gcj 4.4 failures on fash.d.o but not on fano.d.o

[Petr Salinger <Petr.Salinger@seznam.cz>]

> > * openjdk on both kfreebsd-i386/kfreebsd-amd64
> >   man-power is missing, we use gcj similarly as hppa
>
> gcj 4.4 is currently broken on kfreebsd-amd64 per #576335.

The forked-daapd failed three times on fash, but build for 1st attempt on 
fano. It looks like still the same problem as bellow.

Aurelien prepared fix for get_phys_pages() in eglibc (not yet uploaded), 
it is related ?

Petr


On Thu, 29 Apr 2010, Petr Salinger wrote:

> Hi,
>
> it looks like jetty 6.1.24-2 just run into this problem too.
>
> Please could you investigate the output for
> "ulimit -a -S",  "ulimit -a -H" during build,
> memory sizes, etc.
>
> It also might be related to stack size of pthread,
> fixed in eglibc 2.10.2-7.
>
> Petr
>
> On Fri, 9 Apr 2010, Aurelien Jarno wrote:
>
> > On Thu, Apr 08, 2010 at 09:04:34PM +0200, Petr Salinger wrote:
> > > Hi,
> > >
> > > jakarta-log4j finally succeeded on fano, but previously failed on fash,
> > > brltty finally succeeded on fano, but previously failed on fash,
> > > voms have been tried only on fash (failure), but it builds fine
> > > for me locally.
> > >
> > > Please could you double check/recreate fasch chroot ?
> >
> > Both chroots on fasch and fano where broken/not clean, as the machines
> > were rebooted a few times du to the vnode bug.
> >
> > I have cleaned both chroots yesterday night, so I guess it's fine.
> >
> > > Are there any difference between fano and fash in (ulimit ?) setting ?
> >
> > The main difference is that fano is a virtual machine with only one 
> > CPU,
> > while fasch is a real machine with 4 CPU. Also fasch has more RAM. 
> > Other settings should be the same.