From @masak

<Timbus> m​: sub a {"foo"}; sub b {"bar"~callsame}; &a.wrap(&b); say a();
<camelia> rakudo-moar 90cd58​: OUTPUT«foo␤»
<masak> um.
<Timbus> optimizer ate it
<masak> that would be my guess, too.
* masak submits rakudobug
<moritz> not a bug
* masak doesn't submit rakudobug
<moritz> iirc you need a special pragma if you want to replace lexials
<masak> oh.
<Timbus> that would, not be good for the test suite then
<Timbus> wait is that 'use soft;'
<moritz> m​: use soft; sub a {"foo"}; sub b {"bar"~callsame};
&a.wrap(&b); say a();
<camelia> rakudo-moar 90cd58​: OUTPUT«barfoo␤»
<moritz> Timbus​: aye
<moritz> most of our optimizations rely on the fact the lexicals are
restricted in some ways at compile time
<masak> as they should.
<TimToady> seems to me that &a.wrap(&b); ought to fail if &a has been
hardened at CHECK time, and that's the bug that masak++ should submit
* masak submits &a.wrap(&b) rakudobug

From @smls

On Fri, 30 May 2014 09​:51​:18 -0700, masak wrote​:

<Timbus> m​: sub a {"foo"}; sub b {"bar"~callsame}; &a.wrap(&b); say a();
<camelia> rakudo-moar 90cd58​: OUTPUT«foo␤»

This now works​:

  ➜ sub a {"foo"}; sub b {"bar"~callsame}; &a.wrap(&b); say a();
  barfoo

Is TimToady's recommendation of forbidding `.wrap` on routines that aren't under `use soft`, still relevant?

This is Rakudo version 2017.08-104-g76f1d8970 built on MoarVM version 2017.08.1-148-g1059eed1
implementing Perl 6.c.

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