Outstanding, thank you very much Francisco! Changes merged.
On Thu, Apr 27, 2017 at 11:44 AM, Francisco Javier Ros Muñoz < fj...@doalitic.com> wrote: > Thanks Eric, the following PRs should close the issue: > > https://github.com/apache/libcloud/pull/1047 > https://github.com/apache/libcloud/pull/1048 > > Best, > -- > Francisco J. Ros > > > El 27 abr 2017, a las 19:03, Eric Johnson <erjoh...@apache.org> > escribió: > > > > Hi Francisco, > > > > I agree with you that #2 would be best. I would assume most client code > was > > written based off the docs, so we should make sure it's supported with > the > > code. Especially since it used to work that way. > > > > And thanks for the quick fix in #1046 too! > > > > Thank you, > > Eric > > > > On Thu, Apr 27, 2017 at 8:47 AM, Francisco Javier Ros Muñoz < > > fj...@doalitic.com> wrote: > > > >> Hi again, > >> > >> I’ve found another issue when trying to upgrade to libcloud-2.0.0. > >> > >> GCENodeDriver.create_volume() has an optional ‘location’ parameter that, > >> according to the docs, can be of type ‘str’, ‘GCEZone’ or > ‘NodeLocation’. > >> If something other than a str is passed as argument, the call fails in > >> libcloud-2.0.0 (worked fine in 1.3.0, the bug was introduced when > adding a > >> dict of cached volumes). > >> > >> Before working on a fix, I’d like to ask you first which option you > prefer: > >> 1) Leave the code as-is and fix the documentation; or > >> 2) Leave the documentation as-is and fix the code. > >> > >> I think (2) is better because it maintains api compatibility with > previous > >> libcloud versions. Please let me know your opinion. > >> > >> Btw, when looking into this I noticed that a non-existent method of the > >> same driver was being invoked: https://github.com/apache/ > >> libcloud/pull/1046 > >> > >> Best, > >> -- > >> Francisco J. Ros > >> > >> > >
