In perl.git, the branch yves/fix_131649 has been created
<http://perl5.git.perl.org/perl.git/commitdiff/06335364f4f871829eb590bdb9013f5eacee67b6?hp=0000000000000000000000000000000000000000>
at 06335364f4f871829eb590bdb9013f5eacee67b6 (commit)
- Log -----------------------------------------------------------------
commit 06335364f4f871829eb590bdb9013f5eacee67b6
Author: Yves Orton <[email protected]>
Date: Mon Jun 26 13:19:55 2017 +0200
work in progress fixing extended charclass parsing
the extended charclass parser makes some assumptions during the
first pass which are only true on well structured input, and it
does not properly catch various errors. later on the code assumes
that things the first pass will let through are valid, when in
fact they should trigger errors.
this is a work in progress patch needing review by karl
M pod/perlrecharclass.pod
M regcomp.c
commit 22f8cf2a7f2bf8409d4a533bc56bc8b99f6de4ba
Author: Yves Orton <[email protected]>
Date: Sun Jun 25 22:31:10 2017 +0200
regcomp.c: handle_regex_sets() - add DEBUG_PARSE and fixup 'depth' logic
'depth' is used to track the recursion depth during compilation,
and is used by things like DEBUG_PARSE() to show the compilation
process.
handle_regex_sets() was using its own 'depth' for two different purposes,
which is quite confusing.
At the same time, when we call handle_regex_sets() from reg() it is
important to increment 'depth'.
M regcomp.c
commit 05e580898736fca24304424e8e38015e174c689a
Author: Yves Orton <[email protected]>
Date: Sun Jun 25 22:13:36 2017 +0200
improve documentation of reg_data types
and fix an harmless issue where we use the wrong type for
certain data, although luckily it turns out not to matter.
M regcomp.c
-----------------------------------------------------------------------
--
Perl5 Master Repository
