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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 331 - (show annotations) (download)
Sat Apr 5 15:38:53 2008 UTC (6 years, 4 months ago) by ph10
File size: 152722 byte(s)
Applied Craig's patch to remove the use of push_back().

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. (Subsequently also a second patch to fix the
42 first patch. And a third patch - this was a messy problem.)
43
44 9. Applied Craig's patch to remove the use of push_back().
45
46
47 Version 7.6 28-Jan-08
48 ---------------------
49
50 1. A character class containing a very large number of characters with
51 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
52 overflow.
53
54 2. Patch to cut out the "long long" test in pcrecpp_unittest when
55 HAVE_LONG_LONG is not defined.
56
57 3. Applied Christian Ehrlicher's patch to update the CMake build files to
58 bring them up to date and include new features. This patch includes:
59
60 - Fixed PH's badly added libz and libbz2 support.
61 - Fixed a problem with static linking.
62 - Added pcredemo. [But later removed - see 7 below.]
63 - Fixed dftables problem and added an option.
64 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
65 HAVE_LONG_LONG.
66 - Added readline support for pcretest.
67 - Added an listing of the option settings after cmake has run.
68
69 4. A user submitted a patch to Makefile that makes it easy to create
70 "pcre.dll" under mingw when using Configure/Make. I added stuff to
71 Makefile.am that cause it to include this special target, without
72 affecting anything else. Note that the same mingw target plus all
73 the other distribution libraries and programs are now supported
74 when configuring with CMake (see 6 below) instead of with
75 Configure/Make.
76
77 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
78 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
79 exported in the Windows port". It has not yet been confirmed that the patch
80 solves the problem, but it does no harm.
81
82 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
83 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
84 with CMake, and also correct the comment about stack recursion.
85
86 7. Remove the automatic building of pcredemo from the ./configure system and
87 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
88 of a program that users should build themselves after PCRE is installed, so
89 building it automatically is not really right. What is more, it gave
90 trouble in some build environments.
91
92 8. Further tidies to CMakeLists.txt from Sheri and Christian.
93
94
95 Version 7.5 10-Jan-08
96 ---------------------
97
98 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
99 values in parens when parsing an RE using the C++ wrapper."
100
101 2. Negative specials like \S did not work in character classes in UTF-8 mode.
102 Characters greater than 255 were excluded from the class instead of being
103 included.
104
105 3. The same bug as (2) above applied to negated POSIX classes such as
106 [:^space:].
107
108 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
109 defined or documented. It seems to have been a typo for PCRE_STATIC, so
110 I have changed it.
111
112 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
113 first named subpattern) and a construct such as (?&a) would reference the
114 first named subpattern whose name started with "a" (in other words, the
115 length check was missing). Both these problems are fixed. "Subpattern name
116 expected" is now given for (?&) (a zero-length name), and this patch also
117 makes it give the same error for \k'' (previously it complained that that
118 was a reference to a non-existent subpattern).
119
120 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
121 this is right because (?- can be followed by option settings as well as by
122 digits. I have, however, made the messages clearer.
123
124 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
125 than the number used in the conditional) now cause a compile-time error.
126 This is actually not compatible with Perl, which accepts such patterns, but
127 treats the conditional as always being FALSE (as PCRE used to), but it
128 seems to me that giving a diagnostic is better.
129
130 8. Change "alphameric" to the more common word "alphanumeric" in comments
131 and messages.
132
133 9. Fix two occurrences of "backslash" in comments that should have been
134 "backspace".
135
136 10. Remove two redundant lines of code that can never be obeyed (their function
137 was moved elsewhere).
138
139 11. The program that makes PCRE's Unicode character property table had a bug
140 which caused it to generate incorrect table entries for sequences of
141 characters that have the same character type, but are in different scripts.
142 It amalgamated them into a single range, with the script of the first of
143 them. In other words, some characters were in the wrong script. There were
144 thirteen such cases, affecting characters in the following ranges:
145
146 U+002b0 - U+002c1
147 U+0060c - U+0060d
148 U+0061e - U+00612
149 U+0064b - U+0065e
150 U+0074d - U+0076d
151 U+01800 - U+01805
152 U+01d00 - U+01d77
153 U+01d9b - U+01dbf
154 U+0200b - U+0200f
155 U+030fc - U+030fe
156 U+03260 - U+0327f
157 U+0fb46 - U+0fbb1
158 U+10450 - U+1049d
159
160 12. The -o option (show only the matching part of a line) for pcregrep was not
161 compatible with GNU grep in that, if there was more than one match in a
162 line, it showed only the first of them. It now behaves in the same way as
163 GNU grep.
164
165 13. If the -o and -v options were combined for pcregrep, it printed a blank
166 line for every non-matching line. GNU grep prints nothing, and pcregrep now
167 does the same. The return code can be used to tell if there were any
168 non-matching lines.
169
170 14. Added --file-offsets and --line-offsets to pcregrep.
171
172 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
173 infinitely looping recursion. The bug was that positive lookaheads were not
174 being skipped when checking for a possible empty match (negative lookaheads
175 and both kinds of lookbehind were skipped).
176
177 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
178 inclusion of <windows.h> to before rather than after the definition of
179 INVALID_FILE_ATTRIBUTES (patch from David Byron).
180
181 17. Specifying a possessive quantifier with a specific limit for a Unicode
182 character property caused pcre_compile() to compile bad code, which led at
183 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
184 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
185 caused the error; without that there was no problem.
186
187 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
188
189 19. Added --enable-pcretest-libreadline.
190
191 20. In pcrecpp.cc, the variable 'count' was incremented twice in
192 RE::GlobalReplace(). As a result, the number of replacements returned was
193 double what it should be. I removed one of the increments, but Craig sent a
194 later patch that removed the other one (the right fix) and added unit tests
195 that check the return values (which was not done before).
196
197 21. Several CMake things:
198
199 (1) Arranged that, when cmake is used on Unix, the libraries end up with
200 the names libpcre and libpcreposix, not just pcre and pcreposix.
201
202 (2) The above change means that pcretest and pcregrep are now correctly
203 linked with the newly-built libraries, not previously installed ones.
204
205 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
206
207 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
208 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
209 UTF-8 newline character). The key issue is that the pattern starts .*;
210 this means that the match must be either at the beginning, or after a
211 newline. The bug was in the code for advancing after a failed match and
212 checking that the new position followed a newline. It was not taking
213 account of UTF-8 characters correctly.
214
215 23. PCRE was behaving differently from Perl in the way it recognized POSIX
216 character classes. PCRE was not treating the sequence [:...:] as a
217 character class unless the ... were all letters. Perl, however, seems to
218 allow any characters between [: and :], though of course it rejects as
219 unknown any "names" that contain non-letters, because all the known class
220 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
221 for example, whereas PCRE did not - it did not recognize a POSIX character
222 class. This seemed a bit dangerous, so the code has been changed to be
223 closer to Perl. The behaviour is not identical to Perl, because PCRE will
224 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
225 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
226 Perl does, and where it didn't before.
227
228 24. Rewrite so as to remove the single use of %n from pcregrep because in some
229 Windows environments %n is disabled by default.
230
231
232 Version 7.4 21-Sep-07
233 ---------------------
234
235 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
236 means that a class such as [\s] counted as "explicit reference to CR or
237 LF". That isn't really right - the whole point of the change was to try to
238 help when there was an actual mention of one of the two characters. So now
239 the change happens only if \r or \n (or a literal CR or LF) character is
240 encountered.
241
242 2. The 32-bit options word was also used for 6 internal flags, but the numbers
243 of both had grown to the point where there were only 3 bits left.
244 Fortunately, there was spare space in the data structure, and so I have
245 moved the internal flags into a new 16-bit field to free up more option
246 bits.
247
248 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
249 but did not set the internal JCHANGED flag - either of these is enough to
250 control the way the "get" function works - but the PCRE_INFO_JCHANGED
251 facility is supposed to tell if (?J) was ever used, so now (?J) at the
252 start sets both bits.
253
254 4. Added options (at build time, compile time, exec time) to change \R from
255 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
256
257 5. doc/pcresyntax.html was missing from the distribution.
258
259 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
260 compatibility, even though it is no longer used.
261
262 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
263 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
264 windows.h file is present (where different names are used). [This was
265 reversed later after testing - see 16 below.]
266
267 8. Changed all #include <config.h> to #include "config.h". There were also
268 some further <pcre.h> cases that I changed to "pcre.h".
269
270 9. When pcregrep was used with the --colour option, it missed the line ending
271 sequence off the lines that it output.
272
273 10. It was pointed out to me that arrays of string pointers cause lots of
274 relocations when a shared library is dynamically loaded. A technique of
275 using a single long string with a table of offsets can drastically reduce
276 these. I have refactored PCRE in four places to do this. The result is
277 dramatic:
278
279 Originally: 290
280 After changing UCP table: 187
281 After changing error message table: 43
282 After changing table of "verbs" 36
283 After changing table of Posix names 22
284
285 Thanks to the folks working on Gregex for glib for this insight.
286
287 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
288 unicode-properties was also set.
289
290 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
291
292 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
293 checked only for CRLF.
294
295 14. Added casts to pcretest.c to avoid compiler warnings.
296
297 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
298
299 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
300 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
301 entirely. This removes changes made in 7 above.
302
303 17. The CMake files have been updated, and there is now more information about
304 building with CMake in the NON-UNIX-USE document.
305
306
307 Version 7.3 28-Aug-07
308 ---------------------
309
310 1. In the rejigging of the build system that eventually resulted in 7.1, the
311 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
312 brackets there is not right, since it causes compilers to look for an
313 installed pcre.h, not the version that is in the source that is being
314 compiled (which of course may be different). I have changed it back to:
315
316 #include "pcre.h"
317
318 I have a vague recollection that the change was concerned with compiling in
319 different directories, but in the new build system, that is taken care of
320 by the VPATH setting the Makefile.
321
322 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
323 when the subject happened to end in the byte 0x85 (e.g. if the last
324 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
325 characters but of course it shouldn't be taken as a newline when it is part
326 of another character. The bug was that, for an unlimited repeat of . in
327 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
328 characters when looking for a newline.
329
330 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
331
332 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
333 in debug output.
334
335 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
336 long printing in the pcrecpp unittest when running under MinGW.
337
338 6. ESC_K was left out of the EBCDIC table.
339
340 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
341 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
342 limit also applies to "virtual nesting" when a pattern is recursive, and in
343 this case 1000 isn't so big. I have been able to remove this limit at the
344 expense of backing off one optimization in certain circumstances. Normally,
345 when pcre_exec() would call its internal match() function recursively and
346 immediately return the result unconditionally, it uses a "tail recursion"
347 feature to save stack. However, when a subpattern that can match an empty
348 string has an unlimited repetition quantifier, it no longer makes this
349 optimization. That gives it a stack frame in which to save the data for
350 checking that an empty string has been matched. Previously this was taken
351 from the 1000-entry workspace that had been reserved. So now there is no
352 explicit limit, but more stack is used.
353
354 8. Applied Daniel's patches to solve problems with the import/export magic
355 syntax that is required for Windows, and which was going wrong for the
356 pcreposix and pcrecpp parts of the library. These were overlooked when this
357 problem was solved for the main library.
358
359 9. There were some crude static tests to avoid integer overflow when computing
360 the size of patterns that contain repeated groups with explicit upper
361 limits. As the maximum quantifier is 65535, the maximum group length was
362 set at 30,000 so that the product of these two numbers did not overflow a
363 32-bit integer. However, it turns out that people want to use groups that
364 are longer than 30,000 bytes (though not repeat them that many times).
365 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
366 made it possible to implement the integer overflow checks in a much more
367 dynamic way, which I have now done. The artificial limitation on group
368 length has been removed - we now have only the limit on the total length of
369 the compiled pattern, which depends on the LINK_SIZE setting.
370
371 10. Fixed a bug in the documentation for get/copy named substring when
372 duplicate names are permitted. If none of the named substrings are set, the
373 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
374 empty string.
375
376 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
377 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
378 because the ] is interpreted as the first data character and the
379 terminating ] is not found. PCRE has been made compatible with Perl in this
380 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
381 cause memory overwriting.
382
383 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
384 string has been matched (to stop an infinite loop). It was not recognizing
385 a conditional subpattern that could match an empty string if that
386 subpattern was within another subpattern. For example, it looped when
387 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
388 condition was not nested. This bug has been fixed.
389
390 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
391 past the start of the subject in the presence of bytes with the top bit
392 set, for example "\x8aBCD".
393
394 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
395 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
396
397 14. Optimized (?!) to (*FAIL).
398
399 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
400 This restricts code points to be within the range 0 to 0x10FFFF, excluding
401 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
402 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
403 does: it's just the validity check that is more restrictive.
404
405 16. Inserted checks for integer overflows during escape sequence (backslash)
406 processing, and also fixed erroneous offset values for syntax errors during
407 backslash processing.
408
409 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
410 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
411
412 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
413 caused an overrun.
414
415 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
416 something other than just ASCII characters) inside a group that had an
417 unlimited repeat caused a loop at compile time (while checking to see
418 whether the group could match an empty string).
419
420 20. Debugging a pattern containing \p or \P could cause a crash. For example,
421 [\P{Any}] did so. (Error in the code for printing property names.)
422
423 21. An orphan \E inside a character class could cause a crash.
424
425 22. A repeated capturing bracket such as (A)? could cause a wild memory
426 reference during compilation.
427
428 23. There are several functions in pcre_compile() that scan along a compiled
429 expression for various reasons (e.g. to see if it's fixed length for look
430 behind). There were bugs in these functions when a repeated \p or \P was
431 present in the pattern. These operators have additional parameters compared
432 with \d, etc, and these were not being taken into account when moving along
433 the compiled data. Specifically:
434
435 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
436 length.
437
438 (b) An item such as \pL+ within a repeated group could cause crashes or
439 loops.
440
441 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
442 "reference to non-existent subpattern" error.
443
444 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
445
446 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
447 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
448
449 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
450
451 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
452 character were causing crashes (broken optimization).
453
454 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
455 \p or \P) caused a compile-time loop.
456
457 28. More problems have arisen in unanchored patterns when CRLF is a valid line
458 break. For example, the unstudied pattern [\r\n]A does not match the string
459 "\r\nA" because change 7.0/46 below moves the current point on by two
460 characters after failing to match at the start. However, the pattern \nA
461 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
462 the same is true. There doesn't seem any very clean way out of this, but
463 what I have chosen to do makes the common cases work: PCRE now takes note
464 of whether there can be an explicit match for \r or \n anywhere in the
465 pattern, and if so, 7.0/46 no longer applies. As part of this change,
466 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
467 pattern has explicit CR or LF references.
468
469 29. Added (*CR) etc for changing newline setting at start of pattern.
470
471
472 Version 7.2 19-Jun-07
473 ---------------------
474
475 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
476 which is apparently normally available under Windows.
477
478 2. Re-jig the pcregrep tests with different newline settings in an attempt
479 to make them independent of the local environment's newline setting.
480
481 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
482
483 4. Some of the "internals" tests were previously cut out when the link size
484 was not 2, because the output contained actual offsets. The recent new
485 "Z" feature of pcretest means that these can be cut out, making the tests
486 usable with all link sizes.
487
488 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
489 stack recursion. This gives a massive performance boost under BSD, but just
490 a small improvement under Linux. However, it saves one field in the frame
491 in all cases.
492
493 6. Added more features from the forthcoming Perl 5.10:
494
495 (a) (?-n) (where n is a string of digits) is a relative subroutine or
496 recursion call. It refers to the nth most recently opened parentheses.
497
498 (b) (?+n) is also a relative subroutine call; it refers to the nth next
499 to be opened parentheses.
500
501 (c) Conditions that refer to capturing parentheses can be specified
502 relatively, for example, (?(-2)... or (?(+3)...
503
504 (d) \K resets the start of the current match so that everything before
505 is not part of it.
506
507 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
508
509 (f) \g{name} is another synonym - part of Perl 5.10's unification of
510 reference syntax.
511
512 (g) (?| introduces a group in which the numbering of parentheses in each
513 alternative starts with the same number.
514
515 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
516
517 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
518 PCRE_INFO_JCHANGED.
519
520 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
521 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
522 for detecting groups that can match an empty string.
523
524 9. A pattern with a very large number of alternatives (more than several
525 hundred) was running out of internal workspace during the pre-compile
526 phase, where pcre_compile() figures out how much memory will be needed. A
527 bit of new cunning has reduced the workspace needed for groups with
528 alternatives. The 1000-alternative test pattern now uses 12 bytes of
529 workspace instead of running out of the 4096 that are available.
530
531 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
532
533 11. Applied patch from Google to remove an optimization that didn't quite work.
534 The report of the bug said:
535
536 pcrecpp::RE("a*").FullMatch("aaa") matches, while
537 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
538 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
539
540 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
541 it matched the wrong number of bytes.
542
543
544 Version 7.1 24-Apr-07
545 ---------------------
546
547 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
548 that is more "standard", making use of automake and other Autotools. There
549 is some re-arrangement of the files and adjustment of comments consequent
550 on this.
551
552 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
553 for recursive directory scanning broke on some systems because the files
554 are not scanned in any specific order and on different systems the order
555 was different. A call to "sort" has been inserted into RunGrepTest for the
556 approprate test as a short-term fix. In the longer term there may be an
557 alternative.
558
559 3. I had an email from Eric Raymond about problems translating some of PCRE's
560 man pages to HTML (despite the fact that I distribute HTML pages, some
561 people do their own conversions for various reasons). The problems
562 concerned the use of low-level troff macros .br and .in. I have therefore
563 removed all such uses from the man pages (some were redundant, some could
564 be replaced by .nf/.fi pairs). The 132html script that I use to generate
565 HTML has been updated to handle .nf/.fi and to complain if it encounters
566 .br or .in.
567
568 4. Updated comments in configure.ac that get placed in config.h.in and also
569 arranged for config.h to be included in the distribution, with the name
570 config.h.generic, for the benefit of those who have to compile without
571 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
572
573 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
574 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
575 makevp.bat for the current PCRE, using the additional files
576 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
577
578 6. A Windows user reported a minor discrepancy with test 2, which turned out
579 to be caused by a trailing space on an input line that had got lost in his
580 copy. The trailing space was an accident, so I've just removed it.
581
582 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
583 that is needed.
584
585 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
586 as "const" (a) because they are and (b) because it helps the PHP
587 maintainers who have recently made a script to detect big data structures
588 in the php code that should be moved to the .rodata section. I remembered
589 to update Builducptable as well, so it won't revert if ucptable.h is ever
590 re-created.
591
592 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
593 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
594 order to be able to cut out the UTF-8 tables in the latter when UTF-8
595 support is not required. This saves 1.5-2K of code, which is important in
596 some applications.
597
598 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
599 so as not to refer to the tables, even though these functions will never be
600 called when UTF-8 support is disabled. Otherwise there are problems with a
601 shared library.
602
603 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
604
605 (a) It was defining its arguments as char * instead of void *.
606
607 (b) It was assuming that all moves were upwards in memory; this was true
608 a long time ago when I wrote it, but is no longer the case.
609
610 The emulated memove() is provided for those environments that have neither
611 memmove() nor bcopy(). I didn't think anyone used it these days, but that
612 is clearly not the case, as these two bugs were recently reported.
613
614 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
615 and Detrail to create the HTML documentation, the .txt form of the man
616 pages, and it removes trailing spaces from listed files. It also creates
617 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
618 case, it wraps all the #defines with #ifndefs. This script should be run
619 before "make dist".
620
621 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
622 with Unicode property support.
623
624 (a) For a maximizing quantifier, if the two different cases of the
625 character were of different lengths in their UTF-8 codings (there are
626 some cases like this - I found 11), and the matching function had to
627 back up over a mixture of the two cases, it incorrectly assumed they
628 were both the same length.
629
630 (b) When PCRE was configured to use the heap rather than the stack for
631 recursion during matching, it was not correctly preserving the data for
632 the other case of a UTF-8 character when checking ahead for a match
633 while processing a minimizing repeat. If the check also involved
634 matching a wide character, but failed, corruption could cause an
635 erroneous result when trying to check for a repeat of the original
636 character.
637
638 13. Some tidying changes to the testing mechanism:
639
640 (a) The RunTest script now detects the internal link size and whether there
641 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
642 values substituted by "configure". (The RunGrepTest script already did
643 this for UTF-8.) The configure.ac script no longer substitutes the
644 relevant variables.
645
646 (b) The debugging options /B and /D in pcretest show the compiled bytecode
647 with length and offset values. This means that the output is different
648 for different internal link sizes. Test 2 is skipped for link sizes
649 other than 2 because of this, bypassing the problem. Unfortunately,
650 there was also a test in test 3 (the locale tests) that used /B and
651 failed for link sizes other than 2. Rather than cut the whole test out,
652 I have added a new /Z option to pcretest that replaces the length and
653 offset values with spaces. This is now used to make test 3 independent
654 of link size. (Test 2 will be tidied up later.)
655
656 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
657 segmentation fault instead of returning the appropriate error message.
658
659 15. In multiline mode when the newline sequence was set to "any", the pattern
660 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
661 This doesn't seem right; it now treats the CRLF combination as the line
662 ending, and so does not match in that case. It's only a pattern such as ^$
663 that would hit this one: something like ^ABC$ would have failed after \r
664 and then tried again after \r\n.
665
666 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
667 in an attempt to make files that differ only in their line terminators
668 compare equal. This works on Linux.
669
670 17. Under certain error circumstances pcregrep might try to free random memory
671 as it exited. This is now fixed, thanks to valgrind.
672
673 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
674 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
675 was because its rules for how to advance for /g after matching an empty
676 string at the end of a line did not allow for this case. They now check for
677 it specially.
678
679 20. pcretest is supposed to handle patterns and data of any length, by
680 extending its buffers when necessary. It was getting this wrong when the
681 buffer for a data line had to be extended.
682
683 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
684 CRLF as a newline sequence.
685
686 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
687 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
688 I have nevertheless tidied it up.
689
690 23. Added some casts to kill warnings from HP-UX ia64 compiler.
691
692 24. Added a man page for pcre-config.
693
694
695 Version 7.0 19-Dec-06
696 ---------------------
697
698 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
699 moving to gcc 4.1.1.
700
701 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
702 sys/time.h, which is documented as needed for this function. It doesn't
703 seem to matter on Linux, but it showed up on some releases of OS X.
704
705 3. It seems that there are systems where bytes whose values are greater than
706 127 match isprint() in the "C" locale. The "C" locale should be the
707 default when a C program starts up. In most systems, only ASCII printing
708 characters match isprint(). This difference caused the output from pcretest
709 to vary, making some of the tests fail. I have changed pcretest so that:
710
711 (a) When it is outputting text in the compiled version of a pattern, bytes
712 other than 32-126 are always shown as hex escapes.
713
714 (b) When it is outputting text that is a matched part of a subject string,
715 it does the same, unless a different locale has been set for the match
716 (using the /L modifier). In this case, it uses isprint() to decide.
717
718 4. Fixed a major bug that caused incorrect computation of the amount of memory
719 required for a compiled pattern when options that changed within the
720 pattern affected the logic of the preliminary scan that determines the
721 length. The relevant options are -x, and -i in UTF-8 mode. The result was
722 that the computed length was too small. The symptoms of this bug were
723 either the PCRE error "internal error: code overflow" from pcre_compile(),
724 or a glibc crash with a message such as "pcretest: free(): invalid next
725 size (fast)". Examples of patterns that provoked this bug (shown in
726 pcretest format) are:
727
728 /(?-x: )/x
729 /(?x)(?-x: \s*#\s*)/
730 /((?i)[\x{c0}])/8
731 /(?i:[\x{c0}])/8
732
733 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
734 is now done differently.
735
736 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
737 wrapper classes; (b) implement a new function in the C++ scanner that is
738 more efficient than the old way of doing things because it avoids levels of
739 recursion in the regex matching; (c) add a paragraph to the documentation
740 for the FullMatch() function.
741
742 6. The escape sequence \n was being treated as whatever was defined as
743 "newline". Not only was this contrary to the documentation, which states
744 that \n is character 10 (hex 0A), but it also went horribly wrong when
745 "newline" was defined as CRLF. This has been fixed.
746
747 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
748 was being set to -1 for the "end of line" case (supposedly a value that no
749 character can have). Though this value is never used (the check for end of
750 line is "zero bytes in current character"), it caused compiler complaints.
751 I've changed it to 0xffffffff.
752
753 8. In pcre_version.c, the version string was being built by a sequence of
754 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
755 string (as it is for production releases) called a macro with an empty
756 argument. The C standard says the result of this is undefined. The gcc
757 compiler treats it as an empty string (which was what was wanted) but it is
758 reported that Visual C gives an error. The source has been hacked around to
759 avoid this problem.
760
761 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
762 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
763 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
764 of them did).
765
766 10. Originally, pcretest opened its input and output without "b"; then I was
767 told that "b" was needed in some environments, so it was added for release
768 5.0 to both the input and output. (It makes no difference on Unix-like
769 systems.) Later I was told that it is wrong for the input on Windows. I've
770 now abstracted the modes into two macros, to make it easier to fiddle with
771 them, and removed "b" from the input mode under Windows.
772
773 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
774
775 12. Added -help and --help to pcretest as an official way of being reminded
776 of the options.
777
778 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
779 and pcrecpp.cc because they annoy compilers at high warning levels.
780
781 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
782
783 15. Fixed an occurrence of == in configure.ac that should have been = (shell
784 scripts are not C programs :-) and which was not noticed because it works
785 on Linux.
786
787 16. pcretest is supposed to handle any length of pattern and data line (as one
788 line or as a continued sequence of lines) by extending its input buffer if
789 necessary. This feature was broken for very long pattern lines, leading to
790 a string of junk being passed to pcre_compile() if the pattern was longer
791 than about 50K.
792
793 17. I have done a major re-factoring of the way pcre_compile() computes the
794 amount of memory needed for a compiled pattern. Previously, there was code
795 that made a preliminary scan of the pattern in order to do this. That was
796 OK when PCRE was new, but as the facilities have expanded, it has become
797 harder and harder to keep it in step with the real compile phase, and there
798 have been a number of bugs (see for example, 4 above). I have now found a
799 cunning way of running the real compile function in a "fake" mode that
800 enables it to compute how much memory it would need, while actually only
801 ever using a few hundred bytes of working memory and without too many
802 tests of the mode. This should make future maintenance and development
803 easier. A side effect of this work is that the limit of 200 on the nesting
804 depth of parentheses has been removed (though this was never a serious
805 limitation, I suspect). However, there is a downside: pcre_compile() now
806 runs more slowly than before (30% or more, depending on the pattern). I
807 hope this isn't a big issue. There is no effect on runtime performance.
808
809 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
810 newline (only possible for the last line of a file) and it was a
811 pattern that set a locale (followed by /Lsomething), pcretest crashed.
812
813 19. Added additional timing features to pcretest. (1) The -tm option now times
814 matching only, not compiling. (2) Both -t and -tm can be followed, as a
815 separate command line item, by a number that specifies the number of
816 repeats to use when timing. The default is 50000; this gives better
817 precision, but takes uncomfortably long for very large patterns.
818
819 20. Extended pcre_study() to be more clever in cases where a branch of a
820 subpattern has no definite first character. For example, (a*|b*)[cd] would
821 previously give no result from pcre_study(). Now it recognizes that the
822 first character must be a, b, c, or d.
823
824 21. There was an incorrect error "recursive call could loop indefinitely" if
825 a subpattern (or the entire pattern) that was being tested for matching an
826 empty string contained only one non-empty item after a nested subpattern.
827 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
828 incorrectly, because the \d was being skipped in the check.
829
830 22. The pcretest program now has a new pattern option /B and a command line
831 option -b, which is equivalent to adding /B to every pattern. This causes
832 it to show the compiled bytecode, without the additional information that
833 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
834 is the same as /B/I).
835
836 23. A new optimization is now able automatically to treat some sequences such
837 as a*b as a*+b. More specifically, if something simple (such as a character
838 or a simple class like \d) has an unlimited quantifier, and is followed by
839 something that cannot possibly match the quantified thing, the quantifier
840 is automatically "possessified".
841
842 24. A recursive reference to a subpattern whose number was greater than 39
843 went wrong under certain circumstances in UTF-8 mode. This bug could also
844 have affected the operation of pcre_study().
845
846 25. Realized that a little bit of performance could be had by replacing
847 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
848
849 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
850
851 27. Possessive quantifiers such as a++ were previously implemented by turning
852 them into atomic groups such as ($>a+). Now they have their own opcodes,
853 which improves performance. This includes the automatically created ones
854 from 23 above.
855
856 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
857 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
858 the first matched character to be a colon. This applied both to named and
859 numbered groups.
860
861 29. The ucpinternal.h header file was missing its idempotency #ifdef.
862
863 30. I was sent a "project" file called libpcre.a.dev which I understand makes
864 building PCRE on Windows easier, so I have included it in the distribution.
865
866 31. There is now a check in pcretest against a ridiculously large number being
867 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
868 loop, the loop is abandoned.
869
870 32. Forward references to subpatterns in conditions such as (?(2)...) where
871 subpattern 2 is defined later cause pcre_compile() to search forwards in
872 the pattern for the relevant set of parentheses. This search went wrong
873 when there were unescaped parentheses in a character class, parentheses
874 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
875
876 33. "Subroutine" calls and backreferences were previously restricted to
877 referencing subpatterns earlier in the regex. This restriction has now
878 been removed.
879
880 34. Added a number of extra features that are going to be in Perl 5.10. On the
881 whole, these are just syntactic alternatives for features that PCRE had
882 previously implemented using the Python syntax or my own invention. The
883 other formats are all retained for compatibility.
884
885 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
886 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
887 also .NET compatible.
888
889 (b) A recursion or subroutine call to a named group can now be defined as
890 (?&name) as well as (?P>name).
891
892 (c) A backreference to a named group can now be defined as \k<name> or
893 \k'name' as well as (?P=name). The new forms, as well as being in Perl
894 5.10, are also .NET compatible.
895
896 (d) A conditional reference to a named group can now use the syntax
897 (?(<name>) or (?('name') as well as (?(name).
898
899 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
900 groups (named and numbered) that are never evaluated inline, but can be
901 called as "subroutines" from elsewhere. In effect, the DEFINE condition
902 is always false. There may be only one alternative in such a group.
903
904 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
905 as the simple (?(R). The condition is true only if the most recent
906 recursion is that of the given number or name. It does not search out
907 through the entire recursion stack.
908
909 (g) The escape \gN or \g{N} has been added, where N is a positive or
910 negative number, specifying an absolute or relative reference.
911
912 35. Tidied to get rid of some further signed/unsigned compiler warnings and
913 some "unreachable code" warnings.
914
915 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
916 things, this adds five new scripts.
917
918 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
919 There were also incompatibilities regarding the handling of \Q..\E inside
920 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
921 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
922
923 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
924 matches an empty string, and forcibly breaks the loop. There were bugs in
925 this code in non-simple cases. For a pattern such as ^(a()*)* matched
926 against aaaa the result was just "a" rather than "aaaa", for example. Two
927 separate and independent bugs (that affected different cases) have been
928 fixed.
929
930 39. Refactored the code to abolish the use of different opcodes for small
931 capturing bracket numbers. This is a tidy that I avoided doing when I
932 removed the limit on the number of capturing brackets for 3.5 back in 2001.
933 The new approach is not only tidier, it makes it possible to reduce the
934 memory needed to fix the previous bug (38).
935
936 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
937 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
938 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
939 mode.
940
941 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
942 report.
943
944 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
945 copy construction and assignment in the C++ wrapper.
946
947 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
948 couple of bugs that could have given wrong results in the "--newline=crlf"
949 case.
950
951 44. Added a number of casts and did some reorganization of signed/unsigned int
952 variables following suggestions from Dair Grant. Also renamed the variable
953 "this" as "item" because it is a C++ keyword.
954
955 45. Arranged for dftables to add
956
957 #include "pcre_internal.h"
958
959 to pcre_chartables.c because without it, gcc 4.x may remove the array
960 definition from the final binary if PCRE is built into a static library and
961 dead code stripping is activated.
962
963 46. For an unanchored pattern, if a match attempt fails at the start of a
964 newline sequence, and the newline setting is CRLF or ANY, and the next two
965 characters are CRLF, advance by two characters instead of one.
966
967
968 Version 6.7 04-Jul-06
969 ---------------------
970
971 1. In order to handle tests when input lines are enormously long, pcretest has
972 been re-factored so that it automatically extends its buffers when
973 necessary. The code is crude, but this _is_ just a test program. The
974 default size has been increased from 32K to 50K.
975
976 2. The code in pcre_study() was using the value of the re argument before
977 testing it for NULL. (Of course, in any sensible call of the function, it
978 won't be NULL.)
979
980 3. The memmove() emulation function in pcre_internal.h, which is used on
981 systems that lack both memmove() and bcopy() - that is, hardly ever -
982 was missing a "static" storage class specifier.
983
984 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
985 containing an extended class (one that cannot be represented by a bitmap
986 because it contains high-valued characters or Unicode property items, e.g.
987 [\pZ]). Almost always one would set UTF-8 mode when processing such a
988 pattern, but PCRE should not loop if you do not (it no longer does).
989 [Detail: two cases were found: (a) a repeated subpattern containing an
990 extended class; (b) a recursive reference to a subpattern that followed a
991 previous extended class. It wasn't skipping over the extended class
992 correctly when UTF-8 mode was not set.]
993
994 5. A negated single-character class was not being recognized as fixed-length
995 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
996 compile error "lookbehind assertion is not fixed length".
997
998 6. The RunPerlTest auxiliary script was showing an unexpected difference
999 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1000 write a Perl script that can interpret lines of an input file either as
1001 byte characters or as UTF-8, which is what "perltest" was being required to
1002 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1003 can't do is switch easily at run time between having the "use utf8;" pragma
1004 or not. In the end, I fudged it by using the RunPerlTest script to insert
1005 "use utf8;" explicitly for the UTF-8 tests.
1006
1007 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1008 the end of the subject string, contrary to the documentation and to what
1009 Perl does. This was true of both matching functions. Now it matches only at
1010 the start of the subject and immediately after *internal* newlines.
1011
1012 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1013 a pointer to an int instead of a pointer to an unsigned long int. This
1014 caused problems on 64-bit systems.
1015
1016 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1017 instance of the 'standard' template library not being so standard".
1018
1019 10. There was no check on the number of named subpatterns nor the maximum
1020 length of a subpattern name. The product of these values is used to compute
1021 the size of the memory block for a compiled pattern. By supplying a very
1022 long subpattern name and a large number of named subpatterns, the size
1023 computation could be caused to overflow. This is now prevented by limiting
1024 the length of names to 32 characters, and the number of named subpatterns
1025 to 10,000.
1026
1027 11. Subpatterns that are repeated with specific counts have to be replicated in
1028 the compiled pattern. The size of memory for this was computed from the
1029 length of the subpattern and the repeat count. The latter is limited to
1030 65535, but there was no limit on the former, meaning that integer overflow
1031 could in principle occur. The compiled length of a repeated subpattern is
1032 now limited to 30,000 bytes in order to prevent this.
1033
1034 12. Added the optional facility to have named substrings with the same name.
1035
1036 13. Added the ability to use a named substring as a condition, using the
1037 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1038 are numbers (not recommended). Forward references are permitted.
1039
1040 14. Added forward references in named backreferences (if you see what I mean).
1041
1042 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1043 pattern could run off the end of the subject. For example, the pattern
1044 "(?s)(.{1,5})"8 did this with the subject "ab".
1045
1046 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1047 PCRE_CASELESS was set when matching characters that were quantified with ?
1048 or *.
1049
1050 17. A character class other than a single negated character that had a minimum
1051 but no maximum quantifier - for example [ab]{6,} - was not handled
1052 correctly by pce_dfa_exec(). It would match only one character.
1053
1054 18. A valid (though odd) pattern that looked like a POSIX character
1055 class but used an invalid character after [ (for example [[,abc,]]) caused
1056 pcre_compile() to give the error "Failed: internal error: code overflow" or
1057 in some cases to crash with a glibc free() error. This could even happen if
1058 the pattern terminated after [[ but there just happened to be a sequence of
1059 letters, a binary zero, and a closing ] in the memory that followed.
1060
1061 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1062 over the years. Originally (before any Unicode support), just the bottom 8
1063 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1064 output from "man perlunicode" includes this:
1065
1066 The regular expression compiler produces polymorphic opcodes. That
1067 is, the pattern adapts to the data and automatically switches to
1068 the Unicode character scheme when presented with Unicode data--or
1069 instead uses a traditional byte scheme when presented with byte
1070 data.
1071
1072 Sadly, a wide octal escape does not cause a switch, and in a string with
1073 no other multibyte characters, these octal escapes are treated as before.
1074 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1075 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1076 Unicode string.
1077
1078 I have not perpetrated such confusion in PCRE. Up till now, it took just
1079 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1080 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1081 translate to the appropriate multibyte character.
1082
1083 29. Applied some refactoring to reduce the number of warnings from Microsoft
1084 and Borland compilers. This has included removing the fudge introduced
1085 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1086 a warning about an unused variable.
1087
1088 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1089 characters since release 4.0, because Perl (from release 5.004) does not.
1090 [Or at least, is documented not to: some releases seem to be in conflict
1091 with the documentation.] However, when a pattern was studied with
1092 pcre_study() and all its branches started with \s, PCRE still included VT
1093 as a possible starting character. Of course, this did no harm; it just
1094 caused an unnecessary match attempt.
1095
1096 22. Removed a now-redundant internal flag bit that recorded the fact that case
1097 dependency changed within the pattern. This was once needed for "required
1098 byte" processing, but is no longer used. This recovers a now-scarce options
1099 bit. Also moved the least significant internal flag bit to the most-
1100 significant bit of the word, which was not previously used (hangover from
1101 the days when it was an int rather than a uint) to free up another bit for
1102 the future.
1103
1104 23. Added support for CRLF line endings as well as CR and LF. As well as the
1105 default being selectable at build time, it can now be changed at runtime
1106 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1107 specify that it is scanning data with non-default line endings.
1108
1109 24. Changed the definition of CXXLINK to make it agree with the definition of
1110 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1111
1112 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1113 recursions. This makes a big different to stack usage for some patterns.
1114
1115 26. If a subpattern containing a named recursion or subroutine reference such
1116 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1117 the space required for the compiled pattern went wrong and gave too small a
1118 value. Depending on the environment, this could lead to "Failed: internal
1119 error: code overflow at offset 49" or "glibc detected double free or
1120 corruption" errors.
1121
1122 27. Applied patches from Google (a) to support the new newline modes and (b) to
1123 advance over multibyte UTF-8 characters in GlobalReplace.
1124
1125 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1126 difference for some implementation of PCRE in some Windows version.
1127
1128 29. Added some extra testing facilities to pcretest:
1129
1130 \q<number> in a data line sets the "match limit" value
1131 \Q<number> in a data line sets the "match recursion limt" value
1132 -S <number> sets the stack size, where <number> is in megabytes
1133
1134 The -S option isn't available for Windows.
1135
1136
1137 Version 6.6 06-Feb-06
1138 ---------------------
1139
1140 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1141 in pcreposix.h. I have copied the definition from pcre.h.
1142
1143 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1144 because pcre.h is no longer a built file.
1145
1146 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1147 not normally included in the compiled code.
1148
1149
1150 Version 6.5 01-Feb-06
1151 ---------------------
1152
1153 1. When using the partial match feature with pcre_dfa_exec(), it was not
1154 anchoring the second and subsequent partial matches at the new starting
1155 point. This could lead to incorrect results. For example, with the pattern
1156 /1234/, partially matching against "123" and then "a4" gave a match.
1157
1158 2. Changes to pcregrep:
1159
1160 (a) All non-match returns from pcre_exec() were being treated as failures
1161 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1162 error message is output. Some extra information is given for the
1163 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1164 probably the only errors that are likely to be caused by users (by
1165 specifying a regex that has nested indefinite repeats, for instance).
1166 If there are more than 20 of these errors, pcregrep is abandoned.
1167
1168 (b) A binary zero was treated as data while matching, but terminated the
1169 output line if it was written out. This has been fixed: binary zeroes
1170 are now no different to any other data bytes.
1171
1172 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1173 used to set a locale for matching. The --locale=xxxx long option has
1174 been added (no short equivalent) to specify a locale explicitly on the
1175 pcregrep command, overriding the environment variables.
1176
1177 (d) When -B was used with -n, some line numbers in the output were one less
1178 than they should have been.
1179
1180 (e) Added the -o (--only-matching) option.
1181
1182 (f) If -A or -C was used with -c (count only), some lines of context were
1183 accidentally printed for the final match.
1184
1185 (g) Added the -H (--with-filename) option.
1186
1187 (h) The combination of options -rh failed to suppress file names for files
1188 that were found from directory arguments.
1189
1190 (i) Added the -D (--devices) and -d (--directories) options.
1191
1192 (j) Added the -F (--fixed-strings) option.
1193
1194 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1195
1196 (l) Added the --colo(u)r option.
1197
1198 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1199 is not present by default.
1200
1201 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1202 items such as (?R) or (?1), when the recursion could match a number of
1203 alternatives. If it matched one of the alternatives, but subsequently,
1204 outside the recursion, there was a failure, the code tried to back up into
1205 the recursion. However, because of the way PCRE is implemented, this is not
1206 possible, and the result was an incorrect result from the match.
1207
1208 In order to prevent this happening, the specification of recursion has
1209 been changed so that all such subpatterns are automatically treated as
1210 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1211
1212 4. I had overlooked the fact that, in some locales, there are characters for
1213 which isalpha() is true but neither isupper() nor islower() are true. In
1214 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1215 and ordfeminine) are like this. This affected the treatment of \w and \W
1216 when they appeared in character classes, but not when they appeared outside
1217 a character class. The bit map for "word" characters is now created
1218 separately from the results of isalnum() instead of just taking it from the
1219 upper, lower, and digit maps. (Plus the underscore character, of course.)
1220
1221 5. The above bug also affected the handling of POSIX character classes such as
1222 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1223 permanent tables. Instead, the bit maps for such a class were previously
1224 created as the appropriate unions of the upper, lower, and digit bitmaps.
1225 Now they are created by subtraction from the [[:word:]] class, which has
1226 its own bitmap.
1227
1228 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1229 It is created by subtracting the vertical space characters (\x09, \x0a,
1230 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1231 subtraction was done in the overall bitmap for a character class, meaning
1232 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1233 be recognized. This bug has been fixed.
1234
1235 7. Patches from the folks at Google:
1236
1237 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1238 real life, but is still worth protecting against".
1239
1240 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1241 regular expressions".
1242
1243 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1244 have it.
1245
1246 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1247 "configure" and the latter not, in order to fix a problem somebody had
1248 with compiling the Arg class on HP-UX.
1249
1250 (e) Improve the error-handling of the C++ wrapper a little bit.
1251
1252 (f) New tests for checking recursion limiting.
1253
1254 8. The pcre_memmove() function, which is used only if the environment does not
1255 have a standard memmove() function (and is therefore rarely compiled),
1256 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1257 returning a result (though PCRE never actually uses the result).
1258
1259 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1260 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1261 returned instead of calling malloc() with an overflowing number that would
1262 most likely cause subsequent chaos.
1263
1264 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1265
1266 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1267 with this option is matched, the nmatch and pmatch options of regexec() are
1268 ignored.
1269
1270 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1271 provided in case anyone wants to the the POSIX interface with UTF-8
1272 strings.
1273
1274 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1275 C++ linking (needed for some HP-UX environments).
1276
1277 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1278 (unused parameter) and in the pcre_printint() function (omitted "default"
1279 switch label when the default is to do nothing).
1280
1281 15. Added some code to make it possible, when PCRE is compiled as a C++
1282 library, to replace subject pointers for pcre_exec() with a smart pointer
1283 class, thus making it possible to process discontinuous strings.
1284
1285 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1286 much the same function. They were added by different people who were trying
1287 to make PCRE easy to compile on non-Unix systems. It has been suggested
1288 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1289 for compiling on Windows systems. I have therefore replaced it with
1290 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1291 defaults to "extern" for C or "extern C" for C++, which works fine on
1292 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1293 SCOPE with something explicit in config.h. In addition:
1294
1295 (a) pcreposix.h still had just "extern" instead of either of these macros;
1296 I have replaced it with PCRE_DATA_SCOPE.
1297
1298 (b) Functions such as _pcre_xclass(), which are internal to the library,
1299 but external in the C sense, all had PCRE_EXPORT in their definitions.
1300 This is apparently wrong for the Windows case, so I have removed it.
1301 (It makes no difference on Unix-like systems.)
1302
1303 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1304 of recursive calls to match(). This is different to MATCH_LIMIT because
1305 that limits the total number of calls to match(), not all of which increase
1306 the depth of recursion. Limiting the recursion depth limits the amount of
1307 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1308 when PCRE is compiled, and changed at run time. A patch from Google adds
1309 this functionality to the C++ interface.
1310
1311 18. Changes to the handling of Unicode character properties:
1312
1313 (a) Updated the table to Unicode 4.1.0.
1314
1315 (b) Recognize characters that are not in the table as "Cn" (undefined).
1316
1317 (c) I revised the way the table is implemented to a much improved format
1318 which includes recognition of ranges. It now supports the ranges that
1319 are defined in UnicodeData.txt, and it also amalgamates other
1320 characters into ranges. This has reduced the number of entries in the
1321 table from around 16,000 to around 3,000, thus reducing its size
1322 considerably. I realized I did not need to use a tree structure after
1323 all - a binary chop search is just as efficient. Having reduced the
1324 number of entries, I extended their size from 6 bytes to 8 bytes to
1325 allow for more data.
1326
1327 (d) Added support for Unicode script names via properties such as \p{Han}.
1328
1329 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1330 matching that character.
1331
1332 20. When matching a repeated Unicode property with a minimum greater than zero,
1333 (for example \pL{2,}), PCRE could look past the end of the subject if it
1334 reached it while seeking the minimum number of characters. This could
1335 happen only if some of the characters were more than one byte long, because
1336 there is a check for at least the minimum number of bytes.
1337
1338 21. Refactored the implementation of \p and \P so as to be more general, to
1339 allow for more different types of property in future. This has changed the
1340 compiled form incompatibly. Anybody with saved compiled patterns that use
1341 \p or \P will have to recompile them.
1342
1343 22. Added "Any" and "L&" to the supported property types.
1344
1345 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1346 but give a compile time error if the value is greater than 0xff.
1347
1348 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1349 accidentally not being installed or uninstalled.
1350
1351 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1352 made were to insert the current release number. This seemed silly, because
1353 it made things harder for people building PCRE on systems that don't run
1354 "configure". I have turned pcre.h into a distributed file, no longer built
1355 by "configure", with the version identification directly included. There is
1356 no longer a pcre.h.in file.
1357
1358 However, this change necessitated a change to the pcre-config script as
1359 well. It is built from pcre-config.in, and one of the substitutions was the
1360 release number. I have updated configure.ac so that ./configure now finds
1361 the release number by grepping pcre.h.
1362
1363 26. Added the ability to run the tests under valgrind.
1364
1365
1366 Version 6.4 05-Sep-05
1367 ---------------------
1368
1369 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1370 "--" to be printed when multiple files were scanned, even when none of the
1371 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1372 consider it to be a bug, and have restored the previous behaviour.
1373
1374 2. A couple of code tidies to get rid of compiler warnings.
1375
1376 3. The pcretest program used to cheat by referring to symbols in the library
1377 whose names begin with _pcre_. These are internal symbols that are not
1378 really supposed to be visible externally, and in some environments it is
1379 possible to suppress them. The cheating is now confined to including
1380 certain files from the library's source, which is a bit cleaner.
1381
1382 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1383 file's purpose clearer.
1384
1385 5. Reorganized pcre_ucp_findchar().
1386
1387
1388 Version 6.3 15-Aug-05
1389 ---------------------
1390
1391 1. The file libpcre.pc.in did not have general read permission in the tarball.
1392
1393 2. There were some problems when building without C++ support:
1394
1395 (a) If C++ support was not built, "make install" and "make test" still
1396 tried to test it.
1397
1398 (b) There were problems when the value of CXX was explicitly set. Some
1399 changes have been made to try to fix these, and ...
1400
1401 (c) --disable-cpp can now be used to explicitly disable C++ support.
1402
1403 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1404 backslash in a target when C++ was disabled. This confuses some
1405 versions of "make", apparently. Using an intermediate variable solves
1406 this. (Same for CPP_LOBJ.)
1407
1408 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1409 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1410 necessary on certain architectures.
1411
1412 4. Added a setting of -export-symbols-regex to the link command to remove
1413 those symbols that are exported in the C sense, but actually are local
1414 within the library, and not documented. Their names all begin with
1415 "_pcre_". This is not a perfect job, because (a) we have to except some
1416 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1417 available (and never for static libraries). I have made a note to try to
1418 find a way round (a) in the future.
1419
1420
1421 Version 6.2 01-Aug-05
1422 ---------------------
1423
1424 1. There was no test for integer overflow of quantifier values. A construction
1425 such as {1111111111111111} would give undefined results. What is worse, if
1426 a minimum quantifier for a parenthesized subpattern overflowed and became
1427 negative, the calculation of the memory size went wrong. This could have
1428 led to memory overwriting.
1429
1430 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1431
1432 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1433 operating environments where this matters.
1434
1435 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1436 PCRE options from within the C++ wrapper.
1437
1438 5. Named capturing subpatterns were not being correctly counted when a pattern
1439 was compiled. This caused two problems: (a) If there were more than 100
1440 such subpatterns, the calculation of the memory needed for the whole
1441 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1442 back references of the form \12, where the number was greater than 9, were
1443 not recognized as back references, even though there were sufficient
1444 previous subpatterns.
1445
1446 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1447 versions of gcc, e.g. 2.95.4.
1448
1449
1450 Version 6.1 21-Jun-05
1451 ---------------------
1452
1453 1. There was one reference to the variable "posix" in pcretest.c that was not
1454 surrounded by "#if !defined NOPOSIX".
1455
1456 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1457 the cross-check on the old pcre_info() function, for the benefit of the
1458 cut-down version of PCRE that is currently imported into Exim.
1459
1460 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1461 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1462 patterns that start (?i) but I don't think that matters. The use of (?i) is
1463 just an example; this all applies to the other options as well.
1464
1465 4. Since libtool seems to echo the compile commands it is issuing, the output
1466 from "make" can be reduced a bit by putting "@" in front of each libtool
1467 compile command.
1468
1469 5. Patch from the folks at Google for configure.in to be a bit more thorough
1470 in checking for a suitable C++ installation before trying to compile the
1471 C++ stuff. This should fix a reported problem when a compiler was present,
1472 but no suitable headers.
1473
1474 6. The man pages all had just "PCRE" as their title. I have changed them to
1475 be the relevant file name. I have also arranged that these names are
1476 retained in the file doc/pcre.txt, which is a concatenation in text format
1477 of all the man pages except the little individual ones for each function.
1478
1479 7. The NON-UNIX-USE file had not been updated for the different set of source
1480 files that come with release 6. I also added a few comments about the C++
1481 wrapper.
1482
1483
1484 Version 6.0 07-Jun-05
1485 ---------------------
1486
1487 1. Some minor internal re-organization to help with my DFA experiments.
1488
1489 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1490 didn't matter for the library itself when fully configured, but did matter
1491 when compiling without UCP support, or within Exim, where the ucp files are
1492 not imported.
1493
1494 3. Refactoring of the library code to split up the various functions into
1495 different source modules. The addition of the new DFA matching code (see
1496 below) to a single monolithic source would have made it really too
1497 unwieldy, quite apart from causing all the code to be include in a
1498 statically linked application, when only some functions are used. This is
1499 relevant even without the DFA addition now that patterns can be compiled in
1500 one application and matched in another.
1501
1502 The downside of splitting up is that there have to be some external
1503 functions and data tables that are used internally in different modules of
1504 the library but which are not part of the API. These have all had their
1505 names changed to start with "_pcre_" so that they are unlikely to clash
1506 with other external names.
1507
1508 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1509 a different (DFA) algorithm. Although it is slower than the original
1510 function, it does have some advantages for certain types of matching
1511 problem.
1512
1513 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1514 including restarting after a partial match.
1515
1516 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1517 defined when compiling for Windows was sent to me. I have put it into the
1518 code, though I have no means of testing or verifying it.
1519
1520 7. Added the pcre_refcount() auxiliary function.
1521
1522 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1523 match before or at the first newline in the subject string. In pcretest,
1524 the /f option on a pattern can be used to set this.
1525
1526 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1527 would behave wrongly. This has been present in PCRE since release 4.0.
1528
1529 10. A number of changes to the pcregrep command:
1530
1531 (a) Refactored how -x works; insert ^(...)$ instead of setting
1532 PCRE_ANCHORED and checking the length, in preparation for adding
1533 something similar for -w.
1534
1535 (b) Added the -w (match as a word) option.
1536
1537 (c) Refactored the way lines are read and buffered so as to have more
1538 than one at a time available.
1539
1540 (d) Implemented a pcregrep test script.
1541
1542 (e) Added the -M (multiline match) option. This allows patterns to match
1543 over several lines of the subject. The buffering ensures that at least
1544 8K, or the rest of the document (whichever is the shorter) is available
1545 for matching (and similarly the previous 8K for lookbehind assertions).
1546
1547 (f) Changed the --help output so that it now says
1548
1549 -w, --word-regex(p)
1550
1551 instead of two lines, one with "regex" and the other with "regexp"
1552 because that confused at least one person since the short forms are the
1553 same. (This required a bit of code, as the output is generated
1554 automatically from a table. It wasn't just a text change.)
1555
1556 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1557 option but starts with a hyphen. Could be a pattern or a path name
1558 starting with a hyphen, for instance.
1559
1560 (h) "-" can be given as a file name to represent stdin.
1561
1562 (i) When file names are being printed, "(standard input)" is used for
1563 the standard input, for compatibility with GNU grep. Previously
1564 "<stdin>" was used.
1565
1566 (j) The option --label=xxx can be used to supply a name to be used for
1567 stdin when file names are being printed. There is no short form.
1568
1569 (k) Re-factored the options decoding logic because we are going to add
1570 two more options that take data. Such options can now be given in four
1571 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1572
1573 (l) Added the -A, -B, and -C options for requesting that lines of context
1574 around matches be printed.
1575
1576 (m) Added the -L option to print the names of files that do not contain
1577 any matching lines, that is, the complement of -l.
1578
1579 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1580 continue to scan other files.
1581
1582 (o) The -s option was incorrectly implemented. For compatibility with other
1583 greps, it now suppresses the error message for a non-existent or non-
1584 accessible file (but not the return code). There is a new option called
1585 -q that suppresses the output of matching lines, which was what -s was
1586 previously doing.
1587
1588 (p) Added --include and --exclude options to specify files for inclusion
1589 and exclusion when recursing.
1590
1591 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1592 Hopefully, it now does.
1593
1594 12. Missing cast in pcre_study().
1595
1596 13. Added an "uninstall" target to the makefile.
1597
1598 14. Replaced "extern" in the function prototypes in Makefile.in with
1599 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1600 world, but is set differently for Windows.
1601
1602 15. Added a second compiling function called pcre_compile2(). The only
1603 difference is that it has an extra argument, which is a pointer to an
1604 integer error code. When there is a compile-time failure, this is set
1605 non-zero, in addition to the error test pointer being set to point to an
1606 error message. The new argument may be NULL if no error number is required
1607 (but then you may as well call pcre_compile(), which is now just a
1608 wrapper). This facility is provided because some applications need a
1609 numeric error indication, but it has also enabled me to tidy up the way
1610 compile-time errors are handled in the POSIX wrapper.
1611
1612 16. Added VPATH=.libs to the makefile; this should help when building with one
1613 prefix path and installing with another. (Or so I'm told by someone who
1614 knows more about this stuff than I do.)
1615
1616 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1617 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1618 match everything, including newlines. This is not POSIX-compatible, but
1619 somebody wanted the feature. From pcretest it can be activated by using
1620 both the P and the s flags.
1621
1622 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1623
1624 19. libpcre.pc was being incorrectly installed as executable.
1625
1626 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1627 it now also looks for '\r' so that it will work unmodified on Windows.
1628
1629 21. Added Google's contributed C++ wrapper to the distribution.
1630
1631 22. Added some untidy missing memory free() calls in pcretest, to keep
1632 Electric Fence happy when testing.
1633
1634
1635
1636 Version 5.0 13-Sep-04
1637 ---------------------
1638
1639 1. Internal change: literal characters are no longer packed up into items
1640 containing multiple characters in a single byte-string. Each character
1641 is now matched using a separate opcode. However, there may be more than one
1642 byte in the character in UTF-8 mode.
1643
1644 2. The pcre_callout_block structure has two new fields: pattern_position and
1645 next_item_length. These contain the offset in the pattern to the next match
1646 item, and its length, respectively.
1647
1648 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1649 insertion of callouts before each pattern item. Added the /C option to
1650 pcretest to make use of this.
1651
1652 4. On the advice of a Windows user, the lines
1653
1654 #if defined(_WIN32) || defined(WIN32)
1655 _setmode( _fileno( stdout ), 0x8000 );
1656 #endif /* defined(_WIN32) || defined(WIN32) */
1657
1658 have been added to the source of pcretest. This apparently does useful
1659 magic in relation to line terminators.
1660
1661 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1662 for the benefit of those environments where the "b" makes a difference.
1663
1664 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1665 to know about it. I have put a hack into configure.in that adds in code
1666 to set GCC=yes if CC=icc. This seems to end up at a point in the
1667 generated configure script that is early enough to affect the setting of
1668 compiler options, which is what is needed, but I have no means of testing
1669 whether it really works. (The user who reported this had patched the
1670 generated configure script, which of course I cannot do.)
1671
1672 LATER: After change 22 below (new libtool files), the configure script
1673 seems to know about icc (and also ecc). Therefore, I have commented out
1674 this hack in configure.in.
1675
1676 7. Added support for pkg-config (2 patches were sent in).
1677
1678 8. Negated POSIX character classes that used a combination of internal tables
1679 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1680 [[:^ascii]]. Typically, they would match almost any characters. The other
1681 POSIX classes were not broken in this way.
1682
1683 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1684 to find the match, as PCRE was deluded into thinking that the match had to
1685 start at the start point or following a newline. The same bug applied to
1686 patterns with negative forward assertions or any backward assertions
1687 preceding ".*" at the start, unless the pattern required a fixed first
1688 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1689
1690 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1691 starting at the last subject character, bytes beyond the end of the subject
1692 string were read.
1693
1694 11. Renamed the variable "class" as "classbits" to make life easier for C++
1695 users. (Previously there was a macro definition, but it apparently wasn't
1696 enough.)
1697
1698 12. Added the new field "tables" to the extra data so that tables can be passed
1699 in at exec time, or the internal tables can be re-selected. This allows
1700 a compiled regex to be saved and re-used at a later time by a different
1701 program that might have everything at different addresses.
1702
1703 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1704 -R library as well as a -L library.
1705
1706 14. The debugging options of pcretest (-d on the command line or D on a
1707 pattern) showed incorrect output for anything following an extended class
1708 that contained multibyte characters and which was followed by a quantifier.
1709
1710 15. Added optional support for general category Unicode character properties
1711 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1712 support. It adds about 90K to the size of the library. The meanings of the
1713 inbuilt class escapes such as \d and \s have NOT been changed.
1714
1715 16. Updated pcredemo.c to include calls to free() to release the memory for the
1716 compiled pattern.
1717
1718 17. The generated file chartables.c was being created in the source directory
1719 instead of in the building directory. This caused the build to fail if the
1720 source directory was different from the building directory, and was
1721 read-only.
1722
1723 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1724 file. No doubt somebody will tell me if they don't make sense... Also added
1725 Dan Mooney's comments about building on OpenVMS.
1726
1727 19. Added support for partial matching via the PCRE_PARTIAL option for
1728 pcre_exec() and the \P data escape in pcretest.
1729
1730 20. Extended pcretest with 3 new pattern features:
1731
1732 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1733 write the compiled pattern to the file whose name is "rest-of-line".
1734 This is a straight binary dump of the data, with the saved pointer to
1735 the character tables forced to be NULL. The study data, if any, is
1736 written too. After writing, pcretest reads a new pattern.
1737
1738 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1739 compiled pattern from the given file. There must not be any
1740 occurrences of "<" in the file name (pretty unlikely); if there are,
1741 pcretest will instead treat the initial "<" as a pattern delimiter.
1742 After reading in the pattern, pcretest goes on to read data lines as
1743 usual.
1744
1745 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1746 and 16-bit fields in a compiled pattern, to simulate a pattern that
1747 was compiled on a host of opposite endianness.
1748
1749 21. The pcre-exec() function can now cope with patterns that were compiled on
1750 hosts of opposite endianness, with this restriction:
1751
1752 As for any compiled expression that is saved and used later, the tables
1753 pointer field cannot be preserved; the extra_data field in the arguments
1754 to pcre_exec() should be used to pass in a tables address if a value
1755 other than the default internal tables were used at compile time.
1756
1757 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1758 now diagnosed as an error. Previously, most of the time, a negative number
1759 would have been treated as zero, but if in addition "ovector" was passed as
1760 NULL, a crash could occur.
1761
1762 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1763 new versions from the libtool 1.5 distribution (the last one is a copy of
1764 a file called libtool.m4). This seems to have fixed the need to patch
1765 "configure" to support Darwin 1.3 (which I used to do). However, I still
1766 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1767 workstation).
1768
1769 24. Changed the PCRE licence to be the more standard "BSD" licence.
1770
1771
1772 Version 4.5 01-Dec-03
1773 ---------------------
1774
1775 1. There has been some re-arrangement of the code for the match() function so
1776 that it can be compiled in a version that does not call itself recursively.
1777 Instead, it keeps those local variables that need separate instances for
1778 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1779 needs to "recurse". Keeping track of where control must go is done by means
1780 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1781 hide most of the details from the main code, and operates only if
1782 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1783 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1784 operating.
1785
1786 To make it easier for callers to provide specially tailored get/free
1787 functions for this usage, two new functions, pcre_stack_malloc, and
1788 pcre_stack_free, are used. They are always called in strict stacking order,
1789 and the size of block requested is always the same.
1790
1791 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1792 PCRE has been compiled to use the stack or the heap for recursion. The
1793 -C option of pcretest uses this to show which version is compiled.
1794
1795 A new data escape \S, is added to pcretest; it causes the amounts of store
1796 obtained and freed by both kinds of malloc/free at match time to be added
1797 to the output.
1798
1799 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1800 what's available on my current Linux desktop machine.
1801
1802 3. When matching a UTF-8 string, the test for a valid string at the start has
1803 been extended. If start_offset is not zero, PCRE now checks that it points
1804 to a byte that is the start of a UTF-8 character. If not, it returns
1805 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1806 this is necessary because there may be backward assertions in the pattern.
1807 When matching the same subject several times, it may save resources to use
1808 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1809
1810 4. The code for checking the validity of UTF-8 strings has been tightened so
1811 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1812 containing "overlong sequences".
1813
1814 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1815 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1816 should have been "&", but it just so happened that all the cases this let
1817 through by mistake were picked up later in the function.
1818
1819 6. I had used a variable called "isblank" - this is a C99 function, causing
1820 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1821
1822 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1823 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1824 so the version is shown; (c) stop "make test" echoing "./RunTest".
1825
1826 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1827
1828 9. The prototype for memmove() for systems that don't have it was using
1829 size_t, but the inclusion of the header that defines size_t was later. I've
1830 moved the #includes for the C headers earlier to avoid this.
1831
1832 10. Added some adjustments to the code to make it easier to compiler on certain
1833 special systems:
1834
1835 (a) Some "const" qualifiers were missing.
1836 (b) Added the macro EXPORT before all exported functions; by default this
1837 is defined to be empty.
1838 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1839 that it reads its output file name as an argument instead of writing
1840 to the standard output and assuming this can be redirected.
1841
1842 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1843 class containing characters with values greater than 255, PCRE compilation
1844 went into a loop.
1845
1846 12. A recursive reference to a subpattern that was within another subpattern
1847 that had a minimum quantifier of zero caused PCRE to crash. For example,
1848 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1849 recursion. If the recursively-called subpattern itself had a zero repeat,
1850 that was OK.
1851
1852 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1853 buffer into which it was copied (for escape processing) was still set at
1854 1024, so long lines caused crashes.
1855
1856 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1857 "internal error: code overflow...". This applied to any character class
1858 that was followed by a possessive quantifier.
1859
1860 15. Modified the Makefile to add libpcre.la as a prerequisite for
1861 libpcreposix.la because I was told this is needed for a parallel build to
1862 work.
1863
1864 16. If a pattern that contained .* following optional items at the start was
1865 studied, the wrong optimizing data was generated, leading to matching
1866 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1867 matching string must start with a or b or c. The correct conclusion for
1868 this pattern is that a match can start with any character.
1869
1870
1871 Version 4.4 13-Aug-03
1872 ---------------------
1873
1874 1. In UTF-8 mode, a character class containing characters with values between
1875 127 and 255 was not handled correctly if the compiled pattern was studied.
1876 In fixing this, I have also improved the studying algorithm for such
1877 classes (slightly).
1878
1879 2. Three internal functions had redundant arguments passed to them. Removal
1880 might give a very teeny performance improvement.
1881
1882 3. Documentation bug: the value of the capture_top field in a callout is *one
1883 more than* the number of the hightest numbered captured substring.
1884
1885 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1886 in incorrectly linking with a previously installed version. They now link
1887 explicitly with libpcre.la.
1888
1889 5. configure.in no longer needs to recognize Cygwin specially.
1890
1891 6. A problem in pcre.in for Windows platforms is fixed.
1892
1893 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1894 pcretest, it used to include the size of the study block as part of its
1895 output. Unfortunately, the structure contains a field that has a different
1896 size on different hardware architectures. This meant that the tests that
1897 showed this size failed. As the block is currently always of a fixed size,
1898 this information isn't actually particularly useful in pcretest output, so
1899 I have just removed it.
1900
1901 8. Three pre-processor statements accidentally did not start in column 1.
1902 Sadly, there are *still* compilers around that complain, even though
1903 standard C has not required this for well over a decade. Sigh.
1904
1905 9. In pcretest, the code for checking callouts passed small integers in the
1906 callout_data field, which is a void * field. However, some picky compilers
1907 complained about the casts involved for this on 64-bit systems. Now
1908 pcretest passes the address of the small integer instead, which should get
1909 rid of the warnings.
1910
1911 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1912 both compile and run time, and gives an error if an invalid UTF-8 sequence
1913 is found. There is a option for disabling this check in cases where the
1914 string is known to be correct and/or the maximum performance is wanted.
1915
1916 11. In response to a bug report, I changed one line in Makefile.in from
1917
1918 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1919 to
1920 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1921
1922 to look similar to other lines, but I have no way of telling whether this
1923 is the right thing to do, as I do not use Windows. No doubt I'll get told
1924 if it's wrong...
1925
1926
1927 Version 4.3 21-May-03
1928 ---------------------
1929
1930 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1931 Makefile.
1932
1933 2. Some refactoring to improve the quality of the code:
1934
1935 (i) The utf8_table... variables are now declared "const".
1936
1937 (ii) The code for \cx, which used the "case flipping" table to upper case
1938 lower case letters, now just substracts 32. This is ASCII-specific,
1939 but the whole concept of \cx is ASCII-specific, so it seems
1940 reasonable.
1941
1942 (iii) PCRE was using its character types table to recognize decimal and
1943 hexadecimal digits in the pattern. This is silly, because it handles
1944 only 0-9, a-f, and A-F, but the character types table is locale-
1945 specific, which means strange things might happen. A private
1946 table is now used for this - though it costs 256 bytes, a table is
1947 much faster than multiple explicit tests. Of course, the standard
1948 character types table is still used for matching digits in subject
1949 strings against \d.
1950
1951 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1952 ending in _t are). So I've renamed it as ESC_tee.
1953
1954 3. The first argument for regexec() in the POSIX wrapper should have been
1955 defined as "const".
1956
1957 4. Changed pcretest to use malloc() for its buffers so that they can be
1958 Electric Fenced for debugging.
1959
1960 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1961 to read one or more bytes before the start of the subject string. Often this
1962 had no effect on PCRE's behaviour, but in some circumstances it could
1963 provoke a segmentation fault.
1964
1965 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1966 to try to read one or more bytes before the start of the subject string.
1967
1968 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1969 UTF-8 support could misbehave in various ways if the subject string
1970 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1971 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1972 back over UTF-8 characters.)
1973
1974
1975 Version 4.2 14-Apr-03
1976 ---------------------
1977
1978 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1979
1980 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1981 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1982 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1983 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1984 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1985 and BUILD_EXEEXT
1986 Note: automatic setting of the BUILD variables is not yet working
1987 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1988 compile-time but not at link-time
1989 [LINK]: use for linking executables only
1990 make different versions for Windows and non-Windows
1991 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1992 libraries
1993 [LINK_FOR_BUILD]: new variable
1994 [OBJEXT]: use throughout
1995 [EXEEXT]: use throughout
1996 <winshared>: new target
1997 <wininstall>: new target
1998 <dftables.o>: use native compiler
1999 <dftables>: use native linker
2000 <install>: handle Windows platform correctly
2001 <clean>: ditto
2002 <check>: ditto
2003 copy DLL to top builddir before testing
2004
2005 As part of these changes, -no-undefined was removed again. This was reported
2006 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2007 in any case.
2008
2009 3. Some tidies to get rid of compiler warnings:
2010
2011 . In the match_data structure, match_limit was an unsigned long int, whereas
2012 match_call_count was an int. I've made them both unsigned long ints.
2013
2014 . In pcretest the fact that a const uschar * doesn't automatically cast to
2015 a void * provoked a warning.
2016
2017 . Turning on some more compiler warnings threw up some "shadow" variables
2018 and a few more missing casts.
2019
2020 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2021 option, a class that contained a single character with a value between 128
2022 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2023
2024 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2025 option, a class that contained several characters, but with at least one
2026 whose value was between 128 and 255 caused PCRE to crash.
2027
2028
2029 Version 4.1 12-Mar-03
2030 ---------------------
2031
2032 1. Compiling with gcc -pedantic found a couple of places where casts were
2033 needed, and a string in dftables.c that was longer than standard compilers are
2034 required to support.
2035
2036 2. Compiling with Sun's compiler found a few more places where the code could
2037 be tidied up in order to avoid warnings.
2038
2039 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2040 first of these names is deprecated in the latest Autoconf in favour of the name
2041 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2042 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2043 analogy I have changed it to CFLAGS_FOR_BUILD.
2044
2045 4. Added -no-undefined to the linking command in the Makefile, because this is
2046 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2047 linking step for the pcreposix library.
2048
2049 5. PCRE was failing to diagnose the case of two named groups with the same
2050 name.
2051
2052 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2053 literal character that is needed in the subject for a match, and scans along to
2054 ensure that it is present before embarking on the full matching process. This
2055 saves time in cases of nested unlimited repeats that are never going to match.
2056 Problem: the scan can take a lot of time if the subject is very long (e.g.
2057 megabytes), thus penalizing straightforward matches. It is now done only if the
2058 amount of subject to be scanned is less than 1000 bytes.
2059
2060 7. A lesser problem with the same optimization is that it was recording the
2061 first character of an anchored pattern as "needed", thus provoking a search
2062 right along the subject, even when the first match of the pattern was going to
2063 fail. The "needed" character is now not set for anchored patterns, unless it
2064 follows something in the pattern that is of non-fixed length. Thus, it still
2065 fulfils its original purpose of finding quick non-matches in cases of nested
2066 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2067
2068
2069 Version 4.0 17-Feb-03
2070 ---------------------
2071
2072 1. If a comment in an extended regex that started immediately after a meta-item
2073 extended to the end of string, PCRE compiled incorrect data. This could lead to
2074 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2075
2076 2. Moved to autoconf 2.53 and libtool 1.4.2.
2077
2078 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2079 the special perltest8 script is no longer needed - all the tests can be run
2080 from a single perltest script.
2081
2082 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2083 by \s. It has now been removed in PCRE. This means it isn't recognized as
2084 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2085 class [:space:] *does* include VT, thereby creating a mess.
2086
2087 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2088 space and tab.
2089
2090 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2091 its new features into the main test script, reducing the number of scripts.
2092
2093 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2094 were backward compatible, and made the (?i) apply to the whole pattern, as if
2095 /i were given. Now it behaves more logically, and applies the option setting
2096 only to what follows. PCRE has been changed to follow suit. However, if it
2097 finds options settings right at the start of the pattern, it extracts them into
2098 the global options, as before. Thus, they show up in the info data.
2099
2100 8. Added support for the \Q...\E escape sequence. Characters in between are
2101 treated as literals. This is slightly different from Perl in that $ and @ are
2102 also handled as literals inside the quotes. In Perl, they will cause variable
2103 interpolation. Note the following examples:
2104
2105 Pattern PCRE matches Perl matches
2106
2107 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2108 \Qabc\$xyz\E abc\$xyz abc\$xyz
2109 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2110
2111 For compatibility with Perl, \Q...\E sequences are recognized inside character
2112 classes as well as outside them.
2113
2114 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2115 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2116 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2117 signed/unsigned warnings.
2118
2119 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2120 option for pcretest, so I've replaced it by a simple function that does just
2121 that job.
2122
2123 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2124 "pcregrep -".
2125
2126 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2127 Java package. This provides some syntactic sugar for simple cases of what my
2128 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2129 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2130 item, you can use this simplified notation. Note that only makes sense with
2131 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2132 greediness, whatever the setting of the PCRE_UNGREEDY option.
2133
2134 13. A change of greediness default within a pattern was not taking effect at
2135 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2136 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2137 was abstracted outside.
2138
2139 14. PCRE now supports the \G assertion. It is true when the current matching
2140 position is at the start point of the match. This differs from \A when the
2141 starting offset is non-zero. Used with the /g option of pcretest (or similar
2142 code), it works in the same way as it does for Perl's /g option. If all
2143 alternatives of a regex begin with \G, the expression is anchored to the start
2144 match position, and the "anchored" flag is set in the compiled expression.
2145
2146 15. Some bugs concerning the handling of certain option changes within patterns
2147 have been fixed. These applied to options other than (?ims). For example,
2148 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2149 been the other way round. Some of this was related to change 7 above.
2150
2151 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2152 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2153 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2154 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2155
2156 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2157 mode. Unlike ".", it always matches newline, whatever the setting of
2158 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2159 assertions. Perl allows it, but it doesn't (in general) work because it can't
2160 calculate the length of the lookbehind. At least, that's the case for Perl
2161 5.8.0 - I've been told they are going to document that it doesn't work in
2162 future.
2163
2164 18. Added an error diagnosis for escapes that PCRE does not support: these are
2165 \L, \l, \N, \P, \p, \U, \u, and \X.
2166
2167 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2168 reading past the end of the pattern in cases such as /[abcd/.
2169
2170 20. PCRE was getting more memory than necessary for patterns with classes that
2171 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2172
2173 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2174 compiling PCRE for use with Virtual Pascal.
2175
2176 22. Small fix to the Makefile to make it work properly if the build is done
2177 outside the source tree.
2178
2179 23. Added a new extension: a condition to go with recursion. If a conditional
2180 subpattern starts with (?(R) the "true" branch is used if recursion has
2181 happened, whereas the "false" branch is used only at the top level.
2182
2183 24. When there was a very long string of literal characters (over 255 bytes
2184 without UTF support, over 250 bytes with UTF support), the computation of how
2185 much memory was required could be incorrect, leading to segfaults or other
2186 strange effects.
2187
2188 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2189 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2190 there was a subsequent back reference to those brackets. This meant that, for
2191 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2192 possible to check for precisely this case. All we can do is abandon the
2193 optimization if .* occurs inside capturing brackets when there are any back
2194 references whatsoever. (See below for a better fix that came later.)
2195
2196 26. The handling of the optimization for finding the first character of a
2197 non-anchored pattern, and for finding a character that is required later in the
2198 match were failing in some cases. This didn't break the matching; it just
2199 failed to optimize when it could. The way this is done has been re-implemented.
2200
2201 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2202
2203 28. Added a new feature that provides some of the functionality that Perl
2204 provides with (?{...}). The facility is termed a "callout". The way it is done
2205 in PCRE is for the caller to provide an optional function, by setting
2206 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2207 global variable. By default it is unset, which disables all calling out. To get
2208 the function called, the regex must include (?C) at appropriate points. This
2209 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2210 This provides a means of identifying different callout points. When PCRE
2211 reaches such a point in the regex, if pcre_callout has been set, the external
2212 function is called. It is provided with data in a structure called
2213 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2214 matching continues; if it returns a non-zero value, the match at the current
2215 point fails. However, backtracking will occur if possible. [This was changed
2216 later and other features added - see item 49 below.]
2217
2218 29. pcretest is upgraded to test the callout functionality. It provides a
2219 callout function that displays information. By default, it shows the start of
2220 the match and the current position in the text. There are some new data escapes
2221 to vary what happens:
2222
2223 \C+ in addition, show current contents of captured substrings
2224 \C- do not supply a callout function
2225 \C!n return 1 when callout number n is reached
2226 \C!n!m return 1 when callout number n is reached for the mth time
2227
2228 30. If pcregrep was called with the -l option and just a single file name, it
2229 output "<stdin>" if a match was found, instead of the file name.
2230
2231 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2232 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2233 pcre_exec(). This saves a malloc/free per call. The default value of
2234 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2235 when configuring.
2236
2237 32. The default maximum size of a compiled pattern is 64K. There have been a
2238 few cases of people hitting this limit. The code now uses macros to handle the
2239 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2240 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2241 configuring. Tests 2 and 5 work only with 2-byte links because they output
2242 debugging information about compiled patterns.
2243
2244 33. Internal code re-arrangements:
2245
2246 (a) Moved the debugging function for printing out a compiled regex into
2247 its own source file (printint.c) and used #include to pull it into
2248 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2249 separate copies.
2250
2251 (b) Defined the list of op-code names for debugging as a macro in
2252 internal.h so that it is next to the definition of the opcodes.
2253
2254 (c) Defined a table of op-code lengths for simpler skipping along compiled
2255 code. This is again a macro in internal.h so that it is next to the
2256 definition of the opcodes.
2257
2258 34. Added support for recursive calls to individual subpatterns, along the
2259 lines of Robin Houston's patch (but implemented somewhat differently).
2260
2261 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2262 allow it to read and process whole directories in Win32. This code was
2263 contributed by Lionel Fourquaux; it has not been tested by me.
2264
2265 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2266 used to name a group. Names consist of alphanumerics and underscores, and must
2267 be unique. Back references use the syntax (?P=name) and recursive calls use
2268 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2269 numbers. The function pcre_fullinfo() can be used after compilation to extract
2270 a name/number map. There are three relevant calls:
2271
2272 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2273 PCRE_INFO_NAMECOUNT yields the number of entries
2274 PCRE_INFO_NAMETABLE yields a pointer to the map.
2275
2276 The map is a vector of fixed-size entries. The size of each entry depends on
2277 the length of the longest name used. The first two bytes of each entry are the
2278 group number, most significant byte first. There follows the corresponding
2279 name, zero terminated. The names are in alphabetical order.
2280
2281 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2282 case instead of 255. Making it the same both with and without UTF-8 support
2283 means that the same test output works with both.
2284
2285 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2286 calling malloc() with a zero argument.
2287
2288 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2289 optimization. I've improved things by keeping a bitmap of backreferences with
2290 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2291 fact referenced, the optimization can be applied. It is unlikely that a
2292 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2293 the match to follow \n) will appear inside brackets with a number greater than
2294 31, but if it does, any back reference > 31 suppresses the optimization.
2295
2296 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2297 of disabling numbered capturing parentheses. Any opening parenthesis that is
2298 not followed by ? behaves as if it were followed by ?: but named parentheses
2299 can still be used for capturing (and they will acquire numbers in the usual
2300 way).
2301
2302 41. Redesigned the return codes from the match() function into yes/no/error so
2303 that errors can be passed back from deep inside the nested calls. A malloc
2304 failure while inside a recursive subpattern call now causes the
2305 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2306
2307 42. It is now possible to set a limit on the number of times the match()
2308 function is called in a call to pcre_exec(). This facility makes it possible to
2309 limit the amount of recursion and backtracking, though not in a directly
2310 obvious way, because the match() function is used in a number of different
2311 circumstances. The count starts from zero for each position in the subject
2312 string (for non-anchored patterns). The default limit is, for compatibility, a
2313 large number, namely 10 000 000. You can change this in two ways:
2314
2315 (a) When configuring PCRE before making, you can use --with-match-limit=n
2316 to set a default value for the compiled library.
2317
2318 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2319 a different value is set. See 45 below.
2320
2321 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2322
2323 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2324 of things that can be changed at compile time. The first argument specifies
2325 what is wanted and the second points to where the information is to be placed.
2326 The current list of available information is:
2327
2328 PCRE_CONFIG_UTF8
2329
2330 The output is an integer that is set to one if UTF-8 support is available;
2331 otherwise it is set to zero.
2332
2333 PCRE_CONFIG_NEWLINE
2334
2335 The output is an integer that it set to the value of the code that is used for
2336 newline. It is either LF (10) or CR (13).
2337
2338 PCRE_CONFIG_LINK_SIZE
2339
2340 The output is an integer that contains the number of bytes used for internal
2341 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2342
2343 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2344
2345 The output is an integer that contains the threshold above which the POSIX
2346 interface uses malloc() for output vectors. See item 31 above.
2347
2348 PCRE_CONFIG_MATCH_LIMIT
2349
2350 The output is an unsigned integer that contains the default limit of the number
2351 of match() calls in a pcre_exec() execution. See 42 above.
2352
2353 44. pcretest has been upgraded by the addition of the -C option. This causes it
2354 to extract all the available output from the new pcre_config() function, and to
2355 output it. The program then exits immediately.
2356
2357 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2358 order to support additional features. One way would have been to define
2359 pcre_exec2() (for example) with extra arguments, but this would not have been
2360 extensible, and would also have required all calls to the original function to
2361 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2362 is used for passing in "extra" data from pcre_study().
2363
2364 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2365 contains the following fields:
2366
2367 flags a bitmap indicating which of the following fields are set
2368 study_data opaque data from pcre_study()
2369 match_limit a way of specifying a limit on match() calls for a specific
2370 call to pcre_exec()
2371 callout_data data for callouts (see 49 below)
2372
2373 The flag bits are also defined in pcre.h, and are
2374
2375 PCRE_EXTRA_STUDY_DATA
2376 PCRE_EXTRA_MATCH_LIMIT
2377 PCRE_EXTRA_CALLOUT_DATA
2378
2379 The pcre_study() function now returns one of these new pcre_extra blocks, with
2380 the actual study data pointed to by the study_data field, and the
2381 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2382 before. That is, this change is entirely upwards-compatible and requires no
2383 change to existing code.
2384
2385 If you want to pass in additional data to pcre_exec(), you can either place it
2386 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2387 block.
2388
2389 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2390 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2391 times with different match limits, until it finds the minimum value needed for
2392 pcre_exec() to complete. The value is then output. This can be instructive; for
2393 most simple matches the number is quite small, but for pathological cases it
2394 gets very large very quickly.
2395
2396 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2397 returns the size of the data block pointed to by the study_data field in a
2398 pcre_extra block, that is, the value that was passed as the argument to
2399 pcre_malloc() when PCRE was getting memory in which to place the information
2400 created by pcre_study(). The fourth argument should point to a size_t variable.
2401 pcretest has been extended so that this information is shown after a successful
2402 pcre_study() call when information about the compiled regex is being displayed.
2403
2404 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2405 because what follows is always an absolute path. (Later: it turns out that this
2406 is more than cosmetic for MinGW, because it doesn't like empty path
2407 components.)
2408
2409 49. Some changes have been made to the callout feature (see 28 above):
2410
2411 (i) A callout function now has three choices for what it returns:
2412
2413 0 => success, carry on matching
2414 > 0 => failure at this point, but backtrack if possible
2415 < 0 => serious error, return this value from pcre_exec()
2416
2417 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2418 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2419 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2420 use by callout functions. It will never be used by PCRE itself.
2421
2422 (ii) The pcre_extra structure (see 45 above) has a void * field called
2423 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2424 pcre_callout_block structure has a field of the same name. The contents of
2425 the field passed in the pcre_extra structure are passed to the callout
2426 function in the corresponding field in the callout block. This makes it
2427 easier to use the same callout-containing regex from multiple threads. For
2428 testing, the pcretest program has a new data escape
2429
2430 \C*n pass the number n (may be negative) as callout_data
2431
2432 If the callout function in pcretest receives a non-zero value as
2433 callout_data, it returns that value.
2434
2435 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2436 there were some redundant $(CFLAGS) in commands that are now specified as
2437 $(LINK), which already includes $(CFLAGS).
2438
2439 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2440 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2441 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2442 one-byte characters throughout. Note that case-insensitive matching applies
2443 only to characters whose values are less than 256. PCRE doesn't support the
2444 notion of cases for higher-valued characters.
2445
2446 (i) A character class whose characters are all within 0-255 is handled as
2447 a bit map, and the map is inverted for negative classes. Previously, a
2448 character > 255 always failed to match such a class; however it should
2449 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2450
2451 (ii) A negated character class with a single character < 255 is coded as
2452 "not this character" (OP_NOT). This wasn't working properly when the test
2453 character was multibyte, either singly or repeated.
2454
2455 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2456 mode, for example: \x{100}{2,3}.
2457
2458 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2459 singly or repeated) now correctly test multibyte characters. However,
2460 PCRE doesn't recognize any characters with values greater than 255 as
2461 digits, spaces, or word characters. Such characters always match \D, \S,
2462 and \W, and never match \d, \s, or \w.
2463
2464 (v) Classes may now contain characters and character ranges with values
2465 greater than 255. For example: [ab\x{100}-\x{400}].
2466
2467 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2468 PCRE in UTF-8 mode.
2469
2470 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2471 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2472 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2473 value.)
2474
2475 53. The single man page has become too large. I have therefore split it up into
2476 a number of separate man pages. These also give rise to individual HTML pages;
2477 these are now put in a separate directory, and there is an index.html page that
2478 lists them all. Some hyperlinking between the pages has been installed.
2479
2480 54. Added convenience functions for handling named capturing parentheses.
2481
2482 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2483 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2484 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2485 are faulted.
2486
2487 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2488 calling configure. These values are used when compiling the dftables.c program
2489 which is run to generate the source of the default character tables. They
2490 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2491 you will need to set these values.
2492
2493 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2494
2495
2496 Version 3.9 02-Jan-02
2497 ---------------------
2498
2499 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2500
2501 2. If --disable-static was given, the building process failed when trying to
2502 build pcretest and pcregrep. (For some reason it was using libtool to compile
2503 them, which is not right, as they aren't part of the library.)
2504
2505
2506 Version 3.8 18-Dec-01
2507 ---------------------
2508
2509 1. The experimental UTF-8 code was completely screwed up. It was packing the
2510 bytes in the wrong order. How dumb can you get?
2511
2512
2513 Version 3.7 29-Oct-01
2514 ---------------------
2515
2516 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2517 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2518 this didn't happen under Solaris 8, where I normally test things.
2519
2520 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2521 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2522 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2523 this fix an hour or so after the initial 3.7 release.)
2524
2525
2526 Version 3.6 23-Oct-01
2527 ---------------------
2528
2529 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2530 offsets passed as NULL with zero offset count.
2531
2532 2. The config.guess and config.sub files had not been updated when I moved to
2533 the latest autoconf.
2534
2535
2536 Version 3.5 15-Aug-01
2537 ---------------------
2538
2539 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2540 had been forgotten.
2541
2542 2. By using declared but undefined structures, we can avoid using "void"
2543 definitions in pcre.h while keeping the internal definitions of the structures
2544 private.
2545
2546 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2547 user point of view, this means that both static and shared libraries are built
2548 by default, but this can be individually controlled. More of the work of
2549 handling this static/shared cases is now inside libtool instead of PCRE's make
2550 file.
2551
2552 4. The pcretest utility is now installed along with pcregrep because it is
2553 useful for users (to test regexs) and by doing this, it automatically gets
2554 relinked by libtool. The documentation has been turned into a man page, so
2555 there are now .1, .txt, and .html versions in /doc.
2556
2557 5. Upgrades to pcregrep:
2558 (i) Added long-form option names like gnu grep.
2559 (ii) Added --help to list all options with an explanatory phrase.
2560 (iii) Added -r, --recursive to recurse into sub-directories.
2561 (iv) Added -f, --file to read patterns from a file.
2562
2563 6. pcre_exec() was referring to its "code" argument before testing that
2564 argument for NULL (and giving an error if it was NULL).
2565
2566 7. Upgraded Makefile.in to allow for compiling in a different directory from
2567 the source directory.
2568
2569 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2570 options bits, the pointer it was passed was to an int instead of to an unsigned
2571 long int. This mattered only on 64-bit systems.
2572
2573 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2574 generated) instead of pcre.in, which it its source. Also made the same change
2575 in several of the .c files.
2576
2577 10. A new release of gcc defines printf() as a macro, which broke pcretest
2578 because it had an ifdef in the middle of a string argument for printf(). Fixed
2579 by using separate calls to printf().
2580
2581 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2582 script, to force use of CR or LF instead of \n in the source. On non-Unix
2583 systems, the value can be set in config.h.
2584
2585 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2586 absolute limit. Changed the text of the error message to make this clear, and
2587 likewise updated the man page.
2588
2589 13. The limit of 99 on the number of capturing subpatterns has been removed.
2590 The new limit is 65535, which I hope will not be a "real" limit.
2591
2592
2593 Version 3.4 22-Aug-00
2594 ---------------------
2595
2596 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2597
2598 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2599
2600
2601 Version 3.3 01-Aug-00
2602 ---------------------
2603
2604 1. If an octal character was given, but the value was greater than \377, it
2605 was not getting masked to the least significant bits, as documented. This could
2606 lead to crashes in some systems.
2607
2608 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2609 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2610
2611 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2612 These just pass their arguments on to (pcre_free)(), but they are provided
2613 because some uses of PCRE bind it to non-C systems that can call its functions,
2614 but cannot call free() or pcre_free() directly.
2615
2616 4. Add "make test" as a synonym for "make check". Corrected some comments in
2617 the Makefile.
2618
2619 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2620 Makefile.
2621
2622 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2623 command called pgrep for grepping around the active processes.
2624
2625 7. Added the beginnings of support for UTF-8 character strings.
2626
2627 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2628 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2629 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2630 out for the ar command.)
2631
2632
2633 Version 3.2 12-May-00
2634 ---------------------
2635
2636 This is purely a bug fixing release.
2637
2638 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2639 of ZA. This was just one example of several cases that could provoke this bug,
2640 which was introduced by change 9 of version 2.00. The code for breaking
2641 infinite loops after an iteration that matches an empty string was't working
2642 correctly.
2643
2644 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2645 when matched against abbab (for example). After matching an empty string, it
2646 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2647 caused it to match further down the string than it should.
2648
2649 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2650 was there because it doesn't seem to be needed, and it causes trouble on some
2651 systems, as it is not a Standard C header. It has been removed.
2652
2653 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2654 were reported on the Macintosh. The changes were from
2655
2656 while ((c = *(++ptr)) != 0 && c != '\n');
2657 to
2658 while ((c = *(++ptr)) != 0 && c != '\n') ;
2659
2660 Totally extraordinary, but if that's what it takes...
2661
2662 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2663 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2664 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2665 assumes the way PCRE uses memmove() (always moving upwards).
2666
2667 6. PCRE is being used in one environment where strchr() is not available. There
2668 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2669 faster code anyway.
2670
2671
2672 Version 3.1 09-Feb-00
2673 ---------------------
2674
2675 The only change in this release is the fixing of some bugs in Makefile.in for
2676 the "install" target:
2677
2678 (1) It was failing to install pcreposix.h.
2679
2680 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2681
2682
2683 Version 3.0 01-Feb-00
2684 ---------------------
2685
2686 1. Add support for the /+ modifier to perltest (to output $` like it does in
2687 pcretest).
2688
2689 2. Add support for the /g modifier to perltest.
2690
2691 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2692 matches null strings.
2693
2694 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2695 pattern. Perl treats empty patterns specially - it reuses the most recent
2696 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2697 effect.
2698
2699 5. The POSIX interface was broken in that it was just handing over the POSIX
2700 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2701 required a bigger vector, with some working space on the end. This means that
2702 the POSIX wrapper now has to get and free some memory, and copy the results.
2703
2704 6. Added some simple autoconf support, placing the test data and the
2705 documentation in separate directories, re-organizing some of the
2706 information files, and making it build pcre-config (a GNU standard). Also added
2707 libtool support for building PCRE as a shared library, which is now the
2708 default.
2709
2710 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2711 09 are not valid octal constants. Single digits will be used for minor values
2712 less than 10.
2713
2714 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2715 existing programs that set these in the POSIX interface can use PCRE without
2716 modification.
2717
2718 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2719 return all that pcre_info() returns, plus additional data. The pcre_info()
2720 function is retained for compatibility, but is considered to be obsolete.
2721
2722 10. Added experimental recursion feature (?R) to handle one common case that
2723 Perl 5.6 will be able to do with (?p{...}).
2724
2725 11. Added support for POSIX character classes like [:alpha:], which Perl is
2726 adopting.
2727
2728
2729 Version 2.08 31-Aug-99
2730 ----------------------
2731
2732 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2733 trying to match at the startoffset position, but instead was moving forward to
2734 the next newline as if a previous match had failed.
2735
2736 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2737 and could get into a loop if a null string was matched other than at the start
2738 of the subject.
2739
2740 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2741 be distinguished at compile time, and for completeness also added PCRE_DATE.
2742
2743 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2744 in GnuWin32 environments.
2745
2746
2747 Version 2.07 29-Jul-99
2748 ----------------------
2749
2750 1. The documentation is now supplied in plain text form and HTML as well as in
2751 the form of man page sources.
2752
2753 2. C++ compilers don't like assigning (void *) values to other pointer types.
2754 In particular this affects malloc(). Although there is no problem in Standard
2755 C, I've put in casts to keep C++ compilers happy.
2756
2757 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2758 should be (const char *).
2759
2760 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2761 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2762 However, I haven't made this a standard facility. The documentation doesn't
2763 mention it, and the Makefile doesn't support it.
2764
2765 5. The Makefile now contains an "install" target, with editable destinations at
2766 the top of the file. The pcretest program is not installed.
2767
2768 6. pgrep -V now gives the PCRE version number and date.
2769
2770 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2771 causing the entire string to be ignored, instead of just the last character.
2772
2773 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2774 non-matching string, it can take a very, very long time, even for strings of
2775 quite modest length, because of the nested recursion. PCRE now does better in
2776 some of these cases. It does this by remembering the last required literal
2777 character in the pattern, and pre-searching the subject to ensure it is present
2778 before running the real match. In other words, it applies a heuristic to detect
2779 some types of certain failure quickly, and in the above example, if presented
2780 with a string that has no trailing " it gives "no match" very quickly.
2781
2782 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2783 other alternatives are tried instead.
2784
2785
2786 Version 2.06 09-Jun-99
2787 ----------------------
2788
2789 1. Change pcretest's output for amount of store used to show just the code
2790 space, because the remainder (the data block) varies in size between 32-bit and
2791 64-bit systems.
2792
2793 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2794 start matching at. This allows lookbehinds to work when searching for multiple
2795 occurrences in a string.
2796
2797 3. Added additional options to pcretest for testing multiple occurrences:
2798
2799 /+ outputs the rest of the string that follows a match
2800 /g loops for multiple occurrences, using the new startoffset argument
2801 /G loops for multiple occurrences by passing an incremented pointer
2802
2803 4. PCRE wasn't doing the "first character" optimization for patterns starting
2804 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2805 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2806 the letter 'x'. On long subject strings, this gives a significant speed-up.
2807
2808
2809 Version 2.05 21-Apr-99
2810 ----------------------
2811
2812 1. Changed the type of magic_number from int to long int so that it works
2813 properly on 16-bit systems.
2814
2815 2. Fixed a bug which caused patterns starting with .* not to work correctly
2816 when the subject string contained newline characters. PCRE was assuming
2817 anchoring for such patterns in all cases, which is not correct because .* will
2818 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2819 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2820 must be retried after every newline in the subject.
2821
2822
2823 Version 2.04 18-Feb-99
2824 ----------------------
2825
2826 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2827 computation of the store needed to hold the pattern was incorrect (too large).
2828 If such patterns were nested a few deep, this could multiply and become a real
2829 problem.
2830
2831 2. Added /M option to pcretest to show the memory requirement of a specific
2832 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2833
2834 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2835 compiled in such a way that the backtracking after subsequent failure was
2836 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2837 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2838
2839
2840 Version 2.03 02-Feb-99
2841 ----------------------
2842
2843 1. Fixed typo and small mistake in man page.
2844
2845 2. Added 4th condition (GPL supersedes if conflict) and created separate
2846 LICENCE file containing the conditions.
2847
2848 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2849 Perl, that is the internal \ allows the delimiter to be included in the
2850 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2851 the final delimiter, add \ to the end of the pattern (to test the error).
2852
2853 4. Added the convenience functions for extracting substrings after a successful
2854 match. Updated pcretest to make it able to test these functions.
2855
2856
2857 Version 2.02 14-Jan-99
2858 ----------------------
2859
2860 1. Initialized the working variables associated with each extraction so that
2861 their saving and restoring doesn't refer to uninitialized store.
2862
2863 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2864 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2865 fix the problem.
2866
2867 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2868 calls, and wasn't printing the correct value for compiling calls. Increased the
2869 default value of LOOPREPEAT, and the number of significant figures in the
2870 times.
2871
2872 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2873
2874 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2875 a building problem on Windows NT with a FAT file system.
2876
2877
2878 Version 2.01 21-Oct-98
2879 ----------------------
2880
2881 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2882 to character tables built by pcre_maketables() in the current locale. If NULL
2883 is passed, the default tables are used.
2884
2885
2886 Version 2.00 24-Sep-98
2887 ----------------------
2888
2889 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2890 it any more.
2891
2892 2. Allow quantification of (?>) groups, and make it work correctly.
2893
2894 3. The first character computation wasn't working for (?>) groups.
2895
2896 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2897 end of the subject) and add 5.005's \z, which really does match only at the
2898 very end of the subject.
2899
2900 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2901
2902 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2903 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2904 localized options. All options to pcre_study() were also removed.
2905
2906 7. Add other new features from 5.005:
2907
2908 $(?<= positive lookbehind
2909 $(?<! negative lookbehind
2910 (?imsx-imsx) added the unsetting capability
2911 such a setting is global if at outer level; local otherwise
2912 (?imsx-imsx:) non-capturing groups with option setting
2913 (?(cond)re|re) conditional pattern matching
2914
2915 A backreference to itself in a repeated group matches the previous
2916 captured string.
2917
2918 8. General tidying up of studying (both automatic and via "study")
2919 consequential on the addition of new assertions.
2920
2921 9. As in 5.005, unlimited repeated groups that could match an empty substring
2922 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2923 runtime if any iteration does actually match an empty substring.
2924
2925 10. Include the RunTest script in the distribution.
2926
2927 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2928 discrepancies, some of which were old and were also with respect to 5.004. They
2929 have now been fixed.
2930
2931
2932 Version 1.09 28-Apr-98
2933 ----------------------
2934
2935 1. A negated single character class followed by a quantifier with a minimum
2936 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2937 program crashes, or just wrong answers. This did not apply to negated classes
2938 containing more than one character, or to minima other than one.
2939
2940
2941 Version 1.08 27-Mar-98
2942 ----------------------
2943
2944 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2945
2946 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2947 latter must appear before anything that relies on it in the pattern.
2948
2949
2950 Version 1.07 16-Feb-98
2951 ----------------------
2952
2953 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2954 repeat of a potentially empty string).
2955
2956
2957 Version 1.06 23-Jan-98
2958 ----------------------
2959
2960 1. Added Markus Oberhumer's little patches for C++.
2961
2962 2. Literal strings longer than 255 characters were broken.
2963
2964
2965 Version 1.05 23-Dec-97
2966 ----------------------
2967
2968 1. Negated character classes containing more than one character were failing if
2969 PCRE_CASELESS was set at run time.
2970
2971
2972 Version 1.04 19-Dec-97
2973 ----------------------
2974
2975 1. Corrected the man page, where some "const" qualifiers had been omitted.
2976
2977 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2978 input syntax.
2979
2980 3. Fixed memory leak which occurred when a regex with back references was
2981 matched with an offsets vector that wasn't big enough. The temporary memory
2982 that is used in this case wasn't being freed if the match failed.
2983
2984 4. Tidied pcretest to ensure it frees memory that it gets.
2985
2986 5. Temporary memory was being obtained in the case where the passed offsets
2987 vector was exactly big enough.
2988
2989 6. Corrected definition of offsetof() from change 5 below.
2990
2991 7. I had screwed up change 6 below and broken the rules for the use of
2992 setjmp(). Now fixed.
2993
2994
2995 Version 1.03 18-Dec-97
2996 ----------------------
2997
2998 1. A erroneous regex with a missing opening parenthesis was correctly
2999 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3000 on some systems.
3001
3002 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3003 it was reported that one broken compiler failed on the former because "code" is
3004 also an independent variable.
3005
3006 3. The erroneous regex a[]b caused an array overrun reference.
3007
3008 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3009 fail on data ending with that character. (It was going on too far, and checking
3010 the next character, typically a binary zero.) This was specific to the
3011 optimized code for single-character negative classes.
3012
3013 5. Added a contributed patch from the TIN world which does the following:
3014
3015 + Add an undef for memmove, in case the the system defines a macro for it.
3016
3017 + Add a definition of offsetof(), in case there isn't one. (I don't know
3018 the reason behind this - offsetof() is part of the ANSI standard - but
3019 it does no harm).
3020
3021 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3022 most of the places where whitespace preceded '#'. I have given up and
3023 allowed the remaining 2 cases to be at the margin.
3024
3025 + Rename some variables in pcre to eliminate shadowing. This seems very
3026 pedantic, but does no harm, of course.
3027
3028 6. Moved the call to setjmp() into its own function, to get rid of warnings
3029 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3030
3031 7. Constructs such as \d{8,} were compiling into the equivalent of
3032 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3033 outcome, but in this particular case used more store than had been allocated,
3034 which caused the bug to be discovered because it threw up an internal error.
3035
3036 8. The debugging code in both pcre and pcretest for outputting the compiled
3037 form of a regex was going wrong in the case of back references followed by
3038 curly-bracketed repeats.
3039
3040
3041 Version 1.02 12-Dec-97
3042 ----------------------
3043
3044 1. Typos in pcre.3 and comments in the source fixed.
3045
3046 2. Applied a contributed patch to get rid of places where it used to remove
3047 'const' from variables, and fixed some signed/unsigned and uninitialized
3048 variable warnings.
3049
3050 3. Added the "runtest" target to Makefile.
3051
3052 4. Set default compiler flag to -O2 rather than just -O.
3053
3054
3055 Version 1.01 19-Nov-97
3056 ----------------------
3057
3058 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3059 like /([ab]*)*/, that is, for classes with more than one character in them.
3060
3061 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3062 as /((?>a*))*/ (a PCRE_EXTRA facility).
3063
3064
3065 Version 1.00 18-Nov-97
3066 ----------------------
3067
3068 1. Added compile-time macros to support systems such as SunOS4 which don't have
3069 memmove() or strerror() but have other things that can be used instead.
3070
3071 2. Arranged that "make clean" removes the executables.
3072
3073
3074 Version 0.99 27-Oct-97
3075 ----------------------
3076
3077 1. Fixed bug in code for optimizing classes with only one character. It was
3078 initializing a 32-byte map regardless, which could cause it to run off the end
3079 of the memory it had got.
3080
3081 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3082
3083
3084 Version 0.98 22-Oct-97
3085 ----------------------
3086
3087 1. Fixed bug in code for handling temporary memory usage when there are more
3088 back references than supplied space in the ovector. This could cause segfaults.
3089
3090
3091 Version 0.97 21-Oct-97
3092 ----------------------
3093
3094 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3095
3096 2. Optimized negated single characters not to use a bit map.
3097
3098 3. Brought error texts together as macro definitions; clarified some of them;
3099 fixed one that was wrong - it said "range out of order" when it meant "invalid
3100 escape sequence".
3101
3102 4. Changed some char * arguments to const char *.
3103
3104 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3105
3106 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3107 pcretest.
3108
3109
3110 Version 0.96 16-Oct-97
3111 ----------------------
3112
3113 1. Added a simple "pgrep" utility to the distribution.
3114
3115 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3116 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3117 where "ddd" means "one or more decimal digits".
3118
3119 3. Fixed serious bug. If a pattern had a back reference, but the call to
3120 pcre_exec() didn't supply a large enough ovector to record the related
3121 identifying subpattern, the match always failed. PCRE now remembers the number
3122 of the largest back reference, and gets some temporary memory in which to save
3123 the offsets during matching if necessary, in order to ensure that
3124 backreferences always work.
3125
3126 4. Increased the compatibility with Perl in a number of ways:
3127
3128 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3129 to request this handling. The option can be set at compile or exec time.
3130
3131 (b) $ matches before a terminating newline by default; an option
3132 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3133 mode). The option can be set at compile or exec time.
3134
3135 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3136 the same as Perl's. If the decimal number it represents is less than 10
3137 or there aren't that many previous left capturing parentheses, an octal
3138 escape is read. Inside a character class, it's always an octal escape,
3139 even if it is a single digit.
3140
3141 (d) An escaped but undefined alphabetic character is taken as a literal,
3142 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3143 escapes.
3144
3145 (e) {0} is now permitted. (The previous item is removed from the compiled
3146 pattern).
3147
3148 5. Changed all the names of code files so that the basic parts are no longer
3149 than 10 characters, and abolished the teeny "globals.c" file.
3150
3151 6. Changed the handling of character classes; they are now done with a 32-byte
3152 bit map always.
3153
3154 7. Added the -d and /D options to pcretest to make it possible to look at the
3155 internals of compilation without having to recompile pcre.
3156
3157
3158 Version 0.95 23-Sep-97
3159 ----------------------
3160
3161 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3162 \x20 at the start of a run of normal characters. These were being treated as
3163 real characters, instead of the source characters being re-checked.
3164
3165
3166 Version 0.94 18-Sep-97
3167 ----------------------
3168
3169 1. The functions are now thread-safe, with the caveat that the global variables
3170 containing pointers to malloc() and free() or alternative functions are the
3171 same for all threads.
3172
3173 2. Get pcre_study() to generate a bitmap of initial characters for non-
3174 anchored patterns when this is possible, and use it if passed to pcre_exec().
3175
3176
3177 Version 0.93 15-Sep-97
3178 ----------------------
3179
3180 1. /(b)|(:+)/ was computing an incorrect first character.
3181
3182 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3183 but not actually doing anything yet.
3184
3185 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3186 as Perl does (e.g. [-az] or [az-]).
3187
3188 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3189 all possible positions.
3190
3191 5. Split up into different modules to avoid including unneeded functions in a
3192 compiled binary. However, compile and exec are still in one module. The "study"
3193 function is split off.
3194
3195 6. The character tables are now in a separate module whose source is generated
3196 by an auxiliary program - but can then be edited by hand if required. There are
3197 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3198 toupper() in the code.
3199
3200 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3201 make them global. Abolish the function for setting them, as the caller can now
3202 set them directly.
3203
3204
3205 Version 0.92 11-Sep-97
3206 ----------------------
3207
3208 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3209 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3210
3211 2. Caseless matching was not working in character classes if the characters in
3212 the pattern were in upper case.
3213
3214 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3215
3216 4. Make PCRE_ANCHORED public and accept as a compile option.
3217
3218 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3219 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3220 pass them.
3221
3222 6. Give an error if bad option bits passed at compile or run time.
3223
3224 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3225 pcretest to cause it to pass that flag.
3226
3227 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3228 options, and the first character, if set.
3229
3230 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3231
3232
3233 Version 0.91 10-Sep-97
3234 ----------------------
3235
3236 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3237 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3238
3239 2. PCRE was looping on encountering an indefinitely repeated back reference to
3240 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3241 Perl does - treats the match as successful.
3242
3243 ****

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