jakarta-oro-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Daniel F. Savarese" <...@savarese.org>
Subject Re: Re[2]: substitution: what stands for the whole match
Date Tue, 08 Jan 2002 17:47:27 GMT

In message <13733088228.20011223235313@zolshar.ru>,                  writes:
>As i am an ordinary stuped user, i don't want to know of any subtle matter.
>I just plain know that in perl one can use "$&" in a substution,
>while i can't, and this frustrates me a lot :-)))

Anyone else think $& is a must?  In the end, user needs has to drive
feature development.  Anyone want to implement it?  I guess further
discussion about this should move to oro-dev.  I'm not sure that I'm
going to go ahead and implement it, but if someone else does, I wouldn't
oppose applying the patch.

>And, finally, the real subject.
>I've done a part of regex testing framework , namely a suit based
>on perl5.003_07/t/op/re_tests(class junit.regex.Perl5TestSuite).
>It's a very basic suit, but it runs and gives an excellent result for ORO.
>What i have got:
>package - jakarta-oro-2.0.4.jar
>tests - 274
>unexpected errors - 0 !
>Failures: 20
>ALL failures are of the same type - illegal compilation
>of zero-with infinite quantifiers(see details below), IMHO
>we can ignore them all.

I think Perl 5.6 does something about the zero-width infinite
quanitifiers but Perl 5.003 would happily accept them, but maybe
I'm misremembering.  This is why there are a few Bugzilla tickets that I
claimed weren't bugs but would be fixed later.  They all involved
expressions that could produce repeated matching of nothing during
backtracking of alternations.

>And, the last question, where to upload the stuff.

Do you have a web page somewhere?  The process would basically be this.
The code would have to be reviewed and tried out.  If it's useful (which
it probably is), we'd add it to the repository.  Since I want to spend
what little time I have on starting the Perl 5.6 feature set and writing
a user's guide, you will have effectively volunteered to maintain the
unit tests.  If you would be up to the task, we'd have to make you a


To unsubscribe, e-mail:   <mailto:oro-user-unsubscribe@jakarta.apache.org>
For additional commands, e-mail: <mailto:oro-user-help@jakarta.apache.org>

View raw message