331
1332 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1333 the convenience functions in the pcre_get.c source file.
1334
1335 11. An option change at the start of a pattern that had top-level alternatives
1336 could cause overwriting and/or a crash. This command provoked a crash in
1337 some environments:
1338
1339 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1340
1341 This potential security problem was recorded as CVE-2008-2371.
1342
1343 12. For a pattern where the match had to start at the beginning or immediately
1344 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1345 pcre_dfa_exec() could read past the end of the passed subject if there was
1346 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1347 pcretest so that it places the subject at the end of its malloc-ed buffer.
1348
1349 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1350 exec() might read past the end of the data buffer in UTF-8 mode.
1351
1352 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1353 the data contained the byte 0x85 as part of a UTF-8 character within its
1354 first line. This applied both to normal and DFA matching.
1355
1356 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1357 /^[^d]*?$/8 failed to match "abc".
1358
1359 16. Added a missing copyright notice to pcrecpp_internal.h.
1360
1361 17. Make it more clear in the documentation that values returned from
1362 pcre_exec() in ovector are byte offsets, not character counts.
1363
1364 18. Tidied a few places to stop certain compilers from issuing warnings.
1365
1366 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1367 supplied by Stefan Weber. I made a further small update for 7.8 because
1368 there is a change of source arrangements: the pcre_searchfuncs.c module is
1369 replaced by pcre_ucd.c.
1370
1371
1372 Version 7.7 07-May-08
1373 ---------------------
1374
1375 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1376 a string to a long long, pretend we don't even have a long long." This is
1377 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1378
1379 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1380 pre-7.6 versions, which defined a global no_arg variable instead of putting
1381 it in the RE class. (See also #8 below.)
1382
1383 3. Remove a line of dead code, identified by coverity and reported by Nuno
1384 Lopes.
1385
1386 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1387
1388 (1) The include/exclude patterns were being applied to the whole pathnames
1389 of files, instead of just to the final components.
1390
1391 (2) If there was more than one level of directory, the subdirectories were
1392 skipped unless they satisfied the include/exclude conditions. This is
1393 inconsistent with GNU grep (and could even be seen as contrary to the
1394 pcregrep specification - which I improved to make it absolutely clear).
1395 The action now is always to scan all levels of directory, and just
1396 apply the include/exclude patterns to regular files.
1397
1398 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1399 --exclude_dir in the tests to avoid scanning .svn directories.
1400
1401 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1402 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1403 doesn't support NULs in patterns.
1404
1405 7. Added some missing "const"s to declarations of static tables in
1406 pcre_compile.c and pcre_dfa_exec.c.
1407
1408 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1409 caused by fix #2 above. (Subsequently also a second patch to fix the
1410 first patch. And a third patch - this was a messy problem.)
1411
1412 9. Applied Craig's patch to remove the use of push_back().
1413
1414 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1415 matching function regexec().
1416
1417 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1418 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1419 references. PCRE supports relative numbers with this syntax (I don't think
1420 Oniguruma does).
1421
1422 12. Previously, a group with a zero repeat such as (...){0} was completely
1423 omitted from the compiled regex. However, this means that if the group
1424 was called as a subroutine from elsewhere in the pattern, things went wrong
1425 (an internal error was given). Such groups are now left in the compiled
1426 pattern, with a new opcode that causes them to be skipped at execution
1427 time.
1428
1429 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1430 to the way PCRE behaves:
1431
1432 (a) A lone ] character is dis-allowed (Perl treats it as data).
1433
1434 (b) A back reference to an unmatched subpattern matches an empty string
1435 (Perl fails the current match path).
1436
1437 (c) A data ] in a character class must be notated as \] because if the
1438 first data character in a class is ], it defines an empty class. (In
1439 Perl it is not possible to have an empty class.) The empty class []
1440 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1441 The negative empty class [^] matches any one character, independently
1442 of the DOTALL setting.
1443
1444 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1445 non-existent subpattern following a character class starting with ']' and
1446 containing () gave an internal compiling error instead of "reference to
1447 non-existent subpattern". Fortunately, when the pattern did exist, the
1448 compiled code was correct. (When scanning forwards to check for the
1449 existencd of the subpattern, it was treating the data ']' as terminating
1450 the class, so got the count wrong. When actually compiling, the reference
1451 was subsequently set up correctly.)
1452
1453 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1454 it was being rejected as not supported by pcre_dfa_exec(), even though
1455 other assertions are supported. I have made pcre_dfa_exec() support
1456 (*FAIL).
1457
1458 16. The implementation of 13c above involved the invention of a new opcode,
1459 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1460 cannot be changed at match time, I realized I could make a small
1461 improvement to matching performance by compiling OP_ALLANY instead of
1462 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1463 on the OP_ANY path.
1464
1465 17. Compiling pcretest on Windows with readline support failed without the
1466 following two fixes: (1) Make the unistd.h include conditional on
1467 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1468
1469 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1470 ncurses library to be included for pcretest when ReadLine support is
1471 requested, but also to allow for it to be overridden. This patch came from
1472 Daniel Bergström.
1473
1474 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1475 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1476 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1477 spotting this.
1478
1479
1480 Version 7.6 28-Jan-08
1481 ---------------------
1482
1483 1. A character class containing a very large number of characters with
1484 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1485 overflow.
1486
1487 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1488 HAVE_LONG_LONG is not defined.
1489
1490 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1491 bring them up to date and include new features. This patch includes:
1492
1493 - Fixed PH's badly added libz and libbz2 support.
1494 - Fixed a problem with static linking.
1495 - Added pcredemo. [But later removed - see 7 below.]
1496 - Fixed dftables problem and added an option.
1497 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1498 HAVE_LONG_LONG.
1499 - Added readline support for pcretest.
1500 - Added an listing of the option settings after cmake has run.
1501
1502 4. A user submitted a patch to Makefile that makes it easy to create
1503 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1504 Makefile.am that cause it to include this special target, without
1505 affecting anything else. Note that the same mingw target plus all
1506 the other distribution libraries and programs are now supported
1507 when configuring with CMake (see 6 below) instead of with
1508 Configure/Make.
1509
1510 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1511 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1512 exported in the Windows port". It has not yet been confirmed that the patch
1513 solves the problem, but it does no harm.
1514
1515 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1516 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1517 with CMake, and also correct the comment about stack recursion.
1518
1519 7. Remove the automatic building of pcredemo from the ./configure system and
1520 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1521 of a program that users should build themselves after PCRE is installed, so
1522 building it automatically is not really right. What is more, it gave
1523 trouble in some build environments.
1524
1525 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1526
1527
1528 Version 7.5 10-Jan-08
1529 ---------------------
1530
1531 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1532 values in parens when parsing an RE using the C++ wrapper."
1533
1534 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1535 Characters greater than 255 were excluded from the class instead of being
1536 included.
1537
1538 3. The same bug as (2) above applied to negated POSIX classes such as
1539 [:^space:].
1540
1541 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1542 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1543 I have changed it.
1544
1545 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1546 first named subpattern) and a construct such as (?&a) would reference the
1547 first named subpattern whose name started with "a" (in other words, the
1548 length check was missing). Both these problems are fixed. "Subpattern name
1549 expected" is now given for (?&) (a zero-length name), and this patch also
1550 makes it give the same error for \k'' (previously it complained that that
1551 was a reference to a non-existent subpattern).
1552
1553 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1554 this is right because (?- can be followed by option settings as well as by
1555 digits. I have, however, made the messages clearer.
1556
1557 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1558 than the number used in the conditional) now cause a compile-time error.
1559 This is actually not compatible with Perl, which accepts such patterns, but
1560 treats the conditional as always being FALSE (as PCRE used to), but it
1561 seems to me that giving a diagnostic is better.
1562
1563 8. Change "alphameric" to the more common word "alphanumeric" in comments
1564 and messages.
1565
1566 9. Fix two occurrences of "backslash" in comments that should have been
1567 "backspace".
1568
1569 10. Remove two redundant lines of code that can never be obeyed (their function
1570 was moved elsewhere).
1571
1572 11. The program that makes PCRE's Unicode character property table had a bug
1573 which caused it to generate incorrect table entries for sequences of
1574 characters that have the same character type, but are in different scripts.
1575 It amalgamated them into a single range, with the script of the first of
1576 them. In other words, some characters were in the wrong script. There were
1577 thirteen such cases, affecting characters in the following ranges:
1578
1579 U+002b0 - U+002c1
1580 U+0060c - U+0060d
1581 U+0061e - U+00612
1582 U+0064b - U+0065e
1583 U+0074d - U+0076d
1584 U+01800 - U+01805
1585 U+01d00 - U+01d77
1586 U+01d9b - U+01dbf
1587 U+0200b - U+0200f
1588 U+030fc - U+030fe
1589 U+03260 - U+0327f
1590 U+0fb46 - U+0fbb1
1591 U+10450 - U+1049d
1592
1593 12. The -o option (show only the matching part of a line) for pcregrep was not
1594 compatible with GNU grep in that, if there was more than one match in a
1595 line, it showed only the first of them. It now behaves in the same way as
1596 GNU grep.
1597
1598 13. If the -o and -v options were combined for pcregrep, it printed a blank
1599 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1600 does the same. The return code can be used to tell if there were any
1601 non-matching lines.
1602
1603 14. Added --file-offsets and --line-offsets to pcregrep.
1604
1605 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1606 infinitely looping recursion. The bug was that positive lookaheads were not
1607 being skipped when checking for a possible empty match (negative lookaheads
1608 and both kinds of lookbehind were skipped).
1609
1610 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1611 inclusion of <windows.h> to before rather than after the definition of
1612 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1613
1614 17. Specifying a possessive quantifier with a specific limit for a Unicode
1615 character property caused pcre_compile() to compile bad code, which led at
1616 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1617 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1618 caused the error; without that there was no problem.
1619
1620 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1621
1622 19. Added --enable-pcretest-libreadline.
1623
1624 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1625 RE::GlobalReplace(). As a result, the number of replacements returned was
1626 double what it should be. I removed one of the increments, but Craig sent a
1627 later patch that removed the other one (the right fix) and added unit tests
1628 that check the return values (which was not done before).
1629
1630 21. Several CMake things:
1631
1632 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1633 the names libpcre and libpcreposix, not just pcre and pcreposix.
1634
1635 (2) The above change means that pcretest and pcregrep are now correctly
1636 linked with the newly-built libraries, not previously installed ones.
1637
1638 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1639
1640 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1641 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1642 UTF-8 newline character). The key issue is that the pattern starts .*;
1643 this means that the match must be either at the beginning, or after a
1644 newline. The bug was in the code for advancing after a failed match and
1645 checking that the new position followed a newline. It was not taking
1646 account of UTF-8 characters correctly.
1647
1648 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1649 character classes. PCRE was not treating the sequence [:...:] as a
1650 character class unless the ... were all letters. Perl, however, seems to
1651 allow any characters between [: and :], though of course it rejects as
1652 unknown any "names" that contain non-letters, because all the known class
1653 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1654 for example, whereas PCRE did not - it did not recognize a POSIX character
1655 class. This seemed a bit dangerous, so the code has been changed to be
1656 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1657 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1658 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1659 Perl does, and where it didn't before.
1660
1661 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1662 Windows environments %n is disabled by default.
1663
1664
1665 Version 7.4 21-Sep-07
1666 ---------------------
1667
1668 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1669 means that a class such as [\s] counted as "explicit reference to CR or
1670 LF". That isn't really right - the whole point of the change was to try to
1671 help when there was an actual mention of one of the two characters. So now
1672 the change happens only if \r or \n (or a literal CR or LF) character is
1673 encountered.
1674
1675 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1676 of both had grown to the point where there were only 3 bits left.
1677 Fortunately, there was spare space in the data structure, and so I have
1678 moved the internal flags into a new 16-bit field to free up more option
1679 bits.
1680
1681 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1682 but did not set the internal JCHANGED flag - either of these is enough to
1683 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1684 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1685 start sets both bits.
1686
1687 4. Added options (at build time, compile time, exec time) to change \R from
1688 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1689
1690 5. doc/pcresyntax.html was missing from the distribution.
1691
1692 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1693 compatibility, even though it is no longer used.
1694
1695 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1696 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1697 windows.h file is present (where different names are used). [This was
1698 reversed later after testing - see 16 below.]
1699
1700 8. Changed all #include <config.h> to #include "config.h". There were also
1701 some further <pcre.h> cases that I changed to "pcre.h".
1702
1703 9. When pcregrep was used with the --colour option, it missed the line ending
1704 sequence off the lines that it output.
1705
1706 10. It was pointed out to me that arrays of string pointers cause lots of
1707 relocations when a shared library is dynamically loaded. A technique of
1708 using a single long string with a table of offsets can drastically reduce
1709 these. I have refactored PCRE in four places to do this. The result is
1710 dramatic:
1711
1712 Originally: 290
1713 After changing UCP table: 187
1714 After changing error message table: 43
1715 After changing table of "verbs" 36
1716 After changing table of Posix names 22
1717
1718 Thanks to the folks working on Gregex for glib for this insight.
1719
1720 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1721 unicode-properties was also set.
1722
1723 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1724
1725 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1726 checked only for CRLF.
1727
1728 14. Added casts to pcretest.c to avoid compiler warnings.
1729
1730 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1731
1732 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1733 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1734 entirely. This removes changes made in 7 above.
1735
1736 17. The CMake files have been updated, and there is now more information about
1737 building with CMake in the NON-UNIX-USE document.
1738
1739
1740 Version 7.3 28-Aug-07
1741 ---------------------
1742
1743 1. In the rejigging of the build system that eventually resulted in 7.1, the
1744 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1745 brackets there is not right, since it causes compilers to look for an
1746 installed pcre.h, not the version that is in the source that is being
1747 compiled (which of course may be different). I have changed it back to:
1748
1749 #include "pcre.h"
1750
1751 I have a vague recollection that the change was concerned with compiling in
1752 different directories, but in the new build system, that is taken care of
1753 by the VPATH setting the Makefile.
1754
1755 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1756 when the subject happened to end in the byte 0x85 (e.g. if the last
1757 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1758 characters but of course it shouldn't be taken as a newline when it is part
1759 of another character. The bug was that, for an unlimited repeat of . in
1760 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1761 characters when looking for a newline.
1762
1763 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1764
1765 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1766 in debug output.
1767
1768 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1769 long printing in the pcrecpp unittest when running under MinGW.
1770
1771 6. ESC_K was left out of the EBCDIC table.
1772
1773 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1774 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1775 limit also applies to "virtual nesting" when a pattern is recursive, and in
1776 this case 1000 isn't so big. I have been able to remove this limit at the
1777 expense of backing off one optimization in certain circumstances. Normally,
1778 when pcre_exec() would call its internal match() function recursively and
1779 immediately return the result unconditionally, it uses a "tail recursion"
1780 feature to save stack. However, when a subpattern that can match an empty
1781 string has an unlimited repetition quantifier, it no longer makes this
1782 optimization. That gives it a stack frame in which to save the data for
1783 checking that an empty string has been matched. Previously this was taken
1784 from the 1000-entry workspace that had been reserved. So now there is no
1785 explicit limit, but more stack is used.
1786
1787 8. Applied Daniel's patches to solve problems with the import/export magic
1788 syntax that is required for Windows, and which was going wrong for the
1789 pcreposix and pcrecpp parts of the library. These were overlooked when this
1790 problem was solved for the main library.
1791
1792 9. There were some crude static tests to avoid integer overflow when computing
1793 the size of patterns that contain repeated groups with explicit upper
1794 limits. As the maximum quantifier is 65535, the maximum group length was
1795 set at 30,000 so that the product of these two numbers did not overflow a
1796 32-bit integer. However, it turns out that people want to use groups that
1797 are longer than 30,000 bytes (though not repeat them that many times).
1798 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1799 made it possible to implement the integer overflow checks in a much more
1800 dynamic way, which I have now done. The artificial limitation on group
1801 length has been removed - we now have only the limit on the total length of
1802 the compiled pattern, which depends on the LINK_SIZE setting.
1803
1804 10. Fixed a bug in the documentation for get/copy named substring when
1805 duplicate names are permitted. If none of the named substrings are set, the
1806 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1807 empty string.
1808
1809 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1810 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1811 because the ] is interpreted as the first data character and the
1812 terminating ] is not found. PCRE has been made compatible with Perl in this
1813 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1814 cause memory overwriting.
1815
1816 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1817 string has been matched (to stop an infinite loop). It was not recognizing
1818 a conditional subpattern that could match an empty string if that
1819 subpattern was within another subpattern. For example, it looped when
1820 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1821 condition was not nested. This bug has been fixed.
1822
1823 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1824 past the start of the subject in the presence of bytes with the top bit
1825 set, for example "\x8aBCD".
1826
1827 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1828 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1829
1830 14. Optimized (?!) to (*FAIL).
1831
1832 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1833 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1834 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1835 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1836 does: it's just the validity check that is more restrictive.
1837
1838 16. Inserted checks for integer overflows during escape sequence (backslash)
1839 processing, and also fixed erroneous offset values for syntax errors during
1840 backslash processing.
1841
1842 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1843 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1844
1845 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1846 caused an overrun.
1847
1848 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1849 something other than just ASCII characters) inside a group that had an
1850 unlimited repeat caused a loop at compile time (while checking to see
1851 whether the group could match an empty string).
1852
1853 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1854 [\P{Any}] did so. (Error in the code for printing property names.)
1855
1856 21. An orphan \E inside a character class could cause a crash.
1857
1858 22. A repeated capturing bracket such as (A)? could cause a wild memory
1859 reference during compilation.
1860
1861 23. There are several functions in pcre_compile() that scan along a compiled
1862 expression for various reasons (e.g. to see if it's fixed length for look
1863 behind). There were bugs in these functions when a repeated \p or \P was
1864 present in the pattern. These operators have additional parameters compared
1865 with \d, etc, and these were not being taken into account when moving along
1866 the compiled data. Specifically:
1867
1868 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1869 length.
1870
1871 (b) An item such as \pL+ within a repeated group could cause crashes or
1872 loops.
1873
1874 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1875 "reference to non-existent subpattern" error.
1876
1877 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1878
1879 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1880 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1881
1882 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1883
1884 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1885 character were causing crashes (broken optimization).
1886
1887 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1888 \p or \P) caused a compile-time loop.
1889
1890 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1891 break. For example, the unstudied pattern [\r\n]A does not match the string
1892 "\r\nA" because change 7.0/46 below moves the current point on by two
1893 characters after failing to match at the start. However, the pattern \nA
1894 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1895 the same is true. There doesn't seem any very clean way out of this, but
1896 what I have chosen to do makes the common cases work: PCRE now takes note
1897 of whether there can be an explicit match for \r or \n anywhere in the
1898 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1899 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1900 pattern has explicit CR or LF references.
1901
1902 29. Added (*CR) etc for changing newline setting at start of pattern.
1903
1904
1905 Version 7.2 19-Jun-07
1906 ---------------------
1907
1908 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1909 which is apparently normally available under Windows.
1910
1911 2. Re-jig the pcregrep tests with different newline settings in an attempt
1912 to make them independent of the local environment's newline setting.
1913
1914 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1915
1916 4. Some of the "internals" tests were previously cut out when the link size
1917 was not 2, because the output contained actual offsets. The recent new
1918 "Z" feature of pcretest means that these can be cut out, making the tests
1919 usable with all link sizes.
1920
1921 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1922 stack recursion. This gives a massive performance boost under BSD, but just
1923 a small improvement under Linux. However, it saves one field in the frame
1924 in all cases.
1925
1926 6. Added more features from the forthcoming Perl 5.10:
1927
1928 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1929 recursion call. It refers to the nth most recently opened parentheses.
1930
1931 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1932 to be opened parentheses.
1933
1934 (c) Conditions that refer to capturing parentheses can be specified
1935 relatively, for example, (?(-2)... or (?(+3)...
1936
1937 (d) \K resets the start of the current match so that everything before
1938 is not part of it.
1939
1940 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1941
1942 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1943 reference syntax.
1944
1945 (g) (?| introduces a group in which the numbering of parentheses in each
1946 alternative starts with the same number.
1947
1948 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1949
1950 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1951 PCRE_INFO_JCHANGED.
1952
1953 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1954 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1955 for detecting groups that can match an empty string.
1956
1957 9. A pattern with a very large number of alternatives (more than several
1958 hundred) was running out of internal workspace during the pre-compile
1959 phase, where pcre_compile() figures out how much memory will be needed. A
1960 bit of new cunning has reduced the workspace needed for groups with
1961 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1962 workspace instead of running out of the 4096 that are available.
1963
1964 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1965
1966 11. Applied patch from Google to remove an optimization that didn't quite work.
1967 The report of the bug said:
1968
1969 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1970 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1971 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1972
1973 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1974 it matched the wrong number of bytes.
1975
1976
1977 Version 7.1 24-Apr-07
1978 ---------------------
1979
1980 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1981 that is more "standard", making use of automake and other Autotools. There
1982 is some re-arrangement of the files and adjustment of comments consequent
1983 on this.
1984
1985 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1986 for recursive directory scanning broke on some systems because the files
1987 are not scanned in any specific order and on different systems the order
1988 was different. A call to "sort" has been inserted into RunGrepTest for the
1989 approprate test as a short-term fix. In the longer term there may be an
1990 alternative.
1991
1992 3. I had an email from Eric Raymond about problems translating some of PCRE's
1993 man pages to HTML (despite the fact that I distribute HTML pages, some
1994 people do their own conversions for various reasons). The problems
1995 concerned the use of low-level troff macros .br and .in. I have therefore
1996 removed all such uses from the man pages (some were redundant, some could
1997 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1998 HTML has been updated to handle .nf/.fi and to complain if it encounters
1999 .br or .in.
2000
2001 4. Updated comments in configure.ac that get placed in config.h.in and also
2002 arranged for config.h to be included in the distribution, with the name
2003 config.h.generic, for the benefit of those who have to compile without
2004 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2005
2006 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2007 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2008 makevp.bat for the current PCRE, using the additional files
2009 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2010
2011 6. A Windows user reported a minor discrepancy with test 2, which turned out
2012 to be caused by a trailing space on an input line that had got lost in his
2013 copy. The trailing space was an accident, so I've just removed it.
2014
2015 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2016 that is needed.
2017
2018 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2019 as "const" (a) because they are and (b) because it helps the PHP
2020 maintainers who have recently made a script to detect big data structures
2021 in the php code that should be moved to the .rodata section. I remembered
2022 to update Builducptable as well, so it won't revert if ucptable.h is ever
2023 re-created.
2024
2025 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2026 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2027 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2028 support is not required. This saves 1.5-2K of code, which is important in
2029 some applications.
2030
2031 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2032 so as not to refer to the tables, even though these functions will never be
2033 called when UTF-8 support is disabled. Otherwise there are problems with a
2034 shared library.
2035
2036 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2037
2038 (a) It was defining its arguments as char * instead of void *.
2039
2040 (b) It was assuming that all moves were upwards in memory; this was true
2041 a long time ago when I wrote it, but is no longer the case.
2042
2043 The emulated memove() is provided for those environments that have neither
2044 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2045 is clearly not the case, as these two bugs were recently reported.
2046
2047 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2048 and Detrail to create the HTML documentation, the .txt form of the man
2049 pages, and it removes trailing spaces from listed files. It also creates
2050 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2051 case, it wraps all the #defines with #ifndefs. This script should be run
2052 before "make dist".
2053
2054 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2055 with Unicode property support.
2056
2057 (a) For a maximizing quantifier, if the two different cases of the
2058 character were of different lengths in their UTF-8 codings (there are
2059 some cases like this - I found 11), and the matching function had to
2060 back up over a mixture of the two cases, it incorrectly assumed they
2061 were both the same length.
2062
2063 (b) When PCRE was configured to use the heap rather than the stack for
2064 recursion during matching, it was not correctly preserving the data for
2065 the other case of a UTF-8 character when checking ahead for a match
2066 while processing a minimizing repeat. If the check also involved
2067 matching a wide character, but failed, corruption could cause an
2068 erroneous result when trying to check for a repeat of the original
2069 character.
2070
2071 13. Some tidying changes to the testing mechanism:
2072
2073 (a) The RunTest script now detects the internal link size and whether there
2074 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2075 values substituted by "configure". (The RunGrepTest script already did
2076 this for UTF-8.) The configure.ac script no longer substitutes the
2077 relevant variables.
2078
2079 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2080 with length and offset values. This means that the output is different
2081 for different internal link sizes. Test 2 is skipped for link sizes
2082 other than 2 because of this, bypassing the problem. Unfortunately,
2083 there was also a test in test 3 (the locale tests) that used /B and
2084 failed for link sizes other than 2. Rather than cut the whole test out,
2085 I have added a new /Z option to pcretest that replaces the length and
2086 offset values with spaces. This is now used to make test 3 independent
2087 of link size. (Test 2 will be tidied up later.)
2088
2089 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2090 segmentation fault instead of returning the appropriate error message.
2091
2092 15. In multiline mode when the newline sequence was set to "any", the pattern
2093 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2094 This doesn't seem right; it now treats the CRLF combination as the line
2095 ending, and so does not match in that case. It's only a pattern such as ^$
2096 that would hit this one: something like ^ABC$ would have failed after \r
2097 and then tried again after \r\n.
2098
2099 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2100 in an attempt to make files that differ only in their line terminators
2101 compare equal. This works on Linux.
2102
2103 17. Under certain error circumstances pcregrep might try to free random memory
2104 as it exited. This is now fixed, thanks to valgrind.
2105
2106 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2107 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2108 was because its rules for how to advance for /g after matching an empty
2109 string at the end of a line did not allow for this case. They now check for
2110 it specially.
2111
2112 20. pcretest is supposed to handle patterns and data of any length, by
2113 extending its buffers when necessary. It was getting this wrong when the
2114 buffer for a data line had to be extended.
2115
2116 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2117 CRLF as a newline sequence.
2118
2119 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2120 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2121 I have nevertheless tidied it up.
2122
2123 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2124
2125 24. Added a man page for pcre-config.
2126
2127
2128 Version 7.0 19-Dec-06
2129 ---------------------
2130
2131 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2132 moving to gcc 4.1.1.
2133
2134 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2135 sys/time.h, which is documented as needed for this function. It doesn't
2136 seem to matter on Linux, but it showed up on some releases of OS X.
2137
2138 3. It seems that there are systems where bytes whose values are greater than
2139 127 match isprint() in the "C" locale. The "C" locale should be the
2140 default when a C program starts up. In most systems, only ASCII printing
2141 characters match isprint(). This difference caused the output from pcretest
2142 to vary, making some of the tests fail. I have changed pcretest so that:
2143
2144 (a) When it is outputting text in the compiled version of a pattern, bytes
2145 other than 32-126 are always shown as hex escapes.
2146
2147 (b) When it is outputting text that is a matched part of a subject string,
2148 it does the same, unless a different locale has been set for the match
2149 (using the /L modifier). In this case, it uses isprint() to decide.
2150
2151 4. Fixed a major bug that caused incorrect computation of the amount of memory
2152 required for a compiled pattern when options that changed within the
2153 pattern affected the logic of the preliminary scan that determines the
2154 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2155 that the computed length was too small. The symptoms of this bug were
2156 either the PCRE error "internal error: code overflow" from pcre_compile(),
2157 or a glibc crash with a message such as "pcretest: free(): invalid next
2158 size (fast)". Examples of patterns that provoked this bug (shown in
2159 pcretest format) are:
2160
2161 /(?-x: )/x
2162 /(?x)(?-x: \s*#\s*)/
2163 /((?i)[\x{c0}])/8
2164 /(?i:[\x{c0}])/8
2165
2166 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2167 is now done differently.
2168
2169 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2170 wrapper classes; (b) implement a new function in the C++ scanner that is
2171 more efficient than the old way of doing things because it avoids levels of
2172 recursion in the regex matching; (c) add a paragraph to the documentation
2173 for the FullMatch() function.
2174
2175 6. The escape sequence \n was being treated as whatever was defined as
2176 "newline". Not only was this contrary to the documentation, which states
2177 that \n is character 10 (hex 0A), but it also went horribly wrong when
2178 "newline" was defined as CRLF. This has been fixed.
2179
2180 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2181 was being set to -1 for the "end of line" case (supposedly a value that no
2182 character can have). Though this value is never used (the check for end of
2183 line is "zero bytes in current character"), it caused compiler complaints.
2184 I've changed it to 0xffffffff.
2185
2186 8. In pcre_version.c, the version string was being built by a sequence of
2187 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2188 string (as it is for production releases) called a macro with an empty
2189 argument. The C standard says the result of this is undefined. The gcc
2190 compiler treats it as an empty string (which was what was wanted) but it is
2191 reported that Visual C gives an error. The source has been hacked around to
2192 avoid this problem.
2193
2194 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2195 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2196 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2197 of them did).
2198
2199 10. Originally, pcretest opened its input and output without "b"; then I was
2200 told that "b" was needed in some environments, so it was added for release
2201 5.0 to both the input and output. (It makes no difference on Unix-like
2202 systems.) Later I was told that it is wrong for the input on Windows. I've
2203 now abstracted the modes into two macros, to make it easier to fiddle with
2204 them, and removed "b" from the input mode under Windows.
2205
2206 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2207
2208 12. Added -help and --help to pcretest as an official way of being reminded
2209 of the options.
2210
2211 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2212 and pcrecpp.cc because they annoy compilers at high warning levels.
2213
2214 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2215
2216 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2217 scripts are not C programs :-) and which was not noticed because it works
2218 on Linux.
2219
2220 16. pcretest is supposed to handle any length of pattern and data line (as one
2221 line or as a continued sequence of lines) by extending its input buffer if
2222 necessary. This feature was broken for very long pattern lines, leading to
2223 a string of junk being passed to pcre_compile() if the pattern was longer
2224 than about 50K.
2225
2226 17. I have done a major re-factoring of the way pcre_compile() computes the
2227 amount of memory needed for a compiled pattern. Previously, there was code
2228 that made a preliminary scan of the pattern in order to do this. That was
2229 OK when PCRE was new, but as the facilities have expanded, it has become
2230 harder and harder to keep it in step with the real compile phase, and there
2231 have been a number of bugs (see for example, 4 above). I have now found a
2232 cunning way of running the real compile function in a "fake" mode that
2233 enables it to compute how much memory it would need, while actually only
2234 ever using a few hundred bytes of working memory and without too many
2235 tests of the mode. This should make future maintenance and development
2236 easier. A side effect of this work is that the limit of 200 on the nesting
2237 depth of parentheses has been removed (though this was never a serious
2238 limitation, I suspect). However, there is a downside: pcre_compile() now
2239 runs more slowly than before (30% or more, depending on the pattern). I
2240 hope this isn't a big issue. There is no effect on runtime performance.
2241
2242 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2243 newline (only possible for the last line of a file) and it was a
2244 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2245
2246 19. Added additional timing features to pcretest. (1) The -tm option now times
2247 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2248 separate command line item, by a number that specifies the number of
2249 repeats to use when timing. The default is 50000; this gives better
2250 precision, but takes uncomfortably long for very large patterns.
2251
2252 20. Extended pcre_study() to be more clever in cases where a branch of a
2253 subpattern has no definite first character. For example, (a*|b*)[cd] would
2254 previously give no result from pcre_study(). Now it recognizes that the
2255 first character must be a, b, c, or d.
2256
2257 21. There was an incorrect error "recursive call could loop indefinitely" if
2258 a subpattern (or the entire pattern) that was being tested for matching an
2259 empty string contained only one non-empty item after a nested subpattern.
2260 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2261 incorrectly, because the \d was being skipped in the check.
2262
2263 22. The pcretest program now has a new pattern option /B and a command line
2264 option -b, which is equivalent to adding /B to every pattern. This causes
2265 it to show the compiled bytecode, without the additional information that
2266 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2267 is the same as /B/I).
2268
2269 23. A new optimization is now able automatically to treat some sequences such
2270 as a*b as a*+b. More specifically, if something simple (such as a character
2271 or a simple class like \d) has an unlimited quantifier, and is followed by
2272 something that cannot possibly match the quantified thing, the quantifier
2273 is automatically "possessified".
2274
2275 24. A recursive reference to a subpattern whose number was greater than 39
2276 went wrong under certain circumstances in UTF-8 mode. This bug could also
2277 have affected the operation of pcre_study().
2278
2279 25. Realized that a little bit of performance could be had by replacing
2280 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2281
2282 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2283
2284 27. Possessive quantifiers such as a++ were previously implemented by turning
2285 them into atomic groups such as ($>a+). Now they have their own opcodes,
2286 which improves performance. This includes the automatically created ones
2287 from 23 above.
2288
2289 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2290 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2291 the first matched character to be a colon. This applied both to named and
2292 numbered groups.
2293
2294 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2295
2296 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2297 building PCRE on Windows easier, so I have included it in the distribution.
2298
2299 31. There is now a check in pcretest against a ridiculously large number being
2300 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2301 loop, the loop is abandoned.
2302
2303 32. Forward references to subpatterns in conditions such as (?(2)...) where
2304 subpattern 2 is defined later cause pcre_compile() to search forwards in
2305 the pattern for the relevant set of parentheses. This search went wrong
2306 when there were unescaped parentheses in a character class, parentheses
2307 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2308
2309 33. "Subroutine" calls and backreferences were previously restricted to
2310 referencing subpatterns earlier in the regex. This restriction has now
2311 been removed.
2312
2313 34. Added a number of extra features that are going to be in Perl 5.10. On the
2314 whole, these are just syntactic alternatives for features that PCRE had
2315 previously implemented using the Python syntax or my own invention. The
2316 other formats are all retained for compatibility.
2317
2318 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2319 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2320 also .NET compatible.
2321
2322 (b) A recursion or subroutine call to a named group can now be defined as
2323 (?&name) as well as (?P>name).
2324
2325 (c) A backreference to a named group can now be defined as \k<name> or
2326 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2327 5.10, are also .NET compatible.
2328
2329 (d) A conditional reference to a named group can now use the syntax
2330 (?(<name>) or (?('name') as well as (?(name).
2331
2332 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2333 groups (named and numbered) that are never evaluated inline, but can be
2334 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2335 is always false. There may be only one alternative in such a group.
2336
2337 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2338 as the simple (?(R). The condition is true only if the most recent
2339 recursion is that of the given number or name. It does not search out
2340 through the entire recursion stack.
2341
2342 (g) The escape \gN or \g{N} has been added, where N is a positive or
2343 negative number, specifying an absolute or relative reference.
2344
2345 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2346 some "unreachable code" warnings.
2347
2348 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2349 things, this adds five new scripts.
2350
2351 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2352 There were also incompatibilities regarding the handling of \Q..\E inside
2353 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2354 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2355
2356 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2357 matches an empty string, and forcibly breaks the loop. There were bugs in
2358 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2359 against aaaa the result was just "a" rather than "aaaa", for example. Two
2360 separate and independent bugs (that affected different cases) have been
2361 fixed.
2362
2363 39. Refactored the code to abolish the use of different opcodes for small
2364 capturing bracket numbers. This is a tidy that I avoided doing when I
2365 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2366 The new approach is not only tidier, it makes it possible to reduce the
2367 memory needed to fix the previous bug (38).
2368
2369 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2370 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2371 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2372 mode.
2373
2374 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2375 report.
2376
2377 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2378 copy construction and assignment in the C++ wrapper.
2379
2380 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2381 couple of bugs that could have given wrong results in the "--newline=crlf"
2382 case.
2383
2384 44. Added a number of casts and did some reorganization of signed/unsigned int
2385 variables following suggestions from Dair Grant. Also renamed the variable
2386 "this" as "item" because it is a C++ keyword.
2387
2388 45. Arranged for dftables to add
2389
2390 #include "pcre_internal.h"
2391
2392 to pcre_chartables.c because without it, gcc 4.x may remove the array
2393 definition from the final binary if PCRE is built into a static library and
2394 dead code stripping is activated.
2395
2396 46. For an unanchored pattern, if a match attempt fails at the start of a
2397 newline sequence, and the newline setting is CRLF or ANY, and the next two
2398 characters are CRLF, advance by two characters instead of one.
2399
2400
2401 Version 6.7 04-Jul-06
2402 ---------------------
2403
2404 1. In order to handle tests when input lines are enormously long, pcretest has
2405 been re-factored so that it automatically extends its buffers when
2406 necessary. The code is crude, but this _is_ just a test program. The
2407 default size has been increased from 32K to 50K.
2408
2409 2. The code in pcre_study() was using the value of the re argument before
2410 testing it for NULL. (Of course, in any sensible call of the function, it
2411 won't be NULL.)
2412
2413 3. The memmove() emulation function in pcre_internal.h, which is used on
2414 systems that lack both memmove() and bcopy() - that is, hardly ever -
2415 was missing a "static" storage class specifier.
2416
2417 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2418 containing an extended class (one that cannot be represented by a bitmap
2419 because it contains high-valued characters or Unicode property items, e.g.
2420 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2421 pattern, but PCRE should not loop if you do not (it no longer does).
2422 [Detail: two cases were found: (a) a repeated subpattern containing an
2423 extended class; (b) a recursive reference to a subpattern that followed a
2424 previous extended class. It wasn't skipping over the extended class
2425 correctly when UTF-8 mode was not set.]
2426
2427 5. A negated single-character class was not being recognized as fixed-length
2428 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2429 compile error "lookbehind assertion is not fixed length".
2430
2431 6. The RunPerlTest auxiliary script was showing an unexpected difference
2432 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2433 write a Perl script that can interpret lines of an input file either as
2434 byte characters or as UTF-8, which is what "perltest" was being required to
2435 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2436 can't do is switch easily at run time between having the "use utf8;" pragma
2437 or not. In the end, I fudged it by using the RunPerlTest script to insert
2438 "use utf8;" explicitly for the UTF-8 tests.
2439
2440 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2441 the end of the subject string, contrary to the documentation and to what
2442 Perl does. This was true of both matching functions. Now it matches only at
2443 the start of the subject and immediately after *internal* newlines.
2444
2445 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2446 a pointer to an int instead of a pointer to an unsigned long int. This
2447 caused problems on 64-bit systems.
2448
2449 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2450 instance of the 'standard' template library not being so standard".
2451
2452 10. There was no check on the number of named subpatterns nor the maximum
2453 length of a subpattern name. The product of these values is used to compute
2454 the size of the memory block for a compiled pattern. By supplying a very
2455 long subpattern name and a large number of named subpatterns, the size
2456 computation could be caused to overflow. This is now prevented by limiting
2457 the length of names to 32 characters, and the number of named subpatterns
2458 to 10,000.
2459
2460 11. Subpatterns that are repeated with specific counts have to be replicated in
2461 the compiled pattern. The size of memory for this was computed from the
2462 length of the subpattern and the repeat count. The latter is limited to
2463 65535, but there was no limit on the former, meaning that integer overflow
2464 could in principle occur. The compiled length of a repeated subpattern is
2465 now limited to 30,000 bytes in order to prevent this.
2466
2467 12. Added the optional facility to have named substrings with the same name.
2468
2469 13. Added the ability to use a named substring as a condition, using the
2470 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2471 are numbers (not recommended). Forward references are permitted.
2472
2473 14. Added forward references in named backreferences (if you see what I mean).
2474
2475 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2476 pattern could run off the end of the subject. For example, the pattern
2477 "(?s)(.{1,5})"8 did this with the subject "ab".
2478
2479 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2480 PCRE_CASELESS was set when matching characters that were quantified with ?
2481 or *.
2482
2483 17. A character class other than a single negated character that had a minimum
2484 but no maximum quantifier - for example [ab]{6,} - was not handled
2485 correctly by pce_dfa_exec(). It would match only one character.
2486
2487 18. A valid (though odd) pattern that looked like a POSIX character
2488 class but used an invalid character after [ (for example [[,abc,]]) caused
2489 pcre_compile() to give the error "Failed: internal error: code overflow" or
2490 in some cases to crash with a glibc free() error. This could even happen if
2491 the pattern terminated after [[ but there just happened to be a sequence of
2492 letters, a binary zero, and a closing ] in the memory that followed.
2493
2494 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2495 over the years. Originally (before any Unicode support), just the bottom 8
2496 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2497 output from "man perlunicode" includes this:
2498
2499 The regular expression compiler produces polymorphic opcodes. That
2500 is, the pattern adapts to the data and automatically switches to
2501 the Unicode character scheme when presented with Unicode data--or
2502 instead uses a traditional byte scheme when presented with byte
2503 data.
2504
2505 Sadly, a wide octal escape does not cause a switch, and in a string with
2506 no other multibyte characters, these octal escapes are treated as before.
2507 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2508 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2509 Unicode string.
2510
2511 I have not perpetrated such confusion in PCRE. Up till now, it took just
2512 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2513 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2514 translate to the appropriate multibyte character.
2515
2516 29. Applied some refactoring to reduce the number of warnings from Microsoft
2517 and Borland compilers. This has included removing the fudge introduced
2518 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2519 a warning about an unused variable.
2520
2521 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2522 characters since release 4.0, because Perl (from release 5.004) does not.
2523 [Or at least, is documented not to: some releases seem to be in conflict
2524 with the documentation.] However, when a pattern was studied with
2525 pcre_study() and all its branches started with \s, PCRE still included VT
2526 as a possible starting character. Of course, this did no harm; it just
2527 caused an unnecessary match attempt.
2528
2529 22. Removed a now-redundant internal flag bit that recorded the fact that case
2530 dependency changed within the pattern. This was once needed for "required
2531 byte" processing, but is no longer used. This recovers a now-scarce options
2532 bit. Also moved the least significant internal flag bit to the most-
2533 significant bit of the word, which was not previously used (hangover from
2534 the days when it was an int rather than a uint) to free up another bit for
2535 the future.
2536
2537 23. Added support for CRLF line endings as well as CR and LF. As well as the
2538 default being selectable at build time, it can now be changed at runtime
2539 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2540 specify that it is scanning data with non-default line endings.
2541
2542 24. Changed the definition of CXXLINK to make it agree with the definition of
2543 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2544
2545 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2546 recursions. This makes a big different to stack usage for some patterns.
2547
2548 26. If a subpattern containing a named recursion or subroutine reference such
2549 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2550 the space required for the compiled pattern went wrong and gave too small a
2551 value. Depending on the environment, this could lead to "Failed: internal
2552 error: code overflow at offset 49" or "glibc detected double free or
2553 corruption" errors.
2554
2555 27. Applied patches from Google (a) to support the new newline modes and (b) to
2556 advance over multibyte UTF-8 characters in GlobalReplace.
2557
2558 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2559 difference for some implementation of PCRE in some Windows version.
2560
2561 29. Added some extra testing facilities to pcretest:
2562
2563 \q<number> in a data line sets the "match limit" value
2564 \Q<number> in a data line sets the "match recursion limt" value
2565 -S <number> sets the stack size, where <number> is in megabytes
2566
2567 The -S option isn't available for Windows.
2568
2569
2570 Version 6.6 06-Feb-06
2571 ---------------------
2572
2573 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2574 in pcreposix.h. I have copied the definition from pcre.h.
2575
2576 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2577 because pcre.h is no longer a built file.
2578
2579 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2580 not normally included in the compiled code.
2581
2582
2583 Version 6.5 01-Feb-06
2584 ---------------------
2585
2586 1. When using the partial match feature with pcre_dfa_exec(), it was not
2587 anchoring the second and subsequent partial matches at the new starting
2588 point. This could lead to incorrect results. For example, with the pattern
2589 /1234/, partially matching against "123" and then "a4" gave a match.
2590
2591 2. Changes to pcregrep:
2592
2593 (a) All non-match returns from pcre_exec() were being treated as failures
2594 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2595 error message is output. Some extra information is given for the
2596 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2597 probably the only errors that are likely to be caused by users (by
2598 specifying a regex that has nested indefinite repeats, for instance).
2599 If there are more than 20 of these errors, pcregrep is abandoned.
2600
2601 (b) A binary zero was treated as data while matching, but terminated the
2602 output line if it was written out. This has been fixed: binary zeroes
2603 are now no different to any other data bytes.
2604
2605 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2606 used to set a locale for matching. The --locale=xxxx long option has
2607 been added (no short equivalent) to specify a locale explicitly on the
2608 pcregrep command, overriding the environment variables.
2609
2610 (d) When -B was used with -n, some line numbers in the output were one less
2611 than they should have been.
2612
2613 (e) Added the -o (--only-matching) option.
2614
2615 (f) If -A or -C was used with -c (count only), some lines of context were
2616 accidentally printed for the final match.
2617
2618 (g) Added the -H (--with-filename) option.
2619
2620 (h) The combination of options -rh failed to suppress file names for files
2621 that were found from directory arguments.
2622
2623 (i) Added the -D (--devices) and -d (--directories) options.
2624
2625 (j) Added the -F (--fixed-strings) option.
2626
2627 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2628
2629 (l) Added the --colo(u)r option.
2630
2631 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2632 is not present by default.
2633
2634 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2635 items such as (?R) or (?1), when the recursion could match a number of
2636 alternatives. If it matched one of the alternatives, but subsequently,
2637 outside the recursion, there was a failure, the code tried to back up into
2638 the recursion. However, because of the way PCRE is implemented, this is not
2639 possible, and the result was an incorrect result from the match.
2640
2641 In order to prevent this happening, the specification of recursion has
2642 been changed so that all such subpatterns are automatically treated as
2643 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2644
2645 4. I had overlooked the fact that, in some locales, there are characters for
2646 which isalpha() is true but neither isupper() nor islower() are true. In
2647 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2648 and ordfeminine) are like this. This affected the treatment of \w and \W
2649 when they appeared in character classes, but not when they appeared outside
2650 a character class. The bit map for "word" characters is now created
2651 separately from the results of isalnum() instead of just taking it from the
2652 upper, lower, and digit maps. (Plus the underscore character, of course.)
2653
2654 5. The above bug also affected the handling of POSIX character classes such as
2655 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2656 permanent tables. Instead, the bit maps for such a class were previously
2657 created as the appropriate unions of the upper, lower, and digit bitmaps.
2658 Now they are created by subtraction from the [[:word:]] class, which has
2659 its own bitmap.
2660
2661 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2662 It is created by subtracting the vertical space characters (\x09, \x0a,
2663 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2664 subtraction was done in the overall bitmap for a character class, meaning
2665 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2666 be recognized. This bug has been fixed.
2667
2668 7. Patches from the folks at Google:
2669
2670 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2671 real life, but is still worth protecting against".
2672
2673 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2674 regular expressions".
2675
2676 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2677 have it.
2678
2679 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2680 "configure" and the latter not, in order to fix a problem somebody had
2681 with compiling the Arg class on HP-UX.
2682
2683 (e) Improve the error-handling of the C++ wrapper a little bit.
2684
2685 (f) New tests for checking recursion limiting.
2686
2687 8. The pcre_memmove() function, which is used only if the environment does not
2688 have a standard memmove() function (and is therefore rarely compiled),
2689 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2690 returning a result (though PCRE never actually uses the result).
2691
2692 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2693 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2694 returned instead of calling malloc() with an overflowing number that would
2695 most likely cause subsequent chaos.
2696
2697 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2698
2699 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2700 with this option is matched, the nmatch and pmatch options of regexec() are
2701 ignored.
2702
2703 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2704 provided in case anyone wants to the the POSIX interface with UTF-8
2705 strings.
2706
2707 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2708 C++ linking (needed for some HP-UX environments).
2709
2710 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2711 (unused parameter) and in the pcre_printint() function (omitted "default"
2712 switch label when the default is to do nothing).
2713
2714 15. Added some code to make it possible, when PCRE is compiled as a C++
2715 library, to replace subject pointers for pcre_exec() with a smart pointer
2716 class, thus making it possible to process discontinuous strings.
2717
2718 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2719 much the same function. They were added by different people who were trying
2720 to make PCRE easy to compile on non-Unix systems. It has been suggested
2721 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2722 for compiling on Windows systems. I have therefore replaced it with
2723 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2724 defaults to "extern" for C or "extern C" for C++, which works fine on
2725 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2726 SCOPE with something explicit in config.h. In addition:
2727
2728 (a) pcreposix.h still had just "extern" instead of either of these macros;
2729 I have replaced it with PCRE_DATA_SCOPE.
2730
2731 (b) Functions such as _pcre_xclass(), which are internal to the library,
2732 but external in the C sense, all had PCRE_EXPORT in their definitions.
2733 This is apparently wrong for the Windows case, so I have removed it.
2734 (It makes no difference on Unix-like systems.)
2735
2736 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2737 of recursive calls to match(). This is different to MATCH_LIMIT because
2738 that limits the total number of calls to match(), not all of which increase
2739 the depth of recursion. Limiting the recursion depth limits the amount of
2740 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2741 when PCRE is compiled, and changed at run time. A patch from Google adds
2742 this functionality to the C++ interface.
2743
2744 18. Changes to the handling of Unicode character properties:
2745
2746 (a) Updated the table to Unicode 4.1.0.
2747
2748 (b) Recognize characters that are not in the table as "Cn" (undefined).
2749
2750 (c) I revised the way the table is implemented to a much improved format
2751 which includes recognition of ranges. It now supports the ranges that
2752 are defined in UnicodeData.txt, and it also amalgamates other
2753 characters into ranges. This has reduced the number of entries in the
2754 table from around 16,000 to around 3,000, thus reducing its size
2755 considerably. I realized I did not need to use a tree structure after
2756 all - a binary chop search is just as efficient. Having reduced the
2757 number of entries, I extended their size from 6 bytes to 8 bytes to
2758 allow for more data.
2759
2760 (d) Added support for Unicode script names via properties such as \p{Han}.
2761
2762 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2763 matching that character.
2764
2765 20. When matching a repeated Unicode property with a minimum greater than zero,
2766 (for example \pL{2,}), PCRE could look past the end of the subject if it
2767 reached it while seeking the minimum number of characters. This could
2768 happen only if some of the characters were more than one byte long, because
2769 there is a check for at least the minimum number of bytes.
2770
2771 21. Refactored the implementation of \p and \P so as to be more general, to
2772 allow for more different types of property in future. This has changed the
2773 compiled form incompatibly. Anybody with saved compiled patterns that use
2774 \p or \P will have to recompile them.
2775
2776 22. Added "Any" and "L&" to the supported property types.
2777
2778 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2779 but give a compile time error if the value is greater than 0xff.
2780
2781 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2782 accidentally not being installed or uninstalled.
2783
2784 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2785 made were to insert the current release number. This seemed silly, because
2786 it made things harder for people building PCRE on systems that don't run
2787 "configure". I have turned pcre.h into a distributed file, no longer built
2788 by "configure", with the version identification directly included. There is
2789 no longer a pcre.h.in file.
2790
2791 However, this change necessitated a change to the pcre-config script as
2792 well. It is built from pcre-config.in, and one of the substitutions was the
2793 release number. I have updated configure.ac so that ./configure now finds
2794 the release number by grepping pcre.h.
2795
2796 26. Added the ability to run the tests under valgrind.
2797
2798
2799 Version 6.4 05-Sep-05
2800 ---------------------
2801
2802 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2803 "--" to be printed when multiple files were scanned, even when none of the
2804 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2805 consider it to be a bug, and have restored the previous behaviour.
2806
2807 2. A couple of code tidies to get rid of compiler warnings.
2808
2809 3. The pcretest program used to cheat by referring to symbols in the library
2810 whose names begin with _pcre_. These are internal symbols that are not
2811 really supposed to be visible externally, and in some environments it is
2812 possible to suppress them. The cheating is now confined to including
2813 certain files from the library's source, which is a bit cleaner.
2814
2815 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2816 file's purpose clearer.
2817
2818 5. Reorganized pcre_ucp_findchar().
2819
2820
2821 Version 6.3 15-Aug-05
2822 ---------------------
2823
2824 1. The file libpcre.pc.in did not have general read permission in the tarball.
2825
2826 2. There were some problems when building without C++ support:
2827
2828 (a) If C++ support was not built, "make install" and "make test" still
2829 tried to test it.
2830
2831 (b) There were problems when the value of CXX was explicitly set. Some
2832 changes have been made to try to fix these, and ...
2833
2834 (c) --disable-cpp can now be used to explicitly disable C++ support.
2835
2836 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2837 backslash in a target when C++ was disabled. This confuses some
2838 versions of "make", apparently. Using an intermediate variable solves
2839 this. (Same for CPP_LOBJ.)
2840
2841 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2842 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2843 necessary on certain architectures.
2844
2845 4. Added a setting of -export-symbols-regex to the link command to remove
2846 those symbols that are exported in the C sense, but actually are local
2847 within the library, and not documented. Their names all begin with
2848 "_pcre_". This is not a perfect job, because (a) we have to except some
2849 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2850 available (and never for static libraries). I have made a note to try to
2851 find a way round (a) in the future.
2852
2853
2854 Version 6.2 01-Aug-05
2855 ---------------------
2856
2857 1. There was no test for integer overflow of quantifier values. A construction
2858 such as {1111111111111111} would give undefined results. What is worse, if
2859 a minimum quantifier for a parenthesized subpattern overflowed and became
2860 negative, the calculation of the memory size went wrong. This could have
2861 led to memory overwriting.
2862
2863 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2864
2865 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2866 operating environments where this matters.
2867
2868 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2869 PCRE options from within the C++ wrapper.
2870
2871 5. Named capturing subpatterns were not being correctly counted when a pattern
2872 was compiled. This caused two problems: (a) If there were more than 100
2873 such subpatterns, the calculation of the memory needed for the whole
2874 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2875 back references of the form \12, where the number was greater than 9, were
2876 not recognized as back references, even though there were sufficient
2877 previous subpatterns.
2878
2879 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2880 versions of gcc, e.g. 2.95.4.
2881
2882
2883 Version 6.1 21-Jun-05
2884 ---------------------
2885
2886 1. There was one reference to the variable "posix" in pcretest.c that was not
2887 surrounded by "#if !defined NOPOSIX".
2888
2889 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2890 the cross-check on the old pcre_info() function, for the benefit of the
2891 cut-down version of PCRE that is currently imported into Exim.
2892
2893 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2894 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2895 patterns that start (?i) but I don't think that matters. The use of (?i) is
2896 just an example; this all applies to the other options as well.
2897
2898 4. Since libtool seems to echo the compile commands it is issuing, the output
2899 from "make" can be reduced a bit by putting "@" in front of each libtool
2900 compile command.
2901
2902 5. Patch from the folks at Google for configure.in to be a bit more thorough
2903 in checking for a suitable C++ installation before trying to compile the
2904 C++ stuff. This should fix a reported problem when a compiler was present,
2905 but no suitable headers.
2906
2907 6. The man pages all had just "PCRE" as their title. I have changed them to
2908 be the relevant file name. I have also arranged that these names are
2909 retained in the file doc/pcre.txt, which is a concatenation in text format
2910 of all the man pages except the little individual ones for each function.
2911
2912 7. The NON-UNIX-USE file had not been updated for the different set of source
2913 files that come with release 6. I also added a few comments about the C++
2914 wrapper.
2915
2916
2917 Version 6.0 07-Jun-05
2918 ---------------------
2919
2920 1. Some minor internal re-organization to help with my DFA experiments.
2921
2922 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2923 didn't matter for the library itself when fully configured, but did matter
2924 when compiling without UCP support, or within Exim, where the ucp files are
2925 not imported.
2926
2927 3. Refactoring of the library code to split up the various functions into
2928 different source modules. The addition of the new DFA matching code (see
2929 below) to a single monolithic source would have made it really too
2930 unwieldy, quite apart from causing all the code to be include in a
2931 statically linked application, when only some functions are used. This is
2932 relevant even without the DFA addition now that patterns can be compiled in
2933 one application and matched in another.
2934
2935 The downside of splitting up is that there have to be some external
2936 functions and data tables that are used internally in different modules of
2937 the library but which are not part of the API. These have all had their
2938 names changed to start with "_pcre_" so that they are unlikely to clash
2939 with other external names.
2940
2941 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2942 a different (DFA) algorithm. Although it is slower than the original
2943 function, it does have some advantages for certain types of matching
2944 problem.
2945
2946 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2947 including restarting after a partial match.
2948
2949 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2950 defined when compiling for Windows was sent to me. I have put it into the
2951 code, though I have no means of testing or verifying it.
2952
2953 7. Added the pcre_refcount() auxiliary function.
2954
2955 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2956 match before or at the first newline in the subject string. In pcretest,
2957 the /f option on a pattern can be used to set this.
2958
2959 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2960 would behave wrongly. This has been present in PCRE since release 4.0.
2961
2962 10. A number of changes to the pcregrep command:
2963
2964 (a) Refactored how -x works; insert ^(...)$ instead of setting
2965 PCRE_ANCHORED and checking the length, in preparation for adding
2966 something similar for -w.
2967
2968 (b) Added the -w (match as a word) option.
2969
2970 (c) Refactored the way lines are read and buffered so as to have more
2971 than one at a time available.
2972
2973 (d) Implemented a pcregrep test script.
2974
2975 (e) Added the -M (multiline match) option. This allows patterns to match
2976 over several lines of the subject. The buffering ensures that at least
2977 8K, or the rest of the document (whichever is the shorter) is available
2978 for matching (and similarly the previous 8K for lookbehind assertions).
2979
2980 (f) Changed the --help output so that it now says
2981
2982 -w, --word-regex(p)
2983
2984 instead of two lines, one with "regex" and the other with "regexp"
2985 because that confused at least one person since the short forms are the
2986 same. (This required a bit of code, as the output is generated
2987 automatically from a table. It wasn't just a text change.)
2988
2989 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2990 option but starts with a hyphen. Could be a pattern or a path name
2991 starting with a hyphen, for instance.
2992
2993 (h) "-" can be given as a file name to represent stdin.
2994
2995 (i) When file names are being printed, "(standard input)" is used for
2996 the standard input, for compatibility with GNU grep. Previously
2997 "<stdin>" was used.
2998
2999 (j) The option --label=xxx can be used to supply a name to be used for
3000 stdin when file names are being printed. There is no short form.
3001
3002 (k) Re-factored the options decoding logic because we are going to add
3003 two more options that take data. Such options can now be given in four
3004 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3005
3006 (l) Added the -A, -B, and -C options for requesting that lines of context
3007 around matches be printed.
3008
3009 (m) Added the -L option to print the names of files that do not contain
3010 any matching lines, that is, the complement of -l.
3011
3012 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3013 continue to scan other files.
3014
3015 (o) The -s option was incorrectly implemented. For compatibility with other
3016 greps, it now suppresses the error message for a non-existent or non-
3017 accessible file (but not the return code). There is a new option called
3018 -q that suppresses the output of matching lines, which was what -s was
3019 previously doing.
3020
3021 (p) Added --include and --exclude options to specify files for inclusion
3022 and exclusion when recursing.
3023
3024 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3025 Hopefully, it now does.
3026
3027 12. Missing cast in pcre_study().
3028
3029 13. Added an "uninstall" target to the makefile.
3030
3031 14. Replaced "extern" in the function prototypes in Makefile.in with
3032 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3033 world, but is set differently for Windows.
3034
3035 15. Added a second compiling function called pcre_compile2(). The only
3036 difference is that it has an extra argument, which is a pointer to an
3037 integer error code. When there is a compile-time failure, this is set
3038 non-zero, in addition to the error test pointer being set to point to an
3039 error message. The new argument may be NULL if no error number is required
3040 (but then you may as well call pcre_compile(), which is now just a
3041 wrapper). This facility is provided because some applications need a
3042 numeric error indication, but it has also enabled me to tidy up the way
3043 compile-time errors are handled in the POSIX wrapper.
3044
3045 16. Added VPATH=.libs to the makefile; this should help when building with one
3046 prefix path and installing with another. (Or so I'm told by someone who
3047 knows more about this stuff than I do.)
3048
3049 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3050 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3051 match everything, including newlines. This is not POSIX-compatible, but
3052 somebody wanted the feature. From pcretest it can be activated by using
3053 both the P and the s flags.
3054
3055 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3056
3057 19. libpcre.pc was being incorrectly installed as executable.
3058
3059 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3060 it now also looks for '\r' so that it will work unmodified on Windows.
3061
3062 21. Added Google's contributed C++ wrapper to the distribution.
3063
3064 22. Added some untidy missing memory free() calls in pcretest, to keep
3065 Electric Fence happy when testing.
3066
3067
3068
3069 Version 5.0 13-Sep-04
3070 ---------------------
3071
3072 1. Internal change: literal characters are no longer packed up into items
3073 containing multiple characters in a single byte-string. Each character
3074 is now matched using a separate opcode. However, there may be more than one
3075 byte in the character in UTF-8 mode.
3076
3077 2. The pcre_callout_block structure has two new fields: pattern_position and
3078 next_item_length. These contain the offset in the pattern to the next match
3079 item, and its length, respectively.
3080
3081 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3082 insertion of callouts before each pattern item. Added the /C option to
3083 pcretest to make use of this.
3084
3085 4. On the advice of a Windows user, the lines
3086
3087 #if defined(_WIN32) || defined(WIN32)
3088 _setmode( _fileno( stdout ), 0x8000 );
3089 #endif /* defined(_WIN32) || defined(WIN32) */
3090
3091 have been added to the source of pcretest. This apparently does useful
3092 magic in relation to line terminators.
3093
3094 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3095 for the benefit of those environments where the "b" makes a difference.
3096
3097 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3098 to know about it. I have put a hack into configure.in that adds in code
3099 to set GCC=yes if CC=icc. This seems to end up at a point in the
3100 generated configure script that is early enough to affect the setting of
3101 compiler options, which is what is needed, but I have no means of testing
3102 whether it really works. (The user who reported this had patched the
3103 generated configure script, which of course I cannot do.)
3104
3105 LATER: After change 22 below (new libtool files), the configure script
3106 seems to know about icc (and also ecc). Therefore, I have commented out
3107 this hack in configure.in.
3108
3109 7. Added support for pkg-config (2 patches were sent in).
3110
3111 8. Negated POSIX character classes that used a combination of internal tables
3112 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3113 [[:^ascii]]. Typically, they would match almost any characters. The other
3114 POSIX classes were not broken in this way.
3115
3116 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3117 to find the match, as PCRE was deluded into thinking that the match had to
3118 start at the start point or following a newline. The same bug applied to
3119 patterns with negative forward assertions or any backward assertions
3120 preceding ".*" at the start, unless the pattern required a fixed first
3121 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3122
3123 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3124 starting at the last subject character, bytes beyond the end of the subject
3125 string were read.
3126
3127 11. Renamed the variable "class" as "classbits" to make life easier for C++
3128 users. (Previously there was a macro definition, but it apparently wasn't
3129 enough.)
3130
3131 12. Added the new field "tables" to the extra data so that tables can be passed
3132 in at exec time, or the internal tables can be re-selected. This allows
3133 a compiled regex to be saved and re-used at a later time by a different
3134 program that might have everything at different addresses.
3135
3136 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3137 -R library as well as a -L library.
3138
3139 14. The debugging options of pcretest (-d on the command line or D on a
3140 pattern) showed incorrect output for anything following an extended class
3141 that contained multibyte characters and which was followed by a quantifier.
3142
3143 15. Added optional support for general category Unicode character properties
3144 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3145 support. It adds about 90K to the size of the library. The meanings of the
3146 inbuilt class escapes such as \d and \s have NOT been changed.
3147
3148 16. Updated pcredemo.c to include calls to free() to release the memory for the
3149 compiled pattern.
3150
3151 17. The generated file chartables.c was being created in the source directory
3152 instead of in the building directory. This caused the build to fail if the
3153 source directory was different from the building directory, and was
3154 read-only.
3155
3156 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3157 file. No doubt somebody will tell me if they don't make sense... Also added
3158 Dan Mooney's comments about building on OpenVMS.
3159
3160 19. Added support for partial matching via the PCRE_PARTIAL option for
3161 pcre_exec() and the \P data escape in pcretest.
3162
3163 20. Extended pcretest with 3 new pattern features:
3164
3165 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3166 write the compiled pattern to the file whose name is "rest-of-line".
3167 This is a straight binary dump of the data, with the saved pointer to
3168 the character tables forced to be NULL. The study data, if any, is
3169 written too. After writing, pcretest reads a new pattern.
3170
3171 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3172 compiled pattern from the given file. There must not be any
3173 occurrences of "<" in the file name (pretty unlikely); if there are,
3174 pcretest will instead treat the initial "<" as a pattern delimiter.
3175 After reading in the pattern, pcretest goes on to read data lines as
3176 usual.
3177
3178 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3179 and 16-bit fields in a compiled pattern, to simulate a pattern that
3180 was compiled on a host of opposite endianness.
3181
3182 21. The pcre-exec() function can now cope with patterns that were compiled on
3183 hosts of opposite endianness, with this restriction:
3184
3185 As for any compiled expression that is saved and used later, the tables
3186 pointer field cannot be preserved; the extra_data field in the arguments
3187 to pcre_exec() should be used to pass in a tables address if a value
3188 other than the default internal tables were used at compile time.
3189
3190 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3191 now diagnosed as an error. Previously, most of the time, a negative number
3192 would have been treated as zero, but if in addition "ovector" was passed as
3193 NULL, a crash could occur.
3194
3195 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3196 new versions from the libtool 1.5 distribution (the last one is a copy of
3197 a file called libtool.m4). This seems to have fixed the need to patch
3198 "configure" to support Darwin 1.3 (which I used to do). However, I still
3199 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3200 workstation).
3201
3202 24. Changed the PCRE licence to be the more standard "BSD" licence.
3203
3204
3205 Version 4.5 01-Dec-03
3206 ---------------------
3207
3208 1. There has been some re-arrangement of the code for the match() function so
3209 that it can be compiled in a version that does not call itself recursively.
3210 Instead, it keeps those local variables that need separate instances for
3211 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3212 needs to "recurse". Keeping track of where control must go is done by means
3213 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3214 hide most of the details from the main code, and operates only if
3215 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3216 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3217 operating.
3218
3219 To make it easier for callers to provide specially tailored get/free
3220 functions for this usage, two new functions, pcre_stack_malloc, and
3221 pcre_stack_free, are used. They are always called in strict stacking order,
3222 and the size of block requested is always the same.
3223
3224 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3225 PCRE has been compiled to use the stack or the heap for recursion. The
3226 -C option of pcretest uses this to show which version is compiled.
3227
3228 A new data escape \S, is added to pcretest; it causes the amounts of store
3229 obtained and freed by both kinds of malloc/free at match time to be added
3230 to the output.
3231
3232 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3233 what's available on my current Linux desktop machine.
3234
3235 3. When matching a UTF-8 string, the test for a valid string at the start has
3236 been extended. If start_offset is not zero, PCRE now checks that it points
3237 to a byte that is the start of a UTF-8 character. If not, it returns
3238 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3239 this is necessary because there may be backward assertions in the pattern.
3240 When matching the same subject several times, it may save resources to use
3241 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3242
3243 4. The code for checking the validity of UTF-8 strings has been tightened so
3244 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3245 containing "overlong sequences".
3246
3247 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3248 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3249 should have been "&", but it just so happened that all the cases this let
3250 through by mistake were picked up later in the function.
3251
3252 6. I had used a variable called "isblank" - this is a C99 function, causing
3253 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3254
3255 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3256 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3257 so the version is shown; (c) stop "make test" echoing "./RunTest".
3258
3259 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3260
3261 9. The prototype for memmove() for systems that don't have it was using
3262 size_t, but the inclusion of the header that defines size_t was later. I've
3263 moved the #includes for the C headers earlier to avoid this.
3264
3265 10. Added some adjustments to the code to make it easier to compiler on certain
3266 special systems:
3267
3268 (a) Some "const" qualifiers were missing.
3269 (b) Added the macro EXPORT before all exported functions; by default this
3270 is defined to be empty.
3271 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3272 that it reads its output file name as an argument instead of writing
3273 to the standard output and assuming this can be redirected.
3274
3275 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3276 class containing characters with values greater than 255, PCRE compilation
3277 went into a loop.
3278
3279 12. A recursive reference to a subpattern that was within another subpattern
3280 that had a minimum quantifier of zero caused PCRE to crash. For example,
3281 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3282 recursion. If the recursively-called subpattern itself had a zero repeat,
3283 that was OK.
3284
3285 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3286 buffer into which it was copied (for escape processing) was still set at
3287 1024, so long lines caused crashes.
3288
3289 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3290 "internal error: code overflow...". This applied to any character class
3291 that was followed by a possessive quantifier.
3292
3293 15. Modified the Makefile to add libpcre.la as a prerequisite for
3294 libpcreposix.la because I was told this is needed for a parallel build to
3295 work.
3296
3297 16. If a pattern that contained .* following optional items at the start was
3298 studied, the wrong optimizing data was generated, leading to matching
3299 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3300 matching string must start with a or b or c. The correct conclusion for
3301 this pattern is that a match can start with any character.
3302
3303
3304 Version 4.4 13-Aug-03
3305 ---------------------
3306
3307 1. In UTF-8 mode, a character class containing characters with values between
3308 127 and 255 was not handled correctly if the compiled pattern was studied.
3309 In fixing this, I have also improved the studying algorithm for such
3310 classes (slightly).
3311
3312 2. Three internal functions had redundant arguments passed to them. Removal
3313 might give a very teeny performance improvement.
3314
3315 3. Documentation bug: the value of the capture_top field in a callout is *one
3316 more than* the number of the hightest numbered captured substring.
3317
3318 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3319 in incorrectly linking with a previously installed version. They now link
3320 explicitly with libpcre.la.
3321
3322 5. configure.in no longer needs to recognize Cygwin specially.
3323
3324 6. A problem in pcre.in for Windows platforms is fixed.
3325
3326 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3327 pcretest, it used to include the size of the study block as part of its
3328 output. Unfortunately, the structure contains a field that has a different
3329 size on different hardware architectures. This meant that the tests that
3330 showed this size failed. As the block is currently always of a fixed size,
3331 this information isn't actually particularly useful in pcretest output, so
3332 I have just removed it.
3333
3334 8. Three pre-processor statements accidentally did not start in column 1.
3335 Sadly, there are *still* compilers around that complain, even though
3336 standard C has not required this for well over a decade. Sigh.
3337
3338 9. In pcretest, the code for checking callouts passed small integers in the
3339 callout_data field, which is a void * field. However, some picky compilers
3340 complained about the casts involved for this on 64-bit systems. Now
3341 pcretest passes the address of the small integer instead, which should get
3342 rid of the warnings.
3343
3344 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3345 both compile and run time, and gives an error if an invalid UTF-8 sequence
3346 is found. There is a option for disabling this check in cases where the
3347 string is known to be correct and/or the maximum performance is wanted.
3348
3349 11. In response to a bug report, I changed one line in Makefile.in from
3350
3351 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3352 to
3353 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3354
3355 to look similar to other lines, but I have no way of telling whether this
3356 is the right thing to do, as I do not use Windows. No doubt I'll get told
3357 if it's wrong...
3358
3359
3360 Version 4.3 21-May-03
3361 ---------------------
3362
3363 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3364 Makefile.
3365
3366 2. Some refactoring to improve the quality of the code:
3367
3368 (i) The utf8_table... variables are now declared "const".
3369
3370 (ii) The code for \cx, which used the "case flipping" table to upper case
3371 lower case letters, now just substracts 32. This is ASCII-specific,
3372 but the whole concept of \cx is ASCII-specific, so it seems
3373 reasonable.
3374
3375 (iii) PCRE was using its character types table to recognize decimal and
3376 hexadecimal digits in the pattern. This is silly, because it handles
3377 only 0-9, a-f, and A-F, but the character types table is locale-
3378 specific, which means strange things might happen. A private
3379 table is now used for this - though it costs 256 bytes, a table is
3380 much faster than multiple explicit tests. Of course, the standard
3381 character types table is still used for matching digits in subject
3382 strings against \d.
3383
3384 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3385 ending in _t are). So I've renamed it as ESC_tee.
3386
3387 3. The first argument for regexec() in the POSIX wrapper should have been
3388 defined as "const".
3389
3390 4. Changed pcretest to use malloc() for its buffers so that they can be
3391 Electric Fenced for debugging.
3392
3393 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3394 to read one or more bytes before the start of the subject string. Often this
3395 had no effect on PCRE's behaviour, but in some circumstances it could
3396 provoke a segmentation fault.
3397
3398 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3399 to try to read one or more bytes before the start of the subject string.
3400
3401 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3402 UTF-8 support could misbehave in various ways if the subject string
3403 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3404 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3405 back over UTF-8 characters.)
3406
3407
3408 Version 4.2 14-Apr-03
3409 ---------------------
3410
3411 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3412
3413 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3414 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3415 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3416 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3417 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3418 and BUILD_EXEEXT
3419 Note: automatic setting of the BUILD variables is not yet working
3420 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3421 compile-time but not at link-time
3422 [LINK]: use for linking executables only
3423 make different versions for Windows and non-Windows
3424 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3425 libraries
3426 [LINK_FOR_BUILD]: new variable
3427 [OBJEXT]: use throughout
3428 [EXEEXT]: use throughout
3429 <winshared>: new target
3430 <wininstall>: new target
3431 <dftables.o>: use native compiler
3432 <dftables>: use native linker
3433 <install>: handle Windows platform correctly
3434 <clean>: ditto
3435 <check>: ditto
3436 copy DLL to top builddir before testing
3437
3438 As part of these changes, -no-undefined was removed again. This was reported
3439 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3440 in any case.
3441
3442 3. Some tidies to get rid of compiler warnings:
3443
3444 . In the match_data structure, match_limit was an unsigned long int, whereas
3445 match_call_count was an int. I've made them both unsigned long ints.
3446
3447 . In pcretest the fact that a const uschar * doesn't automatically cast to
3448 a void * provoked a warning.
3449
3450 . Turning on some more compiler warnings threw up some "shadow" variables
3451 and a few more missing casts.
3452
3453 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3454 option, a class that contained a single character with a value between 128
3455 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3456
3457 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3458 option, a class that contained several characters, but with at least one
3459 whose value was between 128 and 255 caused PCRE to crash.
3460
3461
3462 Version 4.1 12-Mar-03
3463 ---------------------
3464
3465 1. Compiling with gcc -pedantic found a couple of places where casts were
3466 needed, and a string in dftables.c that was longer than standard compilers are
3467 required to support.
3468
3469 2. Compiling with Sun's compiler found a few more places where the code could
3470 be tidied up in order to avoid warnings.
3471
3472 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3473 first of these names is deprecated in the latest Autoconf in favour of the name
3474 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3475 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3476 analogy I have changed it to CFLAGS_FOR_BUILD.
3477
3478 4. Added -no-undefined to the linking command in the Makefile, because this is
3479 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3480 linking step for the pcreposix library.
3481
3482 5. PCRE was failing to diagnose the case of two named groups with the same
3483 name.
3484
3485 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3486 literal character that is needed in the subject for a match, and scans along to
3487 ensure that it is present before embarking on the full matching process. This
3488 saves time in cases of nested unlimited repeats that are never going to match.
3489 Problem: the scan can take a lot of time if the subject is very long (e.g.
3490 megabytes), thus penalizing straightforward matches. It is now done only if the
3491 amount of subject to be scanned is less than 1000 bytes.
3492
3493 7. A lesser problem with the same optimization is that it was recording the
3494 first character of an anchored pattern as "needed", thus provoking a search
3495 right along the subject, even when the first match of the pattern was going to
3496 fail. The "needed" character is now not set for anchored patterns, unless it
3497 follows something in the pattern that is of non-fixed length. Thus, it still
3498 fulfils its original purpose of finding quick non-matches in cases of nested
3499 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3500
3501
3502 Version 4.0 17-Feb-03
3503 ---------------------
3504
3505 1. If a comment in an extended regex that started immediately after a meta-item
3506 extended to the end of string, PCRE compiled incorrect data. This could lead to
3507 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3508
3509 2. Moved to autoconf 2.53 and libtool 1.4.2.
3510
3511 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3512 the special perltest8 script is no longer needed - all the tests can be run
3513 from a single perltest script.
3514
3515 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3516 by \s. It has now been removed in PCRE. This means it isn't recognized as
3517 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3518 class [:space:] *does* include VT, thereby creating a mess.
3519
3520 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3521 space and tab.
3522
3523 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3524 its new features into the main test script, reducing the number of scripts.
3525
3526 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3527 were backward compatible, and made the (?i) apply to the whole pattern, as if
3528 /i were given. Now it behaves more logically, and applies the option setting
3529 only to what follows. PCRE has been changed to follow suit. However, if it
3530 finds options settings right at the start of the pattern, it extracts them into
3531 the global options, as before. Thus, they show up in the info data.
3532
3533 8. Added support for the \Q...\E escape sequence. Characters in between are
3534 treated as literals. This is slightly different from Perl in that $ and @ are
3535 also handled as literals inside the quotes. In Perl, they will cause variable
3536 interpolation. Note the following examples:
3537
3538 Pattern PCRE matches Perl matches
3539
3540 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3541 \Qabc\$xyz\E abc\$xyz abc\$xyz
3542 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3543
3544 For compatibility with Perl, \Q...\E sequences are recognized inside character
3545 classes as well as outside them.
3546
3547 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3548 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3549 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3550 signed/unsigned warnings.
3551
3552 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3553 option for pcretest, so I've replaced it by a simple function that does just
3554 that job.
3555
3556 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3557 "pcregrep -".
3558
3559 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3560 Java package. This provides some syntactic sugar for simple cases of what my
3561 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3562 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3563 item, you can use this simplified notation. Note that only makes sense with
3564 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3565 greediness, whatever the setting of the PCRE_UNGREEDY option.
3566
3567 13. A change of greediness default within a pattern was not taking effect at
3568 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3569 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3570 was abstracted outside.
3571
3572 14. PCRE now supports the \G assertion. It is true when the current matching
3573 position is at the start point of the match. This differs from \A when the
3574 starting offset is non-zero. Used with the /g option of pcretest (or similar
3575 code), it works in the same way as it does for Perl's /g option. If all
3576 alternatives of a regex begin with \G, the expression is anchored to the start
3577 match position, and the "anchored" flag is set in the compiled expression.
3578
3579 15. Some bugs concerning the handling of certain option changes within patterns
3580 have been fixed. These applied to options other than (?ims). For example,
3581 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3582 been the other way round. Some of this was related to change 7 above.
3583
3584 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3585 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3586 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3587 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3588
3589 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3590 mode. Unlike ".", it always matches newline, whatever the setting of
3591 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3592 assertions. Perl allows it, but it doesn't (in general) work because it can't
3593 calculate the length of the lookbehind. At least, that's the case for Perl
3594 5.8.0 - I've been told they are going to document that it doesn't work in
3595 future.
3596
3597 18. Added an error diagnosis for escapes that PCRE does not support: these are
3598 \L, \l, \N, \P, \p, \U, \u, and \X.
3599
3600 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3601 reading past the end of the pattern in cases such as /[abcd/.
3602
3603 20. PCRE was getting more memory than necessary for patterns with classes that
3604 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3605
3606 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3607 compiling PCRE for use with Virtual Pascal.
3608
3609 22. Small fix to the Makefile to make it work properly if the build is done
3610 outside the source tree.
3611
3612 23. Added a new extension: a condition to go with recursion. If a conditional
3613 subpattern starts with (?(R) the "true" branch is used if recursion has
3614 happened, whereas the "false" branch is used only at the top level.
3615
3616 24. When there was a very long string of literal characters (over 255 bytes
3617 without UTF support, over 250 bytes with UTF support), the computation of how
3618 much memory was required could be incorrect, leading to segfaults or other
3619 strange effects.
3620
3621 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3622 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3623 there was a subsequent back reference to those brackets. This meant that, for
3624 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3625 possible to check for precisely this case. All we can do is abandon the
3626 optimization if .* occurs inside capturing brackets when there are any back
3627 references whatsoever. (See below for a better fix that came later.)
3628
3629 26. The handling of the optimization for finding the first character of a
3630 non-anchored pattern, and for finding a character that is required later in the
3631 match were failing in some cases. This didn't break the matching; it just
3632 failed to optimize when it could. The way this is done has been re-implemented.
3633
3634 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3635
3636 28. Added a new feature that provides some of the functionality that Perl
3637 provides with (?{...}). The facility is termed a "callout". The way it is done
3638 in PCRE is for the caller to provide an optional function, by setting
3639 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3640 global variable. By default it is unset, which disables all calling out. To get
3641 the function called, the regex must include (?C) at appropriate points. This
3642 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3643 This provides a means of identifying different callout points. When PCRE
3644 reaches such a point in the regex, if pcre_callout has been set, the external
3645 function is called. It is provided with data in a structure called
3646 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3647 matching continues; if it returns a non-zero value, the match at the current
3648 point fails. However, backtracking will occur if possible. [This was changed
3649 later and other features added - see item 49 below.]
3650
3651 29. pcretest is upgraded to test the callout functionality. It provides a
3652 callout function that displays information. By default, it shows the start of
3653 the match and the current position in the text. There are some new data escapes
3654 to vary what happens:
3655
3656 \C+ in addition, show current contents of captured substrings
3657 \C- do not supply a callout function
3658 \C!n return 1 when callout number n is reached
3659 \C!n!m return 1 when callout number n is reached for the mth time
3660
3661 30. If pcregrep was called with the -l option and just a single file name, it
3662 output "<stdin>" if a match was found, instead of the file name.
3663
3664 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3665 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3666 pcre_exec(). This saves a malloc/free per call. The default value of
3667 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3668 when configuring.
3669
3670 32. The default maximum size of a compiled pattern is 64K. There have been a
3671 few cases of people hitting this limit. The code now uses macros to handle the
3672 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3673 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3674 configuring. Tests 2 and 5 work only with 2-byte links because they output
3675 debugging information about compiled patterns.
3676
3677 33. Internal code re-arrangements:
3678
3679 (a) Moved the debugging function for printing out a compiled regex into
3680 its own source file (printint.c) and used #include to pull it into
3681 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3682 separate copies.
3683
3684 (b) Defined the list of op-code names for debugging as a macro in
3685 internal.h so that it is next to the definition of the opcodes.
3686
3687 (c) Defined a table of op-code lengths for simpler skipping along compiled
3688 code. This is again a macro in internal.h so that it is next to the
3689 definition of the opcodes.
3690
3691 34. Added support for recursive calls to individual subpatterns, along the
3692 lines of Robin Houston's patch (but implemented somewhat differently).
3693
3694 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3695 allow it to read and process whole directories in Win32. This code was
3696 contributed by Lionel Fourquaux; it has not been tested by me.
3697
3698 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3699 used to name a group. Names consist of alphanumerics and underscores, and must
3700 be unique. Back references use the syntax (?P=name) and recursive calls use
3701 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3702 numbers. The function pcre_fullinfo() can be used after compilation to extract
3703 a name/number map. There are three relevant calls:
3704
3705 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3706 PCRE_INFO_NAMECOUNT yields the number of entries
3707 PCRE_INFO_NAMETABLE yields a pointer to the map.
3708
3709 The map is a vector of fixed-size entries. The size of each entry depends on
3710 the length of the longest name used. The first two bytes of each entry are the
3711 group number, most significant byte first. There follows the corresponding
3712 name, zero terminated. The names are in alphabetical order.
3713
3714 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3715 case instead of 255. Making it the same both with and without UTF-8 support
3716 means that the same test output works with both.
3717
3718 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3719 calling malloc() with a zero argument.
3720
3721 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3722 optimization. I've improved things by keeping a bitmap of backreferences with
3723 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3724 fact referenced, the optimization can be applied. It is unlikely that a
3725 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3726 the match to follow \n) will appear inside brackets with a number greater than
3727 31, but if it does, any back reference > 31 suppresses the optimization.
3728
3729 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3730 of disabling numbered capturing parentheses. Any opening parenthesis that is
3731 not followed by ? behaves as if it were followed by ?: but named parentheses
3732 can still be used for capturing (and they will acquire numbers in the usual
3733 way).
3734
3735 41. Redesigned the return codes from the match() function into yes/no/error so
3736 that errors can be passed back from deep inside the nested calls. A malloc
3737 failure while inside a recursive subpattern call now causes the
3738 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3739
3740 42. It is now possible to set a limit on the number of times the match()
3741 function is called in a call to pcre_exec(). This facility makes it possible to
3742 limit the amount of recursion and backtracking, though not in a directly
3743 obvious way, because the match() function is used in a number of different
3744 circumstances. The count starts from zero for each position in the subject
3745 string (for non-anchored patterns). The default limit is, for compatibility, a
3746 large number, namely 10 000 000. You can change this in two ways:
3747
3748 (a) When configuring PCRE before making, you can use --with-match-limit=n
3749 to set a default value for the compiled library.
3750
3751 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3752 a different value is set. See 45 below.
3753
3754 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3755
3756 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3757 of things that can be changed at compile time. The first argument specifies
3758 what is wanted and the second points to where the information is to be placed.
3759 The current list of available information is:
3760
3761 PCRE_CONFIG_UTF8
3762
3763 The output is an integer that is set to one if UTF-8 support is available;
3764 otherwise it is set to zero.
3765
3766 PCRE_CONFIG_NEWLINE
3767
3768 The output is an integer that it set to the value of the code that is used for
3769 newline. It is either LF (10) or CR (13).
3770
3771 PCRE_CONFIG_LINK_SIZE
3772
3773 The output is an integer that contains the number of bytes used for internal
3774 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3775
3776 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3777
3778 The output is an integer that contains the threshold above which the POSIX
3779 interface uses malloc() for output vectors. See item 31 above.
3780
3781 PCRE_CONFIG_MATCH_LIMIT
3782
3783 The output is an unsigned integer that contains the default limit of the number
3784 of match() calls in a pcre_exec() execution. See 42 above.
3785
3786 44. pcretest has been upgraded by the addition of the -C option. This causes it
3787 to extract all the available output from the new pcre_config() function, and to
3788 output it. The program then exits immediately.
3789
3790 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3791 order to support additional features. One way would have been to define
3792 pcre_exec2() (for example) with extra arguments, but this would not have been
3793 extensible, and would also have required all calls to the original function to
3794 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3795 is used for passing in "extra" data from pcre_study().
3796
3797 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3798 contains the following fields:
3799
3800 flags a bitmap indicating which of the following fields are set
3801 study_data opaque data from pcre_study()
3802 match_limit a way of specifying a limit on match() calls for a specific
3803 call to pcre_exec()
3804 callout_data data for callouts (see 49 below)
3805
3806 The flag bits are also defined in pcre.h, and are
3807
3808 PCRE_EXTRA_STUDY_DATA
3809 PCRE_EXTRA_MATCH_LIMIT
3810 PCRE_EXTRA_CALLOUT_DATA
3811
3812 The pcre_study() function now returns one of these new pcre_extra blocks, with
3813 the actual study data pointed to by the study_data field, and the
3814 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3815 before. That is, this change is entirely upwards-compatible and requires no
3816 change to existing code.
3817
3818 If you want to pass in additional data to pcre_exec(), you can either place it
3819 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3820 block.
3821
3822 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3823 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3824 times with different match limits, until it finds the minimum value needed for
3825 pcre_exec() to complete. The value is then output. This can be instructive; for
3826 most simple matches the number is quite small, but for pathological cases it
3827 gets very large very quickly.
3828
3829 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3830 returns the size of the data block pointed to by the study_data field in a
3831 pcre_extra block, that is, the value that was passed as the argument to
3832 pcre_malloc() when PCRE was getting memory in which to place the information
3833 created by pcre_study(). The fourth argument should point to a size_t variable.
3834 pcretest has been extended so that this information is shown after a successful
3835 pcre_study() call when information about the compiled regex is being displayed.
3836
3837 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3838 because what follows is always an absolute path. (Later: it turns out that this
3839 is more than cosmetic for MinGW, because it doesn't like empty path
3840 components.)
3841
3842 49. Some changes have been made to the callout feature (see 28 above):
3843
3844 (i) A callout function now has three choices for what it returns:
3845
3846 0 => success, carry on matching
3847 > 0 => failure at this point, but backtrack if possible
3848 < 0 => serious error, return this value from pcre_exec()
3849
3850 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3851 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3852 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3853 use by callout functions. It will never be used by PCRE itself.
3854
3855 (ii) The pcre_extra structure (see 45 above) has a void * field called
3856 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3857 pcre_callout_block structure has a field of the same name. The contents of
3858 the field passed in the pcre_extra structure are passed to the callout
3859 function in the corresponding field in the callout block. This makes it
3860 easier to use the same callout-containing regex from multiple threads. For
3861 testing, the pcretest program has a new data escape
3862
3863 \C*n pass the number n (may be negative) as callout_data
3864
3865 If the callout function in pcretest receives a non-zero value as
3866 callout_data, it returns that value.
3867
3868 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3869 there were some redundant $(CFLAGS) in commands that are now specified as
3870 $(LINK), which already includes $(CFLAGS).
3871
3872 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3873 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3874 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3875 one-byte characters throughout. Note that case-insensitive matching applies
3876 only to characters whose values are less than 256. PCRE doesn't support the
3877 notion of cases for higher-valued characters.
3878
3879 (i) A character class whose characters are all within 0-255 is handled as
3880 a bit map, and the map is inverted for negative classes. Previously, a
3881 character > 255 always failed to match such a class; however it should
3882 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3883
3884 (ii) A negated character class with a single character < 255 is coded as
3885 "not this character" (OP_NOT). This wasn't working properly when the test
3886 character was multibyte, either singly or repeated.
3887
3888 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3889 mode, for example: \x{100}{2,3}.
3890
3891 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3892 singly or repeated) now correctly test multibyte characters. However,
3893 PCRE doesn't recognize any characters with values greater than 255 as
3894 digits, spaces, or word characters. Such characters always match \D, \S,
3895 and \W, and never match \d, \s, or \w.
3896
3897 (v) Classes may now contain characters and character ranges with values
3898 greater than 255. For example: [ab\x{100}-\x{400}].
3899
3900 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3901 PCRE in UTF-8 mode.
3902
3903 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3904 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3905 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3906 value.)
3907
3908 53. The single man page has become too large. I have therefore split it up into
3909 a number of separate man pages. These also give rise to individual HTML pages;
3910 these are now put in a separate directory, and there is an index.html page that
3911 lists them all. Some hyperlinking between the pages has been installed.
3912
3913 54. Added convenience functions for handling named capturing parentheses.
3914
3915 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3916 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3917 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3918 are faulted.
3919
3920 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3921 calling configure. These values are used when compiling the dftables.c program
3922 which is run to generate the source of the default character tables. They
3923 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3924 you will need to set these values.
3925
3926 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3927
3928
3929 Version 3.9 02-Jan-02
3930 ---------------------
3931
3932 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3933
3934 2. If --disable-static was given, the building process failed when trying to
3935 build pcretest and pcregrep. (For some reason it was using libtool to compile
3936 them, which is not right, as they aren't part of the library.)
3937
3938
3939 Version 3.8 18-Dec-01
3940 ---------------------
3941
3942 1. The experimental UTF-8 code was completely screwed up. It was packing the
3943 bytes in the wrong order. How dumb can you get?
3944
3945
3946 Version 3.7 29-Oct-01
3947 ---------------------
3948
3949 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3950 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3951 this didn't happen under Solaris 8, where I normally test things.
3952
3953 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3954 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3955 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3956 this fix an hour or so after the initial 3.7 release.)
3957
3958
3959 Version 3.6 23-Oct-01
3960 ---------------------
3961
3962 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3963 offsets passed as NULL with zero offset count.
3964
3965 2. The config.guess and config.sub files had not been updated when I moved to
3966 the latest autoconf.
3967
3968
3969 Version 3.5 15-Aug-01
3970 ---------------------
3971
3972 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3973 had been forgotten.
3974
3975 2. By using declared but undefined structures, we can avoid using "void"
3976 definitions in pcre.h while keeping the internal definitions of the structures
3977 private.
3978
3979 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3980 user point of view, this means that both static and shared libraries are built
3981 by default, but this can be individually controlled. More of the work of
3982 handling this static/shared cases is now inside libtool instead of PCRE's make
3983 file.
3984
3985 4. The pcretest utility is now installed along with pcregrep because it is
3986 useful for users (to test regexs) and by doing this, it automatically gets
3987 relinked by libtool. The documentation has been turned into a man page, so
3988 there are now .1, .txt, and .html versions in /doc.
3989
3990 5. Upgrades to pcregrep:
3991 (i) Added long-form option names like gnu grep.
3992 (ii) Added --help to list all options with an explanatory phrase.
3993 (iii) Added -r, --recursive to recurse into sub-directories.
3994 (iv) Added -f, --file to read patterns from a file.
3995
3996 6. pcre_exec() was referring to its "code" argument before testing that
3997 argument for NULL (and giving an error if it was NULL).
3998
3999 7. Upgraded Makefile.in to allow for compiling in a different directory from
4000 the source directory.
4001
4002 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4003 options bits, the pointer it was passed was to an int instead of to an unsigned
4004 long int. This mattered only on 64-bit systems.
4005
4006 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4007 generated) instead of pcre.in, which it its source. Also made the same change
4008 in several of the .c files.
4009
4010 10. A new release of gcc defines printf() as a macro, which broke pcretest
4011 because it had an ifdef in the middle of a string argument for printf(). Fixed
4012 by using separate calls to printf().
4013
4014 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4015 script, to force use of CR or LF instead of \n in the source. On non-Unix
4016 systems, the value can be set in config.h.
4017
4018 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4019 absolute limit. Changed the text of the error message to make this clear, and
4020 likewise updated the man page.
4021
4022 13. The limit of 99 on the number of capturing subpatterns has been removed.
4023 The new limit is 65535, which I hope will not be a "real" limit.
4024
4025
4026 Version 3.4 22-Aug-00
4027 ---------------------
4028
4029 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4030
4031 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4032
4033
4034 Version 3.3 01-Aug-00
4035 ---------------------
4036
4037 1. If an octal character was given, but the value was greater than \377, it
4038 was not getting masked to the least significant bits, as documented. This could
4039 lead to crashes in some systems.
4040
4041 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4042 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4043
4044 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4045 These just pass their arguments on to (pcre_free)(), but they are provided
4046 because some uses of PCRE bind it to non-C systems that can call its functions,
4047 but cannot call free() or pcre_free() directly.
4048
4049 4. Add "make test" as a synonym for "make check". Corrected some comments in
4050 the Makefile.
4051
4052 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4053 Makefile.
4054
4055 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4056 command called pgrep for grepping around the active processes.
4057
4058 7. Added the beginnings of support for UTF-8 character strings.
4059
4060 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4061 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4062 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4063 out for the ar command.)
4064
4065
4066 Version 3.2 12-May-00
4067 ---------------------
4068
4069 This is purely a bug fixing release.
4070
4071 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4072 of ZA. This was just one example of several cases that could provoke this bug,
4073 which was introduced by change 9 of version 2.00. The code for breaking
4074 infinite loops after an iteration that matches an empty string was't working
4075 correctly.
4076
4077 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4078 when matched against abbab (for example). After matching an empty string, it
4079 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4080 caused it to match further down the string than it should.
4081
4082 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4083 was there because it doesn't seem to be needed, and it causes trouble on some
4084 systems, as it is not a Standard C header. It has been removed.
4085
4086 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4087 were reported on the Macintosh. The changes were from
4088
4089 while ((c = *(++ptr)) != 0 && c != '\n');
4090 to
4091 while ((c = *(++ptr)) != 0 && c != '\n') ;
4092
4093 Totally extraordinary, but if that's what it takes...
4094
4095 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4096 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4097 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4098 assumes the way PCRE uses memmove() (always moving upwards).
4099
4100 6. PCRE is being used in one environment where strchr() is not available. There
4101 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4102 faster code anyway.
4103
4104
4105 Version 3.1 09-Feb-00
4106 ---------------------
4107
4108 The only change in this release is the fixing of some bugs in Makefile.in for
4109 the "install" target:
4110
4111 (1) It was failing to install pcreposix.h.
4112
4113 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4114
4115
4116 Version 3.0 01-Feb-00
4117 ---------------------
4118
4119 1. Add support for the /+ modifier to perltest (to output $` like it does in
4120 pcretest).
4121
4122 2. Add support for the /g modifier to perltest.
4123
4124 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4125 matches null strings.
4126
4127 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4128 pattern. Perl treats empty patterns specially - it reuses the most recent
4129 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4130 effect.
4131
4132 5. The POSIX interface was broken in that it was just handing over the POSIX
4133 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4134 required a bigger vector, with some working space on the end. This means that
4135 the POSIX wrapper now has to get and free some memory, and copy the results.
4136
4137 6. Added some simple autoconf support, placing the test data and the
4138 documentation in separate directories, re-organizing some of the
4139 information files, and making it build pcre-config (a GNU standard). Also added
4140 libtool support for building PCRE as a shared library, which is now the
4141 default.
4142
4143 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4144 09 are not valid octal constants. Single digits will be used for minor values
4145 less than 10.
4146
4147 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4148 existing programs that set these in the POSIX interface can use PCRE without
4149 modification.
4150
4151 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4152 return all that pcre_info() returns, plus additional data. The pcre_info()
4153 function is retained for compatibility, but is considered to be obsolete.
4154
4155 10. Added experimental recursion feature (?R) to handle one common case that
4156 Perl 5.6 will be able to do with (?p{...}).
4157
4158 11. Added support for POSIX character classes like [:alpha:], which Perl is
4159 adopting.
4160
4161
4162 Version 2.08 31-Aug-99
4163 ----------------------
4164
4165 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4166 trying to match at the startoffset position, but instead was moving forward to
4167 the next newline as if a previous match had failed.
4168
4169 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4170 and could get into a loop if a null string was matched other than at the start
4171 of the subject.
4172
4173 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4174 be distinguished at compile time, and for completeness also added PCRE_DATE.
4175
4176 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4177 in GnuWin32 environments.
4178
4179
4180 Version 2.07 29-Jul-99
4181 ----------------------
4182
4183 1. The documentation is now supplied in plain text form and HTML as well as in
4184 the form of man page sources.
4185
4186 2. C++ compilers don't like assigning (void *) values to other pointer types.
4187 In particular this affects malloc(). Although there is no problem in Standard
4188 C, I've put in casts to keep C++ compilers happy.
4189
4190 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4191 should be (const char *).
4192
4193 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4194 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4195 However, I haven't made this a standard facility. The documentation doesn't
4196 mention it, and the Makefile doesn't support it.
4197
4198 5. The Makefile now contains an "install" target, with editable destinations at
4199 the top of the file. The pcretest program is not installed.
4200
4201 6. pgrep -V now gives the PCRE version number and date.
4202
4203 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4204 causing the entire string to be ignored, instead of just the last character.
4205
4206 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4207 non-matching string, it can take a very, very long time, even for strings of
4208 quite modest length, because of the nested recursion. PCRE now does better in
4209 some of these cases. It does this by remembering the last required literal
4210 character in the pattern, and pre-searching the subject to ensure it is present
4211 before running the real match. In other words, it applies a heuristic to detect
4212 some types of certain failure quickly, and in the above example, if presented
4213 with a string that has no trailing " it gives "no match" very quickly.
4214
4215 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4216 other alternatives are tried instead.
4217
4218
4219 Version 2.06 09-Jun-99
4220 ----------------------
4221
4222 1. Change pcretest's output for amount of store used to show just the code
4223 space, because the remainder (the data block) varies in size between 32-bit and
4224 64-bit systems.
4225
4226 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4227 start matching at. This allows lookbehinds to work when searching for multiple
4228 occurrences in a string.
4229
4230 3. Added additional options to pcretest for testing multiple occurrences:
4231
4232 /+ outputs the rest of the string that follows a match
4233 /g loops for multiple occurrences, using the new startoffset argument
4234 /G loops for multiple occurrences by passing an incremented pointer
4235
4236 4. PCRE wasn't doing the "first character" optimization for patterns starting
4237 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4238 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4239 the letter 'x'. On long subject strings, this gives a significant speed-up.
4240
4241
4242 Version 2.05 21-Apr-99
4243 ----------------------
4244
4245 1. Changed the type of magic_number from int to long int so that it works
4246 properly on 16-bit systems.
4247
4248 2. Fixed a bug which caused patterns starting with .* not to work correctly
4249 when the subject string contained newline characters. PCRE was assuming
4250 anchoring for such patterns in all cases, which is not correct because .* will
4251 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4252 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4253 must be retried after every newline in the subject.
4254
4255
4256 Version 2.04 18-Feb-99
4257 ----------------------
4258
4259 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4260 computation of the store needed to hold the pattern was incorrect (too large).
4261 If such patterns were nested a few deep, this could multiply and become a real
4262 problem.
4263
4264 2. Added /M option to pcretest to show the memory requirement of a specific
4265 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4266
4267 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4268 compiled in such a way that the backtracking after subsequent failure was
4269 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4270 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4271
4272
4273 Version 2.03 02-Feb-99
4274 ----------------------
4275
4276 1. Fixed typo and small mistake in man page.
4277
4278 2. Added 4th condition (GPL supersedes if conflict) and created separate
4279 LICENCE file containing the conditions.
4280
4281 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4282 Perl, that is the internal \ allows the delimiter to be included in the
4283 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4284 the final delimiter, add \ to the end of the pattern (to test the error).
4285
4286 4. Added the convenience functions for extracting substrings after a successful
4287 match. Updated pcretest to make it able to test these functions.
4288
4289
4290 Version 2.02 14-Jan-99
4291 ----------------------
4292
4293 1. Initialized the working variables associated with each extraction so that
4294 their saving and restoring doesn't refer to uninitialized store.
4295
4296 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4297 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4298 fix the problem.
4299
4300 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4301 calls, and wasn't printing the correct value for compiling calls. Increased the
4302 default value of LOOPREPEAT, and the number of significant figures in the
4303 times.
4304
4305 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4306
4307 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4308 a building problem on Windows NT with a FAT file system.
4309
4310
4311 Version 2.01 21-Oct-98
4312 ----------------------
4313
4314 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4315 to character tables built by pcre_maketables() in the current locale. If NULL
4316 is passed, the default tables are used.
4317
4318
4319 Version 2.00 24-Sep-98
4320 ----------------------
4321
4322 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4323 it any more.
4324
4325 2. Allow quantification of (?>) groups, and make it work correctly.
4326
4327 3. The first character computation wasn't working for (?>) groups.
4328
4329 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4330 end of the subject) and add 5.005's \z, which really does match only at the
4331 very end of the subject.
4332
4333 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4334
4335 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4336 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4337 localized options. All options to pcre_study() were also removed.
4338
4339 7. Add other new features from 5.005:
4340
4341 $(?<= positive lookbehind
4342 $(?<! negative lookbehind
4343 (?imsx-imsx) added the unsetting capability
4344 such a setting is global if at outer level; local otherwise
4345 (?imsx-imsx:) non-capturing groups with option setting
4346 (?(cond)re|re) conditional pattern matching
4347
4348 A backreference to itself in a repeated group matches the previous
4349 captured string.
4350
4351 8. General tidying up of studying (both automatic and via "study")
4352 consequential on the addition of new assertions.
4353
4354 9. As in 5.005, unlimited repeated groups that could match an empty substring
4355 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4356 runtime if any iteration does actually match an empty substring.
4357
4358 10. Include the RunTest script in the distribution.
4359
4360 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4361 discrepancies, some of which were old and were also with respect to 5.004. They
4362 have now been fixed.
4363
4364
4365 Version 1.09 28-Apr-98
4366 ----------------------
4367
4368 1. A negated single character class followed by a quantifier with a minimum
4369 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4370 program crashes, or just wrong answers. This did not apply to negated classes
4371 containing more than one character, or to minima other than one.
4372
4373
4374 Version 1.08 27-Mar-98
4375 ----------------------
4376
4377 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4378
4379 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4380 latter must appear before anything that relies on it in the pattern.
4381
4382
4383 Version 1.07 16-Feb-98
4384 ----------------------
4385
4386 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4387 repeat of a potentially empty string).
4388
4389
4390 Version 1.06 23-Jan-98
4391 ----------------------
4392
4393 1. Added Markus Oberhumer's little patches for C++.
4394
4395 2. Literal strings longer than 255 characters were broken.
4396
4397
4398 Version 1.05 23-Dec-97
4399 ----------------------
4400
4401 1. Negated character classes containing more than one character were failing if
4402 PCRE_CASELESS was set at run time.
4403
4404
4405 Version 1.04 19-Dec-97
4406 ----------------------
4407
4408 1. Corrected the man page, where some "const" qualifiers had been omitted.
4409
4410 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4411 input syntax.
4412
4413 3. Fixed memory leak which occurred when a regex with back references was
4414 matched with an offsets vector that wasn't big enough. The temporary memory
4415 that is used in this case wasn't being freed if the match failed.
4416
4417 4. Tidied pcretest to ensure it frees memory that it gets.
4418
4419 5. Temporary memory was being obtained in the case where the passed offsets
4420 vector was exactly big enough.
4421
4422 6. Corrected definition of offsetof() from change 5 below.
4423
4424 7. I had screwed up change 6 below and broken the rules for the use of
4425 setjmp(). Now fixed.
4426
4427
4428 Version 1.03 18-Dec-97
4429 ----------------------
4430
4431 1. A erroneous regex with a missing opening parenthesis was correctly
4432 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4433 on some systems.
4434
4435 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4436 it was reported that one broken compiler failed on the former because "code" is
4437 also an independent variable.
4438
4439 3. The erroneous regex a[]b caused an array overrun reference.
4440
4441 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4442 fail on data ending with that character. (It was going on too far, and checking
4443 the next character, typically a binary zero.) This was specific to the
4444 optimized code for single-character negative classes.
4445
4446 5. Added a contributed patch from the TIN world which does the following:
4447
4448 + Add an undef for memmove, in case the the system defines a macro for it.
4449
4450 + Add a definition of offsetof(), in case there isn't one. (I don't know
4451 the reason behind this - offsetof() is part of the ANSI standard - but
4452 it does no harm).
4453
4454 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4455 most of the places where whitespace preceded '#'. I have given up and
4456 allowed the remaining 2 cases to be at the margin.
4457
4458 + Rename some variables in pcre to eliminate shadowing. This seems very
4459 pedantic, but does no harm, of course.
4460
4461 6. Moved the call to setjmp() into its own function, to get rid of warnings
4462 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4463
4464 7. Constructs such as \d{8,} were compiling into the equivalent of
4465 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4466 outcome, but in this particular case used more store than had been allocated,
4467 which caused the bug to be discovered because it threw up an internal error.
4468
4469 8. The debugging code in both pcre and pcretest for outputting the compiled
4470 form of a regex was going wrong in the case of back references followed by
4471 curly-bracketed repeats.
4472
4473
4474 Version 1.02 12-Dec-97
4475 ----------------------
4476
4477 1. Typos in pcre.3 and comments in the source fixed.
4478
4479 2. Applied a contributed patch to get rid of places where it used to remove
4480 'const' from variables, and fixed some signed/unsigned and uninitialized
4481 variable warnings.
4482
4483 3. Added the "runtest" target to Makefile.
4484
4485 4. Set default compiler flag to -O2 rather than just -O.
4486
4487
4488 Version 1.01 19-Nov-97
4489 ----------------------
4490
4491 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4492 like /([ab]*)*/, that is, for classes with more than one character in them.
4493
4494 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4495 as /((?>a*))*/ (a PCRE_EXTRA facility).
4496
4497
4498 Version 1.00 18-Nov-97
4499 ----------------------
4500
4501 1. Added compile-time macros to support systems such as SunOS4 which don't have
4502 memmove() or strerror() but have other things that can be used instead.
4503
4504 2. Arranged that "make clean" removes the executables.
4505
4506
4507 Version 0.99 27-Oct-97
4508 ----------------------
4509
4510 1. Fixed bug in code for optimizing classes with only one character. It was
4511 initializing a 32-byte map regardless, which could cause it to run off the end
4512 of the memory it had got.
4513
4514 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4515
4516
4517 Version 0.98 22-Oct-97
4518 ----------------------
4519
4520 1. Fixed bug in code for handling temporary memory usage when there are more
4521 back references than supplied space in the ovector. This could cause segfaults.
4522
4523
4524 Version 0.97 21-Oct-97
4525 ----------------------
4526
4527 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4528
4529 2. Optimized negated single characters not to use a bit map.
4530
4531 3. Brought error texts together as macro definitions; clarified some of them;
4532 fixed one that was wrong - it said "range out of order" when it meant "invalid
4533 escape sequence".
4534
4535 4. Changed some char * arguments to const char *.
4536
4537 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4538
4539 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4540 pcretest.
4541
4542
4543 Version 0.96 16-Oct-97
4544 ----------------------
4545
4546 1. Added a simple "pgrep" utility to the distribution.
4547
4548 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4549 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4550 where "ddd" means "one or more decimal digits".
4551
4552 3. Fixed serious bug. If a pattern had a back reference, but the call to
4553 pcre_exec() didn't supply a large enough ovector to record the related
4554 identifying subpattern, the match always failed. PCRE now remembers the number
4555 of the largest back reference, and gets some temporary memory in which to save
4556 the offsets during matching if necessary, in order to ensure that
4557 backreferences always work.
4558
4559 4. Increased the compatibility with Perl in a number of ways:
4560
4561 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4562 to request this handling. The option can be set at compile or exec time.
4563
4564 (b) $ matches before a terminating newline by default; an option
4565 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4566 mode). The option can be set at compile or exec time.
4567
4568 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4569 the same as Perl's. If the decimal number it represents is less than 10
4570 or there aren't that many previous left capturing parentheses, an octal
4571 escape is read. Inside a character class, it's always an octal escape,
4572 even if it is a single digit.
4573
4574 (d) An escaped but undefined alphabetic character is taken as a literal,
4575 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4576 escapes.
4577
4578 (e) {0} is now permitted. (The previous item is removed from the compiled
4579 pattern).
4580
4581 5. Changed all the names of code files so that the basic parts are no longer
4582 than 10 characters, and abolished the teeny "globals.c" file.
4583
4584 6. Changed the handling of character classes; they are now done with a 32-byte
4585 bit map always.
4586
4587 7. Added the -d and /D options to pcretest to make it possible to look at the
4588 internals of compilation without having to recompile pcre.
4589
4590
4591 Version 0.95 23-Sep-97
4592 ----------------------
4593
4594 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4595 \x20 at the start of a run of normal characters. These were being treated as
4596 real characters, instead of the source characters being re-checked.
4597
4598
4599 Version 0.94 18-Sep-97
4600 ----------------------
4601
4602 1. The functions are now thread-safe, with the caveat that the global variables
4603 containing pointers to malloc() and free() or alternative functions are the
4604 same for all threads.
4605
4606 2. Get pcre_study() to generate a bitmap of initial characters for non-
4607 anchored patterns when this is possible, and use it if passed to pcre_exec().
4608
4609
4610 Version 0.93 15-Sep-97
4611 ----------------------
4612
4613 1. /(b)|(:+)/ was computing an incorrect first character.
4614
4615 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4616 but not actually doing anything yet.
4617
4618 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4619 as Perl does (e.g. [-az] or [az-]).
4620
4621 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4622 all possible positions.
4623
4624 5. Split up into different modules to avoid including unneeded functions in a
4625 compiled binary. However, compile and exec are still in one module. The "study"
4626 function is split off.
4627
4628 6. The character tables are now in a separate module whose source is generated
4629 by an auxiliary program - but can then be edited by hand if required. There are
4630 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4631 toupper() in the code.
4632
4633 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4634 make them global. Abolish the function for setting them, as the caller can now
4635 set them directly.
4636
4637
4638 Version 0.92 11-Sep-97
4639 ----------------------
4640
4641 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4642 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4643
4644 2. Caseless matching was not working in character classes if the characters in
4645 the pattern were in upper case.
4646
4647 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4648
4649 4. Make PCRE_ANCHORED public and accept as a compile option.
4650
4651 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4652 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4653 pass them.
4654
4655 6. Give an error if bad option bits passed at compile or run time.
4656
4657 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4658 pcretest to cause it to pass that flag.
4659
4660 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4661 options, and the first character, if set.
4662
4663 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4664
4665
4666 Version 0.91 10-Sep-97
4667 ----------------------
4668
4669 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4670 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4671
4672 2. PCRE was looping on encountering an indefinitely repeated back reference to
4673 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4674 Perl does - treats the match as successful.
4675
4676 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5