[oe] [RFC] Palms commits.

Sergey Lapin slapinid at gmail.com
Tue Jul 10 19:11:48 UTC 2007


Paul Sokolovsky wrote:
> Hello Sergey,
> 
> Tuesday, July 10, 2007, 6:45:11 PM, you wrote:
> 
>> Hi, all!
> 
>> I just had raging argument with Paul Sokolovsky about my commits,
>> so I'd like to RFC them, and also I'd like to ask if all these changes
>> should be passed through RFC.
> 
>> Please, see attachments. Sorry for post-factum.
> 
>> Briefly, these commits generalize palms support and make maintenance
>> easier, also, update Zire 72 configuration.
> 
>   Well, I regret that even after discussion on IRC you still miss the
> point. The talk was not about "palm commits" per se, but about
> refactors to make machines more maintainable. OE obviously has
> noticeable problems with that, and trying to solve them in local
> machine corners won't benefit OE, but in the end just will become the
> same mess but in the other plane.

And I was telling you that in my position, such big start is doubtful,
if not impossible by several reasons, among which I need to learn
infrastructure mechanics in OE, and I need to get more time and effort,
because of busy schedule. So it is not possible _immediately_.

Your argument still amazes me.


> 
>   So, if you have bright ideas how to optimize maintenance, please
> think how it fits with already existing machine support framework
> (task-base), share them with wider audience, and call for feedback.
> I'm sure there're parties interested in the same changes, and getting
> together may lead to much better result for OE as a whole.

I never disagreed on this thing, since it is generic enough as a concept.

> 
>   Of course, you're free to treat those changes as Palm local, but
> what I wrote still holds - machine configs in OE need refactoring and
> elaboration. If you don't want to put a bit more effort into that and
> make your changes useful universally, then someone else will need to
> duplicate large part of your work, and possibly achieve incompatible
> results with your current scheme. As that someone hopefully will
> target it on global level, then palm config would need to be changed
> too, so in the end it will appear turn out that you lost time too.

I understand your point here, but if I prefer only start doing
something when I can handle that thing, and when I can be
responsible for it. If at the moment I can finish 1 little thing,
then that's what I'm going to do. If I have resources to spend on
bigger problem, I go and solve one (or help somebody with even bigger issues).

I have a list of smaller thingies I can do just passing by, and bigger ones,
which I leave for bigger windows of time.

Do you understand my time management better now? That's probably would
clear our discussion?

> 
>   Summing up, by investing 50% more effort into that, there can be 5
> times more outcome. And as a "fresh blood" in OE, you're in good
> position to lead this, as other developers are of course stuck with
> other issues.

Do you count yourself in this 50%, btw?

Best regards,

S.





More information about the Openembedded-devel mailing list