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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 255 - (show annotations) (download)
Wed Sep 19 08:50:04 2007 UTC (6 years, 10 months ago) by ph10
File size: 141451 byte(s)
Add casts to pcretest.c to avoid compiler warnings.

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

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