Front page | perl.perl5.porters |
Postings from February 2022
Re: PSC #055 2022-02-18
Thread Previous
|
Thread Next
From:
demerphq
Date:
February 21, 2022 04:57
Subject:
Re: PSC #055 2022-02-18
Message ID:
CANgJU+XTo=76xxdN6FdkXMbJzqBJmVX79v0P5qQyGT3JV9Atdw@mail.gmail.com
On Mon, 21 Feb 2022 at 04:01, demerphq <demerphq@gmail.com> wrote:
>
>
> On Mon, 21 Feb 2022, 07:24 Neil Bowers, <neilb@neilb.org> wrote:
>
>> De-experimentalising variable-length look-behind (`experimental::vlb`).
>> See https://github.com/Perl/perl5/issues/18756.
>>
>
> I don't think this is ready for prime time yet. The feature is way under
> tested, and without my recent PR it is plain broken, and I'm pretty sure I
> can find other issues and work that needs to be done given a bit of time. I
> have a feeling when I start poking it with a stick later today I'm going to
> find more issues that will need fixing and we are rapidly approaching code
> freezes and etc. I just don't think it's reasonable to expect to get it
> into a shape where I could in good conscience sign off on it in the time
> available.
>
> I'll do my best but I am not optimistic. The regex engine is garly as heck
> and even though I now have rewrapped my brain around the state machine I
> expect the changes needed will be in the compiler (regcomp.c) which is even
> more gnarly than the state machine (regexec.c).
>
>
To repeat what I said in the ticket....
I retract this. I was *wrong*. (Rarely have I been so pleased to say
that!) I thought when I tested yesterday that (?<=x+) was allowed. But
further testing this morning shows it is not. The length restrictions
appear to be working correctly. Also I thought my patch may have broken
(?<=) but they did not.
Therefore and assuming no further issues are revealed in additional tests,
especially nested lookbehind/lookahead tests, I am comfortable with signing
this off.
Karl did a great job with the length restrictions. I apologize for doubting
him.
I will push a de-experimentalizing patch to go with my PR.
Sorry for the noise from my original mail. It was *very* hot here yesterday
and I came to incorrect conclusions while working on the fix for #19168.
cheers,
Yves
--
perl -Mre=debug -e "/just|another|perl|hacker/"
Thread Previous
|
Thread Next