On Fri, Apr 08, 2011 at 12:31:25AM +0100, Neil Jerram wrote:
> Martin Jansa <[email protected]> writes:
> 
> > In short:
> > There are couple of layers
> > oe-core with well tested core stuff and higher quality requirements for 
> > patches
> > meta-oe for recipes/classes and other usefull stuff from "old" oe
> > meta-efl similar rules like meta-oe but having only stuff related to efl/e17
> > meta-shr recipes/classes used only in SHR
> >
> > meta-angstrom recipes/classes used only in Angstrom
> > meta-ti bsp layer for ti devices
> > etc...
> >
> > each layer has own maintainers which apply stuff sent in form of git
> > pull requests, so there is lower chance that someone will accidentaly
> > break something (something like linux kernel is using for subtrees)
> 
> Ah, OK.  And this is what you're overall calling shr-core?

Yes, it will replace shr-unstable when it's ready, but before that, I'm
calling it shr-core (as shr based on openembedded-core). 

Currently I'm pushing most patches to both repositories (old OE master 
and appropriate layer), when I consider shr-core stable enough to be 
called shr-u I won't push it to OE master (and probably nobody else will).

> I suppose, then, that this may be the "happy medium" that I was looking
> for in the previous email?

I hope so.

> And if it is, would you agree that we then probably won't need SHR-T any
> more?  Instead, we can just snapshot SHR-C every now and then.

SHR-T are just snapshots of SHR-U every now and then already, so with
SHR-C becoming new SHR-U later it should stay the same.

Cheers,

-- 
Martin 'JaMa' Jansa     jabber: [email protected]

Attachment: pgpKLDBrGLSK9.pgp
Description: PGP signature

_______________________________________________
Shr-devel mailing list
[email protected]
http://lists.shr-project.org/mailman/listinfo/shr-devel

Reply via email to