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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 288 - (show annotations) (download)
Wed Dec 19 12:21:45 2007 UTC (7 years ago) by ph10
File size: 146380 byte(s)
Fix C++ GlobalReplace() double-counting bug.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.5 12-Nov-07
5 ---------------------
6
7 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
8 values in parens when parsing an RE using the C++ wrapper."
9
10 2. Negative specials like \S did not work in character classes in UTF-8 mode.
11 Characters greater than 255 were excluded from the class instead of being
12 included.
13
14 3. The same bug as (2) above applied to negated POSIX classes such as
15 [:^space:].
16
17 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
18 defined or documented. It seems to have been a typo for PCRE_STATIC, so
19 I have changed it.
20
21 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
22 first named subpattern) and a construct such as (?&a) would reference the
23 first named subpattern whose name started with "a" (in other words, the
24 length check was missing). Both these problems are fixed. "Subpattern name
25 expected" is now given for (?&) (a zero-length name), and this patch also
26 makes it give the same error for \k'' (previously it complained that that
27 was a reference to a non-existent subpattern).
28
29 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
30 this is right because (?- can be followed by option settings as well as by
31 digits. I have, however, made the messages clearer.
32
33 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
34 than the number used in the conditional) now cause a compile-time error.
35 This is actually not compatible with Perl, which accepts such patterns, but
36 treats the conditional as always being FALSE (as PCRE used to), but it
37 seems to me that giving a diagnostic is better.
38
39 8. Change "alphameric" to the more common word "alphanumeric" in comments
40 and messages.
41
42 9. Fix two occurrences of "backslash" in comments that should have been
43 "backspace".
44
45 10. Remove two redundant lines of code that can never be obeyed (their function
46 was moved elsewhere).
47
48 11. The program that makes PCRE's Unicode character property table had a bug
49 which caused it to generate incorrect table entries for sequences of
50 characters that have the same character type, but are in different scripts.
51 It amalgamated them into a single range, with the script of the first of
52 them. In other words, some characters were in the wrong script. There were
53 thirteen such cases, affecting characters in the following ranges:
54
55 U+002b0 - U+002c1
56 U+0060c - U+0060d
57 U+0061e - U+00612
58 U+0064b - U+0065e
59 U+0074d - U+0076d
60 U+01800 - U+01805
61 U+01d00 - U+01d77
62 U+01d9b - U+01dbf
63 U+0200b - U+0200f
64 U+030fc - U+030fe
65 U+03260 - U+0327f
66 U+0fb46 - U+0fbb1
67 U+10450 - U+1049d
68
69 12. The -o option (show only the matching part of a line) for pcregrep was not
70 compatible with GNU grep in that, if there was more than one match in a
71 line, it showed only the first of them. It now behaves in the same way as
72 GNU grep.
73
74 13. If the -o and -v options were combined for pcregrep, it printed a blank
75 line for every non-matching line. GNU grep prints nothing, and pcregrep now
76 does the same. The return code can be used to tell if there were any
77 non-matching lines.
78
79 14. The pattern (?=something)(?R) was not being diagnosed as a potentially
80 infinitely looping recursion. The bug was that positive lookaheads were not
81 being skipped when checking for a possible empty match (negative lookaheads
82 and both kinds of lookbehind were skipped).
83
84 15. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
85 inclusion of <windows.h> to before rather than after the definition of
86 INVALID_FILE_ATTRIBUTES (patch from David Byron).
87
88 16. Specifying a possessive quantifier with a specific limit for a Unicode
89 character property caused pcre_compile() to compile bad code, which led at
90 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
91 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
92 caused the error; without that there was no problem.
93
94 17. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
95
96 18. Added --enable-pcretest-libreadline.
97
98 19. In pcrecpp.cc, the variable 'count' was incremented twice in
99 RE::GlobalReplace(). As a result, the number of replacements returned was
100 double what it should be. I have removed one of the increments.
101
102
103 Version 7.4 21-Sep-07
104 ---------------------
105
106 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
107 means that a class such as [\s] counted as "explicit reference to CR or
108 LF". That isn't really right - the whole point of the change was to try to
109 help when there was an actual mention of one of the two characters. So now
110 the change happens only if \r or \n (or a literal CR or LF) character is
111 encountered.
112
113 2. The 32-bit options word was also used for 6 internal flags, but the numbers
114 of both had grown to the point where there were only 3 bits left.
115 Fortunately, there was spare space in the data structure, and so I have
116 moved the internal flags into a new 16-bit field to free up more option
117 bits.
118
119 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
120 but did not set the internal JCHANGED flag - either of these is enough to
121 control the way the "get" function works - but the PCRE_INFO_JCHANGED
122 facility is supposed to tell if (?J) was ever used, so now (?J) at the
123 start sets both bits.
124
125 4. Added options (at build time, compile time, exec time) to change \R from
126 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
127
128 5. doc/pcresyntax.html was missing from the distribution.
129
130 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
131 compatibility, even though it is no longer used.
132
133 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
134 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
135 windows.h file is present (where different names are used). [This was
136 reversed later after testing - see 16 below.]
137
138 8. Changed all #include <config.h> to #include "config.h". There were also
139 some further <pcre.h> cases that I changed to "pcre.h".
140
141 9. When pcregrep was used with the --colour option, it missed the line ending
142 sequence off the lines that it output.
143
144 10. It was pointed out to me that arrays of string pointers cause lots of
145 relocations when a shared library is dynamically loaded. A technique of
146 using a single long string with a table of offsets can drastically reduce
147 these. I have refactored PCRE in four places to do this. The result is
148 dramatic:
149
150 Originally: 290
151 After changing UCP table: 187
152 After changing error message table: 43
153 After changing table of "verbs" 36
154 After changing table of Posix names 22
155
156 Thanks to the folks working on Gregex for glib for this insight.
157
158 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
159 unicode-properties was also set.
160
161 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
162
163 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
164 checked only for CRLF.
165
166 14. Added casts to pcretest.c to avoid compiler warnings.
167
168 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
169
170 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
171 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
172 entirely. This removes changes made in 7 above.
173
174 17. The CMake files have been updated, and there is now more information about
175 building with CMake in the NON-UNIX-USE document.
176
177
178 Version 7.3 28-Aug-07
179 ---------------------
180
181 1. In the rejigging of the build system that eventually resulted in 7.1, the
182 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
183 brackets there is not right, since it causes compilers to look for an
184 installed pcre.h, not the version that is in the source that is being
185 compiled (which of course may be different). I have changed it back to:
186
187 #include "pcre.h"
188
189 I have a vague recollection that the change was concerned with compiling in
190 different directories, but in the new build system, that is taken care of
191 by the VPATH setting the Makefile.
192
193 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
194 when the subject happened to end in the byte 0x85 (e.g. if the last
195 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
196 characters but of course it shouldn't be taken as a newline when it is part
197 of another character. The bug was that, for an unlimited repeat of . in
198 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
199 characters when looking for a newline.
200
201 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
202
203 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
204 in debug output.
205
206 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
207 long printing in the pcrecpp unittest when running under MinGW.
208
209 6. ESC_K was left out of the EBCDIC table.
210
211 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
212 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
213 limit also applies to "virtual nesting" when a pattern is recursive, and in
214 this case 1000 isn't so big. I have been able to remove this limit at the
215 expense of backing off one optimization in certain circumstances. Normally,
216 when pcre_exec() would call its internal match() function recursively and
217 immediately return the result unconditionally, it uses a "tail recursion"
218 feature to save stack. However, when a subpattern that can match an empty
219 string has an unlimited repetition quantifier, it no longer makes this
220 optimization. That gives it a stack frame in which to save the data for
221 checking that an empty string has been matched. Previously this was taken
222 from the 1000-entry workspace that had been reserved. So now there is no
223 explicit limit, but more stack is used.
224
225 8. Applied Daniel's patches to solve problems with the import/export magic
226 syntax that is required for Windows, and which was going wrong for the
227 pcreposix and pcrecpp parts of the library. These were overlooked when this
228 problem was solved for the main library.
229
230 9. There were some crude static tests to avoid integer overflow when computing
231 the size of patterns that contain repeated groups with explicit upper
232 limits. As the maximum quantifier is 65535, the maximum group length was
233 set at 30,000 so that the product of these two numbers did not overflow a
234 32-bit integer. However, it turns out that people want to use groups that
235 are longer than 30,000 bytes (though not repeat them that many times).
236 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
237 made it possible to implement the integer overflow checks in a much more
238 dynamic way, which I have now done. The artificial limitation on group
239 length has been removed - we now have only the limit on the total length of
240 the compiled pattern, which depends on the LINK_SIZE setting.
241
242 10. Fixed a bug in the documentation for get/copy named substring when
243 duplicate names are permitted. If none of the named substrings are set, the
244 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
245 empty string.
246
247 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
248 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
249 because the ] is interpreted as the first data character and the
250 terminating ] is not found. PCRE has been made compatible with Perl in this
251 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
252 cause memory overwriting.
253
254 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
255 string has been matched (to stop an infinite loop). It was not recognizing
256 a conditional subpattern that could match an empty string if that
257 subpattern was within another subpattern. For example, it looped when
258 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
259 condition was not nested. This bug has been fixed.
260
261 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
262 past the start of the subject in the presence of bytes with the top bit
263 set, for example "\x8aBCD".
264
265 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
266 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
267
268 14. Optimized (?!) to (*FAIL).
269
270 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
271 This restricts code points to be within the range 0 to 0x10FFFF, excluding
272 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
273 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
274 does: it's just the validity check that is more restrictive.
275
276 16. Inserted checks for integer overflows during escape sequence (backslash)
277 processing, and also fixed erroneous offset values for syntax errors during
278 backslash processing.
279
280 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
281 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
282
283 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
284 caused an overrun.
285
286 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
287 something other than just ASCII characters) inside a group that had an
288 unlimited repeat caused a loop at compile time (while checking to see
289 whether the group could match an empty string).
290
291 20. Debugging a pattern containing \p or \P could cause a crash. For example,
292 [\P{Any}] did so. (Error in the code for printing property names.)
293
294 21. An orphan \E inside a character class could cause a crash.
295
296 22. A repeated capturing bracket such as (A)? could cause a wild memory
297 reference during compilation.
298
299 23. There are several functions in pcre_compile() that scan along a compiled
300 expression for various reasons (e.g. to see if it's fixed length for look
301 behind). There were bugs in these functions when a repeated \p or \P was
302 present in the pattern. These operators have additional parameters compared
303 with \d, etc, and these were not being taken into account when moving along
304 the compiled data. Specifically:
305
306 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
307 length.
308
309 (b) An item such as \pL+ within a repeated group could cause crashes or
310 loops.
311
312 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
313 "reference to non-existent subpattern" error.
314
315 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
316
317 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
318 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
319
320 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
321
322 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
323 character were causing crashes (broken optimization).
324
325 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
326 \p or \P) caused a compile-time loop.
327
328 28. More problems have arisen in unanchored patterns when CRLF is a valid line
329 break. For example, the unstudied pattern [\r\n]A does not match the string
330 "\r\nA" because change 7.0/46 below moves the current point on by two
331 characters after failing to match at the start. However, the pattern \nA
332 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
333 the same is true. There doesn't seem any very clean way out of this, but
334 what I have chosen to do makes the common cases work: PCRE now takes note
335 of whether there can be an explicit match for \r or \n anywhere in the
336 pattern, and if so, 7.0/46 no longer applies. As part of this change,
337 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
338 pattern has explicit CR or LF references.
339
340 29. Added (*CR) etc for changing newline setting at start of pattern.
341
342
343 Version 7.2 19-Jun-07
344 ---------------------
345
346 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
347 which is apparently normally available under Windows.
348
349 2. Re-jig the pcregrep tests with different newline settings in an attempt
350 to make them independent of the local environment's newline setting.
351
352 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
353
354 4. Some of the "internals" tests were previously cut out when the link size
355 was not 2, because the output contained actual offsets. The recent new
356 "Z" feature of pcretest means that these can be cut out, making the tests
357 usable with all link sizes.
358
359 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
360 stack recursion. This gives a massive performance boost under BSD, but just
361 a small improvement under Linux. However, it saves one field in the frame
362 in all cases.
363
364 6. Added more features from the forthcoming Perl 5.10:
365
366 (a) (?-n) (where n is a string of digits) is a relative subroutine or
367 recursion call. It refers to the nth most recently opened parentheses.
368
369 (b) (?+n) is also a relative subroutine call; it refers to the nth next
370 to be opened parentheses.
371
372 (c) Conditions that refer to capturing parentheses can be specified
373 relatively, for example, (?(-2)... or (?(+3)...
374
375 (d) \K resets the start of the current match so that everything before
376 is not part of it.
377
378 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
379
380 (f) \g{name} is another synonym - part of Perl 5.10's unification of
381 reference syntax.
382
383 (g) (?| introduces a group in which the numbering of parentheses in each
384 alternative starts with the same number.
385
386 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
387
388 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
389 PCRE_INFO_JCHANGED.
390
391 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
392 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
393 for detecting groups that can match an empty string.
394
395 9. A pattern with a very large number of alternatives (more than several
396 hundred) was running out of internal workspace during the pre-compile
397 phase, where pcre_compile() figures out how much memory will be needed. A
398 bit of new cunning has reduced the workspace needed for groups with
399 alternatives. The 1000-alternative test pattern now uses 12 bytes of
400 workspace instead of running out of the 4096 that are available.
401
402 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
403
404 11. Applied patch from Google to remove an optimization that didn't quite work.
405 The report of the bug said:
406
407 pcrecpp::RE("a*").FullMatch("aaa") matches, while
408 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
409 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
410
411 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
412 it matched the wrong number of bytes.
413
414
415 Version 7.1 24-Apr-07
416 ---------------------
417
418 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
419 that is more "standard", making use of automake and other Autotools. There
420 is some re-arrangement of the files and adjustment of comments consequent
421 on this.
422
423 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
424 for recursive directory scanning broke on some systems because the files
425 are not scanned in any specific order and on different systems the order
426 was different. A call to "sort" has been inserted into RunGrepTest for the
427 approprate test as a short-term fix. In the longer term there may be an
428 alternative.
429
430 3. I had an email from Eric Raymond about problems translating some of PCRE's
431 man pages to HTML (despite the fact that I distribute HTML pages, some
432 people do their own conversions for various reasons). The problems
433 concerned the use of low-level troff macros .br and .in. I have therefore
434 removed all such uses from the man pages (some were redundant, some could
435 be replaced by .nf/.fi pairs). The 132html script that I use to generate
436 HTML has been updated to handle .nf/.fi and to complain if it encounters
437 .br or .in.
438
439 4. Updated comments in configure.ac that get placed in config.h.in and also
440 arranged for config.h to be included in the distribution, with the name
441 config.h.generic, for the benefit of those who have to compile without
442 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
443
444 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
445 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
446 makevp.bat for the current PCRE, using the additional files
447 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
448
449 6. A Windows user reported a minor discrepancy with test 2, which turned out
450 to be caused by a trailing space on an input line that had got lost in his
451 copy. The trailing space was an accident, so I've just removed it.
452
453 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
454 that is needed.
455
456 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
457 as "const" (a) because they are and (b) because it helps the PHP
458 maintainers who have recently made a script to detect big data structures
459 in the php code that should be moved to the .rodata section. I remembered
460 to update Builducptable as well, so it won't revert if ucptable.h is ever
461 re-created.
462
463 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
464 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
465 order to be able to cut out the UTF-8 tables in the latter when UTF-8
466 support is not required. This saves 1.5-2K of code, which is important in
467 some applications.
468
469 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
470 so as not to refer to the tables, even though these functions will never be
471 called when UTF-8 support is disabled. Otherwise there are problems with a
472 shared library.
473
474 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
475
476 (a) It was defining its arguments as char * instead of void *.
477
478 (b) It was assuming that all moves were upwards in memory; this was true
479 a long time ago when I wrote it, but is no longer the case.
480
481 The emulated memove() is provided for those environments that have neither
482 memmove() nor bcopy(). I didn't think anyone used it these days, but that
483 is clearly not the case, as these two bugs were recently reported.
484
485 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
486 and Detrail to create the HTML documentation, the .txt form of the man
487 pages, and it removes trailing spaces from listed files. It also creates
488 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
489 case, it wraps all the #defines with #ifndefs. This script should be run
490 before "make dist".
491
492 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
493 with Unicode property support.
494
495 (a) For a maximizing quantifier, if the two different cases of the
496 character were of different lengths in their UTF-8 codings (there are
497 some cases like this - I found 11), and the matching function had to
498 back up over a mixture of the two cases, it incorrectly assumed they
499 were both the same length.
500
501 (b) When PCRE was configured to use the heap rather than the stack for
502 recursion during matching, it was not correctly preserving the data for
503 the other case of a UTF-8 character when checking ahead for a match
504 while processing a minimizing repeat. If the check also involved
505 matching a wide character, but failed, corruption could cause an
506 erroneous result when trying to check for a repeat of the original
507 character.
508
509 13. Some tidying changes to the testing mechanism:
510
511 (a) The RunTest script now detects the internal link size and whether there
512 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
513 values substituted by "configure". (The RunGrepTest script already did
514 this for UTF-8.) The configure.ac script no longer substitutes the
515 relevant variables.
516
517 (b) The debugging options /B and /D in pcretest show the compiled bytecode
518 with length and offset values. This means that the output is different
519 for different internal link sizes. Test 2 is skipped for link sizes
520 other than 2 because of this, bypassing the problem. Unfortunately,
521 there was also a test in test 3 (the locale tests) that used /B and
522 failed for link sizes other than 2. Rather than cut the whole test out,
523 I have added a new /Z option to pcretest that replaces the length and
524 offset values with spaces. This is now used to make test 3 independent
525 of link size. (Test 2 will be tidied up later.)
526
527 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
528 segmentation fault instead of returning the appropriate error message.
529
530 15. In multiline mode when the newline sequence was set to "any", the pattern
531 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
532 This doesn't seem right; it now treats the CRLF combination as the line
533 ending, and so does not match in that case. It's only a pattern such as ^$
534 that would hit this one: something like ^ABC$ would have failed after \r
535 and then tried again after \r\n.
536
537 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
538 in an attempt to make files that differ only in their line terminators
539 compare equal. This works on Linux.
540
541 17. Under certain error circumstances pcregrep might try to free random memory
542 as it exited. This is now fixed, thanks to valgrind.
543
544 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
545 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
546 was because its rules for how to advance for /g after matching an empty
547 string at the end of a line did not allow for this case. They now check for
548 it specially.
549
550 20. pcretest is supposed to handle patterns and data of any length, by
551 extending its buffers when necessary. It was getting this wrong when the
552 buffer for a data line had to be extended.
553
554 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
555 CRLF as a newline sequence.
556
557 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
558 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
559 I have nevertheless tidied it up.
560
561 23. Added some casts to kill warnings from HP-UX ia64 compiler.
562
563 24. Added a man page for pcre-config.
564
565
566 Version 7.0 19-Dec-06
567 ---------------------
568
569 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
570 moving to gcc 4.1.1.
571
572 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
573 sys/time.h, which is documented as needed for this function. It doesn't
574 seem to matter on Linux, but it showed up on some releases of OS X.
575
576 3. It seems that there are systems where bytes whose values are greater than
577 127 match isprint() in the "C" locale. The "C" locale should be the
578 default when a C program starts up. In most systems, only ASCII printing
579 characters match isprint(). This difference caused the output from pcretest
580 to vary, making some of the tests fail. I have changed pcretest so that:
581
582 (a) When it is outputting text in the compiled version of a pattern, bytes
583 other than 32-126 are always shown as hex escapes.
584
585 (b) When it is outputting text that is a matched part of a subject string,
586 it does the same, unless a different locale has been set for the match
587 (using the /L modifier). In this case, it uses isprint() to decide.
588
589 4. Fixed a major bug that caused incorrect computation of the amount of memory
590 required for a compiled pattern when options that changed within the
591 pattern affected the logic of the preliminary scan that determines the
592 length. The relevant options are -x, and -i in UTF-8 mode. The result was
593 that the computed length was too small. The symptoms of this bug were
594 either the PCRE error "internal error: code overflow" from pcre_compile(),
595 or a glibc crash with a message such as "pcretest: free(): invalid next
596 size (fast)". Examples of patterns that provoked this bug (shown in
597 pcretest format) are:
598
599 /(?-x: )/x
600 /(?x)(?-x: \s*#\s*)/
601 /((?i)[\x{c0}])/8
602 /(?i:[\x{c0}])/8
603
604 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
605 is now done differently.
606
607 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
608 wrapper classes; (b) implement a new function in the C++ scanner that is
609 more efficient than the old way of doing things because it avoids levels of
610 recursion in the regex matching; (c) add a paragraph to the documentation
611 for the FullMatch() function.
612
613 6. The escape sequence \n was being treated as whatever was defined as
614 "newline". Not only was this contrary to the documentation, which states
615 that \n is character 10 (hex 0A), but it also went horribly wrong when
616 "newline" was defined as CRLF. This has been fixed.
617
618 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
619 was being set to -1 for the "end of line" case (supposedly a value that no
620 character can have). Though this value is never used (the check for end of
621 line is "zero bytes in current character"), it caused compiler complaints.
622 I've changed it to 0xffffffff.
623
624 8. In pcre_version.c, the version string was being built by a sequence of
625 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
626 string (as it is for production releases) called a macro with an empty
627 argument. The C standard says the result of this is undefined. The gcc
628 compiler treats it as an empty string (which was what was wanted) but it is
629 reported that Visual C gives an error. The source has been hacked around to
630 avoid this problem.
631
632 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
633 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
634 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
635 of them did).
636
637 10. Originally, pcretest opened its input and output without "b"; then I was
638 told that "b" was needed in some environments, so it was added for release
639 5.0 to both the input and output. (It makes no difference on Unix-like
640 systems.) Later I was told that it is wrong for the input on Windows. I've
641 now abstracted the modes into two macros, to make it easier to fiddle with
642 them, and removed "b" from the input mode under Windows.
643
644 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
645
646 12. Added -help and --help to pcretest as an official way of being reminded
647 of the options.
648
649 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
650 and pcrecpp.cc because they annoy compilers at high warning levels.
651
652 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
653
654 15. Fixed an occurrence of == in configure.ac that should have been = (shell
655 scripts are not C programs :-) and which was not noticed because it works
656 on Linux.
657
658 16. pcretest is supposed to handle any length of pattern and data line (as one
659 line or as a continued sequence of lines) by extending its input buffer if
660 necessary. This feature was broken for very long pattern lines, leading to
661 a string of junk being passed to pcre_compile() if the pattern was longer
662 than about 50K.
663
664 17. I have done a major re-factoring of the way pcre_compile() computes the
665 amount of memory needed for a compiled pattern. Previously, there was code
666 that made a preliminary scan of the pattern in order to do this. That was
667 OK when PCRE was new, but as the facilities have expanded, it has become
668 harder and harder to keep it in step with the real compile phase, and there
669 have been a number of bugs (see for example, 4 above). I have now found a
670 cunning way of running the real compile function in a "fake" mode that
671 enables it to compute how much memory it would need, while actually only
672 ever using a few hundred bytes of working memory and without too many
673 tests of the mode. This should make future maintenance and development
674 easier. A side effect of this work is that the limit of 200 on the nesting
675 depth of parentheses has been removed (though this was never a serious
676 limitation, I suspect). However, there is a downside: pcre_compile() now
677 runs more slowly than before (30% or more, depending on the pattern). I
678 hope this isn't a big issue. There is no effect on runtime performance.
679
680 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
681 newline (only possible for the last line of a file) and it was a
682 pattern that set a locale (followed by /Lsomething), pcretest crashed.
683
684 19. Added additional timing features to pcretest. (1) The -tm option now times
685 matching only, not compiling. (2) Both -t and -tm can be followed, as a
686 separate command line item, by a number that specifies the number of
687 repeats to use when timing. The default is 50000; this gives better
688 precision, but takes uncomfortably long for very large patterns.
689
690 20. Extended pcre_study() to be more clever in cases where a branch of a
691 subpattern has no definite first character. For example, (a*|b*)[cd] would
692 previously give no result from pcre_study(). Now it recognizes that the
693 first character must be a, b, c, or d.
694
695 21. There was an incorrect error "recursive call could loop indefinitely" if
696 a subpattern (or the entire pattern) that was being tested for matching an
697 empty string contained only one non-empty item after a nested subpattern.
698 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
699 incorrectly, because the \d was being skipped in the check.
700
701 22. The pcretest program now has a new pattern option /B and a command line
702 option -b, which is equivalent to adding /B to every pattern. This causes
703 it to show the compiled bytecode, without the additional information that
704 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
705 is the same as /B/I).
706
707 23. A new optimization is now able automatically to treat some sequences such
708 as a*b as a*+b. More specifically, if something simple (such as a character
709 or a simple class like \d) has an unlimited quantifier, and is followed by
710 something that cannot possibly match the quantified thing, the quantifier
711 is automatically "possessified".
712
713 24. A recursive reference to a subpattern whose number was greater than 39
714 went wrong under certain circumstances in UTF-8 mode. This bug could also
715 have affected the operation of pcre_study().
716
717 25. Realized that a little bit of performance could be had by replacing
718 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
719
720 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
721
722 27. Possessive quantifiers such as a++ were previously implemented by turning
723 them into atomic groups such as ($>a+). Now they have their own opcodes,
724 which improves performance. This includes the automatically created ones
725 from 23 above.
726
727 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
728 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
729 the first matched character to be a colon. This applied both to named and
730 numbered groups.
731
732 29. The ucpinternal.h header file was missing its idempotency #ifdef.
733
734 30. I was sent a "project" file called libpcre.a.dev which I understand makes
735 building PCRE on Windows easier, so I have included it in the distribution.
736
737 31. There is now a check in pcretest against a ridiculously large number being
738 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
739 loop, the loop is abandoned.
740
741 32. Forward references to subpatterns in conditions such as (?(2)...) where
742 subpattern 2 is defined later cause pcre_compile() to search forwards in
743 the pattern for the relevant set of parentheses. This search went wrong
744 when there were unescaped parentheses in a character class, parentheses
745 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
746
747 33. "Subroutine" calls and backreferences were previously restricted to
748 referencing subpatterns earlier in the regex. This restriction has now
749 been removed.
750
751 34. Added a number of extra features that are going to be in Perl 5.10. On the
752 whole, these are just syntactic alternatives for features that PCRE had
753 previously implemented using the Python syntax or my own invention. The
754 other formats are all retained for compatibility.
755
756 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
757 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
758 also .NET compatible.
759
760 (b) A recursion or subroutine call to a named group can now be defined as
761 (?&name) as well as (?P>name).
762
763 (c) A backreference to a named group can now be defined as \k<name> or
764 \k'name' as well as (?P=name). The new forms, as well as being in Perl
765 5.10, are also .NET compatible.
766
767 (d) A conditional reference to a named group can now use the syntax
768 (?(<name>) or (?('name') as well as (?(name).
769
770 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
771 groups (named and numbered) that are never evaluated inline, but can be
772 called as "subroutines" from elsewhere. In effect, the DEFINE condition
773 is always false. There may be only one alternative in such a group.
774
775 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
776 as the simple (?(R). The condition is true only if the most recent
777 recursion is that of the given number or name. It does not search out
778 through the entire recursion stack.
779
780 (g) The escape \gN or \g{N} has been added, where N is a positive or
781 negative number, specifying an absolute or relative reference.
782
783 35. Tidied to get rid of some further signed/unsigned compiler warnings and
784 some "unreachable code" warnings.
785
786 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
787 things, this adds five new scripts.
788
789 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
790 There were also incompatibilities regarding the handling of \Q..\E inside
791 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
792 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
793
794 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
795 matches an empty string, and forcibly breaks the loop. There were bugs in
796 this code in non-simple cases. For a pattern such as ^(a()*)* matched
797 against aaaa the result was just "a" rather than "aaaa", for example. Two
798 separate and independent bugs (that affected different cases) have been
799 fixed.
800
801 39. Refactored the code to abolish the use of different opcodes for small
802 capturing bracket numbers. This is a tidy that I avoided doing when I
803 removed the limit on the number of capturing brackets for 3.5 back in 2001.
804 The new approach is not only tidier, it makes it possible to reduce the
805 memory needed to fix the previous bug (38).
806
807 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
808 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
809 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
810 mode.
811
812 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
813 report.
814
815 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
816 copy construction and assignment in the C++ wrapper.
817
818 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
819 couple of bugs that could have given wrong results in the "--newline=crlf"
820 case.
821
822 44. Added a number of casts and did some reorganization of signed/unsigned int
823 variables following suggestions from Dair Grant. Also renamed the variable
824 "this" as "item" because it is a C++ keyword.
825
826 45. Arranged for dftables to add
827
828 #include "pcre_internal.h"
829
830 to pcre_chartables.c because without it, gcc 4.x may remove the array
831 definition from the final binary if PCRE is built into a static library and
832 dead code stripping is activated.
833
834 46. For an unanchored pattern, if a match attempt fails at the start of a
835 newline sequence, and the newline setting is CRLF or ANY, and the next two
836 characters are CRLF, advance by two characters instead of one.
837
838
839 Version 6.7 04-Jul-06
840 ---------------------
841
842 1. In order to handle tests when input lines are enormously long, pcretest has
843 been re-factored so that it automatically extends its buffers when
844 necessary. The code is crude, but this _is_ just a test program. The
845 default size has been increased from 32K to 50K.
846
847 2. The code in pcre_study() was using the value of the re argument before
848 testing it for NULL. (Of course, in any sensible call of the function, it
849 won't be NULL.)
850
851 3. The memmove() emulation function in pcre_internal.h, which is used on
852 systems that lack both memmove() and bcopy() - that is, hardly ever -
853 was missing a "static" storage class specifier.
854
855 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
856 containing an extended class (one that cannot be represented by a bitmap
857 because it contains high-valued characters or Unicode property items, e.g.
858 [\pZ]). Almost always one would set UTF-8 mode when processing such a
859 pattern, but PCRE should not loop if you do not (it no longer does).
860 [Detail: two cases were found: (a) a repeated subpattern containing an
861 extended class; (b) a recursive reference to a subpattern that followed a
862 previous extended class. It wasn't skipping over the extended class
863 correctly when UTF-8 mode was not set.]
864
865 5. A negated single-character class was not being recognized as fixed-length
866 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
867 compile error "lookbehind assertion is not fixed length".
868
869 6. The RunPerlTest auxiliary script was showing an unexpected difference
870 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
871 write a Perl script that can interpret lines of an input file either as
872 byte characters or as UTF-8, which is what "perltest" was being required to
873 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
874 can't do is switch easily at run time between having the "use utf8;" pragma
875 or not. In the end, I fudged it by using the RunPerlTest script to insert
876 "use utf8;" explicitly for the UTF-8 tests.
877
878 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
879 the end of the subject string, contrary to the documentation and to what
880 Perl does. This was true of both matching functions. Now it matches only at
881 the start of the subject and immediately after *internal* newlines.
882
883 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
884 a pointer to an int instead of a pointer to an unsigned long int. This
885 caused problems on 64-bit systems.
886
887 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
888 instance of the 'standard' template library not being so standard".
889
890 10. There was no check on the number of named subpatterns nor the maximum
891 length of a subpattern name. The product of these values is used to compute
892 the size of the memory block for a compiled pattern. By supplying a very
893 long subpattern name and a large number of named subpatterns, the size
894 computation could be caused to overflow. This is now prevented by limiting
895 the length of names to 32 characters, and the number of named subpatterns
896 to 10,000.
897
898 11. Subpatterns that are repeated with specific counts have to be replicated in
899 the compiled pattern. The size of memory for this was computed from the
900 length of the subpattern and the repeat count. The latter is limited to
901 65535, but there was no limit on the former, meaning that integer overflow
902 could in principle occur. The compiled length of a repeated subpattern is
903 now limited to 30,000 bytes in order to prevent this.
904
905 12. Added the optional facility to have named substrings with the same name.
906
907 13. Added the ability to use a named substring as a condition, using the
908 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
909 are numbers (not recommended). Forward references are permitted.
910
911 14. Added forward references in named backreferences (if you see what I mean).
912
913 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
914 pattern could run off the end of the subject. For example, the pattern
915 "(?s)(.{1,5})"8 did this with the subject "ab".
916
917 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
918 PCRE_CASELESS was set when matching characters that were quantified with ?
919 or *.
920
921 17. A character class other than a single negated character that had a minimum
922 but no maximum quantifier - for example [ab]{6,} - was not handled
923 correctly by pce_dfa_exec(). It would match only one character.
924
925 18. A valid (though odd) pattern that looked like a POSIX character
926 class but used an invalid character after [ (for example [[,abc,]]) caused
927 pcre_compile() to give the error "Failed: internal error: code overflow" or
928 in some cases to crash with a glibc free() error. This could even happen if
929 the pattern terminated after [[ but there just happened to be a sequence of
930 letters, a binary zero, and a closing ] in the memory that followed.
931
932 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
933 over the years. Originally (before any Unicode support), just the bottom 8
934 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
935 output from "man perlunicode" includes this:
936
937 The regular expression compiler produces polymorphic opcodes. That
938 is, the pattern adapts to the data and automatically switches to
939 the Unicode character scheme when presented with Unicode data--or
940 instead uses a traditional byte scheme when presented with byte
941 data.
942
943 Sadly, a wide octal escape does not cause a switch, and in a string with
944 no other multibyte characters, these octal escapes are treated as before.
945 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
946 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
947 Unicode string.
948
949 I have not perpetrated such confusion in PCRE. Up till now, it took just
950 the bottom 8 bits, as in old Perl. I have now made octal escapes with
951 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
952 translate to the appropriate multibyte character.
953
954 29. Applied some refactoring to reduce the number of warnings from Microsoft
955 and Borland compilers. This has included removing the fudge introduced
956 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
957 a warning about an unused variable.
958
959 21. PCRE has not included VT (character 0x0b) in the set of whitespace
960 characters since release 4.0, because Perl (from release 5.004) does not.
961 [Or at least, is documented not to: some releases seem to be in conflict
962 with the documentation.] However, when a pattern was studied with
963 pcre_study() and all its branches started with \s, PCRE still included VT
964 as a possible starting character. Of course, this did no harm; it just
965 caused an unnecessary match attempt.
966
967 22. Removed a now-redundant internal flag bit that recorded the fact that case
968 dependency changed within the pattern. This was once needed for "required
969 byte" processing, but is no longer used. This recovers a now-scarce options
970 bit. Also moved the least significant internal flag bit to the most-
971 significant bit of the word, which was not previously used (hangover from
972 the days when it was an int rather than a uint) to free up another bit for
973 the future.
974
975 23. Added support for CRLF line endings as well as CR and LF. As well as the
976 default being selectable at build time, it can now be changed at runtime
977 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
978 specify that it is scanning data with non-default line endings.
979
980 24. Changed the definition of CXXLINK to make it agree with the definition of
981 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
982
983 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
984 recursions. This makes a big different to stack usage for some patterns.
985
986 26. If a subpattern containing a named recursion or subroutine reference such
987 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
988 the space required for the compiled pattern went wrong and gave too small a
989 value. Depending on the environment, this could lead to "Failed: internal
990 error: code overflow at offset 49" or "glibc detected double free or
991 corruption" errors.
992
993 27. Applied patches from Google (a) to support the new newline modes and (b) to
994 advance over multibyte UTF-8 characters in GlobalReplace.
995
996 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
997 difference for some implementation of PCRE in some Windows version.
998
999 29. Added some extra testing facilities to pcretest:
1000
1001 \q<number> in a data line sets the "match limit" value
1002 \Q<number> in a data line sets the "match recursion limt" value
1003 -S <number> sets the stack size, where <number> is in megabytes
1004
1005 The -S option isn't available for Windows.
1006
1007
1008 Version 6.6 06-Feb-06
1009 ---------------------
1010
1011 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1012 in pcreposix.h. I have copied the definition from pcre.h.
1013
1014 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1015 because pcre.h is no longer a built file.
1016
1017 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1018 not normally included in the compiled code.
1019
1020
1021 Version 6.5 01-Feb-06
1022 ---------------------
1023
1024 1. When using the partial match feature with pcre_dfa_exec(), it was not
1025 anchoring the second and subsequent partial matches at the new starting
1026 point. This could lead to incorrect results. For example, with the pattern
1027 /1234/, partially matching against "123" and then "a4" gave a match.
1028
1029 2. Changes to pcregrep:
1030
1031 (a) All non-match returns from pcre_exec() were being treated as failures
1032 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1033 error message is output. Some extra information is given for the
1034 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1035 probably the only errors that are likely to be caused by users (by
1036 specifying a regex that has nested indefinite repeats, for instance).
1037 If there are more than 20 of these errors, pcregrep is abandoned.
1038
1039 (b) A binary zero was treated as data while matching, but terminated the
1040 output line if it was written out. This has been fixed: binary zeroes
1041 are now no different to any other data bytes.
1042
1043 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1044 used to set a locale for matching. The --locale=xxxx long option has
1045 been added (no short equivalent) to specify a locale explicitly on the
1046 pcregrep command, overriding the environment variables.
1047
1048 (d) When -B was used with -n, some line numbers in the output were one less
1049 than they should have been.
1050
1051 (e) Added the -o (--only-matching) option.
1052
1053 (f) If -A or -C was used with -c (count only), some lines of context were
1054 accidentally printed for the final match.
1055
1056 (g) Added the -H (--with-filename) option.
1057
1058 (h) The combination of options -rh failed to suppress file names for files
1059 that were found from directory arguments.
1060
1061 (i) Added the -D (--devices) and -d (--directories) options.
1062
1063 (j) Added the -F (--fixed-strings) option.
1064
1065 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1066
1067 (l) Added the --colo(u)r option.
1068
1069 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1070 is not present by default.
1071
1072 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1073 items such as (?R) or (?1), when the recursion could match a number of
1074 alternatives. If it matched one of the alternatives, but subsequently,
1075 outside the recursion, there was a failure, the code tried to back up into
1076 the recursion. However, because of the way PCRE is implemented, this is not
1077 possible, and the result was an incorrect result from the match.
1078
1079 In order to prevent this happening, the specification of recursion has
1080 been changed so that all such subpatterns are automatically treated as
1081 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1082
1083 4. I had overlooked the fact that, in some locales, there are characters for
1084 which isalpha() is true but neither isupper() nor islower() are true. In
1085 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1086 and ordfeminine) are like this. This affected the treatment of \w and \W
1087 when they appeared in character classes, but not when they appeared outside
1088 a character class. The bit map for "word" characters is now created
1089 separately from the results of isalnum() instead of just taking it from the
1090 upper, lower, and digit maps. (Plus the underscore character, of course.)
1091
1092 5. The above bug also affected the handling of POSIX character classes such as
1093 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1094 permanent tables. Instead, the bit maps for such a class were previously
1095 created as the appropriate unions of the upper, lower, and digit bitmaps.
1096 Now they are created by subtraction from the [[:word:]] class, which has
1097 its own bitmap.
1098
1099 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1100 It is created by subtracting the vertical space characters (\x09, \x0a,
1101 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1102 subtraction was done in the overall bitmap for a character class, meaning
1103 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1104 be recognized. This bug has been fixed.
1105
1106 7. Patches from the folks at Google:
1107
1108 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1109 real life, but is still worth protecting against".
1110
1111 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1112 regular expressions".
1113
1114 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1115 have it.
1116
1117 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1118 "configure" and the latter not, in order to fix a problem somebody had
1119 with compiling the Arg class on HP-UX.
1120
1121 (e) Improve the error-handling of the C++ wrapper a little bit.
1122
1123 (f) New tests for checking recursion limiting.
1124
1125 8. The pcre_memmove() function, which is used only if the environment does not
1126 have a standard memmove() function (and is therefore rarely compiled),
1127 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1128 returning a result (though PCRE never actually uses the result).
1129
1130 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1131 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1132 returned instead of calling malloc() with an overflowing number that would
1133 most likely cause subsequent chaos.
1134
1135 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1136
1137 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1138 with this option is matched, the nmatch and pmatch options of regexec() are
1139 ignored.
1140
1141 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1142 provided in case anyone wants to the the POSIX interface with UTF-8
1143 strings.
1144
1145 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1146 C++ linking (needed for some HP-UX environments).
1147
1148 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1149 (unused parameter) and in the pcre_printint() function (omitted "default"
1150 switch label when the default is to do nothing).
1151
1152 15. Added some code to make it possible, when PCRE is compiled as a C++
1153 library, to replace subject pointers for pcre_exec() with a smart pointer
1154 class, thus making it possible to process discontinuous strings.
1155
1156 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1157 much the same function. They were added by different people who were trying
1158 to make PCRE easy to compile on non-Unix systems. It has been suggested
1159 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1160 for compiling on Windows systems. I have therefore replaced it with
1161 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1162 defaults to "extern" for C or "extern C" for C++, which works fine on
1163 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1164 SCOPE with something explicit in config.h. In addition:
1165
1166 (a) pcreposix.h still had just "extern" instead of either of these macros;
1167 I have replaced it with PCRE_DATA_SCOPE.
1168
1169 (b) Functions such as _pcre_xclass(), which are internal to the library,
1170 but external in the C sense, all had PCRE_EXPORT in their definitions.
1171 This is apparently wrong for the Windows case, so I have removed it.
1172 (It makes no difference on Unix-like systems.)
1173
1174 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1175 of recursive calls to match(). This is different to MATCH_LIMIT because
1176 that limits the total number of calls to match(), not all of which increase
1177 the depth of recursion. Limiting the recursion depth limits the amount of
1178 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1179 when PCRE is compiled, and changed at run time. A patch from Google adds
1180 this functionality to the C++ interface.
1181
1182 18. Changes to the handling of Unicode character properties:
1183
1184 (a) Updated the table to Unicode 4.1.0.
1185
1186 (b) Recognize characters that are not in the table as "Cn" (undefined).
1187
1188 (c) I revised the way the table is implemented to a much improved format
1189 which includes recognition of ranges. It now supports the ranges that
1190 are defined in UnicodeData.txt, and it also amalgamates other
1191 characters into ranges. This has reduced the number of entries in the
1192 table from around 16,000 to around 3,000, thus reducing its size
1193 considerably. I realized I did not need to use a tree structure after
1194 all - a binary chop search is just as efficient. Having reduced the
1195 number of entries, I extended their size from 6 bytes to 8 bytes to
1196 allow for more data.
1197
1198 (d) Added support for Unicode script names via properties such as \p{Han}.
1199
1200 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1201 matching that character.
1202
1203 20. When matching a repeated Unicode property with a minimum greater than zero,
1204 (for example \pL{2,}), PCRE could look past the end of the subject if it
1205 reached it while seeking the minimum number of characters. This could
1206 happen only if some of the characters were more than one byte long, because
1207 there is a check for at least the minimum number of bytes.
1208
1209 21. Refactored the implementation of \p and \P so as to be more general, to
1210 allow for more different types of property in future. This has changed the
1211 compiled form incompatibly. Anybody with saved compiled patterns that use
1212 \p or \P will have to recompile them.
1213
1214 22. Added "Any" and "L&" to the supported property types.
1215
1216 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1217 but give a compile time error if the value is greater than 0xff.
1218
1219 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1220 accidentally not being installed or uninstalled.
1221
1222 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1223 made were to insert the current release number. This seemed silly, because
1224 it made things harder for people building PCRE on systems that don't run
1225 "configure". I have turned pcre.h into a distributed file, no longer built
1226 by "configure", with the version identification directly included. There is
1227 no longer a pcre.h.in file.
1228
1229 However, this change necessitated a change to the pcre-config script as
1230 well. It is built from pcre-config.in, and one of the substitutions was the
1231 release number. I have updated configure.ac so that ./configure now finds
1232 the release number by grepping pcre.h.
1233
1234 26. Added the ability to run the tests under valgrind.
1235
1236
1237 Version 6.4 05-Sep-05
1238 ---------------------
1239
1240 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1241 "--" to be printed when multiple files were scanned, even when none of the
1242 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1243 consider it to be a bug, and have restored the previous behaviour.
1244
1245 2. A couple of code tidies to get rid of compiler warnings.
1246
1247 3. The pcretest program used to cheat by referring to symbols in the library
1248 whose names begin with _pcre_. These are internal symbols that are not
1249 really supposed to be visible externally, and in some environments it is
1250 possible to suppress them. The cheating is now confined to including
1251 certain files from the library's source, which is a bit cleaner.
1252
1253 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1254 file's purpose clearer.
1255
1256 5. Reorganized pcre_ucp_findchar().
1257
1258
1259 Version 6.3 15-Aug-05
1260 ---------------------
1261
1262 1. The file libpcre.pc.in did not have general read permission in the tarball.
1263
1264 2. There were some problems when building without C++ support:
1265
1266 (a) If C++ support was not built, "make install" and "make test" still
1267 tried to test it.
1268
1269 (b) There were problems when the value of CXX was explicitly set. Some
1270 changes have been made to try to fix these, and ...
1271
1272 (c) --disable-cpp can now be used to explicitly disable C++ support.
1273
1274 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1275 backslash in a target when C++ was disabled. This confuses some
1276 versions of "make", apparently. Using an intermediate variable solves
1277 this. (Same for CPP_LOBJ.)
1278
1279 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1280 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1281 necessary on certain architectures.
1282
1283 4. Added a setting of -export-symbols-regex to the link command to remove
1284 those symbols that are exported in the C sense, but actually are local
1285 within the library, and not documented. Their names all begin with
1286 "_pcre_". This is not a perfect job, because (a) we have to except some
1287 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1288 available (and never for static libraries). I have made a note to try to
1289 find a way round (a) in the future.
1290
1291
1292 Version 6.2 01-Aug-05
1293 ---------------------
1294
1295 1. There was no test for integer overflow of quantifier values. A construction
1296 such as {1111111111111111} would give undefined results. What is worse, if
1297 a minimum quantifier for a parenthesized subpattern overflowed and became
1298 negative, the calculation of the memory size went wrong. This could have
1299 led to memory overwriting.
1300
1301 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1302
1303 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1304 operating environments where this matters.
1305
1306 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1307 PCRE options from within the C++ wrapper.
1308
1309 5. Named capturing subpatterns were not being correctly counted when a pattern
1310 was compiled. This caused two problems: (a) If there were more than 100
1311 such subpatterns, the calculation of the memory needed for the whole
1312 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1313 back references of the form \12, where the number was greater than 9, were
1314 not recognized as back references, even though there were sufficient
1315 previous subpatterns.
1316
1317 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1318 versions of gcc, e.g. 2.95.4.
1319
1320
1321 Version 6.1 21-Jun-05
1322 ---------------------
1323
1324 1. There was one reference to the variable "posix" in pcretest.c that was not
1325 surrounded by "#if !defined NOPOSIX".
1326
1327 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1328 the cross-check on the old pcre_info() function, for the benefit of the
1329 cut-down version of PCRE that is currently imported into Exim.
1330
1331 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1332 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1333 patterns that start (?i) but I don't think that matters. The use of (?i) is
1334 just an example; this all applies to the other options as well.
1335
1336 4. Since libtool seems to echo the compile commands it is issuing, the output
1337 from "make" can be reduced a bit by putting "@" in front of each libtool
1338 compile command.
1339
1340 5. Patch from the folks at Google for configure.in to be a bit more thorough
1341 in checking for a suitable C++ installation before trying to compile the
1342 C++ stuff. This should fix a reported problem when a compiler was present,
1343 but no suitable headers.
1344
1345 6. The man pages all had just "PCRE" as their title. I have changed them to
1346 be the relevant file name. I have also arranged that these names are
1347 retained in the file doc/pcre.txt, which is a concatenation in text format
1348 of all the man pages except the little individual ones for each function.
1349
1350 7. The NON-UNIX-USE file had not been updated for the different set of source
1351 files that come with release 6. I also added a few comments about the C++
1352 wrapper.
1353
1354
1355 Version 6.0 07-Jun-05
1356 ---------------------
1357
1358 1. Some minor internal re-organization to help with my DFA experiments.
1359
1360 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1361 didn't matter for the library itself when fully configured, but did matter
1362 when compiling without UCP support, or within Exim, where the ucp files are
1363 not imported.
1364
1365 3. Refactoring of the library code to split up the various functions into
1366 different source modules. The addition of the new DFA matching code (see
1367 below) to a single monolithic source would have made it really too
1368 unwieldy, quite apart from causing all the code to be include in a
1369 statically linked application, when only some functions are used. This is
1370 relevant even without the DFA addition now that patterns can be compiled in
1371 one application and matched in another.
1372
1373 The downside of splitting up is that there have to be some external
1374 functions and data tables that are used internally in different modules of
1375 the library but which are not part of the API. These have all had their
1376 names changed to start with "_pcre_" so that they are unlikely to clash
1377 with other external names.
1378
1379 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1380 a different (DFA) algorithm. Although it is slower than the original
1381 function, it does have some advantages for certain types of matching
1382 problem.
1383
1384 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1385 including restarting after a partial match.
1386
1387 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1388 defined when compiling for Windows was sent to me. I have put it into the
1389 code, though I have no means of testing or verifying it.
1390
1391 7. Added the pcre_refcount() auxiliary function.
1392
1393 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1394 match before or at the first newline in the subject string. In pcretest,
1395 the /f option on a pattern can be used to set this.
1396
1397 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1398 would behave wrongly. This has been present in PCRE since release 4.0.
1399
1400 10. A number of changes to the pcregrep command:
1401
1402 (a) Refactored how -x works; insert ^(...)$ instead of setting
1403 PCRE_ANCHORED and checking the length, in preparation for adding
1404 something similar for -w.
1405
1406 (b) Added the -w (match as a word) option.
1407
1408 (c) Refactored the way lines are read and buffered so as to have more
1409 than one at a time available.
1410
1411 (d) Implemented a pcregrep test script.
1412
1413 (e) Added the -M (multiline match) option. This allows patterns to match
1414 over several lines of the subject. The buffering ensures that at least
1415 8K, or the rest of the document (whichever is the shorter) is available
1416 for matching (and similarly the previous 8K for lookbehind assertions).
1417
1418 (f) Changed the --help output so that it now says
1419
1420 -w, --word-regex(p)
1421
1422 instead of two lines, one with "regex" and the other with "regexp"
1423 because that confused at least one person since the short forms are the
1424 same. (This required a bit of code, as the output is generated
1425 automatically from a table. It wasn't just a text change.)
1426
1427 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1428 option but starts with a hyphen. Could be a pattern or a path name
1429 starting with a hyphen, for instance.
1430
1431 (h) "-" can be given as a file name to represent stdin.
1432
1433 (i) When file names are being printed, "(standard input)" is used for
1434 the standard input, for compatibility with GNU grep. Previously
1435 "<stdin>" was used.
1436
1437 (j) The option --label=xxx can be used to supply a name to be used for
1438 stdin when file names are being printed. There is no short form.
1439
1440 (k) Re-factored the options decoding logic because we are going to add
1441 two more options that take data. Such options can now be given in four
1442 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1443
1444 (l) Added the -A, -B, and -C options for requesting that lines of context
1445 around matches be printed.
1446
1447 (m) Added the -L option to print the names of files that do not contain
1448 any matching lines, that is, the complement of -l.
1449
1450 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1451 continue to scan other files.
1452
1453 (o) The -s option was incorrectly implemented. For compatibility with other
1454 greps, it now suppresses the error message for a non-existent or non-
1455 accessible file (but not the return code). There is a new option called
1456 -q that suppresses the output of matching lines, which was what -s was
1457 previously doing.
1458
1459 (p) Added --include and --exclude options to specify files for inclusion
1460 and exclusion when recursing.
1461
1462 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1463 Hopefully, it now does.
1464
1465 12. Missing cast in pcre_study().
1466
1467 13. Added an "uninstall" target to the makefile.
1468
1469 14. Replaced "extern" in the function prototypes in Makefile.in with
1470 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1471 world, but is set differently for Windows.
1472
1473 15. Added a second compiling function called pcre_compile2(). The only
1474 difference is that it has an extra argument, which is a pointer to an
1475 integer error code. When there is a compile-time failure, this is set
1476 non-zero, in addition to the error test pointer being set to point to an
1477 error message. The new argument may be NULL if no error number is required
1478 (but then you may as well call pcre_compile(), which is now just a
1479 wrapper). This facility is provided because some applications need a
1480 numeric error indication, but it has also enabled me to tidy up the way
1481 compile-time errors are handled in the POSIX wrapper.
1482
1483 16. Added VPATH=.libs to the makefile; this should help when building with one
1484 prefix path and installing with another. (Or so I'm told by someone who
1485 knows more about this stuff than I do.)
1486
1487 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1488 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1489 match everything, including newlines. This is not POSIX-compatible, but
1490 somebody wanted the feature. From pcretest it can be activated by using
1491 both the P and the s flags.
1492
1493 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1494
1495 19. libpcre.pc was being incorrectly installed as executable.
1496
1497 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1498 it now also looks for '\r' so that it will work unmodified on Windows.
1499
1500 21. Added Google's contributed C++ wrapper to the distribution.
1501
1502 22. Added some untidy missing memory free() calls in pcretest, to keep
1503 Electric Fence happy when testing.
1504
1505
1506
1507 Version 5.0 13-Sep-04
1508 ---------------------
1509
1510 1. Internal change: literal characters are no longer packed up into items
1511 containing multiple characters in a single byte-string. Each character
1512 is now matched using a separate opcode. However, there may be more than one
1513 byte in the character in UTF-8 mode.
1514
1515 2. The pcre_callout_block structure has two new fields: pattern_position and
1516 next_item_length. These contain the offset in the pattern to the next match
1517 item, and its length, respectively.
1518
1519 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1520 insertion of callouts before each pattern item. Added the /C option to
1521 pcretest to make use of this.
1522
1523 4. On the advice of a Windows user, the lines
1524
1525 #if defined(_WIN32) || defined(WIN32)
1526 _setmode( _fileno( stdout ), 0x8000 );
1527 #endif /* defined(_WIN32) || defined(WIN32) */
1528
1529 have been added to the source of pcretest. This apparently does useful
1530 magic in relation to line terminators.
1531
1532 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1533 for the benefit of those environments where the "b" makes a difference.
1534
1535 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1536 to know about it. I have put a hack into configure.in that adds in code
1537 to set GCC=yes if CC=icc. This seems to end up at a point in the
1538 generated configure script that is early enough to affect the setting of
1539 compiler options, which is what is needed, but I have no means of testing
1540 whether it really works. (The user who reported this had patched the
1541 generated configure script, which of course I cannot do.)
1542
1543 LATER: After change 22 below (new libtool files), the configure script
1544 seems to know about icc (and also ecc). Therefore, I have commented out
1545 this hack in configure.in.
1546
1547 7. Added support for pkg-config (2 patches were sent in).
1548
1549 8. Negated POSIX character classes that used a combination of internal tables
1550 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1551 [[:^ascii]]. Typically, they would match almost any characters. The other
1552 POSIX classes were not broken in this way.
1553
1554 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1555 to find the match, as PCRE was deluded into thinking that the match had to
1556 start at the start point or following a newline. The same bug applied to
1557 patterns with negative forward assertions or any backward assertions
1558 preceding ".*" at the start, unless the pattern required a fixed first
1559 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1560
1561 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1562 starting at the last subject character, bytes beyond the end of the subject
1563 string were read.
1564
1565 11. Renamed the variable "class" as "classbits" to make life easier for C++
1566 users. (Previously there was a macro definition, but it apparently wasn't
1567 enough.)
1568
1569 12. Added the new field "tables" to the extra data so that tables can be passed
1570 in at exec time, or the internal tables can be re-selected. This allows
1571 a compiled regex to be saved and re-used at a later time by a different
1572 program that might have everything at different addresses.
1573
1574 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1575 -R library as well as a -L library.
1576
1577 14. The debugging options of pcretest (-d on the command line or D on a
1578 pattern) showed incorrect output for anything following an extended class
1579 that contained multibyte characters and which was followed by a quantifier.
1580
1581 15. Added optional support for general category Unicode character properties
1582 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1583 support. It adds about 90K to the size of the library. The meanings of the
1584 inbuilt class escapes such as \d and \s have NOT been changed.
1585
1586 16. Updated pcredemo.c to include calls to free() to release the memory for the
1587 compiled pattern.
1588
1589 17. The generated file chartables.c was being created in the source directory
1590 instead of in the building directory. This caused the build to fail if the
1591 source directory was different from the building directory, and was
1592 read-only.
1593
1594 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1595 file. No doubt somebody will tell me if they don't make sense... Also added
1596 Dan Mooney's comments about building on OpenVMS.
1597
1598 19. Added support for partial matching via the PCRE_PARTIAL option for
1599 pcre_exec() and the \P data escape in pcretest.
1600
1601 20. Extended pcretest with 3 new pattern features:
1602
1603 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1604 write the compiled pattern to the file whose name is "rest-of-line".
1605 This is a straight binary dump of the data, with the saved pointer to
1606 the character tables forced to be NULL. The study data, if any, is
1607 written too. After writing, pcretest reads a new pattern.
1608
1609 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1610 compiled pattern from the given file. There must not be any
1611 occurrences of "<" in the file name (pretty unlikely); if there are,
1612 pcretest will instead treat the initial "<" as a pattern delimiter.
1613 After reading in the pattern, pcretest goes on to read data lines as
1614 usual.
1615
1616 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1617 and 16-bit fields in a compiled pattern, to simulate a pattern that
1618 was compiled on a host of opposite endianness.
1619
1620 21. The pcre-exec() function can now cope with patterns that were compiled on
1621 hosts of opposite endianness, with this restriction:
1622
1623 As for any compiled expression that is saved and used later, the tables
1624 pointer field cannot be preserved; the extra_data field in the arguments
1625 to pcre_exec() should be used to pass in a tables address if a value
1626 other than the default internal tables were used at compile time.
1627
1628 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1629 now diagnosed as an error. Previously, most of the time, a negative number
1630 would have been treated as zero, but if in addition "ovector" was passed as
1631 NULL, a crash could occur.
1632
1633 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1634 new versions from the libtool 1.5 distribution (the last one is a copy of
1635 a file called libtool.m4). This seems to have fixed the need to patch
1636 "configure" to support Darwin 1.3 (which I used to do). However, I still
1637 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1638 workstation).
1639
1640 24. Changed the PCRE licence to be the more standard "BSD" licence.
1641
1642
1643 Version 4.5 01-Dec-03
1644 ---------------------
1645
1646 1. There has been some re-arrangement of the code for the match() function so
1647 that it can be compiled in a version that does not call itself recursively.
1648 Instead, it keeps those local variables that need separate instances for
1649 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1650 needs to "recurse". Keeping track of where control must go is done by means
1651 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1652 hide most of the details from the main code, and operates only if
1653 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1654 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1655 operating.
1656
1657 To make it easier for callers to provide specially tailored get/free
1658 functions for this usage, two new functions, pcre_stack_malloc, and
1659 pcre_stack_free, are used. They are always called in strict stacking order,
1660 and the size of block requested is always the same.
1661
1662 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1663 PCRE has been compiled to use the stack or the heap for recursion. The
1664 -C option of pcretest uses this to show which version is compiled.
1665
1666 A new data escape \S, is added to pcretest; it causes the amounts of store
1667 obtained and freed by both kinds of malloc/free at match time to be added
1668 to the output.
1669
1670 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1671 what's available on my current Linux desktop machine.
1672
1673 3. When matching a UTF-8 string, the test for a valid string at the start has
1674 been extended. If start_offset is not zero, PCRE now checks that it points
1675 to a byte that is the start of a UTF-8 character. If not, it returns
1676 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1677 this is necessary because there may be backward assertions in the pattern.
1678 When matching the same subject several times, it may save resources to use
1679 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1680
1681 4. The code for checking the validity of UTF-8 strings has been tightened so
1682 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1683 containing "overlong sequences".
1684
1685 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1686 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1687 should have been "&", but it just so happened that all the cases this let
1688 through by mistake were picked up later in the function.
1689
1690 6. I had used a variable called "isblank" - this is a C99 function, causing
1691 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1692
1693 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1694 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1695 so the version is shown; (c) stop "make test" echoing "./RunTest".
1696
1697 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1698
1699 9. The prototype for memmove() for systems that don't have it was using
1700 size_t, but the inclusion of the header that defines size_t was later. I've
1701 moved the #includes for the C headers earlier to avoid this.
1702
1703 10. Added some adjustments to the code to make it easier to compiler on certain
1704 special systems:
1705
1706 (a) Some "const" qualifiers were missing.
1707 (b) Added the macro EXPORT before all exported functions; by default this
1708 is defined to be empty.
1709 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1710 that it reads its output file name as an argument instead of writing
1711 to the standard output and assuming this can be redirected.
1712
1713 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1714 class containing characters with values greater than 255, PCRE compilation
1715 went into a loop.
1716
1717 12. A recursive reference to a subpattern that was within another subpattern
1718 that had a minimum quantifier of zero caused PCRE to crash. For example,
1719 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1720 recursion. If the recursively-called subpattern itself had a zero repeat,
1721 that was OK.
1722
1723 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1724 buffer into which it was copied (for escape processing) was still set at
1725 1024, so long lines caused crashes.
1726
1727 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1728 "internal error: code overflow...". This applied to any character class
1729 that was followed by a possessive quantifier.
1730
1731 15. Modified the Makefile to add libpcre.la as a prerequisite for
1732 libpcreposix.la because I was told this is needed for a parallel build to
1733 work.
1734
1735 16. If a pattern that contained .* following optional items at the start was
1736 studied, the wrong optimizing data was generated, leading to matching
1737 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1738 matching string must start with a or b or c. The correct conclusion for
1739 this pattern is that a match can start with any character.
1740
1741
1742 Version 4.4 13-Aug-03
1743 ---------------------
1744
1745 1. In UTF-8 mode, a character class containing characters with values between
1746 127 and 255 was not handled correctly if the compiled pattern was studied.
1747 In fixing this, I have also improved the studying algorithm for such
1748 classes (slightly).
1749
1750 2. Three internal functions had redundant arguments passed to them. Removal
1751 might give a very teeny performance improvement.
1752
1753 3. Documentation bug: the value of the capture_top field in a callout is *one
1754 more than* the number of the hightest numbered captured substring.
1755
1756 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1757 in incorrectly linking with a previously installed version. They now link
1758 explicitly with libpcre.la.
1759
1760 5. configure.in no longer needs to recognize Cygwin specially.
1761
1762 6. A problem in pcre.in for Windows platforms is fixed.
1763
1764 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1765 pcretest, it used to include the size of the study block as part of its
1766 output. Unfortunately, the structure contains a field that has a different
1767 size on different hardware architectures. This meant that the tests that
1768 showed this size failed. As the block is currently always of a fixed size,
1769 this information isn't actually particularly useful in pcretest output, so
1770 I have just removed it.
1771
1772 8. Three pre-processor statements accidentally did not start in column 1.
1773 Sadly, there are *still* compilers around that complain, even though
1774 standard C has not required this for well over a decade. Sigh.
1775
1776 9. In pcretest, the code for checking callouts passed small integers in the
1777 callout_data field, which is a void * field. However, some picky compilers
1778 complained about the casts involved for this on 64-bit systems. Now
1779 pcretest passes the address of the small integer instead, which should get
1780 rid of the warnings.
1781
1782 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1783 both compile and run time, and gives an error if an invalid UTF-8 sequence
1784 is found. There is a option for disabling this check in cases where the
1785 string is known to be correct and/or the maximum performance is wanted.
1786
1787 11. In response to a bug report, I changed one line in Makefile.in from
1788
1789 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1790 to
1791 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1792
1793 to look similar to other lines, but I have no way of telling whether this
1794 is the right thing to do, as I do not use Windows. No doubt I'll get told
1795 if it's wrong...
1796
1797
1798 Version 4.3 21-May-03
1799 ---------------------
1800
1801 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1802 Makefile.
1803
1804 2. Some refactoring to improve the quality of the code:
1805
1806 (i) The utf8_table... variables are now declared "const".
1807
1808 (ii) The code for \cx, which used the "case flipping" table to upper case
1809 lower case letters, now just substracts 32. This is ASCII-specific,
1810 but the whole concept of \cx is ASCII-specific, so it seems
1811 reasonable.
1812
1813 (iii) PCRE was using its character types table to recognize decimal and
1814 hexadecimal digits in the pattern. This is silly, because it handles
1815 only 0-9, a-f, and A-F, but the character types table is locale-
1816 specific, which means strange things might happen. A private
1817 table is now used for this - though it costs 256 bytes, a table is
1818 much faster than multiple explicit tests. Of course, the standard
1819 character types table is still used for matching digits in subject
1820 strings against \d.
1821
1822 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1823 ending in _t are). So I've renamed it as ESC_tee.
1824
1825 3. The first argument for regexec() in the POSIX wrapper should have been
1826 defined as "const".
1827
1828 4. Changed pcretest to use malloc() for its buffers so that they can be
1829 Electric Fenced for debugging.
1830
1831 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1832 to read one or more bytes before the start of the subject string. Often this
1833 had no effect on PCRE's behaviour, but in some circumstances it could
1834 provoke a segmentation fault.
1835
1836 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1837 to try to read one or more bytes before the start of the subject string.
1838
1839 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1840 UTF-8 support could misbehave in various ways if the subject string
1841 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1842 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1843 back over UTF-8 characters.)
1844
1845
1846 Version 4.2 14-Apr-03
1847 ---------------------
1848
1849 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1850
1851 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1852 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1853 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1854 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1855 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1856 and BUILD_EXEEXT
1857 Note: automatic setting of the BUILD variables is not yet working
1858 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1859 compile-time but not at link-time
1860 [LINK]: use for linking executables only
1861 make different versions for Windows and non-Windows
1862 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1863 libraries
1864 [LINK_FOR_BUILD]: new variable
1865 [OBJEXT]: use throughout
1866 [EXEEXT]: use throughout
1867 <winshared>: new target
1868 <wininstall>: new target
1869 <dftables.o>: use native compiler
1870 <dftables>: use native linker
1871 <install>: handle Windows platform correctly
1872 <clean>: ditto
1873 <check>: ditto
1874 copy DLL to top builddir before testing
1875
1876 As part of these changes, -no-undefined was removed again. This was reported
1877 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1878 in any case.
1879
1880 3. Some tidies to get rid of compiler warnings:
1881
1882 . In the match_data structure, match_limit was an unsigned long int, whereas
1883 match_call_count was an int. I've made them both unsigned long ints.
1884
1885 . In pcretest the fact that a const uschar * doesn't automatically cast to
1886 a void * provoked a warning.
1887
1888 . Turning on some more compiler warnings threw up some "shadow" variables
1889 and a few more missing casts.
1890
1891 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1892 option, a class that contained a single character with a value between 128
1893 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1894
1895 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1896 option, a class that contained several characters, but with at least one
1897 whose value was between 128 and 255 caused PCRE to crash.
1898
1899
1900 Version 4.1 12-Mar-03
1901 ---------------------
1902
1903 1. Compiling with gcc -pedantic found a couple of places where casts were
1904 needed, and a string in dftables.c that was longer than standard compilers are
1905 required to support.
1906
1907 2. Compiling with Sun's compiler found a few more places where the code could
1908 be tidied up in order to avoid warnings.
1909
1910 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1911 first of these names is deprecated in the latest Autoconf in favour of the name
1912 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1913 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1914 analogy I have changed it to CFLAGS_FOR_BUILD.
1915
1916 4. Added -no-undefined to the linking command in the Makefile, because this is
1917 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1918 linking step for the pcreposix library.
1919
1920 5. PCRE was failing to diagnose the case of two named groups with the same
1921 name.
1922
1923 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1924 literal character that is needed in the subject for a match, and scans along to
1925 ensure that it is present before embarking on the full matching process. This
1926 saves time in cases of nested unlimited repeats that are never going to match.
1927 Problem: the scan can take a lot of time if the subject is very long (e.g.
1928 megabytes), thus penalizing straightforward matches. It is now done only if the
1929 amount of subject to be scanned is less than 1000 bytes.
1930
1931 7. A lesser problem with the same optimization is that it was recording the
1932 first character of an anchored pattern as "needed", thus provoking a search
1933 right along the subject, even when the first match of the pattern was going to
1934 fail. The "needed" character is now not set for anchored patterns, unless it
1935 follows something in the pattern that is of non-fixed length. Thus, it still
1936 fulfils its original purpose of finding quick non-matches in cases of nested
1937 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1938
1939
1940 Version 4.0 17-Feb-03
1941 ---------------------
1942
1943 1. If a comment in an extended regex that started immediately after a meta-item
1944 extended to the end of string, PCRE compiled incorrect data. This could lead to
1945 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
1946
1947 2. Moved to autoconf 2.53 and libtool 1.4.2.
1948
1949 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
1950 the special perltest8 script is no longer needed - all the tests can be run
1951 from a single perltest script.
1952
1953 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
1954 by \s. It has now been removed in PCRE. This means it isn't recognized as
1955 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
1956 class [:space:] *does* include VT, thereby creating a mess.
1957
1958 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
1959 space and tab.
1960
1961 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
1962 its new features into the main test script, reducing the number of scripts.
1963
1964 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
1965 were backward compatible, and made the (?i) apply to the whole pattern, as if
1966 /i were given. Now it behaves more logically, and applies the option setting
1967 only to what follows. PCRE has been changed to follow suit. However, if it
1968 finds options settings right at the start of the pattern, it extracts them into
1969 the global options, as before. Thus, they show up in the info data.
1970
1971 8. Added support for the \Q...\E escape sequence. Characters in between are
1972 treated as literals. This is slightly different from Perl in that $ and @ are
1973 also handled as literals inside the quotes. In Perl, they will cause variable
1974 interpolation. Note the following examples:
1975
1976 Pattern PCRE matches Perl matches
1977
1978 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
1979 \Qabc\$xyz\E abc\$xyz abc\$xyz
1980 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
1981
1982 For compatibility with Perl, \Q...\E sequences are recognized inside character
1983 classes as well as outside them.
1984
1985 9. Re-organized 3 code statements in pcretest to avoid "overflow in
1986 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
1987 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
1988 signed/unsigned warnings.
1989
1990 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
1991 option for pcretest, so I've replaced it by a simple function that does just
1992 that job.
1993
1994 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
1995 "pcregrep -".
1996
1997 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
1998 Java package. This provides some syntactic sugar for simple cases of what my
1999 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2000 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2001 item, you can use this simplified notation. Note that only makes sense with
2002 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2003 greediness, whatever the setting of the PCRE_UNGREEDY option.
2004
2005 13. A change of greediness default within a pattern was not taking effect at
2006 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2007 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2008 was abstracted outside.
2009
2010 14. PCRE now supports the \G assertion. It is true when the current matching
2011 position is at the start point of the match. This differs from \A when the
2012 starting offset is non-zero. Used with the /g option of pcretest (or similar
2013 code), it works in the same way as it does for Perl's /g option. If all
2014 alternatives of a regex begin with \G, the expression is anchored to the start
2015 match position, and the "anchored" flag is set in the compiled expression.
2016
2017 15. Some bugs concerning the handling of certain option changes within patterns
2018 have been fixed. These applied to options other than (?ims). For example,
2019 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2020 been the other way round. Some of this was related to change 7 above.
2021
2022 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2023 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2024 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2025 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2026
2027 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2028 mode. Unlike ".", it always matches newline, whatever the setting of
2029 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2030 assertions. Perl allows it, but it doesn't (in general) work because it can't
2031 calculate the length of the lookbehind. At least, that's the case for Perl
2032 5.8.0 - I've been told they are going to document that it doesn't work in
2033 future.
2034
2035 18. Added an error diagnosis for escapes that PCRE does not support: these are
2036 \L, \l, \N, \P, \p, \U, \u, and \X.
2037
2038 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2039 reading past the end of the pattern in cases such as /[abcd/.
2040
2041 20. PCRE was getting more memory than necessary for patterns with classes that
2042 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2043
2044 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2045 compiling PCRE for use with Virtual Pascal.
2046
2047 22. Small fix to the Makefile to make it work properly if the build is done
2048 outside the source tree.
2049
2050 23. Added a new extension: a condition to go with recursion. If a conditional
2051 subpattern starts with (?(R) the "true" branch is used if recursion has
2052 happened, whereas the "false" branch is used only at the top level.
2053
2054 24. When there was a very long string of literal characters (over 255 bytes
2055 without UTF support, over 250 bytes with UTF support), the computation of how
2056 much memory was required could be incorrect, leading to segfaults or other
2057 strange effects.
2058
2059 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2060 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2061 there was a subsequent back reference to those brackets. This meant that, for
2062 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2063 possible to check for precisely this case. All we can do is abandon the
2064 optimization if .* occurs inside capturing brackets when there are any back
2065 references whatsoever. (See below for a better fix that came later.)
2066
2067 26. The handling of the optimization for finding the first character of a
2068 non-anchored pattern, and for finding a character that is required later in the
2069 match were failing in some cases. This didn't break the matching; it just
2070 failed to optimize when it could. The way this is done has been re-implemented.
2071
2072 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2073
2074 28. Added a new feature that provides some of the functionality that Perl
2075 provides with (?{...}). The facility is termed a "callout". The way it is done
2076 in PCRE is for the caller to provide an optional function, by setting
2077 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2078 global variable. By default it is unset, which disables all calling out. To get
2079 the function called, the regex must include (?C) at appropriate points. This
2080 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2081 This provides a means of identifying different callout points. When PCRE
2082 reaches such a point in the regex, if pcre_callout has been set, the external
2083 function is called. It is provided with data in a structure called
2084 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2085 matching continues; if it returns a non-zero value, the match at the current
2086 point fails. However, backtracking will occur if possible. [This was changed
2087 later and other features added - see item 49 below.]
2088
2089 29. pcretest is upgraded to test the callout functionality. It provides a
2090 callout function that displays information. By default, it shows the start of
2091 the match and the current position in the text. There are some new data escapes
2092 to vary what happens:
2093
2094 \C+ in addition, show current contents of captured substrings
2095 \C- do not supply a callout function
2096 \C!n return 1 when callout number n is reached
2097 \C!n!m return 1 when callout number n is reached for the mth time
2098
2099 30. If pcregrep was called with the -l option and just a single file name, it
2100 output "<stdin>" if a match was found, instead of the file name.
2101
2102 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2103 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2104 pcre_exec(). This saves a malloc/free per call. The default value of
2105 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2106 when configuring.
2107
2108 32. The default maximum size of a compiled pattern is 64K. There have been a
2109 few cases of people hitting this limit. The code now uses macros to handle the
2110 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2111 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2112 configuring. Tests 2 and 5 work only with 2-byte links because they output
2113 debugging information about compiled patterns.
2114
2115 33. Internal code re-arrangements:
2116
2117 (a) Moved the debugging function for printing out a compiled regex into
2118 its own source file (printint.c) and used #include to pull it into
2119 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2120 separate copies.
2121
2122 (b) Defined the list of op-code names for debugging as a macro in
2123 internal.h so that it is next to the definition of the opcodes.
2124
2125 (c) Defined a table of op-code lengths for simpler skipping along compiled
2126 code. This is again a macro in internal.h so that it is next to the
2127 definition of the opcodes.
2128
2129 34. Added support for recursive calls to individual subpatterns, along the
2130 lines of Robin Houston's patch (but implemented somewhat differently).
2131
2132 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2133 allow it to read and process whole directories in Win32. This code was
2134 contributed by Lionel Fourquaux; it has not been tested by me.
2135
2136 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2137 used to name a group. Names consist of alphanumerics and underscores, and must
2138 be unique. Back references use the syntax (?P=name) and recursive calls use
2139 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2140 numbers. The function pcre_fullinfo() can be used after compilation to extract
2141 a name/number map. There are three relevant calls:
2142
2143 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2144 PCRE_INFO_NAMECOUNT yields the number of entries
2145 PCRE_INFO_NAMETABLE yields a pointer to the map.
2146
2147 The map is a vector of fixed-size entries. The size of each entry depends on
2148 the length of the longest name used. The first two bytes of each entry are the
2149 group number, most significant byte first. There follows the corresponding
2150 name, zero terminated. The names are in alphabetical order.
2151
2152 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2153 case instead of 255. Making it the same both with and without UTF-8 support
2154 means that the same test output works with both.
2155
2156 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2157 calling malloc() with a zero argument.
2158
2159 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2160 optimization. I've improved things by keeping a bitmap of backreferences with
2161 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2162 fact referenced, the optimization can be applied. It is unlikely that a
2163 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2164 the match to follow \n) will appear inside brackets with a number greater than
2165 31, but if it does, any back reference > 31 suppresses the optimization.
2166
2167 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2168 of disabling numbered capturing parentheses. Any opening parenthesis that is
2169 not followed by ? behaves as if it were followed by ?: but named parentheses
2170 can still be used for capturing (and they will acquire numbers in the usual
2171 way).
2172
2173 41. Redesigned the return codes from the match() function into yes/no/error so
2174 that errors can be passed back from deep inside the nested calls. A malloc
2175 failure while inside a recursive subpattern call now causes the
2176 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2177
2178 42. It is now possible to set a limit on the number of times the match()
2179 function is called in a call to pcre_exec(). This facility makes it possible to
2180 limit the amount of recursion and backtracking, though not in a directly
2181 obvious way, because the match() function is used in a number of different
2182 circumstances. The count starts from zero for each position in the subject
2183 string (for non-anchored patterns). The default limit is, for compatibility, a
2184 large number, namely 10 000 000. You can change this in two ways:
2185
2186 (a) When configuring PCRE before making, you can use --with-match-limit=n
2187 to set a default value for the compiled library.
2188
2189 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2190 a different value is set. See 45 below.
2191
2192 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2193
2194 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2195 of things that can be changed at compile time. The first argument specifies
2196 what is wanted and the second points to where the information is to be placed.
2197 The current list of available information is:
2198
2199 PCRE_CONFIG_UTF8
2200
2201 The output is an integer that is set to one if UTF-8 support is available;
2202 otherwise it is set to zero.
2203
2204 PCRE_CONFIG_NEWLINE
2205
2206 The output is an integer that it set to the value of the code that is used for
2207 newline. It is either LF (10) or CR (13).
2208
2209 PCRE_CONFIG_LINK_SIZE
2210
2211 The output is an integer that contains the number of bytes used for internal
2212 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2213
2214 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2215
2216 The output is an integer that contains the threshold above which the POSIX
2217 interface uses malloc() for output vectors. See item 31 above.
2218
2219 PCRE_CONFIG_MATCH_LIMIT
2220
2221 The output is an unsigned integer that contains the default limit of the number
2222 of match() calls in a pcre_exec() execution. See 42 above.
2223
2224 44. pcretest has been upgraded by the addition of the -C option. This causes it
2225 to extract all the available output from the new pcre_config() function, and to
2226 output it. The program then exits immediately.
2227
2228 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2229 order to support additional features. One way would have been to define
2230 pcre_exec2() (for example) with extra arguments, but this would not have been
2231 extensible, and would also have required all calls to the original function to
2232 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2233 is used for passing in "extra" data from pcre_study().
2234
2235 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2236 contains the following fields:
2237
2238 flags a bitmap indicating which of the following fields are set
2239 study_data opaque data from pcre_study()
2240 match_limit a way of specifying a limit on match() calls for a specific
2241 call to pcre_exec()
2242 callout_data data for callouts (see 49 below)
2243
2244 The flag bits are also defined in pcre.h, and are
2245
2246 PCRE_EXTRA_STUDY_DATA
2247 PCRE_EXTRA_MATCH_LIMIT
2248 PCRE_EXTRA_CALLOUT_DATA
2249
2250 The pcre_study() function now returns one of these new pcre_extra blocks, with
2251 the actual study data pointed to by the study_data field, and the
2252 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2253 before. That is, this change is entirely upwards-compatible and requires no
2254 change to existing code.
2255
2256 If you want to pass in additional data to pcre_exec(), you can either place it
2257 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2258 block.
2259
2260 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2261 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2262 times with different match limits, until it finds the minimum value needed for
2263 pcre_exec() to complete. The value is then output. This can be instructive; for
2264 most simple matches the number is quite small, but for pathological cases it
2265 gets very large very quickly.
2266
2267 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2268 returns the size of the data block pointed to by the study_data field in a
2269 pcre_extra block, that is, the value that was passed as the argument to
2270 pcre_malloc() when PCRE was getting memory in which to place the information
2271 created by pcre_study(). The fourth argument should point to a size_t variable.
2272 pcretest has been extended so that this information is shown after a successful
2273 pcre_study() call when information about the compiled regex is being displayed.
2274
2275 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2276 because what follows is always an absolute path. (Later: it turns out that this
2277 is more than cosmetic for MinGW, because it doesn't like empty path
2278 components.)
2279
2280 49. Some changes have been made to the callout feature (see 28 above):
2281
2282 (i) A callout function now has three choices for what it returns:
2283
2284 0 => success, carry on matching
2285 > 0 => failure at this point, but backtrack if possible
2286 < 0 => serious error, return this value from pcre_exec()
2287
2288 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2289 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2290 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2291 use by callout functions. It will never be used by PCRE itself.
2292
2293 (ii) The pcre_extra structure (see 45 above) has a void * field called
2294 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2295 pcre_callout_block structure has a field of the same name. The contents of
2296 the field passed in the pcre_extra structure are passed to the callout
2297 function in the corresponding field in the callout block. This makes it
2298 easier to use the same callout-containing regex from multiple threads. For
2299 testing, the pcretest program has a new data escape
2300
2301 \C*n pass the number n (may be negative) as callout_data
2302
2303 If the callout function in pcretest receives a non-zero value as
2304 callout_data, it returns that value.
2305
2306 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2307 there were some redundant $(CFLAGS) in commands that are now specified as
2308 $(LINK), which already includes $(CFLAGS).
2309
2310 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2311 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2312 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2313 one-byte characters throughout. Note that case-insensitive matching applies
2314 only to characters whose values are less than 256. PCRE doesn't support the
2315 notion of cases for higher-valued characters.
2316
2317 (i) A character class whose characters are all within 0-255 is handled as
2318 a bit map, and the map is inverted for negative classes. Previously, a
2319 character > 255 always failed to match such a class; however it should
2320 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2321
2322 (ii) A negated character class with a single character < 255 is coded as
2323 "not this character" (OP_NOT). This wasn't working properly when the test
2324 character was multibyte, either singly or repeated.
2325
2326 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2327 mode, for example: \x{100}{2,3}.
2328
2329 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2330 singly or repeated) now correctly test multibyte characters. However,
2331 PCRE doesn't recognize any characters with values greater than 255 as
2332 digits, spaces, or word characters. Such characters always match \D, \S,
2333 and \W, and never match \d, \s, or \w.
2334
2335 (v) Classes may now contain characters and character ranges with values
2336 greater than 255. For example: [ab\x{100}-\x{400}].
2337
2338 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2339 PCRE in UTF-8 mode.
2340
2341 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2342 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2343 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2344 value.)
2345
2346 53. The single man page has become too large. I have therefore split it up into
2347 a number of separate man pages. These also give rise to individual HTML pages;
2348 these are now put in a separate directory, and there is an index.html page that
2349 lists them all. Some hyperlinking between the pages has been installed.
2350
2351 54. Added convenience functions for handling named capturing parentheses.
2352
2353 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2354 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2355 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2356 are faulted.
2357
2358 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2359 calling configure. These values are used when compiling the dftables.c program
2360 which is run to generate the source of the default character tables. They
2361 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2362 you will need to set these values.
2363
2364 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2365
2366
2367 Version 3.9 02-Jan-02
2368 ---------------------
2369
2370 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2371
2372 2. If --disable-static was given, the building process failed when trying to
2373 build pcretest and pcregrep. (For some reason it was using libtool to compile
2374 them, which is not right, as they aren't part of the library.)
2375
2376
2377 Version 3.8 18-Dec-01
2378 ---------------------
2379
2380 1. The experimental UTF-8 code was completely screwed up. It was packing the
2381 bytes in the wrong order. How dumb can you get?
2382
2383
2384 Version 3.7 29-Oct-01
2385 ---------------------
2386
2387 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2388 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2389 this didn't happen under Solaris 8, where I normally test things.
2390
2391 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2392 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2393 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2394 this fix an hour or so after the initial 3.7 release.)
2395
2396
2397 Version 3.6 23-Oct-01
2398 ---------------------
2399
2400 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2401 offsets passed as NULL with zero offset count.
2402
2403 2. The config.guess and config.sub files had not been updated when I moved to
2404 the latest autoconf.
2405
2406
2407 Version 3.5 15-Aug-01
2408 ---------------------
2409
2410 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2411 had been forgotten.
2412
2413 2. By using declared but undefined structures, we can avoid using "void"
2414 definitions in pcre.h while keeping the internal definitions of the structures
2415 private.
2416
2417 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2418 user point of view, this means that both static and shared libraries are built
2419 by default, but this can be individually controlled. More of the work of
2420 handling this static/shared cases is now inside libtool instead of PCRE's make
2421 file.
2422
2423 4. The pcretest utility is now installed along with pcregrep because it is
2424 useful for users (to test regexs) and by doing this, it automatically gets
2425 relinked by libtool. The documentation has been turned into a man page, so
2426 there are now .1, .txt, and .html versions in /doc.
2427
2428 5. Upgrades to pcregrep:
2429 (i) Added long-form option names like gnu grep.
2430 (ii) Added --help to list all options with an explanatory phrase.
2431 (iii) Added -r, --recursive to recurse into sub-directories.
2432 (iv) Added -f, --file to read patterns from a file.
2433
2434 6. pcre_exec() was referring to its "code" argument before testing that
2435 argument for NULL (and giving an error if it was NULL).
2436
2437 7. Upgraded Makefile.in to allow for compiling in a different directory from
2438 the source directory.
2439
2440 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2441 options bits, the pointer it was passed was to an int instead of to an unsigned
2442 long int. This mattered only on 64-bit systems.
2443
2444 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2445 generated) instead of pcre.in, which it its source. Also made the same change
2446 in several of the .c files.
2447
2448 10. A new release of gcc defines printf() as a macro, which broke pcretest
2449 because it had an ifdef in the middle of a string argument for printf(). Fixed
2450 by using separate calls to printf().
2451
2452 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2453 script, to force use of CR or LF instead of \n in the source. On non-Unix
2454 systems, the value can be set in config.h.
2455
2456 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2457 absolute limit. Changed the text of the error message to make this clear, and
2458 likewise updated the man page.
2459
2460 13. The limit of 99 on the number of capturing subpatterns has been removed.
2461 The new limit is 65535, which I hope will not be a "real" limit.
2462
2463
2464 Version 3.4 22-Aug-00
2465 ---------------------
2466
2467 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2468
2469 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2470
2471
2472 Version 3.3 01-Aug-00
2473 ---------------------
2474
2475 1. If an octal character was given, but the value was greater than \377, it
2476 was not getting masked to the least significant bits, as documented. This could
2477 lead to crashes in some systems.
2478
2479 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2480 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2481
2482 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2483 These just pass their arguments on to (pcre_free)(), but they are provided
2484 because some uses of PCRE bind it to non-C systems that can call its functions,
2485 but cannot call free() or pcre_free() directly.
2486
2487 4. Add "make test" as a synonym for "make check". Corrected some comments in
2488 the Makefile.
2489
2490 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2491 Makefile.
2492
2493 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2494 command called pgrep for grepping around the active processes.
2495
2496 7. Added the beginnings of support for UTF-8 character strings.
2497
2498 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2499 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2500 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2501 out for the ar command.)
2502
2503
2504 Version 3.2 12-May-00
2505 ---------------------
2506
2507 This is purely a bug fixing release.
2508
2509 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2510 of ZA. This was just one example of several cases that could provoke this bug,
2511 which was introduced by change 9 of version 2.00. The code for breaking
2512 infinite loops after an iteration that matches an empty string was't working
2513 correctly.
2514
2515 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2516 when matched against abbab (for example). After matching an empty string, it
2517 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2518 caused it to match further down the string than it should.
2519
2520 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2521 was there because it doesn't seem to be needed, and it causes trouble on some
2522 systems, as it is not a Standard C header. It has been removed.
2523
2524 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2525 were reported on the Macintosh. The changes were from
2526
2527 while ((c = *(++ptr)) != 0 && c != '\n');
2528 to
2529 while ((c = *(++ptr)) != 0 && c != '\n') ;
2530
2531 Totally extraordinary, but if that's what it takes...
2532
2533 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2534 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2535 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2536 assumes the way PCRE uses memmove() (always moving upwards).
2537
2538 6. PCRE is being used in one environment where strchr() is not available. There
2539 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2540 faster code anyway.
2541
2542
2543 Version 3.1 09-Feb-00
2544 ---------------------
2545
2546 The only change in this release is the fixing of some bugs in Makefile.in for
2547 the "install" target:
2548
2549 (1) It was failing to install pcreposix.h.
2550
2551 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2552
2553
2554 Version 3.0 01-Feb-00
2555 ---------------------
2556
2557 1. Add support for the /+ modifier to perltest (to output $` like it does in
2558 pcretest).
2559
2560 2. Add support for the /g modifier to perltest.
2561
2562 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2563 matches null strings.
2564
2565 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2566 pattern. Perl treats empty patterns specially - it reuses the most recent
2567 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2568 effect.
2569
2570 5. The POSIX interface was broken in that it was just handing over the POSIX
2571 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2572 required a bigger vector, with some working space on the end. This means that
2573 the POSIX wrapper now has to get and free some memory, and copy the results.
2574
2575 6. Added some simple autoconf support, placing the test data and the
2576 documentation in separate directories, re-organizing some of the
2577 information files, and making it build pcre-config (a GNU standard). Also added
2578 libtool support for building PCRE as a shared library, which is now the
2579 default.
2580
2581 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2582 09 are not valid octal constants. Single digits will be used for minor values
2583 less than 10.
2584
2585 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2586 existing programs that set these in the POSIX interface can use PCRE without
2587 modification.
2588
2589 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2590 return all that pcre_info() returns, plus additional data. The pcre_info()
2591 function is retained for compatibility, but is considered to be obsolete.
2592
2593 10. Added experimental recursion feature (?R) to handle one common case that
2594 Perl 5.6 will be able to do with (?p{...}).
2595
2596 11. Added support for POSIX character classes like [:alpha:], which Perl is
2597 adopting.
2598
2599
2600 Version 2.08 31-Aug-99
2601 ----------------------
2602
2603 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2604 trying to match at the startoffset position, but instead was moving forward to
2605 the next newline as if a previous match had failed.
2606
2607 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2608 and could get into a loop if a null string was matched other than at the start
2609 of the subject.
2610
2611 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2612 be distinguished at compile time, and for completeness also added PCRE_DATE.
2613
2614 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2615 in GnuWin32 environments.
2616
2617
2618 Version 2.07 29-Jul-99
2619 ----------------------
2620
2621 1. The documentation is now supplied in plain text form and HTML as well as in
2622 the form of man page sources.
2623
2624 2. C++ compilers don't like assigning (void *) values to other pointer types.
2625 In particular this affects malloc(). Although there is no problem in Standard
2626 C, I've put in casts to keep C++ compilers happy.
2627
2628 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2629 should be (const char *).
2630
2631 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2632 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2633 However, I haven't made this a standard facility. The documentation doesn't
2634 mention it, and the Makefile doesn't support it.
2635
2636 5. The Makefile now contains an "install" target, with editable destinations at
2637 the top of the file. The pcretest program is not installed.
2638
2639 6. pgrep -V now gives the PCRE version number and date.
2640
2641 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2642 causing the entire string to be ignored, instead of just the last character.
2643
2644 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2645 non-matching string, it can take a very, very long time, even for strings of
2646 quite modest length, because of the nested recursion. PCRE now does better in
2647 some of these cases. It does this by remembering the last required literal
2648 character in the pattern, and pre-searching the subject to ensure it is present
2649 before running the real match. In other words, it applies a heuristic to detect
2650 some types of certain failure quickly, and in the above example, if presented
2651 with a string that has no trailing " it gives "no match" very quickly.
2652
2653 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2654 other alternatives are tried instead.
2655
2656
2657 Version 2.06 09-Jun-99
2658 ----------------------
2659
2660 1. Change pcretest's output for amount of store used to show just the code
2661 space, because the remainder (the data block) varies in size between 32-bit and
2662 64-bit systems.
2663
2664 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2665 start matching at. This allows lookbehinds to work when searching for multiple
2666 occurrences in a string.
2667
2668 3. Added additional options to pcretest for testing multiple occurrences:
2669
2670 /+ outputs the rest of the string that follows a match
2671 /g loops for multiple occurrences, using the new startoffset argument
2672 /G loops for multiple occurrences by passing an incremented pointer
2673
2674 4. PCRE wasn't doing the "first character" optimization for patterns starting
2675 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2676 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2677 the letter 'x'. On long subject strings, this gives a significant speed-up.
2678
2679
2680 Version 2.05 21-Apr-99
2681 ----------------------
2682
2683 1. Changed the type of magic_number from int to long int so that it works
2684 properly on 16-bit systems.
2685
2686 2. Fixed a bug which caused patterns starting with .* not to work correctly
2687 when the subject string contained newline characters. PCRE was assuming
2688 anchoring for such patterns in all cases, which is not correct because .* will
2689 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2690 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2691 must be retried after every newline in the subject.
2692
2693
2694 Version 2.04 18-Feb-99
2695 ----------------------
2696
2697 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2698 computation of the store needed to hold the pattern was incorrect (too large).
2699 If such patterns were nested a few deep, this could multiply and become a real
2700 problem.
2701
2702 2. Added /M option to pcretest to show the memory requirement of a specific
2703 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2704
2705 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2706 compiled in such a way that the backtracking after subsequent failure was
2707 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2708 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2709
2710
2711 Version 2.03 02-Feb-99
2712 ----------------------
2713
2714 1. Fixed typo and small mistake in man page.
2715
2716 2. Added 4th condition (GPL supersedes if conflict) and created separate
2717 LICENCE file containing the conditions.
2718
2719 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2720 Perl, that is the internal \ allows the delimiter to be included in the
2721 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2722 the final delimiter, add \ to the end of the pattern (to test the error).
2723
2724 4. Added the convenience functions for extracting substrings after a successful
2725 match. Updated pcretest to make it able to test these functions.
2726
2727
2728 Version 2.02 14-Jan-99
2729 ----------------------
2730
2731 1. Initialized the working variables associated with each extraction so that
2732 their saving and restoring doesn't refer to uninitialized store.
2733
2734 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2735 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2736 fix the problem.
2737
2738 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2739 calls, and wasn't printing the correct value for compiling calls. Increased the
2740 default value of LOOPREPEAT, and the number of significant figures in the
2741 times.
2742
2743 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2744
2745 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2746 a building problem on Windows NT with a FAT file system.
2747
2748
2749 Version 2.01 21-Oct-98
2750 ----------------------
2751
2752 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2753 to character tables built by pcre_maketables() in the current locale. If NULL
2754 is passed, the default tables are used.
2755
2756
2757 Version 2.00 24-Sep-98
2758 ----------------------
2759
2760 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2761 it any more.
2762
2763 2. Allow quantification of (?>) groups, and make it work correctly.
2764
2765 3. The first character computation wasn't working for (?>) groups.
2766
2767 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2768 end of the subject) and add 5.005's \z, which really does match only at the
2769 very end of the subject.
2770
2771 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2772
2773 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2774 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2775 localized options. All options to pcre_study() were also removed.
2776
2777 7. Add other new features from 5.005:
2778
2779 $(?<= positive lookbehind
2780 $(?<! negative lookbehind
2781 (?imsx-imsx) added the unsetting capability
2782 such a setting is global if at outer level; local otherwise
2783 (?imsx-imsx:) non-capturing groups with option setting
2784 (?(cond)re|re) conditional pattern matching
2785
2786 A backreference to itself in a repeated group matches the previous
2787 captured string.
2788
2789 8. General tidying up of studying (both automatic and via "study")
2790 consequential on the addition of new assertions.
2791
2792 9. As in 5.005, unlimited repeated groups that could match an empty substring
2793 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2794 runtime if any iteration does actually match an empty substring.
2795
2796 10. Include the RunTest script in the distribution.
2797
2798 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2799 discrepancies, some of which were old and were also with respect to 5.004. They
2800 have now been fixed.
2801
2802
2803 Version 1.09 28-Apr-98
2804 ----------------------
2805
2806 1. A negated single character class followed by a quantifier with a minimum
2807 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2808 program crashes, or just wrong answers. This did not apply to negated classes
2809 containing more than one character, or to minima other than one.
2810
2811
2812 Version 1.08 27-Mar-98
2813 ----------------------
2814
2815 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2816
2817 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2818 latter must appear before anything that relies on it in the pattern.
2819
2820
2821 Version 1.07 16-Feb-98
2822 ----------------------
2823
2824 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2825 repeat of a potentially empty string).
2826
2827
2828 Version 1.06 23-Jan-98
2829 ----------------------
2830
2831 1. Added Markus Oberhumer's little patches for C++.
2832
2833 2. Literal strings longer than 255 characters were broken.
2834
2835
2836 Version 1.05 23-Dec-97
2837 ----------------------
2838
2839 1. Negated character classes containing more than one character were failing if
2840 PCRE_CASELESS was set at run time.
2841
2842
2843 Version 1.04 19-Dec-97
2844 ----------------------
2845
2846 1. Corrected the man page, where some "const" qualifiers had been omitted.
2847
2848 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2849 input syntax.
2850
2851 3. Fixed memory leak which occurred when a regex with back references was
2852 matched with an offsets vector that wasn't big enough. The temporary memory
2853 that is used in this case wasn't being freed if the match failed.
2854
2855 4. Tidied pcretest to ensure it frees memory that it gets.
2856
2857 5. Temporary memory was being obtained in the case where the passed offsets
2858 vector was exactly big enough.
2859
2860 6. Corrected definition of offsetof() from change 5 below.
2861
2862 7. I had screwed up change 6 below and broken the rules for the use of
2863 setjmp(). Now fixed.
2864
2865
2866 Version 1.03 18-Dec-97
2867 ----------------------
2868
2869 1. A erroneous regex with a missing opening parenthesis was correctly
2870 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2871 on some systems.
2872
2873 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2874 it was reported that one broken compiler failed on the former because "code" is
2875 also an independent variable.
2876
2877 3. The erroneous regex a[]b caused an array overrun reference.
2878
2879 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2880 fail on data ending with that character. (It was going on too far, and checking
2881 the next character, typically a binary zero.) This was specific to the
2882 optimized code for single-character negative classes.
2883
2884 5. Added a contributed patch from the TIN world which does the following:
2885
2886 + Add an undef for memmove, in case the the system defines a macro for it.
2887
2888 + Add a definition of offsetof(), in case there isn't one. (I don't know
2889 the reason behind this - offsetof() is part of the ANSI standard - but
2890 it does no harm).
2891
2892 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2893 most of the places where whitespace preceded '#'. I have given up and
2894 allowed the remaining 2 cases to be at the margin.
2895
2896 + Rename some variables in pcre to eliminate shadowing. This seems very
2897 pedantic, but does no harm, of course.
2898
2899 6. Moved the call to setjmp() into its own function, to get rid of warnings
2900 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2901
2902 7. Constructs such as \d{8,} were compiling into the equivalent of
2903 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2904 outcome, but in this particular case used more store than had been allocated,
2905 which caused the bug to be discovered because it threw up an internal error.
2906
2907 8. The debugging code in both pcre and pcretest for outputting the compiled
2908 form of a regex was going wrong in the case of back references followed by
2909 curly-bracketed repeats.
2910
2911
2912 Version 1.02 12-Dec-97
2913 ----------------------
2914
2915 1. Typos in pcre.3 and comments in the source fixed.
2916
2917 2. Applied a contributed patch to get rid of places where it used to remove
2918 'const' from variables, and fixed some signed/unsigned and uninitialized
2919 variable warnings.
2920
2921 3. Added the "runtest" target to Makefile.
2922
2923 4. Set default compiler flag to -O2 rather than just -O.
2924
2925
2926 Version 1.01 19-Nov-97
2927 ----------------------
2928
2929 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2930 like /([ab]*)*/, that is, for classes with more than one character in them.
2931
2932 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2933 as /((?>a*))*/ (a PCRE_EXTRA facility).
2934
2935
2936 Version 1.00 18-Nov-97
2937 ----------------------
2938
2939 1. Added compile-time macros to support systems such as SunOS4 which don't have
2940 memmove() or strerror() but have other things that can be used instead.
2941
2942 2. Arranged that "make clean" removes the executables.
2943
2944
2945 Version 0.99 27-Oct-97
2946 ----------------------
2947
2948 1. Fixed bug in code for optimizing classes with only one character. It was
2949 initializing a 32-byte map regardless, which could cause it to run off the end
2950 of the memory it had got.
2951
2952 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
2953
2954
2955 Version 0.98 22-Oct-97
2956 ----------------------
2957
2958 1. Fixed bug in code for handling temporary memory usage when there are more
2959 back references than supplied space in the ovector. This could cause segfaults.
2960
2961
2962 Version 0.97 21-Oct-97
2963 ----------------------
2964
2965 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
2966
2967 2. Optimized negated single characters not to use a bit map.
2968
2969 3. Brought error texts together as macro definitions; clarified some of them;
2970 fixed one that was wrong - it said "range out of order" when it meant "invalid
2971 escape sequence".
2972
2973 4. Changed some char * arguments to const char *.
2974
2975 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
2976
2977 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
2978 pcretest.
2979
2980
2981 Version 0.96 16-Oct-97
2982 ----------------------
2983
2984 1. Added a simple "pgrep" utility to the distribution.
2985
2986 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
2987 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
2988 where "ddd" means "one or more decimal digits".
2989
2990 3. Fixed serious bug. If a pattern had a back reference, but the call to
2991 pcre_exec() didn't supply a large enough ovector to record the related
2992 identifying subpattern, the match always failed. PCRE now remembers the number
2993 of the largest back reference, and gets some temporary memory in which to save
2994 the offsets during matching if necessary, in order to ensure that
2995 backreferences always work.
2996
2997 4. Increased the compatibility with Perl in a number of ways:
2998
2999 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3000 to request this handling. The option can be set at compile or exec time.
3001
3002 (b) $ matches before a terminating newline by default; an option
3003 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3004 mode). The option can be set at compile or exec time.
3005
3006 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3007 the same as Perl's. If the decimal number it represents is less than 10
3008 or there aren't that many previous left capturing parentheses, an octal
3009 escape is read. Inside a character class, it's always an octal escape,
3010 even if it is a single digit.
3011
3012 (d) An escaped but undefined alphabetic character is taken as a literal,
3013 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3014 escapes.
3015
3016 (e) {0} is now permitted. (The previous item is removed from the compiled
3017 pattern).
3018
3019 5. Changed all the names of code files so that the basic parts are no longer
3020 than 10 characters, and abolished the teeny "globals.c" file.
3021
3022 6. Changed the handling of character classes; they are now done with a 32-byte
3023 bit map always.
3024
3025 7. Added the -d and /D options to pcretest to make it possible to look at the
3026 internals of compilation without having to recompile pcre.
3027
3028
3029 Version 0.95 23-Sep-97
3030 ----------------------
3031
3032 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3033 \x20 at the start of a run of normal characters. These were being treated as
3034 real characters, instead of the source characters being re-checked.
3035
3036
3037 Version 0.94 18-Sep-97
3038 ----------------------
3039
3040 1. The functions are now thread-safe, with the caveat that the global variables
3041 containing pointers to malloc() and free() or alternative functions are the
3042 same for all threads.
3043
3044 2. Get pcre_study() to generate a bitmap of initial characters for non-
3045 anchored patterns when this is possible, and use it if passed to pcre_exec().
3046
3047
3048 Version 0.93 15-Sep-97
3049 ----------------------
3050
3051 1. /(b)|(:+)/ was computing an incorrect first character.
3052
3053 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3054 but not actually doing anything yet.
3055
3056 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3057 as Perl does (e.g. [-az] or [az-]).
3058
3059 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3060 all possible positions.
3061
3062 5. Split up into different modules to avoid including unneeded functions in a
3063 compiled binary. However, compile and exec are still in one module. The "study"
3064 function is split off.
3065
3066 6. The character tables are now in a separate module whose source is generated
3067 by an auxiliary program - but can then be edited by hand if required. There are
3068 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3069 toupper() in the code.
3070
3071 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3072 make them global. Abolish the function for setting them, as the caller can now
3073 set them directly.
3074
3075
3076 Version 0.92 11-Sep-97
3077 ----------------------
3078
3079 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3080 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3081
3082 2. Caseless matching was not working in character classes if the characters in
3083 the pattern were in upper case.
3084
3085 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3086
3087 4. Make PCRE_ANCHORED public and accept as a compile option.
3088
3089 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3090 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3091 pass them.
3092
3093 6. Give an error if bad option bits passed at compile or run time.
3094
3095 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3096 pcretest to cause it to pass that flag.
3097
3098 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3099 options, and the first character, if set.
3100
3101 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3102
3103
3104 Version 0.91 10-Sep-97
3105 ----------------------
3106
3107 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3108 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3109
3110 2. PCRE was looping on encountering an indefinitely repeated back reference to
3111 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3112 Perl does - treats the match as successful.
3113
3114 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12