"Fossies" - the Fresh Open Source Software Archive

Member "tin-2.6.1/pcre/ChangeLog" (23 Aug 2021, 118928 Bytes) of package /linux/misc/tin-2.6.1.tar.xz:

As a special service "Fossies" has tried to format the requested text file into HTML format (style: standard) with prefixed line numbers. Alternatively you can here view or download the uninterpreted source code file.

    1 ChangeLog for PCRE
    2 ------------------
    4 Version 7.0 19-Dec-06
    5 ---------------------
    7  1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
    8     moving to gcc 4.1.1.
   10  2. The -S option for pcretest uses setrlimit(); I had omitted to #include
   11     sys/time.h, which is documented as needed for this function. It doesn't
   12     seem to matter on Linux, but it showed up on some releases of OS X.
   14  3. It seems that there are systems where bytes whose values are greater than
   15     127 match isprint() in the "C" locale. The "C" locale should be the
   16     default when a C program starts up. In most systems, only ASCII printing
   17     characters match isprint(). This difference caused the output from pcretest
   18     to vary, making some of the tests fail. I have changed pcretest so that:
   20     (a) When it is outputting text in the compiled version of a pattern, bytes
   21         other than 32-126 are always shown as hex escapes.
   23     (b) When it is outputting text that is a matched part of a subject string,
   24         it does the same, unless a different locale has been set for the match
   25         (using the /L modifier). In this case, it uses isprint() to decide.
   27  4. Fixed a major bug that caused incorrect computation of the amount of memory
   28     required for a compiled pattern when options that changed within the
   29     pattern affected the logic of the preliminary scan that determines the
   30     length. The relevant options are -x, and -i in UTF-8 mode. The result was
   31     that the computed length was too small. The symptoms of this bug were
   32     either the PCRE error "internal error: code overflow" from pcre_compile(),
   33     or a glibc crash with a message such as "pcretest: free(): invalid next
   34     size (fast)". Examples of patterns that provoked this bug (shown in
   35     pcretest format) are:
   37       /(?-x: )/x
   38       /(?x)(?-x: \s*#\s*)/
   39       /((?i)[\x{c0}])/8
   40       /(?i:[\x{c0}])/8
   42     HOWEVER: Change 17 below makes this fix obsolete as the memory computation
   43     is now done differently.
   45  5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
   46     wrapper classes; (b) implement a new function in the C++ scanner that is
   47     more efficient than the old way of doing things because it avoids levels of
   48     recursion in the regex matching; (c) add a paragraph to the documentation
   49     for the FullMatch() function.
   51  6. The escape sequence \n was being treated as whatever was defined as
   52     "newline". Not only was this contrary to the documentation, which states
   53     that \n is character 10 (hex 0A), but it also went horribly wrong when
   54     "newline" was defined as CRLF. This has been fixed.
   56  7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
   57     was being set to -1 for the "end of line" case (supposedly a value that no
   58     character can have). Though this value is never used (the check for end of
   59     line is "zero bytes in current character"), it caused compiler complaints.
   60     I've changed it to 0xffffffff.
   62  8. In pcre_version.c, the version string was being built by a sequence of
   63     C macros that, in the event of PCRE_PRERELEASE being defined as an empty
   64     string (as it is for production releases) called a macro with an empty
   65     argument. The C standard says the result of this is undefined. The gcc
   66     compiler treats it as an empty string (which was what was wanted) but it is
   67     reported that Visual C gives an error. The source has been hacked around to
   68     avoid this problem.
   70  9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
   71     builds of pcretest, and changed the call to _setmode() to use _O_BINARY
   72     instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
   73     of them did).
   75 10. Originally, pcretest opened its input and output without "b"; then I was
   76     told that "b" was needed in some environments, so it was added for release
   77     5.0 to both the input and output. (It makes no difference on Unix-like
   78     systems.) Later I was told that it is wrong for the input on Windows. I've
   79     now abstracted the modes into two macros, to make it easier to fiddle with
   80     them, and removed "b" from the input mode under Windows.
   82 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
   84 12. Added -help and --help to pcretest as an official way of being reminded
   85     of the options.
   87 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
   88     and pcrecpp.cc because they annoy compilers at high warning levels.
   90 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
   92 15. Fixed an occurrence of == in configure.ac that should have been = (shell
   93     scripts are not C programs :-) and which was not noticed because it works
   94     on Linux.
   96 16. pcretest is supposed to handle any length of pattern and data line (as one
   97     line or as a continued sequence of lines) by extending its input buffer if
   98     necessary. This feature was broken for very long pattern lines, leading to
   99     a string of junk being passed to pcre_compile() if the pattern was longer
  100     than about 50K.
  102 17. I have done a major re-factoring of the way pcre_compile() computes the
  103     amount of memory needed for a compiled pattern. Previously, there was code
  104     that made a preliminary scan of the pattern in order to do this. That was
  105     OK when PCRE was new, but as the facilities have expanded, it has become
  106     harder and harder to keep it in step with the real compile phase, and there
  107     have been a number of bugs (see for example, 4 above). I have now found a
  108     cunning way of running the real compile function in a "fake" mode that
  109     enables it to compute how much memory it would need, while actually only
  110     ever using a few hundred bytes of working memory and without too many
  111     tests of the mode. This should make future maintenance and development
  112     easier. A side effect of this work is that the limit of 200 on the nesting
  113     depth of parentheses has been removed (though this was never a serious
  114     limitation, I suspect). However, there is a downside: pcre_compile() now
  115     runs more slowly than before (30% or more, depending on the pattern). I
  116     hope this isn't a big issue. There is no effect on runtime performance.
  118 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
  119     newline (only possible for the last line of a file) and it was a
  120     pattern that set a locale (followed by /Lsomething), pcretest crashed.
  122 19. Added additional timing features to pcretest. (1) The -tm option now times
  123     matching only, not compiling. (2) Both -t and -tm can be followed, as a
  124     separate command line item, by a number that specifies the number of
  125     repeats to use when timing. The default is 50000; this gives better
  126     precision, but takes uncomfortably long for very large patterns.
  128 20. Extended pcre_study() to be more clever in cases where a branch of a
  129     subpattern has no definite first character. For example, (a*|b*)[cd] would
  130     previously give no result from pcre_study(). Now it recognizes that the
  131     first character must be a, b, c, or d.
  133 21. There was an incorrect error "recursive call could loop indefinitely" if
  134     a subpattern (or the entire pattern) that was being tested for matching an
  135     empty string contained only one non-empty item after a nested subpattern.
  136     For example, the pattern (?>\x{100}*)\d(?R) provoked this error
  137     incorrectly, because the \d was being skipped in the check.
  139 22. The pcretest program now has a new pattern option /B and a command line
  140     option -b, which is equivalent to adding /B to every pattern. This causes
  141     it to show the compiled bytecode, without the additional information that
  142     -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
  143     is the same as /B/I).
  145 23. A new optimization is now able automatically to treat some sequences such
  146     as a*b as a*+b. More specifically, if something simple (such as a character
  147     or a simple class like \d) has an unlimited quantifier, and is followed by
  148     something that cannot possibly match the quantified thing, the quantifier
  149     is automatically "possessified".
  151 24. A recursive reference to a subpattern whose number was greater than 39
  152     went wrong under certain circumstances in UTF-8 mode. This bug could also
  153     have affected the operation of pcre_study().
  155 25. Realized that a little bit of performance could be had by replacing
  156     (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
  158 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
  160 27. Possessive quantifiers such as a++ were previously implemented by turning
  161     them into atomic groups such as ($>a+). Now they have their own opcodes,
  162     which improves performance. This includes the automatically created ones
  163     from 23 above.
  165 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
  166     lookahead was broken if it was not anchored. PCRE was mistakenly expecting
  167     the first matched character to be a colon. This applied both to named and
  168     numbered groups.
  170 29. The ucpinternal.h header file was missing its idempotency #ifdef.
  172 30. I was sent a "project" file called libpcre.a.dev which I understand makes
  173     building PCRE on Windows easier, so I have included it in the distribution.
  175 31. There is now a check in pcretest against a ridiculously large number being
  176     returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
  177     loop, the loop is abandoned.
  179 32. Forward references to subpatterns in conditions such as (?(2)...) where
  180     subpattern 2 is defined later cause pcre_compile() to search forwards in
  181     the pattern for the relevant set of parentheses. This search went wrong
  182     when there were unescaped parentheses in a character class, parentheses
  183     escaped with \Q...\E, or parentheses in a #-comment in /x mode.
  185 33. "Subroutine" calls and backreferences were previously restricted to
  186     referencing subpatterns earlier in the regex. This restriction has now
  187     been removed.
  189 34. Added a number of extra features that are going to be in Perl 5.10. On the
  190     whole, these are just syntactic alternatives for features that PCRE had
  191     previously implemented using the Python syntax or my own invention. The
  192     other formats are all retained for compatibility.
  194     (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
  195         as (?P<name>...). The new forms, as well as being in Perl 5.10, are
  196         also .NET compatible.
  198     (b) A recursion or subroutine call to a named group can now be defined as
  199         (?&name) as well as (?P>name).
  201     (c) A backreference to a named group can now be defined as \k<name> or
  202         \k'name' as well as (?P=name). The new forms, as well as being in Perl
  203         5.10, are also .NET compatible.
  205     (d) A conditional reference to a named group can now use the syntax
  206         (?(<name>) or (?('name') as well as (?(name).
  208     (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
  209         groups (named and numbered) that are never evaluated inline, but can be
  210         called as "subroutines" from elsewhere. In effect, the DEFINE condition
  211         is always false. There may be only one alternative in such a group.
  213     (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
  214         as the simple (?(R). The condition is true only if the most recent
  215         recursion is that of the given number or name. It does not search out
  216         through the entire recursion stack.
  218     (g) The escape \gN or \g{N} has been added, where N is a positive or
  219         negative number, specifying an absolute or relative reference.
  221 35. Tidied to get rid of some further signed/unsigned compiler warnings and
  222     some "unreachable code" warnings.
  224 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
  225     things, this adds five new scripts.
  227 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
  228     There were also incompatibilities regarding the handling of \Q..\E inside
  229     character classes, for example with patterns like [\Qa\E-\Qz\E] where the
  230     hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
  232 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
  233     matches an empty string, and forcibly breaks the loop. There were bugs in
  234     this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
  235     against  aaaa  the result was just "a" rather than "aaaa", for example. Two
  236     separate and independent bugs (that affected different cases) have been
  237     fixed.
  239 39. Refactored the code to abolish the use of different opcodes for small
  240     capturing bracket numbers. This is a tidy that I avoided doing when I
  241     removed the limit on the number of capturing brackets for 3.5 back in 2001.
  242     The new approach is not only tidier, it makes it possible to reduce the
  243     memory needed to fix the previous bug (38).
  245 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
  246     sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
  247     processing dot, circumflex, or dollar metacharacters, or #-comments in /x
  248     mode.
  250 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
  251     report.
  253 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
  254     copy construction and assignment in the C++ wrapper.
  256 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
  257     couple of bugs that could have given wrong results in the "--newline=crlf"
  258     case.
  260 44. Added a number of casts and did some reorganization of signed/unsigned int
  261     variables following suggestions from Dair Grant. Also renamed the variable
  262     "this" as "item" because it is a C++ keyword.
  264 45. Arranged for dftables to add
  266       #include "pcre_internal.h"
  268     to pcre_chartables.c because without it, gcc 4.x may remove the array
  269     definition from the final binary if PCRE is built into a static library and
  270     dead code stripping is activated.
  272 46. For an unanchored pattern, if a match attempt fails at the start of a
  273     newline sequence, and the newline setting is CRLF or ANY, and the next two
  274     characters are CRLF, advance by two characters instead of one.
  277 Version 6.7 04-Jul-06
  278 ---------------------
  280  1. In order to handle tests when input lines are enormously long, pcretest has
  281     been re-factored so that it automatically extends its buffers when
  282     necessary. The code is crude, but this _is_ just a test program. The
  283     default size has been increased from 32K to 50K.
  285  2. The code in pcre_study() was using the value of the re argument before
  286     testing it for NULL. (Of course, in any sensible call of the function, it
  287     won't be NULL.)
  289  3. The memmove() emulation function in pcre_internal.h, which is used on
  290     systems that lack both memmove() and bcopy() - that is, hardly ever -
  291     was missing a "static" storage class specifier.
  293  4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
  294     containing an extended class (one that cannot be represented by a bitmap
  295     because it contains high-valued characters or Unicode property items, e.g.
  296     [\pZ]). Almost always one would set UTF-8 mode when processing such a
  297     pattern, but PCRE should not loop if you do not (it no longer does).
  298     [Detail: two cases were found: (a) a repeated subpattern containing an
  299     extended class; (b) a recursive reference to a subpattern that followed a
  300     previous extended class. It wasn't skipping over the extended class
  301     correctly when UTF-8 mode was not set.]
  303  5. A negated single-character class was not being recognized as fixed-length
  304     in lookbehind assertions such as (?<=[^f]), leading to an incorrect
  305     compile error "lookbehind assertion is not fixed length".
  307  6. The RunPerlTest auxiliary script was showing an unexpected difference
  308     between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
  309     write a Perl script that can interpret lines of an input file either as
  310     byte characters or as UTF-8, which is what "perltest" was being required to
  311     do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
  312     can't do is switch easily at run time between having the "use utf8;" pragma
  313     or not. In the end, I fudged it by using the RunPerlTest script to insert
  314     "use utf8;" explicitly for the UTF-8 tests.
  316  7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
  317     the end of the subject string, contrary to the documentation and to what
  318     Perl does. This was true of both matching functions. Now it matches only at
  319     the start of the subject and immediately after *internal* newlines.
  321  8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
  322     a pointer to an int instead of a pointer to an unsigned long int. This
  323     caused problems on 64-bit systems.
  325  9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
  326     instance of the 'standard' template library not being so standard".
  328 10. There was no check on the number of named subpatterns nor the maximum
  329     length of a subpattern name. The product of these values is used to compute
  330     the size of the memory block for a compiled pattern. By supplying a very
  331     long subpattern name and a large number of named subpatterns, the size
  332     computation could be caused to overflow. This is now prevented by limiting
  333     the length of names to 32 characters, and the number of named subpatterns
  334     to 10,000.
  336 11. Subpatterns that are repeated with specific counts have to be replicated in
  337     the compiled pattern. The size of memory for this was computed from the
  338     length of the subpattern and the repeat count. The latter is limited to
  339     65535, but there was no limit on the former, meaning that integer overflow
  340     could in principle occur. The compiled length of a repeated subpattern is
  341     now limited to 30,000 bytes in order to prevent this.
  343 12. Added the optional facility to have named substrings with the same name.
  345 13. Added the ability to use a named substring as a condition, using the
  346     Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
  347     are numbers (not recommended). Forward references are permitted.
  349 14. Added forward references in named backreferences (if you see what I mean).
  351 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
  352     pattern could run off the end of the subject. For example, the pattern
  353     "(?s)(.{1,5})"8 did this with the subject "ab".
  355 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
  356     PCRE_CASELESS was set when matching characters that were quantified with ?
  357     or *.
  359 17. A character class other than a single negated character that had a minimum
  360     but no maximum quantifier - for example [ab]{6,} - was not handled
  361     correctly by pce_dfa_exec(). It would match only one character.
  363 18. A valid (though odd) pattern that looked like a POSIX character
  364     class but used an invalid character after [ (for example [[,abc,]]) caused
  365     pcre_compile() to give the error "Failed: internal error: code overflow" or
  366     in some cases to crash with a glibc free() error. This could even happen if
  367     the pattern terminated after [[ but there just happened to be a sequence of
  368     letters, a binary zero, and a closing ] in the memory that followed.
  370 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
  371     over the years. Originally (before any Unicode support), just the bottom 8
  372     bits were taken. Thus, for example, \500 really meant \100. Nowadays the
  373     output from "man perlunicode" includes this:
  375       The regular expression compiler produces polymorphic opcodes.  That
  376       is, the pattern adapts to the data and automatically switches to
  377       the Unicode character scheme when presented with Unicode data--or
  378       instead uses a traditional byte scheme when presented with byte
  379       data.
  381     Sadly, a wide octal escape does not cause a switch, and in a string with
  382     no other multibyte characters, these octal escapes are treated as before.
  383     Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
  384     /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
  385     Unicode string.
  387     I have not perpetrated such confusion in PCRE. Up till now, it took just
  388     the bottom 8 bits, as in old Perl. I have now made octal escapes with
  389     values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
  390     translate to the appropriate multibyte character.
  392 29. Applied some refactoring to reduce the number of warnings from Microsoft
  393     and Borland compilers. This has included removing the fudge introduced
  394     seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
  395     a warning about an unused variable.
  397 21. PCRE has not included VT (character 0x0b) in the set of whitespace
  398     characters since release 4.0, because Perl (from release 5.004) does not.
  399     [Or at least, is documented not to: some releases seem to be in conflict
  400     with the documentation.] However, when a pattern was studied with
  401     pcre_study() and all its branches started with \s, PCRE still included VT
  402     as a possible starting character. Of course, this did no harm; it just
  403     caused an unnecessary match attempt.
  405 22. Removed a now-redundant internal flag bit that recorded the fact that case
  406     dependency changed within the pattern. This was once needed for "required
  407     byte" processing, but is no longer used. This recovers a now-scarce options
  408     bit. Also moved the least significant internal flag bit to the most-
  409     significant bit of the word, which was not previously used (hangover from
  410     the days when it was an int rather than a uint) to free up another bit for
  411     the future.
  413 23. Added support for CRLF line endings as well as CR and LF. As well as the
  414     default being selectable at build time, it can now be changed at runtime
  415     via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
  416     specify that it is scanning data with non-default line endings.
  418 24. Changed the definition of CXXLINK to make it agree with the definition of
  419     LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
  421 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
  422     recursions. This makes a big different to stack usage for some patterns.
  424 26. If a subpattern containing a named recursion or subroutine reference such
  425     as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
  426     the space required for the compiled pattern went wrong and gave too small a
  427     value. Depending on the environment, this could lead to "Failed: internal
  428     error: code overflow at offset 49" or "glibc detected double free or
  429     corruption" errors.
  431 27. Applied patches from Google (a) to support the new newline modes and (b) to
  432     advance over multibyte UTF-8 characters in GlobalReplace.
  434 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
  435     difference for some implementation of PCRE in some Windows version.
  437 29. Added some extra testing facilities to pcretest:
  439     \q<number>   in a data line sets the "match limit" value
  440     \Q<number>   in a data line sets the "match recursion limt" value
  441     -S <number>  sets the stack size, where <number> is in megabytes
  443     The -S option isn't available for Windows.
  446 Version 6.6 06-Feb-06
  447 ---------------------
  449  1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
  450     in pcreposix.h. I have copied the definition from pcre.h.
  452  2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
  453     because pcre.h is no longer a built file.
  455  3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
  456     not normally included in the compiled code.
  459 Version 6.5 01-Feb-06
  460 ---------------------
  462  1. When using the partial match feature with pcre_dfa_exec(), it was not
  463     anchoring the second and subsequent partial matches at the new starting
  464     point. This could lead to incorrect results. For example, with the pattern
  465     /1234/, partially matching against "123" and then "a4" gave a match.
  467  2. Changes to pcregrep:
  469     (a) All non-match returns from pcre_exec() were being treated as failures
  470         to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
  471         error message is output. Some extra information is given for the
  473         probably the only errors that are likely to be caused by users (by
  474         specifying a regex that has nested indefinite repeats, for instance).
  475         If there are more than 20 of these errors, pcregrep is abandoned.
  477     (b) A binary zero was treated as data while matching, but terminated the
  478         output line if it was written out. This has been fixed: binary zeroes
  479         are now no different to any other data bytes.
  481     (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
  482         used to set a locale for matching. The --locale=xxxx long option has
  483         been added (no short equivalent) to specify a locale explicitly on the
  484         pcregrep command, overriding the environment variables.
  486     (d) When -B was used with -n, some line numbers in the output were one less
  487         than they should have been.
  489     (e) Added the -o (--only-matching) option.
  491     (f) If -A or -C was used with -c (count only), some lines of context were
  492         accidentally printed for the final match.
  494     (g) Added the -H (--with-filename) option.
  496     (h) The combination of options -rh failed to suppress file names for files
  497         that were found from directory arguments.
  499     (i) Added the -D (--devices) and -d (--directories) options.
  501     (j) Added the -F (--fixed-strings) option.
  503     (k) Allow "-" to be used as a file name for -f as well as for a data file.
  505     (l) Added the --colo(u)r option.
  507     (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
  508         is not present by default.
  510  3. A nasty bug was discovered in the handling of recursive patterns, that is,
  511     items such as (?R) or (?1), when the recursion could match a number of
  512     alternatives. If it matched one of the alternatives, but subsequently,
  513     outside the recursion, there was a failure, the code tried to back up into
  514     the recursion. However, because of the way PCRE is implemented, this is not
  515     possible, and the result was an incorrect result from the match.
  517     In order to prevent this happening, the specification of recursion has
  518     been changed so that all such subpatterns are automatically treated as
  519     atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
  521  4. I had overlooked the fact that, in some locales, there are characters for
  522     which isalpha() is true but neither isupper() nor islower() are true. In
  523     the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
  524     and ordfeminine) are like this. This affected the treatment of \w and \W
  525     when they appeared in character classes, but not when they appeared outside
  526     a character class. The bit map for "word" characters is now created
  527     separately from the results of isalnum() instead of just taking it from the
  528     upper, lower, and digit maps. (Plus the underscore character, of course.)
  530  5. The above bug also affected the handling of POSIX character classes such as
  531     [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
  532     permanent tables. Instead, the bit maps for such a class were previously
  533     created as the appropriate unions of the upper, lower, and digit bitmaps.
  534     Now they are created by subtraction from the [[:word:]] class, which has
  535     its own bitmap.
  537  6. The [[:blank:]] character class matches horizontal, but not vertical space.
  538     It is created by subtracting the vertical space characters (\x09, \x0a,
  539     \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
  540     subtraction was done in the overall bitmap for a character class, meaning
  541     that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
  542     be recognized. This bug has been fixed.
  544  7. Patches from the folks at Google:
  546       (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
  547       real life, but is still worth protecting against".
  549       (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
  550       regular expressions".
  552       (c) pcre_scanner.cc: avoid use of std::count() because not all systems
  553       have it.
  555       (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
  556       "configure" and the latter not, in order to fix a problem somebody had
  557       with compiling the Arg class on HP-UX.
  559       (e) Improve the error-handling of the C++ wrapper a little bit.
  561       (f) New tests for checking recursion limiting.
  563  8. The pcre_memmove() function, which is used only if the environment does not
  564     have a standard memmove() function (and is therefore rarely compiled),
  565     contained two bugs: (a) use of int instead of size_t, and (b) it was not
  566     returning a result (though PCRE never actually uses the result).
  568  9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
  569     large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
  570     returned instead of calling malloc() with an overflowing number that would
  571     most likely cause subsequent chaos.
  573 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
  575 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
  576     with this option is matched, the nmatch and pmatch options of regexec() are
  577     ignored.
  579 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
  580     provided in case anyone wants to the the POSIX interface with UTF-8
  581     strings.
  583 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
  584     C++ linking (needed for some HP-UX environments).
  586 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
  587     (unused parameter) and in the pcre_printint() function (omitted "default"
  588     switch label when the default is to do nothing).
  590 15. Added some code to make it possible, when PCRE is compiled as a C++
  591     library, to replace subject pointers for pcre_exec() with a smart pointer
  592     class, thus making it possible to process discontinuous strings.
  594 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
  595     much the same function. They were added by different people who were trying
  596     to make PCRE easy to compile on non-Unix systems. It has been suggested
  597     that PCRE_EXPORT be abolished now that there is more automatic apparatus
  598     for compiling on Windows systems. I have therefore replaced it with
  599     PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
  600     defaults to "extern" for C or "extern C" for C++, which works fine on
  601     Unix-like systems. It is now possible to override the value of PCRE_DATA_
  602     SCOPE with something explicit in config.h. In addition:
  604     (a) pcreposix.h still had just "extern" instead of either of these macros;
  605         I have replaced it with PCRE_DATA_SCOPE.
  607     (b) Functions such as _pcre_xclass(), which are internal to the library,
  608         but external in the C sense, all had PCRE_EXPORT in their definitions.
  609         This is apparently wrong for the Windows case, so I have removed it.
  610         (It makes no difference on Unix-like systems.)
  612 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
  613     of recursive calls to match(). This is different to MATCH_LIMIT because
  614     that limits the total number of calls to match(), not all of which increase
  615     the depth of recursion. Limiting the recursion depth limits the amount of
  616     stack (or heap if NO_RECURSE is set) that is used. The default can be set
  617     when PCRE is compiled, and changed at run time. A patch from Google adds
  618     this functionality to the C++ interface.
  620 18. Changes to the handling of Unicode character properties:
  622     (a) Updated the table to Unicode 4.1.0.
  624     (b) Recognize characters that are not in the table as "Cn" (undefined).
  626     (c) I revised the way the table is implemented to a much improved format
  627         which includes recognition of ranges. It now supports the ranges that
  628         are defined in UnicodeData.txt, and it also amalgamates other
  629         characters into ranges. This has reduced the number of entries in the
  630         table from around 16,000 to around 3,000, thus reducing its size
  631         considerably. I realized I did not need to use a tree structure after
  632         all - a binary chop search is just as efficient. Having reduced the
  633         number of entries, I extended their size from 6 bytes to 8 bytes to
  634         allow for more data.
  636     (d) Added support for Unicode script names via properties such as \p{Han}.
  638 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
  639     matching that character.
  641 20. When matching a repeated Unicode property with a minimum greater than zero,
  642     (for example \pL{2,}), PCRE could look past the end of the subject if it
  643     reached it while seeking the minimum number of characters. This could
  644     happen only if some of the characters were more than one byte long, because
  645     there is a check for at least the minimum number of bytes.
  647 21. Refactored the implementation of \p and \P so as to be more general, to
  648     allow for more different types of property in future. This has changed the
  649     compiled form incompatibly. Anybody with saved compiled patterns that use
  650     \p or \P will have to recompile them.
  652 22. Added "Any" and "L&" to the supported property types.
  654 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
  655     but give a compile time error if the value is greater than 0xff.
  657 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
  658     accidentally not being installed or uninstalled.
  660 25. The pcre.h file was built from pcre.h.in, but the only changes that were
  661     made were to insert the current release number. This seemed silly, because
  662     it made things harder for people building PCRE on systems that don't run
  663     "configure". I have turned pcre.h into a distributed file, no longer built
  664     by "configure", with the version identification directly included. There is
  665     no longer a pcre.h.in file.
  667     However, this change necessitated a change to the pcre-config script as
  668     well. It is built from pcre-config.in, and one of the substitutions was the
  669     release number. I have updated configure.ac so that ./configure now finds
  670     the release number by grepping pcre.h.
  672 26. Added the ability to run the tests under valgrind.
  675 Version 6.4 05-Sep-05
  676 ---------------------
  678  1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
  679     "--" to be printed when multiple files were scanned, even when none of the
  680     -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
  681     consider it to be a bug, and have restored the previous behaviour.
  683  2. A couple of code tidies to get rid of compiler warnings.
  685  3. The pcretest program used to cheat by referring to symbols in the library
  686     whose names begin with _pcre_. These are internal symbols that are not
  687     really supposed to be visible externally, and in some environments it is
  688     possible to suppress them. The cheating is now confined to including
  689     certain files from the library's source, which is a bit cleaner.
  691  4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
  692     file's purpose clearer.
  694  5. Reorganized pcre_ucp_findchar().
  697 Version 6.3 15-Aug-05
  698 ---------------------
  700  1. The file libpcre.pc.in did not have general read permission in the tarball.
  702  2. There were some problems when building without C++ support:
  704     (a) If C++ support was not built, "make install" and "make test" still
  705         tried to test it.
  707     (b) There were problems when the value of CXX was explicitly set. Some
  708         changes have been made to try to fix these, and ...
  710     (c) --disable-cpp can now be used to explicitly disable C++ support.
  712     (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
  713         backslash in a target when C++ was disabled. This confuses some
  714         versions of "make", apparently. Using an intermediate variable solves
  715         this. (Same for CPP_LOBJ.)
  717  3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
  718     (non-Windows) now includes $(CFLAGS) because these flags are sometimes
  719     necessary on certain architectures.
  721  4. Added a setting of -export-symbols-regex to the link command to remove
  722     those symbols that are exported in the C sense, but actually are local
  723     within the library, and not documented. Their names all begin with
  724     "_pcre_". This is not a perfect job, because (a) we have to except some
  725     symbols that pcretest ("illegally") uses, and (b) the facility isn't always
  726     available (and never for static libraries). I have made a note to try to
  727     find a way round (a) in the future.
  730 Version 6.2 01-Aug-05
  731 ---------------------
  733  1. There was no test for integer overflow of quantifier values. A construction
  734     such as {1111111111111111} would give undefined results. What is worse, if
  735     a minimum quantifier for a parenthesized subpattern overflowed and became
  736     negative, the calculation of the memory size went wrong. This could have
  737     led to memory overwriting.
  739  2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
  741  3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
  742     operating environments where this matters.
  744  4. Applied Giuseppe Maxia's patch to add additional features for controlling
  745     PCRE options from within the C++ wrapper.
  747  5. Named capturing subpatterns were not being correctly counted when a pattern
  748     was compiled. This caused two problems: (a) If there were more than 100
  749     such subpatterns, the calculation of the memory needed for the whole
  750     compiled pattern went wrong, leading to an overflow error. (b) Numerical
  751     back references of the form \12, where the number was greater than 9, were
  752     not recognized as back references, even though there were sufficient
  753     previous subpatterns.
  755  6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
  756     versions of gcc, e.g. 2.95.4.
  759 Version 6.1 21-Jun-05
  760 ---------------------
  762  1. There was one reference to the variable "posix" in pcretest.c that was not
  763     surrounded by "#if !defined NOPOSIX".
  765  2. Make it possible to compile pcretest without DFA support, UTF8 support, or
  766     the cross-check on the old pcre_info() function, for the benefit of the
  767     cut-down version of PCRE that is currently imported into Exim.
  769  3. A (silly) pattern starting with (?i)(?-i) caused an internal space
  770     allocation error. I've done the easy fix, which wastes 2 bytes for sensible
  771     patterns that start (?i) but I don't think that matters. The use of (?i) is
  772     just an example; this all applies to the other options as well.
  774  4. Since libtool seems to echo the compile commands it is issuing, the output
  775     from "make" can be reduced a bit by putting "@" in front of each libtool
  776     compile command.
  778  5. Patch from the folks at Google for configure.in to be a bit more thorough
  779     in checking for a suitable C++ installation before trying to compile the
  780     C++ stuff. This should fix a reported problem when a compiler was present,
  781     but no suitable headers.
  783  6. The man pages all had just "PCRE" as their title. I have changed them to
  784     be the relevant file name. I have also arranged that these names are
  785     retained in the file doc/pcre.txt, which is a concatenation in text format
  786     of all the man pages except the little individual ones for each function.
  788  7. The NON-UNIX-USE file had not been updated for the different set of source
  789     files that come with release 6. I also added a few comments about the C++
  790     wrapper.
  793 Version 6.0 07-Jun-05
  794 ---------------------
  796  1. Some minor internal re-organization to help with my DFA experiments.
  798  2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
  799     didn't matter for the library itself when fully configured, but did matter
  800     when compiling without UCP support, or within Exim, where the ucp files are
  801     not imported.
  803  3. Refactoring of the library code to split up the various functions into
  804     different source modules. The addition of the new DFA matching code (see
  805     below) to a single monolithic source would have made it really too
  806     unwieldy, quite apart from causing all the code to be include in a
  807     statically linked application, when only some functions are used. This is
  808     relevant even without the DFA addition now that patterns can be compiled in
  809     one application and matched in another.
  811     The downside of splitting up is that there have to be some external
  812     functions and data tables that are used internally in different modules of
  813     the library but which are not part of the API. These have all had their
  814     names changed to start with "_pcre_" so that they are unlikely to clash
  815     with other external names.
  817  4. Added an alternate matching function, pcre_dfa_exec(), which matches using
  818     a different (DFA) algorithm. Although it is slower than the original
  819     function, it does have some advantages for certain types of matching
  820     problem.
  822  5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
  823     including restarting after a partial match.
  825  6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
  826     defined when compiling for Windows was sent to me. I have put it into the
  827     code, though I have no means of testing or verifying it.
  829  7. Added the pcre_refcount() auxiliary function.
  831  8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
  832     match before or at the first newline in the subject string. In pcretest,
  833     the /f option on a pattern can be used to set this.
  835  9. A repeated \w when used in UTF-8 mode with characters greater than 256
  836     would behave wrongly. This has been present in PCRE since release 4.0.
  838 10. A number of changes to the pcregrep command:
  840     (a) Refactored how -x works; insert ^(...)$ instead of setting
  841         PCRE_ANCHORED and checking the length, in preparation for adding
  842         something similar for -w.
  844     (b) Added the -w (match as a word) option.
  846     (c) Refactored the way lines are read and buffered so as to have more
  847         than one at a time available.
  849     (d) Implemented a pcregrep test script.
  851     (e) Added the -M (multiline match) option. This allows patterns to match
  852         over several lines of the subject. The buffering ensures that at least
  853         8K, or the rest of the document (whichever is the shorter) is available
  854         for matching (and similarly the previous 8K for lookbehind assertions).
  856     (f) Changed the --help output so that it now says
  858           -w, --word-regex(p)
  860         instead of two lines, one with "regex" and the other with "regexp"
  861         because that confused at least one person since the short forms are the
  862         same. (This required a bit of code, as the output is generated
  863         automatically from a table. It wasn't just a text change.)
  865     (g) -- can be used to terminate pcregrep options if the next thing isn't an
  866         option but starts with a hyphen. Could be a pattern or a path name
  867         starting with a hyphen, for instance.
  869     (h) "-" can be given as a file name to represent stdin.
  871     (i) When file names are being printed, "(standard input)" is used for
  872         the standard input, for compatibility with GNU grep. Previously
  873         "<stdin>" was used.
  875     (j) The option --label=xxx can be used to supply a name to be used for
  876         stdin when file names are being printed. There is no short form.
  878     (k) Re-factored the options decoding logic because we are going to add
  879         two more options that take data. Such options can now be given in four
  880         different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
  882     (l) Added the -A, -B, and -C options for requesting that lines of context
  883         around matches be printed.
  885     (m) Added the -L option to print the names of files that do not contain
  886         any matching lines, that is, the complement of -l.
  888     (n) The return code is 2 if any file cannot be opened, but pcregrep does
  889         continue to scan other files.
  891     (o) The -s option was incorrectly implemented. For compatibility with other
  892         greps, it now suppresses the error message for a non-existent or non-
  893         accessible file (but not the return code). There is a new option called
  894         -q that suppresses the output of matching lines, which was what -s was
  895         previously doing.
  897     (p) Added --include and --exclude options to specify files for inclusion
  898         and exclusion when recursing.
  900 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
  901     Hopefully, it now does.
  903 12. Missing cast in pcre_study().
  905 13. Added an "uninstall" target to the makefile.
  907 14. Replaced "extern" in the function prototypes in Makefile.in with
  908     "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
  909     world, but is set differently for Windows.
  911 15. Added a second compiling function called pcre_compile2(). The only
  912     difference is that it has an extra argument, which is a pointer to an
  913     integer error code. When there is a compile-time failure, this is set
  914     non-zero, in addition to the error test pointer being set to point to an
  915     error message. The new argument may be NULL if no error number is required
  916     (but then you may as well call pcre_compile(), which is now just a
  917     wrapper). This facility is provided because some applications need a
  918     numeric error indication, but it has also enabled me to tidy up the way
  919     compile-time errors are handled in the POSIX wrapper.
  921 16. Added VPATH=.libs to the makefile; this should help when building with one
  922     prefix path and installing with another. (Or so I'm told by someone who
  923     knows more about this stuff than I do.)
  925 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
  926     passes PCRE_DOTALL to the pcre_compile() function, making the "." character
  927     match everything, including newlines. This is not POSIX-compatible, but
  928     somebody wanted the feature. From pcretest it can be activated by using
  929     both the P and the s flags.
  931 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
  933 19. libpcre.pc was being incorrectly installed as executable.
  935 20. A couple of places in pcretest check for end-of-line by looking for '\n';
  936     it now also looks for '\r' so that it will work unmodified on Windows.
  938 21. Added Google's contributed C++ wrapper to the distribution.
  940 22. Added some untidy missing memory free() calls in pcretest, to keep
  941     Electric Fence happy when testing.
  945 Version 5.0 13-Sep-04
  946 ---------------------
  948  1. Internal change: literal characters are no longer packed up into items
  949     containing multiple characters in a single byte-string. Each character
  950     is now matched using a separate opcode. However, there may be more than one
  951     byte in the character in UTF-8 mode.
  953  2. The pcre_callout_block structure has two new fields: pattern_position and
  954     next_item_length. These contain the offset in the pattern to the next match
  955     item, and its length, respectively.
  957  3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
  958     insertion of callouts before each pattern item. Added the /C option to
  959     pcretest to make use of this.
  961  4. On the advice of a Windows user, the lines
  963       #if defined(_WIN32) || defined(WIN32)
  964       _setmode( _fileno( stdout ), 0x8000 );
  965       #endif  /* defined(_WIN32) || defined(WIN32) */
  967     have been added to the source of pcretest. This apparently does useful
  968     magic in relation to line terminators.
  970  5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
  971     for the benefit of those environments where the "b" makes a difference.
  973  6. The icc compiler has the same options as gcc, but "configure" doesn't seem
  974     to know about it. I have put a hack into configure.in that adds in code
  975     to set GCC=yes if CC=icc. This seems to end up at a point in the
  976     generated configure script that is early enough to affect the setting of
  977     compiler options, which is what is needed, but I have no means of testing
  978     whether it really works. (The user who reported this had patched the
  979     generated configure script, which of course I cannot do.)
  981     LATER: After change 22 below (new libtool files), the configure script
  982     seems to know about icc (and also ecc). Therefore, I have commented out
  983     this hack in configure.in.
  985  7. Added support for pkg-config (2 patches were sent in).
  987  8. Negated POSIX character classes that used a combination of internal tables
  988     were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
  989     [[:^ascii]]. Typically, they would match almost any characters. The other
  990     POSIX classes were not broken in this way.
  992  9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
  993     to find the match, as PCRE was deluded into thinking that the match had to
  994     start at the start point or following a newline. The same bug applied to
  995     patterns with negative forward assertions or any backward assertions
  996     preceding ".*" at the start, unless the pattern required a fixed first
  997     character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
  999 10. In UTF-8 mode, when moving forwards in the subject after a failed match
 1000     starting at the last subject character, bytes beyond the end of the subject
 1001     string were read.
 1003 11. Renamed the variable "class" as "classbits" to make life easier for C++
 1004     users. (Previously there was a macro definition, but it apparently wasn't
 1005     enough.)
 1007 12. Added the new field "tables" to the extra data so that tables can be passed
 1008     in at exec time, or the internal tables can be re-selected. This allows
 1009     a compiled regex to be saved and re-used at a later time by a different
 1010     program that might have everything at different addresses.
 1012 13. Modified the pcre-config script so that, when run on Solaris, it shows a
 1013     -R library as well as a -L library.
 1015 14. The debugging options of pcretest (-d on the command line or D on a
 1016     pattern) showed incorrect output for anything following an extended class
 1017     that contained multibyte characters and which was followed by a quantifier.
 1019 15. Added optional support for general category Unicode character properties
 1020     via the \p, \P, and \X escapes. Unicode property support implies UTF-8
 1021     support. It adds about 90K to the size of the library. The meanings of the
 1022     inbuilt class escapes such as \d and \s have NOT been changed.
 1024 16. Updated pcredemo.c to include calls to free() to release the memory for the
 1025     compiled pattern.
 1027 17. The generated file chartables.c was being created in the source directory
 1028     instead of in the building directory. This caused the build to fail if the
 1029     source directory was different from the building directory, and was
 1030     read-only.
 1032 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
 1033     file. No doubt somebody will tell me if they don't make sense... Also added
 1034     Dan Mooney's comments about building on OpenVMS.
 1036 19. Added support for partial matching via the PCRE_PARTIAL option for
 1037     pcre_exec() and the \P data escape in pcretest.
 1039 20. Extended pcretest with 3 new pattern features:
 1041     (i)   A pattern option of the form ">rest-of-line" causes pcretest to
 1042           write the compiled pattern to the file whose name is "rest-of-line".
 1043           This is a straight binary dump of the data, with the saved pointer to
 1044           the character tables forced to be NULL. The study data, if any, is
 1045           written too. After writing, pcretest reads a new pattern.
 1047     (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
 1048           compiled pattern from the given file. There must not be any
 1049           occurrences of "<" in the file name (pretty unlikely); if there are,
 1050           pcretest will instead treat the initial "<" as a pattern delimiter.
 1051           After reading in the pattern, pcretest goes on to read data lines as
 1052           usual.
 1054     (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
 1055           and 16-bit fields in a compiled pattern, to simulate a pattern that
 1056           was compiled on a host of opposite endianness.
 1058 21. The pcre-exec() function can now cope with patterns that were compiled on
 1059     hosts of opposite endianness, with this restriction:
 1061       As for any compiled expression that is saved and used later, the tables
 1062       pointer field cannot be preserved; the extra_data field in the arguments
 1063       to pcre_exec() should be used to pass in a tables address if a value
 1064       other than the default internal tables were used at compile time.
 1066 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
 1067     now diagnosed as an error. Previously, most of the time, a negative number
 1068     would have been treated as zero, but if in addition "ovector" was passed as
 1069     NULL, a crash could occur.
 1071 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
 1072     new versions from the libtool 1.5 distribution (the last one is a copy of
 1073     a file called libtool.m4). This seems to have fixed the need to patch
 1074     "configure" to support Darwin 1.3 (which I used to do). However, I still
 1075     had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
 1076     workstation).
 1078 24. Changed the PCRE licence to be the more standard "BSD" licence.
 1081 Version 4.5 01-Dec-03
 1082 ---------------------
 1084  1. There has been some re-arrangement of the code for the match() function so
 1085     that it can be compiled in a version that does not call itself recursively.
 1086     Instead, it keeps those local variables that need separate instances for
 1087     each "recursion" in a frame on the heap, and gets/frees frames whenever it
 1088     needs to "recurse". Keeping track of where control must go is done by means
 1089     of setjmp/longjmp. The whole thing is implemented by a set of macros that
 1090     hide most of the details from the main code, and operates only if
 1091     NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
 1092     "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
 1093     operating.
 1095     To make it easier for callers to provide specially tailored get/free
 1096     functions for this usage, two new functions, pcre_stack_malloc, and
 1097     pcre_stack_free, are used. They are always called in strict stacking order,
 1098     and the size of block requested is always the same.
 1100     The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
 1101     PCRE has been compiled to use the stack or the heap for recursion. The
 1102     -C option of pcretest uses this to show which version is compiled.
 1104     A new data escape \S, is added to pcretest; it causes the amounts of store
 1105     obtained and freed by both kinds of malloc/free at match time to be added
 1106     to the output.
 1108  2. Changed the locale test to use "fr_FR" instead of "fr" because that's
 1109     what's available on my current Linux desktop machine.
 1111  3. When matching a UTF-8 string, the test for a valid string at the start has
 1112     been extended. If start_offset is not zero, PCRE now checks that it points
 1113     to a byte that is the start of a UTF-8 character. If not, it returns
 1114     PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
 1115     this is necessary because there may be backward assertions in the pattern.
 1116     When matching the same subject several times, it may save resources to use
 1117     PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
 1119  4. The code for checking the validity of UTF-8 strings has been tightened so
 1120     that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
 1121     containing "overlong sequences".
 1123  5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
 1124     I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
 1125     should have been "&", but it just so happened that all the cases this let
 1126     through by mistake were picked up later in the function.
 1128  6. I had used a variable called "isblank" - this is a C99 function, causing
 1129     some compilers to warn. To avoid this, I renamed it (as "blankclass").
 1131  7. Cosmetic: (a) only output another newline at the end of pcretest if it is
 1132     prompting; (b) run "./pcretest /dev/null" at the start of the test script
 1133     so the version is shown; (c) stop "make test" echoing "./RunTest".
 1135  8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
 1137  9. The prototype for memmove() for systems that don't have it was using
 1138     size_t, but the inclusion of the header that defines size_t was later. I've
 1139     moved the #includes for the C headers earlier to avoid this.
 1141 10. Added some adjustments to the code to make it easier to compiler on certain
 1142     special systems:
 1144       (a) Some "const" qualifiers were missing.
 1145       (b) Added the macro EXPORT before all exported functions; by default this
 1146           is defined to be empty.
 1147       (c) Changed the dftables auxiliary program (that builds chartables.c) so
 1148           that it reads its output file name as an argument instead of writing
 1149           to the standard output and assuming this can be redirected.
 1151 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
 1152     class containing characters with values greater than 255, PCRE compilation
 1153     went into a loop.
 1155 12. A recursive reference to a subpattern that was within another subpattern
 1156     that had a minimum quantifier of zero caused PCRE to crash. For example,
 1157     (x(y(?2))z)? provoked this bug with a subject that got as far as the
 1158     recursion. If the recursively-called subpattern itself had a zero repeat,
 1159     that was OK.
 1161 13. In pcretest, the buffer for reading a data line was set at 30K, but the
 1162     buffer into which it was copied (for escape processing) was still set at
 1163     1024, so long lines caused crashes.
 1165 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
 1166     "internal error: code overflow...". This applied to any character class
 1167     that was followed by a possessive quantifier.
 1169 15. Modified the Makefile to add libpcre.la as a prerequisite for
 1170     libpcreposix.la because I was told this is needed for a parallel build to
 1171     work.
 1173 16. If a pattern that contained .* following optional items at the start was
 1174     studied, the wrong optimizing data was generated, leading to matching
 1175     errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
 1176     matching string must start with a or b or c. The correct conclusion for
 1177     this pattern is that a match can start with any character.
 1180 Version 4.4 13-Aug-03
 1181 ---------------------
 1183  1. In UTF-8 mode, a character class containing characters with values between
 1184     127 and 255 was not handled correctly if the compiled pattern was studied.
 1185     In fixing this, I have also improved the studying algorithm for such
 1186     classes (slightly).
 1188  2. Three internal functions had redundant arguments passed to them. Removal
 1189     might give a very teeny performance improvement.
 1191  3. Documentation bug: the value of the capture_top field in a callout is *one
 1192     more than* the number of the hightest numbered captured substring.
 1194  4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
 1195     in incorrectly linking with a previously installed version. They now link
 1196     explicitly with libpcre.la.
 1198  5. configure.in no longer needs to recognize Cygwin specially.
 1200  6. A problem in pcre.in for Windows platforms is fixed.
 1202  7. If a pattern was successfully studied, and the -d (or /D) flag was given to
 1203     pcretest, it used to include the size of the study block as part of its
 1204     output. Unfortunately, the structure contains a field that has a different
 1205     size on different hardware architectures. This meant that the tests that
 1206     showed this size failed. As the block is currently always of a fixed size,
 1207     this information isn't actually particularly useful in pcretest output, so
 1208     I have just removed it.
 1210  8. Three pre-processor statements accidentally did not start in column 1.
 1211     Sadly, there are *still* compilers around that complain, even though
 1212     standard C has not required this for well over a decade. Sigh.
 1214  9. In pcretest, the code for checking callouts passed small integers in the
 1215     callout_data field, which is a void * field. However, some picky compilers
 1216     complained about the casts involved for this on 64-bit systems. Now
 1217     pcretest passes the address of the small integer instead, which should get
 1218     rid of the warnings.
 1220 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
 1221     both compile and run time, and gives an error if an invalid UTF-8 sequence
 1222     is found. There is a option for disabling this check in cases where the
 1223     string is known to be correct and/or the maximum performance is wanted.
 1225 11. In response to a bug report, I changed one line in Makefile.in from
 1227         -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
 1228     to
 1229         -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
 1231     to look similar to other lines, but I have no way of telling whether this
 1232     is the right thing to do, as I do not use Windows. No doubt I'll get told
 1233     if it's wrong...
 1236 Version 4.3 21-May-03
 1237 ---------------------
 1239 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
 1240    Makefile.
 1242 2. Some refactoring to improve the quality of the code:
 1244    (i)   The utf8_table... variables are now declared "const".
 1246    (ii)  The code for \cx, which used the "case flipping" table to upper case
 1247          lower case letters, now just substracts 32. This is ASCII-specific,
 1248          but the whole concept of \cx is ASCII-specific, so it seems
 1249          reasonable.
 1251    (iii) PCRE was using its character types table to recognize decimal and
 1252          hexadecimal digits in the pattern. This is silly, because it handles
 1253          only 0-9, a-f, and A-F, but the character types table is locale-
 1254          specific, which means strange things might happen. A private
 1255          table is now used for this - though it costs 256 bytes, a table is
 1256          much faster than multiple explicit tests. Of course, the standard
 1257          character types table is still used for matching digits in subject
 1258          strings against \d.
 1260    (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
 1261          ending in _t are). So I've renamed it as ESC_tee.
 1263 3. The first argument for regexec() in the POSIX wrapper should have been
 1264    defined as "const".
 1266 4. Changed pcretest to use malloc() for its buffers so that they can be
 1267    Electric Fenced for debugging.
 1269 5. There were several places in the code where, in UTF-8 mode, PCRE would try
 1270    to read one or more bytes before the start of the subject string. Often this
 1271    had no effect on PCRE's behaviour, but in some circumstances it could
 1272    provoke a segmentation fault.
 1274 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
 1275    to try to read one or more bytes before the start of the subject string.
 1277 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
 1278    UTF-8 support could misbehave in various ways if the subject string
 1279    contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
 1280    area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
 1281    back over UTF-8 characters.)
 1284 Version 4.2 14-Apr-03
 1285 ---------------------
 1287 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
 1289 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
 1290      [ON_WINDOWS]: new variable, "#" on non-Windows platforms
 1291      [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
 1292      [WIN_PREFIX]: new variable, "cyg" for Cygwin
 1293      * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
 1294        and BUILD_EXEEXT
 1295      Note: automatic setting of the BUILD variables is not yet working
 1296      set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
 1297        compile-time but not at link-time
 1298      [LINK]: use for linking executables only
 1299      make different versions for Windows and non-Windows
 1300      [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
 1301        libraries
 1302      [LINK_FOR_BUILD]: new variable
 1303      [OBJEXT]: use throughout
 1304      [EXEEXT]: use throughout
 1305      <winshared>: new target
 1306      <wininstall>: new target
 1307      <dftables.o>: use native compiler
 1308      <dftables>: use native linker
 1309      <install>: handle Windows platform correctly
 1310      <clean>: ditto
 1311      <check>: ditto
 1312      copy DLL to top builddir before testing
 1314    As part of these changes, -no-undefined was removed again. This was reported
 1315    to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
 1316    in any case.
 1318 3. Some tidies to get rid of compiler warnings:
 1320    . In the match_data structure, match_limit was an unsigned long int, whereas
 1321      match_call_count was an int. I've made them both unsigned long ints.
 1323    . In pcretest the fact that a const uschar * doesn't automatically cast to
 1324      a void * provoked a warning.
 1326    . Turning on some more compiler warnings threw up some "shadow" variables
 1327      and a few more missing casts.
 1329 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
 1330    option, a class that contained a single character with a value between 128
 1331    and 255 (e.g. /[\xFF]/) caused PCRE to crash.
 1333 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
 1334    option, a class that contained several characters, but with at least one
 1335    whose value was between 128 and 255 caused PCRE to crash.
 1338 Version 4.1 12-Mar-03
 1339 ---------------------
 1341 1. Compiling with gcc -pedantic found a couple of places where casts were
 1342 needed, and a string in dftables.c that was longer than standard compilers are
 1343 required to support.
 1345 2. Compiling with Sun's compiler found a few more places where the code could
 1346 be tidied up in order to avoid warnings.
 1348 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
 1349 first of these names is deprecated in the latest Autoconf in favour of the name
 1350 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
 1351 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
 1352 analogy I have changed it to CFLAGS_FOR_BUILD.
 1354 4. Added -no-undefined to the linking command in the Makefile, because this is
 1355 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
 1356 linking step for the pcreposix library.
 1358 5. PCRE was failing to diagnose the case of two named groups with the same
 1359 name.
 1361 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
 1362 literal character that is needed in the subject for a match, and scans along to
 1363 ensure that it is present before embarking on the full matching process. This
 1364 saves time in cases of nested unlimited repeats that are never going to match.
 1365 Problem: the scan can take a lot of time if the subject is very long (e.g.
 1366 megabytes), thus penalizing straightforward matches. It is now done only if the
 1367 amount of subject to be scanned is less than 1000 bytes.
 1369 7. A lesser problem with the same optimization is that it was recording the
 1370 first character of an anchored pattern as "needed", thus provoking a search
 1371 right along the subject, even when the first match of the pattern was going to
 1372 fail. The "needed" character is now not set for anchored patterns, unless it
 1373 follows something in the pattern that is of non-fixed length. Thus, it still
 1374 fulfils its original purpose of finding quick non-matches in cases of nested
 1375 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
 1378 Version 4.0 17-Feb-03
 1379 ---------------------
 1381 1. If a comment in an extended regex that started immediately after a meta-item
 1382 extended to the end of string, PCRE compiled incorrect data. This could lead to
 1383 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
 1385 2. Moved to autoconf 2.53 and libtool 1.4.2.
 1387 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
 1388 the special perltest8 script is no longer needed - all the tests can be run
 1389 from a single perltest script.
 1391 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
 1392 by \s. It has now been removed in PCRE. This means it isn't recognized as
 1393 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
 1394 class [:space:] *does* include VT, thereby creating a mess.
 1396 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
 1397 space and tab.
 1399 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
 1400 its new features into the main test script, reducing the number of scripts.
 1402 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
 1403 were backward compatible, and made the (?i) apply to the whole pattern, as if
 1404 /i were given. Now it behaves more logically, and applies the option setting
 1405 only to what follows. PCRE has been changed to follow suit. However, if it
 1406 finds options settings right at the start of the pattern, it extracts them into
 1407 the global options, as before. Thus, they show up in the info data.
 1409 8. Added support for the \Q...\E escape sequence. Characters in between are
 1410 treated as literals. This is slightly different from Perl in that $ and @ are
 1411 also handled as literals inside the quotes. In Perl, they will cause variable
 1412 interpolation. Note the following examples:
 1414     Pattern            PCRE matches      Perl matches
 1416     \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
 1417     \Qabc\$xyz\E       abc\$xyz          abc\$xyz
 1418     \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
 1420 For compatibility with Perl, \Q...\E sequences are recognized inside character
 1421 classes as well as outside them.
 1423 9. Re-organized 3 code statements in pcretest to avoid "overflow in
 1424 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
 1425 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
 1426 signed/unsigned warnings.
 1428 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
 1429 option for pcretest, so I've replaced it by a simple function that does just
 1430 that job.
 1432 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
 1433 "pcregrep -".
 1435 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
 1436 Java package. This provides some syntactic sugar for simple cases of what my
 1437 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
 1438 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
 1439 item, you can use this simplified notation. Note that only makes sense with
 1440 greedy quantifiers. Consequently, the use of the possessive quantifier forces
 1441 greediness, whatever the setting of the PCRE_UNGREEDY option.
 1443 13. A change of greediness default within a pattern was not taking effect at
 1444 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
 1445 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
 1446 was abstracted outside.
 1448 14. PCRE now supports the \G assertion. It is true when the current matching
 1449 position is at the start point of the match. This differs from \A when the
 1450 starting offset is non-zero. Used with the /g option of pcretest (or similar
 1451 code), it works in the same way as it does for Perl's /g option. If all
 1452 alternatives of a regex begin with \G, the expression is anchored to the start
 1453 match position, and the "anchored" flag is set in the compiled expression.
 1455 15. Some bugs concerning the handling of certain option changes within patterns
 1456 have been fixed. These applied to options other than (?ims). For example,
 1457 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
 1458 been the other way round. Some of this was related to change 7 above.
 1460 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
 1461 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
 1462 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
 1463 POSIX classes only within a class (e.g. /[[:alpha:]]/).
 1465 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
 1466 mode. Unlike ".", it always matches newline, whatever the setting of
 1467 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
 1468 assertions. Perl allows it, but it doesn't (in general) work because it can't
 1469 calculate the length of the lookbehind. At least, that's the case for Perl
 1470 5.8.0 - I've been told they are going to document that it doesn't work in
 1471 future.
 1473 18. Added an error diagnosis for escapes that PCRE does not support: these are
 1474 \L, \l, \N, \P, \p, \U, \u, and \X.
 1476 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
 1477 reading past the end of the pattern in cases such as /[abcd/.
 1479 20. PCRE was getting more memory than necessary for patterns with classes that
 1480 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
 1482 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
 1483 compiling PCRE for use with Virtual Pascal.
 1485 22. Small fix to the Makefile to make it work properly if the build is done
 1486 outside the source tree.
 1488 23. Added a new extension: a condition to go with recursion. If a conditional
 1489 subpattern starts with (?(R) the "true" branch is used if recursion has
 1490 happened, whereas the "false" branch is used only at the top level.
 1492 24. When there was a very long string of literal characters (over 255 bytes
 1493 without UTF support, over 250 bytes with UTF support), the computation of how
 1494 much memory was required could be incorrect, leading to segfaults or other
 1495 strange effects.
 1497 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
 1498 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
 1499 there was a subsequent back reference to those brackets. This meant that, for
 1500 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
 1501 possible to check for precisely this case. All we can do is abandon the
 1502 optimization if .* occurs inside capturing brackets when there are any back
 1503 references whatsoever. (See below for a better fix that came later.)
 1505 26. The handling of the optimization for finding the first character of a
 1506 non-anchored pattern, and for finding a character that is required later in the
 1507 match were failing in some cases. This didn't break the matching; it just
 1508 failed to optimize when it could. The way this is done has been re-implemented.
 1510 27. Fixed typo in error message for invalid (?R item (it said "(?p").
 1512 28. Added a new feature that provides some of the functionality that Perl
 1513 provides with (?{...}). The facility is termed a "callout". The way it is done
 1514 in PCRE is for the caller to provide an optional function, by setting
 1515 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
 1516 global variable. By default it is unset, which disables all calling out. To get
 1517 the function called, the regex must include (?C) at appropriate points. This
 1518 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
 1519 This provides a means of identifying different callout points. When PCRE
 1520 reaches such a point in the regex, if pcre_callout has been set, the external
 1521 function is called. It is provided with data in a structure called
 1522 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
 1523 matching continues; if it returns a non-zero value, the match at the current
 1524 point fails. However, backtracking will occur if possible. [This was changed
 1525 later and other features added - see item 49 below.]
 1527 29. pcretest is upgraded to test the callout functionality. It provides a
 1528 callout function that displays information. By default, it shows the start of
 1529 the match and the current position in the text. There are some new data escapes
 1530 to vary what happens:
 1532     \C+         in addition, show current contents of captured substrings
 1533     \C-         do not supply a callout function
 1534     \C!n        return 1 when callout number n is reached
 1535     \C!n!m      return 1 when callout number n is reached for the mth time
 1537 30. If pcregrep was called with the -l option and just a single file name, it
 1538 output "<stdin>" if a match was found, instead of the file name.
 1540 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
 1541 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
 1542 pcre_exec(). This saves a malloc/free per call. The default value of
 1543 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
 1544 when configuring.
 1546 32. The default maximum size of a compiled pattern is 64K. There have been a
 1547 few cases of people hitting this limit. The code now uses macros to handle the
 1548 storing of links as offsets within the compiled pattern. It defaults to 2-byte
 1549 links, but this can be changed to 3 or 4 bytes by --with-link-size when
 1550 configuring. Tests 2 and 5 work only with 2-byte links because they output
 1551 debugging information about compiled patterns.
 1553 33. Internal code re-arrangements:
 1555 (a) Moved the debugging function for printing out a compiled regex into
 1556     its own source file (printint.c) and used #include to pull it into
 1557     pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
 1558     separate copies.
 1560 (b) Defined the list of op-code names for debugging as a macro in
 1561     internal.h so that it is next to the definition of the opcodes.
 1563 (c) Defined a table of op-code lengths for simpler skipping along compiled
 1564     code. This is again a macro in internal.h so that it is next to the
 1565     definition of the opcodes.
 1567 34. Added support for recursive calls to individual subpatterns, along the
 1568 lines of Robin Houston's patch (but implemented somewhat differently).
 1570 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
 1571 allow it to read and process whole directories in Win32. This code was
 1572 contributed by Lionel Fourquaux; it has not been tested by me.
 1574 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
 1575 used to name a group. Names consist of alphanumerics and underscores, and must
 1576 be unique. Back references use the syntax (?P=name) and recursive calls use
 1577 (?P>name) which is a PCRE extension to the Python extension. Groups still have
 1578 numbers. The function pcre_fullinfo() can be used after compilation to extract
 1579 a name/number map. There are three relevant calls:
 1581   PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
 1582   PCRE_INFO_NAMECOUNT            yields the number of entries
 1583   PCRE_INFO_NAMETABLE            yields a pointer to the map.
 1585 The map is a vector of fixed-size entries. The size of each entry depends on
 1586 the length of the longest name used. The first two bytes of each entry are the
 1587 group number, most significant byte first. There follows the corresponding
 1588 name, zero terminated. The names are in alphabetical order.
 1590 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
 1591 case instead of 255. Making it the same both with and without UTF-8 support
 1592 means that the same test output works with both.
 1594 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
 1595 calling malloc() with a zero argument.
 1597 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
 1598 optimization. I've improved things by keeping a bitmap of backreferences with
 1599 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
 1600 fact referenced, the optimization can be applied. It is unlikely that a
 1601 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
 1602 the match to follow \n) will appear inside brackets with a number greater than
 1603 31, but if it does, any back reference > 31 suppresses the optimization.
 1605 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
 1606 of disabling numbered capturing parentheses. Any opening parenthesis that is
 1607 not followed by ? behaves as if it were followed by ?: but named parentheses
 1608 can still be used for capturing (and they will acquire numbers in the usual
 1609 way).
 1611 41. Redesigned the return codes from the match() function into yes/no/error so
 1612 that errors can be passed back from deep inside the nested calls. A malloc
 1613 failure while inside a recursive subpattern call now causes the
 1614 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
 1616 42. It is now possible to set a limit on the number of times the match()
 1617 function is called in a call to pcre_exec(). This facility makes it possible to
 1618 limit the amount of recursion and backtracking, though not in a directly
 1619 obvious way, because the match() function is used in a number of different
 1620 circumstances. The count starts from zero for each position in the subject
 1621 string (for non-anchored patterns). The default limit is, for compatibility, a
 1622 large number, namely 10 000 000. You can change this in two ways:
 1624 (a) When configuring PCRE before making, you can use --with-match-limit=n
 1625     to set a default value for the compiled library.
 1627 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
 1628     a different value is set. See 45 below.
 1630 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
 1632 43. Added a new function pcre_config(int, void *) to enable run-time extraction
 1633 of things that can be changed at compile time. The first argument specifies
 1634 what is wanted and the second points to where the information is to be placed.
 1635 The current list of available information is:
 1639 The output is an integer that is set to one if UTF-8 support is available;
 1640 otherwise it is set to zero.
 1644 The output is an integer that it set to the value of the code that is used for
 1645 newline. It is either LF (10) or CR (13).
 1649 The output is an integer that contains the number of bytes used for internal
 1650 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
 1654 The output is an integer that contains the threshold above which the POSIX
 1655 interface uses malloc() for output vectors. See item 31 above.
 1659 The output is an unsigned integer that contains the default limit of the number
 1660 of match() calls in a pcre_exec() execution. See 42 above.
 1662 44. pcretest has been upgraded by the addition of the -C option. This causes it
 1663 to extract all the available output from the new pcre_config() function, and to
 1664 output it. The program then exits immediately.
 1666 45. A need has arisen to pass over additional data with calls to pcre_exec() in
 1667 order to support additional features. One way would have been to define
 1668 pcre_exec2() (for example) with extra arguments, but this would not have been
 1669 extensible, and would also have required all calls to the original function to
 1670 be mapped to the new one. Instead, I have chosen to extend the mechanism that
 1671 is used for passing in "extra" data from pcre_study().
 1673 The pcre_extra structure is now exposed and defined in pcre.h. It currently
 1674 contains the following fields:
 1676   flags         a bitmap indicating which of the following fields are set
 1677   study_data    opaque data from pcre_study()
 1678   match_limit   a way of specifying a limit on match() calls for a specific
 1679                   call to pcre_exec()
 1680   callout_data  data for callouts (see 49 below)
 1682 The flag bits are also defined in pcre.h, and are
 1688 The pcre_study() function now returns one of these new pcre_extra blocks, with
 1689 the actual study data pointed to by the study_data field, and the
 1690 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
 1691 before. That is, this change is entirely upwards-compatible and requires no
 1692 change to existing code.
 1694 If you want to pass in additional data to pcre_exec(), you can either place it
 1695 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
 1696 block.
 1698 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
 1699 data string contains the escape sequence \M, pcretest calls pcre_exec() several
 1700 times with different match limits, until it finds the minimum value needed for
 1701 pcre_exec() to complete. The value is then output. This can be instructive; for
 1702 most simple matches the number is quite small, but for pathological cases it
 1703 gets very large very quickly.
 1705 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
 1706 returns the size of the data block pointed to by the study_data field in a
 1707 pcre_extra block, that is, the value that was passed as the argument to
 1708 pcre_malloc() when PCRE was getting memory in which to place the information
 1709 created by pcre_study(). The fourth argument should point to a size_t variable.
 1710 pcretest has been extended so that this information is shown after a successful
 1711 pcre_study() call when information about the compiled regex is being displayed.
 1713 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
 1714 because what follows is always an absolute path. (Later: it turns out that this
 1715 is more than cosmetic for MinGW, because it doesn't like empty path
 1716 components.)
 1718 49. Some changes have been made to the callout feature (see 28 above):
 1720 (i)  A callout function now has three choices for what it returns:
 1722        0  =>  success, carry on matching
 1723      > 0  =>  failure at this point, but backtrack if possible
 1724      < 0  =>  serious error, return this value from pcre_exec()
 1726      Negative values should normally be chosen from the set of PCRE_ERROR_xxx
 1727      values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
 1728      "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
 1729      use by callout functions. It will never be used by PCRE itself.
 1731 (ii) The pcre_extra structure (see 45 above) has a void * field called
 1732      callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
 1733      pcre_callout_block structure has a field of the same name. The contents of
 1734      the field passed in the pcre_extra structure are passed to the callout
 1735      function in the corresponding field in the callout block. This makes it
 1736      easier to use the same callout-containing regex from multiple threads. For
 1737      testing, the pcretest program has a new data escape
 1739        \C*n        pass the number n (may be negative) as callout_data
 1741      If the callout function in pcretest receives a non-zero value as
 1742      callout_data, it returns that value.
 1744 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
 1745 there were some redundant $(CFLAGS) in commands that are now specified as
 1746 $(LINK), which already includes $(CFLAGS).
 1748 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
 1749 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
 1750 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
 1751 one-byte characters throughout. Note that case-insensitive matching applies
 1752 only to characters whose values are less than 256. PCRE doesn't support the
 1753 notion of cases for higher-valued characters.
 1755 (i)   A character class whose characters are all within 0-255 is handled as
 1756       a bit map, and the map is inverted for negative classes. Previously, a
 1757       character > 255 always failed to match such a class; however it should
 1758       match if the class was a negative one (e.g. [^ab]). This has been fixed.
 1760 (ii)  A negated character class with a single character < 255 is coded as
 1761       "not this character" (OP_NOT). This wasn't working properly when the test
 1762       character was multibyte, either singly or repeated.
 1764 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
 1765       mode, for example: \x{100}{2,3}.
 1767 (iv)  The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
 1768       singly or repeated) now correctly test multibyte characters. However,
 1769       PCRE doesn't recognize any characters with values greater than 255 as
 1770       digits, spaces, or word characters. Such characters always match \D, \S,
 1771       and \W, and never match \d, \s, or \w.
 1773 (v)   Classes may now contain characters and character ranges with values
 1774       greater than 255. For example: [ab\x{100}-\x{400}].
 1776 (vi)  pcregrep now has a --utf-8 option (synonym -u) which makes it call
 1777       PCRE in UTF-8 mode.
 1779 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
 1780 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
 1781 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
 1782 value.)
 1784 53. The single man page has become too large. I have therefore split it up into
 1785 a number of separate man pages. These also give rise to individual HTML pages;
 1786 these are now put in a separate directory, and there is an index.html page that
 1787 lists them all. Some hyperlinking between the pages has been installed.
 1789 54. Added convenience functions for handling named capturing parentheses.
 1791 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
 1792 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
 1793 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
 1794 are faulted.
 1796 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
 1797 calling configure. These values are used when compiling the dftables.c program
 1798 which is run to generate the source of the default character tables. They
 1799 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
 1800 you will need to set these values.
 1802 57. Updated the building process for Windows DLL, as provided by Fred Cox.
 1805 Version 3.9 02-Jan-02
 1806 ---------------------
 1808 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
 1810 2. If --disable-static was given, the building process failed when trying to
 1811 build pcretest and pcregrep. (For some reason it was using libtool to compile
 1812 them, which is not right, as they aren't part of the library.)
 1815 Version 3.8 18-Dec-01
 1816 ---------------------
 1818 1. The experimental UTF-8 code was completely screwed up. It was packing the
 1819 bytes in the wrong order. How dumb can you get?
 1822 Version 3.7 29-Oct-01
 1823 ---------------------
 1825 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
 1826 This caused pcretest, when used on the test data, to segfault. Unfortunately,
 1827 this didn't happen under Solaris 8, where I normally test things.
 1829 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
 1830 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
 1831 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
 1832 this fix an hour or so after the initial 3.7 release.)
 1835 Version 3.6 23-Oct-01
 1836 ---------------------
 1838 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
 1839 offsets passed as NULL with zero offset count.
 1841 2. The config.guess and config.sub files had not been updated when I moved to
 1842 the latest autoconf.
 1845 Version 3.5 15-Aug-01
 1846 ---------------------
 1848 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
 1849 had been forgotten.
 1851 2. By using declared but undefined structures, we can avoid using "void"
 1852 definitions in pcre.h while keeping the internal definitions of the structures
 1853 private.
 1855 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
 1856 user point of view, this means that both static and shared libraries are built
 1857 by default, but this can be individually controlled. More of the work of
 1858 handling this static/shared cases is now inside libtool instead of PCRE's make
 1859 file.
 1861 4. The pcretest utility is now installed along with pcregrep because it is
 1862 useful for users (to test regexs) and by doing this, it automatically gets
 1863 relinked by libtool. The documentation has been turned into a man page, so
 1864 there are now .1, .txt, and .html versions in /doc.
 1866 5. Upgrades to pcregrep:
 1867    (i)   Added long-form option names like gnu grep.
 1868    (ii)  Added --help to list all options with an explanatory phrase.
 1869    (iii) Added -r, --recursive to recurse into sub-directories.
 1870    (iv)  Added -f, --file to read patterns from a file.
 1872 6. pcre_exec() was referring to its "code" argument before testing that
 1873 argument for NULL (and giving an error if it was NULL).
 1875 7. Upgraded Makefile.in to allow for compiling in a different directory from
 1876 the source directory.
 1878 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
 1879 options bits, the pointer it was passed was to an int instead of to an unsigned
 1880 long int. This mattered only on 64-bit systems.
 1882 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
 1883 generated) instead of pcre.in, which it its source. Also made the same change
 1884 in several of the .c files.
 1886 10. A new release of gcc defines printf() as a macro, which broke pcretest
 1887 because it had an ifdef in the middle of a string argument for printf(). Fixed
 1888 by using separate calls to printf().
 1890 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
 1891 script, to force use of CR or LF instead of \n in the source. On non-Unix
 1892 systems, the value can be set in config.h.
 1894 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
 1895 absolute limit. Changed the text of the error message to make this clear, and
 1896 likewise updated the man page.
 1898 13. The limit of 99 on the number of capturing subpatterns has been removed.
 1899 The new limit is 65535, which I hope will not be a "real" limit.
 1902 Version 3.4 22-Aug-00
 1903 ---------------------
 1905 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
 1907 2. Diagnose condition (?(0) as an error instead of crashing on matching.
 1910 Version 3.3 01-Aug-00
 1911 ---------------------
 1913 1. If an octal character was given, but the value was greater than \377, it
 1914 was not getting masked to the least significant bits, as documented. This could
 1915 lead to crashes in some systems.
 1917 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
 1918 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
 1920 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
 1921 These just pass their arguments on to (pcre_free)(), but they are provided
 1922 because some uses of PCRE bind it to non-C systems that can call its functions,
 1923 but cannot call free() or pcre_free() directly.
 1925 4. Add "make test" as a synonym for "make check". Corrected some comments in
 1926 the Makefile.
 1928 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
 1929 Makefile.
 1931 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
 1932 command called pgrep for grepping around the active processes.
 1934 7. Added the beginnings of support for UTF-8 character strings.
 1936 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
 1937 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
 1938 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
 1939 out for the ar command.)
 1942 Version 3.2 12-May-00
 1943 ---------------------
 1945 This is purely a bug fixing release.
 1947 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
 1948 of ZA. This was just one example of several cases that could provoke this bug,
 1949 which was introduced by change 9 of version 2.00. The code for breaking
 1950 infinite loops after an iteration that matches an empty string was't working
 1951 correctly.
 1953 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
 1954 when matched against abbab (for example). After matching an empty string, it
 1955 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
 1956 caused it to match further down the string than it should.
 1958 3. The code contained an inclusion of sys/types.h. It isn't clear why this
 1959 was there because it doesn't seem to be needed, and it causes trouble on some
 1960 systems, as it is not a Standard C header. It has been removed.
 1962 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
 1963 were reported on the Macintosh. The changes were from
 1965   while ((c = *(++ptr)) != 0 && c != '\n');
 1966 to
 1967   while ((c = *(++ptr)) != 0 && c != '\n') ;
 1969 Totally extraordinary, but if that's what it takes...
 1971 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
 1972 available. Added HAVE_BCOPY and an autoconf test for it; if neither
 1973 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
 1974 assumes the way PCRE uses memmove() (always moving upwards).
 1976 6. PCRE is being used in one environment where strchr() is not available. There
 1977 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
 1978 faster code anyway.
 1981 Version 3.1 09-Feb-00
 1982 ---------------------
 1984 The only change in this release is the fixing of some bugs in Makefile.in for
 1985 the "install" target:
 1987 (1) It was failing to install pcreposix.h.
 1989 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
 1992 Version 3.0 01-Feb-00
 1993 ---------------------
 1995 1. Add support for the /+ modifier to perltest (to output $` like it does in
 1996 pcretest).
 1998 2. Add support for the /g modifier to perltest.
 2000 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
 2001 matches null strings.
 2003 4. Fix perltest so that it doesn't do unwanted things when fed an empty
 2004 pattern. Perl treats empty patterns specially - it reuses the most recent
 2005 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
 2006 effect.
 2008 5. The POSIX interface was broken in that it was just handing over the POSIX
 2009 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
 2010 required a bigger vector, with some working space on the end. This means that
 2011 the POSIX wrapper now has to get and free some memory, and copy the results.
 2013 6. Added some simple autoconf support, placing the test data and the
 2014 documentation in separate directories, re-organizing some of the
 2015 information files, and making it build pcre-config (a GNU standard). Also added
 2016 libtool support for building PCRE as a shared library, which is now the
 2017 default.
 2019 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
 2020 09 are not valid octal constants. Single digits will be used for minor values
 2021 less than 10.
 2023 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
 2024 existing programs that set these in the POSIX interface can use PCRE without
 2025 modification.
 2027 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
 2028 return all that pcre_info() returns, plus additional data. The pcre_info()
 2029 function is retained for compatibility, but is considered to be obsolete.
 2031 10. Added experimental recursion feature (?R) to handle one common case that
 2032 Perl 5.6 will be able to do with (?p{...}).
 2034 11. Added support for POSIX character classes like [:alpha:], which Perl is
 2035 adopting.
 2038 Version 2.08 31-Aug-99
 2039 ----------------------
 2041 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
 2042 trying to match at the startoffset position, but instead was moving forward to
 2043 the next newline as if a previous match had failed.
 2045 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
 2046 and could get into a loop if a null string was matched other than at the start
 2047 of the subject.
 2049 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
 2050 be distinguished at compile time, and for completeness also added PCRE_DATE.
 2052 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
 2053 in GnuWin32 environments.
 2056 Version 2.07 29-Jul-99
 2057 ----------------------
 2059 1. The documentation is now supplied in plain text form and HTML as well as in
 2060 the form of man page sources.
 2062 2. C++ compilers don't like assigning (void *) values to other pointer types.
 2063 In particular this affects malloc(). Although there is no problem in Standard
 2064 C, I've put in casts to keep C++ compilers happy.
 2066 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
 2067 should be (const char *).
 2069 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
 2070 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
 2071 However, I haven't made this a standard facility. The documentation doesn't
 2072 mention it, and the Makefile doesn't support it.
 2074 5. The Makefile now contains an "install" target, with editable destinations at
 2075 the top of the file. The pcretest program is not installed.
 2077 6. pgrep -V now gives the PCRE version number and date.
 2079 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
 2080 causing the entire string to be ignored, instead of just the last character.
 2082 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
 2083 non-matching string, it can take a very, very long time, even for strings of
 2084 quite modest length, because of the nested recursion. PCRE now does better in
 2085 some of these cases. It does this by remembering the last required literal
 2086 character in the pattern, and pre-searching the subject to ensure it is present
 2087 before running the real match. In other words, it applies a heuristic to detect
 2088 some types of certain failure quickly, and in the above example, if presented
 2089 with a string that has no trailing " it gives "no match" very quickly.
 2091 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
 2092 other alternatives are tried instead.
 2095 Version 2.06 09-Jun-99
 2096 ----------------------
 2098 1. Change pcretest's output for amount of store used to show just the code
 2099 space, because the remainder (the data block) varies in size between 32-bit and
 2100 64-bit systems.
 2102 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
 2103 start matching at. This allows lookbehinds to work when searching for multiple
 2104 occurrences in a string.
 2106 3. Added additional options to pcretest for testing multiple occurrences:
 2108    /+   outputs the rest of the string that follows a match
 2109    /g   loops for multiple occurrences, using the new startoffset argument
 2110    /G   loops for multiple occurrences by passing an incremented pointer
 2112 4. PCRE wasn't doing the "first character" optimization for patterns starting
 2113 with \b or \B, though it was doing it for other lookbehind assertions. That is,
 2114 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
 2115 the letter 'x'. On long subject strings, this gives a significant speed-up.
 2118 Version 2.05 21-Apr-99
 2119 ----------------------
 2121 1. Changed the type of magic_number from int to long int so that it works
 2122 properly on 16-bit systems.
 2124 2. Fixed a bug which caused patterns starting with .* not to work correctly
 2125 when the subject string contained newline characters. PCRE was assuming
 2126 anchoring for such patterns in all cases, which is not correct because .* will
 2127 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
 2128 DOTALL is set at top level; otherwise it knows that patterns starting with .*
 2129 must be retried after every newline in the subject.
 2132 Version 2.04 18-Feb-99
 2133 ----------------------
 2135 1. For parenthesized subpatterns with repeats whose minimum was zero, the
 2136 computation of the store needed to hold the pattern was incorrect (too large).
 2137 If such patterns were nested a few deep, this could multiply and become a real
 2138 problem.
 2140 2. Added /M option to pcretest to show the memory requirement of a specific
 2141 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
 2143 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
 2144 compiled in such a way that the backtracking after subsequent failure was
 2145 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
 2146 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
 2149 Version 2.03 02-Feb-99
 2150 ----------------------
 2152 1. Fixed typo and small mistake in man page.
 2154 2. Added 4th condition (GPL supersedes if conflict) and created separate
 2155 LICENCE file containing the conditions.
 2157 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
 2158 Perl, that is the internal \ allows the delimiter to be included in the
 2159 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
 2160 the final delimiter, add \ to the end of the pattern (to test the error).
 2162 4. Added the convenience functions for extracting substrings after a successful
 2163 match. Updated pcretest to make it able to test these functions.
 2166 Version 2.02 14-Jan-99
 2167 ----------------------
 2169 1. Initialized the working variables associated with each extraction so that
 2170 their saving and restoring doesn't refer to uninitialized store.
 2172 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
 2173 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
 2174 fix the problem.
 2176 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
 2177 calls, and wasn't printing the correct value for compiling calls. Increased the
 2178 default value of LOOPREPEAT, and the number of significant figures in the
 2179 times.
 2181 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
 2183 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
 2184 a building problem on Windows NT with a FAT file system.
 2187 Version 2.01 21-Oct-98
 2188 ----------------------
 2190 1. Changed the API for pcre_compile() to allow for the provision of a pointer
 2191 to character tables built by pcre_maketables() in the current locale. If NULL
 2192 is passed, the default tables are used.
 2195 Version 2.00 24-Sep-98
 2196 ----------------------
 2198 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
 2199 it any more.
 2201 2. Allow quantification of (?>) groups, and make it work correctly.
 2203 3. The first character computation wasn't working for (?>) groups.
 2205 4. Correct the implementation of \Z (it is permitted to match on the \n at the
 2206 end of the subject) and add 5.005's \z, which really does match only at the
 2207 very end of the subject.
 2209 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
 2211 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
 2212 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
 2213 localized options. All options to pcre_study() were also removed.
 2215 7. Add other new features from 5.005:
 2217    $(?<=           positive lookbehind
 2218    $(?<!           negative lookbehind
 2219    (?imsx-imsx)    added the unsetting capability
 2220                    such a setting is global if at outer level; local otherwise
 2221    (?imsx-imsx:)   non-capturing groups with option setting
 2222    (?(cond)re|re)  conditional pattern matching
 2224    A backreference to itself in a repeated group matches the previous
 2225    captured string.
 2227 8. General tidying up of studying (both automatic and via "study")
 2228 consequential on the addition of new assertions.
 2230 9. As in 5.005, unlimited repeated groups that could match an empty substring
 2231 are no longer faulted at compile time. Instead, the loop is forcibly broken at
 2232 runtime if any iteration does actually match an empty substring.
 2234 10. Include the RunTest script in the distribution.
 2236 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
 2237 discrepancies, some of which were old and were also with respect to 5.004. They
 2238 have now been fixed.
 2241 Version 1.09 28-Apr-98
 2242 ----------------------
 2244 1. A negated single character class followed by a quantifier with a minimum
 2245 value of one (e.g.  [^x]{1,6}  ) was not compiled correctly. This could lead to
 2246 program crashes, or just wrong answers. This did not apply to negated classes
 2247 containing more than one character, or to minima other than one.
 2250 Version 1.08 27-Mar-98
 2251 ----------------------
 2253 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
 2255 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
 2256 latter must appear before anything that relies on it in the pattern.
 2259 Version 1.07 16-Feb-98
 2260 ----------------------
 2262 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
 2263 repeat of a potentially empty string).
 2266 Version 1.06 23-Jan-98
 2267 ----------------------
 2269 1. Added Markus Oberhumer's little patches for C++.
 2271 2. Literal strings longer than 255 characters were broken.
 2274 Version 1.05 23-Dec-97
 2275 ----------------------
 2277 1. Negated character classes containing more than one character were failing if
 2278 PCRE_CASELESS was set at run time.
 2281 Version 1.04 19-Dec-97
 2282 ----------------------
 2284 1. Corrected the man page, where some "const" qualifiers had been omitted.
 2286 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
 2287 input syntax.
 2289 3. Fixed memory leak which occurred when a regex with back references was
 2290 matched with an offsets vector that wasn't big enough. The temporary memory
 2291 that is used in this case wasn't being freed if the match failed.
 2293 4. Tidied pcretest to ensure it frees memory that it gets.
 2295 5. Temporary memory was being obtained in the case where the passed offsets
 2296 vector was exactly big enough.
 2298 6. Corrected definition of offsetof() from change 5 below.
 2300 7. I had screwed up change 6 below and broken the rules for the use of
 2301 setjmp(). Now fixed.
 2304 Version 1.03 18-Dec-97
 2305 ----------------------
 2307 1. A erroneous regex with a missing opening parenthesis was correctly
 2308 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
 2309 on some systems.
 2311 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
 2312 it was reported that one broken compiler failed on the former because "code" is
 2313 also an independent variable.
 2315 3. The erroneous regex a[]b caused an array overrun reference.
 2317 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
 2318 fail on data ending with that character. (It was going on too far, and checking
 2319 the next character, typically a binary zero.) This was specific to the
 2320 optimized code for single-character negative classes.
 2322 5. Added a contributed patch from the TIN world which does the following:
 2324   + Add an undef for memmove, in case the the system defines a macro for it.
 2326   + Add a definition of offsetof(), in case there isn't one. (I don't know
 2327     the reason behind this - offsetof() is part of the ANSI standard - but
 2328     it does no harm).
 2330   + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
 2331     most of the places where whitespace preceded '#'. I have given up and
 2332     allowed the remaining 2 cases to be at the margin.
 2334   + Rename some variables in pcre to eliminate shadowing. This seems very
 2335     pedantic, but does no harm, of course.
 2337 6. Moved the call to setjmp() into its own function, to get rid of warnings
 2338 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
 2340 7. Constructs such as \d{8,} were compiling into the equivalent of
 2341 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
 2342 outcome, but in this particular case used more store than had been allocated,
 2343 which caused the bug to be discovered because it threw up an internal error.
 2345 8. The debugging code in both pcre and pcretest for outputting the compiled
 2346 form of a regex was going wrong in the case of back references followed by
 2347 curly-bracketed repeats.
 2350 Version 1.02 12-Dec-97
 2351 ----------------------
 2353 1. Typos in pcre.3 and comments in the source fixed.
 2355 2. Applied a contributed patch to get rid of places where it used to remove
 2356 'const' from variables, and fixed some signed/unsigned and uninitialized
 2357 variable warnings.
 2359 3. Added the "runtest" target to Makefile.
 2361 4. Set default compiler flag to -O2 rather than just -O.
 2364 Version 1.01 19-Nov-97
 2365 ----------------------
 2367 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
 2368 like /([ab]*)*/, that is, for classes with more than one character in them.
 2370 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
 2371 as /((?>a*))*/ (a PCRE_EXTRA facility).
 2374 Version 1.00 18-Nov-97
 2375 ----------------------
 2377 1. Added compile-time macros to support systems such as SunOS4 which don't have
 2378 memmove() or strerror() but have other things that can be used instead.
 2380 2. Arranged that "make clean" removes the executables.
 2383 Version 0.99 27-Oct-97
 2384 ----------------------
 2386 1. Fixed bug in code for optimizing classes with only one character. It was
 2387 initializing a 32-byte map regardless, which could cause it to run off the end
 2388 of the memory it had got.
 2390 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
 2393 Version 0.98 22-Oct-97
 2394 ----------------------
 2396 1. Fixed bug in code for handling temporary memory usage when there are more
 2397 back references than supplied space in the ovector. This could cause segfaults.
 2400 Version 0.97 21-Oct-97
 2401 ----------------------
 2403 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
 2405 2. Optimized negated single characters not to use a bit map.
 2407 3. Brought error texts together as macro definitions; clarified some of them;
 2408 fixed one that was wrong - it said "range out of order" when it meant "invalid
 2409 escape sequence".
 2411 4. Changed some char * arguments to const char *.
 2413 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
 2415 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
 2416 pcretest.
 2419 Version 0.96 16-Oct-97
 2420 ----------------------
 2422 1. Added a simple "pgrep" utility to the distribution.
 2424 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
 2425 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
 2426 where "ddd" means "one or more decimal digits".
 2428 3. Fixed serious bug. If a pattern had a back reference, but the call to
 2429 pcre_exec() didn't supply a large enough ovector to record the related
 2430 identifying subpattern, the match always failed. PCRE now remembers the number
 2431 of the largest back reference, and gets some temporary memory in which to save
 2432 the offsets during matching if necessary, in order to ensure that
 2433 backreferences always work.
 2435 4. Increased the compatibility with Perl in a number of ways:
 2437   (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
 2438       to request this handling. The option can be set at compile or exec time.
 2440   (b) $ matches before a terminating newline by default; an option
 2441       PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
 2442       mode). The option can be set at compile or exec time.
 2444   (c) The handling of \ followed by a digit other than 0 is now supposed to be
 2445       the same as Perl's. If the decimal number it represents is less than 10
 2446       or there aren't that many previous left capturing parentheses, an octal
 2447       escape is read. Inside a character class, it's always an octal escape,
 2448       even if it is a single digit.
 2450   (d) An escaped but undefined alphabetic character is taken as a literal,
 2451       unless PCRE_EXTRA is set. Currently this just reserves the remaining
 2452       escapes.
 2454   (e) {0} is now permitted. (The previous item is removed from the compiled
 2455       pattern).
 2457 5. Changed all the names of code files so that the basic parts are no longer
 2458 than 10 characters, and abolished the teeny "globals.c" file.
 2460 6. Changed the handling of character classes; they are now done with a 32-byte
 2461 bit map always.
 2463 7. Added the -d and /D options to pcretest to make it possible to look at the
 2464 internals of compilation without having to recompile pcre.
 2467 Version 0.95 23-Sep-97
 2468 ----------------------
 2470 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
 2471 \x20 at the start of a run of normal characters. These were being treated as
 2472 real characters, instead of the source characters being re-checked.
 2475 Version 0.94 18-Sep-97
 2476 ----------------------
 2478 1. The functions are now thread-safe, with the caveat that the global variables
 2479 containing pointers to malloc() and free() or alternative functions are the
 2480 same for all threads.
 2482 2. Get pcre_study() to generate a bitmap of initial characters for non-
 2483 anchored patterns when this is possible, and use it if passed to pcre_exec().
 2486 Version 0.93 15-Sep-97
 2487 ----------------------
 2489 1. /(b)|(:+)/ was computing an incorrect first character.
 2491 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
 2492 but not actually doing anything yet.
 2494 3. Treat "-" characters in classes that cannot be part of ranges as literals,
 2495 as Perl does (e.g. [-az] or [az-]).
 2497 4. Set the anchored flag if a branch starts with .* or .*? because that tests
 2498 all possible positions.
 2500 5. Split up into different modules to avoid including unneeded functions in a
 2501 compiled binary. However, compile and exec are still in one module. The "study"
 2502 function is split off.
 2504 6. The character tables are now in a separate module whose source is generated
 2505 by an auxiliary program - but can then be edited by hand if required. There are
 2506 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
 2507 toupper() in the code.
 2509 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
 2510 make them global. Abolish the function for setting them, as the caller can now
 2511 set them directly.
 2514 Version 0.92 11-Sep-97
 2515 ----------------------
 2517 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
 2518 (e.g. /a{1,3}/) was broken (I mis-optimized it).
 2520 2. Caseless matching was not working in character classes if the characters in
 2521 the pattern were in upper case.
 2523 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
 2525 4. Make PCRE_ANCHORED public and accept as a compile option.
 2527 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
 2528 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
 2529 pass them.
 2531 6. Give an error if bad option bits passed at compile or run time.
 2533 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
 2534 pcretest to cause it to pass that flag.
 2536 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
 2537 options, and the first character, if set.
 2539 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
 2542 Version 0.91 10-Sep-97
 2543 ----------------------
 2545 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
 2546 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
 2548 2. PCRE was looping on encountering an indefinitely repeated back reference to
 2549 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
 2550 Perl does - treats the match as successful.
 2552 ****