Dear all,
<advertisement>
In case you are near Orsay next wednesday, you are welcome to join the
party: http://Nicolas.Thiery.name/habilitation/
</advertisement>
Back to sage development:
On Mon, Dec 1, 2008 at 8:50 PM, Vinzent Steinberg
> Sage has problems with reviewing patches:
> http://sagemath.blogspot.com/2008/11/sage-patch-review.html
Summary of the above: some patches on the Sage trac wait very long
before getting reviewed, which slows the development pace of Sage. So
the participants wonder whether there should be a patch review
manager, or maybe one per field of interest.
Somehow that's what sage-combinat is aiming at in the area of
algebraic combinatorics. Here is the typical life-cycle of one of our
patches:
- Creation one someone's machine
- Push to the sage-combinat server as soon as it is stable enough to
not break (at least not obviously) other functionalities. From this
moment on, it is applied (and maybe even used :-)) by everyone in
our community. Michael is further working on having the patches be
nightly tested. This makes for continuous integration test.
In case something goes wrong, it's easy to disable temporarily the
patch using guards.
- Cycle of further development, (potential) reviews and comments,
addendum by other developers, ...
- When the patch is ready: manual post to trac, and usual official
review and integration.
This last step is not working well right now: many patches just lie
there for too long before being posted to trac. In general, the
overhead in patch management is quite high for those patches which are
essentially trivial.
So, I am wondering whether it would be acceptable, for the trivial
patches and those which do not interact much with the rest of sage, to
have the review process occur solely within the sage-combinat
community. Namely, consider a patch written by nt:
this_trivial_change-nt.patch
- When nt deems it ready, (s)he renames it as:
this_trivial_change-final.patch
- When someone has reviewed it, (s)he renames it to:
this_trivial_change-reviewed.patch
- At this point, a script picks it up automatically and post it to
trac as [with review] using as comment the (short?) log of the
development of the patch + name of developers and reviewers.
- Usual integration
The more involved patches (like the one I am working on for
categories) would of course follow the usual path.
What do you think?
Cheers,
Nicolas
--
Nicolas M. Thiéry "Isil" <[EMAIL PROTECTED]>
http://Nicolas.Thiery.name/
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
