876429 - BaselineCompiler: Compile JSOP_SETCALL

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

[Jan de Mooij [:jandem]]

Attachment: Patch

SETCALL is emitted for things like f() = 1 or f()++, it just throws a ReferenceError.

The patch fixes the number of uses/defs of this op, the Baseline part is just a VM call.

Uncommon, but I'd like Baseline to compile everything, so that includes a bunch of these weird ops.

Comment 1

[Jan de Mooij [:jandem]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/87b0f81d811f

Comment 2

[Ed Morley [:emorley]]

Backed out for jsreftest failures:
https://tbpl.mozilla.org/?tree=Mozilla-Inbound&rev=87b0f81d811f&jobname=jsreftest

https://hg.mozilla.org/integration/mozilla-inbound/rev/9417ffbafc41

Comment 3

[Jan de Mooij [:jandem]]

Attachment: Follow-up fix

Bleh, SETCALL can also be emitted for destructing assignment. This used to magically work due to JSOP_SETCALL's nuses/ndefs, this patch handles the case explicitly.

Comment 4

[Brian Hackett [Laid off!]]

Comment on attachment 755872 [details] [diff] [review]
Follow-up fix

Review of attachment 755872 [details] [diff] [review]:
-----------------------------------------------------------------

::: js/src/jit-test/tests/baseline/setcall.js
@@ +35,5 @@
> +    var v1, v2, v3;
> +    var z = [1, 2, 3];
> +    for (var i=0; i<15; i++) {
> +       if (i > 12)
> +           [v1, v2, g(), v3] = z

If SETCALL exists for webcompat, I guess I don't understand why this isn't a syntax error since the web doesn't yet use destructuring assignments.

Comment 5

[Jan de Mooij [:jandem]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/8323f6dbe1a1

(Try: https://tbpl.mozilla.org/?tree=Try&rev=220e45bc7f96)

(In reply to Brian Hackett (:bhackett) from comment #4)
> If SETCALL exists for webcompat, I guess I don't understand why this isn't a
> syntax error since the web doesn't yet use destructuring assignments.

Yeah, I wondered the same :)

Comment 6

[Jeff Walden [:Waldo]]

Easily done if someone wants to do it -- Parser<FullParseHandler>::makeSetCall(ParseNode *pn, unsigned msg) if you care.  You know you want to... ;-)

Comment 7

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/8323f6dbe1a1