Andrea Giammarchi

@dpgeorge I can confirm the jsffi does what it needs to do on both main and workers (whre it's implicit, but that's another story): ```html MicroPython - Latest function ownKeys(py_or_js)...

@dpgeorge I am not sure but it looks to me that `proxy_call_python` should *not* add the `length` property with value `0` when no arguments (which also I am not sure...

@dpgeorge > Can I just cut-and-paste verbatim your LiteralMap code into the code for this PR? if you want to cut and paste there is another version in here with...

> We could add it as a submodule I meant as *npm* module but I guess the toolchain here is quite different so ... fair enough. If you want to...

@dpgeorge > OK, for now I will leave LiteralMap out and just keep using {} as the base for conversion from a Python dict to JS entity. the idea is...

> I also retained the ability to do interpreter.globals.get("..."). this looks like the best form all worlds, as I was hoping *LiteralMap* idea would've brought in Pyodide too ... if...

@dpgeorge I've just updated polyscript and I've noticed one improvement and some "*regression*" in there ... ```html function keys(ref) { console.log('a' in ref, ref.a); // true, 1 console.log(Object.keys(ref)); // []...

this is something we can play with already, if that `._module` is *private not really* accessible on our side (I didn't - rightly so - find it in docs, I...

FWIWI it looks like *better-sqlite* has a dedicated `.pragma` command to execute *PRAGMA* related queries ... maybe that's the missing part here?

FYI `vc4-fkms-v3d` fixed the issue out of the box. The latest `vc4-kms-v3d` on 4.9 never showed up a thing on the display, only HDMI