Skip Menu |

Preferred bug tracker

Please email the preferred bug tracker to report your issue.

This queue is for tickets about the PPIx-Regexp CPAN distribution.

Report information
The Basics
Id: 82140
Status: resolved
Priority: 0/
Queue: PPIx-Regexp
People
Owner: Nobody in particular
Requestors: CHORNY [...] cpan.org
Cc:
AdminCc:
Bug Information
Severity: Important
Broken in: 0.028
Fixed in: (no value)

History Show all quoted text
  Thu Dec 20 15:04:35 2012 CHORNY [...] cpan.org - Ticket created
Subject: incorrect parsing of \?
PPIx::Regexp->new( '/(\?|I)/' ) parses to PPIx::Regexp failures=0 PPIx::Regexp::Token::Structure '' PPIx::Regexp::Structure::Regexp / ... / PPIx::Regexp::Structure::BranchReset (\?| ... ) PPIx::Regexp::Token::Literal 'I' PPIx::Regexp::Token::Modifier '' and requires perl 5.009005. This is incorrect. -- Alexandr Ciornii, http://chorny.net
  Thu Dec 20 17:35:05 2012 wyant [...] cpan.org - Correspondence added
On Thu Dec 20 15:04:35 2012, CHORNY wrote: Show quoted text
> PPIx::Regexp->new( '/(\?|I)/' ) parses to > PPIx::Regexp failures=0 > PPIx::Regexp::Token::Structure '' > PPIx::Regexp::Structure::Regexp / ... / > PPIx::Regexp::Structure::BranchReset (\?| ... ) > PPIx::Regexp::Token::Literal 'I' > PPIx::Regexp::Token::Modifier '' > > and requires perl 5.009005. This is incorrect. >
Well, branch reset requires 5.9.5, but this is clearly not a branch reset. Looks like the back slash is getting ignored somehow. Good catch. Thanks. Judging by the comments, I was trying to handle '?(\?|I)?'.
  Thu Dec 20 17:35:05 2012 The RT System itself - Status changed from 'new' to 'open'
  Thu Dec 20 19:10:38 2012 wyant [...] cpan.org - Correspondence added
Version 0.028_01 just went to PAUSE, and should be available Real Soon Now. Please let me know how you like it. Under the assumption that this turned up in a Padre bug: With version 0.028_01 of PPIx::Regexp, perlcritic will complain that /(\?|I)/ should be written as a character class. The recommendation is [\?I] rather than the more-correct [?I], but that looks like a wart to me rather than an actual bug. Something will have to be done in Perl::Critic, but what kind of support PPIx::Regexp should give Perl::Critic for this is something I will have to think about.
  Thu Dec 20 19:10:38 2012 wyant [...] cpan.org - Status changed from 'open' to 'patched'
  Fri Dec 28 08:20:04 2012 CHORNY [...] cpan.org - Correspondence added
On Thu Dec 20 19:10:38 2012, WYANT wrote: Show quoted text
> Version 0.028_01 just went to PAUSE, and should be available Real Soon > Now. Please let me know how you like it.
Yes, it's fixed. But there are similar bugs in parsing: /(\?>I)/ and /(\?:I)/ -- Alexandr Ciornii, http://chorny.net
  Fri Dec 28 08:20:05 2012 The RT System itself - Status changed from 'patched' to 'open'
  Fri Dec 28 15:33:26 2012 wyant [...] cpan.org - Correspondence added
On Fri Dec 28 08:20:04 2012, CHORNY wrote: Show quoted text
> On Thu Dec 20 19:10:38 2012, WYANT wrote:
> > Version 0.028_01 just went to PAUSE, and should be available Real Soon > > Now. Please let me know how you like it.
> > Yes, it's fixed. > But there are similar bugs in parsing: > /(\?>I)/ and /(\?:I)/ >
Right you are. And more, I suspect. I have worked a bit on your first example, but have decided a refactoring is in order.
  Fri Dec 28 18:23:19 2012 wyant [...] cpan.org - Correspondence added
The /(\?:I)/ case is going to be harder to get completely right, though I can certainly do better than I do now. How much do you care about regular expressions whose delimiters are letters (e.g. 'm xfoox', as a bizarre variant of 'm/foo/')?
  Mon Dec 31 18:27:59 2012 wyant [...] cpan.org - Correspondence added
Version 0.028_02 went to CPAN this afternoon. This contains a serious effort to fix all the '(?...)' parsing. There is a restriction, which I have not documented, but may before 0.029: the use of word characters as regular expression delimiters is not supported. That is, 'm i(?\i:...)i' is not going to parse as one would wish, and I have no real notion, at the moment, how to fix it.
  Mon Dec 31 18:28:17 2012 wyant [...] cpan.org - Status changed from 'open' to 'patched'
  Fri Jan 04 14:13:31 2013 CHORNY [...] cpan.org - Correspondence added
On Fri Dec 28 18:23:19 2012, WYANT wrote: Show quoted text
> The /(\?:I)/ case is going to be harder to get completely right, though > I can certainly do better than I do now. > > How much do you care about regular expressions whose delimiters are > letters (e.g. 'm xfoox', as a bizarre variant of 'm/foo/')?
Personally I don't care, and I think that almost all users of Perl::MinimumVersion will also not care. -- Alexandr Ciornii, http://chorny.net
  Fri Jan 04 14:13:34 2013 The RT System itself - Status changed from 'patched' to 'open'
  Fri Jan 04 14:13:34 2013 CHORNY [...] cpan.org - Status changed from 'open' to 'patched'
  Mon Jan 14 00:36:14 2013 wyant [...] cpan.org - Correspondence added
PPIx::Regexp 0.029 just went to PAUSE, and should appear on a CPAN mirror near you Real Soon Now. It still does not correctly handle regular expressions delimited by alphabetic characters. I thought about documenting this as a restriction, since I'm not sure how to handle it. But we'll see if it comes up as a real-life issue. Maybe an actual use case will stimulate the ol' brain cells.
  Mon Jan 14 00:36:15 2013 The RT System itself - Status changed from 'patched' to 'open'
  Mon Jan 14 00:36:41 2013 wyant [...] cpan.org - Status changed from 'open' to 'patched'
  Fri Jan 25 11:12:20 2013 wyant [...] cpan.org - Correspondence added
There being no further correspondence, I'm calling this one resolved. If you find something else that looks like the same problem, please feel free to re-open the ticket. If it turns out you need PPIx::Regexp to be able to parse regexps that use alphabetic characters, please open a new ticket. Thank you very much for helping to improve PPIx::Regexp.
  Fri Jan 25 11:12:20 2013 The RT System itself - Status changed from 'patched' to 'open'
  Fri Jan 25 11:12:20 2013 wyant [...] cpan.org - Status changed from 'open' to 'resolved'