Front page | perl.perl5.porters |
Postings from January 2012
On 4 January 2012 23:23, demerphq <demerphq@gmail.com> wrote:
> On 4 January 2012 23:14, demerphq <demerphq@gmail.com> wrote:
>> On 4 January 2012 22:40, Ricardo Signes <perl.p5p@rjbs.manxome.org> wrote:
>>>
>>> \F and fc should be landing any time now, as I understand things.
>>>
>>> Karl has expressed concern a few times, because of the problems here:
>>>
>>> ~$ perl -E 'say "f" =~ qr"\L\x{66}"i ? 1 : 0'
>>> 1
>>>
>>> ~$ perl -E 'say "f" =~ qr"\U\x{66}"i ? 1 : 0'
>>> 0
>>>
>>> The \U and \L affect the spelling of \x (making it \X in the latter example),
>>> which is, if not obviously wrong,
>>
>> I am not sure if this has been pointed out before, and I have not
>> actually checked, but I am reasonably confident that this is a
>> (untested presumably) regression that was introduced somewhere in late
>> 5.9 or so.
>
> Gah. Its not. I should have checked first, but the similarity with
> previous bugs related to this patch is pretty interesting. Although I
> stand by my recommendation for solving the problem.
I checked, and doing:
diff --git a/toke.c b/toke.c
index 8fb6164..488759a 100644
--- a/toke.c
+++ b/toke.c
@@ -2925,7 +2925,7 @@ S_scan_const(pTHX_ char *start)
}
/* string-change backslash escapes */
- if (PL_lex_inwhat != OP_TRANS && *s && strchr("lLuUEQ", *s)) {
+ if (PL_lex_inwhat != OP_TRANS && !PL_lex_inpat && *s &&
strchr("lLuUEQ", *s)) {
--s;
break;
}
Should cause the relevant escapes to pass through to the regex engine
where they could be handled by the regex parser.
cheers,
Yves
--
perl -Mre=debug -e "/just|another|perl|hacker/"
Thread Previous
|
Thread Next