Format for JDK 6/7 changeset comments?
Jim Graham
Jim.A.Graham at Sun.COM
Fri Nov 9 21:01:41 UTC 2007
Why not both? Why does it have to be one *or* the other? I really
don't get that aspect of these discussions.
The somewhat vague argument that information in the changesets could
become stale is only mildly interesting. For one thing, that may happen
- what? - once a year or so? If it happens more than that, then we need
to examine our development documentation processes. Second, even if it
happens, exactly what problem might it cause that is so irreversible
that the project will be irreparably harmed? Finally, it should be
common knowledge that the information that exists there should not be
taken as the authoritative answer, just a convenient copy of the
authoritative answer for purposes of browsing.
The vague idea that such duplication of information *might* be wrong and
that this error could somehow cause serious repercussions has to be
weighed against the value of the convenience of not having to correlate
several repositories of information just to understand the changes that
are occurring. How exactly has this fear of potentially inaccurate
information taken over and pushed out all other considerations?
It's not like the information included in either location is somehow
legally binding - it's all there just to help us get work done...
...jim
Andreas Sterbenz wrote:
> Neal Gafter wrote:
>>
>> Anyway, the main issue is that there is a lot of value in having all
>> relevant information about a bug stored in ONE central place, which
>> is the
>> bug database. If you feel differently, then we'll have to agree to
>> disagree.
>> If the result is a situation that is convenient for you at the expense
>> of inconvenience for outside openjdk developers, agreeing to disagree
>> isn't a very palatable solution.
>
> The only thing we are talking about here is whether the changeset
> comments should include the names of the code reviewers. Are you saying
> it would be an inconvenience for non-Sun developers if they don't?
>
> It is obvious that the current situation with the bug database is
> unacceptable, but a solution is already planned
> (http://blogs.sun.com/mr/entry/under_construction). Having easy access
> to the names of the code reviewers seems to be the least problem to me.
>
> Andreas.
More information about the discuss
mailing list