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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 314 - (show annotations) (download)
Fri Jan 25 17:09:21 2008 UTC (6 years, 8 months ago) by ph10
File size: 150677 byte(s)
Wording tidy, as suggested by Sheri.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.6 21-Jan-08
5 ---------------------
6
7 1. A character class containing a very large number of characters with
8 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
9 overflow.
10
11 2. Patch to cut out the "long long" test in pcrecpp_unittest when
12 HAVE_LONG_LONG is not defined.
13
14 3. Applied Christian Ehrlicher's patch to update the CMake build files to
15 bring them up to date and include new features. This patch includes:
16
17 - Fixed PH's badly added libz and libbz2 support.
18 - Fixed a problem with static linking.
19 - Added pcredemo. [But later removed - see 7 below.]
20 - Fixed dftables problem and added an option.
21 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
22 HAVE_LONG_LONG.
23 - Added readline support for pcretest.
24 - Added an listing of the option settings after cmake has run.
25
26 4. A user submitted a patch to Makefile that makes it easy to create
27 "pcre.dll" under mingw when using Configure/Make. I added stuff to
28 Makefile.am that cause it to include this special target, without
29 affecting anything else. Note that the same mingw target plus all
30 the other distribution libraries and programs are now supported
31 when configuring with CMake (see 6 below) instead of with
32 Configure/Make.
33
34 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
35 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
36 exported in the Windows port". It has not yet been confirmed that the patch
37 solves the problem, but it does no harm.
38
39 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
40 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
41 with CMake, and also correct the comment about stack recursion.
42
43 7. Remove the automatic building of pcredemo from the ./configure system and
44 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
45 of a program that users should build themselves after PCRE is installed, so
46 building it automatically is not really right. What is more, it gave
47 trouble in some build environments.
48
49
50 Version 7.5 10-Jan-08
51 ---------------------
52
53 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
54 values in parens when parsing an RE using the C++ wrapper."
55
56 2. Negative specials like \S did not work in character classes in UTF-8 mode.
57 Characters greater than 255 were excluded from the class instead of being
58 included.
59
60 3. The same bug as (2) above applied to negated POSIX classes such as
61 [:^space:].
62
63 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
64 defined or documented. It seems to have been a typo for PCRE_STATIC, so
65 I have changed it.
66
67 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
68 first named subpattern) and a construct such as (?&a) would reference the
69 first named subpattern whose name started with "a" (in other words, the
70 length check was missing). Both these problems are fixed. "Subpattern name
71 expected" is now given for (?&) (a zero-length name), and this patch also
72 makes it give the same error for \k'' (previously it complained that that
73 was a reference to a non-existent subpattern).
74
75 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
76 this is right because (?- can be followed by option settings as well as by
77 digits. I have, however, made the messages clearer.
78
79 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
80 than the number used in the conditional) now cause a compile-time error.
81 This is actually not compatible with Perl, which accepts such patterns, but
82 treats the conditional as always being FALSE (as PCRE used to), but it
83 seems to me that giving a diagnostic is better.
84
85 8. Change "alphameric" to the more common word "alphanumeric" in comments
86 and messages.
87
88 9. Fix two occurrences of "backslash" in comments that should have been
89 "backspace".
90
91 10. Remove two redundant lines of code that can never be obeyed (their function
92 was moved elsewhere).
93
94 11. The program that makes PCRE's Unicode character property table had a bug
95 which caused it to generate incorrect table entries for sequences of
96 characters that have the same character type, but are in different scripts.
97 It amalgamated them into a single range, with the script of the first of
98 them. In other words, some characters were in the wrong script. There were
99 thirteen such cases, affecting characters in the following ranges:
100
101 U+002b0 - U+002c1
102 U+0060c - U+0060d
103 U+0061e - U+00612
104 U+0064b - U+0065e
105 U+0074d - U+0076d
106 U+01800 - U+01805
107 U+01d00 - U+01d77
108 U+01d9b - U+01dbf
109 U+0200b - U+0200f
110 U+030fc - U+030fe
111 U+03260 - U+0327f
112 U+0fb46 - U+0fbb1
113 U+10450 - U+1049d
114
115 12. The -o option (show only the matching part of a line) for pcregrep was not
116 compatible with GNU grep in that, if there was more than one match in a
117 line, it showed only the first of them. It now behaves in the same way as
118 GNU grep.
119
120 13. If the -o and -v options were combined for pcregrep, it printed a blank
121 line for every non-matching line. GNU grep prints nothing, and pcregrep now
122 does the same. The return code can be used to tell if there were any
123 non-matching lines.
124
125 14. Added --file-offsets and --line-offsets to pcregrep.
126
127 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
128 infinitely looping recursion. The bug was that positive lookaheads were not
129 being skipped when checking for a possible empty match (negative lookaheads
130 and both kinds of lookbehind were skipped).
131
132 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
133 inclusion of <windows.h> to before rather than after the definition of
134 INVALID_FILE_ATTRIBUTES (patch from David Byron).
135
136 17. Specifying a possessive quantifier with a specific limit for a Unicode
137 character property caused pcre_compile() to compile bad code, which led at
138 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
139 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
140 caused the error; without that there was no problem.
141
142 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
143
144 19. Added --enable-pcretest-libreadline.
145
146 20. In pcrecpp.cc, the variable 'count' was incremented twice in
147 RE::GlobalReplace(). As a result, the number of replacements returned was
148 double what it should be. I removed one of the increments, but Craig sent a
149 later patch that removed the other one (the right fix) and added unit tests
150 that check the return values (which was not done before).
151
152 21. Several CMake things:
153
154 (1) Arranged that, when cmake is used on Unix, the libraries end up with
155 the names libpcre and libpcreposix, not just pcre and pcreposix.
156
157 (2) The above change means that pcretest and pcregrep are now correctly
158 linked with the newly-built libraries, not previously installed ones.
159
160 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
161
162 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
163 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
164 UTF-8 newline character). The key issue is that the pattern starts .*;
165 this means that the match must be either at the beginning, or after a
166 newline. The bug was in the code for advancing after a failed match and
167 checking that the new position followed a newline. It was not taking
168 account of UTF-8 characters correctly.
169
170 23. PCRE was behaving differently from Perl in the way it recognized POSIX
171 character classes. PCRE was not treating the sequence [:...:] as a
172 character class unless the ... were all letters. Perl, however, seems to
173 allow any characters between [: and :], though of course it rejects as
174 unknown any "names" that contain non-letters, because all the known class
175 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
176 for example, whereas PCRE did not - it did not recognize a POSIX character
177 class. This seemed a bit dangerous, so the code has been changed to be
178 closer to Perl. The behaviour is not identical to Perl, because PCRE will
179 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
180 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
181 Perl does, and where it didn't before.
182
183 24. Rewrite so as to remove the single use of %n from pcregrep because in some
184 Windows environments %n is disabled by default.
185
186
187 Version 7.4 21-Sep-07
188 ---------------------
189
190 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
191 means that a class such as [\s] counted as "explicit reference to CR or
192 LF". That isn't really right - the whole point of the change was to try to
193 help when there was an actual mention of one of the two characters. So now
194 the change happens only if \r or \n (or a literal CR or LF) character is
195 encountered.
196
197 2. The 32-bit options word was also used for 6 internal flags, but the numbers
198 of both had grown to the point where there were only 3 bits left.
199 Fortunately, there was spare space in the data structure, and so I have
200 moved the internal flags into a new 16-bit field to free up more option
201 bits.
202
203 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
204 but did not set the internal JCHANGED flag - either of these is enough to
205 control the way the "get" function works - but the PCRE_INFO_JCHANGED
206 facility is supposed to tell if (?J) was ever used, so now (?J) at the
207 start sets both bits.
208
209 4. Added options (at build time, compile time, exec time) to change \R from
210 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
211
212 5. doc/pcresyntax.html was missing from the distribution.
213
214 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
215 compatibility, even though it is no longer used.
216
217 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
218 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
219 windows.h file is present (where different names are used). [This was
220 reversed later after testing - see 16 below.]
221
222 8. Changed all #include <config.h> to #include "config.h". There were also
223 some further <pcre.h> cases that I changed to "pcre.h".
224
225 9. When pcregrep was used with the --colour option, it missed the line ending
226 sequence off the lines that it output.
227
228 10. It was pointed out to me that arrays of string pointers cause lots of
229 relocations when a shared library is dynamically loaded. A technique of
230 using a single long string with a table of offsets can drastically reduce
231 these. I have refactored PCRE in four places to do this. The result is
232 dramatic:
233
234 Originally: 290
235 After changing UCP table: 187
236 After changing error message table: 43
237 After changing table of "verbs" 36
238 After changing table of Posix names 22
239
240 Thanks to the folks working on Gregex for glib for this insight.
241
242 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
243 unicode-properties was also set.
244
245 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
246
247 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
248 checked only for CRLF.
249
250 14. Added casts to pcretest.c to avoid compiler warnings.
251
252 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
253
254 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
255 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
256 entirely. This removes changes made in 7 above.
257
258 17. The CMake files have been updated, and there is now more information about
259 building with CMake in the NON-UNIX-USE document.
260
261
262 Version 7.3 28-Aug-07
263 ---------------------
264
265 1. In the rejigging of the build system that eventually resulted in 7.1, the
266 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
267 brackets there is not right, since it causes compilers to look for an
268 installed pcre.h, not the version that is in the source that is being
269 compiled (which of course may be different). I have changed it back to:
270
271 #include "pcre.h"
272
273 I have a vague recollection that the change was concerned with compiling in
274 different directories, but in the new build system, that is taken care of
275 by the VPATH setting the Makefile.
276
277 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
278 when the subject happened to end in the byte 0x85 (e.g. if the last
279 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
280 characters but of course it shouldn't be taken as a newline when it is part
281 of another character. The bug was that, for an unlimited repeat of . in
282 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
283 characters when looking for a newline.
284
285 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
286
287 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
288 in debug output.
289
290 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
291 long printing in the pcrecpp unittest when running under MinGW.
292
293 6. ESC_K was left out of the EBCDIC table.
294
295 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
296 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
297 limit also applies to "virtual nesting" when a pattern is recursive, and in
298 this case 1000 isn't so big. I have been able to remove this limit at the
299 expense of backing off one optimization in certain circumstances. Normally,
300 when pcre_exec() would call its internal match() function recursively and
301 immediately return the result unconditionally, it uses a "tail recursion"
302 feature to save stack. However, when a subpattern that can match an empty
303 string has an unlimited repetition quantifier, it no longer makes this
304 optimization. That gives it a stack frame in which to save the data for
305 checking that an empty string has been matched. Previously this was taken
306 from the 1000-entry workspace that had been reserved. So now there is no
307 explicit limit, but more stack is used.
308
309 8. Applied Daniel's patches to solve problems with the import/export magic
310 syntax that is required for Windows, and which was going wrong for the
311 pcreposix and pcrecpp parts of the library. These were overlooked when this
312 problem was solved for the main library.
313
314 9. There were some crude static tests to avoid integer overflow when computing
315 the size of patterns that contain repeated groups with explicit upper
316 limits. As the maximum quantifier is 65535, the maximum group length was
317 set at 30,000 so that the product of these two numbers did not overflow a
318 32-bit integer. However, it turns out that people want to use groups that
319 are longer than 30,000 bytes (though not repeat them that many times).
320 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
321 made it possible to implement the integer overflow checks in a much more
322 dynamic way, which I have now done. The artificial limitation on group
323 length has been removed - we now have only the limit on the total length of
324 the compiled pattern, which depends on the LINK_SIZE setting.
325
326 10. Fixed a bug in the documentation for get/copy named substring when
327 duplicate names are permitted. If none of the named substrings are set, the
328 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
329 empty string.
330
331 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
332 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
333 because the ] is interpreted as the first data character and the
334 terminating ] is not found. PCRE has been made compatible with Perl in this
335 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
336 cause memory overwriting.
337
338 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
339 string has been matched (to stop an infinite loop). It was not recognizing
340 a conditional subpattern that could match an empty string if that
341 subpattern was within another subpattern. For example, it looped when
342 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
343 condition was not nested. This bug has been fixed.
344
345 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
346 past the start of the subject in the presence of bytes with the top bit
347 set, for example "\x8aBCD".
348
349 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
350 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
351
352 14. Optimized (?!) to (*FAIL).
353
354 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
355 This restricts code points to be within the range 0 to 0x10FFFF, excluding
356 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
357 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
358 does: it's just the validity check that is more restrictive.
359
360 16. Inserted checks for integer overflows during escape sequence (backslash)
361 processing, and also fixed erroneous offset values for syntax errors during
362 backslash processing.
363
364 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
365 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
366
367 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
368 caused an overrun.
369
370 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
371 something other than just ASCII characters) inside a group that had an
372 unlimited repeat caused a loop at compile time (while checking to see
373 whether the group could match an empty string).
374
375 20. Debugging a pattern containing \p or \P could cause a crash. For example,
376 [\P{Any}] did so. (Error in the code for printing property names.)
377
378 21. An orphan \E inside a character class could cause a crash.
379
380 22. A repeated capturing bracket such as (A)? could cause a wild memory
381 reference during compilation.
382
383 23. There are several functions in pcre_compile() that scan along a compiled
384 expression for various reasons (e.g. to see if it's fixed length for look
385 behind). There were bugs in these functions when a repeated \p or \P was
386 present in the pattern. These operators have additional parameters compared
387 with \d, etc, and these were not being taken into account when moving along
388 the compiled data. Specifically:
389
390 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
391 length.
392
393 (b) An item such as \pL+ within a repeated group could cause crashes or
394 loops.
395
396 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
397 "reference to non-existent subpattern" error.
398
399 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
400
401 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
402 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
403
404 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
405
406 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
407 character were causing crashes (broken optimization).
408
409 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
410 \p or \P) caused a compile-time loop.
411
412 28. More problems have arisen in unanchored patterns when CRLF is a valid line
413 break. For example, the unstudied pattern [\r\n]A does not match the string
414 "\r\nA" because change 7.0/46 below moves the current point on by two
415 characters after failing to match at the start. However, the pattern \nA
416 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
417 the same is true. There doesn't seem any very clean way out of this, but
418 what I have chosen to do makes the common cases work: PCRE now takes note
419 of whether there can be an explicit match for \r or \n anywhere in the
420 pattern, and if so, 7.0/46 no longer applies. As part of this change,
421 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
422 pattern has explicit CR or LF references.
423
424 29. Added (*CR) etc for changing newline setting at start of pattern.
425
426
427 Version 7.2 19-Jun-07
428 ---------------------
429
430 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
431 which is apparently normally available under Windows.
432
433 2. Re-jig the pcregrep tests with different newline settings in an attempt
434 to make them independent of the local environment's newline setting.
435
436 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
437
438 4. Some of the "internals" tests were previously cut out when the link size
439 was not 2, because the output contained actual offsets. The recent new
440 "Z" feature of pcretest means that these can be cut out, making the tests
441 usable with all link sizes.
442
443 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
444 stack recursion. This gives a massive performance boost under BSD, but just
445 a small improvement under Linux. However, it saves one field in the frame
446 in all cases.
447
448 6. Added more features from the forthcoming Perl 5.10:
449
450 (a) (?-n) (where n is a string of digits) is a relative subroutine or
451 recursion call. It refers to the nth most recently opened parentheses.
452
453 (b) (?+n) is also a relative subroutine call; it refers to the nth next
454 to be opened parentheses.
455
456 (c) Conditions that refer to capturing parentheses can be specified
457 relatively, for example, (?(-2)... or (?(+3)...
458
459 (d) \K resets the start of the current match so that everything before
460 is not part of it.
461
462 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
463
464 (f) \g{name} is another synonym - part of Perl 5.10's unification of
465 reference syntax.
466
467 (g) (?| introduces a group in which the numbering of parentheses in each
468 alternative starts with the same number.
469
470 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
471
472 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
473 PCRE_INFO_JCHANGED.
474
475 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
476 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
477 for detecting groups that can match an empty string.
478
479 9. A pattern with a very large number of alternatives (more than several
480 hundred) was running out of internal workspace during the pre-compile
481 phase, where pcre_compile() figures out how much memory will be needed. A
482 bit of new cunning has reduced the workspace needed for groups with
483 alternatives. The 1000-alternative test pattern now uses 12 bytes of
484 workspace instead of running out of the 4096 that are available.
485
486 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
487
488 11. Applied patch from Google to remove an optimization that didn't quite work.
489 The report of the bug said:
490
491 pcrecpp::RE("a*").FullMatch("aaa") matches, while
492 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
493 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
494
495 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
496 it matched the wrong number of bytes.
497
498
499 Version 7.1 24-Apr-07
500 ---------------------
501
502 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
503 that is more "standard", making use of automake and other Autotools. There
504 is some re-arrangement of the files and adjustment of comments consequent
505 on this.
506
507 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
508 for recursive directory scanning broke on some systems because the files
509 are not scanned in any specific order and on different systems the order
510 was different. A call to "sort" has been inserted into RunGrepTest for the
511 approprate test as a short-term fix. In the longer term there may be an
512 alternative.
513
514 3. I had an email from Eric Raymond about problems translating some of PCRE's
515 man pages to HTML (despite the fact that I distribute HTML pages, some
516 people do their own conversions for various reasons). The problems
517 concerned the use of low-level troff macros .br and .in. I have therefore
518 removed all such uses from the man pages (some were redundant, some could
519 be replaced by .nf/.fi pairs). The 132html script that I use to generate
520 HTML has been updated to handle .nf/.fi and to complain if it encounters
521 .br or .in.
522
523 4. Updated comments in configure.ac that get placed in config.h.in and also
524 arranged for config.h to be included in the distribution, with the name
525 config.h.generic, for the benefit of those who have to compile without
526 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
527
528 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
529 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
530 makevp.bat for the current PCRE, using the additional files
531 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
532
533 6. A Windows user reported a minor discrepancy with test 2, which turned out
534 to be caused by a trailing space on an input line that had got lost in his
535 copy. The trailing space was an accident, so I've just removed it.
536
537 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
538 that is needed.
539
540 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
541 as "const" (a) because they are and (b) because it helps the PHP
542 maintainers who have recently made a script to detect big data structures
543 in the php code that should be moved to the .rodata section. I remembered
544 to update Builducptable as well, so it won't revert if ucptable.h is ever
545 re-created.
546
547 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
548 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
549 order to be able to cut out the UTF-8 tables in the latter when UTF-8
550 support is not required. This saves 1.5-2K of code, which is important in
551 some applications.
552
553 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
554 so as not to refer to the tables, even though these functions will never be
555 called when UTF-8 support is disabled. Otherwise there are problems with a
556 shared library.
557
558 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
559
560 (a) It was defining its arguments as char * instead of void *.
561
562 (b) It was assuming that all moves were upwards in memory; this was true
563 a long time ago when I wrote it, but is no longer the case.
564
565 The emulated memove() is provided for those environments that have neither
566 memmove() nor bcopy(). I didn't think anyone used it these days, but that
567 is clearly not the case, as these two bugs were recently reported.
568
569 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
570 and Detrail to create the HTML documentation, the .txt form of the man
571 pages, and it removes trailing spaces from listed files. It also creates
572 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
573 case, it wraps all the #defines with #ifndefs. This script should be run
574 before "make dist".
575
576 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
577 with Unicode property support.
578
579 (a) For a maximizing quantifier, if the two different cases of the
580 character were of different lengths in their UTF-8 codings (there are
581 some cases like this - I found 11), and the matching function had to
582 back up over a mixture of the two cases, it incorrectly assumed they
583 were both the same length.
584
585 (b) When PCRE was configured to use the heap rather than the stack for
586 recursion during matching, it was not correctly preserving the data for
587 the other case of a UTF-8 character when checking ahead for a match
588 while processing a minimizing repeat. If the check also involved
589 matching a wide character, but failed, corruption could cause an
590 erroneous result when trying to check for a repeat of the original
591 character.
592
593 13. Some tidying changes to the testing mechanism:
594
595 (a) The RunTest script now detects the internal link size and whether there
596 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
597 values substituted by "configure". (The RunGrepTest script already did
598 this for UTF-8.) The configure.ac script no longer substitutes the
599 relevant variables.
600
601 (b) The debugging options /B and /D in pcretest show the compiled bytecode
602 with length and offset values. This means that the output is different
603 for different internal link sizes. Test 2 is skipped for link sizes
604 other than 2 because of this, bypassing the problem. Unfortunately,
605 there was also a test in test 3 (the locale tests) that used /B and
606 failed for link sizes other than 2. Rather than cut the whole test out,
607 I have added a new /Z option to pcretest that replaces the length and
608 offset values with spaces. This is now used to make test 3 independent
609 of link size. (Test 2 will be tidied up later.)
610
611 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
612 segmentation fault instead of returning the appropriate error message.
613
614 15. In multiline mode when the newline sequence was set to "any", the pattern
615 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
616 This doesn't seem right; it now treats the CRLF combination as the line
617 ending, and so does not match in that case. It's only a pattern such as ^$
618 that would hit this one: something like ^ABC$ would have failed after \r
619 and then tried again after \r\n.
620
621 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
622 in an attempt to make files that differ only in their line terminators
623 compare equal. This works on Linux.
624
625 17. Under certain error circumstances pcregrep might try to free random memory
626 as it exited. This is now fixed, thanks to valgrind.
627
628 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
629 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
630 was because its rules for how to advance for /g after matching an empty
631 string at the end of a line did not allow for this case. They now check for
632 it specially.
633
634 20. pcretest is supposed to handle patterns and data of any length, by
635 extending its buffers when necessary. It was getting this wrong when the
636 buffer for a data line had to be extended.
637
638 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
639 CRLF as a newline sequence.
640
641 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
642 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
643 I have nevertheless tidied it up.
644
645 23. Added some casts to kill warnings from HP-UX ia64 compiler.
646
647 24. Added a man page for pcre-config.
648
649
650 Version 7.0 19-Dec-06
651 ---------------------
652
653 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
654 moving to gcc 4.1.1.
655
656 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
657 sys/time.h, which is documented as needed for this function. It doesn't
658 seem to matter on Linux, but it showed up on some releases of OS X.
659
660 3. It seems that there are systems where bytes whose values are greater than
661 127 match isprint() in the "C" locale. The "C" locale should be the
662 default when a C program starts up. In most systems, only ASCII printing
663 characters match isprint(). This difference caused the output from pcretest
664 to vary, making some of the tests fail. I have changed pcretest so that:
665
666 (a) When it is outputting text in the compiled version of a pattern, bytes
667 other than 32-126 are always shown as hex escapes.
668
669 (b) When it is outputting text that is a matched part of a subject string,
670 it does the same, unless a different locale has been set for the match
671 (using the /L modifier). In this case, it uses isprint() to decide.
672
673 4. Fixed a major bug that caused incorrect computation of the amount of memory
674 required for a compiled pattern when options that changed within the
675 pattern affected the logic of the preliminary scan that determines the
676 length. The relevant options are -x, and -i in UTF-8 mode. The result was
677 that the computed length was too small. The symptoms of this bug were
678 either the PCRE error "internal error: code overflow" from pcre_compile(),
679 or a glibc crash with a message such as "pcretest: free(): invalid next
680 size (fast)". Examples of patterns that provoked this bug (shown in
681 pcretest format) are:
682
683 /(?-x: )/x
684 /(?x)(?-x: \s*#\s*)/
685 /((?i)[\x{c0}])/8
686 /(?i:[\x{c0}])/8
687
688 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
689 is now done differently.
690
691 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
692 wrapper classes; (b) implement a new function in the C++ scanner that is
693 more efficient than the old way of doing things because it avoids levels of
694 recursion in the regex matching; (c) add a paragraph to the documentation
695 for the FullMatch() function.
696
697 6. The escape sequence \n was being treated as whatever was defined as
698 "newline". Not only was this contrary to the documentation, which states
699 that \n is character 10 (hex 0A), but it also went horribly wrong when
700 "newline" was defined as CRLF. This has been fixed.
701
702 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
703 was being set to -1 for the "end of line" case (supposedly a value that no
704 character can have). Though this value is never used (the check for end of
705 line is "zero bytes in current character"), it caused compiler complaints.
706 I've changed it to 0xffffffff.
707
708 8. In pcre_version.c, the version string was being built by a sequence of
709 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
710 string (as it is for production releases) called a macro with an empty
711 argument. The C standard says the result of this is undefined. The gcc
712 compiler treats it as an empty string (which was what was wanted) but it is
713 reported that Visual C gives an error. The source has been hacked around to
714 avoid this problem.
715
716 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
717 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
718 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
719 of them did).
720
721 10. Originally, pcretest opened its input and output without "b"; then I was
722 told that "b" was needed in some environments, so it was added for release
723 5.0 to both the input and output. (It makes no difference on Unix-like
724 systems.) Later I was told that it is wrong for the input on Windows. I've
725 now abstracted the modes into two macros, to make it easier to fiddle with
726 them, and removed "b" from the input mode under Windows.
727
728 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
729
730 12. Added -help and --help to pcretest as an official way of being reminded
731 of the options.
732
733 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
734 and pcrecpp.cc because they annoy compilers at high warning levels.
735
736 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
737
738 15. Fixed an occurrence of == in configure.ac that should have been = (shell
739 scripts are not C programs :-) and which was not noticed because it works
740 on Linux.
741
742 16. pcretest is supposed to handle any length of pattern and data line (as one
743 line or as a continued sequence of lines) by extending its input buffer if
744 necessary. This feature was broken for very long pattern lines, leading to
745 a string of junk being passed to pcre_compile() if the pattern was longer
746 than about 50K.
747
748 17. I have done a major re-factoring of the way pcre_compile() computes the
749 amount of memory needed for a compiled pattern. Previously, there was code
750 that made a preliminary scan of the pattern in order to do this. That was
751 OK when PCRE was new, but as the facilities have expanded, it has become
752 harder and harder to keep it in step with the real compile phase, and there
753 have been a number of bugs (see for example, 4 above). I have now found a
754 cunning way of running the real compile function in a "fake" mode that
755 enables it to compute how much memory it would need, while actually only
756 ever using a few hundred bytes of working memory and without too many
757 tests of the mode. This should make future maintenance and development
758 easier. A side effect of this work is that the limit of 200 on the nesting
759 depth of parentheses has been removed (though this was never a serious
760 limitation, I suspect). However, there is a downside: pcre_compile() now
761 runs more slowly than before (30% or more, depending on the pattern). I
762 hope this isn't a big issue. There is no effect on runtime performance.
763
764 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
765 newline (only possible for the last line of a file) and it was a
766 pattern that set a locale (followed by /Lsomething), pcretest crashed.
767
768 19. Added additional timing features to pcretest. (1) The -tm option now times
769 matching only, not compiling. (2) Both -t and -tm can be followed, as a
770 separate command line item, by a number that specifies the number of
771 repeats to use when timing. The default is 50000; this gives better
772 precision, but takes uncomfortably long for very large patterns.
773
774 20. Extended pcre_study() to be more clever in cases where a branch of a
775 subpattern has no definite first character. For example, (a*|b*)[cd] would
776 previously give no result from pcre_study(). Now it recognizes that the
777 first character must be a, b, c, or d.
778
779 21. There was an incorrect error "recursive call could loop indefinitely" if
780 a subpattern (or the entire pattern) that was being tested for matching an
781 empty string contained only one non-empty item after a nested subpattern.
782 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
783 incorrectly, because the \d was being skipped in the check.
784
785 22. The pcretest program now has a new pattern option /B and a command line
786 option -b, which is equivalent to adding /B to every pattern. This causes
787 it to show the compiled bytecode, without the additional information that
788 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
789 is the same as /B/I).
790
791 23. A new optimization is now able automatically to treat some sequences such
792 as a*b as a*+b. More specifically, if something simple (such as a character
793 or a simple class like \d) has an unlimited quantifier, and is followed by
794 something that cannot possibly match the quantified thing, the quantifier
795 is automatically "possessified".
796
797 24. A recursive reference to a subpattern whose number was greater than 39
798 went wrong under certain circumstances in UTF-8 mode. This bug could also
799 have affected the operation of pcre_study().
800
801 25. Realized that a little bit of performance could be had by replacing
802 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
803
804 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
805
806 27. Possessive quantifiers such as a++ were previously implemented by turning
807 them into atomic groups such as ($>a+). Now they have their own opcodes,
808 which improves performance. This includes the automatically created ones
809 from 23 above.
810
811 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
812 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
813 the first matched character to be a colon. This applied both to named and
814 numbered groups.
815
816 29. The ucpinternal.h header file was missing its idempotency #ifdef.
817
818 30. I was sent a "project" file called libpcre.a.dev which I understand makes
819 building PCRE on Windows easier, so I have included it in the distribution.
820
821 31. There is now a check in pcretest against a ridiculously large number being
822 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
823 loop, the loop is abandoned.
824
825 32. Forward references to subpatterns in conditions such as (?(2)...) where
826 subpattern 2 is defined later cause pcre_compile() to search forwards in
827 the pattern for the relevant set of parentheses. This search went wrong
828 when there were unescaped parentheses in a character class, parentheses
829 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
830
831 33. "Subroutine" calls and backreferences were previously restricted to
832 referencing subpatterns earlier in the regex. This restriction has now
833 been removed.
834
835 34. Added a number of extra features that are going to be in Perl 5.10. On the
836 whole, these are just syntactic alternatives for features that PCRE had
837 previously implemented using the Python syntax or my own invention. The
838 other formats are all retained for compatibility.
839
840 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
841 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
842 also .NET compatible.
843
844 (b) A recursion or subroutine call to a named group can now be defined as
845 (?&name) as well as (?P>name).
846
847 (c) A backreference to a named group can now be defined as \k<name> or
848 \k'name' as well as (?P=name). The new forms, as well as being in Perl
849 5.10, are also .NET compatible.
850
851 (d) A conditional reference to a named group can now use the syntax
852 (?(<name>) or (?('name') as well as (?(name).
853
854 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
855 groups (named and numbered) that are never evaluated inline, but can be
856 called as "subroutines" from elsewhere. In effect, the DEFINE condition
857 is always false. There may be only one alternative in such a group.
858
859 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
860 as the simple (?(R). The condition is true only if the most recent
861 recursion is that of the given number or name. It does not search out
862 through the entire recursion stack.
863
864 (g) The escape \gN or \g{N} has been added, where N is a positive or
865 negative number, specifying an absolute or relative reference.
866
867 35. Tidied to get rid of some further signed/unsigned compiler warnings and
868 some "unreachable code" warnings.
869
870 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
871 things, this adds five new scripts.
872
873 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
874 There were also incompatibilities regarding the handling of \Q..\E inside
875 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
876 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
877
878 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
879 matches an empty string, and forcibly breaks the loop. There were bugs in
880 this code in non-simple cases. For a pattern such as ^(a()*)* matched
881 against aaaa the result was just "a" rather than "aaaa", for example. Two
882 separate and independent bugs (that affected different cases) have been
883 fixed.
884
885 39. Refactored the code to abolish the use of different opcodes for small
886 capturing bracket numbers. This is a tidy that I avoided doing when I
887 removed the limit on the number of capturing brackets for 3.5 back in 2001.
888 The new approach is not only tidier, it makes it possible to reduce the
889 memory needed to fix the previous bug (38).
890
891 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
892 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
893 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
894 mode.
895
896 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
897 report.
898
899 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
900 copy construction and assignment in the C++ wrapper.
901
902 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
903 couple of bugs that could have given wrong results in the "--newline=crlf"
904 case.
905
906 44. Added a number of casts and did some reorganization of signed/unsigned int
907 variables following suggestions from Dair Grant. Also renamed the variable
908 "this" as "item" because it is a C++ keyword.
909
910 45. Arranged for dftables to add
911
912 #include "pcre_internal.h"
913
914 to pcre_chartables.c because without it, gcc 4.x may remove the array
915 definition from the final binary if PCRE is built into a static library and
916 dead code stripping is activated.
917
918 46. For an unanchored pattern, if a match attempt fails at the start of a
919 newline sequence, and the newline setting is CRLF or ANY, and the next two
920 characters are CRLF, advance by two characters instead of one.
921
922
923 Version 6.7 04-Jul-06
924 ---------------------
925
926 1. In order to handle tests when input lines are enormously long, pcretest has
927 been re-factored so that it automatically extends its buffers when
928 necessary. The code is crude, but this _is_ just a test program. The
929 default size has been increased from 32K to 50K.
930
931 2. The code in pcre_study() was using the value of the re argument before
932 testing it for NULL. (Of course, in any sensible call of the function, it
933 won't be NULL.)
934
935 3. The memmove() emulation function in pcre_internal.h, which is used on
936 systems that lack both memmove() and bcopy() - that is, hardly ever -
937 was missing a "static" storage class specifier.
938
939 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
940 containing an extended class (one that cannot be represented by a bitmap
941 because it contains high-valued characters or Unicode property items, e.g.
942 [\pZ]). Almost always one would set UTF-8 mode when processing such a
943 pattern, but PCRE should not loop if you do not (it no longer does).
944 [Detail: two cases were found: (a) a repeated subpattern containing an
945 extended class; (b) a recursive reference to a subpattern that followed a
946 previous extended class. It wasn't skipping over the extended class
947 correctly when UTF-8 mode was not set.]
948
949 5. A negated single-character class was not being recognized as fixed-length
950 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
951 compile error "lookbehind assertion is not fixed length".
952
953 6. The RunPerlTest auxiliary script was showing an unexpected difference
954 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
955 write a Perl script that can interpret lines of an input file either as
956 byte characters or as UTF-8, which is what "perltest" was being required to
957 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
958 can't do is switch easily at run time between having the "use utf8;" pragma
959 or not. In the end, I fudged it by using the RunPerlTest script to insert
960 "use utf8;" explicitly for the UTF-8 tests.
961
962 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
963 the end of the subject string, contrary to the documentation and to what
964 Perl does. This was true of both matching functions. Now it matches only at
965 the start of the subject and immediately after *internal* newlines.
966
967 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
968 a pointer to an int instead of a pointer to an unsigned long int. This
969 caused problems on 64-bit systems.
970
971 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
972 instance of the 'standard' template library not being so standard".
973
974 10. There was no check on the number of named subpatterns nor the maximum
975 length of a subpattern name. The product of these values is used to compute
976 the size of the memory block for a compiled pattern. By supplying a very
977 long subpattern name and a large number of named subpatterns, the size
978 computation could be caused to overflow. This is now prevented by limiting
979 the length of names to 32 characters, and the number of named subpatterns
980 to 10,000.
981
982 11. Subpatterns that are repeated with specific counts have to be replicated in
983 the compiled pattern. The size of memory for this was computed from the
984 length of the subpattern and the repeat count. The latter is limited to
985 65535, but there was no limit on the former, meaning that integer overflow
986 could in principle occur. The compiled length of a repeated subpattern is
987 now limited to 30,000 bytes in order to prevent this.
988
989 12. Added the optional facility to have named substrings with the same name.
990
991 13. Added the ability to use a named substring as a condition, using the
992 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
993 are numbers (not recommended). Forward references are permitted.
994
995 14. Added forward references in named backreferences (if you see what I mean).
996
997 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
998 pattern could run off the end of the subject. For example, the pattern
999 "(?s)(.{1,5})"8 did this with the subject "ab".
1000
1001 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1002 PCRE_CASELESS was set when matching characters that were quantified with ?
1003 or *.
1004
1005 17. A character class other than a single negated character that had a minimum
1006 but no maximum quantifier - for example [ab]{6,} - was not handled
1007 correctly by pce_dfa_exec(). It would match only one character.
1008
1009 18. A valid (though odd) pattern that looked like a POSIX character
1010 class but used an invalid character after [ (for example [[,abc,]]) caused
1011 pcre_compile() to give the error "Failed: internal error: code overflow" or
1012 in some cases to crash with a glibc free() error. This could even happen if
1013 the pattern terminated after [[ but there just happened to be a sequence of
1014 letters, a binary zero, and a closing ] in the memory that followed.
1015
1016 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1017 over the years. Originally (before any Unicode support), just the bottom 8
1018 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1019 output from "man perlunicode" includes this:
1020
1021 The regular expression compiler produces polymorphic opcodes. That
1022 is, the pattern adapts to the data and automatically switches to
1023 the Unicode character scheme when presented with Unicode data--or
1024 instead uses a traditional byte scheme when presented with byte
1025 data.
1026
1027 Sadly, a wide octal escape does not cause a switch, and in a string with
1028 no other multibyte characters, these octal escapes are treated as before.
1029 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1030 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1031 Unicode string.
1032
1033 I have not perpetrated such confusion in PCRE. Up till now, it took just
1034 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1035 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1036 translate to the appropriate multibyte character.
1037
1038 29. Applied some refactoring to reduce the number of warnings from Microsoft
1039 and Borland compilers. This has included removing the fudge introduced
1040 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1041 a warning about an unused variable.
1042
1043 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1044 characters since release 4.0, because Perl (from release 5.004) does not.
1045 [Or at least, is documented not to: some releases seem to be in conflict
1046 with the documentation.] However, when a pattern was studied with
1047 pcre_study() and all its branches started with \s, PCRE still included VT
1048 as a possible starting character. Of course, this did no harm; it just
1049 caused an unnecessary match attempt.
1050
1051 22. Removed a now-redundant internal flag bit that recorded the fact that case
1052 dependency changed within the pattern. This was once needed for "required
1053 byte" processing, but is no longer used. This recovers a now-scarce options
1054 bit. Also moved the least significant internal flag bit to the most-
1055 significant bit of the word, which was not previously used (hangover from
1056 the days when it was an int rather than a uint) to free up another bit for
1057 the future.
1058
1059 23. Added support for CRLF line endings as well as CR and LF. As well as the
1060 default being selectable at build time, it can now be changed at runtime
1061 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1062 specify that it is scanning data with non-default line endings.
1063
1064 24. Changed the definition of CXXLINK to make it agree with the definition of
1065 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1066
1067 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1068 recursions. This makes a big different to stack usage for some patterns.
1069
1070 26. If a subpattern containing a named recursion or subroutine reference such
1071 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1072 the space required for the compiled pattern went wrong and gave too small a
1073 value. Depending on the environment, this could lead to "Failed: internal
1074 error: code overflow at offset 49" or "glibc detected double free or
1075 corruption" errors.
1076
1077 27. Applied patches from Google (a) to support the new newline modes and (b) to
1078 advance over multibyte UTF-8 characters in GlobalReplace.
1079
1080 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1081 difference for some implementation of PCRE in some Windows version.
1082
1083 29. Added some extra testing facilities to pcretest:
1084
1085 \q<number> in a data line sets the "match limit" value
1086 \Q<number> in a data line sets the "match recursion limt" value
1087 -S <number> sets the stack size, where <number> is in megabytes
1088
1089 The -S option isn't available for Windows.
1090
1091
1092 Version 6.6 06-Feb-06
1093 ---------------------
1094
1095 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1096 in pcreposix.h. I have copied the definition from pcre.h.
1097
1098 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1099 because pcre.h is no longer a built file.
1100
1101 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1102 not normally included in the compiled code.
1103
1104
1105 Version 6.5 01-Feb-06
1106 ---------------------
1107
1108 1. When using the partial match feature with pcre_dfa_exec(), it was not
1109 anchoring the second and subsequent partial matches at the new starting
1110 point. This could lead to incorrect results. For example, with the pattern
1111 /1234/, partially matching against "123" and then "a4" gave a match.
1112
1113 2. Changes to pcregrep:
1114
1115 (a) All non-match returns from pcre_exec() were being treated as failures
1116 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1117 error message is output. Some extra information is given for the
1118 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1119 probably the only errors that are likely to be caused by users (by
1120 specifying a regex that has nested indefinite repeats, for instance).
1121 If there are more than 20 of these errors, pcregrep is abandoned.
1122
1123 (b) A binary zero was treated as data while matching, but terminated the
1124 output line if it was written out. This has been fixed: binary zeroes
1125 are now no different to any other data bytes.
1126
1127 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1128 used to set a locale for matching. The --locale=xxxx long option has
1129 been added (no short equivalent) to specify a locale explicitly on the
1130 pcregrep command, overriding the environment variables.
1131
1132 (d) When -B was used with -n, some line numbers in the output were one less
1133 than they should have been.
1134
1135 (e) Added the -o (--only-matching) option.
1136
1137 (f) If -A or -C was used with -c (count only), some lines of context were
1138 accidentally printed for the final match.
1139
1140 (g) Added the -H (--with-filename) option.
1141
1142 (h) The combination of options -rh failed to suppress file names for files
1143 that were found from directory arguments.
1144
1145 (i) Added the -D (--devices) and -d (--directories) options.
1146
1147 (j) Added the -F (--fixed-strings) option.
1148
1149 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1150
1151 (l) Added the --colo(u)r option.
1152
1153 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1154 is not present by default.
1155
1156 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1157 items such as (?R) or (?1), when the recursion could match a number of
1158 alternatives. If it matched one of the alternatives, but subsequently,
1159 outside the recursion, there was a failure, the code tried to back up into
1160 the recursion. However, because of the way PCRE is implemented, this is not
1161 possible, and the result was an incorrect result from the match.
1162
1163 In order to prevent this happening, the specification of recursion has
1164 been changed so that all such subpatterns are automatically treated as
1165 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1166
1167 4. I had overlooked the fact that, in some locales, there are characters for
1168 which isalpha() is true but neither isupper() nor islower() are true. In
1169 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1170 and ordfeminine) are like this. This affected the treatment of \w and \W
1171 when they appeared in character classes, but not when they appeared outside
1172 a character class. The bit map for "word" characters is now created
1173 separately from the results of isalnum() instead of just taking it from the
1174 upper, lower, and digit maps. (Plus the underscore character, of course.)
1175
1176 5. The above bug also affected the handling of POSIX character classes such as
1177 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1178 permanent tables. Instead, the bit maps for such a class were previously
1179 created as the appropriate unions of the upper, lower, and digit bitmaps.
1180 Now they are created by subtraction from the [[:word:]] class, which has
1181 its own bitmap.
1182
1183 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1184 It is created by subtracting the vertical space characters (\x09, \x0a,
1185 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1186 subtraction was done in the overall bitmap for a character class, meaning
1187 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1188 be recognized. This bug has been fixed.
1189
1190 7. Patches from the folks at Google:
1191
1192 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1193 real life, but is still worth protecting against".
1194
1195 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1196 regular expressions".
1197
1198 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1199 have it.
1200
1201 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1202 "configure" and the latter not, in order to fix a problem somebody had
1203 with compiling the Arg class on HP-UX.
1204
1205 (e) Improve the error-handling of the C++ wrapper a little bit.
1206
1207 (f) New tests for checking recursion limiting.
1208
1209 8. The pcre_memmove() function, which is used only if the environment does not
1210 have a standard memmove() function (and is therefore rarely compiled),
1211 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1212 returning a result (though PCRE never actually uses the result).
1213
1214 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1215 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1216 returned instead of calling malloc() with an overflowing number that would
1217 most likely cause subsequent chaos.
1218
1219 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1220
1221 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1222 with this option is matched, the nmatch and pmatch options of regexec() are
1223 ignored.
1224
1225 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1226 provided in case anyone wants to the the POSIX interface with UTF-8
1227 strings.
1228
1229 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1230 C++ linking (needed for some HP-UX environments).
1231
1232 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1233 (unused parameter) and in the pcre_printint() function (omitted "default"
1234 switch label when the default is to do nothing).
1235
1236 15. Added some code to make it possible, when PCRE is compiled as a C++
1237 library, to replace subject pointers for pcre_exec() with a smart pointer
1238 class, thus making it possible to process discontinuous strings.
1239
1240 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1241 much the same function. They were added by different people who were trying
1242 to make PCRE easy to compile on non-Unix systems. It has been suggested
1243 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1244 for compiling on Windows systems. I have therefore replaced it with
1245 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1246 defaults to "extern" for C or "extern C" for C++, which works fine on
1247 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1248 SCOPE with something explicit in config.h. In addition:
1249
1250 (a) pcreposix.h still had just "extern" instead of either of these macros;
1251 I have replaced it with PCRE_DATA_SCOPE.
1252
1253 (b) Functions such as _pcre_xclass(), which are internal to the library,
1254 but external in the C sense, all had PCRE_EXPORT in their definitions.
1255 This is apparently wrong for the Windows case, so I have removed it.
1256 (It makes no difference on Unix-like systems.)
1257
1258 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1259 of recursive calls to match(). This is different to MATCH_LIMIT because
1260 that limits the total number of calls to match(), not all of which increase
1261 the depth of recursion. Limiting the recursion depth limits the amount of
1262 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1263 when PCRE is compiled, and changed at run time. A patch from Google adds
1264 this functionality to the C++ interface.
1265
1266 18. Changes to the handling of Unicode character properties:
1267
1268 (a) Updated the table to Unicode 4.1.0.
1269
1270 (b) Recognize characters that are not in the table as "Cn" (undefined).
1271
1272 (c) I revised the way the table is implemented to a much improved format
1273 which includes recognition of ranges. It now supports the ranges that
1274 are defined in UnicodeData.txt, and it also amalgamates other
1275 characters into ranges. This has reduced the number of entries in the
1276 table from around 16,000 to around 3,000, thus reducing its size
1277 considerably. I realized I did not need to use a tree structure after
1278 all - a binary chop search is just as efficient. Having reduced the
1279 number of entries, I extended their size from 6 bytes to 8 bytes to
1280 allow for more data.
1281
1282 (d) Added support for Unicode script names via properties such as \p{Han}.
1283
1284 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1285 matching that character.
1286
1287 20. When matching a repeated Unicode property with a minimum greater than zero,
1288 (for example \pL{2,}), PCRE could look past the end of the subject if it
1289 reached it while seeking the minimum number of characters. This could
1290 happen only if some of the characters were more than one byte long, because
1291 there is a check for at least the minimum number of bytes.
1292
1293 21. Refactored the implementation of \p and \P so as to be more general, to
1294 allow for more different types of property in future. This has changed the
1295 compiled form incompatibly. Anybody with saved compiled patterns that use
1296 \p or \P will have to recompile them.
1297
1298 22. Added "Any" and "L&" to the supported property types.
1299
1300 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1301 but give a compile time error if the value is greater than 0xff.
1302
1303 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1304 accidentally not being installed or uninstalled.
1305
1306 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1307 made were to insert the current release number. This seemed silly, because
1308 it made things harder for people building PCRE on systems that don't run
1309 "configure". I have turned pcre.h into a distributed file, no longer built
1310 by "configure", with the version identification directly included. There is
1311 no longer a pcre.h.in file.
1312
1313 However, this change necessitated a change to the pcre-config script as
1314 well. It is built from pcre-config.in, and one of the substitutions was the
1315 release number. I have updated configure.ac so that ./configure now finds
1316 the release number by grepping pcre.h.
1317
1318 26. Added the ability to run the tests under valgrind.
1319
1320
1321 Version 6.4 05-Sep-05
1322 ---------------------
1323
1324 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1325 "--" to be printed when multiple files were scanned, even when none of the
1326 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1327 consider it to be a bug, and have restored the previous behaviour.
1328
1329 2. A couple of code tidies to get rid of compiler warnings.
1330
1331 3. The pcretest program used to cheat by referring to symbols in the library
1332 whose names begin with _pcre_. These are internal symbols that are not
1333 really supposed to be visible externally, and in some environments it is
1334 possible to suppress them. The cheating is now confined to including
1335 certain files from the library's source, which is a bit cleaner.
1336
1337 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1338 file's purpose clearer.
1339
1340 5. Reorganized pcre_ucp_findchar().
1341
1342
1343 Version 6.3 15-Aug-05
1344 ---------------------
1345
1346 1. The file libpcre.pc.in did not have general read permission in the tarball.
1347
1348 2. There were some problems when building without C++ support:
1349
1350 (a) If C++ support was not built, "make install" and "make test" still
1351 tried to test it.
1352
1353 (b) There were problems when the value of CXX was explicitly set. Some
1354 changes have been made to try to fix these, and ...
1355
1356 (c) --disable-cpp can now be used to explicitly disable C++ support.
1357
1358 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1359 backslash in a target when C++ was disabled. This confuses some
1360 versions of "make", apparently. Using an intermediate variable solves
1361 this. (Same for CPP_LOBJ.)
1362
1363 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1364 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1365 necessary on certain architectures.
1366
1367 4. Added a setting of -export-symbols-regex to the link command to remove
1368 those symbols that are exported in the C sense, but actually are local
1369 within the library, and not documented. Their names all begin with
1370 "_pcre_". This is not a perfect job, because (a) we have to except some
1371 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1372 available (and never for static libraries). I have made a note to try to
1373 find a way round (a) in the future.
1374
1375
1376 Version 6.2 01-Aug-05
1377 ---------------------
1378
1379 1. There was no test for integer overflow of quantifier values. A construction
1380 such as {1111111111111111} would give undefined results. What is worse, if
1381 a minimum quantifier for a parenthesized subpattern overflowed and became
1382 negative, the calculation of the memory size went wrong. This could have
1383 led to memory overwriting.
1384
1385 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1386
1387 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1388 operating environments where this matters.
1389
1390 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1391 PCRE options from within the C++ wrapper.
1392
1393 5. Named capturing subpatterns were not being correctly counted when a pattern
1394 was compiled. This caused two problems: (a) If there were more than 100
1395 such subpatterns, the calculation of the memory needed for the whole
1396 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1397 back references of the form \12, where the number was greater than 9, were
1398 not recognized as back references, even though there were sufficient
1399 previous subpatterns.
1400
1401 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1402 versions of gcc, e.g. 2.95.4.
1403
1404
1405 Version 6.1 21-Jun-05
1406 ---------------------
1407
1408 1. There was one reference to the variable "posix" in pcretest.c that was not
1409 surrounded by "#if !defined NOPOSIX".
1410
1411 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1412 the cross-check on the old pcre_info() function, for the benefit of the
1413 cut-down version of PCRE that is currently imported into Exim.
1414
1415 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1416 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1417 patterns that start (?i) but I don't think that matters. The use of (?i) is
1418 just an example; this all applies to the other options as well.
1419
1420 4. Since libtool seems to echo the compile commands it is issuing, the output
1421 from "make" can be reduced a bit by putting "@" in front of each libtool
1422 compile command.
1423
1424 5. Patch from the folks at Google for configure.in to be a bit more thorough
1425 in checking for a suitable C++ installation before trying to compile the
1426 C++ stuff. This should fix a reported problem when a compiler was present,
1427 but no suitable headers.
1428
1429 6. The man pages all had just "PCRE" as their title. I have changed them to
1430 be the relevant file name. I have also arranged that these names are
1431 retained in the file doc/pcre.txt, which is a concatenation in text format
1432 of all the man pages except the little individual ones for each function.
1433
1434 7. The NON-UNIX-USE file had not been updated for the different set of source
1435 files that come with release 6. I also added a few comments about the C++
1436 wrapper.
1437
1438
1439 Version 6.0 07-Jun-05
1440 ---------------------
1441
1442 1. Some minor internal re-organization to help with my DFA experiments.
1443
1444 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1445 didn't matter for the library itself when fully configured, but did matter
1446 when compiling without UCP support, or within Exim, where the ucp files are
1447 not imported.
1448
1449 3. Refactoring of the library code to split up the various functions into
1450 different source modules. The addition of the new DFA matching code (see
1451 below) to a single monolithic source would have made it really too
1452 unwieldy, quite apart from causing all the code to be include in a
1453 statically linked application, when only some functions are used. This is
1454 relevant even without the DFA addition now that patterns can be compiled in
1455 one application and matched in another.
1456
1457 The downside of splitting up is that there have to be some external
1458 functions and data tables that are used internally in different modules of
1459 the library but which are not part of the API. These have all had their
1460 names changed to start with "_pcre_" so that they are unlikely to clash
1461 with other external names.
1462
1463 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1464 a different (DFA) algorithm. Although it is slower than the original
1465 function, it does have some advantages for certain types of matching
1466 problem.
1467
1468 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1469 including restarting after a partial match.
1470
1471 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1472 defined when compiling for Windows was sent to me. I have put it into the
1473 code, though I have no means of testing or verifying it.
1474
1475 7. Added the pcre_refcount() auxiliary function.
1476
1477 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1478 match before or at the first newline in the subject string. In pcretest,
1479 the /f option on a pattern can be used to set this.
1480
1481 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1482 would behave wrongly. This has been present in PCRE since release 4.0.
1483
1484 10. A number of changes to the pcregrep command:
1485
1486 (a) Refactored how -x works; insert ^(...)$ instead of setting
1487 PCRE_ANCHORED and checking the length, in preparation for adding
1488 something similar for -w.
1489
1490 (b) Added the -w (match as a word) option.
1491
1492 (c) Refactored the way lines are read and buffered so as to have more
1493 than one at a time available.
1494
1495 (d) Implemented a pcregrep test script.
1496
1497 (e) Added the -M (multiline match) option. This allows patterns to match
1498 over several lines of the subject. The buffering ensures that at least
1499 8K, or the rest of the document (whichever is the shorter) is available
1500 for matching (and similarly the previous 8K for lookbehind assertions).
1501
1502 (f) Changed the --help output so that it now says
1503
1504 -w, --word-regex(p)
1505
1506 instead of two lines, one with "regex" and the other with "regexp"
1507 because that confused at least one person since the short forms are the
1508 same. (This required a bit of code, as the output is generated
1509 automatically from a table. It wasn't just a text change.)
1510
1511 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1512 option but starts with a hyphen. Could be a pattern or a path name
1513 starting with a hyphen, for instance.
1514
1515 (h) "-" can be given as a file name to represent stdin.
1516
1517 (i) When file names are being printed, "(standard input)" is used for
1518 the standard input, for compatibility with GNU grep. Previously
1519 "<stdin>" was used.
1520
1521 (j) The option --label=xxx can be used to supply a name to be used for
1522 stdin when file names are being printed. There is no short form.
1523
1524 (k) Re-factored the options decoding logic because we are going to add
1525 two more options that take data. Such options can now be given in four
1526 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1527
1528 (l) Added the -A, -B, and -C options for requesting that lines of context
1529 around matches be printed.
1530
1531 (m) Added the -L option to print the names of files that do not contain
1532 any matching lines, that is, the complement of -l.
1533
1534 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1535 continue to scan other files.
1536
1537 (o) The -s option was incorrectly implemented. For compatibility with other
1538 greps, it now suppresses the error message for a non-existent or non-
1539 accessible file (but not the return code). There is a new option called
1540 -q that suppresses the output of matching lines, which was what -s was
1541 previously doing.
1542
1543 (p) Added --include and --exclude options to specify files for inclusion
1544 and exclusion when recursing.
1545
1546 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1547 Hopefully, it now does.
1548
1549 12. Missing cast in pcre_study().
1550
1551 13. Added an "uninstall" target to the makefile.
1552
1553 14. Replaced "extern" in the function prototypes in Makefile.in with
1554 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1555 world, but is set differently for Windows.
1556
1557 15. Added a second compiling function called pcre_compile2(). The only
1558 difference is that it has an extra argument, which is a pointer to an
1559 integer error code. When there is a compile-time failure, this is set
1560 non-zero, in addition to the error test pointer being set to point to an
1561 error message. The new argument may be NULL if no error number is required
1562 (but then you may as well call pcre_compile(), which is now just a
1563 wrapper). This facility is provided because some applications need a
1564 numeric error indication, but it has also enabled me to tidy up the way
1565 compile-time errors are handled in the POSIX wrapper.
1566
1567 16. Added VPATH=.libs to the makefile; this should help when building with one
1568 prefix path and installing with another. (Or so I'm told by someone who
1569 knows more about this stuff than I do.)
1570
1571 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1572 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1573 match everything, including newlines. This is not POSIX-compatible, but
1574 somebody wanted the feature. From pcretest it can be activated by using
1575 both the P and the s flags.
1576
1577 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1578
1579 19. libpcre.pc was being incorrectly installed as executable.
1580
1581 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1582 it now also looks for '\r' so that it will work unmodified on Windows.
1583
1584 21. Added Google's contributed C++ wrapper to the distribution.
1585
1586 22. Added some untidy missing memory free() calls in pcretest, to keep
1587 Electric Fence happy when testing.
1588
1589
1590
1591 Version 5.0 13-Sep-04
1592 ---------------------
1593
1594 1. Internal change: literal characters are no longer packed up into items
1595 containing multiple characters in a single byte-string. Each character
1596 is now matched using a separate opcode. However, there may be more than one
1597 byte in the character in UTF-8 mode.
1598
1599 2. The pcre_callout_block structure has two new fields: pattern_position and
1600 next_item_length. These contain the offset in the pattern to the next match
1601 item, and its length, respectively.
1602
1603 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1604 insertion of callouts before each pattern item. Added the /C option to
1605 pcretest to make use of this.
1606
1607 4. On the advice of a Windows user, the lines
1608
1609 #if defined(_WIN32) || defined(WIN32)
1610 _setmode( _fileno( stdout ), 0x8000 );
1611 #endif /* defined(_WIN32) || defined(WIN32) */
1612
1613 have been added to the source of pcretest. This apparently does useful
1614 magic in relation to line terminators.
1615
1616 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1617 for the benefit of those environments where the "b" makes a difference.
1618
1619 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1620 to know about it. I have put a hack into configure.in that adds in code
1621 to set GCC=yes if CC=icc. This seems to end up at a point in the
1622 generated configure script that is early enough to affect the setting of
1623 compiler options, which is what is needed, but I have no means of testing
1624 whether it really works. (The user who reported this had patched the
1625 generated configure script, which of course I cannot do.)
1626
1627 LATER: After change 22 below (new libtool files), the configure script
1628 seems to know about icc (and also ecc). Therefore, I have commented out
1629 this hack in configure.in.
1630
1631 7. Added support for pkg-config (2 patches were sent in).
1632
1633 8. Negated POSIX character classes that used a combination of internal tables
1634 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1635 [[:^ascii]]. Typically, they would match almost any characters. The other
1636 POSIX classes were not broken in this way.
1637
1638 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1639 to find the match, as PCRE was deluded into thinking that the match had to
1640 start at the start point or following a newline. The same bug applied to
1641 patterns with negative forward assertions or any backward assertions
1642 preceding ".*" at the start, unless the pattern required a fixed first
1643 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1644
1645 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1646 starting at the last subject character, bytes beyond the end of the subject
1647 string were read.
1648
1649 11. Renamed the variable "class" as "classbits" to make life easier for C++
1650 users. (Previously there was a macro definition, but it apparently wasn't
1651 enough.)
1652
1653 12. Added the new field "tables" to the extra data so that tables can be passed
1654 in at exec time, or the internal tables can be re-selected. This allows
1655 a compiled regex to be saved and re-used at a later time by a different
1656 program that might have everything at different addresses.
1657
1658 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1659 -R library as well as a -L library.
1660
1661 14. The debugging options of pcretest (-d on the command line or D on a
1662 pattern) showed incorrect output for anything following an extended class
1663 that contained multibyte characters and which was followed by a quantifier.
1664
1665 15. Added optional support for general category Unicode character properties
1666 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1667 support. It adds about 90K to the size of the library. The meanings of the
1668 inbuilt class escapes such as \d and \s have NOT been changed.
1669
1670 16. Updated pcredemo.c to include calls to free() to release the memory for the
1671 compiled pattern.
1672
1673 17. The generated file chartables.c was being created in the source directory
1674 instead of in the building directory. This caused the build to fail if the
1675 source directory was different from the building directory, and was
1676 read-only.
1677
1678 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1679 file. No doubt somebody will tell me if they don't make sense... Also added
1680 Dan Mooney's comments about building on OpenVMS.
1681
1682 19. Added support for partial matching via the PCRE_PARTIAL option for
1683 pcre_exec() and the \P data escape in pcretest.
1684
1685 20. Extended pcretest with 3 new pattern features:
1686
1687 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1688 write the compiled pattern to the file whose name is "rest-of-line".
1689 This is a straight binary dump of the data, with the saved pointer to
1690 the character tables forced to be NULL. The study data, if any, is
1691 written too. After writing, pcretest reads a new pattern.
1692
1693 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1694 compiled pattern from the given file. There must not be any
1695 occurrences of "<" in the file name (pretty unlikely); if there are,
1696 pcretest will instead treat the initial "<" as a pattern delimiter.
1697 After reading in the pattern, pcretest goes on to read data lines as
1698 usual.
1699
1700 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1701 and 16-bit fields in a compiled pattern, to simulate a pattern that
1702 was compiled on a host of opposite endianness.
1703
1704 21. The pcre-exec() function can now cope with patterns that were compiled on
1705 hosts of opposite endianness, with this restriction:
1706
1707 As for any compiled expression that is saved and used later, the tables
1708 pointer field cannot be preserved; the extra_data field in the arguments
1709 to pcre_exec() should be used to pass in a tables address if a value
1710 other than the default internal tables were used at compile time.
1711
1712 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1713 now diagnosed as an error. Previously, most of the time, a negative number
1714 would have been treated as zero, but if in addition "ovector" was passed as
1715 NULL, a crash could occur.
1716
1717 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1718 new versions from the libtool 1.5 distribution (the last one is a copy of
1719 a file called libtool.m4). This seems to have fixed the need to patch
1720 "configure" to support Darwin 1.3 (which I used to do). However, I still
1721 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1722 workstation).
1723
1724 24. Changed the PCRE licence to be the more standard "BSD" licence.
1725
1726
1727 Version 4.5 01-Dec-03
1728 ---------------------
1729
1730 1. There has been some re-arrangement of the code for the match() function so
1731 that it can be compiled in a version that does not call itself recursively.
1732 Instead, it keeps those local variables that need separate instances for
1733 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1734 needs to "recurse". Keeping track of where control must go is done by means
1735 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1736 hide most of the details from the main code, and operates only if
1737 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1738 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1739 operating.
1740
1741 To make it easier for callers to provide specially tailored get/free
1742 functions for this usage, two new functions, pcre_stack_malloc, and
1743 pcre_stack_free, are used. They are always called in strict stacking order,
1744 and the size of block requested is always the same.
1745
1746 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1747 PCRE has been compiled to use the stack or the heap for recursion. The
1748 -C option of pcretest uses this to show which version is compiled.
1749
1750 A new data escape \S, is added to pcretest; it causes the amounts of store
1751 obtained and freed by both kinds of malloc/free at match time to be added
1752 to the output.
1753
1754 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1755 what's available on my current Linux desktop machine.
1756
1757 3. When matching a UTF-8 string, the test for a valid string at the start has
1758 been extended. If start_offset is not zero, PCRE now checks that it points
1759 to a byte that is the start of a UTF-8 character. If not, it returns
1760 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1761 this is necessary because there may be backward assertions in the pattern.
1762 When matching the same subject several times, it may save resources to use
1763 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1764
1765 4. The code for checking the validity of UTF-8 strings has been tightened so
1766 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1767 containing "overlong sequences".
1768
1769 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1770 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1771 should have been "&", but it just so happened that all the cases this let
1772 through by mistake were picked up later in the function.
1773
1774 6. I had used a variable called "isblank" - this is a C99 function, causing
1775 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1776
1777 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1778 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1779 so the version is shown; (c) stop "make test" echoing "./RunTest".
1780
1781 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1782
1783 9. The prototype for memmove() for systems that don't have it was using
1784 size_t, but the inclusion of the header that defines size_t was later. I've
1785 moved the #includes for the C headers earlier to avoid this.
1786
1787 10. Added some adjustments to the code to make it easier to compiler on certain
1788 special systems:
1789
1790 (a) Some "const" qualifiers were missing.
1791 (b) Added the macro EXPORT before all exported functions; by default this
1792 is defined to be empty.
1793 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1794 that it reads its output file name as an argument instead of writing
1795 to the standard output and assuming this can be redirected.
1796
1797 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1798 class containing characters with values greater than 255, PCRE compilation
1799 went into a loop.
1800
1801 12. A recursive reference to a subpattern that was within another subpattern
1802 that had a minimum quantifier of zero caused PCRE to crash. For example,
1803 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1804 recursion. If the recursively-called subpattern itself had a zero repeat,
1805 that was OK.
1806
1807 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1808 buffer into which it was copied (for escape processing) was still set at
1809 1024, so long lines caused crashes.
1810
1811 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1812 "internal error: code overflow...". This applied to any character class
1813 that was followed by a possessive quantifier.
1814
1815 15. Modified the Makefile to add libpcre.la as a prerequisite for
1816 libpcreposix.la because I was told this is needed for a parallel build to
1817 work.
1818
1819 16. If a pattern that contained .* following optional items at the start was
1820 studied, the wrong optimizing data was generated, leading to matching
1821 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1822 matching string must start with a or b or c. The correct conclusion for
1823 this pattern is that a match can start with any character.
1824
1825
1826 Version 4.4 13-Aug-03
1827 ---------------------
1828
1829 1. In UTF-8 mode, a character class containing characters with values between
1830 127 and 255 was not handled correctly if the compiled pattern was studied.
1831 In fixing this, I have also improved the studying algorithm for such
1832 classes (slightly).
1833
1834 2. Three internal functions had redundant arguments passed to them. Removal
1835 might give a very teeny performance improvement.
1836
1837 3. Documentation bug: the value of the capture_top field in a callout is *one
1838 more than* the number of the hightest numbered captured substring.
1839
1840 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1841 in incorrectly linking with a previously installed version. They now link
1842 explicitly with libpcre.la.
1843
1844 5. configure.in no longer needs to recognize Cygwin specially.
1845
1846 6. A problem in pcre.in for Windows platforms is fixed.
1847
1848 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1849 pcretest, it used to include the size of the study block as part of its
1850 output. Unfortunately, the structure contains a field that has a different
1851 size on different hardware architectures. This meant that the tests that
1852 showed this size failed. As the block is currently always of a fixed size,
1853 this information isn't actually particularly useful in pcretest output, so
1854 I have just removed it.
1855
1856 8. Three pre-processor statements accidentally did not start in column 1.
1857 Sadly, there are *still* compilers around that complain, even though
1858 standard C has not required this for well over a decade. Sigh.
1859
1860 9. In pcretest, the code for checking callouts passed small integers in the
1861 callout_data field, which is a void * field. However, some picky compilers
1862 complained about the casts involved for this on 64-bit systems. Now
1863 pcretest passes the address of the small integer instead, which should get
1864 rid of the warnings.
1865
1866 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1867 both compile and run time, and gives an error if an invalid UTF-8 sequence
1868 is found. There is a option for disabling this check in cases where the
1869 string is known to be correct and/or the maximum performance is wanted.
1870
1871 11. In response to a bug report, I changed one line in Makefile.in from
1872
1873 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1874 to
1875 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1876
1877 to look similar to other lines, but I have no way of telling whether this
1878 is the right thing to do, as I do not use Windows. No doubt I'll get told
1879 if it's wrong...
1880
1881
1882 Version 4.3 21-May-03
1883 ---------------------
1884
1885 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1886 Makefile.
1887
1888 2. Some refactoring to improve the quality of the code:
1889
1890 (i) The utf8_table... variables are now declared "const".
1891
1892 (ii) The code for \cx, which used the "case flipping" table to upper case
1893 lower case letters, now just substracts 32. This is ASCII-specific,
1894 but the whole concept of \cx is ASCII-specific, so it seems
1895 reasonable.
1896
1897 (iii) PCRE was using its character types table to recognize decimal and
1898 hexadecimal digits in the pattern. This is silly, because it handles
1899 only 0-9, a-f, and A-F, but the character types table is locale-
1900 specific, which means strange things might happen. A private
1901 table is now used for this - though it costs 256 bytes, a table is
1902 much faster than multiple explicit tests. Of course, the standard
1903 character types table is still used for matching digits in subject
1904 strings against \d.
1905
1906 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1907 ending in _t are). So I've renamed it as ESC_tee.
1908
1909 3. The first argument for regexec() in the POSIX wrapper should have been
1910 defined as "const".
1911
1912 4. Changed pcretest to use malloc() for its buffers so that they can be
1913 Electric Fenced for debugging.
1914
1915 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1916 to read one or more bytes before the start of the subject string. Often this
1917 had no effect on PCRE's behaviour, but in some circumstances it could
1918 provoke a segmentation fault.
1919
1920 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1921 to try to read one or more bytes before the start of the subject string.
1922
1923 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1924 UTF-8 support could misbehave in various ways if the subject string
1925 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1926 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1927 back over UTF-8 characters.)
1928
1929
1930 Version 4.2 14-Apr-03
1931 ---------------------
1932
1933 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1934
1935 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1936 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1937 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1938 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1939 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1940 and BUILD_EXEEXT
1941 Note: automatic setting of the BUILD variables is not yet working
1942 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1943 compile-time but not at link-time
1944 [LINK]: use for linking executables only
1945 make different versions for Windows and non-Windows
1946 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1947 libraries
1948 [LINK_FOR_BUILD]: new variable
1949 [OBJEXT]: use throughout
1950 [EXEEXT]: use throughout
1951 <winshared>: new target
1952 <wininstall>: new target
1953 <dftables.o>: use native compiler
1954 <dftables>: use native linker
1955 <install>: handle Windows platform correctly
1956 <clean>: ditto
1957 <check>: ditto
1958 copy DLL to top builddir before testing
1959
1960 As part of these changes, -no-undefined was removed again. This was reported
1961 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1962 in any case.
1963
1964 3. Some tidies to get rid of compiler warnings:
1965
1966 . In the match_data structure, match_limit was an unsigned long int, whereas
1967 match_call_count was an int. I've made them both unsigned long ints.
1968
1969 . In pcretest the fact that a const uschar * doesn't automatically cast to
1970 a void * provoked a warning.
1971
1972 . Turning on some more compiler warnings threw up some "shadow" variables
1973 and a few more missing casts.
1974
1975 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1976 option, a class that contained a single character with a value between 128
1977 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1978
1979 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1980 option, a class that contained several characters, but with at least one
1981 whose value was between 128 and 255 caused PCRE to crash.
1982
1983
1984 Version 4.1 12-Mar-03
1985 ---------------------
1986
1987 1. Compiling with gcc -pedantic found a couple of places where casts were
1988 needed, and a string in dftables.c that was longer than standard compilers are
1989 required to support.
1990
1991 2. Compiling with Sun's compiler found a few more places where the code could
1992 be tidied up in order to avoid warnings.
1993
1994 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1995 first of these names is deprecated in the latest Autoconf in favour of the name
1996 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1997 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1998 analogy I have changed it to CFLAGS_FOR_BUILD.
1999
2000 4. Added -no-undefined to the linking command in the Makefile, because this is
2001 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2002 linking step for the pcreposix library.
2003
2004 5. PCRE was failing to diagnose the case of two named groups with the same
2005 name.
2006
2007 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2008 literal character that is needed in the subject for a match, and scans along to
2009 ensure that it is present before embarking on the full matching process. This
2010 saves time in cases of nested unlimited repeats that are never going to match.
2011 Problem: the scan can take a lot of time if the subject is very long (e.g.
2012 megabytes), thus penalizing straightforward matches. It is now done only if the
2013 amount of subject to be scanned is less than 1000 bytes.
2014
2015 7. A lesser problem with the same optimization is that it was recording the
2016 first character of an anchored pattern as "needed", thus provoking a search
2017 right along the subject, even when the first match of the pattern was going to
2018 fail. The "needed" character is now not set for anchored patterns, unless it
2019 follows something in the pattern that is of non-fixed length. Thus, it still
2020 fulfils its original purpose of finding quick non-matches in cases of nested
2021 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2022
2023
2024 Version 4.0 17-Feb-03
2025 ---------------------
2026
2027 1. If a comment in an extended regex that started immediately after a meta-item
2028 extended to the end of string, PCRE compiled incorrect data. This could lead to
2029 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2030
2031 2. Moved to autoconf 2.53 and libtool 1.4.2.
2032
2033 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2034 the special perltest8 script is no longer needed - all the tests can be run
2035 from a single perltest script.
2036
2037 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2038 by \s. It has now been removed in PCRE. This means it isn't recognized as
2039 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2040 class [:space:] *does* include VT, thereby creating a mess.
2041
2042 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2043 space and tab.
2044
2045 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2046 its new features into the main test script, reducing the number of scripts.
2047
2048 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2049 were backward compatible, and made the (?i) apply to the whole pattern, as if
2050 /i were given. Now it behaves more logically, and applies the option setting
2051 only to what follows. PCRE has been changed to follow suit. However, if it
2052 finds options settings right at the start of the pattern, it extracts them into
2053 the global options, as before. Thus, they show up in the info data.
2054
2055 8. Added support for the \Q...\E escape sequence. Characters in between are
2056 treated as literals. This is slightly different from Perl in that $ and @ are
2057 also handled as literals inside the quotes. In Perl, they will cause variable
2058 interpolation. Note the following examples:
2059
2060 Pattern PCRE matches Perl matches
2061
2062 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2063 \Qabc\$xyz\E abc\$xyz abc\$xyz
2064 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2065
2066 For compatibility with Perl, \Q...\E sequences are recognized inside character
2067 classes as well as outside them.
2068
2069 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2070 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2071 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2072 signed/unsigned warnings.
2073
2074 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2075 option for pcretest, so I've replaced it by a simple function that does just
2076 that job.
2077
2078 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2079 "pcregrep -".
2080
2081 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2082 Java package. This provides some syntactic sugar for simple cases of what my
2083 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2084 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2085 item, you can use this simplified notation. Note that only makes sense with
2086 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2087 greediness, whatever the setting of the PCRE_UNGREEDY option.
2088
2089 13. A change of greediness default within a pattern was not taking effect at
2090 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2091 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2092 was abstracted outside.
2093
2094 14. PCRE now supports the \G assertion. It is true when the current matching
2095 position is at the start point of the match. This differs from \A when the
2096 starting offset is non-zero. Used with the /g option of pcretest (or similar
2097 code), it works in the same way as it does for Perl's /g option. If all
2098 alternatives of a regex begin with \G, the expression is anchored to the start
2099 match position, and the "anchored" flag is set in the compiled expression.
2100
2101 15. Some bugs concerning the handling of certain option changes within patterns
2102 have been fixed. These applied to options other than (?ims). For example,
2103 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2104 been the other way round. Some of this was related to change 7 above.
2105
2106 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2107 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2108 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2109 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2110
2111 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2112 mode. Unlike ".", it always matches newline, whatever the setting of
2113 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2114 assertions. Perl allows it, but it doesn't (in general) work because it can't
2115 calculate the length of the lookbehind. At least, that's the case for Perl
2116 5.8.0 - I've been told they are going to document that it doesn't work in
2117 future.
2118
2119 18. Added an error diagnosis for escapes that PCRE does not support: these are
2120 \L, \l, \N, \P, \p, \U, \u, and \X.
2121
2122 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2123 reading past the end of the pattern in cases such as /[abcd/.
2124
2125 20. PCRE was getting more memory than necessary for patterns with classes that
2126 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2127
2128 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2129 compiling PCRE for use with Virtual Pascal.
2130
2131 22. Small fix to the Makefile to make it work properly if the build is done
2132 outside the source tree.
2133
2134 23. Added a new extension: a condition to go with recursion. If a conditional
2135 subpattern starts with (?(R) the "true" branch is used if recursion has
2136 happened, whereas the "false" branch is used only at the top level.
2137
2138 24. When there was a very long string of literal characters (over 255 bytes
2139 without UTF support, over 250 bytes with UTF support), the computation of how
2140 much memory was required could be incorrect, leading to segfaults or other
2141 strange effects.
2142
2143 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2144 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2145 there was a subsequent back reference to those brackets. This meant that, for
2146 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2147 possible to check for precisely this case. All we can do is abandon the
2148 optimization if .* occurs inside capturing brackets when there are any back
2149 references whatsoever. (See below for a better fix that came later.)
2150
2151 26. The handling of the optimization for finding the first character of a
2152 non-anchored pattern, and for finding a character that is required later in the
2153 match were failing in some cases. This didn't break the matching; it just
2154 failed to optimize when it could. The way this is done has been re-implemented.
2155
2156 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2157
2158 28. Added a new feature that provides some of the functionality that Perl
2159 provides with (?{...}). The facility is termed a "callout". The way it is done
2160 in PCRE is for the caller to provide an optional function, by setting
2161 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2162 global variable. By default it is unset, which disables all calling out. To get
2163 the function called, the regex must include (?C) at appropriate points. This
2164 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2165 This provides a means of identifying different callout points. When PCRE
2166 reaches such a point in the regex, if pcre_callout has been set, the external
2167 function is called. It is provided with data in a structure called
2168 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2169 matching continues; if it returns a non-zero value, the match at the current
2170 point fails. However, backtracking will occur if possible. [This was changed
2171 later and other features added - see item 49 below.]
2172
2173 29. pcretest is upgraded to test the callout functionality. It provides a
2174 callout function that displays information. By default, it shows the start of
2175 the match and the current position in the text. There are some new data escapes
2176 to vary what happens:
2177
2178 \C+ in addition, show current contents of captured substrings
2179 \C- do not supply a callout function
2180 \C!n return 1 when callout number n is reached
2181 \C!n!m return 1 when callout number n is reached for the mth time
2182
2183 30. If pcregrep was called with the -l option and just a single file name, it
2184 output "<stdin>" if a match was found, instead of the file name.
2185
2186 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2187 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2188 pcre_exec(). This saves a malloc/free per call. The default value of
2189 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2190 when configuring.
2191
2192 32. The default maximum size of a compiled pattern is 64K. There have been a
2193 few cases of people hitting this limit. The code now uses macros to handle the
2194 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2195 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2196 configuring. Tests 2 and 5 work only with 2-byte links because they output
2197 debugging information about compiled patterns.
2198
2199 33. Internal code re-arrangements:
2200
2201 (a) Moved the debugging function for printing out a compiled regex into
2202 its own source file (printint.c) and used #include to pull it into
2203 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2204 separate copies.
2205
2206 (b) Defined the list of op-code names for debugging as a macro in
2207 internal.h so that it is next to the definition of the opcodes.
2208
2209 (c) Defined a table of op-code lengths for simpler skipping along compiled
2210 code. This is again a macro in internal.h so that it is next to the
2211 definition of the opcodes.
2212
2213 34. Added support for recursive calls to individual subpatterns, along the
2214 lines of Robin Houston's patch (but implemented somewhat differently).
2215
2216 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2217 allow it to read and process whole directories in Win32. This code was
2218 contributed by Lionel Fourquaux; it has not been tested by me.
2219
2220 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2221 used to name a group. Names consist of alphanumerics and underscores, and must
2222 be unique. Back references use the syntax (?P=name) and recursive calls use
2223 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2224 numbers. The function pcre_fullinfo() can be used after compilation to extract
2225 a name/number map. There are three relevant calls:
2226
2227 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2228 PCRE_INFO_NAMECOUNT yields the number of entries
2229 PCRE_INFO_NAMETABLE yields a pointer to the map.
2230
2231 The map is a vector of fixed-size entries. The size of each entry depends on
2232 the length of the longest name used. The first two bytes of each entry are the
2233 group number, most significant byte first. There follows the corresponding
2234 name, zero terminated. The names are in alphabetical order.
2235
2236 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2237 case instead of 255. Making it the same both with and without UTF-8 support
2238 means that the same test output works with both.
2239
2240 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2241 calling malloc() with a zero argument.
2242
2243 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2244 optimization. I've improved things by keeping a bitmap of backreferences with
2245 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2246 fact referenced, the optimization can be applied. It is unlikely that a
2247 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2248 the match to follow \n) will appear inside brackets with a number greater than
2249 31, but if it does, any back reference > 31 suppresses the optimization.
2250
2251 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2252 of disabling numbered capturing parentheses. Any opening parenthesis that is
2253 not followed by ? behaves as if it were followed by ?: but named parentheses
2254 can still be used for capturing (and they will acquire numbers in the usual
2255 way).
2256
2257 41. Redesigned the return codes from the match() function into yes/no/error so
2258 that errors can be passed back from deep inside the nested calls. A malloc
2259 failure while inside a recursive subpattern call now causes the
2260 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2261
2262 42. It is now possible to set a limit on the number of times the match()
2263 function is called in a call to pcre_exec(). This facility makes it possible to
2264 limit the amount of recursion and backtracking, though not in a directly
2265 obvious way, because the match() function is used in a number of different
2266 circumstances. The count starts from zero for each position in the subject
2267 string (for non-anchored patterns). The default limit is, for compatibility, a
2268 large number, namely 10 000 000. You can change this in two ways:
2269
2270 (a) When configuring PCRE before making, you can use --with-match-limit=n
2271 to set a default value for the compiled library.
2272
2273 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2274 a different value is set. See 45 below.
2275
2276 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2277
2278 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2279 of things that can be changed at compile time. The first argument specifies
2280 what is wanted and the second points to where the information is to be placed.
2281 The current list of available information is:
2282
2283 PCRE_CONFIG_UTF8
2284
2285 The output is an integer that is set to one if UTF-8 support is available;
2286 otherwise it is set to zero.
2287
2288 PCRE_CONFIG_NEWLINE
2289
2290 The output is an integer that it set to the value of the code that is used for
2291 newline. It is either LF (10) or CR (13).
2292
2293 PCRE_CONFIG_LINK_SIZE
2294
2295 The output is an integer that contains the number of bytes used for internal
2296 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2297
2298 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2299
2300 The output is an integer that contains the threshold above which the POSIX
2301 interface uses malloc() for output vectors. See item 31 above.
2302
2303 PCRE_CONFIG_MATCH_LIMIT
2304
2305 The output is an unsigned integer that contains the default limit of the number
2306 of match() calls in a pcre_exec() execution. See 42 above.
2307
2308 44. pcretest has been upgraded by the addition of the -C option. This causes it
2309 to extract all the available output from the new pcre_config() function, and to
2310 output it. The program then exits immediately.
2311
2312 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2313 order to support additional features. One way would have been to define
2314 pcre_exec2() (for example) with extra arguments, but this would not have been
2315 extensible, and would also have required all calls to the original function to
2316 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2317 is used for passing in "extra" data from pcre_study().
2318
2319 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2320 contains the following fields:
2321
2322 flags a bitmap indicating which of the following fields are set
2323 study_data opaque data from pcre_study()
2324 match_limit a way of specifying a limit on match() calls for a specific
2325 call to pcre_exec()
2326 callout_data data for callouts (see 49 below)
2327
2328 The flag bits are also defined in pcre.h, and are
2329
2330 PCRE_EXTRA_STUDY_DATA
2331 PCRE_EXTRA_MATCH_LIMIT
2332 PCRE_EXTRA_CALLOUT_DATA
2333
2334 The pcre_study() function now returns one of these new pcre_extra blocks, with
2335 the actual study data pointed to by the study_data field, and the
2336 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2337 before. That is, this change is entirely upwards-compatible and requires no
2338 change to existing code.
2339
2340 If you want to pass in additional data to pcre_exec(), you can either place it
2341 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2342 block.
2343
2344 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2345 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2346 times with different match limits, until it finds the minimum value needed for
2347 pcre_exec() to complete. The value is then output. This can be instructive; for
2348 most simple matches the number is quite small, but for pathological cases it
2349 gets very large very quickly.
2350
2351 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2352 returns the size of the data block pointed to by the study_data field in a
2353 pcre_extra block, that is, the value that was passed as the argument to
2354 pcre_malloc() when PCRE was getting memory in which to place the information
2355 created by pcre_study(). The fourth argument should point to a size_t variable.
2356 pcretest has been extended so that this information is shown after a successful
2357 pcre_study() call when information about the compiled regex is being displayed.
2358
2359 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2360 because what follows is always an absolute path. (Later: it turns out that this
2361 is more than cosmetic for MinGW, because it doesn't like empty path
2362 components.)
2363
2364 49. Some changes have been made to the callout feature (see 28 above):
2365
2366 (i) A callout function now has three choices for what it returns:
2367
2368 0 => success, carry on matching
2369 > 0 => failure at this point, but backtrack if possible
2370 < 0 => serious error, return this value from pcre_exec()
2371
2372 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2373 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2374 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2375 use by callout functions. It will never be used by PCRE itself.
2376
2377 (ii) The pcre_extra structure (see 45 above) has a void * field called
2378 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2379 pcre_callout_block structure has a field of the same name. The contents of
2380 the field passed in the pcre_extra structure are passed to the callout
2381 function in the corresponding field in the callout block. This makes it
2382 easier to use the same callout-containing regex from multiple threads. For
2383 testing, the pcretest program has a new data escape
2384
2385 \C*n pass the number n (may be negative) as callout_data
2386
2387 If the callout function in pcretest receives a non-zero value as
2388 callout_data, it returns that value.
2389
2390 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2391 there were some redundant $(CFLAGS) in commands that are now specified as
2392 $(LINK), which already includes $(CFLAGS).
2393
2394 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2395 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2396 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2397 one-byte characters throughout. Note that case-insensitive matching applies
2398 only to characters whose values are less than 256. PCRE doesn't support the
2399 notion of cases for higher-valued characters.
2400
2401 (i) A character class whose characters are all within 0-255 is handled as
2402 a bit map, and the map is inverted for negative classes. Previously, a
2403 character > 255 always failed to match such a class; however it should
2404 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2405
2406 (ii) A negated character class with a single character < 255 is coded as
2407 "not this character" (OP_NOT). This wasn't working properly when the test
2408 character was multibyte, either singly or repeated.
2409
2410 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2411 mode, for example: \x{100}{2,3}.
2412
2413 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2414 singly or repeated) now correctly test multibyte characters. However,
2415 PCRE doesn't recognize any characters with values greater than 255 as
2416 digits, spaces, or word characters. Such characters always match \D, \S,
2417 and \W, and never match \d, \s, or \w.
2418
2419 (v) Classes may now contain characters and character ranges with values
2420 greater than 255. For example: [ab\x{100}-\x{400}].
2421
2422 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2423 PCRE in UTF-8 mode.
2424
2425 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2426 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2427 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2428 value.)
2429
2430 53. The single man page has become too large. I have therefore split it up into
2431 a number of separate man pages. These also give rise to individual HTML pages;
2432 these are now put in a separate directory, and there is an index.html page that
2433 lists them all. Some hyperlinking between the pages has been installed.
2434
2435 54. Added convenience functions for handling named capturing parentheses.
2436
2437 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2438 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2439 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2440 are faulted.
2441
2442 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2443 calling configure. These values are used when compiling the dftables.c program
2444 which is run to generate the source of the default character tables. They
2445 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2446 you will need to set these values.
2447
2448 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2449
2450
2451 Version 3.9 02-Jan-02
2452 ---------------------
2453
2454 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2455
2456 2. If --disable-static was given, the building process failed when trying to
2457 build pcretest and pcregrep. (For some reason it was using libtool to compile
2458 them, which is not right, as they aren't part of the library.)
2459
2460
2461 Version 3.8 18-Dec-01
2462 ---------------------
2463
2464 1. The experimental UTF-8 code was completely screwed up. It was packing the
2465 bytes in the wrong order. How dumb can you get?
2466
2467
2468 Version 3.7 29-Oct-01
2469 ---------------------
2470
2471 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2472 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2473 this didn't happen under Solaris 8, where I normally test things.
2474
2475 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2476 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2477 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2478 this fix an hour or so after the initial 3.7 release.)
2479
2480
2481 Version 3.6 23-Oct-01
2482 ---------------------
2483
2484 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2485 offsets passed as NULL with zero offset count.
2486
2487 2. The config.guess and config.sub files had not been updated when I moved to
2488 the latest autoconf.
2489
2490
2491 Version 3.5 15-Aug-01
2492 ---------------------
2493
2494 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2495 had been forgotten.
2496
2497 2. By using declared but undefined structures, we can avoid using "void"
2498 definitions in pcre.h while keeping the internal definitions of the structures
2499 private.
2500
2501 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2502 user point of view, this means that both static and shared libraries are built
2503 by default, but this can be individually controlled. More of the work of
2504 handling this static/shared cases is now inside libtool instead of PCRE's make
2505 file.
2506
2507 4. The pcretest utility is now installed along with pcregrep because it is
2508 useful for users (to test regexs) and by doing this, it automatically gets
2509 relinked by libtool. The documentation has been turned into a man page, so
2510 there are now .1, .txt, and .html versions in /doc.
2511
2512 5. Upgrades to pcregrep:
2513 (i) Added long-form option names like gnu grep.
2514 (ii) Added --help to list all options with an explanatory phrase.
2515 (iii) Added -r, --recursive to recurse into sub-directories.
2516 (iv) Added -f, --file to read patterns from a file.
2517
2518 6. pcre_exec() was referring to its "code" argument before testing that
2519 argument for NULL (and giving an error if it was NULL).
2520
2521 7. Upgraded Makefile.in to allow for compiling in a different directory from
2522 the source directory.
2523
2524 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2525 options bits, the pointer it was passed was to an int instead of to an unsigned
2526 long int. This mattered only on 64-bit systems.
2527
2528 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2529 generated) instead of pcre.in, which it its source. Also made the same change
2530 in several of the .c files.
2531
2532 10. A new release of gcc defines printf() as a macro, which broke pcretest
2533 because it had an ifdef in the middle of a string argument for printf(). Fixed
2534 by using separate calls to printf().
2535
2536 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2537 script, to force use of CR or LF instead of \n in the source. On non-Unix
2538 systems, the value can be set in config.h.
2539
2540 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2541 absolute limit. Changed the text of the error message to make this clear, and
2542 likewise updated the man page.
2543
2544 13. The limit of 99 on the number of capturing subpatterns has been removed.
2545 The new limit is 65535, which I hope will not be a "real" limit.
2546
2547
2548 Version 3.4 22-Aug-00
2549 ---------------------
2550
2551 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2552
2553 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2554
2555
2556 Version 3.3 01-Aug-00
2557 ---------------------
2558
2559 1. If an octal character was given, but the value was greater than \377, it
2560 was not getting masked to the least significant bits, as documented. This could
2561 lead to crashes in some systems.
2562
2563 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2564 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2565
2566 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2567 These just pass their arguments on to (pcre_free)(), but they are provided
2568 because some uses of PCRE bind it to non-C systems that can call its functions,
2569 but cannot call free() or pcre_free() directly.
2570
2571 4. Add "make test" as a synonym for "make check". Corrected some comments in
2572 the Makefile.
2573
2574 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2575 Makefile.
2576
2577 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2578 command called pgrep for grepping around the active processes.
2579
2580 7. Added the beginnings of support for UTF-8 character strings.
2581
2582 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2583 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2584 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2585 out for the ar command.)
2586
2587
2588 Version 3.2 12-May-00
2589 ---------------------
2590
2591 This is purely a bug fixing release.
2592
2593 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2594 of ZA. This was just one example of several cases that could provoke this bug,
2595 which was introduced by change 9 of version 2.00. The code for breaking
2596 infinite loops after an iteration that matches an empty string was't working
2597 correctly.
2598
2599 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2600 when matched against abbab (for example). After matching an empty string, it
2601 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2602 caused it to match further down the string than it should.
2603
2604 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2605 was there because it doesn't seem to be needed, and it causes trouble on some
2606 systems, as it is not a Standard C header. It has been removed.
2607
2608 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2609 were reported on the Macintosh. The changes were from
2610
2611 while ((c = *(++ptr)) != 0 && c != '\n');
2612 to
2613 while ((c = *(++ptr)) != 0 && c != '\n') ;
2614
2615 Totally extraordinary, but if that's what it takes...
2616
2617 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2618 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2619 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2620 assumes the way PCRE uses memmove() (always moving upwards).
2621
2622 6. PCRE is being used in one environment where strchr() is not available. There
2623 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2624 faster code anyway.
2625
2626
2627 Version 3.1 09-Feb-00
2628 ---------------------
2629
2630 The only change in this release is the fixing of some bugs in Makefile.in for
2631 the "install" target:
2632
2633 (1) It was failing to install pcreposix.h.
2634
2635 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2636
2637
2638 Version 3.0 01-Feb-00
2639 ---------------------
2640
2641 1. Add support for the /+ modifier to perltest (to output $` like it does in
2642 pcretest).
2643
2644 2. Add support for the /g modifier to perltest.
2645
2646 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2647 matches null strings.
2648
2649 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2650 pattern. Perl treats empty patterns specially - it reuses the most recent
2651 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2652 effect.
2653
2654 5. The POSIX interface was broken in that it was just handing over the POSIX
2655 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2656 required a bigger vector, with some working space on the end. This means that
2657 the POSIX wrapper now has to get and free some memory, and copy the results.
2658
2659 6. Added some simple autoconf support, placing the test data and the
2660 documentation in separate directories, re-organizing some of the
2661 information files, and making it build pcre-config (a GNU standard). Also added
2662 libtool support for building PCRE as a shared library, which is now the
2663 default.
2664
2665 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2666 09 are not valid octal constants. Single digits will be used for minor values
2667 less than 10.
2668
2669 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2670 existing programs that set these in the POSIX interface can use PCRE without
2671 modification.
2672
2673 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2674 return all that pcre_info() returns, plus additional data. The pcre_info()
2675 function is retained for compatibility, but is considered to be obsolete.
2676
2677 10. Added experimental recursion feature (?R) to handle one common case that
2678 Perl 5.6 will be able to do with (?p{...}).
2679
2680 11. Added support for POSIX character classes like [:alpha:], which Perl is
2681 adopting.
2682
2683
2684 Version 2.08 31-Aug-99
2685 ----------------------
2686
2687 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2688 trying to match at the startoffset position, but instead was moving forward to
2689 the next newline as if a previous match had failed.
2690
2691 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2692 and could get into a loop if a null string was matched other than at the start
2693 of the subject.
2694
2695 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2696 be distinguished at compile time, and for completeness also added PCRE_DATE.
2697
2698 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2699 in GnuWin32 environments.
2700
2701
2702 Version 2.07 29-Jul-99
2703 ----------------------
2704
2705 1. The documentation is now supplied in plain text form and HTML as well as in
2706 the form of man page sources.
2707
2708 2. C++ compilers don't like assigning (void *) values to other pointer types.
2709 In particular this affects malloc(). Although there is no problem in Standard
2710 C, I've put in casts to keep C++ compilers happy.
2711
2712 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2713 should be (const char *).
2714
2715 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2716 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2717 However, I haven't made this a standard facility. The documentation doesn't
2718 mention it, and the Makefile doesn't support it.
2719
2720 5. The Makefile now contains an "install" target, with editable destinations at
2721 the top of the file. The pcretest program is not installed.
2722
2723 6. pgrep -V now gives the PCRE version number and date.
2724
2725 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2726 causing the entire string to be ignored, instead of just the last character.
2727
2728 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2729 non-matching string, it can take a very, very long time, even for strings of
2730 quite modest length, because of the nested recursion. PCRE now does better in
2731 some of these cases. It does this by remembering the last required literal
2732 character in the pattern, and pre-searching the subject to ensure it is present
2733 before running the real match. In other words, it applies a heuristic to detect
2734 some types of certain failure quickly, and in the above example, if presented
2735 with a string that has no trailing " it gives "no match" very quickly.
2736
2737 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2738 other alternatives are tried instead.
2739
2740
2741 Version 2.06 09-Jun-99
2742 ----------------------
2743
2744 1. Change pcretest's output for amount of store used to show just the code
2745 space, because the remainder (the data block) varies in size between 32-bit and
2746 64-bit systems.
2747
2748 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2749 start matching at. This allows lookbehinds to work when searching for multiple
2750 occurrences in a string.
2751
2752 3. Added additional options to pcretest for testing multiple occurrences:
2753
2754 /+ outputs the rest of the string that follows a match
2755 /g loops for multiple occurrences, using the new startoffset argument
2756 /G loops for multiple occurrences by passing an incremented pointer
2757
2758 4. PCRE wasn't doing the "first character" optimization for patterns starting
2759 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2760 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2761 the letter 'x'. On long subject strings, this gives a significant speed-up.
2762
2763
2764 Version 2.05 21-Apr-99
2765 ----------------------
2766
2767 1. Changed the type of magic_number from int to long int so that it works
2768 properly on 16-bit systems.
2769
2770 2. Fixed a bug which caused patterns starting with .* not to work correctly
2771 when the subject string contained newline characters. PCRE was assuming
2772 anchoring for such patterns in all cases, which is not correct because .* will
2773 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2774 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2775 must be retried after every newline in the subject.
2776
2777
2778 Version 2.04 18-Feb-99
2779 ----------------------
2780
2781 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2782 computation of the store needed to hold the pattern was incorrect (too large).
2783 If such patterns were nested a few deep, this could multiply and become a real
2784 problem.
2785
2786 2. Added /M option to pcretest to show the memory requirement of a specific
2787 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2788
2789 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2790 compiled in such a way that the backtracking after subsequent failure was
2791 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2792 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2793
2794
2795 Version 2.03 02-Feb-99
2796 ----------------------
2797
2798 1. Fixed typo and small mistake in man page.
2799
2800 2. Added 4th condition (GPL supersedes if conflict) and created separate
2801 LICENCE file containing the conditions.
2802
2803 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2804 Perl, that is the internal \ allows the delimiter to be included in the
2805 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2806 the final delimiter, add \ to the end of the pattern (to test the error).
2807
2808 4. Added the convenience functions for extracting substrings after a successful
2809 match. Updated pcretest to make it able to test these functions.
2810
2811
2812 Version 2.02 14-Jan-99
2813 ----------------------
2814
2815 1. Initialized the working variables associated with each extraction so that
2816 their saving and restoring doesn't refer to uninitialized store.
2817
2818 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2819 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2820 fix the problem.
2821
2822 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2823 calls, and wasn't printing the correct value for compiling calls. Increased the
2824 default value of LOOPREPEAT, and the number of significant figures in the
2825 times.
2826
2827 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2828
2829 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2830 a building problem on Windows NT with a FAT file system.
2831
2832
2833 Version 2.01 21-Oct-98
2834 ----------------------
2835
2836 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2837 to character tables built by pcre_maketables() in the current locale. If NULL
2838 is passed, the default tables are used.
2839
2840
2841 Version 2.00 24-Sep-98
2842 ----------------------
2843
2844 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2845 it any more.
2846
2847 2. Allow quantification of (?>) groups, and make it work correctly.
2848
2849 3. The first character computation wasn't working for (?>) groups.
2850
2851 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2852 end of the subject) and add 5.005's \z, which really does match only at the
2853 very end of the subject.
2854
2855 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2856
2857 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2858 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2859 localized options. All options to pcre_study() were also removed.
2860
2861 7. Add other new features from 5.005:
2862
2863 $(?<= positive lookbehind
2864 $(?<! negative lookbehind
2865 (?imsx-imsx) added the unsetting capability
2866 such a setting is global if at outer level; local otherwise
2867 (?imsx-imsx:) non-capturing groups with option setting
2868 (?(cond)re|re) conditional pattern matching
2869
2870 A backreference to itself in a repeated group matches the previous
2871 captured string.
2872
2873 8. General tidying up of studying (both automatic and via "study")
2874 consequential on the addition of new assertions.
2875
2876 9. As in 5.005, unlimited repeated groups that could match an empty substring
2877 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2878 runtime if any iteration does actually match an empty substring.
2879
2880 10. Include the RunTest script in the distribution.
2881
2882 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2883 discrepancies, some of which were old and were also with respect to 5.004. They
2884 have now been fixed.
2885
2886
2887 Version 1.09 28-Apr-98
2888 ----------------------
2889
2890 1. A negated single character class followed by a quantifier with a minimum
2891 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2892 program crashes, or just wrong answers. This did not apply to negated classes
2893 containing more than one character, or to minima other than one.
2894
2895
2896 Version 1.08 27-Mar-98
2897 ----------------------
2898
2899 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2900
2901 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2902 latter must appear before anything that relies on it in the pattern.
2903
2904
2905 Version 1.07 16-Feb-98
2906 ----------------------
2907
2908 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2909 repeat of a potentially empty string).
2910
2911
2912 Version 1.06 23-Jan-98
2913 ----------------------
2914
2915 1. Added Markus Oberhumer's little patches for C++.
2916
2917 2. Literal strings longer than 255 characters were broken.
2918
2919
2920 Version 1.05 23-Dec-97
2921 ----------------------
2922
2923 1. Negated character classes containing more than one character were failing if
2924 PCRE_CASELESS was set at run time.
2925
2926
2927 Version 1.04 19-Dec-97
2928 ----------------------
2929
2930 1. Corrected the man page, where some "const" qualifiers had been omitted.
2931
2932 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2933 input syntax.
2934
2935 3. Fixed memory leak which occurred when a regex with back references was
2936 matched with an offsets vector that wasn't big enough. The temporary memory
2937 that is used in this case wasn't being freed if the match failed.
2938
2939 4. Tidied pcretest to ensure it frees memory that it gets.
2940
2941 5. Temporary memory was being obtained in the case where the passed offsets
2942 vector was exactly big enough.
2943
2944 6. Corrected definition of offsetof() from change 5 below.
2945
2946 7. I had screwed up change 6 below and broken the rules for the use of
2947 setjmp(). Now fixed.
2948
2949
2950 Version 1.03 18-Dec-97
2951 ----------------------
2952
2953 1. A erroneous regex with a missing opening parenthesis was correctly
2954 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2955 on some systems.
2956
2957 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2958 it was reported that one broken compiler failed on the former because "code" is
2959 also an independent variable.
2960
2961 3. The erroneous regex a[]b caused an array overrun reference.
2962
2963 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2964 fail on data ending with that character. (It was going on too far, and checking
2965 the next character, typically a binary zero.) This was specific to the
2966 optimized code for single-character negative classes.
2967
2968 5. Added a contributed patch from the TIN world which does the following:
2969
2970 + Add an undef for memmove, in case the the system defines a macro for it.
2971
2972 + Add a definition of offsetof(), in case there isn't one. (I don't know
2973 the reason behind this - offsetof() is part of the ANSI standard - but
2974 it does no harm).
2975
2976 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2977 most of the places where whitespace preceded '#'. I have given up and
2978 allowed the remaining 2 cases to be at the margin.
2979
2980 + Rename some variables in pcre to eliminate shadowing. This seems very
2981 pedantic, but does no harm, of course.
2982
2983 6. Moved the call to setjmp() into its own function, to get rid of warnings
2984 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2985
2986 7. Constructs such as \d{8,} were compiling into the equivalent of
2987 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2988 outcome, but in this particular case used more store than had been allocated,
2989 which caused the bug to be discovered because it threw up an internal error.
2990
2991 8. The debugging code in both pcre and pcretest for outputting the compiled
2992 form of a regex was going wrong in the case of back references followed by
2993 curly-bracketed repeats.
2994
2995
2996 Version 1.02 12-Dec-97
2997 ----------------------
2998
2999 1. Typos in pcre.3 and comments in the source fixed.
3000
3001 2. Applied a contributed patch to get rid of places where it used to remove
3002 'const' from variables, and fixed some signed/unsigned and uninitialized
3003 variable warnings.
3004
3005 3. Added the "runtest" target to Makefile.
3006
3007 4. Set default compiler flag to -O2 rather than just -O.
3008
3009
3010 Version 1.01 19-Nov-97
3011 ----------------------
3012
3013 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3014 like /([ab]*)*/, that is, for classes with more than one character in them.
3015
3016 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3017 as /((?>a*))*/ (a PCRE_EXTRA facility).
3018
3019
3020 Version 1.00 18-Nov-97
3021 ----------------------
3022
3023 1. Added compile-time macros to support systems such as SunOS4 which don't have
3024 memmove() or strerror() but have other things that can be used instead.
3025
3026 2. Arranged that "make clean" removes the executables.
3027
3028
3029 Version 0.99 27-Oct-97
3030 ----------------------
3031
3032 1. Fixed bug in code for optimizing classes with only one character. It was
3033 initializing a 32-byte map regardless, which could cause it to run off the end
3034 of the memory it had got.
3035
3036 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3037
3038
3039 Version 0.98 22-Oct-97
3040 ----------------------
3041
3042 1. Fixed bug in code for handling temporary memory usage when there are more
3043 back references than supplied space in the ovector. This could cause segfaults.
3044
3045
3046 Version 0.97 21-Oct-97
3047 ----------------------
3048
3049 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3050
3051 2. Optimized negated single characters not to use a bit map.
3052
3053 3. Brought error texts together as macro definitions; clarified some of them;
3054 fixed one that was wrong - it said "range out of order" when it meant "invalid
3055 escape sequence".
3056
3057 4. Changed some char * arguments to const char *.
3058
3059 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3060
3061 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3062 pcretest.
3063
3064
3065 Version 0.96 16-Oct-97
3066 ----------------------
3067
3068 1. Added a simple "pgrep" utility to the distribution.
3069
3070 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3071 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3072 where "ddd" means "one or more decimal digits".
3073
3074 3. Fixed serious bug. If a pattern had a back reference, but the call to
3075 pcre_exec() didn't supply a large enough ovector to record the related
3076 identifying subpattern, the match always failed. PCRE now remembers the number
3077 of the largest back reference, and gets some temporary memory in which to save
3078 the offsets during matching if necessary, in order to ensure that
3079 backreferences always work.
3080
3081 4. Increased the compatibility with Perl in a number of ways:
3082
3083 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3084 to request this handling. The option can be set at compile or exec time.
3085
3086 (b) $ matches before a terminating newline by default; an option
3087 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3088 mode). The option can be set at compile or exec time.
3089
3090 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3091 the same as Perl's. If the decimal number it represents is less than 10
3092 or there aren't that many previous left capturing parentheses, an octal
3093 escape is read. Inside a character class, it's always an octal escape,
3094 even if it is a single digit.
3095
3096 (d) An escaped but undefined alphabetic character is taken as a literal,
3097 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3098 escapes.
3099
3100 (e) {0} is now permitted. (The previous item is removed from the compiled
3101 pattern).
3102
3103 5. Changed all the names of code files so that the basic parts are no longer
3104 than 10 characters, and abolished the teeny "globals.c" file.
3105
3106 6. Changed the handling of character classes; they are now done with a 32-byte
3107 bit map always.
3108
3109 7. Added the -d and /D options to pcretest to make it possible to look at the
3110 internals of compilation without having to recompile pcre.
3111
3112
3113 Version 0.95 23-Sep-97
3114 ----------------------
3115
3116 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3117 \x20 at the start of a run of normal characters. These were being treated as
3118 real characters, instead of the source characters being re-checked.
3119
3120
3121 Version 0.94 18-Sep-97
3122 ----------------------
3123
3124 1. The functions are now thread-safe, with the caveat that the global variables
3125 containing pointers to malloc() and free() or alternative functions are the
3126 same for all threads.
3127
3128 2. Get pcre_study() to generate a bitmap of initial characters for non-
3129 anchored patterns when this is possible, and use it if passed to pcre_exec().
3130
3131
3132 Version 0.93 15-Sep-97
3133 ----------------------
3134
3135 1. /(b)|(:+)/ was computing an incorrect first character.
3136
3137 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3138 but not actually doing anything yet.
3139
3140 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3141 as Perl does (e.g. [-az] or [az-]).
3142
3143 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3144 all possible positions.
3145
3146 5. Split up into different modules to avoid including unneeded functions in a
3147 compiled binary. However, compile and exec are still in one module. The "study"
3148 function is split off.
3149
3150 6. The character tables are now in a separate module whose source is generated
3151 by an auxiliary program - but can then be edited by hand if required. There are
3152 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3153 toupper() in the code.
3154
3155 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3156 make them global. Abolish the function for setting them, as the caller can now
3157 set them directly.
3158
3159
3160 Version 0.92 11-Sep-97
3161 ----------------------
3162
3163 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3164 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3165
3166 2. Caseless matching was not working in character classes if the characters in
3167 the pattern were in upper case.
3168
3169 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3170
3171 4. Make PCRE_ANCHORED public and accept as a compile option.
3172
3173 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3174 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3175 pass them.
3176
3177 6. Give an error if bad option bits passed at compile or run time.
3178
3179 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3180 pcretest to cause it to pass that flag.
3181
3182 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3183 options, and the first character, if set.
3184
3185 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3186
3187
3188 Version 0.91 10-Sep-97
3189 ----------------------
3190
3191 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3192 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3193
3194 2. PCRE was looping on encountering an indefinitely repeated back reference to
3195 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3196 Perl does - treats the match as successful.
3197
3198 ****

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