From @AlexDaniel

Code​:
sub foo { say 'X' };
&foo.wrap​:{ callsame; callsame; callsame };
foo;

Result​:
X

I expected this​:
X
X
X

Another interesting code snippet​:
sub foo { say 'X' };
&foo.wrap​:{ callsame; callsame; nextsame; say 'y' };
foo;

Which, surprisingly, prints this​:
X
y

IRC log​: http://irclog.perlgeek.de/perl6/2015-08-10#i_11036336 and
http://irclog.perlgeek.de/perl6/2015-08-11#i_11036346

From @jnthn

On Mon Aug 10 17​:19​:45 2015, alex.jakimenko@​gmail.com wrote​:

Code​:
sub foo { say 'X' };
&foo.wrap​:{ callsame; callsame; callsame };
foo;

Result​:
X

I expected this​:
X
X
X

Then you expected wrong. Discussion here​:

http://irclog.perlgeek.de/perl6/2015-11-12#i_11527573

This ticket did uncover that there isn't a way to actually do what you wanted, though, so we have nextcallee, which does. I've added a test for that in S12-methods/defer-call.t. The remaining open question I've asked (waiting for answer) is what callsame should do on an exhausted dispatch iterator.

Another interesting code snippet​:
sub foo { say 'X' };
&foo.wrap​:{ callsame; callsame; nextsame; say 'y' };
foo;

Which, surprisingly, prints this​:
X
y

I think it's dubious that nextsame lets control go on to the say 'y' even if the iterator is exhausted, so asked about this one too. Will update tickets and tests with the answers.

/jnthn

The RT System itself - Status changed from 'new' to 'open'

From @jnthn

On Fri Nov 13 09​:05​:08 2015, jnthn@​jnthn.net wrote​:

This ticket did uncover that there isn't a way to actually do what you
wanted, though, so we have nextcallee, which does. I've added a test
for that in S12-methods/defer-call.t. The remaining open question I've
asked (waiting for answer) is what callsame should do on an exhausted
dispatch iterator.

Larry answered that callsame/callwith returning Nil is the right thing, so those semantics stand.

Another interesting code snippet​:
sub foo { say 'X' };
&foo.wrap​:{ callsame; callsame; nextsame; say 'y' };
foo;

Which, surprisingly, prints this​:
X
y

I think it's dubious that nextsame lets control go on to the say 'y'
even if the iterator is exhausted, so asked about this one too. Will
update tickets and tests with the answers.

We now will never reach that 'y' even if we don't have anywhere to go; the nextsame will call a return of Nil.

Tests for these updated semantics added in S12-methods/defer-[call|next].t.

/jnthn

@jnthn - Status changed from 'open' to 'resolved'