GSoC Project : Package Micro-python

Wed Mar 24 18:43:30 UTC 2021

On Wed, Mar 24, 2021 at 11:38 AM Eshan Dhawan <eshandhawan51 at gmail.com> wrote:
>
>
>
> On Wed, Mar 24, 2021 at 12:34 AM Gedare Bloom <gedare at rtems.org> wrote:
>>
>> On Tue, Mar 23, 2021 at 12:16 PM Eshan Dhawan <eshandhawan51 at gmail.com> wrote:
>> >
>> >
>> > Apologies for the late reply.
>> >
>> > On Mon, Mar 22, 2021 at 10:27 PM Joel Sherrill <joel at rtems.org> wrote:
>> >>
>> >>
>> >>
>> >> On Mon, Mar 22, 2021 at 11:55 AM Gedare Bloom <gedare at rtems.org> wrote:
>> >>>
>> >>> On Mon, Mar 22, 2021 at 10:50 AM Joel Sherrill <joel at rtems.org> wrote:
>> >>> >
>> >>> >
>> >>> >
>> >>> > On Mon, Mar 22, 2021 at 11:30 AM Gedare Bloom <gedare at rtems.org> wrote:
>> >>> >>
>> >>> >> On Sat, Mar 20, 2021 at 12:33 PM Eshan Dhawan <eshandhawan51 at gmail.com> wrote:
>> >>> >> >
>> >>> >> > Hello Everyone,
>> >>> >> > I wanted to take Packaging Micro Python up as GSOC project this summer and the project will also include packaging LUA and picoC
>> >>> >> > The ticket for Micro Python  : https://devel.rtems.org/ticket/4349
>> >>> >> > What would be the complete Scope of the project?
>> >>> >> > And what would be a good starting point?
>> >>> >> >
>> >>> >>
>> >>> >> Well, I guess Joel must have described the task, so I'll leave it to
>> >>> >> him to fill in some more details.
>> >>> >>
>> >>> >> Adding RSB packages may be not sufficient coding work for GSoC. It is
>> >>> >> important in the proposal to identify what would be the coding
>> >>> >> activities involved in this project. For example, we know from
>> >>> >> experience that Lua can just be built from some minor tailoring of its
>> >>> >> Makefile, so the package is very straightforward. However, the
>> >>> >> projects you mention are scripting environments, so maybe creating a
>> >>> >> framework in RTEMS for a "shell/intepreter" that can be built as an
>> >>> >> add-on by RSB would be a proper way to scope this effort
>> >
>> > Packaging might not be a lot of coding part but adding its documentation and its example would be a very iterative and time consuming process.
>>
>> Remember that code is what counts, while we expect the other stuff to
>> come along too, you don't want to be doing 90% doco and 10% code. Just
>> keep it in mind.
>
> What would be a good inclusion to this project ?
> I was thinking long double support since I worked on porting POSIX functions I might find it easier.
> But it might interfere with matt's project if I understand that project correctly.

Right, please don't include that. You'll want to think/talk through
(with Joel, maybe) what could be good code contributions. If the RSB
packaging is fairly minimal, then creating a suite of examples might
be one way to increase the SLOC contributions. I also think there is
merit to the idea of creating a "plug-in" way to add shells to RTEMS.
Maybe even refactoring our current shell out to a add-on package then.
Just a thought.

>>
>>
>> >>>
>> >>> >
>> >>> >
>> >>> > I agree that Lua and Micropython should build easy but I had more
>> >>> > in mind.
>> >>> >
>> >>> > The full project was language stacks for RTEMS with a better user
>> >>> > experience for Micropython, Lua, Tcl, etc although I am not sure what
>> >>> > etc would entail. I am not sure all three can be completed in the new
>> >>> > GSoC timeframe. All would follow the same pattern:
>> >
>> > Etc can be managed while framing the proposal according to how time is being managed.
>> >>>
>> >>> >
>> >>> > + RSB package offering a reasonable default and access to configuration
>> >>> > + Examples including at least bare embedded, use of custom commands,
>> >>> > and integrating with RTEMS shell commands Perhaps  interactive use with
>> >>> > command line history and editing integrated if we have that as a library now.
>> >>> > + Documentation specific to RTEMS and the examples
>> >>> >
>> >>> > I imagined completely parallel kits for each embedded language we wanted
>> >>> > to support.
>> >>> >
>> >>> > Does that help? Should he plan on Micropython and Lua?
>> >>> >
>> >>>
>> >>> Sure. Lua should be easy way to get started and develop the
>> >>> framework/infrastructure side in Phase 1. Phase 2 could be extension
>> >>> to micropython / other scripting languages.
>> >
>> > Since all the languages will have a similar pattern complex work can be put in phase 2.
>> > From my past experience, it is the part when most work is done :)
>>
>> True, but for repeat students, we do expect a bit more acceleration in
>> the first phase. Usually, we want to see code merged in phase 1 by
>> repeat students. Just a reminder that the bar is higher :)
>
> :)
>>
>>
>> >>
>> >>
>> >> OK.
>> >>>
>> >>>
>> >>> I'm not sure about the RSB design of things, and whether they should
>> >>> be parallel or capable of integration. Would anyone want to use
>> >>> multiple interpreters in the same application? If so, they should
>> >>> build together to avoid conflicts. If not, parallel is fine.
>> >
>> > building them can be set to build flags,
>> > but there still needs to be a way if we want to build the package other than the default way.
>> > (any ideas on how to do that )
>> >>
>> >>
>> >> I don't see any reason on our side why that shouldn't work but we
>> >> can't guarantee they don't have symbol conflicts. And I'm not sure
>> >> it would make much sense to integrate both at the same time.
>> >>
>> >> I'd think you could install both but we'd focus on only using one
>> >> at a time.
>> >>
>> >> --joel
>> >>>
>> >>>
>> >>> > --joel
>> >>> >
>> >>> >>
>> >>> >> > Thanks
>> >>> >> > - Eshan
>> >>> >> > _______________________________________________
>> >>> >> > devel mailing list
>> >>> >> > devel at rtems.org
>> >>> >> > http://lists.rtems.org/mailman/listinfo/devel
>> >>> >> _______________________________________________
>> >>> >> devel mailing list
>> >>> >> devel at rtems.org
>> >>> >> http://lists.rtems.org/mailman/listinfo/devel