=== The Problem with Patch Review ===

Mandatory patch review was introduced at SD5 and it should be
considered a success since it certainly seems to have helped to keep
the code quality up and prevent many potential bugs from slipping into
Sage. Unfortunately we did not have any formal mechanism in place and
there has been an ever growing number of patches that sit unreviewed
in trac and bit rot. This is unacceptable and at Dev1 in Seattle we
had an extended discussion between all present developers at the daily
status meeting and below is my recollection of what we intended to do.
Please take a look a correct errors, clear up inconsistencies and make
additional suggestions. Eventually we should stick the result of the
discussion in the wiki and also add it to the documentation.

Cheers,

Michael

=== A Potential Solution - Draft ===

 0) Credit reviewer added to release notes

 1) Weekly meeting of patch editorial board. Responsibility:

 -  People [who exactly - i.e. volunteers for now] get together in IRC
and look at the newly opened tickets in trac with a patch and without
a reviewer. Volunteer review as is will go on, but the purpose of the
editor's meeting is to establish a "no patch left behind" policy.
  - deal with negative reviews (invalidate, bring up the issue on
[sage-devel])
  - keep track of deadlines for reviews and reassign as needed

 2) New status titles:

  "[with patch, under review by $FOO before $DATE]"

 3) Let trac users add themselves to categories like "number theory"
so that they are aware of all tickets in that category. Do we need to
split up categories? Martin Albrecht mentioned that "commutative
algebra" for example is too broad. This requires fixes to trac or the
setup of a large number of Google groups specific to each category
that interested parties can subscribe to.

 4) Editor vs. Reviewer:

 - An editor is the person who is in charge of finding a reviewer,
asking the person if they are willing to review, setting a deadline
and enforcing it.
 - A reviewer is the person doing the review as usual. Upon the
request of the editor the reviewer can accept or reject the review
request. If the person agrees a deadline is set, which is usually the
next editors meeting [see below].
 - Editor and reviewer can be the same person, but the reviewer should
not be the patch author.
--~--~---------~--~----~------------~-------~--~----~
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
-~----------~----~----~----~------~----~------~--~---

Reply via email to