Send revision number in the reply to a successful merge

I've created a patch for this that should perform the task. I'm hoping I've approached it a better way this time, by using the VCSHandler class.

This is really important for Launchpad to be able to graph test suite runs versus revision numbers -- unless there's a workaround.

This won't help launchpad graph that because it only gives details on successful runs. What specifically do you need to achieve - what question are you trying to answer?

I'll look at the patch after europython.

On Wed, Jul 11, 2007 at 07:13:24AM -0000, Robert Collins wrote:
> This won't help launchpad graph that because it only gives details on
> successful runs.

I specifically only want to graph successful runs.

> What specifically do you need to achieve - what question are you
> trying to answer?

I want us to track which revisions landed impacted our test suite
positive or negatively.
--
Christian Robottom Reis | http://async.com.br/~kiko/ | [+55 16] 3376 0125

On Wed, 2007-07-11 at 12:28 +0000, Christian Reis wrote:
> On Wed, Jul 11, 2007 at 07:13:24AM -0000, Robert Collins wrote:
> > This won't help launchpad graph that because it only gives details on
> > successful runs.
>
> I specifically only want to graph successful runs.
>
> > What specifically do you need to achieve - what question are you
> > trying to answer?
>
> I want us to track which revisions landed impacted our test suite
> positive or negatively.

Ok, thats interesting.

In bzr we do that too, but we do it outside of the commits landing on
pqm, because to get useful stats you usually need to run several times
and chop off outliers. (e.g. system backups can make the tests slow
during pqm committing, even though its not actually meaningful).

All the same, we will get the revision number in there for you shortly.

-Rob
--
GPG key available at: <http://www.robertcollins.net/keys.txt>.

Yeah. Maybe the fluctation in our case is significant, but maybe it's
not -- so far it's hard to tell with the graph Tom's produced, but it
doesn't /seem/ to be, at least. But having the revision number there
allows us to speculate on that, since we can look at the commit and
sanely decide whether that should have impacted or not. And of course
you also check if the slow-down or speed-up is transient.

Has this been completed?

On October 26, 2007, Joey Stanford wrote:
> Has this been completed?

No

Fix sent to the Bazaar ML.