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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 385 - (hide annotations) (download)
Sun Mar 8 16:56:58 2009 UTC (5 years, 6 months ago) by ph10
File size: 162256 byte(s)
Document named subpattern behaviour in the presence of duplicate numbers.

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

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