Bug#522099: [perl #64326] panic: MUTEX_LOCK (22) [op.c:453] in Perl 5.10.0 on Debian

[Dave Mitchell]

On Mon, Apr 20, 2009 at 12:58:33AM +0100, Dave Mitchell wrote:
> On Thu, Apr 02, 2009 at 09:07:10PM +0300, Niko Tyni wrote:
> > On Tue, Mar 31, 2009 at 11:38:15AM -0700, David F. Skoll wrote:
> > > In Perl 5.10.0, it does not seem possible to create an interpreter,
> > > destroy it, and create another in the same process by following the
> > > perlembed instructions.  The attached test case works fine under
> > > Debian Etch (perl 5.8.8); running "./configure && make && ./embperl"
> > > prints "I'm very happy!" twice.  On lenny running Perl 5.10.0, the first
> > > one works, but then it prints:
> > > 
> > > panic: MUTEX_LOCK (22) [op.c:453] at script.pl line 1.
> > > BEGIN failed--compilation aborted at script.pl line 1.
> > > panic: MUTEX_LOCK (22) [op.c:453] at script.pl line 1.
> > 
> > This is also <http://bugs.debian.org/522099>, and as discussed there,
> > it seems to be a misunderstanding about the needed arguments for
> > PERL_SYS_INIT3(). The test case calls the macro (twice) with the
> > perl_parse() arguments, but apparently the "real" main() argc, argv and
> > env are the right ones.
> > 
> > If this is correct, maybe the attached documentation patch would help
> > to avoid similar confusion in the future?
> 
> Thanks, applied as 174c68d32e9e7f904e35e6bf157e59aa67b468ab

Or, since I forgot to push, then later rebased, that's actually
31f6f5aa00011d88da3c64e4e1cb9629c8334dda

-- 
The Enterprise's efficient long-range scanners detect a temporal vortex
distortion in good time, allowing it to be safely avoided via a minor
course correction.
    -- Things That Never Happen in "Star Trek" #21