Robert Bradshaw <rober...@math.washington.edu> writes:
> On Tue, Feb 28, 2012 at 5:48 PM, Michael Orlitzky <mich...@orlitzky.com> 
> wrote:
>> On 02/28/2012 07:33 PM, Keshav Kini wrote:
>>> It's not much consolation right now, but this will become easier once we
>>> switch to git, as branches can be automatically checked for whether they
>>> are still mergeable into trunk or not. Well, the same could be done with
>>> patches I guess but it would require a lot more coding on someone's
>>> part.
>>
>> The patchbot could theoretically do it, but it would take a very small
>> number of false positives to become annoying.
>
> It does do it, and runs the tests too. There is the problem of
> defining "trunk" with our current workflow, but anyone can run a
> patchbot against the latest release, or any alpha.

There's another more fundamental problem as well - deciding which commit
to apply the patch to in the first place. As far as I know, the patchbot
does not look at the Parent ID headers of Mercurial-produced patch
files, and even if it did, those often do not even correspond to a node
ID in the version of Sage the patchbot has, since 1) node IDs in
development releases are recomputed each time Jeroen creates a new dev
release because he starts from scratch each time (yet another problem
with this method); 2) people may be basing patches upon other patches,
and the node ID of a patch may differ depending on when, where, and how
you qpush it, qrefresh it, or otherwise fiddle with it; 3) we have no
control over how developers generate their patches anyway, and so they
could contain garbage data in the headers for all we know, or even no
headers (as is the case if the patch is uploaded directly from the mq
folder rather than by using `hg export`).

-Keshav

----
Join us in #sagemath on irc.freenode.net !

-- 
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to 
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Reply via email to