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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 245 - (show annotations) (download)
Fri Sep 14 10:06:05 2007 UTC (6 years, 7 months ago) by ph10
File size: 141290 byte(s)
Skip some tests when \R is defaulted to ANYCRLF.

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

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