Hi Clerezza team,

I haven't had time to look at all the details, busy organizing an
internal conference this week...but it looks like there were -1 vetoes
on commits recently, so let me try to explain a bit better how that
works.

If someone vetoes a commit with a valid technical justification, and
espcially if another committer supports the veto, the commit MUST be
reverted.

Moving to a branch is a good idea in this case, so that people can see
the code in question and evaluate it before eventually moving back to
trunk.

This is clearly explained at
http://www.apache.org/foundation/voting.html : "for code-modification
votes, +1 votes are in favour of the proposal, but -1 votes are vetos
and kill the proposal dead until all vetoers withdraw their -1 votes."

In this case, IIUC someone went ahead on some commits without asking
for a vote, and that's probably where the misunderstanding is: any
"important" change should be discussed on the dev list (not just in a
JIRA issue) and applied only if people approve.

The problem is in defining "important", but apparently those changes
have an "important" effect so it's much better to have them in a
branch, or as patches, before applying them to trunk.

If it's hard to agree on what is "important" it might be useful to
change to RTC mode (Review-Then-Commit) for critical parts of the
codebase instead of CTR - maybe just temporarily until the release.
See http://www.apache.org/foundation/glossary.html for RTC and CTR.

thanks for listening,
-Bertrand (with mentor hat on ;-)

Reply via email to