[U-Boot] Patchwork question

[U-Boot] Patchwork question

[Anton Staaf]

I'm new to patchwork, and am wondering.

Who should change the state of patches there? Should I mark patches
that have been applied as Accepted?  Or should the tree maintainer do
that?

Also, should I delegate patches to custodians that I believe should be
looking at them?  Or should custodians delegate patches to themselves?

Thanks,
    Anton

[U-Boot] Patchwork question

[Albert ARIBAUD]

Hi Anton,

Le 28/10/2011 18:52, Anton Staaf a ?crit :
> I'm new to patchwork, and am wondering.
>
> Who should change the state of patches there? Should I mark patches
> that have been applied as Accepted?  Or should the tree maintainer do
> that?

State changing should be done by the person who causes the change -- 
change requester or custodian -- as this is the only person that really 
knows for sure what the state change should be.

Wolfgang uses tools to automate transitions to "Applied" when he applies 
a patch or pull request -- I am looking into using these too.

> Also, should I delegate patches to custodians that I believe should be
> looking at them?  Or should custodians delegate patches to themselves?

Le the custodians delegate as they see fit.

> Thanks,
>      Anton

Amicalement,
-- 
Albert.

[U-Boot] Patchwork question

[Anton Staaf]

On Fri, Oct 28, 2011 at 10:16 AM, Albert ARIBAUD
<albert.u.boot@aribaud.net> wrote:
> Hi Anton,
>
> Le 28/10/2011 18:52, Anton Staaf a ?crit :
>>
>> I'm new to patchwork, and am wondering.
>>
>> Who should change the state of patches there? Should I mark patches
>> that have been applied as Accepted? ?Or should the tree maintainer do
>> that?
>
> State changing should be done by the person who causes the change -- change
> requester or custodian -- as this is the only person that really knows for
> sure what the state change should be.
>
> Wolfgang uses tools to automate transitions to "Applied" when he applies a
> patch or pull request -- I am looking into using these too.

That makes sense to me, the reason I asked is that all of my patches are still
sitting in the new state, even though most of them have been applied already
(or rejected) for a while...

>> Also, should I delegate patches to custodians that I believe should be
>> looking at them? ?Or should custodians delegate patches to themselves?
>
> Le the custodians delegate as they see fit.

Will do.

Thanks,
    Anton

>> Thanks,
>> ? ? Anton
>
> Amicalement,
> --
> Albert.
>

[U-Boot] Patchwork question

[Wolfgang Denk]

Dear Anton Staaf,

In message <CAF6FioWUw7=LC08CWQxrwQMj=4H0ES9O1AsdubnE8POBZpieeA@mail.gmail.com> you wrote:
>
> That makes sense to me, the reason I asked is that all of my patches are still
> sitting in the new state, even though most of them have been applied already
> (or rejected) for a while...

Patchwork bugs...

If you see such cases, please correct the state.  It really helps me.

Thanks!

Best regards,

Wolfgang Denk

-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
"...all the  good  computer  designs  are  bootlegged;  the  formally
planned  products,  if  they  are built at all, are dogs!" - David E.
Lundstrom, "A Few Good Men From Univac", MIT Press, 1987

[U-Boot] Patchwork question

[Wolfgang Denk]

Dear Albert ARIBAUD,

In message <4EAAE366.7050204@aribaud.net> you wrote:
> 
> State changing should be done by the person who causes the change -- >
> change requester or custodian -- as this is the only person that really >
> knows for sure what the state change should be.

Well, Patchwork often guesses wrong, and manual cleanup is needed.  If
a patch submitter sees his patch has been committed, and it is still
marked as "new" in PW then such cleanup would help us maintainers.

> Wolfgang uses tools to automate transitions to "Applied" when he applies
> a patch or pull request -- I am looking into using these too.

...but this suferes from the known PW issues.

> > Also, should I delegate patches to custodians that I believe should be
> > looking at them?  Or should custodians delegate patches to themselves?
>
> Le the custodians delegate as they see fit.

Fact is that many don;t do this, or not regularly at least.  I'd
rather see all patches assigned to someone, even if incorrectly.  The
custodian will recognize soon enough that it's not for him, and
reassign. 

If a patch is assigned, it's on somebody's todo list, asking for
action.  If it's left unassigned, it's in nowhere land where nobody
feels responsible.

Best regards,

Wolfgang Denk

-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
I don't care if you *ARE* on a bondage-and-discipline  post-technical
system  pawned off by the nation's largest oughta-be-illegal monopoly
who cannot escape the sins of their forefathers -- namely, using  the
wrong  slash for directories when the C language and its brethren use
it for something else that's very important.
         -- Tom Christiansen in <55oabg$1j1$1@csnews.cs.colorado.edu>

[U-Boot] Patchwork question

[Wolfgang Denk]

Dear Anton Staaf,

In message <CAF6FioXiOGjHVXXVK1QLRxvw3+GZQANAYFQS8xhdfhh+fYN5kA@mail.gmail.com> you wrote:
> 
> Who should change the state of patches there? Should I mark patches
> that have been applied as Accepted?  Or should the tree maintainer do
> that?

Both :-)

> Also, should I delegate patches to custodians that I believe should be
> looking at them?  Or should custodians delegate patches to themselves?

Both :-)

The thing is, Patchwork has a number of shortcomings, and a number of
bugs.  Some patches get mis-identified, so earlier submissions
(already marked as superseded) will set to "accepted", while the
latest versions remain in state "new".  Manual cleanup is needed then.

Every help with this is more than welcome.

Ditto for assigning patches to custoians (and don;t be afraid to guess
wrong - if one custodian feels a patch is not for his plate, he can
easily re-assign it).  Better a wrong assignment than none [um...
unless you assign them incorrectly to myself ;-) ]

Best regards,

Wolfgang Denk

-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
The human race has one really effective weapon, and that is laughter.
                                                         - Mark Twain