Front page | perl.perl5.porters |
Postings from September 2010
Re: fresh_perl failure
Thread Previous
|
Thread Next
From:
Dave Mitchell
Date:
September 20, 2010 14:23
Subject:
Re: fresh_perl failure
Message ID:
20100920212343.GF3932@iabyn.com
On Mon, Sep 20, 2010 at 04:45:06PM +0100, Nicholas Clark wrote:
> On 32 bit systems (for me - tried x86 FreeBSD and x86 Debian) I see test
> 53 of t/run/fresh_perl.t fail.
Hey boys and girls, gather round! It's time to play
"Global destruction Whack-a-Mole"(tm)!
I couldn't reproduce it myself, but I think its due to the
random ordering of the IO destructors. Fixed (hopefully) with commit
68b590d93559da1f7a0d3956202180d507013365
which now saves PL_defoutgv and PL_errgv to last.
> So it seems that the value from &afh is no longer getting destroyed (for some
> ordering of GVs)
It was getting destroyed; there just wasn't any STDOUT left to accept the
destructor message.
> Also the above output is exactly correct - there is no longer the "at ..."
> line after the stringification of the reference in the output of warn.
> I'm not sure if I like that change.
That appears to have happened somewhere between 5.13.0 and 5.13.1
--
"You may not work around any technical limitations in the software"
-- Windows Vista license
Thread Previous
|
Thread Next