On Thu, Oct 13, 2011 at 2:20 PM, Dave Martin <dave.martin@linaro.org> wrote:
On Thu, Oct 13, 2011 at 01:16:30PM +1300, Michael Hudson-Doyle wrote:What's there looks OK; I have a couple of suggestions, though:
> Hi all,
>
> I've done a stealth deploy (stealth as in "no links point to this page")
> of the first cut of the kernel ci view I've been working on:
>
> http://validation.linaro.org/lava-server/kernel-ci-views/index
>
> It's very much an alpha-style view at the moment -- in particular the
> computation of the new passes and failures is a bit wonky -- but I've
> been working away for a while on this and it seemed like time to
> actually get a live page out to the public and stop emailing screenshots
> to slightly randomly chosen people.
>
> If you see problems, you can file bugs at:
>
> https://bugs.launchpad.net/lava-kernel-ci-views
>
> or make more general comments here.
* Display the total number of tests, passes and failures as well as
the delta. The delta is certainly useful and we should have it,
but without the absolute numbers, misleading conclusions may be
drawn about the health of a branch.
* I think we should display the branch that was tested as well as the
commit. This is valuable context, and there's no easy way to
reconstruct it just from the commit.
* Link the branch and commit directly to gitweb (as discussed the the
mail I sent a short time ago) -- this should be pretty easy; we
just need a rule for manufacturing the appropriate web link per
git server.