On Sun, Oct 18, 2009 at 5:56 PM, Adam Kocoloski <[email protected]> wrote: > On Oct 18, 2009, at 11:48 AM, Benoit Chesneau wrote: > >> On Sun, Oct 18, 2009 at 4:55 PM, Adam Kocoloski <[email protected]> >> wrote: >>> >>> Caveat: I haven't built any "real" system myself that relies on OAuth, >>> I've >>> only helped out with the Canonical deployment. But as far as I know, >>> trunk >>> and the 0.10.x _branch_ fully support OAuth. The 0.10.0 _release_ has a >>> bug >>> that breaks pull replication from an OAuth-protected server, but >>> otherwise >>> it works. (More precisely, if you're running 0.10.0 and you replicate >>> from >>> an OAuth-protected server of any version, it will break. The bug affects >>> the replicator's HTTP client, not the server). >> >> mmm that could be annoying. Is this fixed in trunk ? > > Yes. I tried to say that above, but maybe I wasn't clear. Sorry about > that. > >>> OAuth requests with query-string parameters seem to be a tricky issue in >>> general. I believe a recent version of the Python OAuth library had a >>> bug >>> that would cause it to generate bad Authorization headers when >>> query-strings >>> params were used. I think the latest version of that library has fixed >>> that >>> bug. >>> >>> I'm not sure what you mean by "would it be possible to have devs in sync >>> in >>> a 0.10.1?" Coiuld you elaborate? Best, >>> >> >> Was just thinking that if there were fix on 0.10 in trunk, maybe they >> could be in 0.10.1 branch. depending if such version will happen or >> not of course :) >> >> >> thanks for the answers anyway, it helps >> >> - benoît > > Hi Benoit, I think we've got you covered. I'm assuming an 0.10.1 release, > if there is one, will be built from the 0.10.x branch, and so I've > backported the OAuth replication fix from trunk to 0.10.x. Cheers, > > Adam
Thanks ! - benoit
