Re: [PATCH OSSTEST v4] mg-all-branch-statuses: Show how up to date each branch is

[Ian Campbell <ian.campbell@citrix.com>]

On Fri, 2015-07-03 at 15:43 +0100, Ian Jackson wrote:
> Ian Campbell writes ("Re: [PATCH OSSTEST v4] mg-all-branch-statuses: Show how up to date each branch is"):
> > On Fri, 2015-07-03 at 14:40 +0100, Ian Jackson wrote:
> > > Do you want to provide a cron job to run this weekly, or something ?
> > 
> > Given the "NN days" I was thinking more like daily, otherwise "NN days"
> > is more like "+= up to 7".
> 
> Sure.
> 
> > Where would be a good place to run it from? Is $HOME/testing.git#master
> > always up to date after an osstest harness pass?
> 
> No, it isn't - it's only updated by the machinery in cr-for-branches.
> cr-for-branches enters $HOME/testing.git and updates it each time it's
> run, but...
> 
> > (I think we can tolerate the occasional update under foot).
> 
> ... that's not ideal.
> 
> Perhaps we should make a new $HOME/for-maintjobs.git which can run
> other kinds of not-particularly-long-running jobs.
> 
> You could use cr-for-branches from $HOME/testing.git to invoke it.
> cr-for-branches will do any necessary update and take the lock.
>   BRANCHES=maintjobs ./cr-for-branches . -w ./cr-all-branch-statuses ''
> which would run (in $HOME/./for-maintjobs.git):
>   ./cr-all-branch-statuses maintjobs

Should it not be:

BRANCHES=all-branch-statuses ./cr-for-branches . -w ./cr-maintjobs ''
which ends up calling
./cr-maintjobs all-branch-status
which in turn calls (from a "case $1") the new mg- from this patch?

Or maybe we rename it to mj- (for maintjob) and have cr-maintjobs call
mj-$1.

Would using ~osstest/daemons-testing.git for this be an intolerable
bodge?

Ian.