On Sun Mar 23 12:39:07 2014, haarg wrote:
Show quoted text> On Sat Mar 22 07:52:28 2014, REHSACK wrote:
> > Haarg@ wasn't satisfied by naming implementations by author.
> >
> > This ticket is for finding better names and associations.
>
> I have two problems with the current names.
>
> First is that they provide no useful information about their different
> behaviors. You would have to reference the documentation every time
> to know what the intention is.
>
Okay, now we know why you're unlucky. I'm open for proposals to match the intension better (and I'm happy to find a way to kick the "sno" impl!).
Show quoted text> Second is that they are inaccurate. The behavior change that lead to
> this issue is RT#40905. The change was made in 2009 by the original
> author, Tassilo, for version 0.24. It was then accidentally reverted
> during a maintainership shuffle for Adam Kennedy's 0.24 and 0.26
> releases. The new behavior was re-introduced in version 0.28,
> released 2 weeks after the new 0.24. So the behavior change is
> attributable to Tassilo if anyone.
This is wrong. The change Tassilo made in 0.24 is:
void
any (code,...)
SV *code;
PROTOTYPE: &@
CODE:
{
dMULTICALL;
register int i;
GV *gv;
HV *stash;
I32 gimme = G_SCALAR;
- SV **args = &PL_stack_base[ax];
CV *cv;
if (items <= 1)
XSRETURN_UNDEF;
cv = sv_2cv(code, &stash, &gv, 0);
PUSH_MULTICALL(cv);
SAVESPTR(GvSV(PL_defgv));
for(i = 1 ; i < items ; ++i) {
- GvSV(PL_defgv) = args[i];
+ GvSV(PL_defgv) = PL_stack_base[ax+i];
MULTICALL;
if (SvTRUE(*PL_stack_sp)) {
POP_MULTICALL;
XSRETURN_YES;
}
}
POP_MULTICALL;
XSRETURN_NO;
}
Same for most other functions. The change from "undef" to "true/false" has been introduced by Alias.
RT for rt.cpan.org