On Sep 19, 2017 01:22, "Greg Sheremeta" <gshereme(a)redhat.com> wrote:
On Mon, Sep 18, 2017 at 3:42 PM, Roy Golan <rgolan(a)redhat.com> wrote:
On Mon, 18 Sep 2017 at 22:30 Yaniv Kaul <ykaul(a)redhat.com> wrote:
> "Anyone can send a patch
That's no longer true. We have a whitelist. 
Small correction, anyone can send a patch, only people from the whitelist
can trigger CI jobs on it, for reasons discussed before, mostly security
Initially a patch should be sent as draft"
I think we should edit that to be more along the lines of "consider
initially posting as a draft" with guidelines to assist the decision.
A draft is hidden from the public, why is it better to send as such?
I've sent draft patches for 2 reasons.
1. I made progress on something and want to preserve it, but it's so WIP
that I wouldn't want anyone to see it. That might be because it could
confuse people, or it might be that the code is a prototype and/or so
terrible that I'd be embarrassed if anyone saw it :D Lately I'm more likely
to 'git format-patch | gdrive upload -' if it's something in this category.
2. I don't want to waste CI resources on something. Sometimes related to 1.
I see few advantages and they all drawn from the assumption the
patchset is always some sort of work in progress in really most of the
1. It doesn't invoke automation and waste resources. First the developer
should run it and be passed the checkstyle/pep/other errors locally.
2. Default reviewers feature hopefully will put the reviewers in place
automatically so it not hidden.
Hmm, I believe the "hopefully" doesn't work. A few weeks ago I got a
notification that I had a draft to look at [because I was a default
reviewer], but when I followed the link, I received a "not found" error.
3. After the patch is bit more mature it is worth publishing to get
reviews. Half baked or controversial patches may be costly to review. After
they are published the reviewer can expect higher quality and can estimate
better the effort in review
IMHO we don't use this practice enough.
>  https://www.ovirt.org/develop/dev-process/working-with-gerrit/
> Devel mailing list
Devel mailing list
Devel mailing list