From @masak

<hoelzro> in Perl 5, one can assign to $? in an END block to alter a
program's exit status. Is there an analogous construct in Perl 6?
<masak> hoelzro​: `exit($status);`
<masak> hm, `END exit(5)` hangs locally.
<hoelzro> masak​: that's what I tried, and also discovered =)
<masak> r​: END exit(5)
<camelia> rakudo-{parrot,moar} e56d94​: OUTPUT«(timeout)»
<camelia> ..rakudo-jvm e56d94​: OUTPUT«(timeout)Unhandled exception​:
java.lang.StackOverflowError [...]
* masak submits rakudobug

Moar and Parrot loop forever; the JVM blows the call stack. My guess
is that a call to &exit tries to run all the registered END blocks or
something.

I don't recall the spec mentioning this, but it feels fairly obvious
to me that we should support calling &exit from an END block, and that
it should Just Work. We probably should have wording in S32 to that
effect.

From @lizmat

On 21 Jul 2014, at 09​:25, Carl Mäsak (via RT) <perl6-bugs-followup@​perl.org> wrote​:

# New Ticket Created by "Carl Mäsak"
# Please include the string​: [perl #​122355]
# in the subject line of all future correspondence about this issue.
# <URL​: https://rt-archive.perl.org/perl6/Ticket/Display.html?id=122355 >

<hoelzro> in Perl 5, one can assign to $? in an END block to alter a
program's exit status. Is there an analogous construct in Perl 6?
<masak> hoelzro​: `exit($status);`
<masak> hm, `END exit(5)` hangs locally.
<hoelzro> masak​: that's what I tried, and also discovered =)
<masak> r​: END exit(5)
<camelia> rakudo-{parrot,moar} e56d94​: OUTPUT«(timeout)»
<camelia> ..rakudo-jvm e56d94​: OUTPUT«(timeout)Unhandled exception​:
java.lang.StackOverflowError [...]
* masak submits rakudobug

Moar and Parrot loop forever; the JVM blows the call stack. My guess
is that a call to &exit tries to run all the registered END blocks or
something.

I don't recall the spec mentioning this, but it feels fairly obvious
to me that we should support calling &exit from an END block, and that
it should Just Work. We probably should have wording in S32 to that
effect.

FWIW, I think the problem is that running the END blocks does not remove them from the list of END blocks.

Each END block should only be run *once*. Hence, execution of an END block should remove it from the list *before* executing it.

Pretty sure that would solve this problem.

Liz

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

From @usev6

This seems to work now​:

$ perl6-m -e 'END exit(5)'; echo $?
5
$ perl6-p -e 'END exit(5)'; echo $?
5
$ perl6-j -e 'END exit(5)'; echo $?
5

I added a test to S04-phasers/end.t with the following commit​: Raku/roast@7050e78bc7

I'm closing the ticket now.

@usev6 - Status changed from 'open' to 'resolved'