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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 289 - (show annotations) (download)
Sun Dec 23 12:17:20 2007 UTC (7 years ago) by ph10
File size: 146438 byte(s)
Tidies for 7.5-RC1 release.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.5 23-Dec-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. Added --file-offsets and --line-offsets to pcregrep.
80
81 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
82 infinitely looping recursion. The bug was that positive lookaheads were not
83 being skipped when checking for a possible empty match (negative lookaheads
84 and both kinds of lookbehind were skipped).
85
86 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
87 inclusion of <windows.h> to before rather than after the definition of
88 INVALID_FILE_ATTRIBUTES (patch from David Byron).
89
90 17. Specifying a possessive quantifier with a specific limit for a Unicode
91 character property caused pcre_compile() to compile bad code, which led at
92 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
93 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
94 caused the error; without that there was no problem.
95
96 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
97
98 19. Added --enable-pcretest-libreadline.
99
100 20. In pcrecpp.cc, the variable 'count' was incremented twice in
101 RE::GlobalReplace(). As a result, the number of replacements returned was
102 double what it should be. I have removed one of the increments.
103
104
105 Version 7.4 21-Sep-07
106 ---------------------
107
108 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
109 means that a class such as [\s] counted as "explicit reference to CR or
110 LF". That isn't really right - the whole point of the change was to try to
111 help when there was an actual mention of one of the two characters. So now
112 the change happens only if \r or \n (or a literal CR or LF) character is
113 encountered.
114
115 2. The 32-bit options word was also used for 6 internal flags, but the numbers
116 of both had grown to the point where there were only 3 bits left.
117 Fortunately, there was spare space in the data structure, and so I have
118 moved the internal flags into a new 16-bit field to free up more option
119 bits.
120
121 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
122 but did not set the internal JCHANGED flag - either of these is enough to
123 control the way the "get" function works - but the PCRE_INFO_JCHANGED
124 facility is supposed to tell if (?J) was ever used, so now (?J) at the
125 start sets both bits.
126
127 4. Added options (at build time, compile time, exec time) to change \R from
128 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
129
130 5. doc/pcresyntax.html was missing from the distribution.
131
132 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
133 compatibility, even though it is no longer used.
134
135 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
136 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
137 windows.h file is present (where different names are used). [This was
138 reversed later after testing - see 16 below.]
139
140 8. Changed all #include <config.h> to #include "config.h". There were also
141 some further <pcre.h> cases that I changed to "pcre.h".
142
143 9. When pcregrep was used with the --colour option, it missed the line ending
144 sequence off the lines that it output.
145
146 10. It was pointed out to me that arrays of string pointers cause lots of
147 relocations when a shared library is dynamically loaded. A technique of
148 using a single long string with a table of offsets can drastically reduce
149 these. I have refactored PCRE in four places to do this. The result is
150 dramatic:
151
152 Originally: 290
153 After changing UCP table: 187
154 After changing error message table: 43
155 After changing table of "verbs" 36
156 After changing table of Posix names 22
157
158 Thanks to the folks working on Gregex for glib for this insight.
159
160 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
161 unicode-properties was also set.
162
163 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
164
165 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
166 checked only for CRLF.
167
168 14. Added casts to pcretest.c to avoid compiler warnings.
169
170 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
171
172 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
173 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
174 entirely. This removes changes made in 7 above.
175
176 17. The CMake files have been updated, and there is now more information about
177 building with CMake in the NON-UNIX-USE document.
178
179
180 Version 7.3 28-Aug-07
181 ---------------------
182
183 1. In the rejigging of the build system that eventually resulted in 7.1, the
184 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
185 brackets there is not right, since it causes compilers to look for an
186 installed pcre.h, not the version that is in the source that is being
187 compiled (which of course may be different). I have changed it back to:
188
189 #include "pcre.h"
190
191 I have a vague recollection that the change was concerned with compiling in
192 different directories, but in the new build system, that is taken care of
193 by the VPATH setting the Makefile.
194
195 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
196 when the subject happened to end in the byte 0x85 (e.g. if the last
197 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
198 characters but of course it shouldn't be taken as a newline when it is part
199 of another character. The bug was that, for an unlimited repeat of . in
200 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
201 characters when looking for a newline.
202
203 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
204
205 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
206 in debug output.
207
208 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
209 long printing in the pcrecpp unittest when running under MinGW.
210
211 6. ESC_K was left out of the EBCDIC table.
212
213 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
214 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
215 limit also applies to "virtual nesting" when a pattern is recursive, and in
216 this case 1000 isn't so big. I have been able to remove this limit at the
217 expense of backing off one optimization in certain circumstances. Normally,
218 when pcre_exec() would call its internal match() function recursively and
219 immediately return the result unconditionally, it uses a "tail recursion"
220 feature to save stack. However, when a subpattern that can match an empty
221 string has an unlimited repetition quantifier, it no longer makes this
222 optimization. That gives it a stack frame in which to save the data for
223 checking that an empty string has been matched. Previously this was taken
224 from the 1000-entry workspace that had been reserved. So now there is no
225 explicit limit, but more stack is used.
226
227 8. Applied Daniel's patches to solve problems with the import/export magic
228 syntax that is required for Windows, and which was going wrong for the
229 pcreposix and pcrecpp parts of the library. These were overlooked when this
230 problem was solved for the main library.
231
232 9. There were some crude static tests to avoid integer overflow when computing
233 the size of patterns that contain repeated groups with explicit upper
234 limits. As the maximum quantifier is 65535, the maximum group length was
235 set at 30,000 so that the product of these two numbers did not overflow a
236 32-bit integer. However, it turns out that people want to use groups that
237 are longer than 30,000 bytes (though not repeat them that many times).
238 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
239 made it possible to implement the integer overflow checks in a much more
240 dynamic way, which I have now done. The artificial limitation on group
241 length has been removed - we now have only the limit on the total length of
242 the compiled pattern, which depends on the LINK_SIZE setting.
243
244 10. Fixed a bug in the documentation for get/copy named substring when
245 duplicate names are permitted. If none of the named substrings are set, the
246 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
247 empty string.
248
249 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
250 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
251 because the ] is interpreted as the first data character and the
252 terminating ] is not found. PCRE has been made compatible with Perl in this
253 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
254 cause memory overwriting.
255
256 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
257 string has been matched (to stop an infinite loop). It was not recognizing
258 a conditional subpattern that could match an empty string if that
259 subpattern was within another subpattern. For example, it looped when
260 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
261 condition was not nested. This bug has been fixed.
262
263 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
264 past the start of the subject in the presence of bytes with the top bit
265 set, for example "\x8aBCD".
266
267 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
268 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
269
270 14. Optimized (?!) to (*FAIL).
271
272 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
273 This restricts code points to be within the range 0 to 0x10FFFF, excluding
274 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
275 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
276 does: it's just the validity check that is more restrictive.
277
278 16. Inserted checks for integer overflows during escape sequence (backslash)
279 processing, and also fixed erroneous offset values for syntax errors during
280 backslash processing.
281
282 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
283 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
284
285 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
286 caused an overrun.
287
288 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
289 something other than just ASCII characters) inside a group that had an
290 unlimited repeat caused a loop at compile time (while checking to see
291 whether the group could match an empty string).
292
293 20. Debugging a pattern containing \p or \P could cause a crash. For example,
294 [\P{Any}] did so. (Error in the code for printing property names.)
295
296 21. An orphan \E inside a character class could cause a crash.
297
298 22. A repeated capturing bracket such as (A)? could cause a wild memory
299 reference during compilation.
300
301 23. There are several functions in pcre_compile() that scan along a compiled
302 expression for various reasons (e.g. to see if it's fixed length for look
303 behind). There were bugs in these functions when a repeated \p or \P was
304 present in the pattern. These operators have additional parameters compared
305 with \d, etc, and these were not being taken into account when moving along
306 the compiled data. Specifically:
307
308 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
309 length.
310
311 (b) An item such as \pL+ within a repeated group could cause crashes or
312 loops.
313
314 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
315 "reference to non-existent subpattern" error.
316
317 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
318
319 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
320 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
321
322 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
323
324 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
325 character were causing crashes (broken optimization).
326
327 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
328 \p or \P) caused a compile-time loop.
329
330 28. More problems have arisen in unanchored patterns when CRLF is a valid line
331 break. For example, the unstudied pattern [\r\n]A does not match the string
332 "\r\nA" because change 7.0/46 below moves the current point on by two
333 characters after failing to match at the start. However, the pattern \nA
334 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
335 the same is true. There doesn't seem any very clean way out of this, but
336 what I have chosen to do makes the common cases work: PCRE now takes note
337 of whether there can be an explicit match for \r or \n anywhere in the
338 pattern, and if so, 7.0/46 no longer applies. As part of this change,
339 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
340 pattern has explicit CR or LF references.
341
342 29. Added (*CR) etc for changing newline setting at start of pattern.
343
344
345 Version 7.2 19-Jun-07
346 ---------------------
347
348 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
349 which is apparently normally available under Windows.
350
351 2. Re-jig the pcregrep tests with different newline settings in an attempt
352 to make them independent of the local environment's newline setting.
353
354 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
355
356 4. Some of the "internals" tests were previously cut out when the link size
357 was not 2, because the output contained actual offsets. The recent new
358 "Z" feature of pcretest means that these can be cut out, making the tests
359 usable with all link sizes.
360
361 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
362 stack recursion. This gives a massive performance boost under BSD, but just
363 a small improvement under Linux. However, it saves one field in the frame
364 in all cases.
365
366 6. Added more features from the forthcoming Perl 5.10:
367
368 (a) (?-n) (where n is a string of digits) is a relative subroutine or
369 recursion call. It refers to the nth most recently opened parentheses.
370
371 (b) (?+n) is also a relative subroutine call; it refers to the nth next
372 to be opened parentheses.
373
374 (c) Conditions that refer to capturing parentheses can be specified
375 relatively, for example, (?(-2)... or (?(+3)...
376
377 (d) \K resets the start of the current match so that everything before
378 is not part of it.
379
380 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
381
382 (f) \g{name} is another synonym - part of Perl 5.10's unification of
383 reference syntax.
384
385 (g) (?| introduces a group in which the numbering of parentheses in each
386 alternative starts with the same number.
387
388 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
389
390 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
391 PCRE_INFO_JCHANGED.
392
393 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
394 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
395 for detecting groups that can match an empty string.
396
397 9. A pattern with a very large number of alternatives (more than several
398 hundred) was running out of internal workspace during the pre-compile
399 phase, where pcre_compile() figures out how much memory will be needed. A
400 bit of new cunning has reduced the workspace needed for groups with
401 alternatives. The 1000-alternative test pattern now uses 12 bytes of
402 workspace instead of running out of the 4096 that are available.
403
404 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
405
406 11. Applied patch from Google to remove an optimization that didn't quite work.
407 The report of the bug said:
408
409 pcrecpp::RE("a*").FullMatch("aaa") matches, while
410 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
411 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
412
413 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
414 it matched the wrong number of bytes.
415
416
417 Version 7.1 24-Apr-07
418 ---------------------
419
420 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
421 that is more "standard", making use of automake and other Autotools. There
422 is some re-arrangement of the files and adjustment of comments consequent
423 on this.
424
425 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
426 for recursive directory scanning broke on some systems because the files
427 are not scanned in any specific order and on different systems the order
428 was different. A call to "sort" has been inserted into RunGrepTest for the
429 approprate test as a short-term fix. In the longer term there may be an
430 alternative.
431
432 3. I had an email from Eric Raymond about problems translating some of PCRE's
433 man pages to HTML (despite the fact that I distribute HTML pages, some
434 people do their own conversions for various reasons). The problems
435 concerned the use of low-level troff macros .br and .in. I have therefore
436 removed all such uses from the man pages (some were redundant, some could
437 be replaced by .nf/.fi pairs). The 132html script that I use to generate
438 HTML has been updated to handle .nf/.fi and to complain if it encounters
439 .br or .in.
440
441 4. Updated comments in configure.ac that get placed in config.h.in and also
442 arranged for config.h to be included in the distribution, with the name
443 config.h.generic, for the benefit of those who have to compile without
444 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
445
446 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
447 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
448 makevp.bat for the current PCRE, using the additional files
449 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
450
451 6. A Windows user reported a minor discrepancy with test 2, which turned out
452 to be caused by a trailing space on an input line that had got lost in his
453 copy. The trailing space was an accident, so I've just removed it.
454
455 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
456 that is needed.
457
458 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
459 as "const" (a) because they are and (b) because it helps the PHP
460 maintainers who have recently made a script to detect big data structures
461 in the php code that should be moved to the .rodata section. I remembered
462 to update Builducptable as well, so it won't revert if ucptable.h is ever
463 re-created.
464
465 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
466 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
467 order to be able to cut out the UTF-8 tables in the latter when UTF-8
468 support is not required. This saves 1.5-2K of code, which is important in
469 some applications.
470
471 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
472 so as not to refer to the tables, even though these functions will never be
473 called when UTF-8 support is disabled. Otherwise there are problems with a
474 shared library.
475
476 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
477
478 (a) It was defining its arguments as char * instead of void *.
479
480 (b) It was assuming that all moves were upwards in memory; this was true
481 a long time ago when I wrote it, but is no longer the case.
482
483 The emulated memove() is provided for those environments that have neither
484 memmove() nor bcopy(). I didn't think anyone used it these days, but that
485 is clearly not the case, as these two bugs were recently reported.
486
487 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
488 and Detrail to create the HTML documentation, the .txt form of the man
489 pages, and it removes trailing spaces from listed files. It also creates
490 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
491 case, it wraps all the #defines with #ifndefs. This script should be run
492 before "make dist".
493
494 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
495 with Unicode property support.
496
497 (a) For a maximizing quantifier, if the two different cases of the
498 character were of different lengths in their UTF-8 codings (there are
499 some cases like this - I found 11), and the matching function had to
500 back up over a mixture of the two cases, it incorrectly assumed they
501 were both the same length.
502
503 (b) When PCRE was configured to use the heap rather than the stack for
504 recursion during matching, it was not correctly preserving the data for
505 the other case of a UTF-8 character when checking ahead for a match
506 while processing a minimizing repeat. If the check also involved
507 matching a wide character, but failed, corruption could cause an
508 erroneous result when trying to check for a repeat of the original
509 character.
510
511 13. Some tidying changes to the testing mechanism:
512
513 (a) The RunTest script now detects the internal link size and whether there
514 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
515 values substituted by "configure". (The RunGrepTest script already did
516 this for UTF-8.) The configure.ac script no longer substitutes the
517 relevant variables.
518
519 (b) The debugging options /B and /D in pcretest show the compiled bytecode
520 with length and offset values. This means that the output is different
521 for different internal link sizes. Test 2 is skipped for link sizes
522 other than 2 because of this, bypassing the problem. Unfortunately,
523 there was also a test in test 3 (the locale tests) that used /B and
524 failed for link sizes other than 2. Rather than cut the whole test out,
525 I have added a new /Z option to pcretest that replaces the length and
526 offset values with spaces. This is now used to make test 3 independent
527 of link size. (Test 2 will be tidied up later.)
528
529 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
530 segmentation fault instead of returning the appropriate error message.
531
532 15. In multiline mode when the newline sequence was set to "any", the pattern
533 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
534 This doesn't seem right; it now treats the CRLF combination as the line
535 ending, and so does not match in that case. It's only a pattern such as ^$
536 that would hit this one: something like ^ABC$ would have failed after \r
537 and then tried again after \r\n.
538
539 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
540 in an attempt to make files that differ only in their line terminators
541 compare equal. This works on Linux.
542
543 17. Under certain error circumstances pcregrep might try to free random memory
544 as it exited. This is now fixed, thanks to valgrind.
545
546 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
547 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
548 was because its rules for how to advance for /g after matching an empty
549 string at the end of a line did not allow for this case. They now check for
550 it specially.
551
552 20. pcretest is supposed to handle patterns and data of any length, by
553 extending its buffers when necessary. It was getting this wrong when the
554 buffer for a data line had to be extended.
555
556 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
557 CRLF as a newline sequence.
558
559 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
560 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
561 I have nevertheless tidied it up.
562
563 23. Added some casts to kill warnings from HP-UX ia64 compiler.
564
565 24. Added a man page for pcre-config.
566
567
568 Version 7.0 19-Dec-06
569 ---------------------
570
571 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
572 moving to gcc 4.1.1.
573
574 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
575 sys/time.h, which is documented as needed for this function. It doesn't
576 seem to matter on Linux, but it showed up on some releases of OS X.
577
578 3. It seems that there are systems where bytes whose values are greater than
579 127 match isprint() in the "C" locale. The "C" locale should be the
580 default when a C program starts up. In most systems, only ASCII printing
581 characters match isprint(). This difference caused the output from pcretest
582 to vary, making some of the tests fail. I have changed pcretest so that:
583
584 (a) When it is outputting text in the compiled version of a pattern, bytes
585 other than 32-126 are always shown as hex escapes.
586
587 (b) When it is outputting text that is a matched part of a subject string,
588 it does the same, unless a different locale has been set for the match
589 (using the /L modifier). In this case, it uses isprint() to decide.
590
591 4. Fixed a major bug that caused incorrect computation of the amount of memory
592 required for a compiled pattern when options that changed within the
593 pattern affected the logic of the preliminary scan that determines the
594 length. The relevant options are -x, and -i in UTF-8 mode. The result was
595 that the computed length was too small. The symptoms of this bug were
596 either the PCRE error "internal error: code overflow" from pcre_compile(),
597 or a glibc crash with a message such as "pcretest: free(): invalid next
598 size (fast)". Examples of patterns that provoked this bug (shown in
599 pcretest format) are:
600
601 /(?-x: )/x
602 /(?x)(?-x: \s*#\s*)/
603 /((?i)[\x{c0}])/8
604 /(?i:[\x{c0}])/8
605
606 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
607 is now done differently.
608
609 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
610 wrapper classes; (b) implement a new function in the C++ scanner that is
611 more efficient than the old way of doing things because it avoids levels of
612 recursion in the regex matching; (c) add a paragraph to the documentation
613 for the FullMatch() function.
614
615 6. The escape sequence \n was being treated as whatever was defined as
616 "newline". Not only was this contrary to the documentation, which states
617 that \n is character 10 (hex 0A), but it also went horribly wrong when
618 "newline" was defined as CRLF. This has been fixed.
619
620 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
621 was being set to -1 for the "end of line" case (supposedly a value that no
622 character can have). Though this value is never used (the check for end of
623 line is "zero bytes in current character"), it caused compiler complaints.
624 I've changed it to 0xffffffff.
625
626 8. In pcre_version.c, the version string was being built by a sequence of
627 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
628 string (as it is for production releases) called a macro with an empty
629 argument. The C standard says the result of this is undefined. The gcc
630 compiler treats it as an empty string (which was what was wanted) but it is
631 reported that Visual C gives an error. The source has been hacked around to
632 avoid this problem.
633
634 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
635 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
636 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
637 of them did).
638
639 10. Originally, pcretest opened its input and output without "b"; then I was
640 told that "b" was needed in some environments, so it was added for release
641 5.0 to both the input and output. (It makes no difference on Unix-like
642 systems.) Later I was told that it is wrong for the input on Windows. I've
643 now abstracted the modes into two macros, to make it easier to fiddle with
644 them, and removed "b" from the input mode under Windows.
645
646 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
647
648 12. Added -help and --help to pcretest as an official way of being reminded
649 of the options.
650
651 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
652 and pcrecpp.cc because they annoy compilers at high warning levels.
653
654 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
655
656 15. Fixed an occurrence of == in configure.ac that should have been = (shell
657 scripts are not C programs :-) and which was not noticed because it works
658 on Linux.
659
660 16. pcretest is supposed to handle any length of pattern and data line (as one
661 line or as a continued sequence of lines) by extending its input buffer if
662 necessary. This feature was broken for very long pattern lines, leading to
663 a string of junk being passed to pcre_compile() if the pattern was longer
664 than about 50K.
665
666 17. I have done a major re-factoring of the way pcre_compile() computes the
667 amount of memory needed for a compiled pattern. Previously, there was code
668 that made a preliminary scan of the pattern in order to do this. That was
669 OK when PCRE was new, but as the facilities have expanded, it has become
670 harder and harder to keep it in step with the real compile phase, and there
671 have been a number of bugs (see for example, 4 above). I have now found a
672 cunning way of running the real compile function in a "fake" mode that
673 enables it to compute how much memory it would need, while actually only
674 ever using a few hundred bytes of working memory and without too many
675 tests of the mode. This should make future maintenance and development
676 easier. A side effect of this work is that the limit of 200 on the nesting
677 depth of parentheses has been removed (though this was never a serious
678 limitation, I suspect). However, there is a downside: pcre_compile() now
679 runs more slowly than before (30% or more, depending on the pattern). I
680 hope this isn't a big issue. There is no effect on runtime performance.
681
682 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
683 newline (only possible for the last line of a file) and it was a
684 pattern that set a locale (followed by /Lsomething), pcretest crashed.
685
686 19. Added additional timing features to pcretest. (1) The -tm option now times
687 matching only, not compiling. (2) Both -t and -tm can be followed, as a
688 separate command line item, by a number that specifies the number of
689 repeats to use when timing. The default is 50000; this gives better
690 precision, but takes uncomfortably long for very large patterns.
691
692 20. Extended pcre_study() to be more clever in cases where a branch of a
693 subpattern has no definite first character. For example, (a*|b*)[cd] would
694 previously give no result from pcre_study(). Now it recognizes that the
695 first character must be a, b, c, or d.
696
697 21. There was an incorrect error "recursive call could loop indefinitely" if
698 a subpattern (or the entire pattern) that was being tested for matching an
699 empty string contained only one non-empty item after a nested subpattern.
700 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
701 incorrectly, because the \d was being skipped in the check.
702
703 22. The pcretest program now has a new pattern option /B and a command line
704 option -b, which is equivalent to adding /B to every pattern. This causes
705 it to show the compiled bytecode, without the additional information that
706 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
707 is the same as /B/I).
708
709 23. A new optimization is now able automatically to treat some sequences such
710 as a*b as a*+b. More specifically, if something simple (such as a character
711 or a simple class like \d) has an unlimited quantifier, and is followed by
712 something that cannot possibly match the quantified thing, the quantifier
713 is automatically "possessified".
714
715 24. A recursive reference to a subpattern whose number was greater than 39
716 went wrong under certain circumstances in UTF-8 mode. This bug could also
717 have affected the operation of pcre_study().
718
719 25. Realized that a little bit of performance could be had by replacing
720 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
721
722 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
723
724 27. Possessive quantifiers such as a++ were previously implemented by turning
725 them into atomic groups such as ($>a+). Now they have their own opcodes,
726 which improves performance. This includes the automatically created ones
727 from 23 above.
728
729 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
730 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
731 the first matched character to be a colon. This applied both to named and
732 numbered groups.
733
734 29. The ucpinternal.h header file was missing its idempotency #ifdef.
735
736 30. I was sent a "project" file called libpcre.a.dev which I understand makes
737 building PCRE on Windows easier, so I have included it in the distribution.
738
739 31. There is now a check in pcretest against a ridiculously large number being
740 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
741 loop, the loop is abandoned.
742
743 32. Forward references to subpatterns in conditions such as (?(2)...) where
744 subpattern 2 is defined later cause pcre_compile() to search forwards in
745 the pattern for the relevant set of parentheses. This search went wrong
746 when there were unescaped parentheses in a character class, parentheses
747 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
748
749 33. "Subroutine" calls and backreferences were previously restricted to
750 referencing subpatterns earlier in the regex. This restriction has now
751 been removed.
752
753 34. Added a number of extra features that are going to be in Perl 5.10. On the
754 whole, these are just syntactic alternatives for features that PCRE had
755 previously implemented using the Python syntax or my own invention. The
756 other formats are all retained for compatibility.
757
758 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
759 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
760 also .NET compatible.
761
762 (b) A recursion or subroutine call to a named group can now be defined as
763 (?&name) as well as (?P>name).
764
765 (c) A backreference to a named group can now be defined as \k<name> or
766 \k'name' as well as (?P=name). The new forms, as well as being in Perl
767 5.10, are also .NET compatible.
768
769 (d) A conditional reference to a named group can now use the syntax
770 (?(<name>) or (?('name') as well as (?(name).
771
772 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
773 groups (named and numbered) that are never evaluated inline, but can be
774 called as "subroutines" from elsewhere. In effect, the DEFINE condition
775 is always false. There may be only one alternative in such a group.
776
777 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
778 as the simple (?(R). The condition is true only if the most recent
779 recursion is that of the given number or name. It does not search out
780 through the entire recursion stack.
781
782 (g) The escape \gN or \g{N} has been added, where N is a positive or
783 negative number, specifying an absolute or relative reference.
784
785 35. Tidied to get rid of some further signed/unsigned compiler warnings and
786 some "unreachable code" warnings.
787
788 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
789 things, this adds five new scripts.
790
791 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
792 There were also incompatibilities regarding the handling of \Q..\E inside
793 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
794 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
795
796 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
797 matches an empty string, and forcibly breaks the loop. There were bugs in
798 this code in non-simple cases. For a pattern such as ^(a()*)* matched
799 against aaaa the result was just "a" rather than "aaaa", for example. Two
800 separate and independent bugs (that affected different cases) have been
801 fixed.
802
803 39. Refactored the code to abolish the use of different opcodes for small
804 capturing bracket numbers. This is a tidy that I avoided doing when I
805 removed the limit on the number of capturing brackets for 3.5 back in 2001.
806 The new approach is not only tidier, it makes it possible to reduce the
807 memory needed to fix the previous bug (38).
808
809 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
810 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
811 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
812 mode.
813
814 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
815 report.
816
817 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
818 copy construction and assignment in the C++ wrapper.
819
820 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
821 couple of bugs that could have given wrong results in the "--newline=crlf"
822 case.
823
824 44. Added a number of casts and did some reorganization of signed/unsigned int
825 variables following suggestions from Dair Grant. Also renamed the variable
826 "this" as "item" because it is a C++ keyword.
827
828 45. Arranged for dftables to add
829
830 #include "pcre_internal.h"
831
832 to pcre_chartables.c because without it, gcc 4.x may remove the array
833 definition from the final binary if PCRE is built into a static library and
834 dead code stripping is activated.
835
836 46. For an unanchored pattern, if a match attempt fails at the start of a
837 newline sequence, and the newline setting is CRLF or ANY, and the next two
838 characters are CRLF, advance by two characters instead of one.
839
840
841 Version 6.7 04-Jul-06
842 ---------------------
843
844 1. In order to handle tests when input lines are enormously long, pcretest has
845 been re-factored so that it automatically extends its buffers when
846 necessary. The code is crude, but this _is_ just a test program. The
847 default size has been increased from 32K to 50K.
848
849 2. The code in pcre_study() was using the value of the re argument before
850 testing it for NULL. (Of course, in any sensible call of the function, it
851 won't be NULL.)
852
853 3. The memmove() emulation function in pcre_internal.h, which is used on
854 systems that lack both memmove() and bcopy() - that is, hardly ever -
855 was missing a "static" storage class specifier.
856
857 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
858 containing an extended class (one that cannot be represented by a bitmap
859 because it contains high-valued characters or Unicode property items, e.g.
860 [\pZ]). Almost always one would set UTF-8 mode when processing such a
861 pattern, but PCRE should not loop if you do not (it no longer does).
862 [Detail: two cases were found: (a) a repeated subpattern containing an
863 extended class; (b) a recursive reference to a subpattern that followed a
864 previous extended class. It wasn't skipping over the extended class
865 correctly when UTF-8 mode was not set.]
866
867 5. A negated single-character class was not being recognized as fixed-length
868 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
869 compile error "lookbehind assertion is not fixed length".
870
871 6. The RunPerlTest auxiliary script was showing an unexpected difference
872 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
873 write a Perl script that can interpret lines of an input file either as
874 byte characters or as UTF-8, which is what "perltest" was being required to
875 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
876 can't do is switch easily at run time between having the "use utf8;" pragma
877 or not. In the end, I fudged it by using the RunPerlTest script to insert
878 "use utf8;" explicitly for the UTF-8 tests.
879
880 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
881 the end of the subject string, contrary to the documentation and to what
882 Perl does. This was true of both matching functions. Now it matches only at
883 the start of the subject and immediately after *internal* newlines.
884
885 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
886 a pointer to an int instead of a pointer to an unsigned long int. This
887 caused problems on 64-bit systems.
888
889 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
890 instance of the 'standard' template library not being so standard".
891
892 10. There was no check on the number of named subpatterns nor the maximum
893 length of a subpattern name. The product of these values is used to compute
894 the size of the memory block for a compiled pattern. By supplying a very
895 long subpattern name and a large number of named subpatterns, the size
896 computation could be caused to overflow. This is now prevented by limiting
897 the length of names to 32 characters, and the number of named subpatterns
898 to 10,000.
899
900 11. Subpatterns that are repeated with specific counts have to be replicated in
901 the compiled pattern. The size of memory for this was computed from the
902 length of the subpattern and the repeat count. The latter is limited to
903 65535, but there was no limit on the former, meaning that integer overflow
904 could in principle occur. The compiled length of a repeated subpattern is
905 now limited to 30,000 bytes in order to prevent this.
906
907 12. Added the optional facility to have named substrings with the same name.
908
909 13. Added the ability to use a named substring as a condition, using the
910 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
911 are numbers (not recommended). Forward references are permitted.
912
913 14. Added forward references in named backreferences (if you see what I mean).
914
915 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
916 pattern could run off the end of the subject. For example, the pattern
917 "(?s)(.{1,5})"8 did this with the subject "ab".
918
919 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
920 PCRE_CASELESS was set when matching characters that were quantified with ?
921 or *.
922
923 17. A character class other than a single negated character that had a minimum
924 but no maximum quantifier - for example [ab]{6,} - was not handled
925 correctly by pce_dfa_exec(). It would match only one character.
926
927 18. A valid (though odd) pattern that looked like a POSIX character
928 class but used an invalid character after [ (for example [[,abc,]]) caused
929 pcre_compile() to give the error "Failed: internal error: code overflow" or
930 in some cases to crash with a glibc free() error. This could even happen if
931 the pattern terminated after [[ but there just happened to be a sequence of
932 letters, a binary zero, and a closing ] in the memory that followed.
933
934 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
935 over the years. Originally (before any Unicode support), just the bottom 8
936 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
937 output from "man perlunicode" includes this:
938
939 The regular expression compiler produces polymorphic opcodes. That
940 is, the pattern adapts to the data and automatically switches to
941 the Unicode character scheme when presented with Unicode data--or
942 instead uses a traditional byte scheme when presented with byte
943 data.
944
945 Sadly, a wide octal escape does not cause a switch, and in a string with
946 no other multibyte characters, these octal escapes are treated as before.
947 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
948 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
949 Unicode string.
950
951 I have not perpetrated such confusion in PCRE. Up till now, it took just
952 the bottom 8 bits, as in old Perl. I have now made octal escapes with
953 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
954 translate to the appropriate multibyte character.
955
956 29. Applied some refactoring to reduce the number of warnings from Microsoft
957 and Borland compilers. This has included removing the fudge introduced
958 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
959 a warning about an unused variable.
960
961 21. PCRE has not included VT (character 0x0b) in the set of whitespace
962 characters since release 4.0, because Perl (from release 5.004) does not.
963 [Or at least, is documented not to: some releases seem to be in conflict
964 with the documentation.] However, when a pattern was studied with
965 pcre_study() and all its branches started with \s, PCRE still included VT
966 as a possible starting character. Of course, this did no harm; it just
967 caused an unnecessary match attempt.
968
969 22. Removed a now-redundant internal flag bit that recorded the fact that case
970 dependency changed within the pattern. This was once needed for "required
971 byte" processing, but is no longer used. This recovers a now-scarce options
972 bit. Also moved the least significant internal flag bit to the most-
973 significant bit of the word, which was not previously used (hangover from
974 the days when it was an int rather than a uint) to free up another bit for
975 the future.
976
977 23. Added support for CRLF line endings as well as CR and LF. As well as the
978 default being selectable at build time, it can now be changed at runtime
979 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
980 specify that it is scanning data with non-default line endings.
981
982 24. Changed the definition of CXXLINK to make it agree with the definition of
983 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
984
985 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
986 recursions. This makes a big different to stack usage for some patterns.
987
988 26. If a subpattern containing a named recursion or subroutine reference such
989 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
990 the space required for the compiled pattern went wrong and gave too small a
991 value. Depending on the environment, this could lead to "Failed: internal
992 error: code overflow at offset 49" or "glibc detected double free or
993 corruption" errors.
994
995 27. Applied patches from Google (a) to support the new newline modes and (b) to
996 advance over multibyte UTF-8 characters in GlobalReplace.
997
998 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
999 difference for some implementation of PCRE in some Windows version.
1000
1001 29. Added some extra testing facilities to pcretest:
1002
1003 \q<number> in a data line sets the "match limit" value
1004 \Q<number> in a data line sets the "match recursion limt" value
1005 -S <number> sets the stack size, where <number> is in megabytes
1006
1007 The -S option isn't available for Windows.
1008
1009
1010 Version 6.6 06-Feb-06
1011 ---------------------
1012
1013 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1014 in pcreposix.h. I have copied the definition from pcre.h.
1015
1016 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1017 because pcre.h is no longer a built file.
1018
1019 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1020 not normally included in the compiled code.
1021
1022
1023 Version 6.5 01-Feb-06
1024 ---------------------
1025
1026 1. When using the partial match feature with pcre_dfa_exec(), it was not
1027 anchoring the second and subsequent partial matches at the new starting
1028 point. This could lead to incorrect results. For example, with the pattern
1029 /1234/, partially matching against "123" and then "a4" gave a match.
1030
1031 2. Changes to pcregrep:
1032
1033 (a) All non-match returns from pcre_exec() were being treated as failures
1034 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1035 error message is output. Some extra information is given for the
1036 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1037 probably the only errors that are likely to be caused by users (by
1038 specifying a regex that has nested indefinite repeats, for instance).
1039 If there are more than 20 of these errors, pcregrep is abandoned.
1040
1041 (b) A binary zero was treated as data while matching, but terminated the
1042 output line if it was written out. This has been fixed: binary zeroes
1043 are now no different to any other data bytes.
1044
1045 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1046 used to set a locale for matching. The --locale=xxxx long option has
1047 been added (no short equivalent) to specify a locale explicitly on the
1048 pcregrep command, overriding the environment variables.
1049
1050 (d) When -B was used with -n, some line numbers in the output were one less
1051 than they should have been.
1052
1053 (e) Added the -o (--only-matching) option.
1054
1055 (f) If -A or -C was used with -c (count only), some lines of context were
1056 accidentally printed for the final match.
1057
1058 (g) Added the -H (--with-filename) option.
1059
1060 (h) The combination of options -rh failed to suppress file names for files
1061 that were found from directory arguments.
1062
1063 (i) Added the -D (--devices) and -d (--directories) options.
1064
1065 (j) Added the -F (--fixed-strings) option.
1066
1067 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1068
1069 (l) Added the --colo(u)r option.
1070
1071 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1072 is not present by default.
1073
1074 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1075 items such as (?R) or (?1), when the recursion could match a number of
1076 alternatives. If it matched one of the alternatives, but subsequently,
1077 outside the recursion, there was a failure, the code tried to back up into
1078 the recursion. However, because of the way PCRE is implemented, this is not
1079 possible, and the result was an incorrect result from the match.
1080
1081 In order to prevent this happening, the specification of recursion has
1082 been changed so that all such subpatterns are automatically treated as
1083 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1084
1085 4. I had overlooked the fact that, in some locales, there are characters for
1086 which isalpha() is true but neither isupper() nor islower() are true. In
1087 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1088 and ordfeminine) are like this. This affected the treatment of \w and \W
1089 when they appeared in character classes, but not when they appeared outside
1090 a character class. The bit map for "word" characters is now created
1091 separately from the results of isalnum() instead of just taking it from the
1092 upper, lower, and digit maps. (Plus the underscore character, of course.)
1093
1094 5. The above bug also affected the handling of POSIX character classes such as
1095 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1096 permanent tables. Instead, the bit maps for such a class were previously
1097 created as the appropriate unions of the upper, lower, and digit bitmaps.
1098 Now they are created by subtraction from the [[:word:]] class, which has
1099 its own bitmap.
1100
1101 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1102 It is created by subtracting the vertical space characters (\x09, \x0a,
1103 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1104 subtraction was done in the overall bitmap for a character class, meaning
1105 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1106 be recognized. This bug has been fixed.
1107
1108 7. Patches from the folks at Google:
1109
1110 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1111 real life, but is still worth protecting against".
1112
1113 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1114 regular expressions".
1115
1116 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1117 have it.
1118
1119 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1120 "configure" and the latter not, in order to fix a problem somebody had
1121 with compiling the Arg class on HP-UX.
1122
1123 (e) Improve the error-handling of the C++ wrapper a little bit.
1124
1125 (f) New tests for checking recursion limiting.
1126
1127 8. The pcre_memmove() function, which is used only if the environment does not
1128 have a standard memmove() function (and is therefore rarely compiled),
1129 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1130 returning a result (though PCRE never actually uses the result).
1131
1132 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1133 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1134 returned instead of calling malloc() with an overflowing number that would
1135 most likely cause subsequent chaos.
1136
1137 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1138
1139 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1140 with this option is matched, the nmatch and pmatch options of regexec() are
1141 ignored.
1142
1143 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1144 provided in case anyone wants to the the POSIX interface with UTF-8
1145 strings.
1146
1147 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1148 C++ linking (needed for some HP-UX environments).
1149
1150 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1151 (unused parameter) and in the pcre_printint() function (omitted "default"
1152 switch label when the default is to do nothing).
1153
1154 15. Added some code to make it possible, when PCRE is compiled as a C++
1155 library, to replace subject pointers for pcre_exec() with a smart pointer
1156 class, thus making it possible to process discontinuous strings.
1157
1158 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1159 much the same function. They were added by different people who were trying
1160 to make PCRE easy to compile on non-Unix systems. It has been suggested
1161 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1162 for compiling on Windows systems. I have therefore replaced it with
1163 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1164 defaults to "extern" for C or "extern C" for C++, which works fine on
1165 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1166 SCOPE with something explicit in config.h. In addition:
1167
1168 (a) pcreposix.h still had just "extern" instead of either of these macros;
1169 I have replaced it with PCRE_DATA_SCOPE.
1170
1171 (b) Functions such as _pcre_xclass(), which are internal to the library,
1172 but external in the C sense, all had PCRE_EXPORT in their definitions.
1173 This is apparently wrong for the Windows case, so I have removed it.
1174 (It makes no difference on Unix-like systems.)
1175
1176 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1177 of recursive calls to match(). This is different to MATCH_LIMIT because
1178 that limits the total number of calls to match(), not all of which increase
1179 the depth of recursion. Limiting the recursion depth limits the amount of
1180 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1181 when PCRE is compiled, and changed at run time. A patch from Google adds
1182 this functionality to the C++ interface.
1183
1184 18. Changes to the handling of Unicode character properties:
1185
1186 (a) Updated the table to Unicode 4.1.0.
1187
1188 (b) Recognize characters that are not in the table as "Cn" (undefined).
1189
1190 (c) I revised the way the table is implemented to a much improved format
1191 which includes recognition of ranges. It now supports the ranges that
1192 are defined in UnicodeData.txt, and it also amalgamates other
1193 characters into ranges. This has reduced the number of entries in the
1194 table from around 16,000 to around 3,000, thus reducing its size
1195 considerably. I realized I did not need to use a tree structure after
1196 all - a binary chop search is just as efficient. Having reduced the
1197 number of entries, I extended their size from 6 bytes to 8 bytes to
1198 allow for more data.
1199
1200 (d) Added support for Unicode script names via properties such as \p{Han}.
1201
1202 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1203 matching that character.
1204
1205 20. When matching a repeated Unicode property with a minimum greater than zero,
1206 (for example \pL{2,}), PCRE could look past the end of the subject if it
1207 reached it while seeking the minimum number of characters. This could
1208 happen only if some of the characters were more than one byte long, because
1209 there is a check for at least the minimum number of bytes.
1210
1211 21. Refactored the implementation of \p and \P so as to be more general, to
1212 allow for more different types of property in future. This has changed the
1213 compiled form incompatibly. Anybody with saved compiled patterns that use
1214 \p or \P will have to recompile them.
1215
1216 22. Added "Any" and "L&" to the supported property types.
1217
1218 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1219 but give a compile time error if the value is greater than 0xff.
1220
1221 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1222 accidentally not being installed or uninstalled.
1223
1224 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1225 made were to insert the current release number. This seemed silly, because
1226 it made things harder for people building PCRE on systems that don't run
1227 "configure". I have turned pcre.h into a distributed file, no longer built
1228 by "configure", with the version identification directly included. There is
1229 no longer a pcre.h.in file.
1230
1231 However, this change necessitated a change to the pcre-config script as
1232 well. It is built from pcre-config.in, and one of the substitutions was the
1233 release number. I have updated configure.ac so that ./configure now finds
1234 the release number by grepping pcre.h.
1235
1236 26. Added the ability to run the tests under valgrind.
1237
1238
1239 Version 6.4 05-Sep-05
1240 ---------------------
1241
1242 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1243 "--" to be printed when multiple files were scanned, even when none of the
1244 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1245 consider it to be a bug, and have restored the previous behaviour.
1246
1247 2. A couple of code tidies to get rid of compiler warnings.
1248
1249 3. The pcretest program used to cheat by referring to symbols in the library
1250 whose names begin with _pcre_. These are internal symbols that are not
1251 really supposed to be visible externally, and in some environments it is
1252 possible to suppress them. The cheating is now confined to including
1253 certain files from the library's source, which is a bit cleaner.
1254
1255 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1256 file's purpose clearer.
1257
1258 5. Reorganized pcre_ucp_findchar().
1259
1260
1261 Version 6.3 15-Aug-05
1262 ---------------------
1263
1264 1. The file libpcre.pc.in did not have general read permission in the tarball.
1265
1266 2. There were some problems when building without C++ support:
1267
1268 (a) If C++ support was not built, "make install" and "make test" still
1269 tried to test it.
1270
1271 (b) There were problems when the value of CXX was explicitly set. Some
1272 changes have been made to try to fix these, and ...
1273
1274 (c) --disable-cpp can now be used to explicitly disable C++ support.
1275
1276 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1277 backslash in a target when C++ was disabled. This confuses some
1278 versions of "make", apparently. Using an intermediate variable solves
1279 this. (Same for CPP_LOBJ.)
1280
1281 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1282 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1283 necessary on certain architectures.
1284
1285 4. Added a setting of -export-symbols-regex to the link command to remove
1286 those symbols that are exported in the C sense, but actually are local
1287 within the library, and not documented. Their names all begin with
1288 "_pcre_". This is not a perfect job, because (a) we have to except some
1289 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1290 available (and never for static libraries). I have made a note to try to
1291 find a way round (a) in the future.
1292
1293
1294 Version 6.2 01-Aug-05
1295 ---------------------
1296
1297 1. There was no test for integer overflow of quantifier values. A construction
1298 such as {1111111111111111} would give undefined results. What is worse, if
1299 a minimum quantifier for a parenthesized subpattern overflowed and became
1300 negative, the calculation of the memory size went wrong. This could have
1301 led to memory overwriting.
1302
1303 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1304
1305 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1306 operating environments where this matters.
1307
1308 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1309 PCRE options from within the C++ wrapper.
1310
1311 5. Named capturing subpatterns were not being correctly counted when a pattern
1312 was compiled. This caused two problems: (a) If there were more than 100
1313 such subpatterns, the calculation of the memory needed for the whole
1314 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1315 back references of the form \12, where the number was greater than 9, were
1316 not recognized as back references, even though there were sufficient
1317 previous subpatterns.
1318
1319 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1320 versions of gcc, e.g. 2.95.4.
1321
1322
1323 Version 6.1 21-Jun-05
1324 ---------------------
1325
1326 1. There was one reference to the variable "posix" in pcretest.c that was not
1327 surrounded by "#if !defined NOPOSIX".
1328
1329 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1330 the cross-check on the old pcre_info() function, for the benefit of the
1331 cut-down version of PCRE that is currently imported into Exim.
1332
1333 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1334 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1335 patterns that start (?i) but I don't think that matters. The use of (?i) is
1336 just an example; this all applies to the other options as well.
1337
1338 4. Since libtool seems to echo the compile commands it is issuing, the output
1339 from "make" can be reduced a bit by putting "@" in front of each libtool
1340 compile command.
1341
1342 5. Patch from the folks at Google for configure.in to be a bit more thorough
1343 in checking for a suitable C++ installation before trying to compile the
1344 C++ stuff. This should fix a reported problem when a compiler was present,
1345 but no suitable headers.
1346
1347 6. The man pages all had just "PCRE" as their title. I have changed them to
1348 be the relevant file name. I have also arranged that these names are
1349 retained in the file doc/pcre.txt, which is a concatenation in text format
1350 of all the man pages except the little individual ones for each function.
1351
1352 7. The NON-UNIX-USE file had not been updated for the different set of source
1353 files that come with release 6. I also added a few comments about the C++
1354 wrapper.
1355
1356
1357 Version 6.0 07-Jun-05
1358 ---------------------
1359
1360 1. Some minor internal re-organization to help with my DFA experiments.
1361
1362 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1363 didn't matter for the library itself when fully configured, but did matter
1364 when compiling without UCP support, or within Exim, where the ucp files are
1365 not imported.
1366
1367 3. Refactoring of the library code to split up the various functions into
1368 different source modules. The addition of the new DFA matching code (see
1369 below) to a single monolithic source would have made it really too
1370 unwieldy, quite apart from causing all the code to be include in a
1371 statically linked application, when only some functions are used. This is
1372 relevant even without the DFA addition now that patterns can be compiled in
1373 one application and matched in another.
1374
1375 The downside of splitting up is that there have to be some external
1376 functions and data tables that are used internally in different modules of
1377 the library but which are not part of the API. These have all had their
1378 names changed to start with "_pcre_" so that they are unlikely to clash
1379 with other external names.
1380
1381 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1382 a different (DFA) algorithm. Although it is slower than the original
1383 function, it does have some advantages for certain types of matching
1384 problem.
1385
1386 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1387 including restarting after a partial match.
1388
1389 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1390 defined when compiling for Windows was sent to me. I have put it into the
1391 code, though I have no means of testing or verifying it.
1392
1393 7. Added the pcre_refcount() auxiliary function.
1394
1395 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1396 match before or at the first newline in the subject string. In pcretest,
1397 the /f option on a pattern can be used to set this.
1398
1399 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1400 would behave wrongly. This has been present in PCRE since release 4.0.
1401
1402 10. A number of changes to the pcregrep command:
1403
1404 (a) Refactored how -x works; insert ^(...)$ instead of setting
1405 PCRE_ANCHORED and checking the length, in preparation for adding
1406 something similar for -w.
1407
1408 (b) Added the -w (match as a word) option.
1409
1410 (c) Refactored the way lines are read and buffered so as to have more
1411 than one at a time available.
1412
1413 (d) Implemented a pcregrep test script.
1414
1415 (e) Added the -M (multiline match) option. This allows patterns to match
1416 over several lines of the subject. The buffering ensures that at least
1417 8K, or the rest of the document (whichever is the shorter) is available
1418 for matching (and similarly the previous 8K for lookbehind assertions).
1419
1420 (f) Changed the --help output so that it now says
1421
1422 -w, --word-regex(p)
1423
1424 instead of two lines, one with "regex" and the other with "regexp"
1425 because that confused at least one person since the short forms are the
1426 same. (This required a bit of code, as the output is generated
1427 automatically from a table. It wasn't just a text change.)
1428
1429 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1430 option but starts with a hyphen. Could be a pattern or a path name
1431 starting with a hyphen, for instance.
1432
1433 (h) "-" can be given as a file name to represent stdin.
1434
1435 (i) When file names are being printed, "(standard input)" is used for
1436 the standard input, for compatibility with GNU grep. Previously
1437 "<stdin>" was used.
1438
1439 (j) The option --label=xxx can be used to supply a name to be used for
1440 stdin when file names are being printed. There is no short form.
1441
1442 (k) Re-factored the options decoding logic because we are going to add
1443 two more options that take data. Such options can now be given in four
1444 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1445
1446 (l) Added the -A, -B, and -C options for requesting that lines of context
1447 around matches be printed.
1448
1449 (m) Added the -L option to print the names of files that do not contain
1450 any matching lines, that is, the complement of -l.
1451
1452 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1453 continue to scan other files.
1454
1455 (o) The -s option was incorrectly implemented. For compatibility with other
1456 greps, it now suppresses the error message for a non-existent or non-
1457 accessible file (but not the return code). There is a new option called
1458 -q that suppresses the output of matching lines, which was what -s was
1459 previously doing.
1460
1461 (p) Added --include and --exclude options to specify files for inclusion
1462 and exclusion when recursing.
1463
1464 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1465 Hopefully, it now does.
1466
1467 12. Missing cast in pcre_study().
1468
1469 13. Added an "uninstall" target to the makefile.
1470
1471 14. Replaced "extern" in the function prototypes in Makefile.in with
1472 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1473 world, but is set differently for Windows.
1474
1475 15. Added a second compiling function called pcre_compile2(). The only
1476 difference is that it has an extra argument, which is a pointer to an
1477 integer error code. When there is a compile-time failure, this is set
1478 non-zero, in addition to the error test pointer being set to point to an
1479 error message. The new argument may be NULL if no error number is required
1480 (but then you may as well call pcre_compile(), which is now just a
1481 wrapper). This facility is provided because some applications need a
1482 numeric error indication, but it has also enabled me to tidy up the way
1483 compile-time errors are handled in the POSIX wrapper.
1484
1485 16. Added VPATH=.libs to the makefile; this should help when building with one
1486 prefix path and installing with another. (Or so I'm told by someone who
1487 knows more about this stuff than I do.)
1488
1489 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1490 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1491 match everything, including newlines. This is not POSIX-compatible, but
1492 somebody wanted the feature. From pcretest it can be activated by using
1493 both the P and the s flags.
1494
1495 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1496
1497 19. libpcre.pc was being incorrectly installed as executable.
1498
1499 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1500 it now also looks for '\r' so that it will work unmodified on Windows.
1501
1502 21. Added Google's contributed C++ wrapper to the distribution.
1503
1504 22. Added some untidy missing memory free() calls in pcretest, to keep
1505 Electric Fence happy when testing.
1506
1507
1508
1509 Version 5.0 13-Sep-04
1510 ---------------------
1511
1512 1. Internal change: literal characters are no longer packed up into items
1513 containing multiple characters in a single byte-string. Each character
1514 is now matched using a separate opcode. However, there may be more than one
1515 byte in the character in UTF-8 mode.
1516
1517 2. The pcre_callout_block structure has two new fields: pattern_position and
1518 next_item_length. These contain the offset in the pattern to the next match
1519 item, and its length, respectively.
1520
1521 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1522 insertion of callouts before each pattern item. Added the /C option to
1523 pcretest to make use of this.
1524
1525 4. On the advice of a Windows user, the lines
1526
1527 #if defined(_WIN32) || defined(WIN32)
1528 _setmode( _fileno( stdout ), 0x8000 );
1529 #endif /* defined(_WIN32) || defined(WIN32) */
1530
1531 have been added to the source of pcretest. This apparently does useful
1532 magic in relation to line terminators.
1533
1534 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1535 for the benefit of those environments where the "b" makes a difference.
1536
1537 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1538 to know about it. I have put a hack into configure.in that adds in code
1539 to set GCC=yes if CC=icc. This seems to end up at a point in the
1540 generated configure script that is early enough to affect the setting of
1541 compiler options, which is what is needed, but I have no means of testing
1542 whether it really works. (The user who reported this had patched the
1543 generated configure script, which of course I cannot do.)
1544
1545 LATER: After change 22 below (new libtool files), the configure script
1546 seems to know about icc (and also ecc). Therefore, I have commented out
1547 this hack in configure.in.
1548
1549 7. Added support for pkg-config (2 patches were sent in).
1550
1551 8. Negated POSIX character classes that used a combination of internal tables
1552 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1553 [[:^ascii]]. Typically, they would match almost any characters. The other
1554 POSIX classes were not broken in this way.
1555
1556 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1557 to find the match, as PCRE was deluded into thinking that the match had to
1558 start at the start point or following a newline. The same bug applied to
1559 patterns with negative forward assertions or any backward assertions
1560 preceding ".*" at the start, unless the pattern required a fixed first
1561 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1562
1563 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1564 starting at the last subject character, bytes beyond the end of the subject
1565 string were read.
1566
1567 11. Renamed the variable "class" as "classbits" to make life easier for C++
1568 users. (Previously there was a macro definition, but it apparently wasn't
1569 enough.)
1570
1571 12. Added the new field "tables" to the extra data so that tables can be passed
1572 in at exec time, or the internal tables can be re-selected. This allows
1573 a compiled regex to be saved and re-used at a later time by a different
1574 program that might have everything at different addresses.
1575
1576 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1577 -R library as well as a -L library.
1578
1579 14. The debugging options of pcretest (-d on the command line or D on a
1580 pattern) showed incorrect output for anything following an extended class
1581 that contained multibyte characters and which was followed by a quantifier.
1582
1583 15. Added optional support for general category Unicode character properties
1584 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1585 support. It adds about 90K to the size of the library. The meanings of the
1586 inbuilt class escapes such as \d and \s have NOT been changed.
1587
1588 16. Updated pcredemo.c to include calls to free() to release the memory for the
1589 compiled pattern.
1590
1591 17. The generated file chartables.c was being created in the source directory
1592 instead of in the building directory. This caused the build to fail if the
1593 source directory was different from the building directory, and was
1594 read-only.
1595
1596 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1597 file. No doubt somebody will tell me if they don't make sense... Also added
1598 Dan Mooney's comments about building on OpenVMS.
1599
1600 19. Added support for partial matching via the PCRE_PARTIAL option for
1601 pcre_exec() and the \P data escape in pcretest.
1602
1603 20. Extended pcretest with 3 new pattern features:
1604
1605 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1606 write the compiled pattern to the file whose name is "rest-of-line".
1607 This is a straight binary dump of the data, with the saved pointer to
1608 the character tables forced to be NULL. The study data, if any, is
1609 written too. After writing, pcretest reads a new pattern.
1610
1611 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1612 compiled pattern from the given file. There must not be any
1613 occurrences of "<" in the file name (pretty unlikely); if there are,
1614 pcretest will instead treat the initial "<" as a pattern delimiter.
1615 After reading in the pattern, pcretest goes on to read data lines as
1616 usual.
1617
1618 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1619 and 16-bit fields in a compiled pattern, to simulate a pattern that
1620 was compiled on a host of opposite endianness.
1621
1622 21. The pcre-exec() function can now cope with patterns that were compiled on
1623 hosts of opposite endianness, with this restriction:
1624
1625 As for any compiled expression that is saved and used later, the tables
1626 pointer field cannot be preserved; the extra_data field in the arguments
1627 to pcre_exec() should be used to pass in a tables address if a value
1628 other than the default internal tables were used at compile time.
1629
1630 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1631 now diagnosed as an error. Previously, most of the time, a negative number
1632 would have been treated as zero, but if in addition "ovector" was passed as
1633 NULL, a crash could occur.
1634
1635 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1636 new versions from the libtool 1.5 distribution (the last one is a copy of
1637 a file called libtool.m4). This seems to have fixed the need to patch
1638 "configure" to support Darwin 1.3 (which I used to do). However, I still
1639 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1640 workstation).
1641
1642 24. Changed the PCRE licence to be the more standard "BSD" licence.
1643
1644
1645 Version 4.5 01-Dec-03
1646 ---------------------
1647
1648 1. There has been some re-arrangement of the code for the match() function so
1649 that it can be compiled in a version that does not call itself recursively.
1650 Instead, it keeps those local variables that need separate instances for
1651 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1652 needs to "recurse". Keeping track of where control must go is done by means
1653 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1654 hide most of the details from the main code, and operates only if
1655 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1656 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1657 operating.
1658
1659 To make it easier for callers to provide specially tailored get/free
1660 functions for this usage, two new functions, pcre_stack_malloc, and
1661 pcre_stack_free, are used. They are always called in strict stacking order,
1662 and the size of block requested is always the same.
1663
1664 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1665 PCRE has been compiled to use the stack or the heap for recursion. The
1666 -C option of pcretest uses this to show which version is compiled.
1667
1668 A new data escape \S, is added to pcretest; it causes the amounts of store
1669 obtained and freed by both kinds of malloc/free at match time to be added
1670 to the output.
1671
1672 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1673 what's available on my current Linux desktop machine.
1674
1675 3. When matching a UTF-8 string, the test for a valid string at the start has
1676 been extended. If start_offset is not zero, PCRE now checks that it points
1677 to a byte that is the start of a UTF-8 character. If not, it returns
1678 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1679 this is necessary because there may be backward assertions in the pattern.
1680 When matching the same subject several times, it may save resources to use
1681 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1682
1683 4. The code for checking the validity of UTF-8 strings has been tightened so
1684 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1685 containing "overlong sequences".
1686
1687 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1688 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1689 should have been "&", but it just so happened that all the cases this let
1690 through by mistake were picked up later in the function.
1691
1692 6. I had used a variable called "isblank" - this is a C99 function, causing
1693 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1694
1695 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1696 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1697 so the version is shown; (c) stop "make test" echoing "./RunTest".
1698
1699 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1700
1701 9. The prototype for memmove() for systems that don't have it was using
1702 size_t, but the inclusion of the header that defines size_t was later. I've
1703 moved the #includes for the C headers earlier to avoid this.
1704
1705 10. Added some adjustments to the code to make it easier to compiler on certain
1706 special systems:
1707
1708 (a) Some "const" qualifiers were missing.
1709 (b) Added the macro EXPORT before all exported functions; by default this
1710 is defined to be empty.
1711 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1712 that it reads its output file name as an argument instead of writing
1713 to the standard output and assuming this can be redirected.
1714
1715 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1716 class containing characters with values greater than 255, PCRE compilation
1717 went into a loop.
1718
1719 12. A recursive reference to a subpattern that was within another subpattern
1720 that had a minimum quantifier of zero caused PCRE to crash. For example,
1721 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1722 recursion. If the recursively-called subpattern itself had a zero repeat,
1723 that was OK.
1724
1725 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1726 buffer into which it was copied (for escape processing) was still set at
1727 1024, so long lines caused crashes.
1728
1729 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1730 "internal error: code overflow...". This applied to any character class
1731 that was followed by a possessive quantifier.
1732
1733 15. Modified the Makefile to add libpcre.la as a prerequisite for
1734 libpcreposix.la because I was told this is needed for a parallel build to
1735 work.
1736
1737 16. If a pattern that contained .* following optional items at the start was
1738 studied, the wrong optimizing data was generated, leading to matching
1739 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1740 matching string must start with a or b or c. The correct conclusion for
1741 this pattern is that a match can start with any character.
1742
1743
1744 Version 4.4 13-Aug-03
1745 ---------------------
1746
1747 1. In UTF-8 mode, a character class containing characters with values between
1748 127 and 255 was not handled correctly if the compiled pattern was studied.
1749 In fixing this, I have also improved the studying algorithm for such
1750 classes (slightly).
1751
1752 2. Three internal functions had redundant arguments passed to them. Removal
1753 might give a very teeny performance improvement.
1754
1755 3. Documentation bug: the value of the capture_top field in a callout is *one
1756 more than* the number of the hightest numbered captured substring.
1757
1758 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1759 in incorrectly linking with a previously installed version. They now link
1760 explicitly with libpcre.la.
1761
1762 5. configure.in no longer needs to recognize Cygwin specially.
1763
1764 6. A problem in pcre.in for Windows platforms is fixed.
1765
1766 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1767 pcretest, it used to include the size of the study block as part of its
1768 output. Unfortunately, the structure contains a field that has a different
1769 size on different hardware architectures. This meant that the tests that
1770 showed this size failed. As the block is currently always of a fixed size,
1771 this information isn't actually particularly useful in pcretest output, so
1772 I have just removed it.
1773
1774 8. Three pre-processor statements accidentally did not start in column 1.
1775 Sadly, there are *still* compilers around that complain, even though
1776 standard C has not required this for well over a decade. Sigh.
1777
1778 9. In pcretest, the code for checking callouts passed small integers in the
1779 callout_data field, which is a void * field. However, some picky compilers
1780 complained about the casts involved for this on 64-bit systems. Now
1781 pcretest passes the address of the small integer instead, which should get
1782 rid of the warnings.
1783
1784 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1785 both compile and run time, and gives an error if an invalid UTF-8 sequence
1786 is found. There is a option for disabling this check in cases where the
1787 string is known to be correct and/or the maximum performance is wanted.
1788
1789 11. In response to a bug report, I changed one line in Makefile.in from
1790
1791 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1792 to
1793 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1794
1795 to look similar to other lines, but I have no way of telling whether this
1796 is the right thing to do, as I do not use Windows. No doubt I'll get told
1797 if it's wrong...
1798
1799
1800 Version 4.3 21-May-03
1801 ---------------------
1802
1803 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1804 Makefile.
1805
1806 2. Some refactoring to improve the quality of the code:
1807
1808 (i) The utf8_table... variables are now declared "const".
1809
1810 (ii) The code for \cx, which used the "case flipping" table to upper case
1811 lower case letters, now just substracts 32. This is ASCII-specific,
1812 but the whole concept of \cx is ASCII-specific, so it seems
1813 reasonable.
1814
1815 (iii) PCRE was using its character types table to recognize decimal and
1816 hexadecimal digits in the pattern. This is silly, because it handles
1817 only 0-9, a-f, and A-F, but the character types table is locale-
1818 specific, which means strange things might happen. A private
1819 table is now used for this - though it costs 256 bytes, a table is
1820 much faster than multiple explicit tests. Of course, the standard
1821 character types table is still used for matching digits in subject
1822 strings against \d.
1823
1824 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1825 ending in _t are). So I've renamed it as ESC_tee.
1826
1827 3. The first argument for regexec() in the POSIX wrapper should have been
1828 defined as "const".
1829
1830 4. Changed pcretest to use malloc() for its buffers so that they can be
1831 Electric Fenced for debugging.
1832
1833 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1834 to read one or more bytes before the start of the subject string. Often this
1835 had no effect on PCRE's behaviour, but in some circumstances it could
1836 provoke a segmentation fault.
1837
1838 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1839 to try to read one or more bytes before the start of the subject string.
1840
1841 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1842 UTF-8 support could misbehave in various ways if the subject string
1843 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1844 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1845 back over UTF-8 characters.)
1846
1847
1848 Version 4.2 14-Apr-03
1849 ---------------------
1850
1851 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1852
1853 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1854 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1855 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1856 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1857 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1858 and BUILD_EXEEXT
1859 Note: automatic setting of the BUILD variables is not yet working
1860 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1861 compile-time but not at link-time
1862 [LINK]: use for linking executables only
1863 make different versions for Windows and non-Windows
1864 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1865 libraries
1866 [LINK_FOR_BUILD]: new variable
1867 [OBJEXT]: use throughout
1868 [EXEEXT]: use throughout
1869 <winshared>: new target
1870 <wininstall>: new target
1871 <dftables.o>: use native compiler
1872 <dftables>: use native linker
1873 <install>: handle Windows platform correctly
1874 <clean>: ditto
1875 <check>: ditto
1876 copy DLL to top builddir before testing
1877
1878 As part of these changes, -no-undefined was removed again. This was reported
1879 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1880 in any case.
1881
1882 3. Some tidies to get rid of compiler warnings:
1883
1884 . In the match_data structure, match_limit was an unsigned long int, whereas
1885 match_call_count was an int. I've made them both unsigned long ints.
1886
1887 . In pcretest the fact that a const uschar * doesn't automatically cast to
1888 a void * provoked a warning.
1889
1890 . Turning on some more compiler warnings threw up some "shadow" variables
1891 and a few more missing casts.
1892
1893 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1894 option, a class that contained a single character with a value between 128
1895 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1896
1897 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1898 option, a class that contained several characters, but with at least one
1899 whose value was between 128 and 255 caused PCRE to crash.
1900
1901
1902 Version 4.1 12-Mar-03
1903 ---------------------
1904
1905 1. Compiling with gcc -pedantic found a couple of places where casts were
1906 needed, and a string in dftables.c that was longer than standard compilers are
1907 required to support.
1908
1909 2. Compiling with Sun's compiler found a few more places where the code could
1910 be tidied up in order to avoid warnings.
1911
1912 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1913 first of these names is deprecated in the latest Autoconf in favour of the name
1914 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1915 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1916 analogy I have changed it to CFLAGS_FOR_BUILD.
1917
1918 4. Added -no-undefined to the linking command in the Makefile, because this is
1919 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1920 linking step for the pcreposix library.
1921
1922 5. PCRE was failing to diagnose the case of two named groups with the same
1923 name.
1924
1925 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1926 literal character that is needed in the subject for a match, and scans along to
1927 ensure that it is present before embarking on the full matching process. This
1928 saves time in cases of nested unlimited repeats that are never going to match.
1929 Problem: the scan can take a lot of time if the subject is very long (e.g.
1930 megabytes), thus penalizing straightforward matches. It is now done only if the
1931 amount of subject to be scanned is less than 1000 bytes.
1932
1933 7. A lesser problem with the same optimization is that it was recording the
1934 first character of an anchored pattern as "needed", thus provoking a search
1935 right along the subject, even when the first match of the pattern was going to
1936 fail. The "needed" character is now not set for anchored patterns, unless it
1937 follows something in the pattern that is of non-fixed length. Thus, it still
1938 fulfils its original purpose of finding quick non-matches in cases of nested
1939 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1940
1941
1942 Version 4.0 17-Feb-03
1943 ---------------------
1944
1945 1. If a comment in an extended regex that started immediately after a meta-item
1946 extended to the end of string, PCRE compiled incorrect data. This could lead to
1947 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
1948
1949 2. Moved to autoconf 2.53 and libtool 1.4.2.
1950
1951 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
1952 the special perltest8 script is no longer needed - all the tests can be run
1953 from a single perltest script.
1954
1955 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
1956 by \s. It has now been removed in PCRE. This means it isn't recognized as
1957 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
1958 class [:space:] *does* include VT, thereby creating a mess.
1959
1960 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
1961 space and tab.
1962
1963 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
1964 its new features into the main test script, reducing the number of scripts.
1965
1966 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
1967 were backward compatible, and made the (?i) apply to the whole pattern, as if
1968 /i were given. Now it behaves more logically, and applies the option setting
1969 only to what follows. PCRE has been changed to follow suit. However, if it
1970 finds options settings right at the start of the pattern, it extracts them into
1971 the global options, as before. Thus, they show up in the info data.
1972
1973 8. Added support for the \Q...\E escape sequence. Characters in between are
1974 treated as literals. This is slightly different from Perl in that $ and @ are
1975 also handled as literals inside the quotes. In Perl, they will cause variable
1976 interpolation. Note the following examples:
1977
1978 Pattern PCRE matches Perl matches
1979
1980 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
1981 \Qabc\$xyz\E abc\$xyz abc\$xyz
1982 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
1983
1984 For compatibility with Perl, \Q...\E sequences are recognized inside character
1985 classes as well as outside them.
1986
1987 9. Re-organized 3 code statements in pcretest to avoid "overflow in
1988 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
1989 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
1990 signed/unsigned warnings.
1991
1992 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
1993 option for pcretest, so I've replaced it by a simple function that does just
1994 that job.
1995
1996 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
1997 "pcregrep -".
1998
1999 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2000 Java package. This provides some syntactic sugar for simple cases of what my
2001 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2002 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2003 item, you can use this simplified notation. Note that only makes sense with
2004 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2005 greediness, whatever the setting of the PCRE_UNGREEDY option.
2006
2007 13. A change of greediness default within a pattern was not taking effect at
2008 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2009 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2010 was abstracted outside.
2011
2012 14. PCRE now supports the \G assertion. It is true when the current matching
2013 position is at the start point of the match. This differs from \A when the
2014 starting offset is non-zero. Used with the /g option of pcretest (or similar
2015 code), it works in the same way as it does for Perl's /g option. If all
2016 alternatives of a regex begin with \G, the expression is anchored to the start
2017 match position, and the "anchored" flag is set in the compiled expression.
2018
2019 15. Some bugs concerning the handling of certain option changes within patterns
2020 have been fixed. These applied to options other than (?ims). For example,
2021 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2022 been the other way round. Some of this was related to change 7 above.
2023
2024 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2025 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2026 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2027 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2028
2029 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2030 mode. Unlike ".", it always matches newline, whatever the setting of
2031 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2032 assertions. Perl allows it, but it doesn't (in general) work because it can't
2033 calculate the length of the lookbehind. At least, that's the case for Perl
2034 5.8.0 - I've been told they are going to document that it doesn't work in
2035 future.
2036
2037 18. Added an error diagnosis for escapes that PCRE does not support: these are
2038 \L, \l, \N, \P, \p, \U, \u, and \X.
2039
2040 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2041 reading past the end of the pattern in cases such as /[abcd/.
2042
2043 20. PCRE was getting more memory than necessary for patterns with classes that
2044 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2045
2046 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2047 compiling PCRE for use with Virtual Pascal.
2048
2049 22. Small fix to the Makefile to make it work properly if the build is done
2050 outside the source tree.
2051
2052 23. Added a new extension: a condition to go with recursion. If a conditional
2053 subpattern starts with (?(R) the "true" branch is used if recursion has
2054 happened, whereas the "false" branch is used only at the top level.
2055
2056 24. When there was a very long string of literal characters (over 255 bytes
2057 without UTF support, over 250 bytes with UTF support), the computation of how
2058 much memory was required could be incorrect, leading to segfaults or other
2059 strange effects.
2060
2061 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2062 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2063 there was a subsequent back reference to those brackets. This meant that, for
2064 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2065 possible to check for precisely this case. All we can do is abandon the
2066 optimization if .* occurs inside capturing brackets when there are any back
2067 references whatsoever. (See below for a better fix that came later.)
2068
2069 26. The handling of the optimization for finding the first character of a
2070 non-anchored pattern, and for finding a character that is required later in the
2071 match were failing in some cases. This didn't break the matching; it just
2072 failed to optimize when it could. The way this is done has been re-implemented.
2073
2074 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2075
2076 28. Added a new feature that provides some of the functionality that Perl
2077 provides with (?{...}). The facility is termed a "callout". The way it is done
2078 in PCRE is for the caller to provide an optional function, by setting
2079 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2080 global variable. By default it is unset, which disables all calling out. To get
2081 the function called, the regex must include (?C) at appropriate points. This
2082 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2083 This provides a means of identifying different callout points. When PCRE
2084 reaches such a point in the regex, if pcre_callout has been set, the external
2085 function is called. It is provided with data in a structure called
2086 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2087 matching continues; if it returns a non-zero value, the match at the current
2088 point fails. However, backtracking will occur if possible. [This was changed
2089 later and other features added - see item 49 below.]
2090
2091 29. pcretest is upgraded to test the callout functionality. It provides a
2092 callout function that displays information. By default, it shows the start of
2093 the match and the current position in the text. There are some new data escapes
2094 to vary what happens:
2095
2096 \C+ in addition, show current contents of captured substrings
2097 \C- do not supply a callout function
2098 \C!n return 1 when callout number n is reached
2099 \C!n!m return 1 when callout number n is reached for the mth time
2100
2101 30. If pcregrep was called with the -l option and just a single file name, it
2102 output "<stdin>" if a match was found, instead of the file name.
2103
2104 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2105 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2106 pcre_exec(). This saves a malloc/free per call. The default value of
2107 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2108 when configuring.
2109
2110 32. The default maximum size of a compiled pattern is 64K. There have been a
2111 few cases of people hitting this limit. The code now uses macros to handle the
2112 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2113 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2114 configuring. Tests 2 and 5 work only with 2-byte links because they output
2115 debugging information about compiled patterns.
2116
2117 33. Internal code re-arrangements:
2118
2119 (a) Moved the debugging function for printing out a compiled regex into
2120 its own source file (printint.c) and used #include to pull it into
2121 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2122 separate copies.
2123
2124 (b) Defined the list of op-code names for debugging as a macro in
2125 internal.h so that it is next to the definition of the opcodes.
2126
2127 (c) Defined a table of op-code lengths for simpler skipping along compiled
2128 code. This is again a macro in internal.h so that it is next to the
2129 definition of the opcodes.
2130
2131 34. Added support for recursive calls to individual subpatterns, along the
2132 lines of Robin Houston's patch (but implemented somewhat differently).
2133
2134 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2135 allow it to read and process whole directories in Win32. This code was
2136 contributed by Lionel Fourquaux; it has not been tested by me.
2137
2138 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2139 used to name a group. Names consist of alphanumerics and underscores, and must
2140 be unique. Back references use the syntax (?P=name) and recursive calls use
2141 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2142 numbers. The function pcre_fullinfo() can be used after compilation to extract
2143 a name/number map. There are three relevant calls:
2144
2145 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2146 PCRE_INFO_NAMECOUNT yields the number of entries
2147 PCRE_INFO_NAMETABLE yields a pointer to the map.
2148
2149 The map is a vector of fixed-size entries. The size of each entry depends on
2150 the length of the longest name used. The first two bytes of each entry are the
2151 group number, most significant byte first. There follows the corresponding
2152 name, zero terminated. The names are in alphabetical order.
2153
2154 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2155 case instead of 255. Making it the same both with and without UTF-8 support
2156 means that the same test output works with both.
2157
2158 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2159 calling malloc() with a zero argument.
2160
2161 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2162 optimization. I've improved things by keeping a bitmap of backreferences with
2163 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2164 fact referenced, the optimization can be applied. It is unlikely that a
2165 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2166 the match to follow \n) will appear inside brackets with a number greater than
2167 31, but if it does, any back reference > 31 suppresses the optimization.
2168
2169 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2170 of disabling numbered capturing parentheses. Any opening parenthesis that is
2171 not followed by ? behaves as if it were followed by ?: but named parentheses
2172 can still be used for capturing (and they will acquire numbers in the usual
2173 way).
2174
2175 41. Redesigned the return codes from the match() function into yes/no/error so
2176 that errors can be passed back from deep inside the nested calls. A malloc
2177 failure while inside a recursive subpattern call now causes the
2178 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2179
2180 42. It is now possible to set a limit on the number of times the match()
2181 function is called in a call to pcre_exec(). This facility makes it possible to
2182 limit the amount of recursion and backtracking, though not in a directly
2183 obvious way, because the match() function is used in a number of different
2184 circumstances. The count starts from zero for each position in the subject
2185 string (for non-anchored patterns). The default limit is, for compatibility, a
2186 large number, namely 10 000 000. You can change this in two ways:
2187
2188 (a) When configuring PCRE before making, you can use --with-match-limit=n
2189 to set a default value for the compiled library.
2190
2191 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2192 a different value is set. See 45 below.
2193
2194 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2195
2196 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2197 of things that can be changed at compile time. The first argument specifies
2198 what is wanted and the second points to where the information is to be placed.
2199 The current list of available information is:
2200
2201 PCRE_CONFIG_UTF8
2202
2203 The output is an integer that is set to one if UTF-8 support is available;
2204 otherwise it is set to zero.
2205
2206 PCRE_CONFIG_NEWLINE
2207
2208 The output is an integer that it set to the value of the code that is used for
2209 newline. It is either LF (10) or CR (13).
2210
2211 PCRE_CONFIG_LINK_SIZE
2212
2213 The output is an integer that contains the number of bytes used for internal
2214 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2215
2216 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2217
2218 The output is an integer that contains the threshold above which the POSIX
2219 interface uses malloc() for output vectors. See item 31 above.
2220
2221 PCRE_CONFIG_MATCH_LIMIT
2222
2223 The output is an unsigned integer that contains the default limit of the number
2224 of match() calls in a pcre_exec() execution. See 42 above.
2225
2226 44. pcretest has been upgraded by the addition of the -C option. This causes it
2227 to extract all the available output from the new pcre_config() function, and to
2228 output it. The program then exits immediately.
2229
2230 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2231 order to support additional features. One way would have been to define
2232 pcre_exec2() (for example) with extra arguments, but this would not have been
2233 extensible, and would also have required all calls to the original function to
2234 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2235 is used for passing in "extra" data from pcre_study().
2236
2237 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2238 contains the following fields:
2239
2240 flags a bitmap indicating which of the following fields are set
2241 study_data opaque data from pcre_study()
2242 match_limit a way of specifying a limit on match() calls for a specific
2243 call to pcre_exec()
2244 callout_data data for callouts (see 49 below)
2245
2246 The flag bits are also defined in pcre.h, and are
2247
2248 PCRE_EXTRA_STUDY_DATA
2249 PCRE_EXTRA_MATCH_LIMIT
2250 PCRE_EXTRA_CALLOUT_DATA
2251
2252 The pcre_study() function now returns one of these new pcre_extra blocks, with
2253 the actual study data pointed to by the study_data field, and the
2254 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2255 before. That is, this change is entirely upwards-compatible and requires no
2256 change to existing code.
2257
2258 If you want to pass in additional data to pcre_exec(), you can either place it
2259 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2260 block.
2261
2262 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2263 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2264 times with different match limits, until it finds the minimum value needed for
2265 pcre_exec() to complete. The value is then output. This can be instructive; for
2266 most simple matches the number is quite small, but for pathological cases it
2267 gets very large very quickly.
2268
2269 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2270 returns the size of the data block pointed to by the study_data field in a
2271 pcre_extra block, that is, the value that was passed as the argument to
2272 pcre_malloc() when PCRE was getting memory in which to place the information
2273 created by pcre_study(). The fourth argument should point to a size_t variable.
2274 pcretest has been extended so that this information is shown after a successful
2275 pcre_study() call when information about the compiled regex is being displayed.
2276
2277 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2278 because what follows is always an absolute path. (Later: it turns out that this
2279 is more than cosmetic for MinGW, because it doesn't like empty path
2280 components.)
2281
2282 49. Some changes have been made to the callout feature (see 28 above):
2283
2284 (i) A callout function now has three choices for what it returns:
2285
2286 0 => success, carry on matching
2287 > 0 => failure at this point, but backtrack if possible
2288 < 0 => serious error, return this value from pcre_exec()
2289
2290 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2291 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2292 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2293 use by callout functions. It will never be used by PCRE itself.
2294
2295 (ii) The pcre_extra structure (see 45 above) has a void * field called
2296 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2297 pcre_callout_block structure has a field of the same name. The contents of
2298 the field passed in the pcre_extra structure are passed to the callout
2299 function in the corresponding field in the callout block. This makes it
2300 easier to use the same callout-containing regex from multiple threads. For
2301 testing, the pcretest program has a new data escape
2302
2303 \C*n pass the number n (may be negative) as callout_data
2304
2305 If the callout function in pcretest receives a non-zero value as
2306 callout_data, it returns that value.
2307
2308 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2309 there were some redundant $(CFLAGS) in commands that are now specified as
2310 $(LINK), which already includes $(CFLAGS).
2311
2312 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2313 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2314 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2315 one-byte characters throughout. Note that case-insensitive matching applies
2316 only to characters whose values are less than 256. PCRE doesn't support the
2317 notion of cases for higher-valued characters.
2318
2319 (i) A character class whose characters are all within 0-255 is handled as
2320 a bit map, and the map is inverted for negative classes. Previously, a
2321 character > 255 always failed to match such a class; however it should
2322 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2323
2324 (ii) A negated character class with a single character < 255 is coded as
2325 "not this character" (OP_NOT). This wasn't working properly when the test
2326 character was multibyte, either singly or repeated.
2327
2328 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2329 mode, for example: \x{100}{2,3}.
2330
2331 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2332 singly or repeated) now correctly test multibyte characters. However,
2333 PCRE doesn't recognize any characters with values greater than 255 as
2334 digits, spaces, or word characters. Such characters always match \D, \S,
2335 and \W, and never match \d, \s, or \w.
2336
2337 (v) Classes may now contain characters and character ranges with values
2338 greater than 255. For example: [ab\x{100}-\x{400}].
2339
2340 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2341 PCRE in UTF-8 mode.
2342
2343 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2344 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2345 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2346 value.)
2347
2348 53. The single man page has become too large. I have therefore split it up into
2349 a number of separate man pages. These also give rise to individual HTML pages;
2350 these are now put in a separate directory, and there is an index.html page that
2351 lists them all. Some hyperlinking between the pages has been installed.
2352
2353 54. Added convenience functions for handling named capturing parentheses.
2354
2355 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2356 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2357 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2358 are faulted.
2359
2360 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2361 calling configure. These values are used when compiling the dftables.c program
2362 which is run to generate the source of the default character tables. They
2363 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2364 you will need to set these values.
2365
2366 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2367
2368
2369 Version 3.9 02-Jan-02
2370 ---------------------
2371
2372 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2373
2374 2. If --disable-static was given, the building process failed when trying to
2375 build pcretest and pcregrep. (For some reason it was using libtool to compile
2376 them, which is not right, as they aren't part of the library.)
2377
2378
2379 Version 3.8 18-Dec-01
2380 ---------------------
2381
2382 1. The experimental UTF-8 code was completely screwed up. It was packing the
2383 bytes in the wrong order. How dumb can you get?
2384
2385
2386 Version 3.7 29-Oct-01
2387 ---------------------
2388
2389 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2390 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2391 this didn't happen under Solaris 8, where I normally test things.
2392
2393 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2394 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2395 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2396 this fix an hour or so after the initial 3.7 release.)
2397
2398
2399 Version 3.6 23-Oct-01
2400 ---------------------
2401
2402 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2403 offsets passed as NULL with zero offset count.
2404
2405 2. The config.guess and config.sub files had not been updated when I moved to
2406 the latest autoconf.
2407
2408
2409 Version 3.5 15-Aug-01
2410 ---------------------
2411
2412 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2413 had been forgotten.
2414
2415 2. By using declared but undefined structures, we can avoid using "void"
2416 definitions in pcre.h while keeping the internal definitions of the structures
2417 private.
2418
2419 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2420 user point of view, this means that both static and shared libraries are built
2421 by default, but this can be individually controlled. More of the work of
2422 handling this static/shared cases is now inside libtool instead of PCRE's make
2423 file.
2424
2425 4. The pcretest utility is now installed along with pcregrep because it is
2426 useful for users (to test regexs) and by doing this, it automatically gets
2427 relinked by libtool. The documentation has been turned into a man page, so
2428 there are now .1, .txt, and .html versions in /doc.
2429
2430 5. Upgrades to pcregrep:
2431 (i) Added long-form option names like gnu grep.
2432 (ii) Added --help to list all options with an explanatory phrase.
2433 (iii) Added -r, --recursive to recurse into sub-directories.
2434 (iv) Added -f, --file to read patterns from a file.
2435
2436 6. pcre_exec() was referring to its "code" argument before testing that
2437 argument for NULL (and giving an error if it was NULL).
2438
2439 7. Upgraded Makefile.in to allow for compiling in a different directory from
2440 the source directory.
2441
2442 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2443 options bits, the pointer it was passed was to an int instead of to an unsigned
2444 long int. This mattered only on 64-bit systems.
2445
2446 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2447 generated) instead of pcre.in, which it its source. Also made the same change
2448 in several of the .c files.
2449
2450 10. A new release of gcc defines printf() as a macro, which broke pcretest
2451 because it had an ifdef in the middle of a string argument for printf(). Fixed
2452 by using separate calls to printf().
2453
2454 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2455 script, to force use of CR or LF instead of \n in the source. On non-Unix
2456 systems, the value can be set in config.h.
2457
2458 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2459 absolute limit. Changed the text of the error message to make this clear, and
2460 likewise updated the man page.
2461
2462 13. The limit of 99 on the number of capturing subpatterns has been removed.
2463 The new limit is 65535, which I hope will not be a "real" limit.
2464
2465
2466 Version 3.4 22-Aug-00
2467 ---------------------
2468
2469 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2470
2471 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2472
2473
2474 Version 3.3 01-Aug-00
2475 ---------------------
2476
2477 1. If an octal character was given, but the value was greater than \377, it
2478 was not getting masked to the least significant bits, as documented. This could
2479 lead to crashes in some systems.
2480
2481 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2482 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2483
2484 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2485 These just pass their arguments on to (pcre_free)(), but they are provided
2486 because some uses of PCRE bind it to non-C systems that can call its functions,
2487 but cannot call free() or pcre_free() directly.
2488
2489 4. Add "make test" as a synonym for "make check". Corrected some comments in
2490 the Makefile.
2491
2492 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2493 Makefile.
2494
2495 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2496 command called pgrep for grepping around the active processes.
2497
2498 7. Added the beginnings of support for UTF-8 character strings.
2499
2500 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2501 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2502 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2503 out for the ar command.)
2504
2505
2506 Version 3.2 12-May-00
2507 ---------------------
2508
2509 This is purely a bug fixing release.
2510
2511 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2512 of ZA. This was just one example of several cases that could provoke this bug,
2513 which was introduced by change 9 of version 2.00. The code for breaking
2514 infinite loops after an iteration that matches an empty string was't working
2515 correctly.
2516
2517 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2518 when matched against abbab (for example). After matching an empty string, it
2519 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2520 caused it to match further down the string than it should.
2521
2522 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2523 was there because it doesn't seem to be needed, and it causes trouble on some
2524 systems, as it is not a Standard C header. It has been removed.
2525
2526 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2527 were reported on the Macintosh. The changes were from
2528
2529 while ((c = *(++ptr)) != 0 && c != '\n');
2530 to
2531 while ((c = *(++ptr)) != 0 && c != '\n') ;
2532
2533 Totally extraordinary, but if that's what it takes...
2534
2535 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2536 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2537 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2538 assumes the way PCRE uses memmove() (always moving upwards).
2539
2540 6. PCRE is being used in one environment where strchr() is not available. There
2541 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2542 faster code anyway.
2543
2544
2545 Version 3.1 09-Feb-00
2546 ---------------------
2547
2548 The only change in this release is the fixing of some bugs in Makefile.in for
2549 the "install" target:
2550
2551 (1) It was failing to install pcreposix.h.
2552
2553 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2554
2555
2556 Version 3.0 01-Feb-00
2557 ---------------------
2558
2559 1. Add support for the /+ modifier to perltest (to output $` like it does in
2560 pcretest).
2561
2562 2. Add support for the /g modifier to perltest.
2563
2564 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2565 matches null strings.
2566
2567 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2568 pattern. Perl treats empty patterns specially - it reuses the most recent
2569 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2570 effect.
2571
2572 5. The POSIX interface was broken in that it was just handing over the POSIX
2573 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2574 required a bigger vector, with some working space on the end. This means that
2575 the POSIX wrapper now has to get and free some memory, and copy the results.
2576
2577 6. Added some simple autoconf support, placing the test data and the
2578 documentation in separate directories, re-organizing some of the
2579 information files, and making it build pcre-config (a GNU standard). Also added
2580 libtool support for building PCRE as a shared library, which is now the
2581 default.
2582
2583 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2584 09 are not valid octal constants. Single digits will be used for minor values
2585 less than 10.
2586
2587 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2588 existing programs that set these in the POSIX interface can use PCRE without
2589 modification.
2590
2591 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2592 return all that pcre_info() returns, plus additional data. The pcre_info()
2593 function is retained for compatibility, but is considered to be obsolete.
2594
2595 10. Added experimental recursion feature (?R) to handle one common case that
2596 Perl 5.6 will be able to do with (?p{...}).
2597
2598 11. Added support for POSIX character classes like [:alpha:], which Perl is
2599 adopting.
2600
2601
2602 Version 2.08 31-Aug-99
2603 ----------------------
2604
2605 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2606 trying to match at the startoffset position, but instead was moving forward to
2607 the next newline as if a previous match had failed.
2608
2609 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2610 and could get into a loop if a null string was matched other than at the start
2611 of the subject.
2612
2613 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2614 be distinguished at compile time, and for completeness also added PCRE_DATE.
2615
2616 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2617 in GnuWin32 environments.
2618
2619
2620 Version 2.07 29-Jul-99
2621 ----------------------
2622
2623 1. The documentation is now supplied in plain text form and HTML as well as in
2624 the form of man page sources.
2625
2626 2. C++ compilers don't like assigning (void *) values to other pointer types.
2627 In particular this affects malloc(). Although there is no problem in Standard
2628 C, I've put in casts to keep C++ compilers happy.
2629
2630 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2631 should be (const char *).
2632
2633 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2634 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2635 However, I haven't made this a standard facility. The documentation doesn't
2636 mention it, and the Makefile doesn't support it.
2637
2638 5. The Makefile now contains an "install" target, with editable destinations at
2639 the top of the file. The pcretest program is not installed.
2640
2641 6. pgrep -V now gives the PCRE version number and date.
2642
2643 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2644 causing the entire string to be ignored, instead of just the last character.
2645
2646 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2647 non-matching string, it can take a very, very long time, even for strings of
2648 quite modest length, because of the nested recursion. PCRE now does better in
2649 some of these cases. It does this by remembering the last required literal
2650 character in the pattern, and pre-searching the subject to ensure it is present
2651 before running the real match. In other words, it applies a heuristic to detect
2652 some types of certain failure quickly, and in the above example, if presented
2653 with a string that has no trailing " it gives "no match" very quickly.
2654
2655 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2656 other alternatives are tried instead.
2657
2658
2659 Version 2.06 09-Jun-99
2660 ----------------------
2661
2662 1. Change pcretest's output for amount of store used to show just the code
2663 space, because the remainder (the data block) varies in size between 32-bit and
2664 64-bit systems.
2665
2666 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2667 start matching at. This allows lookbehinds to work when searching for multiple
2668 occurrences in a string.
2669
2670 3. Added additional options to pcretest for testing multiple occurrences:
2671
2672 /+ outputs the rest of the string that follows a match
2673 /g loops for multiple occurrences, using the new startoffset argument
2674 /G loops for multiple occurrences by passing an incremented pointer
2675
2676 4. PCRE wasn't doing the "first character" optimization for patterns starting
2677 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2678 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2679 the letter 'x'. On long subject strings, this gives a significant speed-up.
2680
2681
2682 Version 2.05 21-Apr-99
2683 ----------------------
2684
2685 1. Changed the type of magic_number from int to long int so that it works
2686 properly on 16-bit systems.
2687
2688 2. Fixed a bug which caused patterns starting with .* not to work correctly
2689 when the subject string contained newline characters. PCRE was assuming
2690 anchoring for such patterns in all cases, which is not correct because .* will
2691 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2692 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2693 must be retried after every newline in the subject.
2694
2695
2696 Version 2.04 18-Feb-99
2697 ----------------------
2698
2699 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2700 computation of the store needed to hold the pattern was incorrect (too large).
2701 If such patterns were nested a few deep, this could multiply and become a real
2702 problem.
2703
2704 2. Added /M option to pcretest to show the memory requirement of a specific
2705 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2706
2707 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2708 compiled in such a way that the backtracking after subsequent failure was
2709 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2710 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2711
2712
2713 Version 2.03 02-Feb-99
2714 ----------------------
2715
2716 1. Fixed typo and small mistake in man page.
2717
2718 2. Added 4th condition (GPL supersedes if conflict) and created separate
2719 LICENCE file containing the conditions.
2720
2721 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2722 Perl, that is the internal \ allows the delimiter to be included in the
2723 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2724 the final delimiter, add \ to the end of the pattern (to test the error).
2725
2726 4. Added the convenience functions for extracting substrings after a successful
2727 match. Updated pcretest to make it able to test these functions.
2728
2729
2730 Version 2.02 14-Jan-99
2731 ----------------------
2732
2733 1. Initialized the working variables associated with each extraction so that
2734 their saving and restoring doesn't refer to uninitialized store.
2735
2736 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2737 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2738 fix the problem.
2739
2740 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2741 calls, and wasn't printing the correct value for compiling calls. Increased the
2742 default value of LOOPREPEAT, and the number of significant figures in the
2743 times.
2744
2745 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2746
2747 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2748 a building problem on Windows NT with a FAT file system.
2749
2750
2751 Version 2.01 21-Oct-98
2752 ----------------------
2753
2754 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2755 to character tables built by pcre_maketables() in the current locale. If NULL
2756 is passed, the default tables are used.
2757
2758
2759 Version 2.00 24-Sep-98
2760 ----------------------
2761
2762 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2763 it any more.
2764
2765 2. Allow quantification of (?>) groups, and make it work correctly.
2766
2767 3. The first character computation wasn't working for (?>) groups.
2768
2769 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2770 end of the subject) and add 5.005's \z, which really does match only at the
2771 very end of the subject.
2772
2773 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2774
2775 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2776 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2777 localized options. All options to pcre_study() were also removed.
2778
2779 7. Add other new features from 5.005:
2780
2781 $(?<= positive lookbehind
2782 $(?<! negative lookbehind
2783 (?imsx-imsx) added the unsetting capability
2784 such a setting is global if at outer level; local otherwise
2785 (?imsx-imsx:) non-capturing groups with option setting
2786 (?(cond)re|re) conditional pattern matching
2787
2788 A backreference to itself in a repeated group matches the previous
2789 captured string.
2790
2791 8. General tidying up of studying (both automatic and via "study")
2792 consequential on the addition of new assertions.
2793
2794 9. As in 5.005, unlimited repeated groups that could match an empty substring
2795 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2796 runtime if any iteration does actually match an empty substring.
2797
2798 10. Include the RunTest script in the distribution.
2799
2800 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2801 discrepancies, some of which were old and were also with respect to 5.004. They
2802 have now been fixed.
2803
2804
2805 Version 1.09 28-Apr-98
2806 ----------------------
2807
2808 1. A negated single character class followed by a quantifier with a minimum
2809 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2810 program crashes, or just wrong answers. This did not apply to negated classes
2811 containing more than one character, or to minima other than one.
2812
2813
2814 Version 1.08 27-Mar-98
2815 ----------------------
2816
2817 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2818
2819 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2820 latter must appear before anything that relies on it in the pattern.
2821
2822
2823 Version 1.07 16-Feb-98
2824 ----------------------
2825
2826 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2827 repeat of a potentially empty string).
2828
2829
2830 Version 1.06 23-Jan-98
2831 ----------------------
2832
2833 1. Added Markus Oberhumer's little patches for C++.
2834
2835 2. Literal strings longer than 255 characters were broken.
2836
2837
2838 Version 1.05 23-Dec-97
2839 ----------------------
2840
2841 1. Negated character classes containing more than one character were failing if
2842 PCRE_CASELESS was set at run time.
2843
2844
2845 Version 1.04 19-Dec-97
2846 ----------------------
2847
2848 1. Corrected the man page, where some "const" qualifiers had been omitted.
2849
2850 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2851 input syntax.
2852
2853 3. Fixed memory leak which occurred when a regex with back references was
2854 matched with an offsets vector that wasn't big enough. The temporary memory
2855 that is used in this case wasn't being freed if the match failed.
2856
2857 4. Tidied pcretest to ensure it frees memory that it gets.
2858
2859 5. Temporary memory was being obtained in the case where the passed offsets
2860 vector was exactly big enough.
2861
2862 6. Corrected definition of offsetof() from change 5 below.
2863
2864 7. I had screwed up change 6 below and broken the rules for the use of
2865 setjmp(). Now fixed.
2866
2867
2868 Version 1.03 18-Dec-97
2869 ----------------------
2870
2871 1. A erroneous regex with a missing opening parenthesis was correctly
2872 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2873 on some systems.
2874
2875 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2876 it was reported that one broken compiler failed on the former because "code" is
2877 also an independent variable.
2878
2879 3. The erroneous regex a[]b caused an array overrun reference.
2880
2881 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2882 fail on data ending with that character. (It was going on too far, and checking
2883 the next character, typically a binary zero.) This was specific to the
2884 optimized code for single-character negative classes.
2885
2886 5. Added a contributed patch from the TIN world which does the following:
2887
2888 + Add an undef for memmove, in case the the system defines a macro for it.
2889
2890 + Add a definition of offsetof(), in case there isn't one. (I don't know
2891 the reason behind this - offsetof() is part of the ANSI standard - but
2892 it does no harm).
2893
2894 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2895 most of the places where whitespace preceded '#'. I have given up and
2896 allowed the remaining 2 cases to be at the margin.
2897
2898 + Rename some variables in pcre to eliminate shadowing. This seems very
2899 pedantic, but does no harm, of course.
2900
2901 6. Moved the call to setjmp() into its own function, to get rid of warnings
2902 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2903
2904 7. Constructs such as \d{8,} were compiling into the equivalent of
2905 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2906 outcome, but in this particular case used more store than had been allocated,
2907 which caused the bug to be discovered because it threw up an internal error.
2908
2909 8. The debugging code in both pcre and pcretest for outputting the compiled
2910 form of a regex was going wrong in the case of back references followed by
2911 curly-bracketed repeats.
2912
2913
2914 Version 1.02 12-Dec-97
2915 ----------------------
2916
2917 1. Typos in pcre.3 and comments in the source fixed.
2918
2919 2. Applied a contributed patch to get rid of places where it used to remove
2920 'const' from variables, and fixed some signed/unsigned and uninitialized
2921 variable warnings.
2922
2923 3. Added the "runtest" target to Makefile.
2924
2925 4. Set default compiler flag to -O2 rather than just -O.
2926
2927
2928 Version 1.01 19-Nov-97
2929 ----------------------
2930
2931 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2932 like /([ab]*)*/, that is, for classes with more than one character in them.
2933
2934 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2935 as /((?>a*))*/ (a PCRE_EXTRA facility).
2936
2937
2938 Version 1.00 18-Nov-97
2939 ----------------------
2940
2941 1. Added compile-time macros to support systems such as SunOS4 which don't have
2942 memmove() or strerror() but have other things that can be used instead.
2943
2944 2. Arranged that "make clean" removes the executables.
2945
2946
2947 Version 0.99 27-Oct-97
2948 ----------------------
2949
2950 1. Fixed bug in code for optimizing classes with only one character. It was
2951 initializing a 32-byte map regardless, which could cause it to run off the end
2952 of the memory it had got.
2953
2954 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
2955
2956
2957 Version 0.98 22-Oct-97
2958 ----------------------
2959
2960 1. Fixed bug in code for handling temporary memory usage when there are more
2961 back references than supplied space in the ovector. This could cause segfaults.
2962
2963
2964 Version 0.97 21-Oct-97
2965 ----------------------
2966
2967 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
2968
2969 2. Optimized negated single characters not to use a bit map.
2970
2971 3. Brought error texts together as macro definitions; clarified some of them;
2972 fixed one that was wrong - it said "range out of order" when it meant "invalid
2973 escape sequence".
2974
2975 4. Changed some char * arguments to const char *.
2976
2977 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
2978
2979 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
2980 pcretest.
2981
2982
2983 Version 0.96 16-Oct-97
2984 ----------------------
2985
2986 1. Added a simple "pgrep" utility to the distribution.
2987
2988 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
2989 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
2990 where "ddd" means "one or more decimal digits".
2991
2992 3. Fixed serious bug. If a pattern had a back reference, but the call to
2993 pcre_exec() didn't supply a large enough ovector to record the related
2994 identifying subpattern, the match always failed. PCRE now remembers the number
2995 of the largest back reference, and gets some temporary memory in which to save
2996 the offsets during matching if necessary, in order to ensure that
2997 backreferences always work.
2998
2999 4. Increased the compatibility with Perl in a number of ways:
3000
3001 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3002 to request this handling. The option can be set at compile or exec time.
3003
3004 (b) $ matches before a terminating newline by default; an option
3005 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3006 mode). The option can be set at compile or exec time.
3007
3008 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3009 the same as Perl's. If the decimal number it represents is less than 10
3010 or there aren't that many previous left capturing parentheses, an octal
3011 escape is read. Inside a character class, it's always an octal escape,
3012 even if it is a single digit.
3013
3014 (d) An escaped but undefined alphabetic character is taken as a literal,
3015 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3016 escapes.
3017
3018 (e) {0} is now permitted. (The previous item is removed from the compiled
3019 pattern).
3020
3021 5. Changed all the names of code files so that the basic parts are no longer
3022 than 10 characters, and abolished the teeny "globals.c" file.
3023
3024 6. Changed the handling of character classes; they are now done with a 32-byte
3025 bit map always.
3026
3027 7. Added the -d and /D options to pcretest to make it possible to look at the
3028 internals of compilation without having to recompile pcre.
3029
3030
3031 Version 0.95 23-Sep-97
3032 ----------------------
3033
3034 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3035 \x20 at the start of a run of normal characters. These were being treated as
3036 real characters, instead of the source characters being re-checked.
3037
3038
3039 Version 0.94 18-Sep-97
3040 ----------------------
3041
3042 1. The functions are now thread-safe, with the caveat that the global variables
3043 containing pointers to malloc() and free() or alternative functions are the
3044 same for all threads.
3045
3046 2. Get pcre_study() to generate a bitmap of initial characters for non-
3047 anchored patterns when this is possible, and use it if passed to pcre_exec().
3048
3049
3050 Version 0.93 15-Sep-97
3051 ----------------------
3052
3053 1. /(b)|(:+)/ was computing an incorrect first character.
3054
3055 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3056 but not actually doing anything yet.
3057
3058 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3059 as Perl does (e.g. [-az] or [az-]).
3060
3061 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3062 all possible positions.
3063
3064 5. Split up into different modules to avoid including unneeded functions in a
3065 compiled binary. However, compile and exec are still in one module. The "study"
3066 function is split off.
3067
3068 6. The character tables are now in a separate module whose source is generated
3069 by an auxiliary program - but can then be edited by hand if required. There are
3070 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3071 toupper() in the code.
3072
3073 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3074 make them global. Abolish the function for setting them, as the caller can now
3075 set them directly.
3076
3077
3078 Version 0.92 11-Sep-97
3079 ----------------------
3080
3081 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3082 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3083
3084 2. Caseless matching was not working in character classes if the characters in
3085 the pattern were in upper case.
3086
3087 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3088
3089 4. Make PCRE_ANCHORED public and accept as a compile option.
3090
3091 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3092 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3093 pass them.
3094
3095 6. Give an error if bad option bits passed at compile or run time.
3096
3097 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3098 pcretest to cause it to pass that flag.
3099
3100 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3101 options, and the first character, if set.
3102
3103 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3104
3105
3106 Version 0.91 10-Sep-97
3107 ----------------------
3108
3109 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3110 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3111
3112 2. PCRE was looping on encountering an indefinitely repeated back reference to
3113 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3114 Perl does - treats the match as successful.
3115
3116 ****

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