patchwork for netdev

patchwork for netdev

[David Miller]

During the kernel summit, Paul Mackerras gave an excellent presention
on the patchwork web based tool they are using for powerpc platform
development.

In a nutshell, it's a web tool that gets fed by a mailing list and
records patches.  Administrators for a "project" can change the
state of the patches, edit them, put them into bundles, etc.
And everyone can look at the state of the patch queue and also
look up older patches that were rejected, deferred, accepted, etc.

It looks like it can streamline our workflow as well as give people
a place to see what state their patch is in.  So I've asked the
ozlabs folks to create a "netdev" project.

See for yourself:

	http://patchwork.ozlabs.org/project/netdev/list/

Currently I am the admin for this project.

It might be nice if Jeff Garzik created an account for himself,
so I can get him marked as an admin as well, since driver patches
get funneled there too.

John Linville, on the other hand, may wish to ask the ozlabs folks
to add a linux-wireless project and admin that project.

John, I'll send you a contact under private cover that you can use
if interested.

Re: patchwork for netdev

[Randy.Dunlap]

On Thu, 18 Sep 2008, David Miller wrote:

> During the kernel summit, Paul Mackerras gave an excellent presention
> on the patchwork web based tool they are using for powerpc platform
> development.
> 
> In a nutshell, it's a web tool that gets fed by a mailing list and
> records patches.  Administrators for a "project" can change the
> state of the patches, edit them, put them into bundles, etc.
> And everyone can look at the state of the patch queue and also
> look up older patches that were rejected, deferred, accepted, etc.
> 
> It looks like it can streamline our workflow as well as give people
> a place to see what state their patch is in.  So I've asked the
> ozlabs folks to create a "netdev" project.
> 
> See for yourself:
> 
> 	http://patchwork.ozlabs.org/project/netdev/list/
> 
> Currently I am the admin for this project.
> 
> It might be nice if Jeff Garzik created an account for himself,
> so I can get him marked as an admin as well, since driver patches
> get funneled there too.
> 
> John Linville, on the other hand, may wish to ask the ozlabs folks
> to add a linux-wireless project and admin that project.
> 
> John, I'll send you a contact under private cover that you can use
> if interested.

Maybe patchwork for wireless would satisfy one of my concerns,
but it would be nice if John would ack patches like DaveM and Jeff do... :)

Thanks,
-- 
~Randy

Re: patchwork for netdev

[David Miller]

From: "Randy.Dunlap" <rdunlap@xenotime.net>
Date: Thu, 18 Sep 2008 17:01:52 -0700 (PDT)

> Maybe patchwork for wireless would satisfy one of my concerns,
> but it would be nice if John would ack patches like DaveM and Jeff do... :)

It's so much less work for me to click a bunch of buttons on a web
page and work in batch mode, than it is to give personalized
responses.  When you're applying 40, 50 patches in a day, every minute
in the workflow starts to count.

I was hoping those ACK emails would be one of the things I could stop
doing if patchwork is effective.  Not immediately, but eventually.

Re: patchwork for netdev

[Randy.Dunlap]

On Thu, 18 Sep 2008, David Miller wrote:

> From: "Randy.Dunlap" <rdunlap@xenotime.net>
> Date: Thu, 18 Sep 2008 17:01:52 -0700 (PDT)
> 
> > Maybe patchwork for wireless would satisfy one of my concerns,
> > but it would be nice if John would ack patches like DaveM and Jeff do... :)
> 
> It's so much less work for me to click a bunch of buttons on a web
> page and work in batch mode, than it is to give personalized
> responses.  When you're applying 40, 50 patches in a day, every minute
> in the workflow starts to count.

Yes, understood.

> I was hoping those ACK emails would be one of the things I could stop
> doing if patchwork is effective.  Not immediately, but eventually.

Well, that's a reasonable goal.  OTOH, if everyone wants to know
if their patch was applied, then having 50 people look at a web page
(or at a git tree) is more total overhead than one person generating
an email.  IMHO.  But maybe it depends on which side is being optimized.

Thanks,
-- 
~Randy

Re: patchwork for netdev

[David Miller]

From: "Randy.Dunlap" <rdunlap@xenotime.net>
Date: Thu, 18 Sep 2008 17:09:14 -0700 (PDT)

> On Thu, 18 Sep 2008, David Miller wrote:
> 
> > I was hoping those ACK emails would be one of the things I could stop
> > doing if patchwork is effective.  Not immediately, but eventually.
> 
> Well, that's a reasonable goal.  OTOH, if everyone wants to know
> if their patch was applied, then having 50 people look at a web page
> (or at a git tree) is more total overhead than one person generating
> an email.  IMHO.  But maybe it depends on which side is being optimized.

If the choke point is me, which it is, then...

Re: patchwork for netdev

[Randy.Dunlap]

On Thu, 18 Sep 2008, David Miller wrote:

> From: "Randy.Dunlap" <rdunlap@xenotime.net>
> Date: Thu, 18 Sep 2008 17:09:14 -0700 (PDT)
> 
> > On Thu, 18 Sep 2008, David Miller wrote:
> > 
> > > I was hoping those ACK emails would be one of the things I could stop
> > > doing if patchwork is effective.  Not immediately, but eventually.
> > 
> > Well, that's a reasonable goal.  OTOH, if everyone wants to know
> > if their patch was applied, then having 50 people look at a web page
> > (or at a git tree) is more total overhead than one person generating
> > an email.  IMHO.  But maybe it depends on which side is being optimized.
> 
> If the choke point is me, which it is, then...

then you get to choose, I suppose.  Meaning look at the patchwork
queue, which I did, and it looks good.

-- 
~Randy

Re: patchwork for netdev

[Rick Jones]

> I was hoping those ACK emails would be one of the things I could stop
> doing if patchwork is effective.  Not immediately, but eventually.

Can't patchwork do that as those clicks to change state take place?

rick jones

Re: patchwork for netdev

[David Miller]

From: Rick Jones <rick.jones2@hp.com>
Date: Thu, 18 Sep 2008 17:28:02 -0700

> > I was hoping those ACK emails would be one of the things I could stop
> > doing if patchwork is effective.  Not immediately, but eventually.
> 
> Can't patchwork do that as those clicks to change state take place?

I honestly don't know, but my guess would be know.

We could ask the patchwork author for this feature, as this
stuff is actively maintained.

Re: patchwork for netdev

[Christoph Hellwig]

On Thu, Sep 18, 2008 at 05:09:14PM -0700, Randy.Dunlap wrote:
> > I was hoping those ACK emails would be one of the things I could stop
> > doing if patchwork is effective.  Not immediately, but eventually.
> 
> Well, that's a reasonable goal.  OTOH, if everyone wants to know
> if their patch was applied, then having 50 people look at a web page
> (or at a git tree) is more total overhead than one person generating
> an email.  IMHO.  But maybe it depends on which side is being optimized.

Could patchworks please generate the ACK mail automatically? 

Re: patchwork for netdev

[Simon Horman]

On Thu, Sep 18, 2008 at 07:11:25PM -0700, David Miller wrote:
> From: Rick Jones <rick.jones2@hp.com>
> Date: Thu, 18 Sep 2008 17:28:02 -0700
> 
> > > I was hoping those ACK emails would be one of the things I could stop
> > > doing if patchwork is effective.  Not immediately, but eventually.
> > 
> > Can't patchwork do that as those clicks to change state take place?
> 
> I honestly don't know, but my guess would be know.
> 
> We could ask the patchwork author for this feature, as this
> stuff is actively maintained.

I think that the "thanks applied" feedback that occurs on netdev
is quite useful, as is the kind of automated feedback that
goes on from Andrew Morton's scripts in mm. To be honest I find
it quite frustrating when other maintainers in other arenas just
silently drop/ignore/apply-later (who knows which, or when?).

By all means I think that streamlining processes is good.
But if a change could be made to patchwork allow the "thanks applied"
feedback to continue, that would be doubly good.

-- 
Simon Horman
  VA Linux Systems Japan K.K., Sydney, Australia Satellite Office
  H: www.vergenet.net/~horms/             W: www.valinux.co.jp/en

Re: patchwork for netdev

[David Miller]

From: Simon Horman <horms@verge.net.au>
Date: Tue, 23 Sep 2008 11:08:16 +1000

> But if a change could be made to patchwork allow the "thanks applied"
> feedback to continue, that would be doubly good.

Sure and we can make a request to Jeremy for the feature.

Re: patchwork for netdev

[Benjamin Herrenschmidt]

On Mon, 2008-09-22 at 19:44 -0700, David Miller wrote:
> From: Simon Horman <horms@verge.net.au>
> Date: Tue, 23 Sep 2008 11:08:16 +1000
> 
> > But if a change could be made to patchwork allow the "thanks applied"
> > feedback to continue, that would be doubly good.
> 
> Sure and we can make a request to Jeremy for the feature.

One thing we talked about was to send a submitter an email on
a state change of his patch... that may do the trick.

Jeremy, I CC'ed you but you may want to catch up with the whole thread
first.

Cheers,
Ben.

Re: patchwork for netdev

[Jeremy Kerr]

Hi folks,

> Jeremy, I CC'ed you but you may want to catch up with the whole
> thread first.

Yep, I'd like to get patchwork extended to allow 'feedback' posts to the 
lists. This is what I'm thinking:

 * Add an 'Ack' (and possibly 'Nack' and 'Comment') button for all
   logged-in users, which will send an appropriate message to the list.

 * When a maintainer uses this functionality, patchwork can update the
   state accordingly.

However, I still need to work out the exact implementation - Ideally, 
the message could be constructed with the user's own mail client (and 
saved in the standard 'sent' folder). This requires a little 
integration hackery though.

But yeah - I'm workin' on it! :)

Cheers,


Jeremy