lst_ho...@kwsoft.de:
> Zitat von lst_ho...@kwsoft.de:
>
> > Zitat von Wietse Venema <wie...@porcupine.org>:
> >
> >> lst_ho...@kwsoft.de:
> >>>>> If you have time, can you try:
> >>>>>
> >>>>> stuff = 0.7 * ((double) var_psc_pre_queue_limit);
> >>>>> stuff = 0.9 * ((double) var_psc_pre_queue_limit);
> >>>>>
> >>>>> Thanks,
> >>>>
> >>>> In case that does not help, declaring the hiwat and lowat variables
> >>>> as double instead of int (in postscreen.c and postscreen.h) could
> >>>> make a difference.
> >>>>
> >>>> Perhaps the bug is triggered when the target of assignment is int.
> >>>>
> >>>> Either way it would be a compiler bug.
> >>>
> >>> Luckily we do not need aCC compiled Postfix on HPUX but i wonder what
> >>> else this compiler would break. I have to talk to the devs if they are
> >>> aware of this strange behaviour. Maybe a recommendation should be
> >>> included to better use gcc on HPUX.
> >>
> >> Were you able to make it work? How? I'd rather not have other people
> >> go through this same thing again.
> >
> > As said this HPUX machine is not a mail server but a developer
> > machine. Beside the HPUX aCC compiler there is also the GNU
> > toolchain with gcc. Using gcc (version 4.2.4) yield to correct
> > results. I will have a look if i can test tomorrow the other
> > workarounds (declaring as double) or maybe get a update for aCC. I
> > let you know if i can get it to work with the HP compiler.
>
>
> After this mess i installed the latest aCC Update (aCC 6.06 --> aCC
> 6.15) which include the ANSI-C compiler from HP and started from
> scratch with the RC3 source.
>
> Eh voila:
>
The logging looks OK.
Thanks, I will leave in the panic() check for this compiler bug.
The solution is available so there is no need to jump hoops in
Postfix.
Wietse
PS: WATCHDOG_PIPE is now turned on by default for all systems.
