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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 257 - (show annotations) (download)
Wed Sep 19 09:11:19 2007 UTC (7 years ago) by ph10
File size: 141805 byte(s)
Craig's patch to remove the checks for windows.h and instead check for 
_strtoi64 explicitly, and avoid using snprintf() at all.

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

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