On Wed, Aug 04, 2010 at 20:48:49 +0000, Petr Ročkai wrote: > Also, it would be nice, if someone could get us a regression test > for the said issue
[snip rough test recipe, for which thanks] Much more than "it would be nice", we should be more disciplined about testing. I propose we adopt a policy of not applying issue resolution patches until we have the corresponding regression test. In keeping with the informal use-your-common-sense meta-policy, this policy does not apply if the issue is not testable. That said, I'm not sure about patches that fix minor UI issues like patch307 or implement minor UI features like patch308; in those cases I'm guilty of being too excited with writing the code and not supplying a test. Should these fall under the policy? I suppose we could take a principled no-exceptions stance to avoid the danger of us weaseling out of tests? Thoughts? -- Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow> For a faster response, please try +44 (0)1273 64 2905.
signature.asc
Description: Digital signature
_______________________________________________ darcs-users mailing list darcs-users@darcs.net http://lists.osuosl.org/mailman/listinfo/darcs-users
