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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 342 - (hide annotations) (download)
Sun Apr 20 17:10:13 2008 UTC (6 years, 8 months ago) by ph10
File size: 155483 byte(s)
Slight performance improvement by using the new OP_ALLANY opcode for cases of 
the metacharacter "." when DOTALL is set. Also, some tidies consequent upon its 
invention.

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

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