Autrijus,

> > We see the problems using perl 5.8.1, so the cause is no doubt 
> > in pp or perl (or both) code which hasn't changed recently.
> 
> This worries me -- the bug Vlad reported is specific to pre-5.8.1
> perls, caused by an unneccessary line in myldr/internals.c:
> 
>     SvSETMAGIC(GvSV(tmpgv));
> 
> But if something happens to perl 5.8.1, then it must be another bug
> altogether.

I was referring to either or both bugs involving loss of 
very short command-line parameters in @ARGV, or outright 
failure of dynamic loading with empty command line (due to 
apparent corruption of internal env). I was merely guessing 
that there was a correlation with Vlad's bug.

> As I (well, really mostly Alan Stewart) have cleaned up the environment
> processing logic greatly in PAR 0.81 (and now PAR 0.82), I'm curious about
> whether the heisenbug you mentioned still exists -- I tried to duplicate
> it here and can no longer see it, but perhaps I didn't try hard enough. :-)

Yes, when we looked into it before, we had to try hard. It 
would appear and disappear, sometimes in connection with 
small changes to the env before running.

Unfortunately, I don't expect to have time to investigate 
this again in the near future, though I'll try to do so at 
some point and report back.

As I reported before, here, in our application where those 
bugs were causing problems, we ultimately added a dummy 
(no-op) argument to the perl script, for the sole purpose 
of guaranteeing that the pp command line is always long 
enough to miss hitting the sporadic problems.

-- Mark

Reply via email to