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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 381 - (hide annotations) (download)
Tue Mar 3 16:08:23 2009 UTC (5 years, 5 months ago) by ph10
File size: 162010 byte(s)
Fix bug with (?(?=.*b)b|^) thinking it must match at start of line; also fix 
bug causing a crash when auto-callout is used with a conditional assertion.

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

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