Hi,
2011/10/21 Simon McVittie <s...@debian.org>:
> On Fri, 21 Oct 2011 at 13:33:14 +0900, Nobuhiro Iwamatsu wrote:
>> 3 of 59 tests failed
>
> test-channel, test-contact-lists and test-contacts timed out.
Yes.
>
> How fast is your sh4 buildd, compared with the slower/more embedded official
> architectures like mips/mipsel/armel?
SH4 is very slow compared with other inclusion architecture.
A CPU clock is 600Mhz.
Now, there is not a CPU with the clock beyond this of SH4.
These CPUs are used also embedded.
However, since these are used for servers, they use CPU faster than
CPU for other embedded for buildd.
>
> It'd probably be sufficient to increase the arbitrary timeout in the tests
> (which is only there to make sure the tests terminate eventually, even if they
> deadlock or otherwise stall).
>
> I see the last succesful build (0.11.11-2) took 3.5 hours, and 0.16.0 took
> 4 hours before failing. By way of comparison, 0.16.0 takes about an hour on
> mips/mipsel/armel, and 11 minutes on i386. Is sh4 really an order of
> magnitude slower than our slowest official architectures?!
>
Yes, SH4 is slower than the architecture currently supported officially.
As you have pointed out, I think that a problem is solved by extending
the time of timeout.
Would you correspond this?
Best regards,
Nobuhiro
--
Nobuhiro Iwamatsu
iwamatsu at {nigauri.org / debian.org}
GPG ID: 40AD1FA6
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
