On 04/17/2010 10:56 PM, Tres Seaver wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Lennart Regebro wrote:
On Sat, Apr 17, 2010 at 19:17, Tres Seaver<tseaver@palladion.com> wrote:
I'm ambivalent about testing the Sphinx code snippets on each test run. I want those snippets to be *much* less comprehensive than they are currently, and am pretty sure that drift in the non-executable bits is at least as important a problem as drift in the snippets.
Well, unless they take of time, but they shouldn't, really. I think it's positive to get a quick feedback if you are breaking the documentation.
Again, I don't mind that part, but I want to break the cycle of jamming crap (for documentation purposes) into the docs for purposes of getting test coverage.
I think Lennart was arguing for a safety belt: if we're not supposed to break those tests *at all* within bug fixes then I want to know right away if I did so accidentally. I guess your coverage argument was that if the unit tests do the coverage anyway then we won't miss anything, but we won't be 100% anyway and as the tests in the docs do exist and should run quickly, the safety belt would be cheap. I'm +0.5 after some pondering. -- Christian Theune · ct@gocept.com gocept gmbh & co. kg · forsterstraße 29 · 06112 halle (saale) · germany http://gocept.com · tel +49 345 1229889 0 · fax +49 345 1229889 1 Zope and Plone consulting and development