Patchwork 3 branch

Stephen Finucane stephen at
Fri Oct 18 17:08:36 AEDT 2019

On Fri, 2019-10-18 at 14:21 +1100, Daniel Axtens wrote:
> Hi all,
> We're getting pretty close to a 2.2 release, but there's a few things
> kicking around already that I want to consider for patchwork 3. I don't
> want to delay 2.2 forever and squeeze everything in, but I also don't
> want the patches to rot while we sort out the final 2.2 stuff.

I'm not entirely sure how much value there is in doing this on the main
repo rather than a personal clone (which is what I do), but it doesn't
affect me too much so ¯\_(ツ)_/¯?

> So I'm starting a topic branch for Patchwork 3: patchwork3 on the usual
> repository. I will consider patches labelled with pw3 or patchwork3 for
> the branch.
> Rather than trying to do One Big Merge, I'm probably going to keep
> rebasing it on master. Because that involves rewriting history, I'm
> going to be a bit more relaxed about taking patches during the review
> process and replacing them with subsequent spins if needed. (This is in
> contrast to the current approach where we take patches after review is
> complete and do not rewrite history.) This applies to patches for this
> branch only, and is not a general change in policy!

All good. Just to note though that I'd like to see most things that do
get pushed to this branch come up for review here before they're merged
back to master, even if the smaller ones are merged immediately after.

> I would encourage you to send patches for pw3 sooner rather than later.
> As I will be doing the rebasing, this is also in your interest,
> especially if you want database changes and don't enjoy rebasing
> migrations :)
> So far I have pushed my XML-RPC removal series, and I'm going to send
> and push some patches to drop python2 soonish. After that, the big
> database reshuffle!

Yay, database reshuffle! Looking forward to fixing that at last. As
with the XML-RPC API removal though, could we hold off on the python2
removal series? We can and should drop testing of Python 2/Django < 2.2
immediately, but I expect the full Python 2 removal series to be quite
an invasive one that will hamper clean backporting of inevitable
bugfixes from 'master' to 'stable/2.2'. It would make life
significantly easier for both of us if we were able to wait until
v2.2.0 had been around for at least a few weeks on instances like and before we did this, if
only to maximise the chances of bugs being spotted. Just my 2c though
and, unlike dropping XML-RPC without a pwclient rework, not something
I'm going to block on.


> Regards,
> Daniel
> _______________________________________________
> Patchwork mailing list
> Patchwork at

More information about the Patchwork mailing list