Bug #123860 for Regexp-Grammars: This program doesn’t stop running

Mon Dec 11 04:13:13 2017 alex.karelas [...] gmail.com - Ticket created

Subject:	This program doesn’t stop running
Date:	Mon, 11 Dec 2017 11:12:57 +0200
To:	bug-Regexp-Grammars [...] rt.cpan.org
From:	Alexander Karelas <alex.karelas [...] gmail.com>

Which means i cannot use regexes inside require blocks, i think? But thats bad because how else can i say: i want this word to match one of these regexes but also not match one of those regexes? I see no other way to do it Show quoted text

> #!/usr/bin/env perl > > use v5.16; > use warnings; > use lib 'local/lib/perl5'; > > use Regexp::Grammars; > use Test::More; > use DDP; > use Data::Dumper; > > my $grammar = qr/ > > \A <item> \z > > <token: item> > (\S+) > <require: (?{ no Regexp::Grammars; $CAPTURE != m!! })> > > /; > > my $matches = 'Peter' =~ $grammar; > > my $struct = { > matches => $matches, > hashref => \%/, > errors => \@!, > }; > > p $struct;

Mon Dec 11 20:10:48 2017 damian [...] conway.org - Correspondence added

Subject:	Re: [rt.cpan.org #123860] This program doesn’t stop running
Date:	Tue, 12 Dec 2017 12:09:19 +1100
To:	bug-Regexp-Grammars [...] rt.cpan.org
From:	Damian Conway <damian [...] conway.org>

Hi Alexander. First of all, unless I've misunderstood your intent, I'm fairly sure you need: <require: (?{ no Regexp::Grammars; $CAPTURE !~ m!! })> That is: !~ instead of != Second, if that's the case, there's seems to be no point in doing a negative match against an empty regex, since an empty regex is special-cased to rematch the previous successful regex (which would either be in one of the module's you loaded--most likely inside Regexp::Grammars itself--or else a null match, if there was in fact no preceding successful match). Neither of those is a desirable outcome as far as I can see. I'd have expected you to want to write something like: <require: (?{ no Regexp::Grammars; $CAPTURE !~ m!Paul! })> ...which works perfectly well in my own testing. But perhaps I have misunderstood your intentions, in which case I'd be happy to have them explained to me and to help you achieve your goals with Regexp::Grammars. Damian

Mon Dec 11 20:10:49 2017 The RT System itself - Status changed from 'new' to 'open'

Tue Dec 12 00:55:08 2017 alex.karelas [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #123860] This program doesn’t stop running
Date:	Tue, 12 Dec 2017 07:54:31 +0200
To:	bug-Regexp-Grammars [...] rt.cpan.org
From:	Alexander Karelas <alex.karelas [...] gmail.com>

Indeed that works, but I thought that sibce empty regex doesnt work, that means a regex is not supposed to be there and evil things might happen in complex situations. Is that not the case? Sent from my iPhone Show quoted text

> On 12 Dec 2017, at 3:10 AM, damian@conway.org via RT <bug-Regexp-Grammars@rt.cpan.org> wrote: > > <URL: https://rt.cpan.org/Ticket/Display.html?id=123860 > > > Hi Alexander. > > First of all, unless I've misunderstood your intent, I'm fairly sure you > need: > > <require: (?{ no Regexp::Grammars; $CAPTURE !~ m!! })> > > That is: !~ instead of != > > Second, if that's the case, there's seems to be no point in doing a > negative match against an empty regex, since an empty regex is > special-cased to rematch the previous successful regex (which would > either be in one of the module's you loaded--most likely inside > Regexp::Grammars itself--or else a null match, if there was in fact no > preceding successful match). Neither of those is a desirable outcome as > far as I can see. > > I'd have expected you to want to write something like: > > <require: (?{ no Regexp::Grammars; $CAPTURE !~ m!Paul! })> > > ...which works perfectly well in my own testing. > > But perhaps I have misunderstood your intentions, in which case > I'd be happy to have them explained to me and to help you achieve > your goals with Regexp::Grammars. > > Damian >

Tue Dec 12 02:35:38 2017 DCONWAY [...] cpan.org - Status changed from 'open' to 'rejected'

Tue Dec 12 02:36:18 2017 damian [...] conway.org - Correspondence added

Subject:	Re: [rt.cpan.org #123860] This program doesn’t stop running
Date:	Tue, 12 Dec 2017 18:35:21 +1100
To:	bug-Regexp-Grammars [...] rt.cpan.org
From:	Damian Conway <damian [...] conway.org>

Show quoted text

> Indeed that works, but I thought that since empty regex doesnt work, > that means a regex is not supposed to be there and evil things might > happen in complex situations. Is that not the case?

Evil things might indeed happen in complex situations. They often do. And nested regex matches certainly qualify as "complex situations". :-) In practical terms, a <require:...> is probably not the right tool here. You will probably do better with a preliminary negative look-ahead on the regex you want not to match. Like so: my $grammar = qr/ \A <item> \z <token: item> *(?! Paul | Joh?n | Ringo+ | [George] \b ) # None of these allowed* (\S++) /; Hope this helps, Damian