On May 1, 3:57 pm, Bill Hart <[EMAIL PROTECTED]> wrote:
> Hi Tim,
<SNIP>
> The main focus seems to be on fixing bugs related to failing doctests.
I forgot to make a very important point here: Failing doctests clearly
show that regression testing works. We now have a little under 55,000
or so input doctests that take about 90 minutes of CPU time on
sage.math to run. And the goal is to get coverage up to 100% which
would translate to over 100,000 inputs. That is certainly not a
panacea to the corner case problem for example unless we pay special
attention to it, but that is a hell of a lot more testing than many
other Open Source math packages out there.
For a while now I have not merged a patch if it causes doctest
failures. If you check trac you will see that many times a patch gets
postponed with a comment to fix doctests. Due to parallel doctesting
written by Gary I now run tests after every intrusive or maybe two or
three small, unrelated patches. We used to just merge and then fifteen
or twentyfive patches larger realize that something went wrong and
then start guessing what went wrong. Obviously that is on sage.math
only since that is my primary development box, so once we have people
test on random configs and William on 18+ test boxen things turn up
which then end up in trac and are usually discussed on sage-devel, so
these get a lot more visibility while discussion of other bugs nearly
happens exclusively in IRC where mostly dev people hang out.
3.0 has been much more stable and bug free than I anticipated after
merging 270 tickets. So I really think that we are moving in the right
direction and we will only get better at this and refine our approach
to release management.
<SNIP>
Cheers,
Michael
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
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
-~----------~----~----~----~------~----~------~--~---
