[U-Boot] [STATUS] "Quality" of patches / testing.

[Andreas Bießmann]

Dear Simon,

Am 18.10.2011 08:51, schrieb Simon Schwarz:
> Hi Wolfgang,
> 
> On 10/18/2011 08:23 AM, Wolfgang Denk wrote:
> [SNIP]
>>
>> Why is nobody doing this?
> 
> One of my big problems with this was that you not only have to run a 
> MAKEALL with your own changes but also have to do that with the former 
> state of the repo to identifie the differences.

We should first get a state of "all boards build clean" for a sort of
toolchains (I think arm is now at this state after a lot of tumult in
the last two releases).

> So one suggestion I have is to do at least a reference build of each 
> custodian repo and store the output somewhere so one can diff with these.

This is some kind of CI as suggested by Lukasz Majewski. I also favor CI
for fast feedback to the submitter if his change breaks something.

There was already a suggestion by Graeme Russ to run checkpatch on every
submitted patch. While this requires really low computing power it will
produce a lot of mail overhead. The second suggestion by Graeme was to
add an marker into the patch which states this patch as checkpatch-clean.
If we would build some CI system with automatic build we will need even
more computing power to build each patch with all/a subset of MAKEALL.
Not to mention the mail overhead for clean/broken patches. Therefor some
marker in the patch could do the job here again.

I guess it will be doable to have some scripts/prepare-patch which runs
 a) git format-patch
 b) checkpatch on the patch
 c) (configurable subset of) MAKEALL on some clean tree with that patch
    applied
 d) append the results to the patch

This tool would implement some kind of CI but utilize the computing
power of submitter.

best regards

Andreas Bießmann