[Buildbot-devel] Waterfall Reload Issue in Firefox

[Jean-Paul Calderone]

On Wed, 28 Mar 2007 17:03:13 -0700, Oz <oz at santacruzgames.com> wrote:
>Charles Lepple wrote:
>> On 3/28/07, Oz <oz at santacruzgames.com> wrote:
>>> >> Here's a minor issue with the Waterfall that is quite annoying. If you
>>> >> fetch one project's waterfall (say for buildbot master instance A) and
>>> >> then fetch another project's waterfall (say for buildbot master
>>> instance
>>> >> B), then A's waterfall will continue to render (with B's URL in the
>>> >> address bar) until you reload B a second time.
>>
>> Wait... how do the URLs differ? Are they actually served off of the
>> same port? How are the buildbot masters related? Are they two separate
>> instances (as your example hints), or are they just two different
>> waterfalls pointing at the same buildbot master?
>>
>> (I haven't actually set up two buildbot masters on the same machine,
>> but it's something I'd like to do for a second project.)
>
>Per your last email, twistd (the Twisted daemon) 2.1.0
>
>The URLs are:
>  - http://myserver/build/projectA/
>  - http://myserver/build/projectB/
>
>They are mapped to unique ports via Apache's httpd.conf:
>
>ProxyPass /build/projectA/ http://myserver:7010/
>ProxyPassReverse /build/projectA/ http://myserver:7010/
>
>ProxyPass /build/projectB/ http://myserver:7011/
>ProxyPassReverse /build/projectB/ http://myserver:7011/
>
>And two instances of the buildbot daemon are running in the background
>as two independent masters. I do share a bunch of code between their
>config though via a "common" module.
>

This sounds like either an Apache bug (or misconfiguration, but it's
a pretty insane failure for a misconfiguration) or a Firefox bug (I'd
actually bet on this).  Aside from high energy cosmic rays entangling
bits of your system memory together, there's really no way the second
buildmaster could even have the data in order to serve up the waterfall
from the first master.

Much more likely is that the waterfall (or some other subtle aspect of
your system) is tickling a memory corruption bug in Firefox resulting in
it misrendering the second page.

I've never seen this particular failure, but I've definitely seen a lot
of cases where Firefox does something random and ridiculous and obviously
totally insanely wrong in a similar way.

Jean-Paul