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