Re: [Monotone-devel] testresult certs, proposed default hook change

[Nathaniel Smith]

On Wed, Aug 17, 2005 at 02:17:35PM -0500, Timothy Brownawell wrote:
> Are there any objections to changing it to
> 
> function accept_testresult_change(old_results, new_results)
>    return true
> end
> 
> ? This way people are only affected by testresult certs they want to be
> affected by, rather than by all testresult certs. This would let people
> actually *use* them, rather than having to avoid them in order to not
> cause trouble.

Fine by me for now.  The better solution is to let people define
per-MT which tests they want to require no-regressions on... (maybe
just have a file in MT with one testkey per line, and have that hook
open and read it?)

-- Nathaniel

-- 
In mathematics, it's not enough to read the words
you have to hear the music