Historically the SE client areas (swing/awt/2d) have required 2 reviewers
for everything that's a significant change.
test, doc, trivial fixes can have one reviewer.
I can't say this has been 100% enforced, but we've tried.
This has been policy for a long time - well before openjdk.
Its not primarily to keep people informed, more that it
keeps reviewers honest :-), and you greatly increase
the chances of catching issues with 2 reviewers.
Also we've 'tagged' as reviewers people who actually know the code.
Someone jumping in on a list and saying 'looks good to me'
is maybe reassuring but I'd really like someone familiar
with the area and motivated rather than just anyone who
has the title 'jdk7u reviewer'.
> in that case, I'd suggest that we go for the stricter interpretation,
and require two jdk7u reviewers.
BTW this could mean that the original 8 fix which is back ported may
need additional review.
-phil.
On 6/21/2013 1:09 AM, Dalibor Topic wrote:
On 6/19/13 1:21 PM, Jim Laskey (Oracle) wrote:
Just want to pipe in on 'challenge'. The five person Nashorn team adopted a
two reviewer policy from the start. The feeling was that it kept team members
abreast of what was 'going on' as much as adding more critical review. It gets
tough at vacation time, but we've made it work (with few thin herd/one
liner/one reviewer occurrences.)
Thanks, Jim - in that case, I'd suggest that we go for the stricter
interpretation, and require two jdk7u reviewers.
cheers,
dalibor topic
Cheers,
-- Jim
On 2013-06-19, at 7:44 AM, Alan Bateman <[email protected]> wrote:
On 19/06/2013 10:52, Dalibor Topic wrote:
:
They should both have the Reviewer role.
Would changing "At least two reviewers are required for phase 2." in
http://openjdk.java.net/projects/jdk7u/phase2/phase2-approval-template.html
to "At least two reviewers, from the set of Reviewers listed on the JDK 7 Update
census page [0] are required for phase 2." make it clear? If so,
I'll formally propose it for discussion later today.
[0] http://openjdk.java.net/census#jdk7u
Thanks for getting the clarification on this moving.
The only thing I will say is that for some areas (niche areas in particular)
that having two reviewers with reviewer role on the jdk7u project might be
challenge. One thing to consider is whether it might make sense to relax it a
bit to allow one of the reviewers be someone with reviewer role on the jdk8
project.
-Alan.
