Skip to contentSkip to conversation

Upstreaming patches

DISCUSSION Philip Balister 4 participants 4 messages Apr 11, 2011 – Apr 12, 2011

lore
Share & tools
lore
#0 lore raw

Saul Wold asked me about upstreaming some of the patches in OE. Apparently, Intel has some people who can look at submitting some of our patches upstream.

The question is, what do we do about attribution? The problem is that is not simple for many cases working out who created the patch. Rather than spend time chasing through the revision history, it would be simpler to send the patches upstream mentioning the source as derived from OE.

My thoughts are no reasonable OE contributor would object to someone else upstreaming their patches :), but I thought I should ask people what they thought.

Obviously, if there is a complex patch the situation is different, but I suspect most of the easy patches do not fall into this category.

Comments?

Philip

#1 lore raw
Quoted 13 lines · Philip Balister

Saul Wold asked me about upstreaming some of the patches in OE. Apparently, Intel has some people who can look at submitting some of our patches upstream.

The question is, what do we do about attribution? The problem is that is not simple for many cases working out who created the patch. Rather than spend time chasing through the revision history, it would be simpler to send the patches upstream mentioning the source as derived from OE.

My thoughts are no reasonable OE contributor would object to someone else upstreaming their patches :), but I thought I should ask people what they thought.

Quoted 7 lines

Obviously, if there is a complex patch the situation is different, but I suspect most of the easy patches do not fall into this category.

Comments?

Philip

Sounds good. I think omission of attribution if no original author can be identified is fine. Its more important that the work not be lost.

Please note that many patches submitted upstream have already been ignored by upstream.

-Graham

#2 lore raw

Am Dienstag, den 12.04.2011, 09:29 +0930 schrieb Graham Gower:

Quoted 4 lines · Philip Balister
Quoted 3 lines

Saul Wold asked me about upstreaming some of the patches in OE. Apparently, Intel has some people who can look at submitting some of our patches upstream.

That is great. It is indeed very time consuming to find out the correct place to send patches and register at the ticketing/bug tracking system or – most of the time – subscribe to the corresponding mailing list.

Quoted 9 lines
Quoted 9 lines

The question is, what do we do about attribution? The problem is that is not simple for many cases working out who created the patch. Rather than spend time chasing through the revision history, it would be simpler to send the patches upstream mentioning the source as derived from OE.

My thoughts are no reasonable OE contributor would object to someone else upstreaming their patches :), but I thought I should ask people what they thought.

Quoted 7 lines
Quoted 4 lines

Obviously, if there is a complex patch the situation is different, but I suspect most of the easy patches do not fall into this category.

Comments?

Sounds good. I think omission of attribution if no original author can be identified is fine. Its more important that the work not be lost.

In my opinion at least a

        git log --follow recipe/path/to/patch/file.patch

should be tried. Additionally to make things maybe easier it could be searched for patches with a patch header [1] and track their status.

Quoted 2 lines

Please note that many patches submitted upstream have already been ignored by upstream.

Hopefully those patches have a patch header and the report can be updated easily.

Thanks,

Paul

[1] http://openembedded.org/index.php/Push_patches_upstream

#3 lore raw
In reply to Graham Gower
Quoted 5 lines · Graham Gower

Sounds good. I think omission of attribution if no original author can be identified is fine. Its more important that the work not be lost.

Please note that many patches submitted upstream have already been ignored by upstream.

No harm in trying again.

/Esben