On Mon, Sep 01, 2008 at 09:22:40AM +0200, Lucas Nussbaum wrote:
> On 30/08/08 at 17:43 +0200, Gonéri Le Bouder wrote:
> > On Sat, Aug 30, 2008 at 02:00:06PM +0200, gregor herrmann wrote:
> > > On Sat, 30 Aug 2008 07:28:45 -0300, Martín Ferrari wrote:
(...)
> I think that you are both missing the point.
>
> PET and buildstat are competitors, because they both try to:
> - run tests
> - display information in an useful way
>
> As a result, it's likely that none of them will succeed, and that we
> will have to live with several tools doing basically the same thing for
> years, like lintian and linda.
We are both willing to merge our tools. buildstat works today and is
already able to accept QA resultat from a standardized format. I think
a meeting is absolutly needed here with the three parties.
IMO Buildstat is a DB + an interface to collect QA check results from 3th
parties tools.
Creating such schema wasn't that easy, I had to recreate it from scratch
two times yet.
(...)
> UDD isn't the ultimate solution, but it at least forces a clean
> workflow:
For me, the best solution to get a working solution quickly is to act
like that. It's not what most of you want. I guess you wish to see
buildstat out of the scheam, but it works fine.
> Test1 Test2 Test3 Test4
> \ \ / /
> `---------\ \ / /-----'
buildstat (input interface)
buildstat(DB) <------> U D D
> | | |
> | DDPO-like web interface |
> | ` DDPO-by-mail |
| PET interface |
| buildstat interface |
> | ` other scripts |
Of course, I share the final goal of having a Ultimate Database but
stopping what works fine today is not a good option for me.
If we do a meeting we Buildstat, PET and UDD teams, we will able to
quickly prepare a clean roadmap and, more important, avoid futur wast of
time.
Cheers,
Gonéri Le Bouder
Attachment:
signature.asc
Description: Digital signature