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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 263 - (show annotations) (download)
Mon Nov 12 16:53:25 2007 UTC (6 years, 10 months ago) by ph10
File size: 142116 byte(s)
Apply Craig's patch, which makes it possible to "ignore" values in parens 
when parsing an RE using the c++ wrapper.

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

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