[Ecls-list] [cl-test-grid] cl-test-grid incremental reports

Juan Jose Garcia-Ripoll juanjose.garciaripoll at gmail.com
Thu Jan 17 08:49:56 UTC 2013


On Mon, Jan 14, 2013 at 9:28 PM, Anton Vodonosov <avodonosov at yandex.ru>wrote:

>
> Sorry it took long to reply.
>
> I want to say that the approach is not so manual.
>
> I run few commands (git pull, ecl-rebuild.sh, run-agent.sh)
> and computer tests 800+ libraries on the new ECL.
>
> Then I run another command from slime and have a
> comparison of these results with results
> of the old version.


Thanks for the long answer, Anton. I suspect this is the only reasonable
way to go. Right now I have an automated script that runs the tests every
night there are new commits, comparing with the last version. This leads to
strange results, as sometimes the script may decide to run the tests in
between two commits, not at a point where ECL is stable enough.

I am pondering to change the version number system, changing the commit id
with a GMT date of the commit. The only reason why I am unsure about this
is because unlike commit ids, I cannot see anywhere something that warrants
that the commit date is stable across git implementations and when
pulling/pushing patches between computers.

It is for this reason that right now, to sort commits, I am using a stupid
hack that checks out a file and uses its date to sort the commit IIRC.
Unfortunately the machine that does this died during Xmas, and I am in the
process of recovering the data, together with all of ECL's homepage, test
reports, etc.

Juanjo

-- 
Instituto de Física Fundamental, CSIC
c/ Serrano, 113b, Madrid 28006 (Spain)
http://juanjose.garciaripoll.googlepages.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/ecl-devel/attachments/20130117/1ad8d651/attachment.html>


More information about the ecl-devel mailing list