[hibernate-dev] Retrospective on "Pull Requests": a waste of time?
Hardy Ferentschik
hardy at hibernate.org
Tue Jul 9 15:39:12 EDT 2013
I basically like what I hear. Some wise words :-)
On 9 Jan 2013, at 9:00 PM, Emmanuel Bernard <emmanuel at hibernate.org> wrote:
> There has been a tendency to let PR sit a bit longer than it should as
> we all try to get our stuff done before diving into other's PRs.
> I have been particularly guilty and Hibernate OGM is a particularly bad
> example. I did not see too much lagging PRs on other projects
Right. I feel Search for example is working quite well.
> Should we have a team member on watch whose priority for a week is
> addressing pull requests?
Not sure. The idea has some merits, but I am not sure that it is necessary.
> I also think opening issues to things that are not fully on the topic is
> a good strategy to keep the cycle on a given PR short.
What is the definition of not fully on topic. I would not suggest a change in
class X for a pull request where only class Y and Z are affected. However,
if class X is touched and I see a potential improvement I think it can be considered
being part of the topic. Boy Scout rule number one:" Always leave the campground
cleaner than you found it." I truly believe in this one, but of course sometimes a
potential improvement would have too big of a ripple effect to be pursued.
> That does not
> mean one won't work directly on these after the PR is done
Here we have to disagree. Unless you do it here and now the chances are slim
you are following up.
> About the preview, I have to disagree with Sanne and Hardy, I do like
> them and find them to be the least worse tool to show a preview and get
> feedback. I'm sorry but I have done it many times on JIRA and via emails
> and the feedback is not the same by far.
My experience with asking for feedback on my feature branches is actually
quite good. Kudos to Gunnar and Sanne on their helps. I guess if I felt I would be
left hanging on this type of feedback, I would create these "preview" pull requests
as well.
> Like many rules, they are meant to be broken and good judgement is highly
> preferable
That's the part I like best.
+100
--Hardy
More information about the hibernate-dev
mailing list