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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 399 - (hide annotations) (download)
Sat Mar 21 12:34:15 2009 UTC (5 years, 4 months ago) by ph10
File size: 163599 byte(s)
Further fix to auto-callout with conditional groups whose condition is an 
assertion.

1 nigel 3 ChangeLog for PCRE
2     ------------------
3    
4 ph10 376 Version 7.9 xx-xxx-09
5     ---------------------
6    
7 ph10 392 1. When building with support for bzlib/zlib (pcregrep) and/or readline
8 ph10 376 (pcretest), all targets were linked against these libraries. This included
9     libpcre, libpcreposix, and libpcrecpp, even though they do not use these
10     libraries. This caused unwanted dependencies to be created. This problem
11 ph10 392 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
12 ph10 376 pcretest is linked with readline.
13 ph10 392
14 ph10 376 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
15     "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
16     moved outside it again, because FALSE and TRUE are already defined in AIX,
17     but BOOL is not.
18 ph10 392
19     3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
20 ph10 376 PCRE_MATCH_LIMIT_RETURSION values as ints, when they should be long ints.
21 ph10 392
22 ph10 376 4. The pcregrep documentation said spaces were inserted as well as colons (or
23     hyphens) following file names and line numbers when outputting matching
24 ph10 392 lines. This is not true; no spaces are inserted. I have also clarified the
25 ph10 376 wording for the --colour (or --color) option.
26 ph10 392
27 ph10 378 5. In pcregrep, when --colour was used with -o, the list of matching strings
28     was not coloured; this is different to GNU grep, so I have changed it to be
29     the same.
30 ph10 392
31 ph10 378 6. When --colo(u)r was used in pcregrep, only the first matching substring in
32 ph10 392 each matching line was coloured. Now it goes on to look for further matches
33     of any of the test patterns, which is the same behaviour as GNU grep.
34    
35     7. A pattern that could match an empty string could cause pcregrep to loop; it
36     doesn't make sense to accept an empty string match in pcregrep, so I have
37 ph10 379 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
38     seems to be how GNU grep behaves.
39 ph10 392
40 ph10 380 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
41 ph10 392 start or after a newline", because the conditional assertion was not being
42     correctly handled. The rule now is that both the assertion and what follows
43     in the first alternative must satisfy the test.
44    
45 ph10 399 9. If auto-callout was enabled in a pattern with a conditional group whose
46     condition was an assertion, PCRE could crash during matching, both with
47     pcre_exec() and pcre_dfa_exec().
48 ph10 392
49     10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
50     used for matching.
51    
52     11. Unicode property support in character classes was not working for
53 ph10 385 characters (bytes) greater than 127 when not in UTF-8 mode.
54 ph10 388
55 ph10 392 12. Added the -M command line option to pcretest.
56    
57 ph10 388 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
58 ph10 389
59     15. Added the PCRE_NO_START_OPTIMIZE match-time option.
60 ph10 390
61 ph10 392 16. Added comments and documentation about mis-use of no_arg in the C++
62 ph10 390 wrapper.
63 ph10 392
64 ph10 391 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
65 ph10 392 from Martin Jerabek that uses macro names for all relevant character and
66 ph10 391 string constants.
67 ph10 393
68     18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
69     SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
70     SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
71     these, but not everybody uses configure.
72 ph10 396
73     19. A conditional group that had only one branch was not being correctly
74     recognized as an item that could match an empty string. This meant that an
75     enclosing group might also not be so recognized, causing infinite looping
76     (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
77     with the subject "ab", where knowledge that the repeated group can match
78     nothing is needed in order to break the loop.
79 ph10 397
80     20. If a pattern that was compiled with callouts was matched using pcre_dfa_
81     exec(), but without supplying a callout function, matching went wrong.
82 ph10 376
83 ph10 392
84 ph10 374 Version 7.8 05-Sep-08
85 ph10 349 ---------------------
86    
87     1. Replaced UCP searching code with optimized version as implemented for Ad
88     Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
89 ph10 351 stage table and inline lookup instead of a function, giving speed ups of 2
90     to 5 times on some simple patterns that I tested. Permission was given to
91     distribute the MultiStage2.py script that generates the tables (it's not in
92     the tarball, but is in the Subversion repository).
93 ph10 349
94 ph10 351 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
95     scripts.
96 ph10 358
97 ph10 353 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
98     a group with a zero qualifier. The result of the study could be incorrect,
99 ph10 358 or the function might crash, depending on the pattern.
100    
101     4. Caseless matching was not working for non-ASCII characters in back
102 ph10 354 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
103 ph10 358 It now works when Unicode Property Support is available.
104    
105 ph10 355 5. In pcretest, an escape such as \x{de} in the data was always generating
106     a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
107     non-UTF-8 mode. If the value is greater than 255, it gives a warning about
108 ph10 358 truncation.
109 ph10 349
110 ph10 358 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
111    
112 ph10 357 7. Added two (int) casts to pcregrep when printing the difference of two
113     pointers, in case they are 64-bit values.
114 ph10 371
115     8. Added comments about Mac OS X stack usage to the pcrestack man page and to
116     test 2 if it fails.
117    
118 ph10 359 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
119 ph10 371 and a #define of that name to empty if it is not externally set. This is to
120     allow users of MSVC to set it if necessary.
121    
122     10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
123 ph10 359 the convenience functions in the pcre_get.c source file.
124 ph10 371
125 ph10 360 11. An option change at the start of a pattern that had top-level alternatives
126     could cause overwriting and/or a crash. This command provoked a crash in
127 ph10 371 some environments:
128    
129     printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
130    
131 ph10 360 This potential security problem was recorded as CVE-2008-2371.
132 ph10 371
133 ph10 361 12. For a pattern where the match had to start at the beginning or immediately
134     after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
135 ph10 371 pcre_dfa_exec() could read past the end of the passed subject if there was
136 ph10 361 no match. To help with detecting such bugs (e.g. with valgrind), I modified
137     pcretest so that it places the subject at the end of its malloc-ed buffer.
138 ph10 371
139 ph10 364 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
140 ph10 371 exec() might read past the end of the data buffer in UTF-8 mode.
141    
142 ph10 364 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
143 ph10 371 the data contained the byte 0x85 as part of a UTF-8 character within its
144     first line. This applied both to normal and DFA matching.
145    
146 ph10 366 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
147 ph10 371 /^[^d]*?$/8 failed to match "abc".
148 ph10 351
149 ph10 371 16. Added a missing copyright notice to pcrecpp_internal.h.
150    
151     17. Make it more clear in the documentation that values returned from
152 ph10 368 pcre_exec() in ovector are byte offsets, not character counts.
153 ph10 357
154 ph10 371 18. Tidied a few places to stop certain compilers from issuing warnings.
155 ph10 368
156 ph10 373 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
157     supplied by Stefan Weber. I made a further small update for 7.8 because
158 ph10 374 there is a change of source arrangements: the pcre_searchfuncs.c module is
159     replaced by pcre_ucd.c.
160 ph10 371
161 ph10 373
162 ph10 347 Version 7.7 07-May-08
163 ph10 321 ---------------------
164    
165     1. Applied Craig's patch to sort out a long long problem: "If we can't convert
166 ph10 345 a string to a long long, pretend we don't even have a long long." This is
167 ph10 321 done by checking for the strtoq, strtoll, and _strtoi64 functions.
168 ph10 345
169 ph10 322 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
170     pre-7.6 versions, which defined a global no_arg variable instead of putting
171 ph10 332 it in the RE class. (See also #8 below.)
172 ph10 345
173     3. Remove a line of dead code, identified by coverity and reported by Nuno
174     Lopes.
175    
176 ph10 324 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
177 ph10 321
178 ph10 324 (1) The include/exclude patterns were being applied to the whole pathnames
179 ph10 345 of files, instead of just to the final components.
180    
181 ph10 324 (2) If there was more than one level of directory, the subdirectories were
182     skipped unless they satisfied the include/exclude conditions. This is
183     inconsistent with GNU grep (and could even be seen as contrary to the
184     pcregrep specification - which I improved to make it absolutely clear).
185     The action now is always to scan all levels of directory, and just
186     apply the include/exclude patterns to regular files.
187 ph10 345
188 ph10 325 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
189 ph10 345 --exclude_dir in the tests to avoid scanning .svn directories.
190    
191 ph10 326 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
192 ph10 345 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
193 ph10 326 doesn't support NULs in patterns.
194 ph10 345
195     7. Added some missing "const"s to declarations of static tables in
196     pcre_compile.c and pcre_dfa_exec.c.
197    
198 ph10 328 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
199 ph10 329 caused by fix #2 above. (Subsequently also a second patch to fix the
200 ph10 345 first patch. And a third patch - this was a messy problem.)
201 ph10 321
202 ph10 345 9. Applied Craig's patch to remove the use of push_back().
203    
204 ph10 332 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
205     matching function regexec().
206 ph10 345
207     11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
208     which, however, unlike Perl's \g{...}, are subroutine calls, not back
209 ph10 333 references. PCRE supports relative numbers with this syntax (I don't think
210     Oniguruma does).
211 ph10 345
212     12. Previously, a group with a zero repeat such as (...){0} was completely
213 ph10 335 omitted from the compiled regex. However, this means that if the group
214     was called as a subroutine from elsewhere in the pattern, things went wrong
215 ph10 345 (an internal error was given). Such groups are now left in the compiled
216     pattern, with a new opcode that causes them to be skipped at execution
217 ph10 335 time.
218 ph10 345
219 ph10 341 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
220     to the way PCRE behaves:
221 ph10 345
222 ph10 336 (a) A lone ] character is dis-allowed (Perl treats it as data).
223 ph10 345
224     (b) A back reference to an unmatched subpattern matches an empty string
225 ph10 336 (Perl fails the current match path).
226 ph10 345
227 ph10 341 (c) A data ] in a character class must be notated as \] because if the
228 ph10 345 first data character in a class is ], it defines an empty class. (In
229 ph10 341 Perl it is not possible to have an empty class.) The empty class []
230     never matches; it forces failure and is equivalent to (*FAIL) or (?!).
231     The negative empty class [^] matches any one character, independently
232     of the DOTALL setting.
233 ph10 345
234     14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
235 ph10 340 non-existent subpattern following a character class starting with ']' and
236     containing () gave an internal compiling error instead of "reference to
237     non-existent subpattern". Fortunately, when the pattern did exist, the
238 ph10 345 compiled code was correct. (When scanning forwards to check for the
239 ph10 340 existencd of the subpattern, it was treating the data ']' as terminating
240 ph10 345 the class, so got the count wrong. When actually compiling, the reference
241 ph10 340 was subsequently set up correctly.)
242 ph10 345
243 ph10 341 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
244 ph10 345 it was being rejected as not supported by pcre_dfa_exec(), even though
245     other assertions are supported. I have made pcre_dfa_exec() support
246     (*FAIL).
247    
248 ph10 342 16. The implementation of 13c above involved the invention of a new opcode,
249 ph10 345 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
250     cannot be changed at match time, I realized I could make a small
251     improvement to matching performance by compiling OP_ALLANY instead of
252     OP_ANY for "." when DOTALL was set, and then removing the runtime tests
253     on the OP_ANY path.
254    
255     17. Compiling pcretest on Windows with readline support failed without the
256 ph10 343 following two fixes: (1) Make the unistd.h include conditional on
257 ph10 345 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
258    
259 ph10 344 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
260 ph10 345 ncurses library to be included for pcretest when ReadLine support is
261     requested, but also to allow for it to be overridden. This patch came from
262     Daniel Bergström.
263 ph10 347
264 ph10 346 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
265 ph10 347 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
266     any errors with the current Unicode tables. Thanks to Peter Kankowski for
267     spotting this.
268 ph10 324
269 ph10 332
270 ph10 319 Version 7.6 28-Jan-08
271 ph10 300 ---------------------
272    
273 ph10 302 1. A character class containing a very large number of characters with
274     codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
275     overflow.
276 ph10 309
277     2. Patch to cut out the "long long" test in pcrecpp_unittest when
278     HAVE_LONG_LONG is not defined.
279    
280 ph10 303 3. Applied Christian Ehrlicher's patch to update the CMake build files to
281 ph10 304 bring them up to date and include new features. This patch includes:
282 ph10 309
283 ph10 304 - Fixed PH's badly added libz and libbz2 support.
284     - Fixed a problem with static linking.
285 ph10 312 - Added pcredemo. [But later removed - see 7 below.]
286 ph10 304 - Fixed dftables problem and added an option.
287     - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
288     HAVE_LONG_LONG.
289     - Added readline support for pcretest.
290 ph10 309 - Added an listing of the option settings after cmake has run.
291    
292 ph10 314 4. A user submitted a patch to Makefile that makes it easy to create
293     "pcre.dll" under mingw when using Configure/Make. I added stuff to
294     Makefile.am that cause it to include this special target, without
295     affecting anything else. Note that the same mingw target plus all
296     the other distribution libraries and programs are now supported
297     when configuring with CMake (see 6 below) instead of with
298     Configure/Make.
299 ph10 309
300 ph10 308 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
301     This is an attempt to solve the reported problem "pcrecpp::no_arg is not
302 ph10 309 exported in the Windows port". It has not yet been confirmed that the patch
303     solves the problem, but it does no harm.
304 ph10 313
305 ph10 311 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
306 ph10 319 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
307 ph10 314 with CMake, and also correct the comment about stack recursion.
308 ph10 313
309 ph10 312 7. Remove the automatic building of pcredemo from the ./configure system and
310 ph10 313 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
311     of a program that users should build themselves after PCRE is installed, so
312     building it automatically is not really right. What is more, it gave
313 ph10 312 trouble in some build environments.
314 ph10 300
315 ph10 319 8. Further tidies to CMakeLists.txt from Sheri and Christian.
316 ph10 308
317 ph10 319
318 ph10 298 Version 7.5 10-Jan-08
319 ph10 263 ---------------------
320    
321     1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
322     values in parens when parsing an RE using the C++ wrapper."
323 ph10 286
324 ph10 264 2. Negative specials like \S did not work in character classes in UTF-8 mode.
325     Characters greater than 255 were excluded from the class instead of being
326     included.
327 ph10 286
328     3. The same bug as (2) above applied to negated POSIX classes such as
329 ph10 265 [:^space:].
330 ph10 286
331 ph10 267 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
332     defined or documented. It seems to have been a typo for PCRE_STATIC, so
333 ph10 286 I have changed it.
334    
335 ph10 268 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
336 ph10 286 first named subpattern) and a construct such as (?&a) would reference the
337     first named subpattern whose name started with "a" (in other words, the
338 ph10 272 length check was missing). Both these problems are fixed. "Subpattern name
339     expected" is now given for (?&) (a zero-length name), and this patch also
340     makes it give the same error for \k'' (previously it complained that that
341     was a reference to a non-existent subpattern).
342 ph10 286
343 ph10 269 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
344 ph10 286 this is right because (?- can be followed by option settings as well as by
345 ph10 269 digits. I have, however, made the messages clearer.
346 ph10 286
347 ph10 270 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
348     than the number used in the conditional) now cause a compile-time error.
349     This is actually not compatible with Perl, which accepts such patterns, but
350     treats the conditional as always being FALSE (as PCRE used to), but it
351     seems to me that giving a diagnostic is better.
352 ph10 286
353 ph10 274 8. Change "alphameric" to the more common word "alphanumeric" in comments
354 ph10 275 and messages.
355 ph10 286
356     9. Fix two occurrences of "backslash" in comments that should have been
357     "backspace".
358    
359     10. Remove two redundant lines of code that can never be obeyed (their function
360     was moved elsewhere).
361    
362     11. The program that makes PCRE's Unicode character property table had a bug
363     which caused it to generate incorrect table entries for sequences of
364 ph10 277 characters that have the same character type, but are in different scripts.
365 ph10 286 It amalgamated them into a single range, with the script of the first of
366 ph10 277 them. In other words, some characters were in the wrong script. There were
367     thirteen such cases, affecting characters in the following ranges:
368 ph10 286
369 ph10 277 U+002b0 - U+002c1
370     U+0060c - U+0060d
371 ph10 286 U+0061e - U+00612
372 ph10 277 U+0064b - U+0065e
373     U+0074d - U+0076d
374     U+01800 - U+01805
375     U+01d00 - U+01d77
376     U+01d9b - U+01dbf
377     U+0200b - U+0200f
378     U+030fc - U+030fe
379     U+03260 - U+0327f
380     U+0fb46 - U+0fbb1
381     U+10450 - U+1049d
382 ph10 286
383 ph10 279 12. The -o option (show only the matching part of a line) for pcregrep was not
384     compatible with GNU grep in that, if there was more than one match in a
385     line, it showed only the first of them. It now behaves in the same way as
386     GNU grep.
387 ph10 286
388     13. If the -o and -v options were combined for pcregrep, it printed a blank
389     line for every non-matching line. GNU grep prints nothing, and pcregrep now
390     does the same. The return code can be used to tell if there were any
391     non-matching lines.
392    
393 ph10 289 14. Added --file-offsets and --line-offsets to pcregrep.
394    
395     15. The pattern (?=something)(?R) was not being diagnosed as a potentially
396 ph10 282 infinitely looping recursion. The bug was that positive lookaheads were not
397 ph10 286 being skipped when checking for a possible empty match (negative lookaheads
398     and both kinds of lookbehind were skipped).
399    
400 ph10 289 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
401 ph10 284 inclusion of <windows.h> to before rather than after the definition of
402     INVALID_FILE_ATTRIBUTES (patch from David Byron).
403 ph10 263
404 ph10 289 17. Specifying a possessive quantifier with a specific limit for a Unicode
405 ph10 286 character property caused pcre_compile() to compile bad code, which led at
406     runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
407     are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
408     caused the error; without that there was no problem.
409 ph10 263
410 ph10 289 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
411 ph10 286
412 ph10 289 19. Added --enable-pcretest-libreadline.
413 ph10 286
414 ph10 289 20. In pcrecpp.cc, the variable 'count' was incremented twice in
415 ph10 288 RE::GlobalReplace(). As a result, the number of replacements returned was
416 ph10 298 double what it should be. I removed one of the increments, but Craig sent a
417     later patch that removed the other one (the right fix) and added unit tests
418     that check the return values (which was not done before).
419 ph10 292
420 ph10 291 21. Several CMake things:
421 ph10 287
422 ph10 291 (1) Arranged that, when cmake is used on Unix, the libraries end up with
423 ph10 292 the names libpcre and libpcreposix, not just pcre and pcreposix.
424 ph10 288
425 ph10 292 (2) The above change means that pcretest and pcregrep are now correctly
426     linked with the newly-built libraries, not previously installed ones.
427 ph10 291
428 ph10 292 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
429 ph10 298
430     22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
431     crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
432     UTF-8 newline character). The key issue is that the pattern starts .*;
433     this means that the match must be either at the beginning, or after a
434     newline. The bug was in the code for advancing after a failed match and
435     checking that the new position followed a newline. It was not taking
436 ph10 294 account of UTF-8 characters correctly.
437 ph10 298
438     23. PCRE was behaving differently from Perl in the way it recognized POSIX
439     character classes. PCRE was not treating the sequence [:...:] as a
440     character class unless the ... were all letters. Perl, however, seems to
441     allow any characters between [: and :], though of course it rejects as
442     unknown any "names" that contain non-letters, because all the known class
443     names consist only of letters. Thus, Perl gives an error for [[:1234:]],
444     for example, whereas PCRE did not - it did not recognize a POSIX character
445     class. This seemed a bit dangerous, so the code has been changed to be
446     closer to Perl. The behaviour is not identical to Perl, because PCRE will
447     diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
448     treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
449     Perl does, and where it didn't before.
450    
451 ph10 296 24. Rewrite so as to remove the single use of %n from pcregrep because in some
452     Windows environments %n is disabled by default.
453 ph10 292
454    
455 ph10 260 Version 7.4 21-Sep-07
456 ph10 230 ---------------------
457    
458     1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
459 ph10 231 means that a class such as [\s] counted as "explicit reference to CR or
460     LF". That isn't really right - the whole point of the change was to try to
461     help when there was an actual mention of one of the two characters. So now
462     the change happens only if \r or \n (or a literal CR or LF) character is
463 ph10 230 encountered.
464 ph10 231
465     2. The 32-bit options word was also used for 6 internal flags, but the numbers
466     of both had grown to the point where there were only 3 bits left.
467     Fortunately, there was spare space in the data structure, and so I have
468     moved the internal flags into a new 16-bit field to free up more option
469 ph10 230 bits.
470 ph10 231
471     3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
472     but did not set the internal JCHANGED flag - either of these is enough to
473     control the way the "get" function works - but the PCRE_INFO_JCHANGED
474     facility is supposed to tell if (?J) was ever used, so now (?J) at the
475 ph10 230 start sets both bits.
476    
477 ph10 231 4. Added options (at build time, compile time, exec time) to change \R from
478     matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
479 ph10 230
480 ph10 243 5. doc/pcresyntax.html was missing from the distribution.
481    
482     6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
483 ph10 233 compatibility, even though it is no longer used.
484 ph10 243
485     7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
486 ph10 254 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
487 ph10 259 windows.h file is present (where different names are used). [This was
488 ph10 257 reversed later after testing - see 16 below.]
489 ph10 243
490     8. Changed all #include <config.h> to #include "config.h". There were also
491 ph10 236 some further <pcre.h> cases that I changed to "pcre.h".
492 ph10 243
493 ph10 239 9. When pcregrep was used with the --colour option, it missed the line ending
494 ph10 243 sequence off the lines that it output.
495    
496     10. It was pointed out to me that arrays of string pointers cause lots of
497     relocations when a shared library is dynamically loaded. A technique of
498     using a single long string with a table of offsets can drastically reduce
499     these. I have refactored PCRE in four places to do this. The result is
500 ph10 240 dramatic:
501 ph10 243
502 ph10 240 Originally: 290
503     After changing UCP table: 187
504 ph10 243 After changing error message table: 43
505 ph10 240 After changing table of "verbs" 36
506     After changing table of Posix names 22
507 ph10 243
508 ph10 240 Thanks to the folks working on Gregex for glib for this insight.
509 ph10 247
510 ph10 244 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
511 ph10 247 unicode-properties was also set.
512    
513 ph10 248 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
514 ph10 231
515 ph10 253 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
516     checked only for CRLF.
517 ph10 233
518 ph10 259 14. Added casts to pcretest.c to avoid compiler warnings.
519    
520 ph10 256 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
521 ph10 253
522 ph10 257 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
523 ph10 259 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
524     entirely. This removes changes made in 7 above.
525 ph10 256
526 ph10 261 17. The CMake files have been updated, and there is now more information about
527     building with CMake in the NON-UNIX-USE document.
528 ph10 257
529 ph10 261
530 ph10 228 Version 7.3 28-Aug-07
531 ph10 157 ---------------------
532    
533 ph10 189 1. In the rejigging of the build system that eventually resulted in 7.1, the
534     line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
535     brackets there is not right, since it causes compilers to look for an
536     installed pcre.h, not the version that is in the source that is being
537     compiled (which of course may be different). I have changed it back to:
538 ph10 197
539 ph10 189 #include "pcre.h"
540 ph10 197
541     I have a vague recollection that the change was concerned with compiling in
542     different directories, but in the new build system, that is taken care of
543     by the VPATH setting the Makefile.
544    
545 ph10 190 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
546     when the subject happened to end in the byte 0x85 (e.g. if the last
547     character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
548     characters but of course it shouldn't be taken as a newline when it is part
549     of another character. The bug was that, for an unlimited repeat of . in
550     not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
551     characters when looking for a newline.
552 ph10 197
553     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
554    
555     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
556     in debug output.
557    
558 ph10 201 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
559     long printing in the pcrecpp unittest when running under MinGW.
560 ph10 189
561 ph10 197 6. ESC_K was left out of the EBCDIC table.
562 ph10 189
563 ph10 197 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
564     parentheses; I made it 1000, which seemed large enough. Unfortunately, the
565     limit also applies to "virtual nesting" when a pattern is recursive, and in
566     this case 1000 isn't so big. I have been able to remove this limit at the
567     expense of backing off one optimization in certain circumstances. Normally,
568     when pcre_exec() would call its internal match() function recursively and
569     immediately return the result unconditionally, it uses a "tail recursion"
570     feature to save stack. However, when a subpattern that can match an empty
571     string has an unlimited repetition quantifier, it no longer makes this
572     optimization. That gives it a stack frame in which to save the data for
573     checking that an empty string has been matched. Previously this was taken
574     from the 1000-entry workspace that had been reserved. So now there is no
575     explicit limit, but more stack is used.
576 ph10 189
577 ph10 200 8. Applied Daniel's patches to solve problems with the import/export magic
578     syntax that is required for Windows, and which was going wrong for the
579     pcreposix and pcrecpp parts of the library. These were overlooked when this
580     problem was solved for the main library.
581 ph10 197
582 ph10 202 9. There were some crude static tests to avoid integer overflow when computing
583     the size of patterns that contain repeated groups with explicit upper
584     limits. As the maximum quantifier is 65535, the maximum group length was
585     set at 30,000 so that the product of these two numbers did not overflow a
586     32-bit integer. However, it turns out that people want to use groups that
587     are longer than 30,000 bytes (though not repeat them that many times).
588     Change 7.0/17 (the refactoring of the way the pattern size is computed) has
589     made it possible to implement the integer overflow checks in a much more
590     dynamic way, which I have now done. The artificial limitation on group
591     length has been removed - we now have only the limit on the total length of
592     the compiled pattern, which depends on the LINK_SIZE setting.
593 ph10 208
594     10. Fixed a bug in the documentation for get/copy named substring when
595     duplicate names are permitted. If none of the named substrings are set, the
596     functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
597     empty string.
598    
599     11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
600     instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
601     because the ] is interpreted as the first data character and the
602     terminating ] is not found. PCRE has been made compatible with Perl in this
603     regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
604     cause memory overwriting.
605    
606 ph10 206 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
607     string has been matched (to stop an infinite loop). It was not recognizing
608 ph10 208 a conditional subpattern that could match an empty string if that
609 ph10 206 subpattern was within another subpattern. For example, it looped when
610 ph10 208 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
611 ph10 206 condition was not nested. This bug has been fixed.
612 ph10 208
613 ph10 207 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
614     past the start of the subject in the presence of bytes with the top bit
615     set, for example "\x8aBCD".
616 ph10 211
617 ph10 210 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
618     (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
619 ph10 200
620 ph10 211 14. Optimized (?!) to (*FAIL).
621 ph10 202
622 ph10 212 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
623     This restricts code points to be within the range 0 to 0x10FFFF, excluding
624     the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
625     full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
626     does: it's just the validity check that is more restrictive.
627 ph10 220
628     16. Inserted checks for integer overflows during escape sequence (backslash)
629     processing, and also fixed erroneous offset values for syntax errors during
630     backslash processing.
631    
632 ph10 214 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
633 ph10 220 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
634    
635 ph10 215 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
636     caused an overrun.
637 ph10 220
638     19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
639     something other than just ASCII characters) inside a group that had an
640     unlimited repeat caused a loop at compile time (while checking to see
641     whether the group could match an empty string).
642    
643 ph10 217 20. Debugging a pattern containing \p or \P could cause a crash. For example,
644     [\P{Any}] did so. (Error in the code for printing property names.)
645 ph10 210
646 ph10 220 21. An orphan \E inside a character class could cause a crash.
647    
648     22. A repeated capturing bracket such as (A)? could cause a wild memory
649 ph10 218 reference during compilation.
650 ph10 220
651     23. There are several functions in pcre_compile() that scan along a compiled
652     expression for various reasons (e.g. to see if it's fixed length for look
653 ph10 218 behind). There were bugs in these functions when a repeated \p or \P was
654 ph10 220 present in the pattern. These operators have additional parameters compared
655     with \d, etc, and these were not being taken into account when moving along
656 ph10 218 the compiled data. Specifically:
657 ph10 220
658     (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
659     length.
660    
661     (b) An item such as \pL+ within a repeated group could cause crashes or
662 ph10 218 loops.
663 ph10 220
664 ph10 218 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
665 ph10 220 "reference to non-existent subpattern" error.
666    
667 ph10 221 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
668    
669 ph10 219 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
670 ph10 220 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
671 ph10 211
672 ph10 222 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
673 ph10 218
674 ph10 227 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
675 ph10 223 character were causing crashes (broken optimization).
676 ph10 227
677 ph10 224 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
678 ph10 227 \p or \P) caused a compile-time loop.
679    
680 ph10 226 28. More problems have arisen in unanchored patterns when CRLF is a valid line
681     break. For example, the unstudied pattern [\r\n]A does not match the string
682     "\r\nA" because change 7.0/46 below moves the current point on by two
683     characters after failing to match at the start. However, the pattern \nA
684 ph10 227 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
685     the same is true. There doesn't seem any very clean way out of this, but
686     what I have chosen to do makes the common cases work: PCRE now takes note
687     of whether there can be an explicit match for \r or \n anywhere in the
688     pattern, and if so, 7.0/46 no longer applies. As part of this change,
689     there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
690     pattern has explicit CR or LF references.
691 ph10 222
692 ph10 227 29. Added (*CR) etc for changing newline setting at start of pattern.
693 ph10 223
694 ph10 227
695 ph10 189 Version 7.2 19-Jun-07
696     ---------------------
697    
698 ph10 157 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
699     which is apparently normally available under Windows.
700 ph10 159
701     2. Re-jig the pcregrep tests with different newline settings in an attempt
702     to make them independent of the local environment's newline setting.
703 ph10 157
704 ph10 160 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
705 ph10 165
706 ph10 161 4. Some of the "internals" tests were previously cut out when the link size
707     was not 2, because the output contained actual offsets. The recent new
708 ph10 165 "Z" feature of pcretest means that these can be cut out, making the tests
709     usable with all link sizes.
710    
711 ph10 164 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
712     stack recursion. This gives a massive performance boost under BSD, but just
713 ph10 165 a small improvement under Linux. However, it saves one field in the frame
714 ph10 164 in all cases.
715 ph10 172
716 ph10 166 6. Added more features from the forthcoming Perl 5.10:
717 ph10 172
718 ph10 166 (a) (?-n) (where n is a string of digits) is a relative subroutine or
719     recursion call. It refers to the nth most recently opened parentheses.
720 ph10 172
721 ph10 166 (b) (?+n) is also a relative subroutine call; it refers to the nth next
722 ph10 172 to be opened parentheses.
723    
724     (c) Conditions that refer to capturing parentheses can be specified
725 ph10 167 relatively, for example, (?(-2)... or (?(+3)...
726 ph10 172
727 ph10 168 (d) \K resets the start of the current match so that everything before
728 ph10 172 is not part of it.
729    
730 ph10 171 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
731 ph10 172
732 ph10 171 (f) \g{name} is another synonym - part of Perl 5.10's unification of
733 ph10 172 reference syntax.
734 ph10 182
735 ph10 175 (g) (?| introduces a group in which the numbering of parentheses in each
736 ph10 182 alternative starts with the same number.
737 ph10 172
738 ph10 182 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
739    
740 ph10 172 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
741     PCRE_INFO_JCHANGED.
742    
743     8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
744     terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
745 ph10 170 for detecting groups that can match an empty string.
746 ph10 159
747 ph10 172 9. A pattern with a very large number of alternatives (more than several
748     hundred) was running out of internal workspace during the pre-compile
749     phase, where pcre_compile() figures out how much memory will be needed. A
750     bit of new cunning has reduced the workspace needed for groups with
751     alternatives. The 1000-alternative test pattern now uses 12 bytes of
752     workspace instead of running out of the 4096 that are available.
753 ph10 182
754 ph10 176 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
755 ph10 172
756 ph10 179 11. Applied patch from Google to remove an optimization that didn't quite work.
757     The report of the bug said:
758 ph10 182
759 ph10 179 pcrecpp::RE("a*").FullMatch("aaa") matches, while
760     pcrecpp::RE("a*?").FullMatch("aaa") does not, and
761 ph10 182 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
762 ph10 185
763 ph10 184 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
764 ph10 185 it matched the wrong number of bytes.
765 ph10 172
766 ph10 179
767 ph10 155 Version 7.1 24-Apr-07
768 ph10 98 ---------------------
769    
770 ph10 111 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
771 ph10 122 that is more "standard", making use of automake and other Autotools. There
772 ph10 99 is some re-arrangement of the files and adjustment of comments consequent
773     on this.
774 ph10 111
775     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
776     for recursive directory scanning broke on some systems because the files
777     are not scanned in any specific order and on different systems the order
778     was different. A call to "sort" has been inserted into RunGrepTest for the
779     approprate test as a short-term fix. In the longer term there may be an
780 ph10 100 alternative.
781 ph10 111
782 ph10 100 3. I had an email from Eric Raymond about problems translating some of PCRE's
783 ph10 111 man pages to HTML (despite the fact that I distribute HTML pages, some
784     people do their own conversions for various reasons). The problems
785     concerned the use of low-level troff macros .br and .in. I have therefore
786     removed all such uses from the man pages (some were redundant, some could
787 ph10 113 be replaced by .nf/.fi pairs). The 132html script that I use to generate
788     HTML has been updated to handle .nf/.fi and to complain if it encounters
789     .br or .in.
790 ph10 111
791 ph10 100 4. Updated comments in configure.ac that get placed in config.h.in and also
792 ph10 123 arranged for config.h to be included in the distribution, with the name
793 ph10 111 config.h.generic, for the benefit of those who have to compile without
794     Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
795    
796     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
797     Weber: (1) pcre_internal.h was missing some function renames; (2) updated
798 ph10 127 makevp.bat for the current PCRE, using the additional files
799 ph10 135 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
800 ph10 111
801     6. A Windows user reported a minor discrepancy with test 2, which turned out
802     to be caused by a trailing space on an input line that had got lost in his
803 ph10 102 copy. The trailing space was an accident, so I've just removed it.
804 ph10 111
805 ph10 104 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
806 ph10 111 that is needed.
807    
808 ph10 105 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
809     as "const" (a) because they are and (b) because it helps the PHP
810     maintainers who have recently made a script to detect big data structures
811 ph10 111 in the php code that should be moved to the .rodata section. I remembered
812     to update Builducptable as well, so it won't revert if ucptable.h is ever
813 ph10 105 re-created.
814 ph10 111
815     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
816 ph10 107 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
817 ph10 111 order to be able to cut out the UTF-8 tables in the latter when UTF-8
818     support is not required. This saves 1.5-2K of code, which is important in
819 ph10 107 some applications.
820 ph10 111
821 ph10 110 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
822 ph10 111 so as not to refer to the tables, even though these functions will never be
823     called when UTF-8 support is disabled. Otherwise there are problems with a
824     shared library.
825    
826 ph10 118 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
827    
828     (a) It was defining its arguments as char * instead of void *.
829    
830 ph10 123 (b) It was assuming that all moves were upwards in memory; this was true
831     a long time ago when I wrote it, but is no longer the case.
832    
833 ph10 118 The emulated memove() is provided for those environments that have neither
834 ph10 123 memmove() nor bcopy(). I didn't think anyone used it these days, but that
835 ph10 118 is clearly not the case, as these two bugs were recently reported.
836 ph10 123
837 ph10 111 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
838 ph10 123 and Detrail to create the HTML documentation, the .txt form of the man
839     pages, and it removes trailing spaces from listed files. It also creates
840     pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
841     case, it wraps all the #defines with #ifndefs. This script should be run
842 ph10 111 before "make dist".
843 ph10 123
844 ph10 115 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
845     with Unicode property support.
846 ph10 123
847     (a) For a maximizing quantifier, if the two different cases of the
848     character were of different lengths in their UTF-8 codings (there are
849     some cases like this - I found 11), and the matching function had to
850 ph10 115 back up over a mixture of the two cases, it incorrectly assumed they
851     were both the same length.
852 ph10 123
853     (b) When PCRE was configured to use the heap rather than the stack for
854     recursion during matching, it was not correctly preserving the data for
855     the other case of a UTF-8 character when checking ahead for a match
856     while processing a minimizing repeat. If the check also involved
857 ph10 115 matching a wide character, but failed, corruption could cause an
858     erroneous result when trying to check for a repeat of the original
859     character.
860 ph10 123
861 ph10 116 13. Some tidying changes to the testing mechanism:
862 ph10 98
863 ph10 116 (a) The RunTest script now detects the internal link size and whether there
864     is UTF-8 and UCP support by running ./pcretest -C instead of relying on
865 ph10 123 values substituted by "configure". (The RunGrepTest script already did
866     this for UTF-8.) The configure.ac script no longer substitutes the
867     relevant variables.
868    
869 ph10 116 (b) The debugging options /B and /D in pcretest show the compiled bytecode
870     with length and offset values. This means that the output is different
871     for different internal link sizes. Test 2 is skipped for link sizes
872     other than 2 because of this, bypassing the problem. Unfortunately,
873     there was also a test in test 3 (the locale tests) that used /B and
874 ph10 123 failed for link sizes other than 2. Rather than cut the whole test out,
875     I have added a new /Z option to pcretest that replaces the length and
876     offset values with spaces. This is now used to make test 3 independent
877 ph10 122 of link size. (Test 2 will be tidied up later.)
878 ph10 123
879     14. If erroroffset was passed as NULL to pcre_compile, it provoked a
880 ph10 122 segmentation fault instead of returning the appropriate error message.
881 ph10 134
882 ph10 131 15. In multiline mode when the newline sequence was set to "any", the pattern
883 ph10 134 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
884     This doesn't seem right; it now treats the CRLF combination as the line
885     ending, and so does not match in that case. It's only a pattern such as ^$
886     that would hit this one: something like ^ABC$ would have failed after \r
887     and then tried again after \r\n.
888    
889 ph10 131 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
890 ph10 134 in an attempt to make files that differ only in their line terminators
891     compare equal. This works on Linux.
892 ph10 142
893 ph10 141 17. Under certain error circumstances pcregrep might try to free random memory
894     as it exited. This is now fixed, thanks to valgrind.
895 ph10 142
896 ph10 141 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
897 ph10 142 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
898     was because its rules for how to advance for /g after matching an empty
899 ph10 143 string at the end of a line did not allow for this case. They now check for
900     it specially.
901 ph10 150
902     20. pcretest is supposed to handle patterns and data of any length, by
903     extending its buffers when necessary. It was getting this wrong when the
904 ph10 147 buffer for a data line had to be extended.
905 ph10 150
906 ph10 149 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
907 ph10 150 CRLF as a newline sequence.
908 ph10 152
909 ph10 151 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
910 ph10 152 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
911     I have nevertheless tidied it up.
912 ph10 111
913 ph10 152 23. Added some casts to kill warnings from HP-UX ia64 compiler.
914 ph10 123
915 ph10 153 24. Added a man page for pcre-config.
916 ph10 152
917 ph10 153
918 nigel 93 Version 7.0 19-Dec-06
919     ---------------------
920    
921     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
922     moving to gcc 4.1.1.
923    
924     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
925     sys/time.h, which is documented as needed for this function. It doesn't
926     seem to matter on Linux, but it showed up on some releases of OS X.
927    
928     3. It seems that there are systems where bytes whose values are greater than
929     127 match isprint() in the "C" locale. The "C" locale should be the
930     default when a C program starts up. In most systems, only ASCII printing
931     characters match isprint(). This difference caused the output from pcretest
932     to vary, making some of the tests fail. I have changed pcretest so that:
933    
934     (a) When it is outputting text in the compiled version of a pattern, bytes
935     other than 32-126 are always shown as hex escapes.
936    
937     (b) When it is outputting text that is a matched part of a subject string,
938     it does the same, unless a different locale has been set for the match
939     (using the /L modifier). In this case, it uses isprint() to decide.
940    
941     4. Fixed a major bug that caused incorrect computation of the amount of memory
942     required for a compiled pattern when options that changed within the
943     pattern affected the logic of the preliminary scan that determines the
944     length. The relevant options are -x, and -i in UTF-8 mode. The result was
945     that the computed length was too small. The symptoms of this bug were
946     either the PCRE error "internal error: code overflow" from pcre_compile(),
947     or a glibc crash with a message such as "pcretest: free(): invalid next
948     size (fast)". Examples of patterns that provoked this bug (shown in
949     pcretest format) are:
950    
951     /(?-x: )/x
952     /(?x)(?-x: \s*#\s*)/
953     /((?i)[\x{c0}])/8
954     /(?i:[\x{c0}])/8
955    
956     HOWEVER: Change 17 below makes this fix obsolete as the memory computation
957     is now done differently.
958    
959     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
960     wrapper classes; (b) implement a new function in the C++ scanner that is
961     more efficient than the old way of doing things because it avoids levels of
962     recursion in the regex matching; (c) add a paragraph to the documentation
963     for the FullMatch() function.
964    
965     6. The escape sequence \n was being treated as whatever was defined as
966     "newline". Not only was this contrary to the documentation, which states
967     that \n is character 10 (hex 0A), but it also went horribly wrong when
968     "newline" was defined as CRLF. This has been fixed.
969    
970     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
971     was being set to -1 for the "end of line" case (supposedly a value that no
972     character can have). Though this value is never used (the check for end of
973     line is "zero bytes in current character"), it caused compiler complaints.
974     I've changed it to 0xffffffff.
975    
976     8. In pcre_version.c, the version string was being built by a sequence of
977     C macros that, in the event of PCRE_PRERELEASE being defined as an empty
978     string (as it is for production releases) called a macro with an empty
979     argument. The C standard says the result of this is undefined. The gcc
980     compiler treats it as an empty string (which was what was wanted) but it is
981     reported that Visual C gives an error. The source has been hacked around to
982     avoid this problem.
983    
984     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
985     builds of pcretest, and changed the call to _setmode() to use _O_BINARY
986     instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
987     of them did).
988    
989     10. Originally, pcretest opened its input and output without "b"; then I was
990     told that "b" was needed in some environments, so it was added for release
991     5.0 to both the input and output. (It makes no difference on Unix-like
992     systems.) Later I was told that it is wrong for the input on Windows. I've
993     now abstracted the modes into two macros, to make it easier to fiddle with
994     them, and removed "b" from the input mode under Windows.
995    
996     11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
997    
998     12. Added -help and --help to pcretest as an official way of being reminded
999     of the options.
1000    
1001     13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1002     and pcrecpp.cc because they annoy compilers at high warning levels.
1003    
1004     14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1005    
1006     15. Fixed an occurrence of == in configure.ac that should have been = (shell
1007     scripts are not C programs :-) and which was not noticed because it works
1008     on Linux.
1009    
1010     16. pcretest is supposed to handle any length of pattern and data line (as one
1011     line or as a continued sequence of lines) by extending its input buffer if
1012     necessary. This feature was broken for very long pattern lines, leading to
1013     a string of junk being passed to pcre_compile() if the pattern was longer
1014     than about 50K.
1015    
1016     17. I have done a major re-factoring of the way pcre_compile() computes the
1017     amount of memory needed for a compiled pattern. Previously, there was code
1018     that made a preliminary scan of the pattern in order to do this. That was
1019     OK when PCRE was new, but as the facilities have expanded, it has become
1020     harder and harder to keep it in step with the real compile phase, and there
1021     have been a number of bugs (see for example, 4 above). I have now found a
1022     cunning way of running the real compile function in a "fake" mode that
1023     enables it to compute how much memory it would need, while actually only
1024     ever using a few hundred bytes of working memory and without too many
1025     tests of the mode. This should make future maintenance and development
1026     easier. A side effect of this work is that the limit of 200 on the nesting
1027     depth of parentheses has been removed (though this was never a serious
1028     limitation, I suspect). However, there is a downside: pcre_compile() now
1029     runs more slowly than before (30% or more, depending on the pattern). I
1030     hope this isn't a big issue. There is no effect on runtime performance.
1031    
1032     18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1033     newline (only possible for the last line of a file) and it was a
1034     pattern that set a locale (followed by /Lsomething), pcretest crashed.
1035    
1036     19. Added additional timing features to pcretest. (1) The -tm option now times
1037     matching only, not compiling. (2) Both -t and -tm can be followed, as a
1038     separate command line item, by a number that specifies the number of
1039     repeats to use when timing. The default is 50000; this gives better
1040     precision, but takes uncomfortably long for very large patterns.
1041    
1042     20. Extended pcre_study() to be more clever in cases where a branch of a
1043     subpattern has no definite first character. For example, (a*|b*)[cd] would
1044     previously give no result from pcre_study(). Now it recognizes that the
1045     first character must be a, b, c, or d.
1046    
1047     21. There was an incorrect error "recursive call could loop indefinitely" if
1048     a subpattern (or the entire pattern) that was being tested for matching an
1049     empty string contained only one non-empty item after a nested subpattern.
1050     For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1051     incorrectly, because the \d was being skipped in the check.
1052    
1053     22. The pcretest program now has a new pattern option /B and a command line
1054     option -b, which is equivalent to adding /B to every pattern. This causes
1055     it to show the compiled bytecode, without the additional information that
1056     -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1057     is the same as /B/I).
1058    
1059     23. A new optimization is now able automatically to treat some sequences such
1060     as a*b as a*+b. More specifically, if something simple (such as a character
1061     or a simple class like \d) has an unlimited quantifier, and is followed by
1062     something that cannot possibly match the quantified thing, the quantifier
1063     is automatically "possessified".
1064    
1065     24. A recursive reference to a subpattern whose number was greater than 39
1066     went wrong under certain circumstances in UTF-8 mode. This bug could also
1067     have affected the operation of pcre_study().
1068    
1069     25. Realized that a little bit of performance could be had by replacing
1070     (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1071    
1072     26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1073    
1074     27. Possessive quantifiers such as a++ were previously implemented by turning
1075     them into atomic groups such as ($>a+). Now they have their own opcodes,
1076     which improves performance. This includes the automatically created ones
1077     from 23 above.
1078    
1079     28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1080     lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1081     the first matched character to be a colon. This applied both to named and
1082     numbered groups.
1083    
1084     29. The ucpinternal.h header file was missing its idempotency #ifdef.
1085    
1086     30. I was sent a "project" file called libpcre.a.dev which I understand makes
1087     building PCRE on Windows easier, so I have included it in the distribution.
1088    
1089     31. There is now a check in pcretest against a ridiculously large number being
1090     returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1091     loop, the loop is abandoned.
1092    
1093     32. Forward references to subpatterns in conditions such as (?(2)...) where
1094     subpattern 2 is defined later cause pcre_compile() to search forwards in
1095     the pattern for the relevant set of parentheses. This search went wrong
1096     when there were unescaped parentheses in a character class, parentheses
1097     escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1098    
1099     33. "Subroutine" calls and backreferences were previously restricted to
1100     referencing subpatterns earlier in the regex. This restriction has now
1101     been removed.
1102    
1103     34. Added a number of extra features that are going to be in Perl 5.10. On the
1104     whole, these are just syntactic alternatives for features that PCRE had
1105     previously implemented using the Python syntax or my own invention. The
1106     other formats are all retained for compatibility.
1107    
1108     (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1109     as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1110     also .NET compatible.
1111    
1112     (b) A recursion or subroutine call to a named group can now be defined as
1113     (?&name) as well as (?P>name).
1114    
1115     (c) A backreference to a named group can now be defined as \k<name> or
1116     \k'name' as well as (?P=name). The new forms, as well as being in Perl
1117     5.10, are also .NET compatible.
1118    
1119     (d) A conditional reference to a named group can now use the syntax
1120     (?(<name>) or (?('name') as well as (?(name).
1121    
1122     (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1123     groups (named and numbered) that are never evaluated inline, but can be
1124     called as "subroutines" from elsewhere. In effect, the DEFINE condition
1125     is always false. There may be only one alternative in such a group.
1126    
1127     (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1128     as the simple (?(R). The condition is true only if the most recent
1129     recursion is that of the given number or name. It does not search out
1130     through the entire recursion stack.
1131    
1132     (g) The escape \gN or \g{N} has been added, where N is a positive or
1133     negative number, specifying an absolute or relative reference.
1134    
1135     35. Tidied to get rid of some further signed/unsigned compiler warnings and
1136     some "unreachable code" warnings.
1137    
1138     36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1139     things, this adds five new scripts.
1140    
1141     37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1142     There were also incompatibilities regarding the handling of \Q..\E inside
1143     character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1144     hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1145    
1146     38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1147     matches an empty string, and forcibly breaks the loop. There were bugs in
1148     this code in non-simple cases. For a pattern such as ^(a()*)* matched
1149     against aaaa the result was just "a" rather than "aaaa", for example. Two
1150     separate and independent bugs (that affected different cases) have been
1151     fixed.
1152    
1153     39. Refactored the code to abolish the use of different opcodes for small
1154     capturing bracket numbers. This is a tidy that I avoided doing when I
1155     removed the limit on the number of capturing brackets for 3.5 back in 2001.
1156     The new approach is not only tidier, it makes it possible to reduce the
1157     memory needed to fix the previous bug (38).
1158    
1159     40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1160     sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1161     processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1162     mode.
1163    
1164     41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1165     report.
1166    
1167     42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1168     copy construction and assignment in the C++ wrapper.
1169    
1170     43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1171     couple of bugs that could have given wrong results in the "--newline=crlf"
1172     case.
1173    
1174     44. Added a number of casts and did some reorganization of signed/unsigned int
1175     variables following suggestions from Dair Grant. Also renamed the variable
1176     "this" as "item" because it is a C++ keyword.
1177    
1178     45. Arranged for dftables to add
1179    
1180     #include "pcre_internal.h"
1181    
1182     to pcre_chartables.c because without it, gcc 4.x may remove the array
1183     definition from the final binary if PCRE is built into a static library and
1184     dead code stripping is activated.
1185    
1186     46. For an unanchored pattern, if a match attempt fails at the start of a
1187     newline sequence, and the newline setting is CRLF or ANY, and the next two
1188     characters are CRLF, advance by two characters instead of one.
1189    
1190    
1191 nigel 91 Version 6.7 04-Jul-06
1192     ---------------------
1193    
1194     1. In order to handle tests when input lines are enormously long, pcretest has
1195     been re-factored so that it automatically extends its buffers when
1196     necessary. The code is crude, but this _is_ just a test program. The
1197     default size has been increased from 32K to 50K.
1198    
1199     2. The code in pcre_study() was using the value of the re argument before
1200     testing it for NULL. (Of course, in any sensible call of the function, it
1201     won't be NULL.)
1202    
1203     3. The memmove() emulation function in pcre_internal.h, which is used on
1204     systems that lack both memmove() and bcopy() - that is, hardly ever -
1205     was missing a "static" storage class specifier.
1206    
1207     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1208     containing an extended class (one that cannot be represented by a bitmap
1209     because it contains high-valued characters or Unicode property items, e.g.
1210     [\pZ]). Almost always one would set UTF-8 mode when processing such a
1211     pattern, but PCRE should not loop if you do not (it no longer does).
1212     [Detail: two cases were found: (a) a repeated subpattern containing an
1213     extended class; (b) a recursive reference to a subpattern that followed a
1214     previous extended class. It wasn't skipping over the extended class
1215     correctly when UTF-8 mode was not set.]
1216    
1217     5. A negated single-character class was not being recognized as fixed-length
1218     in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1219     compile error "lookbehind assertion is not fixed length".
1220    
1221     6. The RunPerlTest auxiliary script was showing an unexpected difference
1222     between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1223     write a Perl script that can interpret lines of an input file either as
1224     byte characters or as UTF-8, which is what "perltest" was being required to
1225     do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1226     can't do is switch easily at run time between having the "use utf8;" pragma
1227     or not. In the end, I fudged it by using the RunPerlTest script to insert
1228     "use utf8;" explicitly for the UTF-8 tests.
1229    
1230     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1231     the end of the subject string, contrary to the documentation and to what
1232     Perl does. This was true of both matching functions. Now it matches only at
1233     the start of the subject and immediately after *internal* newlines.
1234    
1235     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1236     a pointer to an int instead of a pointer to an unsigned long int. This
1237     caused problems on 64-bit systems.
1238    
1239     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1240     instance of the 'standard' template library not being so standard".
1241    
1242     10. There was no check on the number of named subpatterns nor the maximum
1243     length of a subpattern name. The product of these values is used to compute
1244     the size of the memory block for a compiled pattern. By supplying a very
1245     long subpattern name and a large number of named subpatterns, the size
1246     computation could be caused to overflow. This is now prevented by limiting
1247     the length of names to 32 characters, and the number of named subpatterns
1248     to 10,000.
1249    
1250     11. Subpatterns that are repeated with specific counts have to be replicated in
1251     the compiled pattern. The size of memory for this was computed from the
1252     length of the subpattern and the repeat count. The latter is limited to
1253     65535, but there was no limit on the former, meaning that integer overflow
1254     could in principle occur. The compiled length of a repeated subpattern is
1255     now limited to 30,000 bytes in order to prevent this.
1256    
1257     12. Added the optional facility to have named substrings with the same name.
1258    
1259     13. Added the ability to use a named substring as a condition, using the
1260     Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1261     are numbers (not recommended). Forward references are permitted.
1262    
1263     14. Added forward references in named backreferences (if you see what I mean).
1264    
1265     15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1266     pattern could run off the end of the subject. For example, the pattern
1267     "(?s)(.{1,5})"8 did this with the subject "ab".
1268    
1269     16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1270     PCRE_CASELESS was set when matching characters that were quantified with ?
1271     or *.
1272    
1273     17. A character class other than a single negated character that had a minimum
1274     but no maximum quantifier - for example [ab]{6,} - was not handled
1275     correctly by pce_dfa_exec(). It would match only one character.
1276    
1277     18. A valid (though odd) pattern that looked like a POSIX character
1278     class but used an invalid character after [ (for example [[,abc,]]) caused
1279     pcre_compile() to give the error "Failed: internal error: code overflow" or
1280     in some cases to crash with a glibc free() error. This could even happen if
1281     the pattern terminated after [[ but there just happened to be a sequence of
1282     letters, a binary zero, and a closing ] in the memory that followed.
1283    
1284     19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1285     over the years. Originally (before any Unicode support), just the bottom 8
1286     bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1287     output from "man perlunicode" includes this:
1288    
1289     The regular expression compiler produces polymorphic opcodes. That
1290     is, the pattern adapts to the data and automatically switches to
1291     the Unicode character scheme when presented with Unicode data--or
1292     instead uses a traditional byte scheme when presented with byte
1293     data.
1294    
1295     Sadly, a wide octal escape does not cause a switch, and in a string with
1296     no other multibyte characters, these octal escapes are treated as before.
1297     Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1298     /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1299     Unicode string.
1300    
1301     I have not perpetrated such confusion in PCRE. Up till now, it took just
1302     the bottom 8 bits, as in old Perl. I have now made octal escapes with
1303     values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1304     translate to the appropriate multibyte character.
1305    
1306     29. Applied some refactoring to reduce the number of warnings from Microsoft
1307     and Borland compilers. This has included removing the fudge introduced
1308     seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1309     a warning about an unused variable.
1310    
1311     21. PCRE has not included VT (character 0x0b) in the set of whitespace
1312     characters since release 4.0, because Perl (from release 5.004) does not.
1313     [Or at least, is documented not to: some releases seem to be in conflict
1314     with the documentation.] However, when a pattern was studied with
1315     pcre_study() and all its branches started with \s, PCRE still included VT
1316     as a possible starting character. Of course, this did no harm; it just
1317     caused an unnecessary match attempt.
1318    
1319     22. Removed a now-redundant internal flag bit that recorded the fact that case
1320     dependency changed within the pattern. This was once needed for "required
1321     byte" processing, but is no longer used. This recovers a now-scarce options
1322     bit. Also moved the least significant internal flag bit to the most-
1323     significant bit of the word, which was not previously used (hangover from
1324     the days when it was an int rather than a uint) to free up another bit for
1325     the future.
1326    
1327     23. Added support for CRLF line endings as well as CR and LF. As well as the
1328     default being selectable at build time, it can now be changed at runtime
1329     via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1330     specify that it is scanning data with non-default line endings.
1331    
1332     24. Changed the definition of CXXLINK to make it agree with the definition of
1333     LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1334    
1335     25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1336     recursions. This makes a big different to stack usage for some patterns.
1337    
1338     26. If a subpattern containing a named recursion or subroutine reference such
1339     as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1340     the space required for the compiled pattern went wrong and gave too small a
1341     value. Depending on the environment, this could lead to "Failed: internal
1342     error: code overflow at offset 49" or "glibc detected double free or
1343     corruption" errors.
1344    
1345     27. Applied patches from Google (a) to support the new newline modes and (b) to
1346     advance over multibyte UTF-8 characters in GlobalReplace.
1347    
1348     28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1349     difference for some implementation of PCRE in some Windows version.
1350    
1351     29. Added some extra testing facilities to pcretest:
1352    
1353     \q<number> in a data line sets the "match limit" value
1354     \Q<number> in a data line sets the "match recursion limt" value
1355     -S <number> sets the stack size, where <number> is in megabytes
1356    
1357     The -S option isn't available for Windows.
1358    
1359    
1360 nigel 89 Version 6.6 06-Feb-06
1361     ---------------------
1362    
1363     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1364     in pcreposix.h. I have copied the definition from pcre.h.
1365    
1366     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1367     because pcre.h is no longer a built file.
1368    
1369     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1370     not normally included in the compiled code.
1371    
1372    
1373 nigel 87 Version 6.5 01-Feb-06
1374     ---------------------
1375    
1376     1. When using the partial match feature with pcre_dfa_exec(), it was not
1377     anchoring the second and subsequent partial matches at the new starting
1378     point. This could lead to incorrect results. For example, with the pattern
1379     /1234/, partially matching against "123" and then "a4" gave a match.
1380    
1381     2. Changes to pcregrep:
1382    
1383     (a) All non-match returns from pcre_exec() were being treated as failures
1384     to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1385     error message is output. Some extra information is given for the
1386     PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1387     probably the only errors that are likely to be caused by users (by
1388     specifying a regex that has nested indefinite repeats, for instance).
1389     If there are more than 20 of these errors, pcregrep is abandoned.
1390    
1391     (b) A binary zero was treated as data while matching, but terminated the
1392     output line if it was written out. This has been fixed: binary zeroes
1393     are now no different to any other data bytes.
1394    
1395     (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1396     used to set a locale for matching. The --locale=xxxx long option has
1397     been added (no short equivalent) to specify a locale explicitly on the
1398     pcregrep command, overriding the environment variables.
1399    
1400     (d) When -B was used with -n, some line numbers in the output were one less
1401     than they should have been.
1402    
1403     (e) Added the -o (--only-matching) option.
1404    
1405     (f) If -A or -C was used with -c (count only), some lines of context were
1406     accidentally printed for the final match.
1407    
1408     (g) Added the -H (--with-filename) option.
1409    
1410     (h) The combination of options -rh failed to suppress file names for files
1411     that were found from directory arguments.
1412    
1413     (i) Added the -D (--devices) and -d (--directories) options.
1414    
1415     (j) Added the -F (--fixed-strings) option.
1416    
1417     (k) Allow "-" to be used as a file name for -f as well as for a data file.
1418    
1419     (l) Added the --colo(u)r option.
1420    
1421     (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1422     is not present by default.
1423    
1424     3. A nasty bug was discovered in the handling of recursive patterns, that is,
1425     items such as (?R) or (?1), when the recursion could match a number of
1426     alternatives. If it matched one of the alternatives, but subsequently,
1427     outside the recursion, there was a failure, the code tried to back up into
1428     the recursion. However, because of the way PCRE is implemented, this is not
1429     possible, and the result was an incorrect result from the match.
1430    
1431     In order to prevent this happening, the specification of recursion has
1432     been changed so that all such subpatterns are automatically treated as
1433     atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1434    
1435     4. I had overlooked the fact that, in some locales, there are characters for
1436     which isalpha() is true but neither isupper() nor islower() are true. In
1437     the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1438     and ordfeminine) are like this. This affected the treatment of \w and \W
1439     when they appeared in character classes, but not when they appeared outside
1440     a character class. The bit map for "word" characters is now created
1441     separately from the results of isalnum() instead of just taking it from the
1442     upper, lower, and digit maps. (Plus the underscore character, of course.)
1443    
1444     5. The above bug also affected the handling of POSIX character classes such as
1445     [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1446     permanent tables. Instead, the bit maps for such a class were previously
1447     created as the appropriate unions of the upper, lower, and digit bitmaps.
1448     Now they are created by subtraction from the [[:word:]] class, which has
1449     its own bitmap.
1450    
1451     6. The [[:blank:]] character class matches horizontal, but not vertical space.
1452     It is created by subtracting the vertical space characters (\x09, \x0a,
1453     \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1454     subtraction was done in the overall bitmap for a character class, meaning
1455     that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1456     be recognized. This bug has been fixed.
1457    
1458     7. Patches from the folks at Google:
1459    
1460     (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1461     real life, but is still worth protecting against".
1462    
1463     (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1464     regular expressions".
1465    
1466     (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1467     have it.
1468    
1469     (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1470     "configure" and the latter not, in order to fix a problem somebody had
1471     with compiling the Arg class on HP-UX.
1472    
1473     (e) Improve the error-handling of the C++ wrapper a little bit.
1474    
1475     (f) New tests for checking recursion limiting.
1476    
1477     8. The pcre_memmove() function, which is used only if the environment does not
1478     have a standard memmove() function (and is therefore rarely compiled),
1479     contained two bugs: (a) use of int instead of size_t, and (b) it was not
1480     returning a result (though PCRE never actually uses the result).
1481    
1482     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1483     large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1484     returned instead of calling malloc() with an overflowing number that would
1485     most likely cause subsequent chaos.
1486    
1487     10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1488    
1489     11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1490     with this option is matched, the nmatch and pmatch options of regexec() are
1491     ignored.
1492    
1493     12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1494     provided in case anyone wants to the the POSIX interface with UTF-8
1495     strings.
1496    
1497     13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1498     C++ linking (needed for some HP-UX environments).
1499    
1500     14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1501     (unused parameter) and in the pcre_printint() function (omitted "default"
1502     switch label when the default is to do nothing).
1503    
1504     15. Added some code to make it possible, when PCRE is compiled as a C++
1505     library, to replace subject pointers for pcre_exec() with a smart pointer
1506     class, thus making it possible to process discontinuous strings.
1507    
1508     16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1509     much the same function. They were added by different people who were trying
1510     to make PCRE easy to compile on non-Unix systems. It has been suggested
1511     that PCRE_EXPORT be abolished now that there is more automatic apparatus
1512     for compiling on Windows systems. I have therefore replaced it with
1513     PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1514     defaults to "extern" for C or "extern C" for C++, which works fine on
1515     Unix-like systems. It is now possible to override the value of PCRE_DATA_
1516     SCOPE with something explicit in config.h. In addition:
1517    
1518     (a) pcreposix.h still had just "extern" instead of either of these macros;
1519     I have replaced it with PCRE_DATA_SCOPE.
1520    
1521     (b) Functions such as _pcre_xclass(), which are internal to the library,
1522     but external in the C sense, all had PCRE_EXPORT in their definitions.
1523     This is apparently wrong for the Windows case, so I have removed it.
1524     (It makes no difference on Unix-like systems.)
1525    
1526     17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1527     of recursive calls to match(). This is different to MATCH_LIMIT because
1528     that limits the total number of calls to match(), not all of which increase
1529     the depth of recursion. Limiting the recursion depth limits the amount of
1530     stack (or heap if NO_RECURSE is set) that is used. The default can be set
1531     when PCRE is compiled, and changed at run time. A patch from Google adds
1532     this functionality to the C++ interface.
1533    
1534     18. Changes to the handling of Unicode character properties:
1535    
1536     (a) Updated the table to Unicode 4.1.0.
1537    
1538     (b) Recognize characters that are not in the table as "Cn" (undefined).
1539    
1540     (c) I revised the way the table is implemented to a much improved format
1541     which includes recognition of ranges. It now supports the ranges that
1542     are defined in UnicodeData.txt, and it also amalgamates other
1543     characters into ranges. This has reduced the number of entries in the
1544     table from around 16,000 to around 3,000, thus reducing its size
1545     considerably. I realized I did not need to use a tree structure after
1546     all - a binary chop search is just as efficient. Having reduced the
1547     number of entries, I extended their size from 6 bytes to 8 bytes to
1548     allow for more data.
1549    
1550     (d) Added support for Unicode script names via properties such as \p{Han}.
1551    
1552     19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1553     matching that character.
1554    
1555     20. When matching a repeated Unicode property with a minimum greater than zero,
1556     (for example \pL{2,}), PCRE could look past the end of the subject if it
1557     reached it while seeking the minimum number of characters. This could
1558     happen only if some of the characters were more than one byte long, because
1559     there is a check for at least the minimum number of bytes.
1560    
1561     21. Refactored the implementation of \p and \P so as to be more general, to
1562     allow for more different types of property in future. This has changed the
1563     compiled form incompatibly. Anybody with saved compiled patterns that use
1564     \p or \P will have to recompile them.
1565    
1566     22. Added "Any" and "L&" to the supported property types.
1567    
1568     23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1569     but give a compile time error if the value is greater than 0xff.
1570    
1571     24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1572     accidentally not being installed or uninstalled.
1573    
1574     25. The pcre.h file was built from pcre.h.in, but the only changes that were
1575     made were to insert the current release number. This seemed silly, because
1576     it made things harder for people building PCRE on systems that don't run
1577     "configure". I have turned pcre.h into a distributed file, no longer built
1578     by "configure", with the version identification directly included. There is
1579     no longer a pcre.h.in file.
1580    
1581     However, this change necessitated a change to the pcre-config script as
1582     well. It is built from pcre-config.in, and one of the substitutions was the
1583     release number. I have updated configure.ac so that ./configure now finds
1584     the release number by grepping pcre.h.
1585    
1586     26. Added the ability to run the tests under valgrind.
1587    
1588    
1589 nigel 85 Version 6.4 05-Sep-05
1590     ---------------------
1591    
1592     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1593     "--" to be printed when multiple files were scanned, even when none of the
1594     -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1595     consider it to be a bug, and have restored the previous behaviour.
1596    
1597     2. A couple of code tidies to get rid of compiler warnings.
1598    
1599     3. The pcretest program used to cheat by referring to symbols in the library
1600     whose names begin with _pcre_. These are internal symbols that are not
1601     really supposed to be visible externally, and in some environments it is
1602     possible to suppress them. The cheating is now confined to including
1603     certain files from the library's source, which is a bit cleaner.
1604    
1605     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1606     file's purpose clearer.
1607    
1608     5. Reorganized pcre_ucp_findchar().
1609    
1610    
1611 nigel 83 Version 6.3 15-Aug-05
1612     ---------------------
1613    
1614     1. The file libpcre.pc.in did not have general read permission in the tarball.
1615    
1616     2. There were some problems when building without C++ support:
1617    
1618     (a) If C++ support was not built, "make install" and "make test" still
1619     tried to test it.
1620    
1621     (b) There were problems when the value of CXX was explicitly set. Some
1622     changes have been made to try to fix these, and ...
1623    
1624     (c) --disable-cpp can now be used to explicitly disable C++ support.
1625    
1626     (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1627     backslash in a target when C++ was disabled. This confuses some
1628     versions of "make", apparently. Using an intermediate variable solves
1629     this. (Same for CPP_LOBJ.)
1630    
1631     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1632     (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1633     necessary on certain architectures.
1634    
1635     4. Added a setting of -export-symbols-regex to the link command to remove
1636     those symbols that are exported in the C sense, but actually are local
1637     within the library, and not documented. Their names all begin with
1638     "_pcre_". This is not a perfect job, because (a) we have to except some
1639     symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1640     available (and never for static libraries). I have made a note to try to
1641     find a way round (a) in the future.
1642    
1643    
1644 nigel 81 Version 6.2 01-Aug-05
1645     ---------------------
1646    
1647     1. There was no test for integer overflow of quantifier values. A construction
1648     such as {1111111111111111} would give undefined results. What is worse, if
1649     a minimum quantifier for a parenthesized subpattern overflowed and became
1650     negative, the calculation of the memory size went wrong. This could have
1651     led to memory overwriting.
1652    
1653     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1654    
1655     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1656     operating environments where this matters.
1657    
1658     4. Applied Giuseppe Maxia's patch to add additional features for controlling
1659     PCRE options from within the C++ wrapper.
1660    
1661     5. Named capturing subpatterns were not being correctly counted when a pattern
1662     was compiled. This caused two problems: (a) If there were more than 100
1663     such subpatterns, the calculation of the memory needed for the whole
1664     compiled pattern went wrong, leading to an overflow error. (b) Numerical
1665     back references of the form \12, where the number was greater than 9, were
1666     not recognized as back references, even though there were sufficient
1667     previous subpatterns.
1668    
1669     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1670     versions of gcc, e.g. 2.95.4.
1671    
1672    
1673 nigel 79 Version 6.1 21-Jun-05
1674     ---------------------
1675    
1676     1. There was one reference to the variable "posix" in pcretest.c that was not
1677     surrounded by "#if !defined NOPOSIX".
1678    
1679     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1680     the cross-check on the old pcre_info() function, for the benefit of the
1681     cut-down version of PCRE that is currently imported into Exim.
1682    
1683     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1684     allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1685     patterns that start (?i) but I don't think that matters. The use of (?i) is
1686     just an example; this all applies to the other options as well.
1687    
1688     4. Since libtool seems to echo the compile commands it is issuing, the output
1689     from "make" can be reduced a bit by putting "@" in front of each libtool
1690     compile command.
1691    
1692     5. Patch from the folks at Google for configure.in to be a bit more thorough
1693     in checking for a suitable C++ installation before trying to compile the
1694     C++ stuff. This should fix a reported problem when a compiler was present,
1695     but no suitable headers.
1696    
1697     6. The man pages all had just "PCRE" as their title. I have changed them to
1698     be the relevant file name. I have also arranged that these names are
1699     retained in the file doc/pcre.txt, which is a concatenation in text format
1700     of all the man pages except the little individual ones for each function.
1701    
1702     7. The NON-UNIX-USE file had not been updated for the different set of source
1703     files that come with release 6. I also added a few comments about the C++
1704     wrapper.
1705    
1706    
1707 nigel 77 Version 6.0 07-Jun-05
1708     ---------------------
1709    
1710     1. Some minor internal re-organization to help with my DFA experiments.
1711    
1712     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1713     didn't matter for the library itself when fully configured, but did matter
1714     when compiling without UCP support, or within Exim, where the ucp files are
1715     not imported.
1716    
1717     3. Refactoring of the library code to split up the various functions into
1718     different source modules. The addition of the new DFA matching code (see
1719     below) to a single monolithic source would have made it really too
1720     unwieldy, quite apart from causing all the code to be include in a
1721     statically linked application, when only some functions are used. This is
1722     relevant even without the DFA addition now that patterns can be compiled in
1723     one application and matched in another.
1724    
1725     The downside of splitting up is that there have to be some external
1726     functions and data tables that are used internally in different modules of
1727     the library but which are not part of the API. These have all had their
1728     names changed to start with "_pcre_" so that they are unlikely to clash
1729     with other external names.
1730    
1731     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1732     a different (DFA) algorithm. Although it is slower than the original
1733     function, it does have some advantages for certain types of matching
1734     problem.
1735    
1736     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1737     including restarting after a partial match.
1738    
1739     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1740     defined when compiling for Windows was sent to me. I have put it into the
1741     code, though I have no means of testing or verifying it.
1742    
1743     7. Added the pcre_refcount() auxiliary function.
1744    
1745     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1746     match before or at the first newline in the subject string. In pcretest,
1747     the /f option on a pattern can be used to set this.
1748    
1749     9. A repeated \w when used in UTF-8 mode with characters greater than 256
1750     would behave wrongly. This has been present in PCRE since release 4.0.
1751    
1752     10. A number of changes to the pcregrep command:
1753    
1754     (a) Refactored how -x works; insert ^(...)$ instead of setting
1755     PCRE_ANCHORED and checking the length, in preparation for adding
1756     something similar for -w.
1757    
1758     (b) Added the -w (match as a word) option.
1759    
1760     (c) Refactored the way lines are read and buffered so as to have more
1761     than one at a time available.
1762    
1763     (d) Implemented a pcregrep test script.
1764    
1765     (e) Added the -M (multiline match) option. This allows patterns to match
1766     over several lines of the subject. The buffering ensures that at least
1767     8K, or the rest of the document (whichever is the shorter) is available
1768     for matching (and similarly the previous 8K for lookbehind assertions).
1769    
1770     (f) Changed the --help output so that it now says
1771    
1772     -w, --word-regex(p)
1773    
1774     instead of two lines, one with "regex" and the other with "regexp"
1775     because that confused at least one person since the short forms are the
1776     same. (This required a bit of code, as the output is generated
1777     automatically from a table. It wasn't just a text change.)
1778    
1779     (g) -- can be used to terminate pcregrep options if the next thing isn't an
1780     option but starts with a hyphen. Could be a pattern or a path name
1781     starting with a hyphen, for instance.
1782    
1783     (h) "-" can be given as a file name to represent stdin.
1784    
1785     (i) When file names are being printed, "(standard input)" is used for
1786     the standard input, for compatibility with GNU grep. Previously
1787     "<stdin>" was used.
1788    
1789     (j) The option --label=xxx can be used to supply a name to be used for
1790     stdin when file names are being printed. There is no short form.
1791    
1792     (k) Re-factored the options decoding logic because we are going to add
1793     two more options that take data. Such options can now be given in four
1794     different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1795    
1796     (l) Added the -A, -B, and -C options for requesting that lines of context
1797     around matches be printed.
1798    
1799     (m) Added the -L option to print the names of files that do not contain
1800     any matching lines, that is, the complement of -l.
1801    
1802     (n) The return code is 2 if any file cannot be opened, but pcregrep does
1803     continue to scan other files.
1804    
1805     (o) The -s option was incorrectly implemented. For compatibility with other
1806     greps, it now suppresses the error message for a non-existent or non-
1807     accessible file (but not the return code). There is a new option called
1808     -q that suppresses the output of matching lines, which was what -s was
1809     previously doing.
1810    
1811     (p) Added --include and --exclude options to specify files for inclusion
1812     and exclusion when recursing.
1813    
1814     11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1815     Hopefully, it now does.
1816    
1817     12. Missing cast in pcre_study().
1818    
1819     13. Added an "uninstall" target to the makefile.
1820    
1821     14. Replaced "extern" in the function prototypes in Makefile.in with
1822     "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1823     world, but is set differently for Windows.
1824    
1825     15. Added a second compiling function called pcre_compile2(). The only
1826     difference is that it has an extra argument, which is a pointer to an
1827     integer error code. When there is a compile-time failure, this is set
1828     non-zero, in addition to the error test pointer being set to point to an
1829     error message. The new argument may be NULL if no error number is required
1830     (but then you may as well call pcre_compile(), which is now just a
1831     wrapper). This facility is provided because some applications need a
1832     numeric error indication, but it has also enabled me to tidy up the way
1833     compile-time errors are handled in the POSIX wrapper.
1834    
1835     16. Added VPATH=.libs to the makefile; this should help when building with one
1836     prefix path and installing with another. (Or so I'm told by someone who
1837     knows more about this stuff than I do.)
1838    
1839     17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1840     passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1841     match everything, including newlines. This is not POSIX-compatible, but
1842     somebody wanted the feature. From pcretest it can be activated by using
1843     both the P and the s flags.
1844    
1845     18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1846    
1847     19. libpcre.pc was being incorrectly installed as executable.
1848    
1849     20. A couple of places in pcretest check for end-of-line by looking for '\n';
1850     it now also looks for '\r' so that it will work unmodified on Windows.
1851    
1852     21. Added Google's contributed C++ wrapper to the distribution.
1853    
1854     22. Added some untidy missing memory free() calls in pcretest, to keep
1855     Electric Fence happy when testing.
1856    
1857    
1858    
1859 nigel 75 Version 5.0 13-Sep-04
1860     ---------------------
1861    
1862     1. Internal change: literal characters are no longer packed up into items
1863     containing multiple characters in a single byte-string. Each character
1864     is now matched using a separate opcode. However, there may be more than one
1865     byte in the character in UTF-8 mode.
1866    
1867     2. The pcre_callout_block structure has two new fields: pattern_position and
1868     next_item_length. These contain the offset in the pattern to the next match
1869     item, and its length, respectively.
1870    
1871     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1872     insertion of callouts before each pattern item. Added the /C option to
1873     pcretest to make use of this.
1874    
1875     4. On the advice of a Windows user, the lines
1876    
1877     #if defined(_WIN32) || defined(WIN32)
1878     _setmode( _fileno( stdout ), 0x8000 );
1879     #endif /* defined(_WIN32) || defined(WIN32) */
1880    
1881     have been added to the source of pcretest. This apparently does useful
1882     magic in relation to line terminators.
1883    
1884     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1885     for the benefit of those environments where the "b" makes a difference.
1886    
1887     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1888     to know about it. I have put a hack into configure.in that adds in code
1889     to set GCC=yes if CC=icc. This seems to end up at a point in the
1890     generated configure script that is early enough to affect the setting of
1891     compiler options, which is what is needed, but I have no means of testing
1892     whether it really works. (The user who reported this had patched the
1893     generated configure script, which of course I cannot do.)
1894    
1895     LATER: After change 22 below (new libtool files), the configure script
1896     seems to know about icc (and also ecc). Therefore, I have commented out
1897     this hack in configure.in.
1898    
1899     7. Added support for pkg-config (2 patches were sent in).
1900    
1901     8. Negated POSIX character classes that used a combination of internal tables
1902     were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1903     [[:^ascii]]. Typically, they would match almost any characters. The other
1904     POSIX classes were not broken in this way.
1905    
1906     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1907     to find the match, as PCRE was deluded into thinking that the match had to
1908     start at the start point or following a newline. The same bug applied to
1909     patterns with negative forward assertions or any backward assertions
1910     preceding ".*" at the start, unless the pattern required a fixed first
1911     character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1912    
1913     10. In UTF-8 mode, when moving forwards in the subject after a failed match
1914     starting at the last subject character, bytes beyond the end of the subject
1915     string were read.
1916    
1917     11. Renamed the variable "class" as "classbits" to make life easier for C++
1918     users. (Previously there was a macro definition, but it apparently wasn't
1919     enough.)
1920    
1921     12. Added the new field "tables" to the extra data so that tables can be passed
1922     in at exec time, or the internal tables can be re-selected. This allows
1923     a compiled regex to be saved and re-used at a later time by a different
1924     program that might have everything at different addresses.
1925    
1926     13. Modified the pcre-config script so that, when run on Solaris, it shows a
1927     -R library as well as a -L library.
1928    
1929     14. The debugging options of pcretest (-d on the command line or D on a
1930     pattern) showed incorrect output for anything following an extended class
1931     that contained multibyte characters and which was followed by a quantifier.
1932    
1933     15. Added optional support for general category Unicode character properties
1934     via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1935     support. It adds about 90K to the size of the library. The meanings of the
1936     inbuilt class escapes such as \d and \s have NOT been changed.
1937    
1938     16. Updated pcredemo.c to include calls to free() to release the memory for the
1939     compiled pattern.
1940    
1941     17. The generated file chartables.c was being created in the source directory
1942     instead of in the building directory. This caused the build to fail if the
1943     source directory was different from the building directory, and was
1944     read-only.
1945    
1946     18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1947     file. No doubt somebody will tell me if they don't make sense... Also added
1948     Dan Mooney's comments about building on OpenVMS.
1949    
1950     19. Added support for partial matching via the PCRE_PARTIAL option for
1951     pcre_exec() and the \P data escape in pcretest.
1952    
1953     20. Extended pcretest with 3 new pattern features:
1954    
1955     (i) A pattern option of the form ">rest-of-line" causes pcretest to
1956     write the compiled pattern to the file whose name is "rest-of-line".
1957     This is a straight binary dump of the data, with the saved pointer to
1958     the character tables forced to be NULL. The study data, if any, is
1959     written too. After writing, pcretest reads a new pattern.
1960    
1961     (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1962     compiled pattern from the given file. There must not be any
1963     occurrences of "<" in the file name (pretty unlikely); if there are,
1964     pcretest will instead treat the initial "<" as a pattern delimiter.
1965     After reading in the pattern, pcretest goes on to read data lines as
1966     usual.
1967    
1968     (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1969     and 16-bit fields in a compiled pattern, to simulate a pattern that
1970     was compiled on a host of opposite endianness.
1971    
1972     21. The pcre-exec() function can now cope with patterns that were compiled on
1973     hosts of opposite endianness, with this restriction:
1974    
1975     As for any compiled expression that is saved and used later, the tables
1976     pointer field cannot be preserved; the extra_data field in the arguments
1977     to pcre_exec() should be used to pass in a tables address if a value
1978     other than the default internal tables were used at compile time.
1979    
1980     22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1981     now diagnosed as an error. Previously, most of the time, a negative number
1982     would have been treated as zero, but if in addition "ovector" was passed as
1983     NULL, a crash could occur.
1984    
1985     23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1986     new versions from the libtool 1.5 distribution (the last one is a copy of
1987     a file called libtool.m4). This seems to have fixed the need to patch
1988     "configure" to support Darwin 1.3 (which I used to do). However, I still
1989     had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1990     workstation).
1991    
1992     24. Changed the PCRE licence to be the more standard "BSD" licence.
1993    
1994    
1995 nigel 73 Version 4.5 01-Dec-03
1996     ---------------------
1997    
1998     1. There has been some re-arrangement of the code for the match() function so
1999     that it can be compiled in a version that does not call itself recursively.
2000     Instead, it keeps those local variables that need separate instances for
2001     each "recursion" in a frame on the heap, and gets/frees frames whenever it
2002     needs to "recurse". Keeping track of where control must go is done by means
2003     of setjmp/longjmp. The whole thing is implemented by a set of macros that
2004     hide most of the details from the main code, and operates only if
2005     NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2006     "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2007     operating.
2008    
2009     To make it easier for callers to provide specially tailored get/free
2010     functions for this usage, two new functions, pcre_stack_malloc, and
2011     pcre_stack_free, are used. They are always called in strict stacking order,
2012     and the size of block requested is always the same.
2013    
2014     The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2015     PCRE has been compiled to use the stack or the heap for recursion. The
2016     -C option of pcretest uses this to show which version is compiled.
2017    
2018     A new data escape \S, is added to pcretest; it causes the amounts of store
2019     obtained and freed by both kinds of malloc/free at match time to be added
2020     to the output.
2021    
2022     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2023     what's available on my current Linux desktop machine.
2024    
2025     3. When matching a UTF-8 string, the test for a valid string at the start has
2026     been extended. If start_offset is not zero, PCRE now checks that it points
2027     to a byte that is the start of a UTF-8 character. If not, it returns
2028     PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2029     this is necessary because there may be backward assertions in the pattern.
2030     When matching the same subject several times, it may save resources to use
2031     PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2032    
2033     4. The code for checking the validity of UTF-8 strings has been tightened so
2034     that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2035     containing "overlong sequences".
2036    
2037     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2038     I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2039     should have been "&", but it just so happened that all the cases this let
2040     through by mistake were picked up later in the function.
2041    
2042     6. I had used a variable called "isblank" - this is a C99 function, causing
2043     some compilers to warn. To avoid this, I renamed it (as "blankclass").
2044    
2045     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2046     prompting; (b) run "./pcretest /dev/null" at the start of the test script
2047     so the version is shown; (c) stop "make test" echoing "./RunTest".
2048    
2049     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2050    
2051     9. The prototype for memmove() for systems that don't have it was using
2052     size_t, but the inclusion of the header that defines size_t was later. I've
2053     moved the #includes for the C headers earlier to avoid this.
2054    
2055     10. Added some adjustments to the code to make it easier to compiler on certain
2056     special systems:
2057    
2058     (a) Some "const" qualifiers were missing.
2059     (b) Added the macro EXPORT before all exported functions; by default this
2060     is defined to be empty.
2061     (c) Changed the dftables auxiliary program (that builds chartables.c) so
2062     that it reads its output file name as an argument instead of writing
2063     to the standard output and assuming this can be redirected.
2064    
2065     11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2066     class containing characters with values greater than 255, PCRE compilation
2067     went into a loop.
2068    
2069     12. A recursive reference to a subpattern that was within another subpattern
2070     that had a minimum quantifier of zero caused PCRE to crash. For example,
2071     (x(y(?2))z)? provoked this bug with a subject that got as far as the
2072     recursion. If the recursively-called subpattern itself had a zero repeat,
2073     that was OK.
2074    
2075     13. In pcretest, the buffer for reading a data line was set at 30K, but the
2076     buffer into which it was copied (for escape processing) was still set at
2077     1024, so long lines caused crashes.
2078    
2079     14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2080     "internal error: code overflow...". This applied to any character class
2081     that was followed by a possessive quantifier.
2082    
2083     15. Modified the Makefile to add libpcre.la as a prerequisite for
2084     libpcreposix.la because I was told this is needed for a parallel build to
2085     work.
2086    
2087     16. If a pattern that contained .* following optional items at the start was
2088     studied, the wrong optimizing data was generated, leading to matching
2089     errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2090     matching string must start with a or b or c. The correct conclusion for
2091     this pattern is that a match can start with any character.
2092    
2093    
2094 nigel 71 Version 4.4 13-Aug-03
2095     ---------------------
2096    
2097     1. In UTF-8 mode, a character class containing characters with values between
2098     127 and 255 was not handled correctly if the compiled pattern was studied.
2099     In fixing this, I have also improved the studying algorithm for such
2100     classes (slightly).
2101    
2102     2. Three internal functions had redundant arguments passed to them. Removal
2103     might give a very teeny performance improvement.
2104    
2105     3. Documentation bug: the value of the capture_top field in a callout is *one
2106     more than* the number of the hightest numbered captured substring.
2107    
2108     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2109     in incorrectly linking with a previously installed version. They now link
2110     explicitly with libpcre.la.
2111    
2112     5. configure.in no longer needs to recognize Cygwin specially.
2113    
2114     6. A problem in pcre.in for Windows platforms is fixed.
2115    
2116     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2117     pcretest, it used to include the size of the study block as part of its
2118     output. Unfortunately, the structure contains a field that has a different
2119     size on different hardware architectures. This meant that the tests that
2120     showed this size failed. As the block is currently always of a fixed size,
2121     this information isn't actually particularly useful in pcretest output, so
2122     I have just removed it.
2123    
2124     8. Three pre-processor statements accidentally did not start in column 1.
2125     Sadly, there are *still* compilers around that complain, even though
2126     standard C has not required this for well over a decade. Sigh.
2127    
2128     9. In pcretest, the code for checking callouts passed small integers in the
2129     callout_data field, which is a void * field. However, some picky compilers
2130     complained about the casts involved for this on 64-bit systems. Now
2131     pcretest passes the address of the small integer instead, which should get
2132     rid of the warnings.
2133    
2134     10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2135     both compile and run time, and gives an error if an invalid UTF-8 sequence
2136     is found. There is a option for disabling this check in cases where the
2137     string is known to be correct and/or the maximum performance is wanted.
2138    
2139     11. In response to a bug report, I changed one line in Makefile.in from
2140    
2141     -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2142     to
2143     -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2144    
2145     to look similar to other lines, but I have no way of telling whether this
2146     is the right thing to do, as I do not use Windows. No doubt I'll get told
2147     if it's wrong...
2148    
2149    
2150 nigel 69 Version 4.3 21-May-03
2151     ---------------------
2152    
2153     1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2154     Makefile.
2155    
2156     2. Some refactoring to improve the quality of the code:
2157    
2158     (i) The utf8_table... variables are now declared "const".
2159    
2160     (ii) The code for \cx, which used the "case flipping" table to upper case
2161     lower case letters, now just substracts 32. This is ASCII-specific,
2162     but the whole concept of \cx is ASCII-specific, so it seems
2163     reasonable.
2164    
2165     (iii) PCRE was using its character types table to recognize decimal and
2166     hexadecimal digits in the pattern. This is silly, because it handles
2167     only 0-9, a-f, and A-F, but the character types table is locale-
2168     specific, which means strange things might happen. A private
2169     table is now used for this - though it costs 256 bytes, a table is
2170     much faster than multiple explicit tests. Of course, the standard
2171     character types table is still used for matching digits in subject
2172     strings against \d.
2173    
2174     (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2175     ending in _t are). So I've renamed it as ESC_tee.
2176    
2177     3. The first argument for regexec() in the POSIX wrapper should have been
2178     defined as "const".
2179    
2180     4. Changed pcretest to use malloc() for its buffers so that they can be
2181     Electric Fenced for debugging.
2182    
2183     5. There were several places in the code where, in UTF-8 mode, PCRE would try
2184     to read one or more bytes before the start of the subject string. Often this
2185     had no effect on PCRE's behaviour, but in some circumstances it could
2186     provoke a segmentation fault.
2187    
2188     6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2189     to try to read one or more bytes before the start of the subject string.
2190    
2191     7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2192     UTF-8 support could misbehave in various ways if the subject string
2193     contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2194     area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2195     back over UTF-8 characters.)
2196    
2197    
2198 nigel 67 Version 4.2 14-Apr-03
2199     ---------------------
2200    
2201     1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2202    
2203     2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2204     [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2205     [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2206     [WIN_PREFIX]: new variable, "cyg" for Cygwin
2207     * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2208     and BUILD_EXEEXT
2209     Note: automatic setting of the BUILD variables is not yet working
2210     set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2211     compile-time but not at link-time
2212     [LINK]: use for linking executables only
2213     make different versions for Windows and non-Windows
2214     [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2215     libraries
2216     [LINK_FOR_BUILD]: new variable
2217     [OBJEXT]: use throughout
2218     [EXEEXT]: use throughout
2219     <winshared>: new target
2220     <wininstall>: new target
2221     <dftables.o>: use native compiler
2222     <dftables>: use native linker
2223     <install>: handle Windows platform correctly
2224     <clean>: ditto
2225     <check>: ditto
2226     copy DLL to top builddir before testing
2227    
2228     As part of these changes, -no-undefined was removed again. This was reported
2229     to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2230     in any case.
2231    
2232     3. Some tidies to get rid of compiler warnings:
2233    
2234     . In the match_data structure, match_limit was an unsigned long int, whereas
2235     match_call_count was an int. I've made them both unsigned long ints.
2236    
2237     . In pcretest the fact that a const uschar * doesn't automatically cast to
2238     a void * provoked a warning.
2239    
2240     . Turning on some more compiler warnings threw up some "shadow" variables
2241     and a few more missing casts.
2242    
2243     4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2244     option, a class that contained a single character with a value between 128
2245     and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2246    
2247     5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2248     option, a class that contained several characters, but with at least one
2249     whose value was between 128 and 255 caused PCRE to crash.
2250    
2251    
2252 nigel 65 Version 4.1 12-Mar-03
2253     ---------------------
2254 nigel 63
2255 nigel 65 1. Compiling with gcc -pedantic found a couple of places where casts were
2256     needed, and a string in dftables.c that was longer than standard compilers are
2257     required to support.
2258    
2259     2. Compiling with Sun's compiler found a few more places where the code could
2260     be tidied up in order to avoid warnings.
2261    
2262     3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2263     first of these names is deprecated in the latest Autoconf in favour of the name
2264     CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2265     compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2266     analogy I have changed it to CFLAGS_FOR_BUILD.
2267    
2268     4. Added -no-undefined to the linking command in the Makefile, because this is
2269     apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2270     linking step for the pcreposix library.
2271    
2272     5. PCRE was failing to diagnose the case of two named groups with the same
2273     name.
2274    
2275     6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2276     literal character that is needed in the subject for a match, and scans along to
2277     ensure that it is present before embarking on the full matching process. This
2278     saves time in cases of nested unlimited repeats that are never going to match.
2279     Problem: the scan can take a lot of time if the subject is very long (e.g.
2280     megabytes), thus penalizing straightforward matches. It is now done only if the
2281     amount of subject to be scanned is less than 1000 bytes.
2282    
2283     7. A lesser problem with the same optimization is that it was recording the
2284     first character of an anchored pattern as "needed", thus provoking a search
2285     right along the subject, even when the first match of the pattern was going to
2286     fail. The "needed" character is now not set for anchored patterns, unless it
2287     follows something in the pattern that is of non-fixed length. Thus, it still
2288     fulfils its original purpose of finding quick non-matches in cases of nested
2289     unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2290    
2291    
2292     Version 4.0 17-Feb-03
2293     ---------------------
2294    
2295 nigel 63 1. If a comment in an extended regex that started immediately after a meta-item
2296     extended to the end of string, PCRE compiled incorrect data. This could lead to
2297     all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2298    
2299     2. Moved to autoconf 2.53 and libtool 1.4.2.
2300    
2301     3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2302     the special perltest8 script is no longer needed - all the tests can be run
2303     from a single perltest script.
2304    
2305     4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2306     by \s. It has now been removed in PCRE. This means it isn't recognized as
2307     whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2308     class [:space:] *does* include VT, thereby creating a mess.
2309    
2310     5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2311     space and tab.
2312    
2313     6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2314     its new features into the main test script, reducing the number of scripts.
2315    
2316     7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2317     were backward compatible, and made the (?i) apply to the whole pattern, as if
2318     /i were given. Now it behaves more logically, and applies the option setting
2319     only to what follows. PCRE has been changed to follow suit. However, if it
2320     finds options settings right at the start of the pattern, it extracts them into
2321     the global options, as before. Thus, they show up in the info data.
2322    
2323     8. Added support for the \Q...\E escape sequence. Characters in between are
2324     treated as literals. This is slightly different from Perl in that $ and @ are
2325     also handled as literals inside the quotes. In Perl, they will cause variable
2326     interpolation. Note the following examples:
2327    
2328     Pattern PCRE matches Perl matches
2329    
2330     \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2331     \Qabc\$xyz\E abc\$xyz abc\$xyz
2332     \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2333    
2334     For compatibility with Perl, \Q...\E sequences are recognized inside character
2335     classes as well as outside them.
2336    
2337     9. Re-organized 3 code statements in pcretest to avoid "overflow in
2338     floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2339     (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2340     signed/unsigned warnings.
2341    
2342     10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2343     option for pcretest, so I've replaced it by a simple function that does just
2344     that job.
2345    
2346     11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2347     "pcregrep -".
2348    
2349     12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2350     Java package. This provides some syntactic sugar for simple cases of what my
2351     documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2352     as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2353     item, you can use this simplified notation. Note that only makes sense with
2354     greedy quantifiers. Consequently, the use of the possessive quantifier forces
2355     greediness, whatever the setting of the PCRE_UNGREEDY option.
2356    
2357     13. A change of greediness default within a pattern was not taking effect at
2358     the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2359     subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2360     was abstracted outside.
2361    
2362     14. PCRE now supports the \G assertion. It is true when the current matching
2363     position is at the start point of the match. This differs from \A when the
2364     starting offset is non-zero. Used with the /g option of pcretest (or similar
2365     code), it works in the same way as it does for Perl's /g option. If all
2366     alternatives of a regex begin with \G, the expression is anchored to the start
2367     match position, and the "anchored" flag is set in the compiled expression.
2368    
2369     15. Some bugs concerning the handling of certain option changes within patterns
2370     have been fixed. These applied to options other than (?ims). For example,
2371     "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2372     been the other way round. Some of this was related to change 7 above.
2373    
2374     16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2375     features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2376     and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2377     POSIX classes only within a class (e.g. /[[:alpha:]]/).
2378    
2379     17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2380     mode. Unlike ".", it always matches newline, whatever the setting of
2381     PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2382     assertions. Perl allows it, but it doesn't (in general) work because it can't
2383     calculate the length of the lookbehind. At least, that's the case for Perl
2384     5.8.0 - I've been told they are going to document that it doesn't work in
2385     future.
2386    
2387     18. Added an error diagnosis for escapes that PCRE does not support: these are
2388     \L, \l, \N, \P, \p, \U, \u, and \X.
2389    
2390     19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2391     reading past the end of the pattern in cases such as /[abcd/.
2392    
2393     20. PCRE was getting more memory than necessary for patterns with classes that
2394     contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2395    
2396     21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2397     compiling PCRE for use with Virtual Pascal.
2398    
2399     22. Small fix to the Makefile to make it work properly if the build is done
2400     outside the source tree.
2401    
2402     23. Added a new extension: a condition to go with recursion. If a conditional
2403     subpattern starts with (?(R) the "true" branch is used if recursion has
2404     happened, whereas the "false" branch is used only at the top level.
2405    
2406     24. When there was a very long string of literal characters (over 255 bytes
2407     without UTF support, over 250 bytes with UTF support), the computation of how
2408     much memory was required could be incorrect, leading to segfaults or other
2409     strange effects.
2410    
2411     25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2412     start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2413     there was a subsequent back reference to those brackets. This meant that, for
2414     example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2415     possible to check for precisely this case. All we can do is abandon the
2416     optimization if .* occurs inside capturing brackets when there are any back
2417     references whatsoever. (See below for a better fix that came later.)
2418    
2419     26. The handling of the optimization for finding the first character of a
2420     non-anchored pattern, and for finding a character that is required later in the
2421     match were failing in some cases. This didn't break the matching; it just
2422     failed to optimize when it could. The way this is done has been re-implemented.
2423    
2424     27. Fixed typo in error message for invalid (?R item (it said "(?p").
2425    
2426     28. Added a new feature that provides some of the functionality that Perl
2427     provides with (?{...}). The facility is termed a "callout". The way it is done
2428     in PCRE is for the caller to provide an optional function, by setting
2429     pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2430     global variable. By default it is unset, which disables all calling out. To get
2431     the function called, the regex must include (?C) at appropriate points. This
2432     is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2433     This provides a means of identifying different callout points. When PCRE
2434     reaches such a point in the regex, if pcre_callout has been set, the external
2435     function is called. It is provided with data in a structure called
2436     pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2437     matching continues; if it returns a non-zero value, the match at the current
2438     point fails. However, backtracking will occur if possible. [This was changed
2439     later and other features added - see item 49 below.]
2440    
2441     29. pcretest is upgraded to test the callout functionality. It provides a
2442     callout function that displays information. By default, it shows the start of
2443     the match and the current position in the text. There are some new data escapes
2444     to vary what happens:
2445    
2446     \C+ in addition, show current contents of captured substrings
2447     \C- do not supply a callout function
2448     \C!n return 1 when callout number n is reached
2449     \C!n!m return 1 when callout number n is reached for the mth time
2450    
2451     30. If pcregrep was called with the -l option and just a single file name, it
2452     output "<stdin>" if a match was found, instead of the file name.
2453    
2454     31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2455     slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2456     pcre_exec(). This saves a malloc/free per call. The default value of
2457     POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2458     when configuring.
2459    
2460     32. The default maximum size of a compiled pattern is 64K. There have been a
2461     few cases of people hitting this limit. The code now uses macros to handle the
2462     storing of links as offsets within the compiled pattern. It defaults to 2-byte
2463     links, but this can be changed to 3 or 4 bytes by --with-link-size when
2464     configuring. Tests 2 and 5 work only with 2-byte links because they output
2465     debugging information about compiled patterns.
2466    
2467     33. Internal code re-arrangements:
2468    
2469     (a) Moved the debugging function for printing out a compiled regex into
2470     its own source file (printint.c) and used #include to pull it into
2471     pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2472     separate copies.
2473    
2474     (b) Defined the list of op-code names for debugging as a macro in
2475     internal.h so that it is next to the definition of the opcodes.
2476    
2477     (c) Defined a table of op-code lengths for simpler skipping along compiled
2478     code. This is again a macro in internal.h so that it is next to the
2479     definition of the opcodes.
2480    
2481     34. Added support for recursive calls to individual subpatterns, along the
2482     lines of Robin Houston's patch (but implemented somewhat differently).
2483    
2484     35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2485     allow it to read and process whole directories in Win32. This code was
2486     contributed by Lionel Fourquaux; it has not been tested by me.
2487    
2488     36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2489     used to name a group. Names consist of alphanumerics and underscores, and must
2490     be unique. Back references use the syntax (?P=name) and recursive calls use
2491     (?P>name) which is a PCRE extension to the Python extension. Groups still have
2492     numbers. The function pcre_fullinfo() can be used after compilation to extract
2493     a name/number map. There are three relevant calls:
2494    
2495     PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2496     PCRE_INFO_NAMECOUNT yields the number of entries
2497     PCRE_INFO_NAMETABLE yields a pointer to the map.
2498    
2499     The map is a vector of fixed-size entries. The size of each entry depends on
2500     the length of the longest name used. The first two bytes of each entry are the
2501     group number, most significant byte first. There follows the corresponding
2502     name, zero terminated. The names are in alphabetical order.
2503    
2504     37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2505     case instead of 255. Making it the same both with and without UTF-8 support
2506     means that the same test output works with both.
2507    
2508     38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2509     calling malloc() with a zero argument.
2510    
2511     39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2512     optimization. I've improved things by keeping a bitmap of backreferences with
2513     numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2514     fact referenced, the optimization can be applied. It is unlikely that a
2515     relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2516     the match to follow \n) will appear inside brackets with a number greater than
2517     31, but if it does, any back reference > 31 suppresses the optimization.
2518    
2519     40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2520     of disabling numbered capturing parentheses. Any opening parenthesis that is
2521     not followed by ? behaves as if it were followed by ?: but named parentheses
2522     can still be used for capturing (and they will acquire numbers in the usual
2523     way).
2524    
2525     41. Redesigned the return codes from the match() function into yes/no/error so
2526     that errors can be passed back from deep inside the nested calls. A malloc
2527     failure while inside a recursive subpattern call now causes the
2528     PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2529    
2530     42. It is now possible to set a limit on the number of times the match()
2531     function is called in a call to pcre_exec(). This facility makes it possible to
2532     limit the amount of recursion and backtracking, though not in a directly
2533     obvious way, because the match() function is used in a number of different
2534     circumstances. The count starts from zero for each position in the subject
2535     string (for non-anchored patterns). The default limit is, for compatibility, a
2536     large number, namely 10 000 000. You can change this in two ways:
2537    
2538     (a) When configuring PCRE before making, you can use --with-match-limit=n
2539     to set a default value for the compiled library.
2540    
2541     (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2542     a different value is set. See 45 below.
2543    
2544     If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2545    
2546     43. Added a new function pcre_config(int, void *) to enable run-time extraction
2547     of things that can be changed at compile time. The first argument specifies
2548     what is wanted and the second points to where the information is to be placed.
2549     The current list of available information is:
2550    
2551     PCRE_CONFIG_UTF8
2552    
2553     The output is an integer that is set to one if UTF-8 support is available;
2554     otherwise it is set to zero.
2555    
2556     PCRE_CONFIG_NEWLINE
2557    
2558     The output is an integer that it set to the value of the code that is used for
2559     newline. It is either LF (10) or CR (13).
2560    
2561     PCRE_CONFIG_LINK_SIZE
2562    
2563     The output is an integer that contains the number of bytes used for internal
2564     linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2565    
2566     PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2567    
2568     The output is an integer that contains the threshold above which the POSIX
2569     interface uses malloc() for output vectors. See item 31 above.
2570    
2571     PCRE_CONFIG_MATCH_LIMIT
2572    
2573     The output is an unsigned integer that contains the default limit of the number
2574     of match() calls in a pcre_exec() execution. See 42 above.
2575    
2576     44. pcretest has been upgraded by the addition of the -C option. This causes it
2577     to extract all the available output from the new pcre_config() function, and to
2578     output it. The program then exits immediately.
2579    
2580     45. A need has arisen to pass over additional data with calls to pcre_exec() in
2581     order to support additional features. One way would have been to define
2582     pcre_exec2() (for example) with extra arguments, but this would not have been
2583     extensible, and would also have required all calls to the original function to
2584     be mapped to the new one. Instead, I have chosen to extend the mechanism that
2585     is used for passing in "extra" data from pcre_study().
2586    
2587     The pcre_extra structure is now exposed and defined in pcre.h. It currently
2588     contains the following fields:
2589    
2590     flags a bitmap indicating which of the following fields are set
2591     study_data opaque data from pcre_study()
2592     match_limit a way of specifying a limit on match() calls for a specific
2593     call to pcre_exec()
2594     callout_data data for callouts (see 49 below)
2595    
2596     The flag bits are also defined in pcre.h, and are
2597    
2598     PCRE_EXTRA_STUDY_DATA
2599     PCRE_EXTRA_MATCH_LIMIT
2600     PCRE_EXTRA_CALLOUT_DATA
2601    
2602     The pcre_study() function now returns one of these new pcre_extra blocks, with
2603     the actual study data pointed to by the study_data field, and the
2604     PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2605     before. That is, this change is entirely upwards-compatible and requires no
2606     change to existing code.
2607    
2608     If you want to pass in additional data to pcre_exec(), you can either place it
2609     in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2610     block.
2611    
2612     46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2613     data string contains the escape sequence \M, pcretest calls pcre_exec() several
2614     times with different match limits, until it finds the minimum value needed for
2615     pcre_exec() to complete. The value is then output. This can be instructive; for
2616     most simple matches the number is quite small, but for pathological cases it
2617     gets very large very quickly.
2618    
2619     47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2620     returns the size of the data block pointed to by the study_data field in a
2621     pcre_extra block, that is, the value that was passed as the argument to
2622     pcre_malloc() when PCRE was getting memory in which to place the information
2623     created by pcre_study(). The fourth argument should point to a size_t variable.
2624     pcretest has been extended so that this information is shown after a successful
2625     pcre_study() call when information about the compiled regex is being displayed.
2626    
2627     48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2628     because what follows is always an absolute path. (Later: it turns out that this
2629     is more than cosmetic for MinGW, because it doesn't like empty path
2630     components.)
2631    
2632     49. Some changes have been made to the callout feature (see 28 above):
2633    
2634     (i) A callout function now has three choices for what it returns:
2635    
2636     0 => success, carry on matching
2637     > 0 => failure at this point, but backtrack if possible
2638     < 0 => serious error, return this value from pcre_exec()
2639    
2640     Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2641     values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2642     "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2643     use by callout functions. It will never be used by PCRE itself.
2644    
2645     (ii) The pcre_extra structure (see 45 above) has a void * field called
2646     callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2647     pcre_callout_block structure has a field of the same name. The contents of
2648     the field passed in the pcre_extra structure are passed to the callout
2649     function in the corresponding field in the callout block. This makes it
2650     easier to use the same callout-containing regex from multiple threads. For
2651     testing, the pcretest program has a new data escape
2652    
2653     \C*n pass the number n (may be negative) as callout_data
2654    
2655     If the callout function in pcretest receives a non-zero value as
2656     callout_data, it returns that value.
2657    
2658     50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2659     there were some redundant $(CFLAGS) in commands that are now specified as
2660     $(LINK), which already includes $(CFLAGS).
2661    
2662     51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2663     has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2664     with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2665     one-byte characters throughout. Note that case-insensitive matching applies
2666     only to characters whose values are less than 256. PCRE doesn't support the
2667     notion of cases for higher-valued characters.
2668    
2669     (i) A character class whose characters are all within 0-255 is handled as
2670     a bit map, and the map is inverted for negative classes. Previously, a
2671     character > 255 always failed to match such a class; however it should
2672     match if the class was a negative one (e.g. [^ab]). This has been fixed.
2673    
2674     (ii) A negated character class with a single character < 255 is coded as
2675     "not this character" (OP_NOT). This wasn't working properly when the test
2676     character was multibyte, either singly or repeated.
2677    
2678     (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2679     mode, for example: \x{100}{2,3}.
2680    
2681     (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2682     singly or repeated) now correctly test multibyte characters. However,
2683     PCRE doesn't recognize any characters with values greater than 255 as
2684     digits, spaces, or word characters. Such characters always match \D, \S,
2685     and \W, and never match \d, \s, or \w.
2686    
2687     (v) Classes may now contain characters and character ranges with values
2688     greater than 255. For example: [ab\x{100}-\x{400}].
2689    
2690     (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2691     PCRE in UTF-8 mode.
2692    
2693     52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2694     PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2695     retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2696     value.)
2697    
2698     53. The single man page has become too large. I have therefore split it up into
2699     a number of separate man pages. These also give rise to individual HTML pages;
2700     these are now put in a separate directory, and there is an index.html page that
2701     lists them all. Some hyperlinking between the pages has been installed.
2702    
2703     54. Added convenience functions for handling named capturing parentheses.
2704    
2705     55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2706     aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2707     true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2708     are faulted.
2709    
2710     56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2711     calling configure. These values are used when compiling the dftables.c program
2712     which is run to generate the source of the default character tables. They
2713     default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2714     you will need to set these values.
2715    
2716     57. Updated the building process for Windows DLL, as provided by Fred Cox.
2717    
2718    
2719     Version 3.9 02-Jan-02
2720 nigel 61 ---------------------
2721    
2722     1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2723    
2724     2. If --disable-static was given, the building process failed when trying to
2725     build pcretest and pcregrep. (For some reason it was using libtool to compile
2726     them, which is not right, as they aren't part of the library.)
2727    
2728    
2729 nigel 59 Version 3.8 18-Dec-01
2730     ---------------------
2731    
2732     1. The experimental UTF-8 code was completely screwed up. It was packing the
2733     bytes in the wrong order. How dumb can you get?
2734    
2735    
2736 nigel 57 Version 3.7 29-Oct-01
2737     ---------------------
2738    
2739     1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2740     This caused pcretest, when used on the test data, to segfault. Unfortunately,
2741     this didn't happen under Solaris 8, where I normally test things.
2742    
2743 nigel 59 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2744     doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2745     isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2746     this fix an hour or so after the initial 3.7 release.)
2747 nigel 57
2748 nigel 59
2749 nigel 55 Version 3.6 23-Oct-01
2750     ---------------------
2751    
2752     1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2753     offsets passed as NULL with zero offset count.
2754    
2755     2. The config.guess and config.sub files had not been updated when I moved to
2756     the latest autoconf.
2757    
2758    
2759 nigel 53 Version 3.5 15-Aug-01
2760     ---------------------
2761 nigel 5
2762 nigel 53 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2763     had been forgotten.
2764    
2765     2. By using declared but undefined structures, we can avoid using "void"
2766     definitions in pcre.h while keeping the internal definitions of the structures
2767     private.
2768    
2769     3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2770     user point of view, this means that both static and shared libraries are built
2771     by default, but this can be individually controlled. More of the work of
2772     handling this static/shared cases is now inside libtool instead of PCRE's make
2773     file.
2774    
2775     4. The pcretest utility is now installed along with pcregrep because it is
2776     useful for users (to test regexs) and by doing this, it automatically gets
2777     relinked by libtool. The documentation has been turned into a man page, so
2778     there are now .1, .txt, and .html versions in /doc.
2779    
2780     5. Upgrades to pcregrep:
2781     (i) Added long-form option names like gnu grep.
2782     (ii) Added --help to list all options with an explanatory phrase.
2783     (iii) Added -r, --recursive to recurse into sub-directories.
2784     (iv) Added -f, --file to read patterns from a file.
2785    
2786     6. pcre_exec() was referring to its "code" argument before testing that
2787     argument for NULL (and giving an error if it was NULL).
2788    
2789     7. Upgraded Makefile.in to allow for compiling in a different directory from
2790     the source directory.
2791    
2792     8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2793     options bits, the pointer it was passed was to an int instead of to an unsigned
2794     long int. This mattered only on 64-bit systems.
2795    
2796     9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2797     generated) instead of pcre.in, which it its source. Also made the same change
2798     in several of the .c files.
2799    
2800     10. A new release of gcc defines printf() as a macro, which broke pcretest
2801     because it had an ifdef in the middle of a string argument for printf(). Fixed
2802     by using separate calls to printf().
2803    
2804     11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2805     script, to force use of CR or LF instead of \n in the source. On non-Unix
2806     systems, the value can be set in config.h.
2807    
2808     12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2809     absolute limit. Changed the text of the error message to make this clear, and
2810     likewise updated the man page.
2811    
2812     13. The limit of 99 on the number of capturing subpatterns has been removed.
2813     The new limit is 65535, which I hope will not be a "real" limit.
2814    
2815    
2816 nigel 51 Version 3.4 22-Aug-00
2817     ---------------------
2818    
2819     1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2820    
2821     2. Diagnose condition (?(0) as an error instead of crashing on matching.
2822    
2823    
2824 nigel 49 Version 3.3 01-Aug-00
2825     ---------------------
2826    
2827     1. If an octal character was given, but the value was greater than \377, it
2828     was not getting masked to the least significant bits, as documented. This could
2829     lead to crashes in some systems.
2830    
2831     2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2832     the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2833    
2834     3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2835     These just pass their arguments on to (pcre_free)(), but they are provided
2836     because some uses of PCRE bind it to non-C systems that can call its functions,
2837     but cannot call free() or pcre_free() directly.
2838    
2839     4. Add "make test" as a synonym for "make check". Corrected some comments in
2840     the Makefile.
2841    
2842     5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2843     Makefile.
2844    
2845     6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2846     command called pgrep for grepping around the active processes.
2847    
2848     7. Added the beginnings of support for UTF-8 character strings.
2849    
2850     8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2851     RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2852     the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2853     out for the ar command.)
2854    
2855    
2856 nigel 47 Version 3.2 12-May-00
2857     ---------------------
2858    
2859     This is purely a bug fixing release.
2860    
2861     1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2862     of ZA. This was just one example of several cases that could provoke this bug,
2863     which was introduced by change 9 of version 2.00. The code for breaking
2864     infinite loops after an iteration that matches an empty string was't working
2865     correctly.
2866    
2867     2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2868     when matched against abbab (for example). After matching an empty string, it
2869     wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2870     caused it to match further down the string than it should.
2871    
2872     3. The code contained an inclusion of sys/types.h. It isn't clear why this
2873     was there because it doesn't seem to be needed, and it causes trouble on some
2874     systems, as it is not a Standard C header. It has been removed.
2875    
2876     4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2877     were reported on the Macintosh. The changes were from
2878    
2879     while ((c = *(++ptr)) != 0 && c != '\n');
2880     to
2881     while ((c = *(++ptr)) != 0 && c != '\n') ;
2882    
2883     Totally extraordinary, but if that's what it takes...
2884    
2885     5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2886     available. Added HAVE_BCOPY and an autoconf test for it; if neither
2887     HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2888     assumes the way PCRE uses memmove() (always moving upwards).
2889    
2890     6. PCRE is being used in one environment where strchr() is not available. There
2891     was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2892     faster code anyway.
2893    
2894    
2895 nigel 45 Version 3.1 09-Feb-00
2896     ---------------------
2897    
2898     The only change in this release is the fixing of some bugs in Makefile.in for
2899     the "install" target:
2900    
2901     (1) It was failing to install pcreposix.h.
2902    
2903     (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2904    
2905    
2906 nigel 43 Version 3.0 01-Feb-00
2907     ---------------------
2908 nigel 41
2909     1. Add support for the /+ modifier to perltest (to output $` like it does in
2910     pcretest).
2911    
2912     2. Add support for the /g modifier to perltest.
2913    
2914     3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2915     matches null strings.
2916    
2917     4. Fix perltest so that it doesn't do unwanted things when fed an empty
2918     pattern. Perl treats empty patterns specially - it reuses the most recent
2919     pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2920     effect.
2921    
2922     5. The POSIX interface was broken in that it was just handing over the POSIX
2923     captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2924     required a bigger vector, with some working space on the end. This means that
2925     the POSIX wrapper now has to get and free some memory, and copy the results.
2926    
2927 nigel 43 6. Added some simple autoconf support, placing the test data and the
2928     documentation in separate directories, re-organizing some of the
2929     information files, and making it build pcre-config (a GNU standard). Also added
2930     libtool support for building PCRE as a shared library, which is now the
2931     default.
2932 nigel 41
2933 nigel 43 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2934     09 are not valid octal constants. Single digits will be used for minor values
2935     less than 10.
2936    
2937     8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2938     existing programs that set these in the POSIX interface can use PCRE without
2939     modification.
2940    
2941     9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2942     return all that pcre_info() returns, plus additional data. The pcre_info()
2943     function is retained for compatibility, but is considered to be obsolete.
2944    
2945     10. Added experimental recursion feature (?R) to handle one common case that
2946     Perl 5.6 will be able to do with (?p{...}).
2947    
2948     11. Added support for POSIX character classes like [:alpha:], which Perl is
2949     adopting.
2950    
2951    
2952 nigel 39 Version 2.08 31-Aug-99
2953     ----------------------
2954    
2955     1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2956     trying to match at the startoffset position, but instead was moving forward to
2957     the next newline as if a previous match had failed.
2958    
2959     2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2960     and could get into a loop if a null string was matched other than at the start
2961     of the subject.
2962    
2963     3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2964     be distinguished at compile time, and for completeness also added PCRE_DATE.
2965    
2966     5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2967     in GnuWin32 environments.
2968    
2969    
2970 nigel 37 Version 2.07 29-Jul-99
2971     ----------------------
2972    
2973     1. The documentation is now supplied in plain text form and HTML as well as in
2974     the form of man page sources.
2975    
2976     2. C++ compilers don't like assigning (void *) values to other pointer types.
2977     In particular this affects malloc(). Although there is no problem in Standard
2978     C, I've put in casts to keep C++ compilers happy.
2979    
2980     3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2981     should be (const char *).
2982    
2983     4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2984     be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2985     However, I haven't made this a standard facility. The documentation doesn't
2986     mention it, and the Makefile doesn't support it.
2987    
2988     5. The Makefile now contains an "install" target, with editable destinations at
2989     the top of the file. The pcretest program is not installed.
2990    
2991     6. pgrep -V now gives the PCRE version number and date.
2992    
2993     7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2994     causing the entire string to be ignored, instead of just the last character.
2995    
2996     8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2997     non-matching string, it can take a very, very long time, even for strings of
2998     quite modest length, because of the nested recursion. PCRE now does better in
2999     some of these cases. It does this by remembering the last required literal
3000     character in the pattern, and pre-searching the subject to ensure it is present
3001     before running the real match. In other words, it applies a heuristic to detect
3002     some types of certain failure quickly, and in the above example, if presented
3003     with a string that has no trailing " it gives "no match" very quickly.
3004    
3005     9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3006     other alternatives are tried instead.
3007    
3008    
3009 nigel 35 Version 2.06 09-Jun-99
3010     ----------------------
3011    
3012     1. Change pcretest's output for amount of store used to show just the code
3013     space, because the remainder (the data block) varies in size between 32-bit and
3014     64-bit systems.
3015    
3016     2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3017     start matching at. This allows lookbehinds to work when searching for multiple
3018     occurrences in a string.
3019    
3020     3. Added additional options to pcretest for testing multiple occurrences:
3021    
3022     /+ outputs the rest of the string that follows a match
3023     /g loops for multiple occurrences, using the new startoffset argument
3024     /G loops for multiple occurrences by passing an incremented pointer
3025    
3026     4. PCRE wasn't doing the "first character" optimization for patterns starting
3027     with \b or \B, though it was doing it for other lookbehind assertions. That is,
3028     it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3029     the letter 'x'. On long subject strings, this gives a significant speed-up.
3030    
3031    
3032 nigel 33 Version 2.05 21-Apr-99
3033     ----------------------
3034    
3035     1. Changed the type of magic_number from int to long int so that it works
3036     properly on 16-bit systems.
3037    
3038     2. Fixed a bug which caused patterns starting with .* not to work correctly
3039     when the subject string contained newline characters. PCRE was assuming
3040     anchoring for such patterns in all cases, which is not correct because .* will
3041     not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3042     DOTALL is set at top level; otherwise it knows that patterns starting with .*
3043     must be retried after every newline in the subject.
3044    
3045    
3046 nigel 31 Version 2.04 18-Feb-99
3047     ----------------------
3048    
3049     1. For parenthesized subpatterns with repeats whose minimum was zero, the
3050     computation of the store needed to hold the pattern was incorrect (too large).
3051     If such patterns were nested a few deep, this could multiply and become a real
3052     problem.
3053    
3054     2. Added /M option to pcretest to show the memory requirement of a specific
3055     pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3056    
3057     3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3058     compiled in such a way that the backtracking after subsequent failure was
3059     pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3060     ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3061    
3062    
3063 nigel 29 Version 2.03 02-Feb-99
3064     ----------------------
3065    
3066     1. Fixed typo and small mistake in man page.
3067    
3068 nigel 31 2. Added 4th condition (GPL supersedes if conflict) and created separate
3069     LICENCE file containing the conditions.
3070 nigel 29
3071     3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3072     Perl, that is the internal \ allows the delimiter to be included in the
3073     pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3074     the final delimiter, add \ to the end of the pattern (to test the error).
3075    
3076     4. Added the convenience functions for extracting substrings after a successful
3077     match. Updated pcretest to make it able to test these functions.
3078    
3079    
3080 nigel 27 Version 2.02 14-Jan-99
3081     ----------------------
3082    
3083     1. Initialized the working variables associated with each extraction so that
3084     their saving and restoring doesn't refer to uninitialized store.
3085    
3086     2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3087     compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3088     fix the problem.
3089    
3090     3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3091     calls, and wasn't printing the correct value for compiling calls. Increased the
3092     default value of LOOPREPEAT, and the number of significant figures in the
3093     times.
3094    
3095     4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3096    
3097     5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3098     a building problem on Windows NT with a FAT file system.
3099    
3100    
3101 nigel 25 Version 2.01 21-Oct-98
3102     ----------------------
3103    
3104     1. Changed the API for pcre_compile() to allow for the provision of a pointer
3105     to character tables built by pcre_maketables() in the current locale. If NULL
3106     is passed, the default tables are used.
3107    
3108    
3109 nigel 23 Version 2.00 24-Sep-98
3110 nigel 21 ----------------------
3111    
3112 nigel 23 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3113     it any more.
3114    
3115     2. Allow quantification of (?>) groups, and make it work correctly.
3116    
3117     3. The first character computation wasn't working for (?>) groups.
3118    
3119     4. Correct the implementation of \Z (it is permitted to match on the \n at the
3120     end of the subject) and add 5.005's \z, which really does match only at the
3121     very end of the subject.
3122    
3123     5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3124    
3125     6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3126     DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3127     localized options. All options to pcre_study() were also removed.
3128    
3129     7. Add other new features from 5.005:
3130    
3131     $(?<= positive lookbehind
3132     $(?<! negative lookbehind
3133     (?imsx-imsx) added the unsetting capability
3134     such a setting is global if at outer level; local otherwise
3135     (?imsx-imsx:) non-capturing groups with option setting
3136     (?(cond)re|re) conditional pattern matching
3137    
3138     A backreference to itself in a repeated group matches the previous
3139     captured string.
3140    
3141     8. General tidying up of studying (both automatic and via "study")
3142     consequential on the addition of new assertions.
3143    
3144     9. As in 5.005, unlimited repeated groups that could match an empty substring
3145     are no longer faulted at compile time. Instead, the loop is forcibly broken at
3146     runtime if any iteration does actually match an empty substring.
3147    
3148     10. Include the RunTest script in the distribution.
3149    
3150     11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3151     discrepancies, some of which were old and were also with respect to 5.004. They
3152     have now been fixed.
3153    
3154    
3155     Version 1.09 28-Apr-98
3156     ----------------------
3157    
3158 nigel 21 1. A negated single character class followed by a quantifier with a minimum
3159     value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3160     program crashes, or just wrong answers. This did not apply to negated classes
3161     containing more than one character, or to minima other than one.
3162    
3163    
3164 nigel 19 Version 1.08 27-Mar-98
3165     ----------------------
3166    
3167     1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3168    
3169     2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3170     latter must appear before anything that relies on it in the pattern.
3171    
3172    
3173 nigel 17 Version 1.07 16-Feb-98
3174     ----------------------
3175    
3176     1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3177     repeat of a potentially empty string).
3178    
3179    
3180 nigel 15 Version 1.06 23-Jan-98
3181     ----------------------
3182    
3183     1. Added Markus Oberhumer's little patches for C++.
3184    
3185     2. Literal strings longer than 255 characters were broken.
3186    
3187    
3188 nigel 13 Version 1.05 23-Dec-97
3189     ----------------------
3190    
3191     1. Negated character classes containing more than one character were failing if
3192     PCRE_CASELESS was set at run time.
3193    
3194    
3195 nigel 11 Version 1.04 19-Dec-97
3196     ----------------------
3197    
3198     1. Corrected the man page, where some "const" qualifiers had been omitted.
3199    
3200     2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3201     input syntax.
3202    
3203     3. Fixed memory leak which occurred when a regex with back references was
3204     matched with an offsets vector that wasn't big enough. The temporary memory
3205     that is used in this case wasn't being freed if the match failed.
3206    
3207     4. Tidied pcretest to ensure it frees memory that it gets.
3208    
3209     5. Temporary memory was being obtained in the case where the passed offsets
3210     vector was exactly big enough.
3211    
3212     6. Corrected definition of offsetof() from change 5 below.
3213    
3214     7. I had screwed up change 6 below and broken the rules for the use of
3215     setjmp(). Now fixed.
3216    
3217    
3218 nigel 9 Version 1.03 18-Dec-97
3219     ----------------------
3220    
3221     1. A erroneous regex with a missing opening parenthesis was correctly
3222     diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3223     on some systems.
3224    
3225     2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3226     it was reported that one broken compiler failed on the former because "code" is
3227     also an independent variable.
3228    
3229     3. The erroneous regex a[]b caused an array overrun reference.
3230    
3231     4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3232     fail on data ending with that character. (It was going on too far, and checking
3233     the next character, typically a binary zero.) This was specific to the
3234     optimized code for single-character negative classes.
3235    
3236     5. Added a contributed patch from the TIN world which does the following:
3237    
3238     + Add an undef for memmove, in case the the system defines a macro for it.
3239    
3240     + Add a definition of offsetof(), in case there isn't one. (I don't know
3241     the reason behind this - offsetof() is part of the ANSI standard - but
3242     it does no harm).
3243    
3244     + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3245     most of the places where whitespace preceded '#'. I have given up and
3246     allowed the remaining 2 cases to be at the margin.
3247    
3248     + Rename some variables in pcre to eliminate shadowing. This seems very
3249     pedantic, but does no harm, of course.
3250    
3251     6. Moved the call to setjmp() into its own function, to get rid of warnings
3252     from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3253    
3254     7. Constructs such as \d{8,} were compiling into the equivalent of
3255 nigel 11 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3256 nigel 9 outcome, but in this particular case used more store than had been allocated,
3257     which caused the bug to be discovered because it threw up an internal error.
3258    
3259     8. The debugging code in both pcre and pcretest for outputting the compiled
3260     form of a regex was going wrong in the case of back references followed by
3261     curly-bracketed repeats.
3262    
3263    
3264 nigel 7 Version 1.02 12-Dec-97
3265     ----------------------
3266    
3267     1. Typos in pcre.3 and comments in the source fixed.
3268    
3269     2. Applied a contributed patch to get rid of places where it used to remove
3270     'const' from variables, and fixed some signed/unsigned and uninitialized
3271     variable warnings.
3272    
3273     3. Added the "runtest" target to Makefile.
3274    
3275     4. Set default compiler flag to -O2 rather than just -O.
3276    
3277    
3278 nigel 5 Version 1.01 19-Nov-97
3279     ----------------------
3280    
3281     1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3282     like /([ab]*)*/, that is, for classes with more than one character in them.
3283    
3284     2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3285     as /((?>a*))*/ (a PCRE_EXTRA facility).
3286    
3287    
3288     Version 1.00 18-Nov-97
3289     ----------------------
3290    
3291     1. Added compile-time macros to support systems such as SunOS4 which don't have
3292     memmove() or strerror() but have other things that can be used instead.
3293    
3294     2. Arranged that "make clean" removes the executables.
3295    
3296    
3297 nigel 3 Version 0.99 27-Oct-97
3298     ----------------------
3299    
3300     1. Fixed bug in code for optimizing classes with only one character. It was
3301     initializing a 32-byte map regardless, which could cause it to run off the end
3302     of the memory it had got.
3303    
3304     2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3305    
3306    
3307     Version 0.98 22-Oct-97
3308     ----------------------
3309    
3310     1. Fixed bug in code for handling temporary memory usage when there are more
3311     back references than supplied space in the ovector. This could cause segfaults.
3312    
3313    
3314     Version 0.97 21-Oct-97
3315     ----------------------
3316    
3317     1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3318    
3319     2. Optimized negated single characters not to use a bit map.
3320    
3321     3. Brought error texts together as macro definitions; clarified some of them;
3322     fixed one that was wrong - it said "range out of order" when it meant "invalid
3323     escape sequence".
3324    
3325     4. Changed some char * arguments to const char *.
3326    
3327     5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3328    
3329     6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3330     pcretest.
3331    
3332    
3333     Version 0.96 16-Oct-97
3334     ----------------------
3335    
3336     1. Added a simple "pgrep" utility to the distribution.
3337    
3338     2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3339     unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3340     where "ddd" means "one or more decimal digits".
3341    
3342     3. Fixed serious bug. If a pattern had a back reference, but the call to
3343     pcre_exec() didn't supply a large enough ovector to record the related
3344     identifying subpattern, the match always failed. PCRE now remembers the number
3345     of the largest back reference, and gets some temporary memory in which to save
3346     the offsets during matching if necessary, in order to ensure that
3347     backreferences always work.
3348    
3349     4. Increased the compatibility with Perl in a number of ways:
3350    
3351     (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3352     to request this handling. The option can be set at compile or exec time.
3353    
3354     (b) $ matches before a terminating newline by default; an option
3355     PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3356     mode). The option can be set at compile or exec time.
3357    
3358     (c) The handling of \ followed by a digit other than 0 is now supposed to be
3359     the same as Perl's. If the decimal number it represents is less than 10
3360     or there aren't that many previous left capturing parentheses, an octal
3361     escape is read. Inside a character class, it's always an octal escape,
3362     even if it is a single digit.
3363    
3364     (d) An escaped but undefined alphabetic character is taken as a literal,
3365     unless PCRE_EXTRA is set. Currently this just reserves the remaining
3366     escapes.
3367    
3368     (e) {0} is now permitted. (The previous item is removed from the compiled
3369     pattern).
3370    
3371     5. Changed all the names of code files so that the basic parts are no longer
3372     than 10 characters, and abolished the teeny "globals.c" file.
3373    
3374     6. Changed the handling of character classes; they are now done with a 32-byte
3375     bit map always.
3376    
3377     7. Added the -d and /D options to pcretest to make it possible to look at the
3378     internals of compilation without having to recompile pcre.
3379    
3380    
3381     Version 0.95 23-Sep-97
3382     ----------------------
3383    
3384     1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3385     \x20 at the start of a run of normal characters. These were being treated as
3386     real characters, instead of the source characters being re-checked.
3387    
3388    
3389     Version 0.94 18-Sep-97
3390     ----------------------
3391    
3392     1. The functions are now thread-safe, with the caveat that the global variables
3393     containing pointers to malloc() and free() or alternative functions are the
3394     same for all threads.
3395    
3396     2. Get pcre_study() to generate a bitmap of initial characters for non-
3397     anchored patterns when this is possible, and use it if passed to pcre_exec().
3398    
3399    
3400     Version 0.93 15-Sep-97
3401     ----------------------
3402    
3403     1. /(b)|(:+)/ was computing an incorrect first character.
3404    
3405     2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3406     but not actually doing anything yet.
3407    
3408     3. Treat "-" characters in classes that cannot be part of ranges as literals,
3409     as Perl does (e.g. [-az] or [az-]).
3410    
3411     4. Set the anchored flag if a branch starts with .* or .*? because that tests
3412     all possible positions.
3413    
3414     5. Split up into different modules to avoid including unneeded functions in a
3415     compiled binary. However, compile and exec are still in one module. The "study"
3416     function is split off.
3417    
3418     6. The character tables are now in a separate module whose source is generated
3419     by an auxiliary program - but can then be edited by hand if required. There are
3420     now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3421     toupper() in the code.
3422    
3423     7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3424     make them global. Abolish the function for setting them, as the caller can now
3425     set them directly.
3426    
3427    
3428     Version 0.92 11-Sep-97
3429     ----------------------
3430    
3431     1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3432     (e.g. /a{1,3}/) was broken (I mis-optimized it).
3433    
3434     2. Caseless matching was not working in character classes if the characters in
3435     the pattern were in upper case.
3436    
3437     3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3438    
3439     4. Make PCRE_ANCHORED public and accept as a compile option.
3440    
3441     5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3442     PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3443     pass them.
3444    
3445     6. Give an error if bad option bits passed at compile or run time.
3446    
3447     7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3448     pcretest to cause it to pass that flag.
3449    
3450     8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3451     options, and the first character, if set.
3452    
3453     9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3454    
3455    
3456     Version 0.91 10-Sep-97
3457     ----------------------
3458    
3459     1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3460     match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3461    
3462     2. PCRE was looping on encountering an indefinitely repeated back reference to
3463     a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3464     Perl does - treats the match as successful.
3465    
3466     ****

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