Re: Re: [RFC] Facility for QA reports
| From: | Tobias Schenck | Date: | Mon, 10 Dec 2001 13:46:50 +0000 |
| Subject: | Re: Re: [RFC] Facility for QA reports | ||
| References: | 1 2 3 | Groups: | php.qa |
| Request: | Send a blank email to php-qa+get-4200@lists.php.net to get a copy of this message | ||
> On Wed, Dec 05, 2001 at 06:56:46PM +0100, Tobias Schenck wrote :
> > I think of simply quering SELECT snapshotversion FROM testresults WHERE
> > testcase_id = 12345 AND result = "failed" and similar statements. This
might
> > make it easier to track down the commit that fixed or caused the bug.
"Markus Fischer" <mfischer@guru.josefine.at> schrieb im Newsbeitrag
news:20011210125851.C28030@guru.josefine.at...
> I'm iffy on this. Someone who thinks he knows it better may
> speak up.
Every right to be iffy about such a lame SQL-statement. The only sense was
trying to point out that one could give it a shot. You can only analyze the
data that has been fed in.
Richard Lynch proposed the following:
>Automate "make test" so that:
>[..]
>Maybe I'm being naive, or idealistic, though...
I can definitely buy into that! My naive and idealistic approach would be to
include a date/timestamp-field php_snap_version in the testrun-table and do
automated make tests on every snap. After every test-build there yould be an
automated report to QA about the failed tests which passed in the snap
before.
I admit that this is simply the view of somebody less concerened with actual
work but with analyzing databases, but on the other hand I'm quite sure we
find a gig of RAM and a cute hard disk for the QA-server.
> > > CREATE TABLE testrun (
> > > id smallint(5) unsigned default NULL auto_increment,
> > > user_id smallint(5) unsigned NOT NULL,
> > > os_id smallint(5) unsigned NOT NULL,
> > > os_version_id smallint(5) unsigned NOT NULL,
> > > distribution_id smallint(5) unsigned NOT NULL,
> > > configure_id smallint(5) unsigned NOT NULL,
>
> php_version char(80) not null,
>
> With the PHP and Zend version nicely formatted so we can
> run a DISTINCT select on thiss table and so the user can
> choose the filter on the query page.
>
> Other ideas?
I believe that "php_version char(80) not null" is fine for filtering. An
additional numerical field might have some advantages for future analysis of
test results.
kind regards,
Tobias Schenck