[L2Ork-dev] Feedback on GSOC draft proposal

[Jonathan Wilkes]

On Sun, Apr 11, 2021 at 2:51 PM Tanmay Gujar <tanmaygujar999 at gmail.com> wrote:
>
> Hello and thanks for the feedback!
> I'll make the changes to the draft so that the REPL would use a socket connection instead.
>
> As for the delete function I was working with the info presented here Pd Internal messages.

Don't use that.

First, skim this:

https://git.purrdata.net/jwilkes/purr-data#core-pd-notes

It has some helpful information about type and file prefixes.

As for Pd's interface:

m_pd.h: the *official* public interface (Purr Data's m_pd.h isn't in
lockstep with Pd Vanilla's m_pd.h, but most things seem to load and
work properly.)

m_imp.h: the internal implementation header file

g_canvas.h: stuff to do with visualizing canvases as well as glist bookkeeping

g_all_guis.h: header for common (though quite complicated, esoteric,
and confusing) interface for most of the core UI widgets

s_stuff.h: misc

I say "official" because libpd reaches into not only some of those
private headers above, but also internal functions in order to give
control of DSP scheduling over to the program that's linking to it.

-Jonathan

>
> Best,
> Tanmay
>
> On Sun, Apr 11, 2021 at 11:41 PM Jonathan Wilkes <jon.w.wilkes at gmail.com> wrote:
>>
>> On Sun, Apr 11, 2021 at 6:25 AM Tanmay Gujar <tanmaygujar999 at gmail.com> wrote:
>> >
>> > Hello!
>> >
>> > I have slightly changed my approach towards building the Terminal REPL. Majorly, I plan to build on top of libpd instead of working with a socket connection to pd.
>>
>> You can do that, but we don't currently ship libpd with Purr Data.
>> (It's in the emscripten branch but even there most of the Web
>> App<->Backend communication doesn't use it.)
>>
>> Furthermore, the socket connection would work for both for a "-nogui"
>> console-based REPL *and* for opening up a REPL somewhere in the GUI
>> (say, a chat window in the Pd Window, or possibly in a canvas window,
>> too.)
>>
>> >
>> > Over the past few days I have been working on a prototype of the same (available at https://github.com/tgujar/pd-REPL). It is by no means complete and I am still working on it. However, it has allowed me to work out some pd specific issues in implementing the REPL. I have documented these in the Inference section in the README.
>>
>> Interesting!
>>
>> I'm not sure what you mean by this: "No delete function in Pd Internal
>> Messages." There is glist_delete for removing an object from a canvas
>> and freeing it. Internally it uses gobj_delete which simply calls the
>> class destructor. glist_delete is in g_canvas.h. So if you just add
>> something like an x_repl.c and x_repl.h to the core you can easily
>> access that function.
>>
>> -Jonathan
>>
>> >
>> > Finally, apologies for being late, I have submitted my draft proposal and would love to have some feedback on it.
>> >
>> > Best,
>> > Tanmay
>> > _______________________________________________
>> > L2Ork-dev mailing list
>> > L2Ork-dev at disis.music.vt.edu
>> > https://disis.music.vt.edu/listinfo/l2ork-dev
>> _______________________________________________
>> L2Ork-dev mailing list
>> L2Ork-dev at disis.music.vt.edu
>> https://disis.music.vt.edu/listinfo/l2ork-dev
>
> _______________________________________________
> L2Ork-dev mailing list
> L2Ork-dev at disis.music.vt.edu
> https://disis.music.vt.edu/listinfo/l2ork-dev