> We should still test that we get the warning if the target doesn't have
> the attribute.  I'll fix the test, thanks.

You probably need to do the same on the 10 branch since it fails there too.

-- 
Eric Botcazou


Reply via email to