From publiustemp-perl6compiler2@yahoo.com

This is blocking me implementing a simple 'plan "no_plan"' for tests.

  $ perl6 -e 'my $foo = "Goodbye cruel world!"; END { say $foo }'
  Scope not found for PAST​::Var '$foo' in

Revision​: 35423

Cheers,
Ovid
--
Buy the book - http://www.oreilly.com/catalog/perlhks/
Tech blog - http://use.perl.org/~Ovid/journal/
Twitter - http://twitter.com/OvidPerl
Official Perl 6 Wiki - http://www.perlfoundation.org/perl6

From @pmichaud

On Mon, Jan 12, 2009 at 12​:57​:44PM -0800, publiustemp-perl6compiler2@​yahoo.com (via RT) wrote​:

This is blocking me implementing a simple 'plan "no_plan"' for tests.

$ perl6 -e 'my $foo = "Goodbye cruel world!"; END { say $foo }'
Scope not found for PAST​::Var '$foo' in

Confirmed -- there are still issues with lexicals in a
variety of control blocks.

However, the easy workaround is to use "our" variables instead
of "my" if at all possible.

Pm

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

From @jnthn

Patrick R. Michaud wrote​:

Confirmed -- there are still issues with lexicals in a
variety of control blocks.

I skimmed the code for that today. I noticed that we compiled the blocks
right away and pushed them onto @​?END at compile time. Is there any
reason not to just compile them with everything else and give them a
loadinit that pushes them onto @​?END?

Jonathan

From @pmichaud

On Tue, Jan 13, 2009 at 05​:17​:16PM +0100, Jonathan Worthington wrote​:

Patrick R. Michaud wrote​:

Confirmed -- there are still issues with lexicals in a
variety of control blocks.

I skimmed the code for that today. I noticed that we compiled the blocks
right away and pushed them onto @​?END at compile time. Is there any
reason not to just compile them with everything else and give them a
loadinit that pushes them onto @​?END?

Sounds reasonable -- feel free to try that.

Pm

From @jnthn

On Tue Jan 13 08​:58​:08 2009, pmichaud wrote​:

On Tue, Jan 13, 2009 at 05​:17​:16PM +0100, Jonathan Worthington wrote​:

Patrick R. Michaud wrote​:

Confirmed -- there are still issues with lexicals in a
variety of control blocks.

I skimmed the code for that today. I noticed that we compiled the
blocks
right away and pushed them onto @​?END at compile time. Is there any
reason not to just compile them with everything else and give them a
loadinit that pushes them onto @​?END?

Sounds reasonable -- feel free to try that.

Did it and it works well, and resolves the original bug report.
Assigning to moritz for tests.

Thanks,

Jonathan

From @moritz

On Thu Jan 22 05​:25​:52 2009, jnthn@​jnthn.net wrote​:

On Tue Jan 13 08​:58​:08 2009, pmichaud wrote​:

On Tue, Jan 13, 2009 at 05​:17​:16PM +0100, Jonathan Worthington wrote​:

Patrick R. Michaud wrote​:

Confirmed -- there are still issues with lexicals in a
variety of control blocks.

I skimmed the code for that today. I noticed that we compiled the
blocks
right away and pushed them onto @​?END at compile time. Is there any
reason not to just compile them with everything else and give them a
loadinit that pushes them onto @​?END?

Sounds reasonable -- feel free to try that.

Did it and it works well, and resolves the original bug report.
Assigning to moritz for tests.

I can neither use outer lexicals in eval(), nor can I spawn (with
reasonable effort) another Rakudo and obtain its output, so all I can
test for atm is that "eval 'my $x; END { $x }'" works.
Which is now tested in S04-closure-traits/end.t.

Cheers,
Moritz

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