Re: Status of PTS SOAP APIs
On Thu, Nov 3, 2011 at 22:01, Stefano Zacchiroli <zack@debian.org> wrote:
> On Thu, Nov 03, 2011 at 09:39:06PM +0100, Sandro Tosi wrote:
>> >From a reportbug POV, they could solve a lot of the problems we have
>> in gathering data, and probably many other services would benefit from
>> having a single point where querying data.
>
> Good to know!
>
> Even though we should keep in mind the data available through PTS SOAP
> API is as good as PTS data, in particular it will be as *old* as PTS
> data, which is updated via cron a few times a day. Would that be good
> enough for reportbug purposes?
Yeah, I knew of this little delays that might be in data being
available on PTS, and that's fine for us (at least better than
checking dpkg status file or so).
>> Now, they have been in alpha stage for a long time, and that doesn't
>> seem have gained them any help. They're languish while they should be
>> used.
>
> True. The usage of it is close to non-existent as far as I know. The API
I looked at PTS log files, and there were a bunch of call in Oct and
Nov 2001 and then we go back to 2008/9 (I think there was some web
package using the PTS api).
> implementation has also been broken for a very long time, with very
> little people noticing. To be honest, I _think_ it is working now,
> probably after Giovanni fixed quite a few bits during DebConf11, but I'm
> not entirely sure. Have you made some test with the current
> soap-alpha.cgi yet?
yep, before sending the email I was playing with that and it worked quite fine.
>> What would be the downsides of just release them (so soap-alpha.cgi ->
>> soap.cgi) and handle bugs/features requests as the come thru? that's
>> the same effect that the alpha stage wanted to reach but without much
>> success.
>>
>> Let's make people use them.
>
> I'm all for it, but to be honest I'm not sure that the fact it is
> currently called "soap-alpha.cgi" instead of "soap.cgi" has much to do
> with the fact it's not being used.
Oh no, but that's for a selfish reason :) If I start using it in
reportbug and then that implementation ends up in stable and then we
change the cgi to soap.api, that would be boring to prepare a stable
update for it
> All that is needed is advertising a bit the existence of the API (blog
> posts and the like) and, more importantly, someone that keeps an eye on
> the code and react to bug reports. I've made clear that, at least as
> long as I'm also DPL, I won't be able to do that.
Mh, isn't all Debian workforce part of QA? <g> I can probably have an
interested look, but don't count me as the sole maintainer.
> Maybe a good start would be to you to start using it and report here
> your experiences (both in term of functionalities and in terms of how
> good the documentation is). If it's decently working, maybe you can blog
> about it and that might attract the interest of others in using it.
I'll start integrating it with reportbug, for the time being in a
separate branch, and see how it's behaving.
> If you think s/-alpha// could help, by all means go ahead (you or anyone
> else interested in trying that strategy).
as you can probably expect, I'm +1 on this: others opinion?
> I'm also open to publish the
> documentation in a more stable place than something with /~zack/ in the
> URL. ... just do that :)
Well, maybe on some part of qa.d.o - but who needs documentation?
Cheers,
--
Sandro Tosi (aka morph, morpheus, matrixhasu)
My website: http://matrixhasu.altervista.org/
Me at Debian: http://wiki.debian.org/SandroTosi
Reply to: