I think David updated his (and medic's) version of npm and that fixed the problem. I think the version he had was not installing devDependencies automatically.
On Fri, Mar 28, 2014 at 11:14 AM, Steven Gill <stevengil...@gmail.com>wrote: > Medic is now showing plugman tests passing. No idea what is going on. They > still pass fine for me. > > On Mar 28, 2014 6:07 AM, "David Kemp" <drk...@chromium.org> wrote: > > > I have just verified the plugman state and on my machine I get the same 7 > > errors that Medic gets. > > All of them are related to version comparisions. > > > > > > > > > > On Thu, Mar 27, 2014 at 6:17 PM, Steven Gill <stevengil...@gmail.com> > > wrote: > > > > > All tests seem to be passing now. > > > > > > Medic is showing plugman tests failing after the tizen commits, but > they > > > are passing fine for me. > > > > > > > > > On Wed, Mar 26, 2014 at 10:05 AM, Andrew Grieve <agri...@chromium.org > > > >wrote: > > > > > > > On Wed, Mar 26, 2014 at 9:05 AM, Josh Soref <jso...@blackberry.com> > > > wrote: > > > > > > > > > https://github.com/blackberry/cordova-cli/tree/cb_6337 > > > > > > > > > > Should cover both of my breaks. > > > > > > > > > > I¹m waiting for our CI tooling to confirm it¹s happy before I > submit > > a > > > > > pull request for it. > > > > > > > > > > Maybe sometime in the near future I¹ll try to rework the tests to > use > > > an > > > > > in-memory-psuedo-filesystem so that we can stop having to add all > > these > > > > > stupid lies. > > > > > > > > > > Tests shouldn¹t be so brittle that changing code which has no > > relation > > > to > > > > > a test results in a broken test. If we¹re lying and pretending that > > > > > there¹s e.g. a "lib/dir", then that directory needs to sufficiently > > > exist > > > > > ‹ adding code which makes a perfectly reasonable assumption ³this > > > > > directory which someone claimed exists exists² shouldn¹t break a > test > > > > > which pretended the directory existed. > > > > > > > > > > > > > Agree 100% with your sentiments here, and also agree that an > in-memory > > > > filesystem is the best way to make our tests decent. > > > > > > > > > > > > > > The CordovaError bit is different (and entirely my fault), I think > I > > > must > > > > > have failed to commit the requires update, because I remember doing > > the > > > > > greps for it... > > > > > > > > > > > > > > > > > > > >
