Mailing List Archive

>>>>> "Gerd" == Gerd Knops <gerti@BITart.com> writes:

Gerd> After countless hours of fruitless debugging of perl code, I finally
Gerd> gave up. MULTIPLICITY does NOT work, somewhere multiple instances of
Gerd> perl interpreters interfere with each other. My understanding of the
Gerd> perl code is not sufficient to even guess, what is causing the
Gerd> trouble.

Gerd> After some changes (and accepting some performance loss) I could
Gerd> 'unnest' my perl contexts, so that now I only need one Interpreter
Gerd> at a time. This allows me also to get rid of EMBED, which is yet
Gerd> another can of worms...

Gerd> To bad, that nobody on this forum could be of any help.

Sorry, that I can't be of any help either, but let me suggest, you put
together a small test case that shows what exactly happens. The test
case should, of course, not rely on external libraries, so everybody
on the list has a chance to play with the test.

In the past it has been proved, that this helps a lot.

Gerd> Gerd Knops

andreas

Andreas writes:

> Sorry, that I can't be of any help either, but let me suggest, you
> put together a small test case that shows what exactly happens. The
> test case should, of course, not rely on external libraries, so
> everybody on the list has a chance to play with the test.

Of course I thought about that, but that is easier said than done.

The package I am working on, hooks perl5 into the NEXTSTEP
Objective-C RunTime system, and the problem I had occurred within
a NEXTSTEP application called InterfaceBuilder, and that after a
lengthy procedure of doing some things. I fear, that if I cut out
the guts for a test, that due to the reduction and simplification of
code the problem wouldn't even show up...

But as soon as some more pressing matters are taken care of (like
getting some alpha code out of the door), I will try.

Thanks for the suggestion!

Gerd Knops