Re: Yorick FTBFS: DM needs access to MIPS box

[Thibaut Paumard <mlotpot.news@free.fr>]

Le 3 mai 10 à 04:30, Shane McDonald a écrit :

> Hi Thibaut:
>
> On Thu, Apr 29, 2010 at 3:45 AM, Thibaut Paumard  
> <mlotpot.news@free.fr> wrote:
> > Hi,
> >
> > I have a build failure on mpis (segfault during test suite):
> > https://buildd.debian.org/fetch.cgi?&pkg=yorick&ver=2.1.06%2Bdfsg-1&arch=mips&stamp=1271731275&file=log
> >
> > > relocate/bin/yorick -batch debian/check.i
> > >  Begin Yorick parser test...
> > >  [...]
> > >  Test catch function...
> > > make: *** [check] Segmentation fault
[...]
> >
> I believe it may be related to the lack of an FPU on the buildd  
> machine.
> In fact, I have been able to come up with a simple testcase that  
> causes
> yorick to fail with a segmentation fault, both on the version in sid
> as well as the version currently in etch.
>
> My simple test is to start yorick up interactively, and enter:
>
>  5.0 / 0.0
>
> at the prompt.  On a mips (or mipsel) box without an FPU,
> this causes a segmentation fault.  With an FPU, it catches
> a SIGFPE and allows me to enter debug mode.
>
> I don't know the configuration of the mips buildd machines,
> but I suspect the ones that yorick has been built on in the past
> have had FPUs, and the one that it was just built on did not.
>
> I have posted a message to the linux-mips mailing list
> asking about this.  Hopefully, somebody there will
> be able to provide some insight.
>
> http://www.linux-mips.org/archives/linux-mips/2010-05/msg00005.html
>
> Shane McDonald

OK, thanks, that is very helpful. Please keep me posted if you ever  
get an answer to your post. In the meantime I will deactivate this  
test on mips since this failure is not a critical bug.

Best regards, Thibaut.