Patchbot hangs when looking at tickets

[Simon King]

Hi!

In the last few days, I noticed that the patchbot produces red blobs,
since it hangs when testing whether or not a dependency has already been
merged. Example:

When testing #12215, the log says
2012-08-18 14:27:31 +0800
Looking at #12215
Looking at #11115
#11115 already applied (5.0.beta0 <= 5.3.beta2)
Looking at #11900
http://trac.sagemath.org/sage_trac/ticket/11900
Traceback (most recent call last):
...
timeout: timed out
2012-08-18 14:27:54 +0800
22 seconds

#11900 is merged, but the patchbot does not recognize it.

How can this be fixed? Is trac too slow? Is the patchbot too impatient?

Cheers,
Simon

[Keshav Kini]

The particular red blob you are quoting the log from was produced by the
patchbot on arando, which I am administrating. It could be that trac is
too slow, but I suspect it might have something to do with the network
connection here in NTU... it doesn't always have such great latency, and
it's probably under a little more stress than earlier this year, since
the school year just started here and students are hanging around the
campus and playing starcraft etc. :)

It would be nice if there were some way to prod the patchbot and make it
rerun tests even if the hashes of the patches were still the same.

-Keshav

----
Join us in #sagemath on irc.freenode.net !

[Simon King]

Hi Keshav,

On 2012-08-18, Keshav Kini <kesha...@gmail.com> wrote:
> It would be nice if there were some way to prod the patchbot and make it
> rerun tests even if the hashes of the patches were still the same.

Aren't the tests repeated if you set ?kick behind the address of the
ticket's patchbot page? Or would this use a cached result?

Anyway, if you have a patchbot page where it says "Retry False", append
?kick to the address and load it in you browser, it says "Retry True".

Cheers,
Simon

[Keshav Kini]

Simon King <simon...@uni-jena.de> writes:
> Aren't the tests repeated if you set ?kick behind the address of the
> ticket's patchbot page? Or would this use a cached result?
>
> Anyway, if you have a patchbot page where it says "Retry False", append
> ?kick to the address and load it in you browser, it says "Retry True".

I wasn't aware of that. Thanks for the tip!