| Note: these documents may be out of date. Do not use as reference! |
To see what is currently happening visit http://www.perl6.org/
Localising Paren Counts in qr()s.
Maintainer: Richard Proctor <richard@waveney.org> Date: 24 Sep 2000 Last Modified: 30 Sep 2000 Mailing List: perl6-language-regex@perl.org Number: 276 Version: 2 Status: Frozen
The Paren Counts and backreferences should be localised in each qr(), to prevent surprises when qr()s are used in combination.
TomCs perl storm #0040 has:
> Figure out way to do > > /$e1 $e2/ > > safely, where $e1 might have '(foo) \1' in it. > and $e2 might have '(bar) \1' in it. Those won't work.
Me: If e1 and e2 are qr// type things the answer might be to localise the backref numbers in each qr// expression. Use of assignment in a regex and named backrefs (RFC 112) would make this a lot safer.
Hugo: I think it is reaonable to ask whether the current handling of qr{} subpatterns is correct:
perl -wle '$a=qr/(a)\1/; $b=qr/(b).*\1/; /$a($b)/g and print join ":", $1, pos for "aabbac"' a:5
I'm tempted to suggest it isn't; that the paren count should be local to each qr{}, so that the above prints 'bb:4'. I think that most people currently construct their qr{} patterns as if they are going to be handled in isolation, without regard to the context in which they are embedded - why else do they override the embedder's flags if not to achieve that?
The problem then becomes: do we provide a mechansim to access the nested backreferences outside of the qr{} in which they were referenced, and if so what syntax do we offer to achieve that? I don't have an answer to the latter, which tempts me to answer 'no' to the former for all the wrong reasons. I suspect (and suggest) that complication is the only reason we don't currently have the behaviour I suggest the rest of the semantics warrant - that backreferences are localised within a qr().
I lie: the other reason qr{} currently doesn't behave like that is that when we interpolate a compiled regexp into a context that requires it be recompiled, we currently ignore the compiled form and act only on the original string. Perhaps this is also an insufficiently intelligent thing to do.
MJD: Interpolated qr() items shouldn't be recompiled anyway. They should be treated as subroutine calls. Unfortunately, this requires a reentrant regex engine, which Perl doesn't have. But I think it's the right way to go, and it would solve the backreference problem, as well as many other related problems.
Me: You can access the nested backreferences outside of the qr{} in which they were referenced by use of the named backref see RFC 112.
The paren count in each qr() is localised to each qr().
There is no way to access the nested backrefernces outside of the qr() by number they may be accessed by name (see RFC 112).
The regex engine must be made re-entrant.
The regex compiler should not need to recompile qr()s when used as part of another regex.
V2 - Added a some more comments under implementation and froze - no significant changes.
The Regex engine must be made re-entrant.
The expansion of variables in regexes must be driven by the regex compiler (Same problem as for RFCs 112, 166 ...)
Hugo:
None of these are necessarily true - we could change the overloading of the Regexp object instead. Currently we have:
my $re = qr{pattern};
print "$re";
.. giving 'pattern' by overloading stringification. If we overload it instead to give '(??{ $re })' (or a moral equivalent) we have a nasty hack, it is true, but it could allow us to defer the much trickier proper solution. Of course it breaks every other use of the string value, and I'm not sure how big a problem that might be.
Perlstorm #0040 from TomC.
RFC 112: Assignment within a regex