:Benjamin Peterson Assignee
	Description • 12 years ago

Now that JägerMonkey is gone, the last "valid" usage of decomposing opcodes is gone. They're merely contributing code bloat now, so let's nuke 'em.

	:Benjamin Peterson Assignee
	Comment 1 • 12 years ago

The main changes are in BytecodeEmitter.cpp. The logic for incrementing and decrementing was (is) quite convoluted. It could perhaps stand to be refactored more. (followup?)

	:Benjamin Peterson Assignee
	Comment 2 • 12 years ago

Remove some debugging code.

	Luke Wagner [:luke]
	Comment 3 • 12 years ago

Comment on attachment 750241 [details] [diff] [review]
zap

Review of attachment 750241 [details] [diff] [review]:
-----------------------------------------------------------------

Nice!  Yes, the emitter could use some further cleanup.  Btw, I recommend fuzzing before landing.

::: js/src/frontend/BytecodeEmitter.cpp
@@ +923,5 @@
>      return EmitAliasedVarOp(cx, op, pn, bce);
>  }
>  
> +static JSOp
> +GetIncDecInfo(ParseNodeKind kind, bool &post)

The usual SM style is to take outparams by *, not & (so that it is clear at the callsite).

@@ +1776,1 @@
>      pn2 = pn->maybeExpr();

Can haz ParseNode *pn2 = pn->maybeExpr()?

	:Benjamin Peterson Assignee
	Comment 4 • 12 years ago

Fuzz please.

	:Benjamin Peterson Assignee
	Comment 5 • 12 years ago

Erk, here's an actually updated patch.

	Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)
	Comment 6 • 12 years ago

Comment on attachment 750839 [details] [diff] [review]
updated for fuzzing

Didn't find anything after a round of overnight fuzzing.

	Christian Holler (:decoder)
	Comment 7 • 12 years ago

Comment on attachment 750839 [details] [diff] [review]
updated for fuzzing

Nothing found here either within 24 hours of fuzzing :)

	:Benjamin Peterson Assignee
	Comment 8 • 12 years ago

https://hg.mozilla.org/integration/mozilla-inbound/rev/4196d8b58814

	Phil Ringnalda (:philor)
	Comment 9 • 12 years ago

https://hg.mozilla.org/mozilla-central/rev/4196d8b58814