On Wed, Jun 26, 2002 at 05:52:25PM +0100, Hugo van der Sanden wrote:
I wrote:
:Attached patch passes all existing tests here, as well as some new ones.
Whoops, crossed in the post. My patch was written against @17356;
it may not be necessary after #17358, but the extra tests might be
worth keeping.
I notice that you've used this code:
SV *nsv = newSVpvn(c, clen);
#17362 adds sv_2mortal() here.
if (PL_encoding)
sv_recode_to_utf8(nsv, PL_encoding);
else
sv_utf8_upgrade(nsv);
Should I be using the same approach in regexec.c:S_to_utf8_substr()
and S_to_byte_substr()? They currently don't know about PL_encoding.
Probably. The PL_encoding is the encoding pragma that'll cause
upgrading to Unicode to assume some other character set than Latin 1
(which sv_utf8_upgrade assumes).
Also, does your version need the additional
SvSetMagicSV(nsv, dstr)
Probably. I'll try whether it disturbs anything, I guess no since
we'd need some magical Unicode string?
that I borrowed from there in my version of the patch?
Hugo
--
$jhi++; # http://www.iki.fi/jhi/
# There is this special biologist word we use for 'stable'.
# It is 'dead'. -- Jack Cohen
