[RFC] push those patches, push'em up

[RFC] push those patches, push'em up

[Rolf Leggewie]

Hi,

OE has quite many patches that are just too valuable to keep to
ourselves.  In the spirit of our bug sprint weekends, I want to propose
patch-pushing weekends.  Maybe we can even replace the former with the
latter.  I'd like to propose the coming weekend as the first of this kind.

To get everybody on the same page and to make sure this is a sustainable
 effort, I'd also like to propose some soft policy for patches committed
to OE.  This should help us quickly find those patches still in need of
some pushing.

 1) first line in a patch starts with upstream: and goes on to list the
    URL where the bug has been reported upstream.
 2) further information can optionally be listed in the following fields
    * status: pending, accepted in XXX, rejected
    * origin: where the patch has been stolen ;-)
    * comment: any further detail such as description or reason for
               application of the patch

Comments?  Volunteers to join the cause?

Regards

Rolf

Re: [RFC] push those patches, push'em up

[Paul Sokolovsky]

Hello,

On Wed, 13 Feb 2008 01:41:02 +0100
Rolf Leggewie <no2spam@nospam.arcornews.de> wrote:

> Hi,
> 
> OE has quite many patches that are just too valuable to keep to
> ourselves.  In the spirit of our bug sprint weekends, I want to
> propose patch-pushing weekends.  Maybe we can even replace the former
> with the latter.  I'd like to propose the coming weekend as the first
> of this kind.
> 
> To get everybody on the same page and to make sure this is a
> sustainable effort, I'd also like to propose some soft policy for
> patches committed to OE.  This should help us quickly find those
> patches still in need of some pushing.
> 
>  1) first line in a patch starts with upstream: and goes on to list
> the URL where the bug has been reported upstream.
>  2) further information can optionally be listed in the following
> fields
>     * status: pending, accepted in XXX, rejected
>     * origin: where the patch has been stolen ;-)
>     * comment: any further detail such as description or reason for
>                application of the patch
> 
> Comments?  Volunteers to join the cause?

Nice idea. Unfortunately, I cannot join at full force, as have my queue
filled with lots of other stuff, by I submitted another GPE patch made
during Angstrom release, and sent nag letter to the GPE list (previous
patches are without response for almost 3 months now).

> 
> Regards
> 
> Rolf

[]


-- 
Best regards,
 Paul                          mailto:pmiscml@gmail.com

Re: [RFC] push those patches, push'em up

[Rolf Leggewie]

Paul Sokolovsky wrote:
> I submitted another GPE patch made during Angstrom release

Yeah, I just saw that.  Great and many thanks.

Re: [RFC] push those patches, push'em up

[Leon Woestenberg]

Rolf,

On Wed, Feb 13, 2008 at 1:41 AM, Rolf Leggewie
<no2spam@nospam.arcornews.de> wrote:
>
>  OE has quite many patches that are just too valuable to keep to
>  ourselves.  In the spirit of our bug sprint weekends, I want to propose
>  patch-pushing weekends.  Maybe we can even replace the former with the
>  latter.  I'd like to propose the coming weekend as the first of this kind.
>
Excellent idea. I think I can join in, see you on #oe.

Everyone, please take care not to corrupt the patches with whitespace, newlines.

Regards,
-- 
Leon

Re: [RFC] push those patches, push'em up

[Michael 'Mickey' Lauer]

On Wednesday 13 February 2008 01:41:02 Rolf Leggewie wrote:
> Hi,
>
> OE has quite many patches that are just too valuable to keep to
> ourselves.  In the spirit of our bug sprint weekends, I want to propose
> patch-pushing weekends.  Maybe we can even replace the former with the
> latter.  I'd like to propose the coming weekend as the first of this kind.

This is an awesome idea, we really need to get this started sooner than later.

> To get everybody on the same page and to make sure this is a sustainable
>  effort, I'd also like to propose some soft policy for patches committed
> to OE.  This should help us quickly find those patches still in need of
> some pushing.
>
>  1) first line in a patch starts with upstream: and goes on to list the
>     URL where the bug has been reported upstream.
>  2) further information can optionally be listed in the following fields
>     * status: pending, accepted in XXX, rejected
>     * origin: where the patch has been stolen ;-)
>     * comment: any further detail such as description or reason for
>                application of the patch

Sounds pretty good to me.

:M:
-- 
Dr. Michael 'Mickey' Lauer | IT-Freelancer | http://www.vanille-media.de

Re: [RFC] push those patches, push'em up

[Holger Freyther]

On Wednesday 13 February 2008 01:41:02 Rolf Leggewie wrote:
> Hi,
>
> OE has quite many patches that are just too valuable to keep to
> ourselves.  In the spirit of our bug sprint weekends, I want to propose
> patch-pushing weekends.  Maybe we can even replace the former with the
> latter.  I'd like to propose the coming weekend as the first of this kind.
>
> To get everybody on the same page and to make sure this is a sustainable
>  effort, I'd also like to propose some soft policy for patches committed
> to OE.  This should help us quickly find those patches still in need of
> some pushing.
>
>  1) first line in a patch starts with upstream: and goes on to list the
>     URL where the bug has been reported upstream.
>  2) further information can optionally be listed in the following fields
>     * status: pending, accepted in XXX, rejected
>     * origin: where the patch has been stolen ;-)
>     * comment: any further detail such as description or reason for
>                application of the patch
>
> Comments?  Volunteers to join the cause?

Hi,
I think we should make the patch header mandantory for new patches and it 
should state:
	-The issue it is fixing and why it is fixing it (comment:)
	-The status: as above, or 'HACK' to mark a patch as such
	-The origin where we got this patch from and then write in the comment if you 
	  change the patch.
	-And if the bug has been filed add the bug URL to the patch as this makes 
	  tracking easy.

I will start doing so.

z.

Re: [RFC] push those patches, push'em up

[Rolf Leggewie]

Looks like somebody thought it was a good idea to have a read-only wiki.
 So, my current draft has to go to
http://oz.leggewie.org/wip/patches-upstream.txt (which is also
read-only, except for me ;-)  Comments welcome.

I hope we can increase the count from

 find packages/ -name *.patch |xargs egrep -l \^upstream\:|wc -l
 find packages/ -name *.diff  |xargs egrep -l \^upstream\:|wc -l

a little more over the next few remaining hours of this first sprint.

Re: [RFC] push those patches, push'em up

[Robert Schuster]

[-- Attachment #1: Type: text/plain, Size: 938 bytes --]

Hi,
while I like the idea I wonder what we do with patches taken from other
famous distributions.

E.g. I like apt-get source <package> to see which flaws the Debianistas
already found and worked-around/fixed and often just include their
patches into OE. Since the Debian team is often interested to get their
patches merged upstream, that means less work for us. :)

Regards
Robert

Rolf Leggewie schrieb:
> Hi,
> 
> OE has quite many patches that are just too valuable to keep to
> ourselves.  In the spirit of our bug sprint weekends, I want to propose
> patch-pushing weekends.  Maybe we can even replace the former with the
> latter.  I'd like to propose the coming weekend as the first of this kind.
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
> 



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 260 bytes --]

Re: [RFC] push those patches, push'em up

[Rolf Leggewie]

Robert Schuster wrote:
> while I like the idea I wonder what we do with patches taken from other
> famous distributions.

Hi Robert,

thank you for the important question.  I guess there is not one single
answer, but I think taking patches from other distributions makes a lot
of sense.  That is where the proposed and optional "origin:"-line would
come into play.

Now what to do about "upstream:".  I guess there are two possibilities here.

1) you can establish that the patch is already pushed upstream.  Then
   all you need to do is include the URL of the bug report somebody else
   made.  You get extra kudos for making a comment in that bug report
   that OE also uses and endorses the patch.  This is the most desirable
   and thus you should search the upstream bug tracker first.
2) you cannot establish that the patch is already pushed upstream.  I'd
   say, it is best then to open a new bug report upstream, include the
   URL to it in "upstream:" and let the upstream people mark it as dupe
   should there have been another report.  Mentioning the debian origin
   should increase the likelihood of it getting applied, I assume.

I just realize now that this approach depends on debian and other
distros using public channels to push their patches upstream.  Do you
happen to know if that is usually the case?

Regards

Rolf