Review process (was: Re: [PATCH] libcli/dns: Time out requests after a while)

Kai Blin kai at
Tue Oct 16 03:53:15 MDT 2012

On 2012-10-16 11:08, Jeremy Allison wrote:

> That's unfair. You posted the patch on the bug, got review and
> it got pushed as per bugzilla proceedure. I wouldn't call doing
> one ping on the list noting it'd already gone into 4.0.0rc having
> to "whine" to get it in. I was already tracking this patch (which
> is why it was still in my inbox) to ensure it got in.

Fair enough, I was overdramatizing a bit on the "whining" part. However,
I was a bit miffed that a patch that I'd have had in master since Sunday
was actually overtaken by the same patch to a bugfix release, which
you'd expect to be the slower process.

I only do my famous mind-reading trick on Thursdays, so I couldn't tell
somebody was already tracking the patch.

> Note you didn't ask anyone in particular to review, just posted
> to the list. If you had explicitly sent to someone requesting
> review then I know (at least for me) it would have gotten in
> faster.

Note that I did CC Matthieu on the first patch and Volker on the second.
I'm sending patches straight from my git tree using "git-format-patch |
git send-email", so the email text is actually the commit message.
Apparently that's not the correct way to do this.

So how does the protocol work? I'm the main person working on the
libcli/dns code. In fact, I'm the maintainer of that code. How do I
pick the reviewer?

Could someone walk me through the whole process here?

1. I have written a patch that I want to get into the tree.
2. I put the patch somewhere for a review. Where?
3. Someone reviews the patch. Who?
4. a) Reviewer pushes patch to autobuild
   b) Reviewer doesn't like the patch, fix patch and goto 1.
5, a) Patch is in master, life is good.
   b) autobuild rejects patch.
      i) autobuild test run was flakey, autobuild had a hangover or
         whatever, is that a goto 4a) or a goto 1.?
     ii) autobuild rejected the patch because it broke a test, fix and
         goto 1.

As you can see I'm a bit fuzzy on the details for steps 2, 3, and 5bi,
I'd be happy if someone could shed some light on this

> As soon as you raised this on the list (even without asking
> a perticular person) it got pushed. I don't think that's too
> bad IMHO.

Sure, that's what I was expecting. Otherwise I'd have just pushed it to
autobuild. I probably should have written up this email right away,
instead of the previous one. Sorry about that.

> Note I have an outstanding patch also waiting a push - but
> this is because I explicitly requested vl to look at it and
> he and I are still discussing this in private email.

Ah, I take that's the famous ninja review? ;)


Kai Blin
Worldforge developer
Wine developer
Samba team member

More information about the samba-technical mailing list