launchpad-dev team mailing list archive

[Gary Poster <gary.poster@xxxxxxxxxxxxx>]

ISTR that you left a player from the "which code base is it" game.  I think this constraint comes from httplib2.

http://code.google.com/p/httplib2/issues/detail?id=125
http://code.google.com/p/httplib2/issues/detail?id=66
https://bugs.launchpad.net/launchpadlib/+bug/459418

https://bugs.launchpad.net/launchpadlib/+bug/513116 seems to want to collect other problems, but doesn't have any so far.

That's all the help I know to give.  Maybe httplib2 can be subclassed to have better behavior, since the maintainer doesn't appear to actually have taken the available fixes, according to the comments in the issues linked above?

Gary


On Sep 26, 2011, at 2:04 PM, Jonathan Lange wrote:

> Hello,
> 
> I'm writing a program with launchpadlib that needs to run through a
> lot of data. As such, I'm parallelizing it. It only does read
> operations against Launchpad, but it turns out that even those read
> operations manage to corrupt the cached WADL.
> 
> Essentially, after running the importer for a little bit, every worker
> starts to raise these errors on login_anonymously:
> 
>    http://paste.ubuntu.com/697373/
> 
> I used strace to confirm that the command is indeed opening up an
> invalid XML file, and that file contains a large chunk of the
> Launchpad WADL.
> 
> My guess is that at some previous point, two workers are trying to
> login_anonymously together, and that somehow corrupts the copy on
> disk. Note that the same filename is used regardless of the name of
> the app.
> 
> Going forward, there seem to be a bunch of bugs and a couple of
> options as to how I proceed.
> 
> It seems to be a bug that a corrupt entry in the cache causes complete
> login failure. I would file it, but I'm a little tired of playing the
> "is it launchpadlib, wadllib, lazr.restful or lazr.restfulclient"
> game.
> 
> It's probably also a bug that launchpadlib allows this corruption in
> the first place.
> 
> As for me, I'm *probably* going to just restructure my program so I
> don't have the workers talking to Launchpad. If someone here can give
> me a clear path to fixing these bugs, and I can have a reasonable hope
> of a prompt release, then I might try fixing them in the underlying
> libraries.
> 
> cheers,
> jml
> 
> _______________________________________________
> Mailing list: https://launchpad.net/~launchpad-dev
> Post to     : launchpad-dev@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~launchpad-dev
> More help   : https://help.launchpad.net/ListHelp