On Sun, 2010-02-28 at 20:33 +0000, Jelmer Vernooij wrote:
> > This is naughty: it should do one of:
> > - fail to import (raise ImportError). This will be logged but the test
> > run will continue
> It does this but raises DependencyNotPresent rather than ImportError,
> which is part of bzrlib.errors. Could bzr perhaps catch this error as
> well?
I don't think the test infrastructure should do that: it won't be known
by other runners at all.
On Sun, 2010-02-28 at 20:33 +0000, Jelmer Vernooij wrote:
> > This is naughty: it should do one of: esent rather than ImportError,
> > - fail to import (raise ImportError). This will be logged but the test
> > run will continue
> It does this but raises DependencyNotPr
> which is part of bzrlib.errors. Could bzr perhaps catch this error as
> well?
I don't think the test infrastructure should do that: it won't be known
by other runners at all.
-Rob