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

Diff of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

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

Legend:
Removed from v.33  
changed lines
  Added in v.282

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12