On Fri, May 3, 2013 at 5:27 AM, Jeroen Demeyer <jdeme...@cage.ugent.be> wrote:
> Dear Sage developers,
>
> I decided to gather some statistics on the reasons why a patch needs work,
> despite having positive review. In other words, on problems that I, as
> release manager, discover when merging a patch.
>
> These statistics are about 82 tickets which all has positive review and that
> I tried to merge between sage-5.9.rc0 and sage-5.10.beta1 inclusive.
>
> Of these 82 tickets, 25 had at least one problem (for tickets with multiple
> problems, I only count the most serious problem):
>  * 8 had problems with the patch itself, like a bad commit message or a
> malformed patch file.
>  * 1 patch needed to be rebased.
>  * 5 tickets had *obvious doctest failures which should have been easily
> detected by the author/reviewer of the patch*.
>  * 2 tickets had doctest failures which would be caught by running --long
> doctests in all of Sage (i.e. make ptestlong).
>  * 4 had problems building the documentation.
>  * 1 was a spkg upgrade causing a doctest failure.
>  * 1 patch had a merge conflict with a different positive_review patch.
>  * 3 had more complicated failures.
>
> Those 5 completely obvious doctest failures are quite frustrating to me,
> because they simply waste my time for something which takes at most 5
> minutes of time for an author/reviewer doing his due diligence.
>
> The 7 "make doc" and "make ptestlong" failures are also failures which could
> (should?) have been caught in the author/review process.
>
> I guess the good news is that 20 of the 25 failures (the first five items in
> the list above) could have been automatically detected by a better patchbot.

With the move to git, I'm envisioning the spkg upgrade and merge
conflict with a different positive_review patch could also be caught
by the patchbot (the latter at a later stage). I don't know what the
more complicated failures were, but perhaps they could have been
caught as well.

Thanks for the extra ping me to finally get to making the patchbot do
ptest long and check docbuilding :)

- Robert

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to