Drop ctx argument from functions that don't need it

[bnoordhuis]

Drop the unused JSContext argument from several oft-used functions.

Non-functional change (strip build/qjs && ls -l build/qjs is the same before and after) but improves legibility, IMO. No change to the public API either.

[saghul]

JS_NewInt32 and js_int32 are the same, what is the reasoning for having "duplicate" implementations? Is it the force-inline?

[saghul]

Shall we drop JS_DupValueRT now?

[saghul]

LGTM with a question.

[bnoordhuis]

JS_NewInt32 and js_int32 are the same, what is the reasoning for having "duplicate" implementations? Is it the force-inline?

No, it's purely stylistic / unclutter.

If we're okay with backwards-incompatible API changes, or versioning the API somehow, we could make the change directly to quickjs.h.

Shall we drop JS_DupValueRT now?

Good point! I was meaning to eradicate the two remaining calls from quickjs.c but I forgot. I'll fix that.

As to your general question: it serves no purpose so... yes?

[saghul]

JS_NewInt32 and js_int32 are the same, what is the reasoning for having "duplicate" implementations? Is it the force-inline?

No, it's purely stylistic / unclutter.

If we're okay with backwards-incompatible API changes, or versioning the API somehow, we could make the change directly to quickjs.h.

Shall we drop JS_DupValueRT now?

Good point! I was meaning to eradicate the two remaining calls from quickjs.c but I forgot. I'll fix that.

As to your general question: it serves no purpose so... yes?

I am somewhere near +0.5, I guess if we are doing a new thing might as well drop the baggage? OTOH, having existing compile as-is looks like a big plus, so let's have it!

[bnoordhuis]

I'll go ahead and merge this and then ponder for a bit whether or not to change the public API.