[coreboot] branching flashrom?

Carl-Daniel Hailfinger c-d.hailfinger.devel.2006 at gmx.net
Mon Jul 21 16:36:31 CEST 2008

On 21.07.2008 16:25, Uwe Hermann wrote:
> On Sun, Jul 20, 2008 at 01:49:04PM +0200, Carl-Daniel Hailfinger wrote:
>> Completely forking the codebase into separate projects is something I'd
>> like to avoid because we already have too few development resources.
> Full ack.
>> Suggestion: After 1.0, we create
>> - a development branch where people can go wild with rewriting stuff
>> - a stable branch with incremental changes where breakage is not allowed.
> Nah, overkill and not really useful IMO.
Hm. There have been quite a few disagreements over design and code questions in the recent past. Branching would allow people who share a common vision to showcase what they intend to do without being limited to single patches.

>> I personally believe more in the incremental approach with minimal changes.
> Ack. No patch should ever break the build or runtime-behaviour of
> flashrom. That doesn't mean we can't do radical changes, such changes
> just have to be done in a manner which doesn't break flashrom, either
> in small incremental steps or in bigger patch-series which do all
> changes at once. But we don't need (or want, IMO) branches for either...

There are philosophical differences as well. Peter and Stefan want to
remove #defines and use magic values directly. I heavily disagree with
that and believe the code is more readable if the meaning of a constant
is visible in the code without having to consult data sheets. I hope
that branching is a way to avoid revert wars or NACKs for design reasons.



More information about the coreboot mailing list