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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 328 - (show annotations) (download)
Wed Mar 26 17:39:06 2008 UTC (6 years, 3 months ago) by ph10
File size: 152547 byte(s)
Patch to pcrecpp.cc to fix an OS X linking problem.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.7 05-Mar-08
5 ---------------------
6
7 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
8 a string to a long long, pretend we don't even have a long long." This is
9 done by checking for the strtoq, strtoll, and _strtoi64 functions.
10
11 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
12 pre-7.6 versions, which defined a global no_arg variable instead of putting
13 it in the RE class.
14
15 3. Remove a line of dead code, identified by coverity and reported by Nuno
16 Lopes.
17
18 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
19
20 (1) The include/exclude patterns were being applied to the whole pathnames
21 of files, instead of just to the final components.
22
23 (2) If there was more than one level of directory, the subdirectories were
24 skipped unless they satisfied the include/exclude conditions. This is
25 inconsistent with GNU grep (and could even be seen as contrary to the
26 pcregrep specification - which I improved to make it absolutely clear).
27 The action now is always to scan all levels of directory, and just
28 apply the include/exclude patterns to regular files.
29
30 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
31 --exclude_dir in the tests to avoid scanning .svn directories.
32
33 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
34 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
35 doesn't support NULs in patterns.
36
37 7. Added some missing "const"s to declarations of static tables in
38 pcre_compile.c and pcre_dfa_exec.c.
39
40 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
41 caused by fix #2 above.
42
43
44 Version 7.6 28-Jan-08
45 ---------------------
46
47 1. A character class containing a very large number of characters with
48 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
49 overflow.
50
51 2. Patch to cut out the "long long" test in pcrecpp_unittest when
52 HAVE_LONG_LONG is not defined.
53
54 3. Applied Christian Ehrlicher's patch to update the CMake build files to
55 bring them up to date and include new features. This patch includes:
56
57 - Fixed PH's badly added libz and libbz2 support.
58 - Fixed a problem with static linking.
59 - Added pcredemo. [But later removed - see 7 below.]
60 - Fixed dftables problem and added an option.
61 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
62 HAVE_LONG_LONG.
63 - Added readline support for pcretest.
64 - Added an listing of the option settings after cmake has run.
65
66 4. A user submitted a patch to Makefile that makes it easy to create
67 "pcre.dll" under mingw when using Configure/Make. I added stuff to
68 Makefile.am that cause it to include this special target, without
69 affecting anything else. Note that the same mingw target plus all
70 the other distribution libraries and programs are now supported
71 when configuring with CMake (see 6 below) instead of with
72 Configure/Make.
73
74 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
75 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
76 exported in the Windows port". It has not yet been confirmed that the patch
77 solves the problem, but it does no harm.
78
79 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
80 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
81 with CMake, and also correct the comment about stack recursion.
82
83 7. Remove the automatic building of pcredemo from the ./configure system and
84 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
85 of a program that users should build themselves after PCRE is installed, so
86 building it automatically is not really right. What is more, it gave
87 trouble in some build environments.
88
89 8. Further tidies to CMakeLists.txt from Sheri and Christian.
90
91
92 Version 7.5 10-Jan-08
93 ---------------------
94
95 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
96 values in parens when parsing an RE using the C++ wrapper."
97
98 2. Negative specials like \S did not work in character classes in UTF-8 mode.
99 Characters greater than 255 were excluded from the class instead of being
100 included.
101
102 3. The same bug as (2) above applied to negated POSIX classes such as
103 [:^space:].
104
105 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
106 defined or documented. It seems to have been a typo for PCRE_STATIC, so
107 I have changed it.
108
109 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
110 first named subpattern) and a construct such as (?&a) would reference the
111 first named subpattern whose name started with "a" (in other words, the
112 length check was missing). Both these problems are fixed. "Subpattern name
113 expected" is now given for (?&) (a zero-length name), and this patch also
114 makes it give the same error for \k'' (previously it complained that that
115 was a reference to a non-existent subpattern).
116
117 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
118 this is right because (?- can be followed by option settings as well as by
119 digits. I have, however, made the messages clearer.
120
121 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
122 than the number used in the conditional) now cause a compile-time error.
123 This is actually not compatible with Perl, which accepts such patterns, but
124 treats the conditional as always being FALSE (as PCRE used to), but it
125 seems to me that giving a diagnostic is better.
126
127 8. Change "alphameric" to the more common word "alphanumeric" in comments
128 and messages.
129
130 9. Fix two occurrences of "backslash" in comments that should have been
131 "backspace".
132
133 10. Remove two redundant lines of code that can never be obeyed (their function
134 was moved elsewhere).
135
136 11. The program that makes PCRE's Unicode character property table had a bug
137 which caused it to generate incorrect table entries for sequences of
138 characters that have the same character type, but are in different scripts.
139 It amalgamated them into a single range, with the script of the first of
140 them. In other words, some characters were in the wrong script. There were
141 thirteen such cases, affecting characters in the following ranges:
142
143 U+002b0 - U+002c1
144 U+0060c - U+0060d
145 U+0061e - U+00612
146 U+0064b - U+0065e
147 U+0074d - U+0076d
148 U+01800 - U+01805
149 U+01d00 - U+01d77
150 U+01d9b - U+01dbf
151 U+0200b - U+0200f
152 U+030fc - U+030fe
153 U+03260 - U+0327f
154 U+0fb46 - U+0fbb1
155 U+10450 - U+1049d
156
157 12. The -o option (show only the matching part of a line) for pcregrep was not
158 compatible with GNU grep in that, if there was more than one match in a
159 line, it showed only the first of them. It now behaves in the same way as
160 GNU grep.
161
162 13. If the -o and -v options were combined for pcregrep, it printed a blank
163 line for every non-matching line. GNU grep prints nothing, and pcregrep now
164 does the same. The return code can be used to tell if there were any
165 non-matching lines.
166
167 14. Added --file-offsets and --line-offsets to pcregrep.
168
169 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
170 infinitely looping recursion. The bug was that positive lookaheads were not
171 being skipped when checking for a possible empty match (negative lookaheads
172 and both kinds of lookbehind were skipped).
173
174 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
175 inclusion of <windows.h> to before rather than after the definition of
176 INVALID_FILE_ATTRIBUTES (patch from David Byron).
177
178 17. Specifying a possessive quantifier with a specific limit for a Unicode
179 character property caused pcre_compile() to compile bad code, which led at
180 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
181 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
182 caused the error; without that there was no problem.
183
184 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
185
186 19. Added --enable-pcretest-libreadline.
187
188 20. In pcrecpp.cc, the variable 'count' was incremented twice in
189 RE::GlobalReplace(). As a result, the number of replacements returned was
190 double what it should be. I removed one of the increments, but Craig sent a
191 later patch that removed the other one (the right fix) and added unit tests
192 that check the return values (which was not done before).
193
194 21. Several CMake things:
195
196 (1) Arranged that, when cmake is used on Unix, the libraries end up with
197 the names libpcre and libpcreposix, not just pcre and pcreposix.
198
199 (2) The above change means that pcretest and pcregrep are now correctly
200 linked with the newly-built libraries, not previously installed ones.
201
202 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
203
204 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
205 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
206 UTF-8 newline character). The key issue is that the pattern starts .*;
207 this means that the match must be either at the beginning, or after a
208 newline. The bug was in the code for advancing after a failed match and
209 checking that the new position followed a newline. It was not taking
210 account of UTF-8 characters correctly.
211
212 23. PCRE was behaving differently from Perl in the way it recognized POSIX
213 character classes. PCRE was not treating the sequence [:...:] as a
214 character class unless the ... were all letters. Perl, however, seems to
215 allow any characters between [: and :], though of course it rejects as
216 unknown any "names" that contain non-letters, because all the known class
217 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
218 for example, whereas PCRE did not - it did not recognize a POSIX character
219 class. This seemed a bit dangerous, so the code has been changed to be
220 closer to Perl. The behaviour is not identical to Perl, because PCRE will
221 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
222 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
223 Perl does, and where it didn't before.
224
225 24. Rewrite so as to remove the single use of %n from pcregrep because in some
226 Windows environments %n is disabled by default.
227
228
229 Version 7.4 21-Sep-07
230 ---------------------
231
232 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
233 means that a class such as [\s] counted as "explicit reference to CR or
234 LF". That isn't really right - the whole point of the change was to try to
235 help when there was an actual mention of one of the two characters. So now
236 the change happens only if \r or \n (or a literal CR or LF) character is
237 encountered.
238
239 2. The 32-bit options word was also used for 6 internal flags, but the numbers
240 of both had grown to the point where there were only 3 bits left.
241 Fortunately, there was spare space in the data structure, and so I have
242 moved the internal flags into a new 16-bit field to free up more option
243 bits.
244
245 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
246 but did not set the internal JCHANGED flag - either of these is enough to
247 control the way the "get" function works - but the PCRE_INFO_JCHANGED
248 facility is supposed to tell if (?J) was ever used, so now (?J) at the
249 start sets both bits.
250
251 4. Added options (at build time, compile time, exec time) to change \R from
252 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
253
254 5. doc/pcresyntax.html was missing from the distribution.
255
256 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
257 compatibility, even though it is no longer used.
258
259 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
260 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
261 windows.h file is present (where different names are used). [This was
262 reversed later after testing - see 16 below.]
263
264 8. Changed all #include <config.h> to #include "config.h". There were also
265 some further <pcre.h> cases that I changed to "pcre.h".
266
267 9. When pcregrep was used with the --colour option, it missed the line ending
268 sequence off the lines that it output.
269
270 10. It was pointed out to me that arrays of string pointers cause lots of
271 relocations when a shared library is dynamically loaded. A technique of
272 using a single long string with a table of offsets can drastically reduce
273 these. I have refactored PCRE in four places to do this. The result is
274 dramatic:
275
276 Originally: 290
277 After changing UCP table: 187
278 After changing error message table: 43
279 After changing table of "verbs" 36
280 After changing table of Posix names 22
281
282 Thanks to the folks working on Gregex for glib for this insight.
283
284 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
285 unicode-properties was also set.
286
287 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
288
289 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
290 checked only for CRLF.
291
292 14. Added casts to pcretest.c to avoid compiler warnings.
293
294 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
295
296 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
297 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
298 entirely. This removes changes made in 7 above.
299
300 17. The CMake files have been updated, and there is now more information about
301 building with CMake in the NON-UNIX-USE document.
302
303
304 Version 7.3 28-Aug-07
305 ---------------------
306
307 1. In the rejigging of the build system that eventually resulted in 7.1, the
308 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
309 brackets there is not right, since it causes compilers to look for an
310 installed pcre.h, not the version that is in the source that is being
311 compiled (which of course may be different). I have changed it back to:
312
313 #include "pcre.h"
314
315 I have a vague recollection that the change was concerned with compiling in
316 different directories, but in the new build system, that is taken care of
317 by the VPATH setting the Makefile.
318
319 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
320 when the subject happened to end in the byte 0x85 (e.g. if the last
321 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
322 characters but of course it shouldn't be taken as a newline when it is part
323 of another character. The bug was that, for an unlimited repeat of . in
324 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
325 characters when looking for a newline.
326
327 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
328
329 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
330 in debug output.
331
332 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
333 long printing in the pcrecpp unittest when running under MinGW.
334
335 6. ESC_K was left out of the EBCDIC table.
336
337 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
338 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
339 limit also applies to "virtual nesting" when a pattern is recursive, and in
340 this case 1000 isn't so big. I have been able to remove this limit at the
341 expense of backing off one optimization in certain circumstances. Normally,
342 when pcre_exec() would call its internal match() function recursively and
343 immediately return the result unconditionally, it uses a "tail recursion"
344 feature to save stack. However, when a subpattern that can match an empty
345 string has an unlimited repetition quantifier, it no longer makes this
346 optimization. That gives it a stack frame in which to save the data for
347 checking that an empty string has been matched. Previously this was taken
348 from the 1000-entry workspace that had been reserved. So now there is no
349 explicit limit, but more stack is used.
350
351 8. Applied Daniel's patches to solve problems with the import/export magic
352 syntax that is required for Windows, and which was going wrong for the
353 pcreposix and pcrecpp parts of the library. These were overlooked when this
354 problem was solved for the main library.
355
356 9. There were some crude static tests to avoid integer overflow when computing
357 the size of patterns that contain repeated groups with explicit upper
358 limits. As the maximum quantifier is 65535, the maximum group length was
359 set at 30,000 so that the product of these two numbers did not overflow a
360 32-bit integer. However, it turns out that people want to use groups that
361 are longer than 30,000 bytes (though not repeat them that many times).
362 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
363 made it possible to implement the integer overflow checks in a much more
364 dynamic way, which I have now done. The artificial limitation on group
365 length has been removed - we now have only the limit on the total length of
366 the compiled pattern, which depends on the LINK_SIZE setting.
367
368 10. Fixed a bug in the documentation for get/copy named substring when
369 duplicate names are permitted. If none of the named substrings are set, the
370 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
371 empty string.
372
373 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
374 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
375 because the ] is interpreted as the first data character and the
376 terminating ] is not found. PCRE has been made compatible with Perl in this
377 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
378 cause memory overwriting.
379
380 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
381 string has been matched (to stop an infinite loop). It was not recognizing
382 a conditional subpattern that could match an empty string if that
383 subpattern was within another subpattern. For example, it looped when
384 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
385 condition was not nested. This bug has been fixed.
386
387 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
388 past the start of the subject in the presence of bytes with the top bit
389 set, for example "\x8aBCD".
390
391 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
392 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
393
394 14. Optimized (?!) to (*FAIL).
395
396 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
397 This restricts code points to be within the range 0 to 0x10FFFF, excluding
398 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
399 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
400 does: it's just the validity check that is more restrictive.
401
402 16. Inserted checks for integer overflows during escape sequence (backslash)
403 processing, and also fixed erroneous offset values for syntax errors during
404 backslash processing.
405
406 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
407 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
408
409 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
410 caused an overrun.
411
412 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
413 something other than just ASCII characters) inside a group that had an
414 unlimited repeat caused a loop at compile time (while checking to see
415 whether the group could match an empty string).
416
417 20. Debugging a pattern containing \p or \P could cause a crash. For example,
418 [\P{Any}] did so. (Error in the code for printing property names.)
419
420 21. An orphan \E inside a character class could cause a crash.
421
422 22. A repeated capturing bracket such as (A)? could cause a wild memory
423 reference during compilation.
424
425 23. There are several functions in pcre_compile() that scan along a compiled
426 expression for various reasons (e.g. to see if it's fixed length for look
427 behind). There were bugs in these functions when a repeated \p or \P was
428 present in the pattern. These operators have additional parameters compared
429 with \d, etc, and these were not being taken into account when moving along
430 the compiled data. Specifically:
431
432 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
433 length.
434
435 (b) An item such as \pL+ within a repeated group could cause crashes or
436 loops.
437
438 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
439 "reference to non-existent subpattern" error.
440
441 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
442
443 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
444 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
445
446 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
447
448 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
449 character were causing crashes (broken optimization).
450
451 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
452 \p or \P) caused a compile-time loop.
453
454 28. More problems have arisen in unanchored patterns when CRLF is a valid line
455 break. For example, the unstudied pattern [\r\n]A does not match the string
456 "\r\nA" because change 7.0/46 below moves the current point on by two
457 characters after failing to match at the start. However, the pattern \nA
458 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
459 the same is true. There doesn't seem any very clean way out of this, but
460 what I have chosen to do makes the common cases work: PCRE now takes note
461 of whether there can be an explicit match for \r or \n anywhere in the
462 pattern, and if so, 7.0/46 no longer applies. As part of this change,
463 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
464 pattern has explicit CR or LF references.
465
466 29. Added (*CR) etc for changing newline setting at start of pattern.
467
468
469 Version 7.2 19-Jun-07
470 ---------------------
471
472 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
473 which is apparently normally available under Windows.
474
475 2. Re-jig the pcregrep tests with different newline settings in an attempt
476 to make them independent of the local environment's newline setting.
477
478 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
479
480 4. Some of the "internals" tests were previously cut out when the link size
481 was not 2, because the output contained actual offsets. The recent new
482 "Z" feature of pcretest means that these can be cut out, making the tests
483 usable with all link sizes.
484
485 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
486 stack recursion. This gives a massive performance boost under BSD, but just
487 a small improvement under Linux. However, it saves one field in the frame
488 in all cases.
489
490 6. Added more features from the forthcoming Perl 5.10:
491
492 (a) (?-n) (where n is a string of digits) is a relative subroutine or
493 recursion call. It refers to the nth most recently opened parentheses.
494
495 (b) (?+n) is also a relative subroutine call; it refers to the nth next
496 to be opened parentheses.
497
498 (c) Conditions that refer to capturing parentheses can be specified
499 relatively, for example, (?(-2)... or (?(+3)...
500
501 (d) \K resets the start of the current match so that everything before
502 is not part of it.
503
504 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
505
506 (f) \g{name} is another synonym - part of Perl 5.10's unification of
507 reference syntax.
508
509 (g) (?| introduces a group in which the numbering of parentheses in each
510 alternative starts with the same number.
511
512 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
513
514 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
515 PCRE_INFO_JCHANGED.
516
517 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
518 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
519 for detecting groups that can match an empty string.
520
521 9. A pattern with a very large number of alternatives (more than several
522 hundred) was running out of internal workspace during the pre-compile
523 phase, where pcre_compile() figures out how much memory will be needed. A
524 bit of new cunning has reduced the workspace needed for groups with
525 alternatives. The 1000-alternative test pattern now uses 12 bytes of
526 workspace instead of running out of the 4096 that are available.
527
528 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
529
530 11. Applied patch from Google to remove an optimization that didn't quite work.
531 The report of the bug said:
532
533 pcrecpp::RE("a*").FullMatch("aaa") matches, while
534 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
535 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
536
537 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
538 it matched the wrong number of bytes.
539
540
541 Version 7.1 24-Apr-07
542 ---------------------
543
544 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
545 that is more "standard", making use of automake and other Autotools. There
546 is some re-arrangement of the files and adjustment of comments consequent
547 on this.
548
549 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
550 for recursive directory scanning broke on some systems because the files
551 are not scanned in any specific order and on different systems the order
552 was different. A call to "sort" has been inserted into RunGrepTest for the
553 approprate test as a short-term fix. In the longer term there may be an
554 alternative.
555
556 3. I had an email from Eric Raymond about problems translating some of PCRE's
557 man pages to HTML (despite the fact that I distribute HTML pages, some
558 people do their own conversions for various reasons). The problems
559 concerned the use of low-level troff macros .br and .in. I have therefore
560 removed all such uses from the man pages (some were redundant, some could
561 be replaced by .nf/.fi pairs). The 132html script that I use to generate
562 HTML has been updated to handle .nf/.fi and to complain if it encounters
563 .br or .in.
564
565 4. Updated comments in configure.ac that get placed in config.h.in and also
566 arranged for config.h to be included in the distribution, with the name
567 config.h.generic, for the benefit of those who have to compile without
568 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
569
570 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
571 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
572 makevp.bat for the current PCRE, using the additional files
573 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
574
575 6. A Windows user reported a minor discrepancy with test 2, which turned out
576 to be caused by a trailing space on an input line that had got lost in his
577 copy. The trailing space was an accident, so I've just removed it.
578
579 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
580 that is needed.
581
582 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
583 as "const" (a) because they are and (b) because it helps the PHP
584 maintainers who have recently made a script to detect big data structures
585 in the php code that should be moved to the .rodata section. I remembered
586 to update Builducptable as well, so it won't revert if ucptable.h is ever
587 re-created.
588
589 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
590 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
591 order to be able to cut out the UTF-8 tables in the latter when UTF-8
592 support is not required. This saves 1.5-2K of code, which is important in
593 some applications.
594
595 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
596 so as not to refer to the tables, even though these functions will never be
597 called when UTF-8 support is disabled. Otherwise there are problems with a
598 shared library.
599
600 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
601
602 (a) It was defining its arguments as char * instead of void *.
603
604 (b) It was assuming that all moves were upwards in memory; this was true
605 a long time ago when I wrote it, but is no longer the case.
606
607 The emulated memove() is provided for those environments that have neither
608 memmove() nor bcopy(). I didn't think anyone used it these days, but that
609 is clearly not the case, as these two bugs were recently reported.
610
611 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
612 and Detrail to create the HTML documentation, the .txt form of the man
613 pages, and it removes trailing spaces from listed files. It also creates
614 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
615 case, it wraps all the #defines with #ifndefs. This script should be run
616 before "make dist".
617
618 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
619 with Unicode property support.
620
621 (a) For a maximizing quantifier, if the two different cases of the
622 character were of different lengths in their UTF-8 codings (there are
623 some cases like this - I found 11), and the matching function had to
624 back up over a mixture of the two cases, it incorrectly assumed they
625 were both the same length.
626
627 (b) When PCRE was configured to use the heap rather than the stack for
628 recursion during matching, it was not correctly preserving the data for
629 the other case of a UTF-8 character when checking ahead for a match
630 while processing a minimizing repeat. If the check also involved
631 matching a wide character, but failed, corruption could cause an
632 erroneous result when trying to check for a repeat of the original
633 character.
634
635 13. Some tidying changes to the testing mechanism:
636
637 (a) The RunTest script now detects the internal link size and whether there
638 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
639 values substituted by "configure". (The RunGrepTest script already did
640 this for UTF-8.) The configure.ac script no longer substitutes the
641 relevant variables.
642
643 (b) The debugging options /B and /D in pcretest show the compiled bytecode
644 with length and offset values. This means that the output is different
645 for different internal link sizes. Test 2 is skipped for link sizes
646 other than 2 because of this, bypassing the problem. Unfortunately,
647 there was also a test in test 3 (the locale tests) that used /B and
648 failed for link sizes other than 2. Rather than cut the whole test out,
649 I have added a new /Z option to pcretest that replaces the length and
650 offset values with spaces. This is now used to make test 3 independent
651 of link size. (Test 2 will be tidied up later.)
652
653 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
654 segmentation fault instead of returning the appropriate error message.
655
656 15. In multiline mode when the newline sequence was set to "any", the pattern
657 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
658 This doesn't seem right; it now treats the CRLF combination as the line
659 ending, and so does not match in that case. It's only a pattern such as ^$
660 that would hit this one: something like ^ABC$ would have failed after \r
661 and then tried again after \r\n.
662
663 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
664 in an attempt to make files that differ only in their line terminators
665 compare equal. This works on Linux.
666
667 17. Under certain error circumstances pcregrep might try to free random memory
668 as it exited. This is now fixed, thanks to valgrind.
669
670 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
671 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
672 was because its rules for how to advance for /g after matching an empty
673 string at the end of a line did not allow for this case. They now check for
674 it specially.
675
676 20. pcretest is supposed to handle patterns and data of any length, by
677 extending its buffers when necessary. It was getting this wrong when the
678 buffer for a data line had to be extended.
679
680 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
681 CRLF as a newline sequence.
682
683 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
684 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
685 I have nevertheless tidied it up.
686
687 23. Added some casts to kill warnings from HP-UX ia64 compiler.
688
689 24. Added a man page for pcre-config.
690
691
692 Version 7.0 19-Dec-06
693 ---------------------
694
695 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
696 moving to gcc 4.1.1.
697
698 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
699 sys/time.h, which is documented as needed for this function. It doesn't
700 seem to matter on Linux, but it showed up on some releases of OS X.
701
702 3. It seems that there are systems where bytes whose values are greater than
703 127 match isprint() in the "C" locale. The "C" locale should be the
704 default when a C program starts up. In most systems, only ASCII printing
705 characters match isprint(). This difference caused the output from pcretest
706 to vary, making some of the tests fail. I have changed pcretest so that:
707
708 (a) When it is outputting text in the compiled version of a pattern, bytes
709 other than 32-126 are always shown as hex escapes.
710
711 (b) When it is outputting text that is a matched part of a subject string,
712 it does the same, unless a different locale has been set for the match
713 (using the /L modifier). In this case, it uses isprint() to decide.
714
715 4. Fixed a major bug that caused incorrect computation of the amount of memory
716 required for a compiled pattern when options that changed within the
717 pattern affected the logic of the preliminary scan that determines the
718 length. The relevant options are -x, and -i in UTF-8 mode. The result was
719 that the computed length was too small. The symptoms of this bug were
720 either the PCRE error "internal error: code overflow" from pcre_compile(),
721 or a glibc crash with a message such as "pcretest: free(): invalid next
722 size (fast)". Examples of patterns that provoked this bug (shown in
723 pcretest format) are:
724
725 /(?-x: )/x
726 /(?x)(?-x: \s*#\s*)/
727 /((?i)[\x{c0}])/8
728 /(?i:[\x{c0}])/8
729
730 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
731 is now done differently.
732
733 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
734 wrapper classes; (b) implement a new function in the C++ scanner that is
735 more efficient than the old way of doing things because it avoids levels of
736 recursion in the regex matching; (c) add a paragraph to the documentation
737 for the FullMatch() function.
738
739 6. The escape sequence \n was being treated as whatever was defined as
740 "newline". Not only was this contrary to the documentation, which states
741 that \n is character 10 (hex 0A), but it also went horribly wrong when
742 "newline" was defined as CRLF. This has been fixed.
743
744 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
745 was being set to -1 for the "end of line" case (supposedly a value that no
746 character can have). Though this value is never used (the check for end of
747 line is "zero bytes in current character"), it caused compiler complaints.
748 I've changed it to 0xffffffff.
749
750 8. In pcre_version.c, the version string was being built by a sequence of
751 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
752 string (as it is for production releases) called a macro with an empty
753 argument. The C standard says the result of this is undefined. The gcc
754 compiler treats it as an empty string (which was what was wanted) but it is
755 reported that Visual C gives an error. The source has been hacked around to
756 avoid this problem.
757
758 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
759 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
760 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
761 of them did).
762
763 10. Originally, pcretest opened its input and output without "b"; then I was
764 told that "b" was needed in some environments, so it was added for release
765 5.0 to both the input and output. (It makes no difference on Unix-like
766 systems.) Later I was told that it is wrong for the input on Windows. I've
767 now abstracted the modes into two macros, to make it easier to fiddle with
768 them, and removed "b" from the input mode under Windows.
769
770 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
771
772 12. Added -help and --help to pcretest as an official way of being reminded
773 of the options.
774
775 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
776 and pcrecpp.cc because they annoy compilers at high warning levels.
777
778 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
779
780 15. Fixed an occurrence of == in configure.ac that should have been = (shell
781 scripts are not C programs :-) and which was not noticed because it works
782 on Linux.
783
784 16. pcretest is supposed to handle any length of pattern and data line (as one
785 line or as a continued sequence of lines) by extending its input buffer if
786 necessary. This feature was broken for very long pattern lines, leading to
787 a string of junk being passed to pcre_compile() if the pattern was longer
788 than about 50K.
789
790 17. I have done a major re-factoring of the way pcre_compile() computes the
791 amount of memory needed for a compiled pattern. Previously, there was code
792 that made a preliminary scan of the pattern in order to do this. That was
793 OK when PCRE was new, but as the facilities have expanded, it has become
794 harder and harder to keep it in step with the real compile phase, and there
795 have been a number of bugs (see for example, 4 above). I have now found a
796 cunning way of running the real compile function in a "fake" mode that
797 enables it to compute how much memory it would need, while actually only
798 ever using a few hundred bytes of working memory and without too many
799 tests of the mode. This should make future maintenance and development
800 easier. A side effect of this work is that the limit of 200 on the nesting
801 depth of parentheses has been removed (though this was never a serious
802 limitation, I suspect). However, there is a downside: pcre_compile() now
803 runs more slowly than before (30% or more, depending on the pattern). I
804 hope this isn't a big issue. There is no effect on runtime performance.
805
806 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
807 newline (only possible for the last line of a file) and it was a
808 pattern that set a locale (followed by /Lsomething), pcretest crashed.
809
810 19. Added additional timing features to pcretest. (1) The -tm option now times
811 matching only, not compiling. (2) Both -t and -tm can be followed, as a
812 separate command line item, by a number that specifies the number of
813 repeats to use when timing. The default is 50000; this gives better
814 precision, but takes uncomfortably long for very large patterns.
815
816 20. Extended pcre_study() to be more clever in cases where a branch of a
817 subpattern has no definite first character. For example, (a*|b*)[cd] would
818 previously give no result from pcre_study(). Now it recognizes that the
819 first character must be a, b, c, or d.
820
821 21. There was an incorrect error "recursive call could loop indefinitely" if
822 a subpattern (or the entire pattern) that was being tested for matching an
823 empty string contained only one non-empty item after a nested subpattern.
824 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
825 incorrectly, because the \d was being skipped in the check.
826
827 22. The pcretest program now has a new pattern option /B and a command line
828 option -b, which is equivalent to adding /B to every pattern. This causes
829 it to show the compiled bytecode, without the additional information that
830 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
831 is the same as /B/I).
832
833 23. A new optimization is now able automatically to treat some sequences such
834 as a*b as a*+b. More specifically, if something simple (such as a character
835 or a simple class like \d) has an unlimited quantifier, and is followed by
836 something that cannot possibly match the quantified thing, the quantifier
837 is automatically "possessified".
838
839 24. A recursive reference to a subpattern whose number was greater than 39
840 went wrong under certain circumstances in UTF-8 mode. This bug could also
841 have affected the operation of pcre_study().
842
843 25. Realized that a little bit of performance could be had by replacing
844 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
845
846 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
847
848 27. Possessive quantifiers such as a++ were previously implemented by turning
849 them into atomic groups such as ($>a+). Now they have their own opcodes,
850 which improves performance. This includes the automatically created ones
851 from 23 above.
852
853 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
854 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
855 the first matched character to be a colon. This applied both to named and
856 numbered groups.
857
858 29. The ucpinternal.h header file was missing its idempotency #ifdef.
859
860 30. I was sent a "project" file called libpcre.a.dev which I understand makes
861 building PCRE on Windows easier, so I have included it in the distribution.
862
863 31. There is now a check in pcretest against a ridiculously large number being
864 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
865 loop, the loop is abandoned.
866
867 32. Forward references to subpatterns in conditions such as (?(2)...) where
868 subpattern 2 is defined later cause pcre_compile() to search forwards in
869 the pattern for the relevant set of parentheses. This search went wrong
870 when there were unescaped parentheses in a character class, parentheses
871 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
872
873 33. "Subroutine" calls and backreferences were previously restricted to
874 referencing subpatterns earlier in the regex. This restriction has now
875 been removed.
876
877 34. Added a number of extra features that are going to be in Perl 5.10. On the
878 whole, these are just syntactic alternatives for features that PCRE had
879 previously implemented using the Python syntax or my own invention. The
880 other formats are all retained for compatibility.
881
882 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
883 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
884 also .NET compatible.
885
886 (b) A recursion or subroutine call to a named group can now be defined as
887 (?&name) as well as (?P>name).
888
889 (c) A backreference to a named group can now be defined as \k<name> or
890 \k'name' as well as (?P=name). The new forms, as well as being in Perl
891 5.10, are also .NET compatible.
892
893 (d) A conditional reference to a named group can now use the syntax
894 (?(<name>) or (?('name') as well as (?(name).
895
896 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
897 groups (named and numbered) that are never evaluated inline, but can be
898 called as "subroutines" from elsewhere. In effect, the DEFINE condition
899 is always false. There may be only one alternative in such a group.
900
901 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
902 as the simple (?(R). The condition is true only if the most recent
903 recursion is that of the given number or name. It does not search out
904 through the entire recursion stack.
905
906 (g) The escape \gN or \g{N} has been added, where N is a positive or
907 negative number, specifying an absolute or relative reference.
908
909 35. Tidied to get rid of some further signed/unsigned compiler warnings and
910 some "unreachable code" warnings.
911
912 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
913 things, this adds five new scripts.
914
915 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
916 There were also incompatibilities regarding the handling of \Q..\E inside
917 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
918 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
919
920 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
921 matches an empty string, and forcibly breaks the loop. There were bugs in
922 this code in non-simple cases. For a pattern such as ^(a()*)* matched
923 against aaaa the result was just "a" rather than "aaaa", for example. Two
924 separate and independent bugs (that affected different cases) have been
925 fixed.
926
927 39. Refactored the code to abolish the use of different opcodes for small
928 capturing bracket numbers. This is a tidy that I avoided doing when I
929 removed the limit on the number of capturing brackets for 3.5 back in 2001.
930 The new approach is not only tidier, it makes it possible to reduce the
931 memory needed to fix the previous bug (38).
932
933 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
934 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
935 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
936 mode.
937
938 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
939 report.
940
941 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
942 copy construction and assignment in the C++ wrapper.
943
944 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
945 couple of bugs that could have given wrong results in the "--newline=crlf"
946 case.
947
948 44. Added a number of casts and did some reorganization of signed/unsigned int
949 variables following suggestions from Dair Grant. Also renamed the variable
950 "this" as "item" because it is a C++ keyword.
951
952 45. Arranged for dftables to add
953
954 #include "pcre_internal.h"
955
956 to pcre_chartables.c because without it, gcc 4.x may remove the array
957 definition from the final binary if PCRE is built into a static library and
958 dead code stripping is activated.
959
960 46. For an unanchored pattern, if a match attempt fails at the start of a
961 newline sequence, and the newline setting is CRLF or ANY, and the next two
962 characters are CRLF, advance by two characters instead of one.
963
964
965 Version 6.7 04-Jul-06
966 ---------------------
967
968 1. In order to handle tests when input lines are enormously long, pcretest has
969 been re-factored so that it automatically extends its buffers when
970 necessary. The code is crude, but this _is_ just a test program. The
971 default size has been increased from 32K to 50K.
972
973 2. The code in pcre_study() was using the value of the re argument before
974 testing it for NULL. (Of course, in any sensible call of the function, it
975 won't be NULL.)
976
977 3. The memmove() emulation function in pcre_internal.h, which is used on
978 systems that lack both memmove() and bcopy() - that is, hardly ever -
979 was missing a "static" storage class specifier.
980
981 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
982 containing an extended class (one that cannot be represented by a bitmap
983 because it contains high-valued characters or Unicode property items, e.g.
984 [\pZ]). Almost always one would set UTF-8 mode when processing such a
985 pattern, but PCRE should not loop if you do not (it no longer does).
986 [Detail: two cases were found: (a) a repeated subpattern containing an
987 extended class; (b) a recursive reference to a subpattern that followed a
988 previous extended class. It wasn't skipping over the extended class
989 correctly when UTF-8 mode was not set.]
990
991 5. A negated single-character class was not being recognized as fixed-length
992 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
993 compile error "lookbehind assertion is not fixed length".
994
995 6. The RunPerlTest auxiliary script was showing an unexpected difference
996 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
997 write a Perl script that can interpret lines of an input file either as
998 byte characters or as UTF-8, which is what "perltest" was being required to
999 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1000 can't do is switch easily at run time between having the "use utf8;" pragma
1001 or not. In the end, I fudged it by using the RunPerlTest script to insert
1002 "use utf8;" explicitly for the UTF-8 tests.
1003
1004 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1005 the end of the subject string, contrary to the documentation and to what
1006 Perl does. This was true of both matching functions. Now it matches only at
1007 the start of the subject and immediately after *internal* newlines.
1008
1009 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1010 a pointer to an int instead of a pointer to an unsigned long int. This
1011 caused problems on 64-bit systems.
1012
1013 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1014 instance of the 'standard' template library not being so standard".
1015
1016 10. There was no check on the number of named subpatterns nor the maximum
1017 length of a subpattern name. The product of these values is used to compute
1018 the size of the memory block for a compiled pattern. By supplying a very
1019 long subpattern name and a large number of named subpatterns, the size
1020 computation could be caused to overflow. This is now prevented by limiting
1021 the length of names to 32 characters, and the number of named subpatterns
1022 to 10,000.
1023
1024 11. Subpatterns that are repeated with specific counts have to be replicated in
1025 the compiled pattern. The size of memory for this was computed from the
1026 length of the subpattern and the repeat count. The latter is limited to
1027 65535, but there was no limit on the former, meaning that integer overflow
1028 could in principle occur. The compiled length of a repeated subpattern is
1029 now limited to 30,000 bytes in order to prevent this.
1030
1031 12. Added the optional facility to have named substrings with the same name.
1032
1033 13. Added the ability to use a named substring as a condition, using the
1034 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1035 are numbers (not recommended). Forward references are permitted.
1036
1037 14. Added forward references in named backreferences (if you see what I mean).
1038
1039 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1040 pattern could run off the end of the subject. For example, the pattern
1041 "(?s)(.{1,5})"8 did this with the subject "ab".
1042
1043 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1044 PCRE_CASELESS was set when matching characters that were quantified with ?
1045 or *.
1046
1047 17. A character class other than a single negated character that had a minimum
1048 but no maximum quantifier - for example [ab]{6,} - was not handled
1049 correctly by pce_dfa_exec(). It would match only one character.
1050
1051 18. A valid (though odd) pattern that looked like a POSIX character
1052 class but used an invalid character after [ (for example [[,abc,]]) caused
1053 pcre_compile() to give the error "Failed: internal error: code overflow" or
1054 in some cases to crash with a glibc free() error. This could even happen if
1055 the pattern terminated after [[ but there just happened to be a sequence of
1056 letters, a binary zero, and a closing ] in the memory that followed.
1057
1058 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1059 over the years. Originally (before any Unicode support), just the bottom 8
1060 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1061 output from "man perlunicode" includes this:
1062
1063 The regular expression compiler produces polymorphic opcodes. That
1064 is, the pattern adapts to the data and automatically switches to
1065 the Unicode character scheme when presented with Unicode data--or
1066 instead uses a traditional byte scheme when presented with byte
1067 data.
1068
1069 Sadly, a wide octal escape does not cause a switch, and in a string with
1070 no other multibyte characters, these octal escapes are treated as before.
1071 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1072 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1073 Unicode string.
1074
1075 I have not perpetrated such confusion in PCRE. Up till now, it took just
1076 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1077 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1078 translate to the appropriate multibyte character.
1079
1080 29. Applied some refactoring to reduce the number of warnings from Microsoft
1081 and Borland compilers. This has included removing the fudge introduced
1082 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1083 a warning about an unused variable.
1084
1085 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1086 characters since release 4.0, because Perl (from release 5.004) does not.
1087 [Or at least, is documented not to: some releases seem to be in conflict
1088 with the documentation.] However, when a pattern was studied with
1089 pcre_study() and all its branches started with \s, PCRE still included VT
1090 as a possible starting character. Of course, this did no harm; it just
1091 caused an unnecessary match attempt.
1092
1093 22. Removed a now-redundant internal flag bit that recorded the fact that case
1094 dependency changed within the pattern. This was once needed for "required
1095 byte" processing, but is no longer used. This recovers a now-scarce options
1096 bit. Also moved the least significant internal flag bit to the most-
1097 significant bit of the word, which was not previously used (hangover from
1098 the days when it was an int rather than a uint) to free up another bit for
1099 the future.
1100
1101 23. Added support for CRLF line endings as well as CR and LF. As well as the
1102 default being selectable at build time, it can now be changed at runtime
1103 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1104 specify that it is scanning data with non-default line endings.
1105
1106 24. Changed the definition of CXXLINK to make it agree with the definition of
1107 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1108
1109 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1110 recursions. This makes a big different to stack usage for some patterns.
1111
1112 26. If a subpattern containing a named recursion or subroutine reference such
1113 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1114 the space required for the compiled pattern went wrong and gave too small a
1115 value. Depending on the environment, this could lead to "Failed: internal
1116 error: code overflow at offset 49" or "glibc detected double free or
1117 corruption" errors.
1118
1119 27. Applied patches from Google (a) to support the new newline modes and (b) to
1120 advance over multibyte UTF-8 characters in GlobalReplace.
1121
1122 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1123 difference for some implementation of PCRE in some Windows version.
1124
1125 29. Added some extra testing facilities to pcretest:
1126
1127 \q<number> in a data line sets the "match limit" value
1128 \Q<number> in a data line sets the "match recursion limt" value
1129 -S <number> sets the stack size, where <number> is in megabytes
1130
1131 The -S option isn't available for Windows.
1132
1133
1134 Version 6.6 06-Feb-06
1135 ---------------------
1136
1137 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1138 in pcreposix.h. I have copied the definition from pcre.h.
1139
1140 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1141 because pcre.h is no longer a built file.
1142
1143 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1144 not normally included in the compiled code.
1145
1146
1147 Version 6.5 01-Feb-06
1148 ---------------------
1149
1150 1. When using the partial match feature with pcre_dfa_exec(), it was not
1151 anchoring the second and subsequent partial matches at the new starting
1152 point. This could lead to incorrect results. For example, with the pattern
1153 /1234/, partially matching against "123" and then "a4" gave a match.
1154
1155 2. Changes to pcregrep:
1156
1157 (a) All non-match returns from pcre_exec() were being treated as failures
1158 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1159 error message is output. Some extra information is given for the
1160 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1161 probably the only errors that are likely to be caused by users (by
1162 specifying a regex that has nested indefinite repeats, for instance).
1163 If there are more than 20 of these errors, pcregrep is abandoned.
1164
1165 (b) A binary zero was treated as data while matching, but terminated the
1166 output line if it was written out. This has been fixed: binary zeroes
1167 are now no different to any other data bytes.
1168
1169 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1170 used to set a locale for matching. The --locale=xxxx long option has
1171 been added (no short equivalent) to specify a locale explicitly on the
1172 pcregrep command, overriding the environment variables.
1173
1174 (d) When -B was used with -n, some line numbers in the output were one less
1175 than they should have been.
1176
1177 (e) Added the -o (--only-matching) option.
1178
1179 (f) If -A or -C was used with -c (count only), some lines of context were
1180 accidentally printed for the final match.
1181
1182 (g) Added the -H (--with-filename) option.
1183
1184 (h) The combination of options -rh failed to suppress file names for files
1185 that were found from directory arguments.
1186
1187 (i) Added the -D (--devices) and -d (--directories) options.
1188
1189 (j) Added the -F (--fixed-strings) option.
1190
1191 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1192
1193 (l) Added the --colo(u)r option.
1194
1195 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1196 is not present by default.
1197
1198 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1199 items such as (?R) or (?1), when the recursion could match a number of
1200 alternatives. If it matched one of the alternatives, but subsequently,
1201 outside the recursion, there was a failure, the code tried to back up into
1202 the recursion. However, because of the way PCRE is implemented, this is not
1203 possible, and the result was an incorrect result from the match.
1204
1205 In order to prevent this happening, the specification of recursion has
1206 been changed so that all such subpatterns are automatically treated as
1207 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1208
1209 4. I had overlooked the fact that, in some locales, there are characters for
1210 which isalpha() is true but neither isupper() nor islower() are true. In
1211 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1212 and ordfeminine) are like this. This affected the treatment of \w and \W
1213 when they appeared in character classes, but not when they appeared outside
1214 a character class. The bit map for "word" characters is now created
1215 separately from the results of isalnum() instead of just taking it from the
1216 upper, lower, and digit maps. (Plus the underscore character, of course.)
1217
1218 5. The above bug also affected the handling of POSIX character classes such as
1219 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1220 permanent tables. Instead, the bit maps for such a class were previously
1221 created as the appropriate unions of the upper, lower, and digit bitmaps.
1222 Now they are created by subtraction from the [[:word:]] class, which has
1223 its own bitmap.
1224
1225 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1226 It is created by subtracting the vertical space characters (\x09, \x0a,
1227 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1228 subtraction was done in the overall bitmap for a character class, meaning
1229 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1230 be recognized. This bug has been fixed.
1231
1232 7. Patches from the folks at Google:
1233
1234 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1235 real life, but is still worth protecting against".
1236
1237 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1238 regular expressions".
1239
1240 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1241 have it.
1242
1243 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1244 "configure" and the latter not, in order to fix a problem somebody had
1245 with compiling the Arg class on HP-UX.
1246
1247 (e) Improve the error-handling of the C++ wrapper a little bit.
1248
1249 (f) New tests for checking recursion limiting.
1250
1251 8. The pcre_memmove() function, which is used only if the environment does not
1252 have a standard memmove() function (and is therefore rarely compiled),
1253 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1254 returning a result (though PCRE never actually uses the result).
1255
1256 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1257 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1258 returned instead of calling malloc() with an overflowing number that would
1259 most likely cause subsequent chaos.
1260
1261 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1262
1263 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1264 with this option is matched, the nmatch and pmatch options of regexec() are
1265 ignored.
1266
1267 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1268 provided in case anyone wants to the the POSIX interface with UTF-8
1269 strings.
1270
1271 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1272 C++ linking (needed for some HP-UX environments).
1273
1274 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1275 (unused parameter) and in the pcre_printint() function (omitted "default"
1276 switch label when the default is to do nothing).
1277
1278 15. Added some code to make it possible, when PCRE is compiled as a C++
1279 library, to replace subject pointers for pcre_exec() with a smart pointer
1280 class, thus making it possible to process discontinuous strings.
1281
1282 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1283 much the same function. They were added by different people who were trying
1284 to make PCRE easy to compile on non-Unix systems. It has been suggested
1285 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1286 for compiling on Windows systems. I have therefore replaced it with
1287 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1288 defaults to "extern" for C or "extern C" for C++, which works fine on
1289 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1290 SCOPE with something explicit in config.h. In addition:
1291
1292 (a) pcreposix.h still had just "extern" instead of either of these macros;
1293 I have replaced it with PCRE_DATA_SCOPE.
1294
1295 (b) Functions such as _pcre_xclass(), which are internal to the library,
1296 but external in the C sense, all had PCRE_EXPORT in their definitions.
1297 This is apparently wrong for the Windows case, so I have removed it.
1298 (It makes no difference on Unix-like systems.)
1299
1300 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1301 of recursive calls to match(). This is different to MATCH_LIMIT because
1302 that limits the total number of calls to match(), not all of which increase
1303 the depth of recursion. Limiting the recursion depth limits the amount of
1304 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1305 when PCRE is compiled, and changed at run time. A patch from Google adds
1306 this functionality to the C++ interface.
1307
1308 18. Changes to the handling of Unicode character properties:
1309
1310 (a) Updated the table to Unicode 4.1.0.
1311
1312 (b) Recognize characters that are not in the table as "Cn" (undefined).
1313
1314 (c) I revised the way the table is implemented to a much improved format
1315 which includes recognition of ranges. It now supports the ranges that
1316 are defined in UnicodeData.txt, and it also amalgamates other
1317 characters into ranges. This has reduced the number of entries in the
1318 table from around 16,000 to around 3,000, thus reducing its size
1319 considerably. I realized I did not need to use a tree structure after
1320 all - a binary chop search is just as efficient. Having reduced the
1321 number of entries, I extended their size from 6 bytes to 8 bytes to
1322 allow for more data.
1323
1324 (d) Added support for Unicode script names via properties such as \p{Han}.
1325
1326 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1327 matching that character.
1328
1329 20. When matching a repeated Unicode property with a minimum greater than zero,
1330 (for example \pL{2,}), PCRE could look past the end of the subject if it
1331 reached it while seeking the minimum number of characters. This could
1332 happen only if some of the characters were more than one byte long, because
1333 there is a check for at least the minimum number of bytes.
1334
1335 21. Refactored the implementation of \p and \P so as to be more general, to
1336 allow for more different types of property in future. This has changed the
1337 compiled form incompatibly. Anybody with saved compiled patterns that use
1338 \p or \P will have to recompile them.
1339
1340 22. Added "Any" and "L&" to the supported property types.
1341
1342 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1343 but give a compile time error if the value is greater than 0xff.
1344
1345 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1346 accidentally not being installed or uninstalled.
1347
1348 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1349 made were to insert the current release number. This seemed silly, because
1350 it made things harder for people building PCRE on systems that don't run
1351 "configure". I have turned pcre.h into a distributed file, no longer built
1352 by "configure", with the version identification directly included. There is
1353 no longer a pcre.h.in file.
1354
1355 However, this change necessitated a change to the pcre-config script as
1356 well. It is built from pcre-config.in, and one of the substitutions was the
1357 release number. I have updated configure.ac so that ./configure now finds
1358 the release number by grepping pcre.h.
1359
1360 26. Added the ability to run the tests under valgrind.
1361
1362
1363 Version 6.4 05-Sep-05
1364 ---------------------
1365
1366 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1367 "--" to be printed when multiple files were scanned, even when none of the
1368 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1369 consider it to be a bug, and have restored the previous behaviour.
1370
1371 2. A couple of code tidies to get rid of compiler warnings.
1372
1373 3. The pcretest program used to cheat by referring to symbols in the library
1374 whose names begin with _pcre_. These are internal symbols that are not
1375 really supposed to be visible externally, and in some environments it is
1376 possible to suppress them. The cheating is now confined to including
1377 certain files from the library's source, which is a bit cleaner.
1378
1379 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1380 file's purpose clearer.
1381
1382 5. Reorganized pcre_ucp_findchar().
1383
1384
1385 Version 6.3 15-Aug-05
1386 ---------------------
1387
1388 1. The file libpcre.pc.in did not have general read permission in the tarball.
1389
1390 2. There were some problems when building without C++ support:
1391
1392 (a) If C++ support was not built, "make install" and "make test" still
1393 tried to test it.
1394
1395 (b) There were problems when the value of CXX was explicitly set. Some
1396 changes have been made to try to fix these, and ...
1397
1398 (c) --disable-cpp can now be used to explicitly disable C++ support.
1399
1400 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1401 backslash in a target when C++ was disabled. This confuses some
1402 versions of "make", apparently. Using an intermediate variable solves
1403 this. (Same for CPP_LOBJ.)
1404
1405 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1406 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1407 necessary on certain architectures.
1408
1409 4. Added a setting of -export-symbols-regex to the link command to remove
1410 those symbols that are exported in the C sense, but actually are local
1411 within the library, and not documented. Their names all begin with
1412 "_pcre_". This is not a perfect job, because (a) we have to except some
1413 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1414 available (and never for static libraries). I have made a note to try to
1415 find a way round (a) in the future.
1416
1417
1418 Version 6.2 01-Aug-05
1419 ---------------------
1420
1421 1. There was no test for integer overflow of quantifier values. A construction
1422 such as {1111111111111111} would give undefined results. What is worse, if
1423 a minimum quantifier for a parenthesized subpattern overflowed and became
1424 negative, the calculation of the memory size went wrong. This could have
1425 led to memory overwriting.
1426
1427 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1428
1429 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1430 operating environments where this matters.
1431
1432 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1433 PCRE options from within the C++ wrapper.
1434
1435 5. Named capturing subpatterns were not being correctly counted when a pattern
1436 was compiled. This caused two problems: (a) If there were more than 100
1437 such subpatterns, the calculation of the memory needed for the whole
1438 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1439 back references of the form \12, where the number was greater than 9, were
1440 not recognized as back references, even though there were sufficient
1441 previous subpatterns.
1442
1443 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1444 versions of gcc, e.g. 2.95.4.
1445
1446
1447 Version 6.1 21-Jun-05
1448 ---------------------
1449
1450 1. There was one reference to the variable "posix" in pcretest.c that was not
1451 surrounded by "#if !defined NOPOSIX".
1452
1453 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1454 the cross-check on the old pcre_info() function, for the benefit of the
1455 cut-down version of PCRE that is currently imported into Exim.
1456
1457 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1458 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1459 patterns that start (?i) but I don't think that matters. The use of (?i) is
1460 just an example; this all applies to the other options as well.
1461
1462 4. Since libtool seems to echo the compile commands it is issuing, the output
1463 from "make" can be reduced a bit by putting "@" in front of each libtool
1464 compile command.
1465
1466 5. Patch from the folks at Google for configure.in to be a bit more thorough
1467 in checking for a suitable C++ installation before trying to compile the
1468 C++ stuff. This should fix a reported problem when a compiler was present,
1469 but no suitable headers.
1470
1471 6. The man pages all had just "PCRE" as their title. I have changed them to
1472 be the relevant file name. I have also arranged that these names are
1473 retained in the file doc/pcre.txt, which is a concatenation in text format
1474 of all the man pages except the little individual ones for each function.
1475
1476 7. The NON-UNIX-USE file had not been updated for the different set of source
1477 files that come with release 6. I also added a few comments about the C++
1478 wrapper.
1479
1480
1481 Version 6.0 07-Jun-05
1482 ---------------------
1483
1484 1. Some minor internal re-organization to help with my DFA experiments.
1485
1486 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1487 didn't matter for the library itself when fully configured, but did matter
1488 when compiling without UCP support, or within Exim, where the ucp files are
1489 not imported.
1490
1491 3. Refactoring of the library code to split up the various functions into
1492 different source modules. The addition of the new DFA matching code (see
1493 below) to a single monolithic source would have made it really too
1494 unwieldy, quite apart from causing all the code to be include in a
1495 statically linked application, when only some functions are used. This is
1496 relevant even without the DFA addition now that patterns can be compiled in
1497 one application and matched in another.
1498
1499 The downside of splitting up is that there have to be some external
1500 functions and data tables that are used internally in different modules of
1501 the library but which are not part of the API. These have all had their
1502 names changed to start with "_pcre_" so that they are unlikely to clash
1503 with other external names.
1504
1505 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1506 a different (DFA) algorithm. Although it is slower than the original
1507 function, it does have some advantages for certain types of matching
1508 problem.
1509
1510 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1511 including restarting after a partial match.
1512
1513 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1514 defined when compiling for Windows was sent to me. I have put it into the
1515 code, though I have no means of testing or verifying it.
1516
1517 7. Added the pcre_refcount() auxiliary function.
1518
1519 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1520 match before or at the first newline in the subject string. In pcretest,
1521 the /f option on a pattern can be used to set this.
1522
1523 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1524 would behave wrongly. This has been present in PCRE since release 4.0.
1525
1526 10. A number of changes to the pcregrep command:
1527
1528 (a) Refactored how -x works; insert ^(...)$ instead of setting
1529 PCRE_ANCHORED and checking the length, in preparation for adding
1530 something similar for -w.
1531
1532 (b) Added the -w (match as a word) option.
1533
1534 (c) Refactored the way lines are read and buffered so as to have more
1535 than one at a time available.
1536
1537 (d) Implemented a pcregrep test script.
1538
1539 (e) Added the -M (multiline match) option. This allows patterns to match
1540 over several lines of the subject. The buffering ensures that at least
1541 8K, or the rest of the document (whichever is the shorter) is available
1542 for matching (and similarly the previous 8K for lookbehind assertions).
1543
1544 (f) Changed the --help output so that it now says
1545
1546 -w, --word-regex(p)
1547
1548 instead of two lines, one with "regex" and the other with "regexp"
1549 because that confused at least one person since the short forms are the
1550 same. (This required a bit of code, as the output is generated
1551 automatically from a table. It wasn't just a text change.)
1552
1553 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1554 option but starts with a hyphen. Could be a pattern or a path name
1555 starting with a hyphen, for instance.
1556
1557 (h) "-" can be given as a file name to represent stdin.
1558
1559 (i) When file names are being printed, "(standard input)" is used for
1560 the standard input, for compatibility with GNU grep. Previously
1561 "<stdin>" was used.
1562
1563 (j) The option --label=xxx can be used to supply a name to be used for
1564 stdin when file names are being printed. There is no short form.
1565
1566 (k) Re-factored the options decoding logic because we are going to add
1567 two more options that take data. Such options can now be given in four
1568 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1569
1570 (l) Added the -A, -B, and -C options for requesting that lines of context
1571 around matches be printed.
1572
1573 (m) Added the -L option to print the names of files that do not contain
1574 any matching lines, that is, the complement of -l.
1575
1576 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1577 continue to scan other files.
1578
1579 (o) The -s option was incorrectly implemented. For compatibility with other
1580 greps, it now suppresses the error message for a non-existent or non-
1581 accessible file (but not the return code). There is a new option called
1582 -q that suppresses the output of matching lines, which was what -s was
1583 previously doing.
1584
1585 (p) Added --include and --exclude options to specify files for inclusion
1586 and exclusion when recursing.
1587
1588 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1589 Hopefully, it now does.
1590
1591 12. Missing cast in pcre_study().
1592
1593 13. Added an "uninstall" target to the makefile.
1594
1595 14. Replaced "extern" in the function prototypes in Makefile.in with
1596 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1597 world, but is set differently for Windows.
1598
1599 15. Added a second compiling function called pcre_compile2(). The only
1600 difference is that it has an extra argument, which is a pointer to an
1601 integer error code. When there is a compile-time failure, this is set
1602 non-zero, in addition to the error test pointer being set to point to an
1603 error message. The new argument may be NULL if no error number is required
1604 (but then you may as well call pcre_compile(), which is now just a
1605 wrapper). This facility is provided because some applications need a
1606 numeric error indication, but it has also enabled me to tidy up the way
1607 compile-time errors are handled in the POSIX wrapper.
1608
1609 16. Added VPATH=.libs to the makefile; this should help when building with one
1610 prefix path and installing with another. (Or so I'm told by someone who
1611 knows more about this stuff than I do.)
1612
1613 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1614 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1615 match everything, including newlines. This is not POSIX-compatible, but
1616 somebody wanted the feature. From pcretest it can be activated by using
1617 both the P and the s flags.
1618
1619 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1620
1621 19. libpcre.pc was being incorrectly installed as executable.
1622
1623 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1624 it now also looks for '\r' so that it will work unmodified on Windows.
1625
1626 21. Added Google's contributed C++ wrapper to the distribution.
1627
1628 22. Added some untidy missing memory free() calls in pcretest, to keep
1629 Electric Fence happy when testing.
1630
1631
1632
1633 Version 5.0 13-Sep-04
1634 ---------------------
1635
1636 1. Internal change: literal characters are no longer packed up into items
1637 containing multiple characters in a single byte-string. Each character
1638 is now matched using a separate opcode. However, there may be more than one
1639 byte in the character in UTF-8 mode.
1640
1641 2. The pcre_callout_block structure has two new fields: pattern_position and
1642 next_item_length. These contain the offset in the pattern to the next match
1643 item, and its length, respectively.
1644
1645 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1646 insertion of callouts before each pattern item. Added the /C option to
1647 pcretest to make use of this.
1648
1649 4. On the advice of a Windows user, the lines
1650
1651 #if defined(_WIN32) || defined(WIN32)
1652 _setmode( _fileno( stdout ), 0x8000 );
1653 #endif /* defined(_WIN32) || defined(WIN32) */
1654
1655 have been added to the source of pcretest. This apparently does useful
1656 magic in relation to line terminators.
1657
1658 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1659 for the benefit of those environments where the "b" makes a difference.
1660
1661 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1662 to know about it. I have put a hack into configure.in that adds in code
1663 to set GCC=yes if CC=icc. This seems to end up at a point in the
1664 generated configure script that is early enough to affect the setting of
1665 compiler options, which is what is needed, but I have no means of testing
1666 whether it really works. (The user who reported this had patched the
1667 generated configure script, which of course I cannot do.)
1668
1669 LATER: After change 22 below (new libtool files), the configure script
1670 seems to know about icc (and also ecc). Therefore, I have commented out
1671 this hack in configure.in.
1672
1673 7. Added support for pkg-config (2 patches were sent in).
1674
1675 8. Negated POSIX character classes that used a combination of internal tables
1676 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1677 [[:^ascii]]. Typically, they would match almost any characters. The other
1678 POSIX classes were not broken in this way.
1679
1680 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1681 to find the match, as PCRE was deluded into thinking that the match had to
1682 start at the start point or following a newline. The same bug applied to
1683 patterns with negative forward assertions or any backward assertions
1684 preceding ".*" at the start, unless the pattern required a fixed first
1685 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1686
1687 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1688 starting at the last subject character, bytes beyond the end of the subject
1689 string were read.
1690
1691 11. Renamed the variable "class" as "classbits" to make life easier for C++
1692 users. (Previously there was a macro definition, but it apparently wasn't
1693 enough.)
1694
1695 12. Added the new field "tables" to the extra data so that tables can be passed
1696 in at exec time, or the internal tables can be re-selected. This allows
1697 a compiled regex to be saved and re-used at a later time by a different
1698 program that might have everything at different addresses.
1699
1700 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1701 -R library as well as a -L library.
1702
1703 14. The debugging options of pcretest (-d on the command line or D on a
1704 pattern) showed incorrect output for anything following an extended class
1705 that contained multibyte characters and which was followed by a quantifier.
1706
1707 15. Added optional support for general category Unicode character properties
1708 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1709 support. It adds about 90K to the size of the library. The meanings of the
1710 inbuilt class escapes such as \d and \s have NOT been changed.
1711
1712 16. Updated pcredemo.c to include calls to free() to release the memory for the
1713 compiled pattern.
1714
1715 17. The generated file chartables.c was being created in the source directory
1716 instead of in the building directory. This caused the build to fail if the
1717 source directory was different from the building directory, and was
1718 read-only.
1719
1720 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1721 file. No doubt somebody will tell me if they don't make sense... Also added
1722 Dan Mooney's comments about building on OpenVMS.
1723
1724 19. Added support for partial matching via the PCRE_PARTIAL option for
1725 pcre_exec() and the \P data escape in pcretest.
1726
1727 20. Extended pcretest with 3 new pattern features:
1728
1729 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1730 write the compiled pattern to the file whose name is "rest-of-line".
1731 This is a straight binary dump of the data, with the saved pointer to
1732 the character tables forced to be NULL. The study data, if any, is
1733 written too. After writing, pcretest reads a new pattern.
1734
1735 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1736 compiled pattern from the given file. There must not be any
1737 occurrences of "<" in the file name (pretty unlikely); if there are,
1738 pcretest will instead treat the initial "<" as a pattern delimiter.
1739 After reading in the pattern, pcretest goes on to read data lines as
1740 usual.
1741
1742 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1743 and 16-bit fields in a compiled pattern, to simulate a pattern that
1744 was compiled on a host of opposite endianness.
1745
1746 21. The pcre-exec() function can now cope with patterns that were compiled on
1747 hosts of opposite endianness, with this restriction:
1748
1749 As for any compiled expression that is saved and used later, the tables
1750 pointer field cannot be preserved; the extra_data field in the arguments
1751 to pcre_exec() should be used to pass in a tables address if a value
1752 other than the default internal tables were used at compile time.
1753
1754 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1755 now diagnosed as an error. Previously, most of the time, a negative number
1756 would have been treated as zero, but if in addition "ovector" was passed as
1757 NULL, a crash could occur.
1758
1759 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1760 new versions from the libtool 1.5 distribution (the last one is a copy of
1761 a file called libtool.m4). This seems to have fixed the need to patch
1762 "configure" to support Darwin 1.3 (which I used to do). However, I still
1763 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1764 workstation).
1765
1766 24. Changed the PCRE licence to be the more standard "BSD" licence.
1767
1768
1769 Version 4.5 01-Dec-03
1770 ---------------------
1771
1772 1. There has been some re-arrangement of the code for the match() function so
1773 that it can be compiled in a version that does not call itself recursively.
1774 Instead, it keeps those local variables that need separate instances for
1775 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1776 needs to "recurse". Keeping track of where control must go is done by means
1777 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1778 hide most of the details from the main code, and operates only if
1779 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1780 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1781 operating.
1782
1783 To make it easier for callers to provide specially tailored get/free
1784 functions for this usage, two new functions, pcre_stack_malloc, and
1785 pcre_stack_free, are used. They are always called in strict stacking order,
1786 and the size of block requested is always the same.
1787
1788 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1789 PCRE has been compiled to use the stack or the heap for recursion. The
1790 -C option of pcretest uses this to show which version is compiled.
1791
1792 A new data escape \S, is added to pcretest; it causes the amounts of store
1793 obtained and freed by both kinds of malloc/free at match time to be added
1794 to the output.
1795
1796 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1797 what's available on my current Linux desktop machine.
1798
1799 3. When matching a UTF-8 string, the test for a valid string at the start has
1800 been extended. If start_offset is not zero, PCRE now checks that it points
1801 to a byte that is the start of a UTF-8 character. If not, it returns
1802 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1803 this is necessary because there may be backward assertions in the pattern.
1804 When matching the same subject several times, it may save resources to use
1805 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1806
1807 4. The code for checking the validity of UTF-8 strings has been tightened so
1808 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1809 containing "overlong sequences".
1810
1811 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1812 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1813 should have been "&", but it just so happened that all the cases this let
1814 through by mistake were picked up later in the function.
1815
1816 6. I had used a variable called "isblank" - this is a C99 function, causing
1817 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1818
1819 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1820 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1821 so the version is shown; (c) stop "make test" echoing "./RunTest".
1822
1823 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1824
1825 9. The prototype for memmove() for systems that don't have it was using
1826 size_t, but the inclusion of the header that defines size_t was later. I've
1827 moved the #includes for the C headers earlier to avoid this.
1828
1829 10. Added some adjustments to the code to make it easier to compiler on certain
1830 special systems:
1831
1832 (a) Some "const" qualifiers were missing.
1833 (b) Added the macro EXPORT before all exported functions; by default this
1834 is defined to be empty.
1835 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1836 that it reads its output file name as an argument instead of writing
1837 to the standard output and assuming this can be redirected.
1838
1839 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1840 class containing characters with values greater than 255, PCRE compilation
1841 went into a loop.
1842
1843 12. A recursive reference to a subpattern that was within another subpattern
1844 that had a minimum quantifier of zero caused PCRE to crash. For example,
1845 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1846 recursion. If the recursively-called subpattern itself had a zero repeat,
1847 that was OK.
1848
1849 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1850 buffer into which it was copied (for escape processing) was still set at
1851 1024, so long lines caused crashes.
1852
1853 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1854 "internal error: code overflow...". This applied to any character class
1855 that was followed by a possessive quantifier.
1856
1857 15. Modified the Makefile to add libpcre.la as a prerequisite for
1858 libpcreposix.la because I was told this is needed for a parallel build to
1859 work.
1860
1861 16. If a pattern that contained .* following optional items at the start was
1862 studied, the wrong optimizing data was generated, leading to matching
1863 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1864 matching string must start with a or b or c. The correct conclusion for
1865 this pattern is that a match can start with any character.
1866
1867
1868 Version 4.4 13-Aug-03
1869 ---------------------
1870
1871 1. In UTF-8 mode, a character class containing characters with values between
1872 127 and 255 was not handled correctly if the compiled pattern was studied.
1873 In fixing this, I have also improved the studying algorithm for such
1874 classes (slightly).
1875
1876 2. Three internal functions had redundant arguments passed to them. Removal
1877 might give a very teeny performance improvement.
1878
1879 3. Documentation bug: the value of the capture_top field in a callout is *one
1880 more than* the number of the hightest numbered captured substring.
1881
1882 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1883 in incorrectly linking with a previously installed version. They now link
1884 explicitly with libpcre.la.
1885
1886 5. configure.in no longer needs to recognize Cygwin specially.
1887
1888 6. A problem in pcre.in for Windows platforms is fixed.
1889
1890 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1891 pcretest, it used to include the size of the study block as part of its
1892 output. Unfortunately, the structure contains a field that has a different
1893 size on different hardware architectures. This meant that the tests that
1894 showed this size failed. As the block is currently always of a fixed size,
1895 this information isn't actually particularly useful in pcretest output, so
1896 I have just removed it.
1897
1898 8. Three pre-processor statements accidentally did not start in column 1.
1899 Sadly, there are *still* compilers around that complain, even though
1900 standard C has not required this for well over a decade. Sigh.
1901
1902 9. In pcretest, the code for checking callouts passed small integers in the
1903 callout_data field, which is a void * field. However, some picky compilers
1904 complained about the casts involved for this on 64-bit systems. Now
1905 pcretest passes the address of the small integer instead, which should get
1906 rid of the warnings.
1907
1908 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1909 both compile and run time, and gives an error if an invalid UTF-8 sequence
1910 is found. There is a option for disabling this check in cases where the
1911 string is known to be correct and/or the maximum performance is wanted.
1912
1913 11. In response to a bug report, I changed one line in Makefile.in from
1914
1915 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1916 to
1917 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1918
1919 to look similar to other lines, but I have no way of telling whether this
1920 is the right thing to do, as I do not use Windows. No doubt I'll get told
1921 if it's wrong...
1922
1923
1924 Version 4.3 21-May-03
1925 ---------------------
1926
1927 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1928 Makefile.
1929
1930 2. Some refactoring to improve the quality of the code:
1931
1932 (i) The utf8_table... variables are now declared "const".
1933
1934 (ii) The code for \cx, which used the "case flipping" table to upper case
1935 lower case letters, now just substracts 32. This is ASCII-specific,
1936 but the whole concept of \cx is ASCII-specific, so it seems
1937 reasonable.
1938
1939 (iii) PCRE was using its character types table to recognize decimal and
1940 hexadecimal digits in the pattern. This is silly, because it handles
1941 only 0-9, a-f, and A-F, but the character types table is locale-
1942 specific, which means strange things might happen. A private
1943 table is now used for this - though it costs 256 bytes, a table is
1944 much faster than multiple explicit tests. Of course, the standard
1945 character types table is still used for matching digits in subject
1946 strings against \d.
1947
1948 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1949 ending in _t are). So I've renamed it as ESC_tee.
1950
1951 3. The first argument for regexec() in the POSIX wrapper should have been
1952 defined as "const".
1953
1954 4. Changed pcretest to use malloc() for its buffers so that they can be
1955 Electric Fenced for debugging.
1956
1957 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1958 to read one or more bytes before the start of the subject string. Often this
1959 had no effect on PCRE's behaviour, but in some circumstances it could
1960 provoke a segmentation fault.
1961
1962 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1963 to try to read one or more bytes before the start of the subject string.
1964
1965 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1966 UTF-8 support could misbehave in various ways if the subject string
1967 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1968 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1969 back over UTF-8 characters.)
1970
1971
1972 Version 4.2 14-Apr-03
1973 ---------------------
1974
1975 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1976
1977 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1978 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1979 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1980 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1981 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1982 and BUILD_EXEEXT
1983 Note: automatic setting of the BUILD variables is not yet working
1984 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1985 compile-time but not at link-time
1986 [LINK]: use for linking executables only
1987 make different versions for Windows and non-Windows
1988 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1989 libraries
1990 [LINK_FOR_BUILD]: new variable
1991 [OBJEXT]: use throughout
1992 [EXEEXT]: use throughout
1993 <winshared>: new target
1994 <wininstall>: new target
1995 <dftables.o>: use native compiler
1996 <dftables>: use native linker
1997 <install>: handle Windows platform correctly
1998 <clean>: ditto
1999 <check>: ditto
2000 copy DLL to top builddir before testing
2001
2002 As part of these changes, -no-undefined was removed again. This was reported
2003 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2004 in any case.
2005
2006 3. Some tidies to get rid of compiler warnings:
2007
2008 . In the match_data structure, match_limit was an unsigned long int, whereas
2009 match_call_count was an int. I've made them both unsigned long ints.
2010
2011 . In pcretest the fact that a const uschar * doesn't automatically cast to
2012 a void * provoked a warning.
2013
2014 . Turning on some more compiler warnings threw up some "shadow" variables
2015 and a few more missing casts.
2016
2017 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2018 option, a class that contained a single character with a value between 128
2019 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2020
2021 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2022 option, a class that contained several characters, but with at least one
2023 whose value was between 128 and 255 caused PCRE to crash.
2024
2025
2026 Version 4.1 12-Mar-03
2027 ---------------------
2028
2029 1. Compiling with gcc -pedantic found a couple of places where casts were
2030 needed, and a string in dftables.c that was longer than standard compilers are
2031 required to support.
2032
2033 2. Compiling with Sun's compiler found a few more places where the code could
2034 be tidied up in order to avoid warnings.
2035
2036 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2037 first of these names is deprecated in the latest Autoconf in favour of the name
2038 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2039 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2040 analogy I have changed it to CFLAGS_FOR_BUILD.
2041
2042 4. Added -no-undefined to the linking command in the Makefile, because this is
2043 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2044 linking step for the pcreposix library.
2045
2046 5. PCRE was failing to diagnose the case of two named groups with the same
2047 name.
2048
2049 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2050 literal character that is needed in the subject for a match, and scans along to
2051 ensure that it is present before embarking on the full matching process. This
2052 saves time in cases of nested unlimited repeats that are never going to match.
2053 Problem: the scan can take a lot of time if the subject is very long (e.g.
2054 megabytes), thus penalizing straightforward matches. It is now done only if the
2055 amount of subject to be scanned is less than 1000 bytes.
2056
2057 7. A lesser problem with the same optimization is that it was recording the
2058 first character of an anchored pattern as "needed", thus provoking a search
2059 right along the subject, even when the first match of the pattern was going to
2060 fail. The "needed" character is now not set for anchored patterns, unless it
2061 follows something in the pattern that is of non-fixed length. Thus, it still
2062 fulfils its original purpose of finding quick non-matches in cases of nested
2063 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2064
2065
2066 Version 4.0 17-Feb-03
2067 ---------------------
2068
2069 1. If a comment in an extended regex that started immediately after a meta-item
2070 extended to the end of string, PCRE compiled incorrect data. This could lead to
2071 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2072
2073 2. Moved to autoconf 2.53 and libtool 1.4.2.
2074
2075 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2076 the special perltest8 script is no longer needed - all the tests can be run
2077 from a single perltest script.
2078
2079 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2080 by \s. It has now been removed in PCRE. This means it isn't recognized as
2081 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2082 class [:space:] *does* include VT, thereby creating a mess.
2083
2084 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2085 space and tab.
2086
2087 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2088 its new features into the main test script, reducing the number of scripts.
2089
2090 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2091 were backward compatible, and made the (?i) apply to the whole pattern, as if
2092 /i were given. Now it behaves more logically, and applies the option setting
2093 only to what follows. PCRE has been changed to follow suit. However, if it
2094 finds options settings right at the start of the pattern, it extracts them into
2095 the global options, as before. Thus, they show up in the info data.
2096
2097 8. Added support for the \Q...\E escape sequence. Characters in between are
2098 treated as literals. This is slightly different from Perl in that $ and @ are
2099 also handled as literals inside the quotes. In Perl, they will cause variable
2100 interpolation. Note the following examples:
2101
2102 Pattern PCRE matches Perl matches
2103
2104 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2105 \Qabc\$xyz\E abc\$xyz abc\$xyz
2106 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2107
2108 For compatibility with Perl, \Q...\E sequences are recognized inside character
2109 classes as well as outside them.
2110
2111 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2112 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2113 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2114 signed/unsigned warnings.
2115
2116 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2117 option for pcretest, so I've replaced it by a simple function that does just
2118 that job.
2119
2120 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2121 "pcregrep -".
2122
2123 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2124 Java package. This provides some syntactic sugar for simple cases of what my
2125 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2126 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2127 item, you can use this simplified notation. Note that only makes sense with
2128 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2129 greediness, whatever the setting of the PCRE_UNGREEDY option.
2130
2131 13. A change of greediness default within a pattern was not taking effect at
2132 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2133 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2134 was abstracted outside.
2135
2136 14. PCRE now supports the \G assertion. It is true when the current matching
2137 position is at the start point of the match. This differs from \A when the
2138 starting offset is non-zero. Used with the /g option of pcretest (or similar
2139 code), it works in the same way as it does for Perl's /g option. If all
2140 alternatives of a regex begin with \G, the expression is anchored to the start
2141 match position, and the "anchored" flag is set in the compiled expression.
2142
2143 15. Some bugs concerning the handling of certain option changes within patterns
2144 have been fixed. These applied to options other than (?ims). For example,
2145 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2146 been the other way round. Some of this was related to change 7 above.
2147
2148 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2149 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2150 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2151 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2152
2153 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2154 mode. Unlike ".", it always matches newline, whatever the setting of
2155 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2156 assertions. Perl allows it, but it doesn't (in general) work because it can't
2157 calculate the length of the lookbehind. At least, that's the case for Perl
2158 5.8.0 - I've been told they are going to document that it doesn't work in
2159 future.
2160
2161 18. Added an error diagnosis for escapes that PCRE does not support: these are
2162 \L, \l, \N, \P, \p, \U, \u, and \X.
2163
2164 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2165 reading past the end of the pattern in cases such as /[abcd/.
2166
2167 20. PCRE was getting more memory than necessary for patterns with classes that
2168 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2169
2170 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2171 compiling PCRE for use with Virtual Pascal.
2172
2173 22. Small fix to the Makefile to make it work properly if the build is done
2174 outside the source tree.
2175
2176 23. Added a new extension: a condition to go with recursion. If a conditional
2177 subpattern starts with (?(R) the "true" branch is used if recursion has
2178 happened, whereas the "false" branch is used only at the top level.
2179
2180 24. When there was a very long string of literal characters (over 255 bytes
2181 without UTF support, over 250 bytes with UTF support), the computation of how
2182 much memory was required could be incorrect, leading to segfaults or other
2183 strange effects.
2184
2185 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2186 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2187 there was a subsequent back reference to those brackets. This meant that, for
2188 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2189 possible to check for precisely this case. All we can do is abandon the
2190 optimization if .* occurs inside capturing brackets when there are any back
2191 references whatsoever. (See below for a better fix that came later.)
2192
2193 26. The handling of the optimization for finding the first character of a
2194 non-anchored pattern, and for finding a character that is required later in the
2195 match were failing in some cases. This didn't break the matching; it just
2196 failed to optimize when it could. The way this is done has been re-implemented.
2197
2198 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2199
2200 28. Added a new feature that provides some of the functionality that Perl
2201 provides with (?{...}). The facility is termed a "callout". The way it is done
2202 in PCRE is for the caller to provide an optional function, by setting
2203 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2204 global variable. By default it is unset, which disables all calling out. To get
2205 the function called, the regex must include (?C) at appropriate points. This
2206 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2207 This provides a means of identifying different callout points. When PCRE
2208 reaches such a point in the regex, if pcre_callout has been set, the external
2209 function is called. It is provided with data in a structure called
2210 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2211 matching continues; if it returns a non-zero value, the match at the current
2212 point fails. However, backtracking will occur if possible. [This was changed
2213 later and other features added - see item 49 below.]
2214
2215 29. pcretest is upgraded to test the callout functionality. It provides a
2216 callout function that displays information. By default, it shows the start of
2217 the match and the current position in the text. There are some new data escapes
2218 to vary what happens:
2219
2220 \C+ in addition, show current contents of captured substrings
2221 \C- do not supply a callout function
2222 \C!n return 1 when callout number n is reached
2223 \C!n!m return 1 when callout number n is reached for the mth time
2224
2225 30. If pcregrep was called with the -l option and just a single file name, it
2226 output "<stdin>" if a match was found, instead of the file name.
2227
2228 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2229 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2230 pcre_exec(). This saves a malloc/free per call. The default value of
2231 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2232 when configuring.
2233
2234 32. The default maximum size of a compiled pattern is 64K. There have been a
2235 few cases of people hitting this limit. The code now uses macros to handle the
2236 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2237 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2238 configuring. Tests 2 and 5 work only with 2-byte links because they output
2239 debugging information about compiled patterns.
2240
2241 33. Internal code re-arrangements:
2242
2243 (a) Moved the debugging function for printing out a compiled regex into
2244 its own source file (printint.c) and used #include to pull it into
2245 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2246 separate copies.
2247
2248 (b) Defined the list of op-code names for debugging as a macro in
2249 internal.h so that it is next to the definition of the opcodes.
2250
2251 (c) Defined a table of op-code lengths for simpler skipping along compiled
2252 code. This is again a macro in internal.h so that it is next to the
2253 definition of the opcodes.
2254
2255 34. Added support for recursive calls to individual subpatterns, along the
2256 lines of Robin Houston's patch (but implemented somewhat differently).
2257
2258 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2259 allow it to read and process whole directories in Win32. This code was
2260 contributed by Lionel Fourquaux; it has not been tested by me.
2261
2262 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2263 used to name a group. Names consist of alphanumerics and underscores, and must
2264 be unique. Back references use the syntax (?P=name) and recursive calls use
2265 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2266 numbers. The function pcre_fullinfo() can be used after compilation to extract
2267 a name/number map. There are three relevant calls:
2268
2269 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2270 PCRE_INFO_NAMECOUNT yields the number of entries
2271 PCRE_INFO_NAMETABLE yields a pointer to the map.
2272
2273 The map is a vector of fixed-size entries. The size of each entry depends on
2274 the length of the longest name used. The first two bytes of each entry are the
2275 group number, most significant byte first. There follows the corresponding
2276 name, zero terminated. The names are in alphabetical order.
2277
2278 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2279 case instead of 255. Making it the same both with and without UTF-8 support
2280 means that the same test output works with both.
2281
2282 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2283 calling malloc() with a zero argument.
2284
2285 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2286 optimization. I've improved things by keeping a bitmap of backreferences with
2287 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2288 fact referenced, the optimization can be applied. It is unlikely that a
2289 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2290 the match to follow \n) will appear inside brackets with a number greater than
2291 31, but if it does, any back reference > 31 suppresses the optimization.
2292
2293 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2294 of disabling numbered capturing parentheses. Any opening parenthesis that is
2295 not followed by ? behaves as if it were followed by ?: but named parentheses
2296 can still be used for capturing (and they will acquire numbers in the usual
2297 way).
2298
2299 41. Redesigned the return codes from the match() function into yes/no/error so
2300 that errors can be passed back from deep inside the nested calls. A malloc
2301 failure while inside a recursive subpattern call now causes the
2302 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2303
2304 42. It is now possible to set a limit on the number of times the match()
2305 function is called in a call to pcre_exec(). This facility makes it possible to
2306 limit the amount of recursion and backtracking, though not in a directly
2307 obvious way, because the match() function is used in a number of different
2308 circumstances. The count starts from zero for each position in the subject
2309 string (for non-anchored patterns). The default limit is, for compatibility, a
2310 large number, namely 10 000 000. You can change this in two ways:
2311
2312 (a) When configuring PCRE before making, you can use --with-match-limit=n
2313 to set a default value for the compiled library.
2314
2315 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2316 a different value is set. See 45 below.
2317
2318 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2319
2320 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2321 of things that can be changed at compile time. The first argument specifies
2322 what is wanted and the second points to where the information is to be placed.
2323 The current list of available information is:
2324
2325 PCRE_CONFIG_UTF8
2326
2327 The output is an integer that is set to one if UTF-8 support is available;
2328 otherwise it is set to zero.
2329
2330 PCRE_CONFIG_NEWLINE
2331
2332 The output is an integer that it set to the value of the code that is used for
2333 newline. It is either LF (10) or CR (13).
2334
2335 PCRE_CONFIG_LINK_SIZE
2336
2337 The output is an integer that contains the number of bytes used for internal
2338 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2339
2340 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2341
2342 The output is an integer that contains the threshold above which the POSIX
2343 interface uses malloc() for output vectors. See item 31 above.
2344
2345 PCRE_CONFIG_MATCH_LIMIT
2346
2347 The output is an unsigned integer that contains the default limit of the number
2348 of match() calls in a pcre_exec() execution. See 42 above.
2349
2350 44. pcretest has been upgraded by the addition of the -C option. This causes it
2351 to extract all the available output from the new pcre_config() function, and to
2352 output it. The program then exits immediately.
2353
2354 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2355 order to support additional features. One way would have been to define
2356 pcre_exec2() (for example) with extra arguments, but this would not have been
2357 extensible, and would also have required all calls to the original function to
2358 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2359 is used for passing in "extra" data from pcre_study().
2360
2361 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2362 contains the following fields:
2363
2364 flags a bitmap indicating which of the following fields are set
2365 study_data opaque data from pcre_study()
2366 match_limit a way of specifying a limit on match() calls for a specific
2367 call to pcre_exec()
2368 callout_data data for callouts (see 49 below)
2369
2370 The flag bits are also defined in pcre.h, and are
2371
2372 PCRE_EXTRA_STUDY_DATA
2373 PCRE_EXTRA_MATCH_LIMIT
2374 PCRE_EXTRA_CALLOUT_DATA
2375
2376 The pcre_study() function now returns one of these new pcre_extra blocks, with
2377 the actual study data pointed to by the study_data field, and the
2378 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2379 before. That is, this change is entirely upwards-compatible and requires no
2380 change to existing code.
2381
2382 If you want to pass in additional data to pcre_exec(), you can either place it
2383 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2384 block.
2385
2386 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2387 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2388 times with different match limits, until it finds the minimum value needed for
2389 pcre_exec() to complete. The value is then output. This can be instructive; for
2390 most simple matches the number is quite small, but for pathological cases it
2391 gets very large very quickly.
2392
2393 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2394 returns the size of the data block pointed to by the study_data field in a
2395 pcre_extra block, that is, the value that was passed as the argument to
2396 pcre_malloc() when PCRE was getting memory in which to place the information
2397 created by pcre_study(). The fourth argument should point to a size_t variable.
2398 pcretest has been extended so that this information is shown after a successful
2399 pcre_study() call when information about the compiled regex is being displayed.
2400
2401 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2402 because what follows is always an absolute path. (Later: it turns out that this
2403 is more than cosmetic for MinGW, because it doesn't like empty path
2404 components.)
2405
2406 49. Some changes have been made to the callout feature (see 28 above):
2407
2408 (i) A callout function now has three choices for what it returns:
2409
2410 0 => success, carry on matching
2411 > 0 => failure at this point, but backtrack if possible
2412 < 0 => serious error, return this value from pcre_exec()
2413
2414 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2415 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2416 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2417 use by callout functions. It will never be used by PCRE itself.
2418
2419 (ii) The pcre_extra structure (see 45 above) has a void * field called
2420 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2421 pcre_callout_block structure has a field of the same name. The contents of
2422 the field passed in the pcre_extra structure are passed to the callout
2423 function in the corresponding field in the callout block. This makes it
2424 easier to use the same callout-containing regex from multiple threads. For
2425 testing, the pcretest program has a new data escape
2426
2427 \C*n pass the number n (may be negative) as callout_data
2428
2429 If the callout function in pcretest receives a non-zero value as
2430 callout_data, it returns that value.
2431
2432 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2433 there were some redundant $(CFLAGS) in commands that are now specified as
2434 $(LINK), which already includes $(CFLAGS).
2435
2436 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2437 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2438 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2439 one-byte characters throughout. Note that case-insensitive matching applies
2440 only to characters whose values are less than 256. PCRE doesn't support the
2441 notion of cases for higher-valued characters.
2442
2443 (i) A character class whose characters are all within 0-255 is handled as
2444 a bit map, and the map is inverted for negative classes. Previously, a
2445 character > 255 always failed to match such a class; however it should
2446 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2447
2448 (ii) A negated character class with a single character < 255 is coded as
2449 "not this character" (OP_NOT). This wasn't working properly when the test
2450 character was multibyte, either singly or repeated.
2451
2452 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2453 mode, for example: \x{100}{2,3}.
2454
2455 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2456 singly or repeated) now correctly test multibyte characters. However,
2457 PCRE doesn't recognize any characters with values greater than 255 as
2458 digits, spaces, or word characters. Such characters always match \D, \S,
2459 and \W, and never match \d, \s, or \w.
2460
2461 (v) Classes may now contain characters and character ranges with values
2462 greater than 255. For example: [ab\x{100}-\x{400}].
2463
2464 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2465 PCRE in UTF-8 mode.
2466
2467 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2468 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2469 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2470 value.)
2471
2472 53. The single man page has become too large. I have therefore split it up into
2473 a number of separate man pages. These also give rise to individual HTML pages;
2474 these are now put in a separate directory, and there is an index.html page that
2475 lists them all. Some hyperlinking between the pages has been installed.
2476
2477 54. Added convenience functions for handling named capturing parentheses.
2478
2479 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2480 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2481 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2482 are faulted.
2483
2484 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2485 calling configure. These values are used when compiling the dftables.c program
2486 which is run to generate the source of the default character tables. They
2487 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2488 you will need to set these values.
2489
2490 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2491
2492
2493 Version 3.9 02-Jan-02
2494 ---------------------
2495
2496 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2497
2498 2. If --disable-static was given, the building process failed when trying to
2499 build pcretest and pcregrep. (For some reason it was using libtool to compile
2500 them, which is not right, as they aren't part of the library.)
2501
2502
2503 Version 3.8 18-Dec-01
2504 ---------------------
2505
2506 1. The experimental UTF-8 code was completely screwed up. It was packing the
2507 bytes in the wrong order. How dumb can you get?
2508
2509
2510 Version 3.7 29-Oct-01
2511 ---------------------
2512
2513 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2514 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2515 this didn't happen under Solaris 8, where I normally test things.
2516
2517 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2518 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2519 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2520 this fix an hour or so after the initial 3.7 release.)
2521
2522
2523 Version 3.6 23-Oct-01
2524 ---------------------
2525
2526 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2527 offsets passed as NULL with zero offset count.
2528
2529 2. The config.guess and config.sub files had not been updated when I moved to
2530 the latest autoconf.
2531
2532
2533 Version 3.5 15-Aug-01
2534 ---------------------
2535
2536 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2537 had been forgotten.
2538
2539 2. By using declared but undefined structures, we can avoid using "void"
2540 definitions in pcre.h while keeping the internal definitions of the structures
2541 private.
2542
2543 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2544 user point of view, this means that both static and shared libraries are built
2545 by default, but this can be individually controlled. More of the work of
2546 handling this static/shared cases is now inside libtool instead of PCRE's make
2547 file.
2548
2549 4. The pcretest utility is now installed along with pcregrep because it is
2550 useful for users (to test regexs) and by doing this, it automatically gets
2551 relinked by libtool. The documentation has been turned into a man page, so
2552 there are now .1, .txt, and .html versions in /doc.
2553
2554 5. Upgrades to pcregrep:
2555 (i) Added long-form option names like gnu grep.
2556 (ii) Added --help to list all options with an explanatory phrase.
2557 (iii) Added -r, --recursive to recurse into sub-directories.
2558 (iv) Added -f, --file to read patterns from a file.
2559
2560 6. pcre_exec() was referring to its "code" argument before testing that
2561 argument for NULL (and giving an error if it was NULL).
2562
2563 7. Upgraded Makefile.in to allow for compiling in a different directory from
2564 the source directory.
2565
2566 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2567 options bits, the pointer it was passed was to an int instead of to an unsigned
2568 long int. This mattered only on 64-bit systems.
2569
2570 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2571 generated) instead of pcre.in, which it its source. Also made the same change
2572 in several of the .c files.
2573
2574 10. A new release of gcc defines printf() as a macro, which broke pcretest
2575 because it had an ifdef in the middle of a string argument for printf(). Fixed
2576 by using separate calls to printf().
2577
2578 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2579 script, to force use of CR or LF instead of \n in the source. On non-Unix
2580 systems, the value can be set in config.h.
2581
2582 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2583 absolute limit. Changed the text of the error message to make this clear, and
2584 likewise updated the man page.
2585
2586 13. The limit of 99 on the number of capturing subpatterns has been removed.
2587 The new limit is 65535, which I hope will not be a "real" limit.
2588
2589
2590 Version 3.4 22-Aug-00
2591 ---------------------
2592
2593 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2594
2595 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2596
2597
2598 Version 3.3 01-Aug-00
2599 ---------------------
2600
2601 1. If an octal character was given, but the value was greater than \377, it
2602 was not getting masked to the least significant bits, as documented. This could
2603 lead to crashes in some systems.
2604
2605 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2606 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2607
2608 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2609 These just pass their arguments on to (pcre_free)(), but they are provided
2610 because some uses of PCRE bind it to non-C systems that can call its functions,
2611 but cannot call free() or pcre_free() directly.
2612
2613 4. Add "make test" as a synonym for "make check". Corrected some comments in
2614 the Makefile.
2615
2616 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2617 Makefile.
2618
2619 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2620 command called pgrep for grepping around the active processes.
2621
2622 7. Added the beginnings of support for UTF-8 character strings.
2623
2624 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2625 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2626 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2627 out for the ar command.)
2628
2629
2630 Version 3.2 12-May-00
2631 ---------------------
2632
2633 This is purely a bug fixing release.
2634
2635 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2636 of ZA. This was just one example of several cases that could provoke this bug,
2637 which was introduced by change 9 of version 2.00. The code for breaking
2638 infinite loops after an iteration that matches an empty string was't working
2639 correctly.
2640
2641 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2642 when matched against abbab (for example). After matching an empty string, it
2643 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2644 caused it to match further down the string than it should.
2645
2646 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2647 was there because it doesn't seem to be needed, and it causes trouble on some
2648 systems, as it is not a Standard C header. It has been removed.
2649
2650 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2651 were reported on the Macintosh. The changes were from
2652
2653 while ((c = *(++ptr)) != 0 && c != '\n');
2654 to
2655 while ((c = *(++ptr)) != 0 && c != '\n') ;
2656
2657 Totally extraordinary, but if that's what it takes...
2658
2659 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2660 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2661 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2662 assumes the way PCRE uses memmove() (always moving upwards).
2663
2664 6. PCRE is being used in one environment where strchr() is not available. There
2665 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2666 faster code anyway.
2667
2668
2669 Version 3.1 09-Feb-00
2670 ---------------------
2671
2672 The only change in this release is the fixing of some bugs in Makefile.in for
2673 the "install" target:
2674
2675 (1) It was failing to install pcreposix.h.
2676
2677 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2678
2679
2680 Version 3.0 01-Feb-00
2681 ---------------------
2682
2683 1. Add support for the /+ modifier to perltest (to output $` like it does in
2684 pcretest).
2685
2686 2. Add support for the /g modifier to perltest.
2687
2688 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2689 matches null strings.
2690
2691 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2692 pattern. Perl treats empty patterns specially - it reuses the most recent
2693 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2694 effect.
2695
2696 5. The POSIX interface was broken in that it was just handing over the POSIX
2697 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2698 required a bigger vector, with some working space on the end. This means that
2699 the POSIX wrapper now has to get and free some memory, and copy the results.
2700
2701 6. Added some simple autoconf support, placing the test data and the
2702 documentation in separate directories, re-organizing some of the
2703 information files, and making it build pcre-config (a GNU standard). Also added
2704 libtool support for building PCRE as a shared library, which is now the
2705 default.
2706
2707 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2708 09 are not valid octal constants. Single digits will be used for minor values
2709 less than 10.
2710
2711 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2712 existing programs that set these in the POSIX interface can use PCRE without
2713 modification.
2714
2715 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2716 return all that pcre_info() returns, plus additional data. The pcre_info()
2717 function is retained for compatibility, but is considered to be obsolete.
2718
2719 10. Added experimental recursion feature (?R) to handle one common case that
2720 Perl 5.6 will be able to do with (?p{...}).
2721
2722 11. Added support for POSIX character classes like [:alpha:], which Perl is
2723 adopting.
2724
2725
2726 Version 2.08 31-Aug-99
2727 ----------------------
2728
2729 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2730 trying to match at the startoffset position, but instead was moving forward to
2731 the next newline as if a previous match had failed.
2732
2733 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2734 and could get into a loop if a null string was matched other than at the start
2735 of the subject.
2736
2737 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2738 be distinguished at compile time, and for completeness also added PCRE_DATE.
2739
2740 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2741 in GnuWin32 environments.
2742
2743
2744 Version 2.07 29-Jul-99
2745 ----------------------
2746
2747 1. The documentation is now supplied in plain text form and HTML as well as in
2748 the form of man page sources.
2749
2750 2. C++ compilers don't like assigning (void *) values to other pointer types.
2751 In particular this affects malloc(). Although there is no problem in Standard
2752 C, I've put in casts to keep C++ compilers happy.
2753
2754 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2755 should be (const char *).
2756
2757 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2758 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2759 However, I haven't made this a standard facility. The documentation doesn't
2760 mention it, and the Makefile doesn't support it.
2761
2762 5. The Makefile now contains an "install" target, with editable destinations at
2763 the top of the file. The pcretest program is not installed.
2764
2765 6. pgrep -V now gives the PCRE version number and date.
2766
2767 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2768 causing the entire string to be ignored, instead of just the last character.
2769
2770 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2771 non-matching string, it can take a very, very long time, even for strings of
2772 quite modest length, because of the nested recursion. PCRE now does better in
2773 some of these cases. It does this by remembering the last required literal
2774 character in the pattern, and pre-searching the subject to ensure it is present
2775 before running the real match. In other words, it applies a heuristic to detect
2776 some types of certain failure quickly, and in the above example, if presented
2777 with a string that has no trailing " it gives "no match" very quickly.
2778
2779 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2780 other alternatives are tried instead.
2781
2782
2783 Version 2.06 09-Jun-99
2784 ----------------------
2785
2786 1. Change pcretest's output for amount of store used to show just the code
2787 space, because the remainder (the data block) varies in size between 32-bit and
2788 64-bit systems.
2789
2790 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2791 start matching at. This allows lookbehinds to work when searching for multiple
2792 occurrences in a string.
2793
2794 3. Added additional options to pcretest for testing multiple occurrences:
2795
2796 /+ outputs the rest of the string that follows a match
2797 /g loops for multiple occurrences, using the new startoffset argument
2798 /G loops for multiple occurrences by passing an incremented pointer
2799
2800 4. PCRE wasn't doing the "first character" optimization for patterns starting
2801 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2802 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2803 the letter 'x'. On long subject strings, this gives a significant speed-up.
2804
2805
2806 Version 2.05 21-Apr-99
2807 ----------------------
2808
2809 1. Changed the type of magic_number from int to long int so that it works
2810 properly on 16-bit systems.
2811
2812 2. Fixed a bug which caused patterns starting with .* not to work correctly
2813 when the subject string contained newline characters. PCRE was assuming
2814 anchoring for such patterns in all cases, which is not correct because .* will
2815 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2816 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2817 must be retried after every newline in the subject.
2818
2819
2820 Version 2.04 18-Feb-99
2821 ----------------------
2822
2823 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2824 computation of the store needed to hold the pattern was incorrect (too large).
2825 If such patterns were nested a few deep, this could multiply and become a real
2826 problem.
2827
2828 2. Added /M option to pcretest to show the memory requirement of a specific
2829 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2830
2831 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2832 compiled in such a way that the backtracking after subsequent failure was
2833 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2834 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2835
2836
2837 Version 2.03 02-Feb-99
2838 ----------------------
2839
2840 1. Fixed typo and small mistake in man page.
2841
2842 2. Added 4th condition (GPL supersedes if conflict) and created separate
2843 LICENCE file containing the conditions.
2844
2845 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2846 Perl, that is the internal \ allows the delimiter to be included in the
2847 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2848 the final delimiter, add \ to the end of the pattern (to test the error).
2849
2850 4. Added the convenience functions for extracting substrings after a successful
2851 match. Updated pcretest to make it able to test these functions.
2852
2853
2854 Version 2.02 14-Jan-99
2855 ----------------------
2856
2857 1. Initialized the working variables associated with each extraction so that
2858 their saving and restoring doesn't refer to uninitialized store.
2859
2860 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2861 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2862 fix the problem.
2863
2864 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2865 calls, and wasn't printing the correct value for compiling calls. Increased the
2866 default value of LOOPREPEAT, and the number of significant figures in the
2867 times.
2868
2869 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2870
2871 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2872 a building problem on Windows NT with a FAT file system.
2873
2874
2875 Version 2.01 21-Oct-98
2876 ----------------------
2877
2878 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2879 to character tables built by pcre_maketables() in the current locale. If NULL
2880 is passed, the default tables are used.
2881
2882
2883 Version 2.00 24-Sep-98
2884 ----------------------
2885
2886 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2887 it any more.
2888
2889 2. Allow quantification of (?>) groups, and make it work correctly.
2890
2891 3. The first character computation wasn't working for (?>) groups.
2892
2893 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2894 end of the subject) and add 5.005's \z, which really does match only at the
2895 very end of the subject.
2896
2897 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2898
2899 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2900 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2901 localized options. All options to pcre_study() were also removed.
2902
2903 7. Add other new features from 5.005:
2904
2905 $(?<= positive lookbehind
2906 $(?<! negative lookbehind
2907 (?imsx-imsx) added the unsetting capability
2908 such a setting is global if at outer level; local otherwise
2909 (?imsx-imsx:) non-capturing groups with option setting
2910 (?(cond)re|re) conditional pattern matching
2911
2912 A backreference to itself in a repeated group matches the previous
2913 captured string.
2914
2915 8. General tidying up of studying (both automatic and via "study")
2916 consequential on the addition of new assertions.
2917
2918 9. As in 5.005, unlimited repeated groups that could match an empty substring
2919 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2920 runtime if any iteration does actually match an empty substring.
2921
2922 10. Include the RunTest script in the distribution.
2923
2924 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2925 discrepancies, some of which were old and were also with respect to 5.004. They
2926 have now been fixed.
2927
2928
2929 Version 1.09 28-Apr-98
2930 ----------------------
2931
2932 1. A negated single character class followed by a quantifier with a minimum
2933 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2934 program crashes, or just wrong answers. This did not apply to negated classes
2935 containing more than one character, or to minima other than one.
2936
2937
2938 Version 1.08 27-Mar-98
2939 ----------------------
2940
2941 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2942
2943 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2944 latter must appear before anything that relies on it in the pattern.
2945
2946
2947 Version 1.07 16-Feb-98
2948 ----------------------
2949
2950 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2951 repeat of a potentially empty string).
2952
2953
2954 Version 1.06 23-Jan-98
2955 ----------------------
2956
2957 1. Added Markus Oberhumer's little patches for C++.
2958
2959 2. Literal strings longer than 255 characters were broken.
2960
2961
2962 Version 1.05 23-Dec-97
2963 ----------------------
2964
2965 1. Negated character classes containing more than one character were failing if
2966 PCRE_CASELESS was set at run time.
2967
2968
2969 Version 1.04 19-Dec-97
2970 ----------------------
2971
2972 1. Corrected the man page, where some "const" qualifiers had been omitted.
2973
2974 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2975 input syntax.
2976
2977 3. Fixed memory leak which occurred when a regex with back references was
2978 matched with an offsets vector that wasn't big enough. The temporary memory
2979 that is used in this case wasn't being freed if the match failed.
2980
2981 4. Tidied pcretest to ensure it frees memory that it gets.
2982
2983 5. Temporary memory was being obtained in the case where the passed offsets
2984 vector was exactly big enough.
2985
2986 6. Corrected definition of offsetof() from change 5 below.
2987
2988 7. I had screwed up change 6 below and broken the rules for the use of
2989 setjmp(). Now fixed.
2990
2991
2992 Version 1.03 18-Dec-97
2993 ----------------------
2994
2995 1. A erroneous regex with a missing opening parenthesis was correctly
2996 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2997 on some systems.
2998
2999 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3000 it was reported that one broken compiler failed on the former because "code" is
3001 also an independent variable.
3002
3003 3. The erroneous regex a[]b caused an array overrun reference.
3004
3005 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3006 fail on data ending with that character. (It was going on too far, and checking
3007 the next character, typically a binary zero.) This was specific to the
3008 optimized code for single-character negative classes.
3009
3010 5. Added a contributed patch from the TIN world which does the following:
3011
3012 + Add an undef for memmove, in case the the system defines a macro for it.
3013
3014 + Add a definition of offsetof(), in case there isn't one. (I don't know
3015 the reason behind this - offsetof() is part of the ANSI standard - but
3016 it does no harm).
3017
3018 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3019 most of the places where whitespace preceded '#'. I have given up and
3020 allowed the remaining 2 cases to be at the margin.
3021
3022 + Rename some variables in pcre to eliminate shadowing. This seems very
3023 pedantic, but does no harm, of course.
3024
3025 6. Moved the call to setjmp() into its own function, to get rid of warnings
3026 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3027
3028 7. Constructs such as \d{8,} were compiling into the equivalent of
3029 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3030 outcome, but in this particular case used more store than had been allocated,
3031 which caused the bug to be discovered because it threw up an internal error.
3032
3033 8. The debugging code in both pcre and pcretest for outputting the compiled
3034 form of a regex was going wrong in the case of back references followed by
3035 curly-bracketed repeats.
3036
3037
3038 Version 1.02 12-Dec-97
3039 ----------------------
3040
3041 1. Typos in pcre.3 and comments in the source fixed.
3042
3043 2. Applied a contributed patch to get rid of places where it used to remove
3044 'const' from variables, and fixed some signed/unsigned and uninitialized
3045 variable warnings.
3046
3047 3. Added the "runtest" target to Makefile.
3048
3049 4. Set default compiler flag to -O2 rather than just -O.
3050
3051
3052 Version 1.01 19-Nov-97
3053 ----------------------
3054
3055 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3056 like /([ab]*)*/, that is, for classes with more than one character in them.
3057
3058 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3059 as /((?>a*))*/ (a PCRE_EXTRA facility).
3060
3061
3062 Version 1.00 18-Nov-97
3063 ----------------------
3064
3065 1. Added compile-time macros to support systems such as SunOS4 which don't have
3066 memmove() or strerror() but have other things that can be used instead.
3067
3068 2. Arranged that "make clean" removes the executables.
3069
3070
3071 Version 0.99 27-Oct-97
3072 ----------------------
3073
3074 1. Fixed bug in code for optimizing classes with only one character. It was
3075 initializing a 32-byte map regardless, which could cause it to run off the end
3076 of the memory it had got.
3077
3078 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3079
3080
3081 Version 0.98 22-Oct-97
3082 ----------------------
3083
3084 1. Fixed bug in code for handling temporary memory usage when there are more
3085 back references than supplied space in the ovector. This could cause segfaults.
3086
3087
3088 Version 0.97 21-Oct-97
3089 ----------------------
3090
3091 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3092
3093 2. Optimized negated single characters not to use a bit map.
3094
3095 3. Brought error texts together as macro definitions; clarified some of them;
3096 fixed one that was wrong - it said "range out of order" when it meant "invalid
3097 escape sequence".
3098
3099 4. Changed some char * arguments to const char *.
3100
3101 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3102
3103 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3104 pcretest.
3105
3106
3107 Version 0.96 16-Oct-97
3108 ----------------------
3109
3110 1. Added a simple "pgrep" utility to the distribution.
3111
3112 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3113 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3114 where "ddd" means "one or more decimal digits".
3115
3116 3. Fixed serious bug. If a pattern had a back reference, but the call to
3117 pcre_exec() didn't supply a large enough ovector to record the related
3118 identifying subpattern, the match always failed. PCRE now remembers the number
3119 of the largest back reference, and gets some temporary memory in which to save
3120 the offsets during matching if necessary, in order to ensure that
3121 backreferences always work.
3122
3123 4. Increased the compatibility with Perl in a number of ways:
3124
3125 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3126 to request this handling. The option can be set at compile or exec time.
3127
3128 (b) $ matches before a terminating newline by default; an option
3129 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3130 mode). The option can be set at compile or exec time.
3131
3132 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3133 the same as Perl's. If the decimal number it represents is less than 10
3134 or there aren't that many previous left capturing parentheses, an octal
3135 escape is read. Inside a character class, it's always an octal escape,
3136 even if it is a single digit.
3137
3138 (d) An escaped but undefined alphabetic character is taken as a literal,
3139 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3140 escapes.
3141
3142 (e) {0} is now permitted. (The previous item is removed from the compiled
3143 pattern).
3144
3145 5. Changed all the names of code files so that the basic parts are no longer
3146 than 10 characters, and abolished the teeny "globals.c" file.
3147
3148 6. Changed the handling of character classes; they are now done with a 32-byte
3149 bit map always.
3150
3151 7. Added the -d and /D options to pcretest to make it possible to look at the
3152 internals of compilation without having to recompile pcre.
3153
3154
3155 Version 0.95 23-Sep-97
3156 ----------------------
3157
3158 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3159 \x20 at the start of a run of normal characters. These were being treated as
3160 real characters, instead of the source characters being re-checked.
3161
3162
3163 Version 0.94 18-Sep-97
3164 ----------------------
3165
3166 1. The functions are now thread-safe, with the caveat that the global variables
3167 containing pointers to malloc() and free() or alternative functions are the
3168 same for all threads.
3169
3170 2. Get pcre_study() to generate a bitmap of initial characters for non-
3171 anchored patterns when this is possible, and use it if passed to pcre_exec().
3172
3173
3174 Version 0.93 15-Sep-97
3175 ----------------------
3176
3177 1. /(b)|(:+)/ was computing an incorrect first character.
3178
3179 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3180 but not actually doing anything yet.
3181
3182 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3183 as Perl does (e.g. [-az] or [az-]).
3184
3185 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3186 all possible positions.
3187
3188 5. Split up into different modules to avoid including unneeded functions in a
3189 compiled binary. However, compile and exec are still in one module. The "study"
3190 function is split off.
3191
3192 6. The character tables are now in a separate module whose source is generated
3193 by an auxiliary program - but can then be edited by hand if required. There are
3194 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3195 toupper() in the code.
3196
3197 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3198 make them global. Abolish the function for setting them, as the caller can now
3199 set them directly.
3200
3201
3202 Version 0.92 11-Sep-97
3203 ----------------------
3204
3205 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3206 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3207
3208 2. Caseless matching was not working in character classes if the characters in
3209 the pattern were in upper case.
3210
3211 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3212
3213 4. Make PCRE_ANCHORED public and accept as a compile option.
3214
3215 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3216 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3217 pass them.
3218
3219 6. Give an error if bad option bits passed at compile or run time.
3220
3221 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3222 pcretest to cause it to pass that flag.
3223
3224 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3225 options, and the first character, if set.
3226
3227 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3228
3229
3230 Version 0.91 10-Sep-97
3231 ----------------------
3232
3233 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3234 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3235
3236 2. PCRE was looping on encountering an indefinitely repeated back reference to
3237 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3238 Perl does - treats the match as successful.
3239
3240 ****

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