> On Jun 15, 2017, at 11:53, Martin von Zweigbergk <martinv...@google.com> > wrote: > > As I said on IRC, I'd consider testing this in test-config.t and only > checking the output of "hg (show)config". Since it will be completely > transparent to the users (such as the status command), I think that's > probably enough. Or at least it probably means we can test less in the > non-config tests. It would also be a natural place for testing setting > of the config at different levels (global/user/repo), even if we don't > care about perfecting the behavior in that regard (which I agree about > leaving for later).
I could see adding that test - but I like having a case that actually demonstrates that this works "in the wild" too. Should I do a v2 with a test-config.t test added? _______________________________________________ Mercurial-devel mailing list Mercurial-devel@mercurial-scm.org https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel