[oe] E-Ten Glofiish M800 support

Koen Kooi k.kooi at student.utwente.nl
Tue Nov 18 19:28:05 UTC 2008


On 18-11-08 19:44, Stefan Schmidt wrote:
> Hello.
>
> On Tue, 2008-11-18 at 17:57, Koen Kooi wrote:
>> On 18-11-08 16:59, Stefan Schmidt wrote:
>>> I just pushed a branch with basic machine and kernel support for the M800 from
>>> E-Ten. Normally Winodws Mobile based device. Linux support is ongoing.
>>>
>>> http://git.openembedded.net/?p=openembedded.git;a=commitdiff;h=0ea0875a3e5456f701df08e0974821a4e8852f55
>>>
>>> As it is just machine, kernel recipe and sane-srcrev it should not break
>>> anything. Still I would like to ask for review before pushing it in.
>>>
>>> How does folks feel about it?
>> Only some small comments:
>>
>> conf/machine/eten-m800.conf:
>> IMAGE_FSTYPES ?= "tar.gz"
>>
>> change that to += so people don't shoot themselves in the foot too easily
>
> Good point. Changed.
>
>> linux-eten_2.6.24+git.bb:
>> git://git.openmoko.org/git/kernel.git;protocol=git;branch=stable \
>>
>> Would it make sense to merge this into the openmoko kernel recipe since
>> it's building from the same tree? I'm trying to cut down the number of
>> linux-foo recipes whenever possible, but it a seperate recipe is easier,
>> go for it.
>
> I thought about this. My point about not doing it is that it could easily be
> that I switch the git repo soon. We just sat up a git tree with WIP stuff. Not
> sure which one I like to have in OE in the long run. I would go for keeping it
> and think about how it turns out.

Do whatever makes you happy :)

> PS: Any people prefer to have patches inlined for easier review, or are all fine
> with a link to webgit here?

Having the patches inline would make it slightly easier to review, but 
people would need to use git send-email properly to avoid e.g .spurious 
nesting.
For small patchsets like this I don't find pasting from gitweb too much 
work, but I can see it getting tedious when >5 points need addressing.
Another option would be to use something like reviewboard or gerrit to 
make pointing out dubious hunks and fixing them easier, but atm we sadly 
don't get enough stuff to review to install such infrastructure.

regards,

Koen





More information about the Openembedded-devel mailing list