2.99_04 -- Win32 test exit code still broken

[Andy Armstrong]

On 12 Oct 2007, at 03:38, David Golden wrote:
> On 10/11/07, Andy Armstrong <n at rciss.us> wrote:
>> On 12 Oct 2007, at 02:58, David Golden wrote:
>>> I think that "-1" is a special wait code and should be checked  
>>> for and
>>> handled explicitly.  (See perldoc -f wait and waidpid).  I'm not
>>> convinced that's an "error" condition.
>>
>> Aha! Penny drops. Yes, I'm sure you're right, thanks :)
>
> Just to clarify -- I think I've determined that this can happen
> somehow when some Perl process/subprocess or DLL gets wedged when
> interrupted with CTRL-C.  Even killing processes doesn't help, but a
> restart does clear it up.
>
> So it would appear to be some sort of Perl bug or strange Windows
> process-spawning interaction, but it's not technically a "fail".

Yes - and I was thinking of -1 being returned from waitpid, but it's  
actually $? that's -1 of course.

Hmm. D'you think we should / can do anything about it?

-- 
Andy Armstrong, Hexten