Using a patch queue?

Using a patch queue?

[Daniel Jacobowitz]

Daniel Berlin offered in February to set up a patch queue.  It's some
custom software that he wrote for GCC, after two consecutive GCC Summits
in which people agreed that they wanted some automated way to keep track of
patches, but no one came up with anything that seemed usable.

Here's the GCC one:
  http://www.dberlin.org/patches/
  http://dberlin.org/patchdirections.html

I've never used it except to play with it, but a lot of GCC contributors do,
as you can see.  I think that's a pretty compelling point in its favor,
since they have a similar workflow to ours.

The patch tracker follows the list (via the web archives, I think) and
collects annotated messages.  You're under no obligation to annotate your
messages; anyone can manually add a URL to the patch tracker via the web
interface.  I believe the first review response removes the patch from the
queue; we might want to save :REVIEWMAIL: for final approval/rejection.
Or it might be useful enough just to track patches which have never
been looked at, which happens quite a lot.

I wouldn't mind having a better tool than my inbox to track down what needs
looking at; I don't have enough time to review everything that needs
reviewing as it is.  Anyone else have an opinion?

-- 
Daniel Jacobowitz
CodeSourcery

Re: Using a patch queue?

[Jim Blandy]

On 3/29/06, Daniel Jacobowitz <drow@false.org> wrote:
> I wouldn't mind having a better tool than my inbox to track down what needs
> looking at; I don't have enough time to review everything that needs
> reviewing as it is.  Anyone else have an opinion?

I'd been thinking of something along those lines for a while.  It
seems really easy to set up, and participation is voluntary, so we
should give it a try.

Re: Using a patch queue?

[Bob Rossi]

On Wed, Mar 29, 2006 at 07:14:59PM -0500, Daniel Jacobowitz wrote:
> Daniel Berlin offered in February to set up a patch queue.  It's some
> custom software that he wrote for GCC, after two consecutive GCC Summits
> in which people agreed that they wanted some automated way to keep track of
> patches, but no one came up with anything that seemed usable.
> 
> Here's the GCC one:
>   http://www.dberlin.org/patches/
>   http://dberlin.org/patchdirections.html
> 
> I've never used it except to play with it, but a lot of GCC contributors do,
> as you can see.  I think that's a pretty compelling point in its favor,
> since they have a similar workflow to ours.
> 
> The patch tracker follows the list (via the web archives, I think) and
> collects annotated messages.  You're under no obligation to annotate your
> messages; anyone can manually add a URL to the patch tracker via the web
> interface.  I believe the first review response removes the patch from the
> queue; we might want to save :REVIEWMAIL: for final approval/rejection.
> Or it might be useful enough just to track patches which have never
> been looked at, which happens quite a lot.
> 
> I wouldn't mind having a better tool than my inbox to track down what needs
> looking at; I don't have enough time to review everything that needs
> reviewing as it is.  Anyone else have an opinion?

For what it's worth, I like the idea a lot.

Bob Rossi

Re: Using a patch queue?

[Ramana Radhakrishnan]

I think its a good idea as it might enable more people to help track
patches down . Folks tend to try and look at patches but this should
help to not allow stuff to slip through the cracks. 

cheers
Ramana

On Wed, 2006-03-29 at 19:14 -0500, Daniel Jacobowitz wrote:
> Daniel Berlin offered in February to set up a patch queue.  It's some
> custom software that he wrote for GCC, after two consecutive GCC Summits
> in which people agreed that they wanted some automated way to keep track of
> patches, but no one came up with anything that seemed usable.
> 
> Here's the GCC one:
>   http://www.dberlin.org/patches/
>   http://dberlin.org/patchdirections.html
> 
> I've never used it except to play with it, but a lot of GCC contributors do,
> as you can see.  I think that's a pretty compelling point in its favor,
> since they have a similar workflow to ours.
> 
> The patch tracker follows the list (via the web archives, I think) and
> collects annotated messages.  You're under no obligation to annotate your
> messages; anyone can manually add a URL to the patch tracker via the web
> interface.  I believe the first review response removes the patch from the
> queue; we might want to save :REVIEWMAIL: for final approval/rejection.
> Or it might be useful enough just to track patches which have never
> been looked at, which happens quite a lot.
> 
> I wouldn't mind having a better tool than my inbox to track down what needs
> looking at; I don't have enough time to review everything that needs
> reviewing as it is.  Anyone else have an opinion?
> 

Patch tracker now available

[Daniel Jacobowitz]

On Wed, Mar 29, 2006 at 07:14:59PM -0500, Daniel Jacobowitz wrote:
> Daniel Berlin offered in February to set up a patch queue.  It's some
> custom software that he wrote for GCC, after two consecutive GCC Summits
> in which people agreed that they wanted some automated way to keep track of
> patches, but no one came up with anything that seemed usable.

Everyone who responded liked this idea, so I've set one up.  It was a
pile of work to get running, so I hope it's useful :-)

It's here:
  http://gdb.false.org/patches

It is certainly not mandatory for patch submitters or reviewers.  It
has both an email interface (it scans the web archives hourly looking
for commands) and a web interface; if I find it useful I can add
patches to it even when they weren't posted with appropriate tags. 
There's some instructions on how to use it by email up top of the list.

If you have any suggestions or problems, please let me know.

If it proves useful, we can make it more prominent, for instance by a
link from the GDB web site.  In the mean time, I encourage patch
submitters to use it.

There is an associated rant, which is that I'd really appreciate more
help in reviewing submitted patches.  I hope that adding tracking will
make it easier to see what needs review.

-- 
Daniel Jacobowitz
CodeSourcery

Re: Patch tracker now available

[PAUL GILLIAM]

On Fri, 2006-06-09 at 10:28 -0400, Daniel Jacobowitz wrote:
. . .
> There is an associated rant, which is that I'd really appreciate more
> help in reviewing submitted patches.  I hope that adding tracking will
> make it easier to see what needs review.

Technically, this is a "cry for help", not a "rant". :-)

If I had more general gdb expertise (and a little more slack in my
schedule) I would feel honor-bound to help.

-=# Paul #=-  

Patch tracker now available

[Nick Roberts]

 > There is an associated rant, which is that I'd really appreciate more
 > help in reviewing submitted patches.  I hope that adding tracking will
 > make it easier to see what needs review.

On that note, at some stage I would like to volunteer as MI maintainer.  This
is really just a statement of intent, as I would like to wait till after the
release of Emacs 22.  If my changes are well received then I think I would be
in a good position, as I would hopefully be getting feedback from the user
community.

To clarify, I would be prepared to review patches on the MI interface itself,
and do not pretend that I could currently cover the parts which delve deep into
the internals of GDB, as Andrew (or Elena) would possibly have done.

Oh yes, about your patch tracker.  I might have inadvertantly deleted a patch.
I didn't realise it would be so easy, maybe a confirmation would be a good
idea.  Perhaps I've disqualified myself already ...oops, sorry!

-- 
Nick                                           http://www.inet.net.nz/~nickrob

Re: Patch tracker now available

[Daniel Jacobowitz]

On Sat, Jun 10, 2006 at 12:32:51PM +1200, Nick Roberts wrote:
>  > There is an associated rant, which is that I'd really appreciate more
>  > help in reviewing submitted patches.  I hope that adding tracking will
>  > make it easier to see what needs review.
> 
> On that note, at some stage I would like to volunteer as MI maintainer.  This
> is really just a statement of intent, as I would like to wait till after the
> release of Emacs 22.  If my changes are well received then I think I would be
> in a good position, as I would hopefully be getting feedback from the user
> community.
> 
> To clarify, I would be prepared to review patches on the MI interface itself,
> and do not pretend that I could currently cover the parts which delve deep into
> the internals of GDB, as Andrew (or Elena) would possibly have done.

I think this would be a good thing.

> Oh yes, about your patch tracker.  I might have inadvertantly deleted a patch.
> I didn't realise it would be so easy, maybe a confirmation would be a good
> idea.  Perhaps I've disqualified myself already ...oops, sorry!

That's two people who've done that today!  I was talking with Dan
Berlin about this and the gcc version now sports an edit box to fix it
from the web interface.  At some point I'll get those changes from him
and update ours.  Meanwhile I poked the database to put them back.
Thanks for letting me know.

-- 
Daniel Jacobowitz
CodeSourcery

Re: Patch tracker now available

[Daniel Jacobowitz]

On Fri, Jun 09, 2006 at 09:09:14PM -0400, Daniel Jacobowitz wrote:
> That's two people who've done that today!  I was talking with Dan
> Berlin about this and the gcc version now sports an edit box to fix it
> from the web interface.  At some point I'll get those changes from him
> and update ours.  Meanwhile I poked the database to put them back.
> Thanks for letting me know.

It's a bit easier now.

By the way, I apologize to anyone who looked at it and said to
themselves "this could be useful, but I can't see what any of the
patches are about!".  I broke summary display.  Now I've unbroken it.

-- 
Daniel Jacobowitz
CodeSourcery