On 12/09/12 11:35, Timo Härkönen wrote:
> Hi
>
> We'll be having a brainstorming meeting to get ideas on how to improve
> the platfrom SDK and it's usability. If you cannot attend the meeting
> and have some input, please send your thoughts by replying to this mail.
>
> Meeting will be held Tuesday 18 September from 9:00 to 10:00 UTC in
> #mer-meeting @freenode
>
> ps. in addtion to sharing thoughts we welcome helping hands to
> implement and test the improvements :)
>
> BR,
>
> Timo

Some thoughts on the matter:

1. Documentation can be improved to explain more clearly how SDK is used
with combination of Scratchbox2 and different targets. For example how
it works with using various Mer core and target related repositories,
and how updating them works (i.e. updating the core SDK vs the target
part). Presenting it more clearly will benefit developers who aren't
familiar with Mer infrastructure, and will help them to dive into using
the SDK for their targets of interest.

2. Installation (including adding more targets) is rather manual, it can
be more automated, even possibly with some optional UI which will
propose potential SB2 targets and other settings.

3. Workflow of development for Mer and derivatives if somewhat obscure
at the moment. Using purely manual method with make, cmake etc. is one
way, but how it combines for example with something created in Qt
Developer? Or should it combine? What paths various developers should
take, some for example are using IDEs for development, others don't (or
use them just for UI prototyping). Describing various approaches would
be useful.

4. Would be good to provide a method how to run an application built for
some target using Qemu and Xephyr (similar to how Harmattan SDK did it).
Otherwise the only option is to build it for x86, but it defeats the
purpose to make sure that other arch builds work (ARMv7hl, ARMv7l and so
on). Providing a streamlined recipe at least to see that something works
in the target arch build without a device will be a good option.

Regards,

Hillel.

Reply via email to