Re: [PATCH] Refactoring to Sv*_set() macros - patch #4

Nicholas Clark <[email protected]> writes:
>On Sat, Apr 16, 2005 at 11:05:02AM -0500, Steve Peters wrote:
>> This is the fourth patch to migrate to the Sv*_set() macros.  This patch
>> takes cares of SVPVX() assignments, mostly.  This does not do anything w=
ith
>> cases where SvPVX() is dereferenced before being assigned.
>
>Thanks applied (24239)
>
>With a little bug fix to Perl_rvpv_dup that only shows up as a core dump
>under ithreads when built with assertions.
>
>
>Grrr. We have an assert() macro that uses PerlIO. And PerlIO doesn't work
>during ithread cloning. This is a stupid state of affairs.

Agreed.
=20=20
Obvious thing to do now stdio and PerlIO can co-exist revert to assert=20
that uses stdio.

Other choice is something like PerlIO_debug() but I thing that uses=20
"our" SV aware sprintf() wrapper so that may not work during clone=20
either.=20