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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 83 - (show annotations) (download)
Sat Feb 24 21:41:06 2007 UTC (7 years, 4 months ago) by nigel
File size: 83115 byte(s)
Load pcre-6.3 into code/trunk.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 6.3 15-Aug-05
5 ---------------------
6
7 1. The file libpcre.pc.in did not have general read permission in the tarball.
8
9 2. There were some problems when building without C++ support:
10
11 (a) If C++ support was not built, "make install" and "make test" still
12 tried to test it.
13
14 (b) There were problems when the value of CXX was explicitly set. Some
15 changes have been made to try to fix these, and ...
16
17 (c) --disable-cpp can now be used to explicitly disable C++ support.
18
19 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
20 backslash in a target when C++ was disabled. This confuses some
21 versions of "make", apparently. Using an intermediate variable solves
22 this. (Same for CPP_LOBJ.)
23
24 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
25 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
26 necessary on certain architectures.
27
28 4. Added a setting of -export-symbols-regex to the link command to remove
29 those symbols that are exported in the C sense, but actually are local
30 within the library, and not documented. Their names all begin with
31 "_pcre_". This is not a perfect job, because (a) we have to except some
32 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
33 available (and never for static libraries). I have made a note to try to
34 find a way round (a) in the future.
35
36
37 Version 6.2 01-Aug-05
38 ---------------------
39
40 1. There was no test for integer overflow of quantifier values. A construction
41 such as {1111111111111111} would give undefined results. What is worse, if
42 a minimum quantifier for a parenthesized subpattern overflowed and became
43 negative, the calculation of the memory size went wrong. This could have
44 led to memory overwriting.
45
46 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
47
48 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
49 operating environments where this matters.
50
51 4. Applied Giuseppe Maxia's patch to add additional features for controlling
52 PCRE options from within the C++ wrapper.
53
54 5. Named capturing subpatterns were not being correctly counted when a pattern
55 was compiled. This caused two problems: (a) If there were more than 100
56 such subpatterns, the calculation of the memory needed for the whole
57 compiled pattern went wrong, leading to an overflow error. (b) Numerical
58 back references of the form \12, where the number was greater than 9, were
59 not recognized as back references, even though there were sufficient
60 previous subpatterns.
61
62 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
63 versions of gcc, e.g. 2.95.4.
64
65
66 Version 6.1 21-Jun-05
67 ---------------------
68
69 1. There was one reference to the variable "posix" in pcretest.c that was not
70 surrounded by "#if !defined NOPOSIX".
71
72 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
73 the cross-check on the old pcre_info() function, for the benefit of the
74 cut-down version of PCRE that is currently imported into Exim.
75
76 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
77 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
78 patterns that start (?i) but I don't think that matters. The use of (?i) is
79 just an example; this all applies to the other options as well.
80
81 4. Since libtool seems to echo the compile commands it is issuing, the output
82 from "make" can be reduced a bit by putting "@" in front of each libtool
83 compile command.
84
85 5. Patch from the folks at Google for configure.in to be a bit more thorough
86 in checking for a suitable C++ installation before trying to compile the
87 C++ stuff. This should fix a reported problem when a compiler was present,
88 but no suitable headers.
89
90 6. The man pages all had just "PCRE" as their title. I have changed them to
91 be the relevant file name. I have also arranged that these names are
92 retained in the file doc/pcre.txt, which is a concatenation in text format
93 of all the man pages except the little individual ones for each function.
94
95 7. The NON-UNIX-USE file had not been updated for the different set of source
96 files that come with release 6. I also added a few comments about the C++
97 wrapper.
98
99
100 Version 6.0 07-Jun-05
101 ---------------------
102
103 1. Some minor internal re-organization to help with my DFA experiments.
104
105 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
106 didn't matter for the library itself when fully configured, but did matter
107 when compiling without UCP support, or within Exim, where the ucp files are
108 not imported.
109
110 3. Refactoring of the library code to split up the various functions into
111 different source modules. The addition of the new DFA matching code (see
112 below) to a single monolithic source would have made it really too
113 unwieldy, quite apart from causing all the code to be include in a
114 statically linked application, when only some functions are used. This is
115 relevant even without the DFA addition now that patterns can be compiled in
116 one application and matched in another.
117
118 The downside of splitting up is that there have to be some external
119 functions and data tables that are used internally in different modules of
120 the library but which are not part of the API. These have all had their
121 names changed to start with "_pcre_" so that they are unlikely to clash
122 with other external names.
123
124 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
125 a different (DFA) algorithm. Although it is slower than the original
126 function, it does have some advantages for certain types of matching
127 problem.
128
129 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
130 including restarting after a partial match.
131
132 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
133 defined when compiling for Windows was sent to me. I have put it into the
134 code, though I have no means of testing or verifying it.
135
136 7. Added the pcre_refcount() auxiliary function.
137
138 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
139 match before or at the first newline in the subject string. In pcretest,
140 the /f option on a pattern can be used to set this.
141
142 9. A repeated \w when used in UTF-8 mode with characters greater than 256
143 would behave wrongly. This has been present in PCRE since release 4.0.
144
145 10. A number of changes to the pcregrep command:
146
147 (a) Refactored how -x works; insert ^(...)$ instead of setting
148 PCRE_ANCHORED and checking the length, in preparation for adding
149 something similar for -w.
150
151 (b) Added the -w (match as a word) option.
152
153 (c) Refactored the way lines are read and buffered so as to have more
154 than one at a time available.
155
156 (d) Implemented a pcregrep test script.
157
158 (e) Added the -M (multiline match) option. This allows patterns to match
159 over several lines of the subject. The buffering ensures that at least
160 8K, or the rest of the document (whichever is the shorter) is available
161 for matching (and similarly the previous 8K for lookbehind assertions).
162
163 (f) Changed the --help output so that it now says
164
165 -w, --word-regex(p)
166
167 instead of two lines, one with "regex" and the other with "regexp"
168 because that confused at least one person since the short forms are the
169 same. (This required a bit of code, as the output is generated
170 automatically from a table. It wasn't just a text change.)
171
172 (g) -- can be used to terminate pcregrep options if the next thing isn't an
173 option but starts with a hyphen. Could be a pattern or a path name
174 starting with a hyphen, for instance.
175
176 (h) "-" can be given as a file name to represent stdin.
177
178 (i) When file names are being printed, "(standard input)" is used for
179 the standard input, for compatibility with GNU grep. Previously
180 "<stdin>" was used.
181
182 (j) The option --label=xxx can be used to supply a name to be used for
183 stdin when file names are being printed. There is no short form.
184
185 (k) Re-factored the options decoding logic because we are going to add
186 two more options that take data. Such options can now be given in four
187 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
188
189 (l) Added the -A, -B, and -C options for requesting that lines of context
190 around matches be printed.
191
192 (m) Added the -L option to print the names of files that do not contain
193 any matching lines, that is, the complement of -l.
194
195 (n) The return code is 2 if any file cannot be opened, but pcregrep does
196 continue to scan other files.
197
198 (o) The -s option was incorrectly implemented. For compatibility with other
199 greps, it now suppresses the error message for a non-existent or non-
200 accessible file (but not the return code). There is a new option called
201 -q that suppresses the output of matching lines, which was what -s was
202 previously doing.
203
204 (p) Added --include and --exclude options to specify files for inclusion
205 and exclusion when recursing.
206
207 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
208 Hopefully, it now does.
209
210 12. Missing cast in pcre_study().
211
212 13. Added an "uninstall" target to the makefile.
213
214 14. Replaced "extern" in the function prototypes in Makefile.in with
215 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
216 world, but is set differently for Windows.
217
218 15. Added a second compiling function called pcre_compile2(). The only
219 difference is that it has an extra argument, which is a pointer to an
220 integer error code. When there is a compile-time failure, this is set
221 non-zero, in addition to the error test pointer being set to point to an
222 error message. The new argument may be NULL if no error number is required
223 (but then you may as well call pcre_compile(), which is now just a
224 wrapper). This facility is provided because some applications need a
225 numeric error indication, but it has also enabled me to tidy up the way
226 compile-time errors are handled in the POSIX wrapper.
227
228 16. Added VPATH=.libs to the makefile; this should help when building with one
229 prefix path and installing with another. (Or so I'm told by someone who
230 knows more about this stuff than I do.)
231
232 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
233 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
234 match everything, including newlines. This is not POSIX-compatible, but
235 somebody wanted the feature. From pcretest it can be activated by using
236 both the P and the s flags.
237
238 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
239
240 19. libpcre.pc was being incorrectly installed as executable.
241
242 20. A couple of places in pcretest check for end-of-line by looking for '\n';
243 it now also looks for '\r' so that it will work unmodified on Windows.
244
245 21. Added Google's contributed C++ wrapper to the distribution.
246
247 22. Added some untidy missing memory free() calls in pcretest, to keep
248 Electric Fence happy when testing.
249
250
251
252 Version 5.0 13-Sep-04
253 ---------------------
254
255 1. Internal change: literal characters are no longer packed up into items
256 containing multiple characters in a single byte-string. Each character
257 is now matched using a separate opcode. However, there may be more than one
258 byte in the character in UTF-8 mode.
259
260 2. The pcre_callout_block structure has two new fields: pattern_position and
261 next_item_length. These contain the offset in the pattern to the next match
262 item, and its length, respectively.
263
264 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
265 insertion of callouts before each pattern item. Added the /C option to
266 pcretest to make use of this.
267
268 4. On the advice of a Windows user, the lines
269
270 #if defined(_WIN32) || defined(WIN32)
271 _setmode( _fileno( stdout ), 0x8000 );
272 #endif /* defined(_WIN32) || defined(WIN32) */
273
274 have been added to the source of pcretest. This apparently does useful
275 magic in relation to line terminators.
276
277 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
278 for the benefit of those environments where the "b" makes a difference.
279
280 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
281 to know about it. I have put a hack into configure.in that adds in code
282 to set GCC=yes if CC=icc. This seems to end up at a point in the
283 generated configure script that is early enough to affect the setting of
284 compiler options, which is what is needed, but I have no means of testing
285 whether it really works. (The user who reported this had patched the
286 generated configure script, which of course I cannot do.)
287
288 LATER: After change 22 below (new libtool files), the configure script
289 seems to know about icc (and also ecc). Therefore, I have commented out
290 this hack in configure.in.
291
292 7. Added support for pkg-config (2 patches were sent in).
293
294 8. Negated POSIX character classes that used a combination of internal tables
295 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
296 [[:^ascii]]. Typically, they would match almost any characters. The other
297 POSIX classes were not broken in this way.
298
299 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
300 to find the match, as PCRE was deluded into thinking that the match had to
301 start at the start point or following a newline. The same bug applied to
302 patterns with negative forward assertions or any backward assertions
303 preceding ".*" at the start, unless the pattern required a fixed first
304 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
305
306 10. In UTF-8 mode, when moving forwards in the subject after a failed match
307 starting at the last subject character, bytes beyond the end of the subject
308 string were read.
309
310 11. Renamed the variable "class" as "classbits" to make life easier for C++
311 users. (Previously there was a macro definition, but it apparently wasn't
312 enough.)
313
314 12. Added the new field "tables" to the extra data so that tables can be passed
315 in at exec time, or the internal tables can be re-selected. This allows
316 a compiled regex to be saved and re-used at a later time by a different
317 program that might have everything at different addresses.
318
319 13. Modified the pcre-config script so that, when run on Solaris, it shows a
320 -R library as well as a -L library.
321
322 14. The debugging options of pcretest (-d on the command line or D on a
323 pattern) showed incorrect output for anything following an extended class
324 that contained multibyte characters and which was followed by a quantifier.
325
326 15. Added optional support for general category Unicode character properties
327 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
328 support. It adds about 90K to the size of the library. The meanings of the
329 inbuilt class escapes such as \d and \s have NOT been changed.
330
331 16. Updated pcredemo.c to include calls to free() to release the memory for the
332 compiled pattern.
333
334 17. The generated file chartables.c was being created in the source directory
335 instead of in the building directory. This caused the build to fail if the
336 source directory was different from the building directory, and was
337 read-only.
338
339 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
340 file. No doubt somebody will tell me if they don't make sense... Also added
341 Dan Mooney's comments about building on OpenVMS.
342
343 19. Added support for partial matching via the PCRE_PARTIAL option for
344 pcre_exec() and the \P data escape in pcretest.
345
346 20. Extended pcretest with 3 new pattern features:
347
348 (i) A pattern option of the form ">rest-of-line" causes pcretest to
349 write the compiled pattern to the file whose name is "rest-of-line".
350 This is a straight binary dump of the data, with the saved pointer to
351 the character tables forced to be NULL. The study data, if any, is
352 written too. After writing, pcretest reads a new pattern.
353
354 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
355 compiled pattern from the given file. There must not be any
356 occurrences of "<" in the file name (pretty unlikely); if there are,
357 pcretest will instead treat the initial "<" as a pattern delimiter.
358 After reading in the pattern, pcretest goes on to read data lines as
359 usual.
360
361 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
362 and 16-bit fields in a compiled pattern, to simulate a pattern that
363 was compiled on a host of opposite endianness.
364
365 21. The pcre-exec() function can now cope with patterns that were compiled on
366 hosts of opposite endianness, with this restriction:
367
368 As for any compiled expression that is saved and used later, the tables
369 pointer field cannot be preserved; the extra_data field in the arguments
370 to pcre_exec() should be used to pass in a tables address if a value
371 other than the default internal tables were used at compile time.
372
373 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
374 now diagnosed as an error. Previously, most of the time, a negative number
375 would have been treated as zero, but if in addition "ovector" was passed as
376 NULL, a crash could occur.
377
378 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
379 new versions from the libtool 1.5 distribution (the last one is a copy of
380 a file called libtool.m4). This seems to have fixed the need to patch
381 "configure" to support Darwin 1.3 (which I used to do). However, I still
382 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
383 workstation).
384
385 24. Changed the PCRE licence to be the more standard "BSD" licence.
386
387
388 Version 4.5 01-Dec-03
389 ---------------------
390
391 1. There has been some re-arrangement of the code for the match() function so
392 that it can be compiled in a version that does not call itself recursively.
393 Instead, it keeps those local variables that need separate instances for
394 each "recursion" in a frame on the heap, and gets/frees frames whenever it
395 needs to "recurse". Keeping track of where control must go is done by means
396 of setjmp/longjmp. The whole thing is implemented by a set of macros that
397 hide most of the details from the main code, and operates only if
398 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
399 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
400 operating.
401
402 To make it easier for callers to provide specially tailored get/free
403 functions for this usage, two new functions, pcre_stack_malloc, and
404 pcre_stack_free, are used. They are always called in strict stacking order,
405 and the size of block requested is always the same.
406
407 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
408 PCRE has been compiled to use the stack or the heap for recursion. The
409 -C option of pcretest uses this to show which version is compiled.
410
411 A new data escape \S, is added to pcretest; it causes the amounts of store
412 obtained and freed by both kinds of malloc/free at match time to be added
413 to the output.
414
415 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
416 what's available on my current Linux desktop machine.
417
418 3. When matching a UTF-8 string, the test for a valid string at the start has
419 been extended. If start_offset is not zero, PCRE now checks that it points
420 to a byte that is the start of a UTF-8 character. If not, it returns
421 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
422 this is necessary because there may be backward assertions in the pattern.
423 When matching the same subject several times, it may save resources to use
424 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
425
426 4. The code for checking the validity of UTF-8 strings has been tightened so
427 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
428 containing "overlong sequences".
429
430 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
431 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
432 should have been "&", but it just so happened that all the cases this let
433 through by mistake were picked up later in the function.
434
435 6. I had used a variable called "isblank" - this is a C99 function, causing
436 some compilers to warn. To avoid this, I renamed it (as "blankclass").
437
438 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
439 prompting; (b) run "./pcretest /dev/null" at the start of the test script
440 so the version is shown; (c) stop "make test" echoing "./RunTest".
441
442 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
443
444 9. The prototype for memmove() for systems that don't have it was using
445 size_t, but the inclusion of the header that defines size_t was later. I've
446 moved the #includes for the C headers earlier to avoid this.
447
448 10. Added some adjustments to the code to make it easier to compiler on certain
449 special systems:
450
451 (a) Some "const" qualifiers were missing.
452 (b) Added the macro EXPORT before all exported functions; by default this
453 is defined to be empty.
454 (c) Changed the dftables auxiliary program (that builds chartables.c) so
455 that it reads its output file name as an argument instead of writing
456 to the standard output and assuming this can be redirected.
457
458 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
459 class containing characters with values greater than 255, PCRE compilation
460 went into a loop.
461
462 12. A recursive reference to a subpattern that was within another subpattern
463 that had a minimum quantifier of zero caused PCRE to crash. For example,
464 (x(y(?2))z)? provoked this bug with a subject that got as far as the
465 recursion. If the recursively-called subpattern itself had a zero repeat,
466 that was OK.
467
468 13. In pcretest, the buffer for reading a data line was set at 30K, but the
469 buffer into which it was copied (for escape processing) was still set at
470 1024, so long lines caused crashes.
471
472 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
473 "internal error: code overflow...". This applied to any character class
474 that was followed by a possessive quantifier.
475
476 15. Modified the Makefile to add libpcre.la as a prerequisite for
477 libpcreposix.la because I was told this is needed for a parallel build to
478 work.
479
480 16. If a pattern that contained .* following optional items at the start was
481 studied, the wrong optimizing data was generated, leading to matching
482 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
483 matching string must start with a or b or c. The correct conclusion for
484 this pattern is that a match can start with any character.
485
486
487 Version 4.4 13-Aug-03
488 ---------------------
489
490 1. In UTF-8 mode, a character class containing characters with values between
491 127 and 255 was not handled correctly if the compiled pattern was studied.
492 In fixing this, I have also improved the studying algorithm for such
493 classes (slightly).
494
495 2. Three internal functions had redundant arguments passed to them. Removal
496 might give a very teeny performance improvement.
497
498 3. Documentation bug: the value of the capture_top field in a callout is *one
499 more than* the number of the hightest numbered captured substring.
500
501 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
502 in incorrectly linking with a previously installed version. They now link
503 explicitly with libpcre.la.
504
505 5. configure.in no longer needs to recognize Cygwin specially.
506
507 6. A problem in pcre.in for Windows platforms is fixed.
508
509 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
510 pcretest, it used to include the size of the study block as part of its
511 output. Unfortunately, the structure contains a field that has a different
512 size on different hardware architectures. This meant that the tests that
513 showed this size failed. As the block is currently always of a fixed size,
514 this information isn't actually particularly useful in pcretest output, so
515 I have just removed it.
516
517 8. Three pre-processor statements accidentally did not start in column 1.
518 Sadly, there are *still* compilers around that complain, even though
519 standard C has not required this for well over a decade. Sigh.
520
521 9. In pcretest, the code for checking callouts passed small integers in the
522 callout_data field, which is a void * field. However, some picky compilers
523 complained about the casts involved for this on 64-bit systems. Now
524 pcretest passes the address of the small integer instead, which should get
525 rid of the warnings.
526
527 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
528 both compile and run time, and gives an error if an invalid UTF-8 sequence
529 is found. There is a option for disabling this check in cases where the
530 string is known to be correct and/or the maximum performance is wanted.
531
532 11. In response to a bug report, I changed one line in Makefile.in from
533
534 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
535 to
536 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
537
538 to look similar to other lines, but I have no way of telling whether this
539 is the right thing to do, as I do not use Windows. No doubt I'll get told
540 if it's wrong...
541
542
543 Version 4.3 21-May-03
544 ---------------------
545
546 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
547 Makefile.
548
549 2. Some refactoring to improve the quality of the code:
550
551 (i) The utf8_table... variables are now declared "const".
552
553 (ii) The code for \cx, which used the "case flipping" table to upper case
554 lower case letters, now just substracts 32. This is ASCII-specific,
555 but the whole concept of \cx is ASCII-specific, so it seems
556 reasonable.
557
558 (iii) PCRE was using its character types table to recognize decimal and
559 hexadecimal digits in the pattern. This is silly, because it handles
560 only 0-9, a-f, and A-F, but the character types table is locale-
561 specific, which means strange things might happen. A private
562 table is now used for this - though it costs 256 bytes, a table is
563 much faster than multiple explicit tests. Of course, the standard
564 character types table is still used for matching digits in subject
565 strings against \d.
566
567 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
568 ending in _t are). So I've renamed it as ESC_tee.
569
570 3. The first argument for regexec() in the POSIX wrapper should have been
571 defined as "const".
572
573 4. Changed pcretest to use malloc() for its buffers so that they can be
574 Electric Fenced for debugging.
575
576 5. There were several places in the code where, in UTF-8 mode, PCRE would try
577 to read one or more bytes before the start of the subject string. Often this
578 had no effect on PCRE's behaviour, but in some circumstances it could
579 provoke a segmentation fault.
580
581 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
582 to try to read one or more bytes before the start of the subject string.
583
584 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
585 UTF-8 support could misbehave in various ways if the subject string
586 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
587 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
588 back over UTF-8 characters.)
589
590
591 Version 4.2 14-Apr-03
592 ---------------------
593
594 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
595
596 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
597 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
598 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
599 [WIN_PREFIX]: new variable, "cyg" for Cygwin
600 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
601 and BUILD_EXEEXT
602 Note: automatic setting of the BUILD variables is not yet working
603 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
604 compile-time but not at link-time
605 [LINK]: use for linking executables only
606 make different versions for Windows and non-Windows
607 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
608 libraries
609 [LINK_FOR_BUILD]: new variable
610 [OBJEXT]: use throughout
611 [EXEEXT]: use throughout
612 <winshared>: new target
613 <wininstall>: new target
614 <dftables.o>: use native compiler
615 <dftables>: use native linker
616 <install>: handle Windows platform correctly
617 <clean>: ditto
618 <check>: ditto
619 copy DLL to top builddir before testing
620
621 As part of these changes, -no-undefined was removed again. This was reported
622 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
623 in any case.
624
625 3. Some tidies to get rid of compiler warnings:
626
627 . In the match_data structure, match_limit was an unsigned long int, whereas
628 match_call_count was an int. I've made them both unsigned long ints.
629
630 . In pcretest the fact that a const uschar * doesn't automatically cast to
631 a void * provoked a warning.
632
633 . Turning on some more compiler warnings threw up some "shadow" variables
634 and a few more missing casts.
635
636 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
637 option, a class that contained a single character with a value between 128
638 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
639
640 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
641 option, a class that contained several characters, but with at least one
642 whose value was between 128 and 255 caused PCRE to crash.
643
644
645 Version 4.1 12-Mar-03
646 ---------------------
647
648 1. Compiling with gcc -pedantic found a couple of places where casts were
649 needed, and a string in dftables.c that was longer than standard compilers are
650 required to support.
651
652 2. Compiling with Sun's compiler found a few more places where the code could
653 be tidied up in order to avoid warnings.
654
655 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
656 first of these names is deprecated in the latest Autoconf in favour of the name
657 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
658 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
659 analogy I have changed it to CFLAGS_FOR_BUILD.
660
661 4. Added -no-undefined to the linking command in the Makefile, because this is
662 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
663 linking step for the pcreposix library.
664
665 5. PCRE was failing to diagnose the case of two named groups with the same
666 name.
667
668 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
669 literal character that is needed in the subject for a match, and scans along to
670 ensure that it is present before embarking on the full matching process. This
671 saves time in cases of nested unlimited repeats that are never going to match.
672 Problem: the scan can take a lot of time if the subject is very long (e.g.
673 megabytes), thus penalizing straightforward matches. It is now done only if the
674 amount of subject to be scanned is less than 1000 bytes.
675
676 7. A lesser problem with the same optimization is that it was recording the
677 first character of an anchored pattern as "needed", thus provoking a search
678 right along the subject, even when the first match of the pattern was going to
679 fail. The "needed" character is now not set for anchored patterns, unless it
680 follows something in the pattern that is of non-fixed length. Thus, it still
681 fulfils its original purpose of finding quick non-matches in cases of nested
682 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
683
684
685 Version 4.0 17-Feb-03
686 ---------------------
687
688 1. If a comment in an extended regex that started immediately after a meta-item
689 extended to the end of string, PCRE compiled incorrect data. This could lead to
690 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
691
692 2. Moved to autoconf 2.53 and libtool 1.4.2.
693
694 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
695 the special perltest8 script is no longer needed - all the tests can be run
696 from a single perltest script.
697
698 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
699 by \s. It has now been removed in PCRE. This means it isn't recognized as
700 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
701 class [:space:] *does* include VT, thereby creating a mess.
702
703 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
704 space and tab.
705
706 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
707 its new features into the main test script, reducing the number of scripts.
708
709 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
710 were backward compatible, and made the (?i) apply to the whole pattern, as if
711 /i were given. Now it behaves more logically, and applies the option setting
712 only to what follows. PCRE has been changed to follow suit. However, if it
713 finds options settings right at the start of the pattern, it extracts them into
714 the global options, as before. Thus, they show up in the info data.
715
716 8. Added support for the \Q...\E escape sequence. Characters in between are
717 treated as literals. This is slightly different from Perl in that $ and @ are
718 also handled as literals inside the quotes. In Perl, they will cause variable
719 interpolation. Note the following examples:
720
721 Pattern PCRE matches Perl matches
722
723 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
724 \Qabc\$xyz\E abc\$xyz abc\$xyz
725 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
726
727 For compatibility with Perl, \Q...\E sequences are recognized inside character
728 classes as well as outside them.
729
730 9. Re-organized 3 code statements in pcretest to avoid "overflow in
731 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
732 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
733 signed/unsigned warnings.
734
735 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
736 option for pcretest, so I've replaced it by a simple function that does just
737 that job.
738
739 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
740 "pcregrep -".
741
742 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
743 Java package. This provides some syntactic sugar for simple cases of what my
744 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
745 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
746 item, you can use this simplified notation. Note that only makes sense with
747 greedy quantifiers. Consequently, the use of the possessive quantifier forces
748 greediness, whatever the setting of the PCRE_UNGREEDY option.
749
750 13. A change of greediness default within a pattern was not taking effect at
751 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
752 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
753 was abstracted outside.
754
755 14. PCRE now supports the \G assertion. It is true when the current matching
756 position is at the start point of the match. This differs from \A when the
757 starting offset is non-zero. Used with the /g option of pcretest (or similar
758 code), it works in the same way as it does for Perl's /g option. If all
759 alternatives of a regex begin with \G, the expression is anchored to the start
760 match position, and the "anchored" flag is set in the compiled expression.
761
762 15. Some bugs concerning the handling of certain option changes within patterns
763 have been fixed. These applied to options other than (?ims). For example,
764 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
765 been the other way round. Some of this was related to change 7 above.
766
767 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
768 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
769 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
770 POSIX classes only within a class (e.g. /[[:alpha:]]/).
771
772 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
773 mode. Unlike ".", it always matches newline, whatever the setting of
774 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
775 assertions. Perl allows it, but it doesn't (in general) work because it can't
776 calculate the length of the lookbehind. At least, that's the case for Perl
777 5.8.0 - I've been told they are going to document that it doesn't work in
778 future.
779
780 18. Added an error diagnosis for escapes that PCRE does not support: these are
781 \L, \l, \N, \P, \p, \U, \u, and \X.
782
783 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
784 reading past the end of the pattern in cases such as /[abcd/.
785
786 20. PCRE was getting more memory than necessary for patterns with classes that
787 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
788
789 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
790 compiling PCRE for use with Virtual Pascal.
791
792 22. Small fix to the Makefile to make it work properly if the build is done
793 outside the source tree.
794
795 23. Added a new extension: a condition to go with recursion. If a conditional
796 subpattern starts with (?(R) the "true" branch is used if recursion has
797 happened, whereas the "false" branch is used only at the top level.
798
799 24. When there was a very long string of literal characters (over 255 bytes
800 without UTF support, over 250 bytes with UTF support), the computation of how
801 much memory was required could be incorrect, leading to segfaults or other
802 strange effects.
803
804 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
805 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
806 there was a subsequent back reference to those brackets. This meant that, for
807 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
808 possible to check for precisely this case. All we can do is abandon the
809 optimization if .* occurs inside capturing brackets when there are any back
810 references whatsoever. (See below for a better fix that came later.)
811
812 26. The handling of the optimization for finding the first character of a
813 non-anchored pattern, and for finding a character that is required later in the
814 match were failing in some cases. This didn't break the matching; it just
815 failed to optimize when it could. The way this is done has been re-implemented.
816
817 27. Fixed typo in error message for invalid (?R item (it said "(?p").
818
819 28. Added a new feature that provides some of the functionality that Perl
820 provides with (?{...}). The facility is termed a "callout". The way it is done
821 in PCRE is for the caller to provide an optional function, by setting
822 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
823 global variable. By default it is unset, which disables all calling out. To get
824 the function called, the regex must include (?C) at appropriate points. This
825 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
826 This provides a means of identifying different callout points. When PCRE
827 reaches such a point in the regex, if pcre_callout has been set, the external
828 function is called. It is provided with data in a structure called
829 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
830 matching continues; if it returns a non-zero value, the match at the current
831 point fails. However, backtracking will occur if possible. [This was changed
832 later and other features added - see item 49 below.]
833
834 29. pcretest is upgraded to test the callout functionality. It provides a
835 callout function that displays information. By default, it shows the start of
836 the match and the current position in the text. There are some new data escapes
837 to vary what happens:
838
839 \C+ in addition, show current contents of captured substrings
840 \C- do not supply a callout function
841 \C!n return 1 when callout number n is reached
842 \C!n!m return 1 when callout number n is reached for the mth time
843
844 30. If pcregrep was called with the -l option and just a single file name, it
845 output "<stdin>" if a match was found, instead of the file name.
846
847 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
848 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
849 pcre_exec(). This saves a malloc/free per call. The default value of
850 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
851 when configuring.
852
853 32. The default maximum size of a compiled pattern is 64K. There have been a
854 few cases of people hitting this limit. The code now uses macros to handle the
855 storing of links as offsets within the compiled pattern. It defaults to 2-byte
856 links, but this can be changed to 3 or 4 bytes by --with-link-size when
857 configuring. Tests 2 and 5 work only with 2-byte links because they output
858 debugging information about compiled patterns.
859
860 33. Internal code re-arrangements:
861
862 (a) Moved the debugging function for printing out a compiled regex into
863 its own source file (printint.c) and used #include to pull it into
864 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
865 separate copies.
866
867 (b) Defined the list of op-code names for debugging as a macro in
868 internal.h so that it is next to the definition of the opcodes.
869
870 (c) Defined a table of op-code lengths for simpler skipping along compiled
871 code. This is again a macro in internal.h so that it is next to the
872 definition of the opcodes.
873
874 34. Added support for recursive calls to individual subpatterns, along the
875 lines of Robin Houston's patch (but implemented somewhat differently).
876
877 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
878 allow it to read and process whole directories in Win32. This code was
879 contributed by Lionel Fourquaux; it has not been tested by me.
880
881 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
882 used to name a group. Names consist of alphanumerics and underscores, and must
883 be unique. Back references use the syntax (?P=name) and recursive calls use
884 (?P>name) which is a PCRE extension to the Python extension. Groups still have
885 numbers. The function pcre_fullinfo() can be used after compilation to extract
886 a name/number map. There are three relevant calls:
887
888 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
889 PCRE_INFO_NAMECOUNT yields the number of entries
890 PCRE_INFO_NAMETABLE yields a pointer to the map.
891
892 The map is a vector of fixed-size entries. The size of each entry depends on
893 the length of the longest name used. The first two bytes of each entry are the
894 group number, most significant byte first. There follows the corresponding
895 name, zero terminated. The names are in alphabetical order.
896
897 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
898 case instead of 255. Making it the same both with and without UTF-8 support
899 means that the same test output works with both.
900
901 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
902 calling malloc() with a zero argument.
903
904 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
905 optimization. I've improved things by keeping a bitmap of backreferences with
906 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
907 fact referenced, the optimization can be applied. It is unlikely that a
908 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
909 the match to follow \n) will appear inside brackets with a number greater than
910 31, but if it does, any back reference > 31 suppresses the optimization.
911
912 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
913 of disabling numbered capturing parentheses. Any opening parenthesis that is
914 not followed by ? behaves as if it were followed by ?: but named parentheses
915 can still be used for capturing (and they will acquire numbers in the usual
916 way).
917
918 41. Redesigned the return codes from the match() function into yes/no/error so
919 that errors can be passed back from deep inside the nested calls. A malloc
920 failure while inside a recursive subpattern call now causes the
921 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
922
923 42. It is now possible to set a limit on the number of times the match()
924 function is called in a call to pcre_exec(). This facility makes it possible to
925 limit the amount of recursion and backtracking, though not in a directly
926 obvious way, because the match() function is used in a number of different
927 circumstances. The count starts from zero for each position in the subject
928 string (for non-anchored patterns). The default limit is, for compatibility, a
929 large number, namely 10 000 000. You can change this in two ways:
930
931 (a) When configuring PCRE before making, you can use --with-match-limit=n
932 to set a default value for the compiled library.
933
934 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
935 a different value is set. See 45 below.
936
937 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
938
939 43. Added a new function pcre_config(int, void *) to enable run-time extraction
940 of things that can be changed at compile time. The first argument specifies
941 what is wanted and the second points to where the information is to be placed.
942 The current list of available information is:
943
944 PCRE_CONFIG_UTF8
945
946 The output is an integer that is set to one if UTF-8 support is available;
947 otherwise it is set to zero.
948
949 PCRE_CONFIG_NEWLINE
950
951 The output is an integer that it set to the value of the code that is used for
952 newline. It is either LF (10) or CR (13).
953
954 PCRE_CONFIG_LINK_SIZE
955
956 The output is an integer that contains the number of bytes used for internal
957 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
958
959 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
960
961 The output is an integer that contains the threshold above which the POSIX
962 interface uses malloc() for output vectors. See item 31 above.
963
964 PCRE_CONFIG_MATCH_LIMIT
965
966 The output is an unsigned integer that contains the default limit of the number
967 of match() calls in a pcre_exec() execution. See 42 above.
968
969 44. pcretest has been upgraded by the addition of the -C option. This causes it
970 to extract all the available output from the new pcre_config() function, and to
971 output it. The program then exits immediately.
972
973 45. A need has arisen to pass over additional data with calls to pcre_exec() in
974 order to support additional features. One way would have been to define
975 pcre_exec2() (for example) with extra arguments, but this would not have been
976 extensible, and would also have required all calls to the original function to
977 be mapped to the new one. Instead, I have chosen to extend the mechanism that
978 is used for passing in "extra" data from pcre_study().
979
980 The pcre_extra structure is now exposed and defined in pcre.h. It currently
981 contains the following fields:
982
983 flags a bitmap indicating which of the following fields are set
984 study_data opaque data from pcre_study()
985 match_limit a way of specifying a limit on match() calls for a specific
986 call to pcre_exec()
987 callout_data data for callouts (see 49 below)
988
989 The flag bits are also defined in pcre.h, and are
990
991 PCRE_EXTRA_STUDY_DATA
992 PCRE_EXTRA_MATCH_LIMIT
993 PCRE_EXTRA_CALLOUT_DATA
994
995 The pcre_study() function now returns one of these new pcre_extra blocks, with
996 the actual study data pointed to by the study_data field, and the
997 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
998 before. That is, this change is entirely upwards-compatible and requires no
999 change to existing code.
1000
1001 If you want to pass in additional data to pcre_exec(), you can either place it
1002 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
1003 block.
1004
1005 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
1006 data string contains the escape sequence \M, pcretest calls pcre_exec() several
1007 times with different match limits, until it finds the minimum value needed for
1008 pcre_exec() to complete. The value is then output. This can be instructive; for
1009 most simple matches the number is quite small, but for pathological cases it
1010 gets very large very quickly.
1011
1012 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
1013 returns the size of the data block pointed to by the study_data field in a
1014 pcre_extra block, that is, the value that was passed as the argument to
1015 pcre_malloc() when PCRE was getting memory in which to place the information
1016 created by pcre_study(). The fourth argument should point to a size_t variable.
1017 pcretest has been extended so that this information is shown after a successful
1018 pcre_study() call when information about the compiled regex is being displayed.
1019
1020 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
1021 because what follows is always an absolute path. (Later: it turns out that this
1022 is more than cosmetic for MinGW, because it doesn't like empty path
1023 components.)
1024
1025 49. Some changes have been made to the callout feature (see 28 above):
1026
1027 (i) A callout function now has three choices for what it returns:
1028
1029 0 => success, carry on matching
1030 > 0 => failure at this point, but backtrack if possible
1031 < 0 => serious error, return this value from pcre_exec()
1032
1033 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
1034 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
1035 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
1036 use by callout functions. It will never be used by PCRE itself.
1037
1038 (ii) The pcre_extra structure (see 45 above) has a void * field called
1039 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
1040 pcre_callout_block structure has a field of the same name. The contents of
1041 the field passed in the pcre_extra structure are passed to the callout
1042 function in the corresponding field in the callout block. This makes it
1043 easier to use the same callout-containing regex from multiple threads. For
1044 testing, the pcretest program has a new data escape
1045
1046 \C*n pass the number n (may be negative) as callout_data
1047
1048 If the callout function in pcretest receives a non-zero value as
1049 callout_data, it returns that value.
1050
1051 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
1052 there were some redundant $(CFLAGS) in commands that are now specified as
1053 $(LINK), which already includes $(CFLAGS).
1054
1055 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
1056 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
1057 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
1058 one-byte characters throughout. Note that case-insensitive matching applies
1059 only to characters whose values are less than 256. PCRE doesn't support the
1060 notion of cases for higher-valued characters.
1061
1062 (i) A character class whose characters are all within 0-255 is handled as
1063 a bit map, and the map is inverted for negative classes. Previously, a
1064 character > 255 always failed to match such a class; however it should
1065 match if the class was a negative one (e.g. [^ab]). This has been fixed.
1066
1067 (ii) A negated character class with a single character < 255 is coded as
1068 "not this character" (OP_NOT). This wasn't working properly when the test
1069 character was multibyte, either singly or repeated.
1070
1071 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
1072 mode, for example: \x{100}{2,3}.
1073
1074 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
1075 singly or repeated) now correctly test multibyte characters. However,
1076 PCRE doesn't recognize any characters with values greater than 255 as
1077 digits, spaces, or word characters. Such characters always match \D, \S,
1078 and \W, and never match \d, \s, or \w.
1079
1080 (v) Classes may now contain characters and character ranges with values
1081 greater than 255. For example: [ab\x{100}-\x{400}].
1082
1083 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
1084 PCRE in UTF-8 mode.
1085
1086 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
1087 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
1088 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
1089 value.)
1090
1091 53. The single man page has become too large. I have therefore split it up into
1092 a number of separate man pages. These also give rise to individual HTML pages;
1093 these are now put in a separate directory, and there is an index.html page that
1094 lists them all. Some hyperlinking between the pages has been installed.
1095
1096 54. Added convenience functions for handling named capturing parentheses.
1097
1098 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
1099 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
1100 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
1101 are faulted.
1102
1103 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
1104 calling configure. These values are used when compiling the dftables.c program
1105 which is run to generate the source of the default character tables. They
1106 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
1107 you will need to set these values.
1108
1109 57. Updated the building process for Windows DLL, as provided by Fred Cox.
1110
1111
1112 Version 3.9 02-Jan-02
1113 ---------------------
1114
1115 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
1116
1117 2. If --disable-static was given, the building process failed when trying to
1118 build pcretest and pcregrep. (For some reason it was using libtool to compile
1119 them, which is not right, as they aren't part of the library.)
1120
1121
1122 Version 3.8 18-Dec-01
1123 ---------------------
1124
1125 1. The experimental UTF-8 code was completely screwed up. It was packing the
1126 bytes in the wrong order. How dumb can you get?
1127
1128
1129 Version 3.7 29-Oct-01
1130 ---------------------
1131
1132 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
1133 This caused pcretest, when used on the test data, to segfault. Unfortunately,
1134 this didn't happen under Solaris 8, where I normally test things.
1135
1136 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
1137 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
1138 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
1139 this fix an hour or so after the initial 3.7 release.)
1140
1141
1142 Version 3.6 23-Oct-01
1143 ---------------------
1144
1145 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
1146 offsets passed as NULL with zero offset count.
1147
1148 2. The config.guess and config.sub files had not been updated when I moved to
1149 the latest autoconf.
1150
1151
1152 Version 3.5 15-Aug-01
1153 ---------------------
1154
1155 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
1156 had been forgotten.
1157
1158 2. By using declared but undefined structures, we can avoid using "void"
1159 definitions in pcre.h while keeping the internal definitions of the structures
1160 private.
1161
1162 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
1163 user point of view, this means that both static and shared libraries are built
1164 by default, but this can be individually controlled. More of the work of
1165 handling this static/shared cases is now inside libtool instead of PCRE's make
1166 file.
1167
1168 4. The pcretest utility is now installed along with pcregrep because it is
1169 useful for users (to test regexs) and by doing this, it automatically gets
1170 relinked by libtool. The documentation has been turned into a man page, so
1171 there are now .1, .txt, and .html versions in /doc.
1172
1173 5. Upgrades to pcregrep:
1174 (i) Added long-form option names like gnu grep.
1175 (ii) Added --help to list all options with an explanatory phrase.
1176 (iii) Added -r, --recursive to recurse into sub-directories.
1177 (iv) Added -f, --file to read patterns from a file.
1178
1179 6. pcre_exec() was referring to its "code" argument before testing that
1180 argument for NULL (and giving an error if it was NULL).
1181
1182 7. Upgraded Makefile.in to allow for compiling in a different directory from
1183 the source directory.
1184
1185 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
1186 options bits, the pointer it was passed was to an int instead of to an unsigned
1187 long int. This mattered only on 64-bit systems.
1188
1189 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
1190 generated) instead of pcre.in, which it its source. Also made the same change
1191 in several of the .c files.
1192
1193 10. A new release of gcc defines printf() as a macro, which broke pcretest
1194 because it had an ifdef in the middle of a string argument for printf(). Fixed
1195 by using separate calls to printf().
1196
1197 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
1198 script, to force use of CR or LF instead of \n in the source. On non-Unix
1199 systems, the value can be set in config.h.
1200
1201 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
1202 absolute limit. Changed the text of the error message to make this clear, and
1203 likewise updated the man page.
1204
1205 13. The limit of 99 on the number of capturing subpatterns has been removed.
1206 The new limit is 65535, which I hope will not be a "real" limit.
1207
1208
1209 Version 3.4 22-Aug-00
1210 ---------------------
1211
1212 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
1213
1214 2. Diagnose condition (?(0) as an error instead of crashing on matching.
1215
1216
1217 Version 3.3 01-Aug-00
1218 ---------------------
1219
1220 1. If an octal character was given, but the value was greater than \377, it
1221 was not getting masked to the least significant bits, as documented. This could
1222 lead to crashes in some systems.
1223
1224 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
1225 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
1226
1227 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
1228 These just pass their arguments on to (pcre_free)(), but they are provided
1229 because some uses of PCRE bind it to non-C systems that can call its functions,
1230 but cannot call free() or pcre_free() directly.
1231
1232 4. Add "make test" as a synonym for "make check". Corrected some comments in
1233 the Makefile.
1234
1235 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
1236 Makefile.
1237
1238 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
1239 command called pgrep for grepping around the active processes.
1240
1241 7. Added the beginnings of support for UTF-8 character strings.
1242
1243 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
1244 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
1245 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
1246 out for the ar command.)
1247
1248
1249 Version 3.2 12-May-00
1250 ---------------------
1251
1252 This is purely a bug fixing release.
1253
1254 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
1255 of ZA. This was just one example of several cases that could provoke this bug,
1256 which was introduced by change 9 of version 2.00. The code for breaking
1257 infinite loops after an iteration that matches an empty string was't working
1258 correctly.
1259
1260 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
1261 when matched against abbab (for example). After matching an empty string, it
1262 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
1263 caused it to match further down the string than it should.
1264
1265 3. The code contained an inclusion of sys/types.h. It isn't clear why this
1266 was there because it doesn't seem to be needed, and it causes trouble on some
1267 systems, as it is not a Standard C header. It has been removed.
1268
1269 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
1270 were reported on the Macintosh. The changes were from
1271
1272 while ((c = *(++ptr)) != 0 && c != '\n');
1273 to
1274 while ((c = *(++ptr)) != 0 && c != '\n') ;
1275
1276 Totally extraordinary, but if that's what it takes...
1277
1278 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
1279 available. Added HAVE_BCOPY and an autoconf test for it; if neither
1280 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
1281 assumes the way PCRE uses memmove() (always moving upwards).
1282
1283 6. PCRE is being used in one environment where strchr() is not available. There
1284 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
1285 faster code anyway.
1286
1287
1288 Version 3.1 09-Feb-00
1289 ---------------------
1290
1291 The only change in this release is the fixing of some bugs in Makefile.in for
1292 the "install" target:
1293
1294 (1) It was failing to install pcreposix.h.
1295
1296 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
1297
1298
1299 Version 3.0 01-Feb-00
1300 ---------------------
1301
1302 1. Add support for the /+ modifier to perltest (to output $` like it does in
1303 pcretest).
1304
1305 2. Add support for the /g modifier to perltest.
1306
1307 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
1308 matches null strings.
1309
1310 4. Fix perltest so that it doesn't do unwanted things when fed an empty
1311 pattern. Perl treats empty patterns specially - it reuses the most recent
1312 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
1313 effect.
1314
1315 5. The POSIX interface was broken in that it was just handing over the POSIX
1316 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
1317 required a bigger vector, with some working space on the end. This means that
1318 the POSIX wrapper now has to get and free some memory, and copy the results.
1319
1320 6. Added some simple autoconf support, placing the test data and the
1321 documentation in separate directories, re-organizing some of the
1322 information files, and making it build pcre-config (a GNU standard). Also added
1323 libtool support for building PCRE as a shared library, which is now the
1324 default.
1325
1326 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
1327 09 are not valid octal constants. Single digits will be used for minor values
1328 less than 10.
1329
1330 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
1331 existing programs that set these in the POSIX interface can use PCRE without
1332 modification.
1333
1334 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
1335 return all that pcre_info() returns, plus additional data. The pcre_info()
1336 function is retained for compatibility, but is considered to be obsolete.
1337
1338 10. Added experimental recursion feature (?R) to handle one common case that
1339 Perl 5.6 will be able to do with (?p{...}).
1340
1341 11. Added support for POSIX character classes like [:alpha:], which Perl is
1342 adopting.
1343
1344
1345 Version 2.08 31-Aug-99
1346 ----------------------
1347
1348 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
1349 trying to match at the startoffset position, but instead was moving forward to
1350 the next newline as if a previous match had failed.
1351
1352 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
1353 and could get into a loop if a null string was matched other than at the start
1354 of the subject.
1355
1356 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
1357 be distinguished at compile time, and for completeness also added PCRE_DATE.
1358
1359 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
1360 in GnuWin32 environments.
1361
1362
1363 Version 2.07 29-Jul-99
1364 ----------------------
1365
1366 1. The documentation is now supplied in plain text form and HTML as well as in
1367 the form of man page sources.
1368
1369 2. C++ compilers don't like assigning (void *) values to other pointer types.
1370 In particular this affects malloc(). Although there is no problem in Standard
1371 C, I've put in casts to keep C++ compilers happy.
1372
1373 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
1374 should be (const char *).
1375
1376 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
1377 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
1378 However, I haven't made this a standard facility. The documentation doesn't
1379 mention it, and the Makefile doesn't support it.
1380
1381 5. The Makefile now contains an "install" target, with editable destinations at
1382 the top of the file. The pcretest program is not installed.
1383
1384 6. pgrep -V now gives the PCRE version number and date.
1385
1386 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
1387 causing the entire string to be ignored, instead of just the last character.
1388
1389 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
1390 non-matching string, it can take a very, very long time, even for strings of
1391 quite modest length, because of the nested recursion. PCRE now does better in
1392 some of these cases. It does this by remembering the last required literal
1393 character in the pattern, and pre-searching the subject to ensure it is present
1394 before running the real match. In other words, it applies a heuristic to detect
1395 some types of certain failure quickly, and in the above example, if presented
1396 with a string that has no trailing " it gives "no match" very quickly.
1397
1398 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
1399 other alternatives are tried instead.
1400
1401
1402 Version 2.06 09-Jun-99
1403 ----------------------
1404
1405 1. Change pcretest's output for amount of store used to show just the code
1406 space, because the remainder (the data block) varies in size between 32-bit and
1407 64-bit systems.
1408
1409 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
1410 start matching at. This allows lookbehinds to work when searching for multiple
1411 occurrences in a string.
1412
1413 3. Added additional options to pcretest for testing multiple occurrences:
1414
1415 /+ outputs the rest of the string that follows a match
1416 /g loops for multiple occurrences, using the new startoffset argument
1417 /G loops for multiple occurrences by passing an incremented pointer
1418
1419 4. PCRE wasn't doing the "first character" optimization for patterns starting
1420 with \b or \B, though it was doing it for other lookbehind assertions. That is,
1421 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
1422 the letter 'x'. On long subject strings, this gives a significant speed-up.
1423
1424
1425 Version 2.05 21-Apr-99
1426 ----------------------
1427
1428 1. Changed the type of magic_number from int to long int so that it works
1429 properly on 16-bit systems.
1430
1431 2. Fixed a bug which caused patterns starting with .* not to work correctly
1432 when the subject string contained newline characters. PCRE was assuming
1433 anchoring for such patterns in all cases, which is not correct because .* will
1434 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
1435 DOTALL is set at top level; otherwise it knows that patterns starting with .*
1436 must be retried after every newline in the subject.
1437
1438
1439 Version 2.04 18-Feb-99
1440 ----------------------
1441
1442 1. For parenthesized subpatterns with repeats whose minimum was zero, the
1443 computation of the store needed to hold the pattern was incorrect (too large).
1444 If such patterns were nested a few deep, this could multiply and become a real
1445 problem.
1446
1447 2. Added /M option to pcretest to show the memory requirement of a specific
1448 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
1449
1450 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
1451 compiled in such a way that the backtracking after subsequent failure was
1452 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
1453 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
1454
1455
1456 Version 2.03 02-Feb-99
1457 ----------------------
1458
1459 1. Fixed typo and small mistake in man page.
1460
1461 2. Added 4th condition (GPL supersedes if conflict) and created separate
1462 LICENCE file containing the conditions.
1463
1464 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
1465 Perl, that is the internal \ allows the delimiter to be included in the
1466 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
1467 the final delimiter, add \ to the end of the pattern (to test the error).
1468
1469 4. Added the convenience functions for extracting substrings after a successful
1470 match. Updated pcretest to make it able to test these functions.
1471
1472
1473 Version 2.02 14-Jan-99
1474 ----------------------
1475
1476 1. Initialized the working variables associated with each extraction so that
1477 their saving and restoring doesn't refer to uninitialized store.
1478
1479 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
1480 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
1481 fix the problem.
1482
1483 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
1484 calls, and wasn't printing the correct value for compiling calls. Increased the
1485 default value of LOOPREPEAT, and the number of significant figures in the
1486 times.
1487
1488 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
1489
1490 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
1491 a building problem on Windows NT with a FAT file system.
1492
1493
1494 Version 2.01 21-Oct-98
1495 ----------------------
1496
1497 1. Changed the API for pcre_compile() to allow for the provision of a pointer
1498 to character tables built by pcre_maketables() in the current locale. If NULL
1499 is passed, the default tables are used.
1500
1501
1502 Version 2.00 24-Sep-98
1503 ----------------------
1504
1505 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
1506 it any more.
1507
1508 2. Allow quantification of (?>) groups, and make it work correctly.
1509
1510 3. The first character computation wasn't working for (?>) groups.
1511
1512 4. Correct the implementation of \Z (it is permitted to match on the \n at the
1513 end of the subject) and add 5.005's \z, which really does match only at the
1514 very end of the subject.
1515
1516 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
1517
1518 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
1519 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
1520 localized options. All options to pcre_study() were also removed.
1521
1522 7. Add other new features from 5.005:
1523
1524 $(?<= positive lookbehind
1525 $(?<! negative lookbehind
1526 (?imsx-imsx) added the unsetting capability
1527 such a setting is global if at outer level; local otherwise
1528 (?imsx-imsx:) non-capturing groups with option setting
1529 (?(cond)re|re) conditional pattern matching
1530
1531 A backreference to itself in a repeated group matches the previous
1532 captured string.
1533
1534 8. General tidying up of studying (both automatic and via "study")
1535 consequential on the addition of new assertions.
1536
1537 9. As in 5.005, unlimited repeated groups that could match an empty substring
1538 are no longer faulted at compile time. Instead, the loop is forcibly broken at
1539 runtime if any iteration does actually match an empty substring.
1540
1541 10. Include the RunTest script in the distribution.
1542
1543 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
1544 discrepancies, some of which were old and were also with respect to 5.004. They
1545 have now been fixed.
1546
1547
1548 Version 1.09 28-Apr-98
1549 ----------------------
1550
1551 1. A negated single character class followed by a quantifier with a minimum
1552 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
1553 program crashes, or just wrong answers. This did not apply to negated classes
1554 containing more than one character, or to minima other than one.
1555
1556
1557 Version 1.08 27-Mar-98
1558 ----------------------
1559
1560 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
1561
1562 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
1563 latter must appear before anything that relies on it in the pattern.
1564
1565
1566 Version 1.07 16-Feb-98
1567 ----------------------
1568
1569 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
1570 repeat of a potentially empty string).
1571
1572
1573 Version 1.06 23-Jan-98
1574 ----------------------
1575
1576 1. Added Markus Oberhumer's little patches for C++.
1577
1578 2. Literal strings longer than 255 characters were broken.
1579
1580
1581 Version 1.05 23-Dec-97
1582 ----------------------
1583
1584 1. Negated character classes containing more than one character were failing if
1585 PCRE_CASELESS was set at run time.
1586
1587
1588 Version 1.04 19-Dec-97
1589 ----------------------
1590
1591 1. Corrected the man page, where some "const" qualifiers had been omitted.
1592
1593 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
1594 input syntax.
1595
1596 3. Fixed memory leak which occurred when a regex with back references was
1597 matched with an offsets vector that wasn't big enough. The temporary memory
1598 that is used in this case wasn't being freed if the match failed.
1599
1600 4. Tidied pcretest to ensure it frees memory that it gets.
1601
1602 5. Temporary memory was being obtained in the case where the passed offsets
1603 vector was exactly big enough.
1604
1605 6. Corrected definition of offsetof() from change 5 below.
1606
1607 7. I had screwed up change 6 below and broken the rules for the use of
1608 setjmp(). Now fixed.
1609
1610
1611 Version 1.03 18-Dec-97
1612 ----------------------
1613
1614 1. A erroneous regex with a missing opening parenthesis was correctly
1615 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
1616 on some systems.
1617
1618 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
1619 it was reported that one broken compiler failed on the former because "code" is
1620 also an independent variable.
1621
1622 3. The erroneous regex a[]b caused an array overrun reference.
1623
1624 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
1625 fail on data ending with that character. (It was going on too far, and checking
1626 the next character, typically a binary zero.) This was specific to the
1627 optimized code for single-character negative classes.
1628
1629 5. Added a contributed patch from the TIN world which does the following:
1630
1631 + Add an undef for memmove, in case the the system defines a macro for it.
1632
1633 + Add a definition of offsetof(), in case there isn't one. (I don't know
1634 the reason behind this - offsetof() is part of the ANSI standard - but
1635 it does no harm).
1636
1637 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
1638 most of the places where whitespace preceded '#'. I have given up and
1639 allowed the remaining 2 cases to be at the margin.
1640
1641 + Rename some variables in pcre to eliminate shadowing. This seems very
1642 pedantic, but does no harm, of course.
1643
1644 6. Moved the call to setjmp() into its own function, to get rid of warnings
1645 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
1646
1647 7. Constructs such as \d{8,} were compiling into the equivalent of
1648 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
1649 outcome, but in this particular case used more store than had been allocated,
1650 which caused the bug to be discovered because it threw up an internal error.
1651
1652 8. The debugging code in both pcre and pcretest for outputting the compiled
1653 form of a regex was going wrong in the case of back references followed by
1654 curly-bracketed repeats.
1655
1656
1657 Version 1.02 12-Dec-97
1658 ----------------------
1659
1660 1. Typos in pcre.3 and comments in the source fixed.
1661
1662 2. Applied a contributed patch to get rid of places where it used to remove
1663 'const' from variables, and fixed some signed/unsigned and uninitialized
1664 variable warnings.
1665
1666 3. Added the "runtest" target to Makefile.
1667
1668 4. Set default compiler flag to -O2 rather than just -O.
1669
1670
1671 Version 1.01 19-Nov-97
1672 ----------------------
1673
1674 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
1675 like /([ab]*)*/, that is, for classes with more than one character in them.
1676
1677 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
1678 as /((?>a*))*/ (a PCRE_EXTRA facility).
1679
1680
1681 Version 1.00 18-Nov-97
1682 ----------------------
1683
1684 1. Added compile-time macros to support systems such as SunOS4 which don't have
1685 memmove() or strerror() but have other things that can be used instead.
1686
1687 2. Arranged that "make clean" removes the executables.
1688
1689
1690 Version 0.99 27-Oct-97
1691 ----------------------
1692
1693 1. Fixed bug in code for optimizing classes with only one character. It was
1694 initializing a 32-byte map regardless, which could cause it to run off the end
1695 of the memory it had got.
1696
1697 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
1698
1699
1700 Version 0.98 22-Oct-97
1701 ----------------------
1702
1703 1. Fixed bug in code for handling temporary memory usage when there are more
1704 back references than supplied space in the ovector. This could cause segfaults.
1705
1706
1707 Version 0.97 21-Oct-97
1708 ----------------------
1709
1710 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
1711
1712 2. Optimized negated single characters not to use a bit map.
1713
1714 3. Brought error texts together as macro definitions; clarified some of them;
1715 fixed one that was wrong - it said "range out of order" when it meant "invalid
1716 escape sequence".
1717
1718 4. Changed some char * arguments to const char *.
1719
1720 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
1721
1722 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
1723 pcretest.
1724
1725
1726 Version 0.96 16-Oct-97
1727 ----------------------
1728
1729 1. Added a simple "pgrep" utility to the distribution.
1730
1731 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
1732 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
1733 where "ddd" means "one or more decimal digits".
1734
1735 3. Fixed serious bug. If a pattern had a back reference, but the call to
1736 pcre_exec() didn't supply a large enough ovector to record the related
1737 identifying subpattern, the match always failed. PCRE now remembers the number
1738 of the largest back reference, and gets some temporary memory in which to save
1739 the offsets during matching if necessary, in order to ensure that
1740 backreferences always work.
1741
1742 4. Increased the compatibility with Perl in a number of ways:
1743
1744 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
1745 to request this handling. The option can be set at compile or exec time.
1746
1747 (b) $ matches before a terminating newline by default; an option
1748 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
1749 mode). The option can be set at compile or exec time.
1750
1751 (c) The handling of \ followed by a digit other than 0 is now supposed to be
1752 the same as Perl's. If the decimal number it represents is less than 10
1753 or there aren't that many previous left capturing parentheses, an octal
1754 escape is read. Inside a character class, it's always an octal escape,
1755 even if it is a single digit.
1756
1757 (d) An escaped but undefined alphabetic character is taken as a literal,
1758 unless PCRE_EXTRA is set. Currently this just reserves the remaining
1759 escapes.
1760
1761 (e) {0} is now permitted. (The previous item is removed from the compiled
1762 pattern).
1763
1764 5. Changed all the names of code files so that the basic parts are no longer
1765 than 10 characters, and abolished the teeny "globals.c" file.
1766
1767 6. Changed the handling of character classes; they are now done with a 32-byte
1768 bit map always.
1769
1770 7. Added the -d and /D options to pcretest to make it possible to look at the
1771 internals of compilation without having to recompile pcre.
1772
1773
1774 Version 0.95 23-Sep-97
1775 ----------------------
1776
1777 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
1778 \x20 at the start of a run of normal characters. These were being treated as
1779 real characters, instead of the source characters being re-checked.
1780
1781
1782 Version 0.94 18-Sep-97
1783 ----------------------
1784
1785 1. The functions are now thread-safe, with the caveat that the global variables
1786 containing pointers to malloc() and free() or alternative functions are the
1787 same for all threads.
1788
1789 2. Get pcre_study() to generate a bitmap of initial characters for non-
1790 anchored patterns when this is possible, and use it if passed to pcre_exec().
1791
1792
1793 Version 0.93 15-Sep-97
1794 ----------------------
1795
1796 1. /(b)|(:+)/ was computing an incorrect first character.
1797
1798 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
1799 but not actually doing anything yet.
1800
1801 3. Treat "-" characters in classes that cannot be part of ranges as literals,
1802 as Perl does (e.g. [-az] or [az-]).
1803
1804 4. Set the anchored flag if a branch starts with .* or .*? because that tests
1805 all possible positions.
1806
1807 5. Split up into different modules to avoid including unneeded functions in a
1808 compiled binary. However, compile and exec are still in one module. The "study"
1809 function is split off.
1810
1811 6. The character tables are now in a separate module whose source is generated
1812 by an auxiliary program - but can then be edited by hand if required. There are
1813 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
1814 toupper() in the code.
1815
1816 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
1817 make them global. Abolish the function for setting them, as the caller can now
1818 set them directly.
1819
1820
1821 Version 0.92 11-Sep-97
1822 ----------------------
1823
1824 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
1825 (e.g. /a{1,3}/) was broken (I mis-optimized it).
1826
1827 2. Caseless matching was not working in character classes if the characters in
1828 the pattern were in upper case.
1829
1830 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
1831
1832 4. Make PCRE_ANCHORED public and accept as a compile option.
1833
1834 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
1835 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
1836 pass them.
1837
1838 6. Give an error if bad option bits passed at compile or run time.
1839
1840 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
1841 pcretest to cause it to pass that flag.
1842
1843 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
1844 options, and the first character, if set.
1845
1846 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
1847
1848
1849 Version 0.91 10-Sep-97
1850 ----------------------
1851
1852 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
1853 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
1854
1855 2. PCRE was looping on encountering an indefinitely repeated back reference to
1856 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
1857 Perl does - treats the match as successful.
1858
1859 ****

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12