Re: Generalised bisection

[Ealdwulf Wuffinga <ealdwulf@googlemail.com>]

On Tue, Mar 10, 2009 at 7:08 AM, Christian Couder
<chriscool@tuxfamily.org> wrote:

> I will try to have a look at the end of this week.
> But do you want it to be integrated with Git or do you want it to be an
> independant project that works with many different version control system?

Hmm. Whatever works, I guess. On the one hand the code does seem
naturally generic. On the other hand, it's good if users don't
have to separately obtain an extra package to use it. Supposing that
the algorithm proves useful, would the git project  be okay with an
extra dependency, or would you want to integrate it? Right now it's in
python, which I understand is an obstacle to integration.

In the short term, I assume the algorithm needs to prove its
usefulness before either being integrated or added as a dependency.
It seems the majority of potential users - developers of code of the
sort likely to have intermittent bugs, such as the kernel or xorg -
use git,
so I would like it to be as easy as possible for git users to try it
out. Maybe it could live in the contrib directory for a while?

Ealdwulf