Tiny improvement for Jira - Was: Re: github OAuth authentication need ?

[Olivier Berger]

Hi.

I think I now have pushed a cleaner version of the Jira remotes.

The WSDL proxy connection will only be attempted once there are actually
bugs in a particular remote's queue, and not at remote declaration time.

I've added a new 'ConnectionFailedExn' exception for connection issues
(like auth issues), which is trapped in order to avoid repeating polling
the same failing endpoint. It's used with failswith instead of die, or
some previous 'return None' instances.

The SOAP login for the WSDL proxy for puppetlabs is now in the new
secrets module : the config in production needs to be adjusted
accordingly.

I've fixed a typo along the way (Eclipse suggests some bugs sometimes,
nice ;-) ... OK, PyDev may not be the best Python dev environment, but
that's not too bad for my needs on bts-link, so far (debugger works).

I hope this is good, now.

Best regards,

Olivier Berger <olivier.berger at telecom-sudparis.eu> writes:

> Sandro Tosi <morph at debian.org> writes:
>
>> On Tue, Dec 8, 2015 at 10:09 AM, Olivier Berger
>> <olivier.berger at telecom-sudparis.eu> wrote:
>>> Hi.
>>>
>>> I've been hit by rate-limiting on puppetlabs' Jira, and made sure to
>>> reset credentials to the normal.
>>>
>>> In the meantime, I've added a tiny bit of exception handling to propery
>>> report the faulty SOAP endpoint.
>>>
>>> You'll find it in cc161f3262df50c0c0f705ac482a4e31975228f7 on the
>>> 'obergix-master' branch.
>>
>> it must not die() but failwith()
>>
>
> You're right. When the login failed, it crashed anyway, but we can do
> better, and just log an error, while the rest of the bugtrackers can be
> initialized and processed.
>
> I'll improve that.
>
> Thanks.
>
>>>
>>> An intereting improvement, though, would be to only initialize the WSDL
>>> proxies for SOAP endpoints only when bugs on such trackers are actually
>>> processed by btspull instead of at every run. It would avoid
>>> authenticating unnecessarily, thus avoiding burning tokens^Wkittens for
>>> nothing. I've put that in my TODO, but anyone faster would be much
>>> welcome ;)
>>
>> that's true for all the remotes, as it's done at initializatiokn time,
>> and they are all init'd
>>
>
> I've pushed an initial attempt at improving that, but I just noticed
> that the error handling wasn't good, so I'll try to integrate better
> exception management after this change too.
>
> Will let you know when looks better.
>
> Best regards,
> -- 
> Olivier BERGER 
> http://www-public.telecom-sudparis.eu/~berger_o/ - OpenPGP-Id: 2048R/5819D7E8
> Ingenieur Recherche - Dept INF
> Institut Mines-Telecom, Telecom SudParis, Evry (France)
>
> _______________________________________________
> Bts-link-devel mailing list
> Bts-link-devel at lists.alioth.debian.org
> http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/bts-link-devel
>

-- 
Olivier BERGER 
http://www-public.telecom-sudparis.eu/~berger_o/ - OpenPGP-Id: 2048R/5819D7E8
Ingenieur Recherche - Dept INF
Institut Mines-Telecom, Telecom SudParis, Evry (France)