Front page | perl.perl5.porters |
Postings from May 2004
On Fri, May 21, 2004 at 10:47:54AM +0200, Rafael Garcia-Suarez wrote:
> tetryl@tokyoprogrammer.com wrote:
> >
> > I love my variable and am looking for various ways of using it.
> > I will show you a bit of code.
> >
> > $ perl -e 'my $x;$x->{foo}' # OK
> > $ perl -e '(my $x)->{foo}' # NG
> >
> > I expected that autovivification would be applied to the
> > variable in the later case as well as in the first case, but
> > the results were different.
>
> If you look at what contexts permit and forbid autovivification, you'll
> see that the situation in perl is currently complex and probably not
> very consistent. Anyway, in general lvalue contexts don't permit it; a
> lexical declaration can be seen as this; and you still can write, in a
> rather obfuscated way,
Thinking about this some more; it could be argued that since op.c has already
gone to the trouble of marking the padsv op with the OPpDEREF flag, that
its a bug in pp_padsv for then ignoring that flag if it happens also to
be marked with OPpLVAL_INTRO. That is to say, ($x)->{foo} autovivifies
but (my $x)->{foo} doesn't. Clearly the OP's patch was wrong because it
threw out the baby with the bathwater, but maybe something like the
following (untested):
--- pp_hot.c Mon May 3 23:57:33 2004
+++ pp_hot.c+ Fri May 21 11:58:22 2004
@@ -205,7 +205,7 @@
if (PL_op->op_flags & OPf_MOD) {
if (PL_op->op_private & OPpLVAL_INTRO)
SAVECLEARSV(PAD_SVl(PL_op->op_targ));
- else if (PL_op->op_private & OPpDEREF) {
+ if (PL_op->op_private & OPpDEREF) {
PUTBACK;
vivify_ref(PAD_SVl(PL_op->op_targ), PL_op->op_private & OPpDEREF);
SPAGAIN;
--
"You're so sadly neglected, and often ignored.
A poor second to Belgium, When going abroad."
-- Monty Python - "Finland"
Thread Previous
|
Thread Next