477143 - jsstack.js static analysis error in js_GetPropertyHelper

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect, P1)

Description

[Jason Orendorff [:jorendorff]]

../jsobj.cpp: In function ‘JSBool js_GetPropertyHelper(JSContext*, JSObject*, jsid, jsval*, JSPropCacheEntry**)’:
../jsobj.cpp:4011: error: cannot call JS_REQUIRES_STACK function Detecting

At this precise moment, it's unclear to me that the new code there is correct, even though I reviewed it myself just a day or two ago.

But anyway I don't think it's necessary for Detecting to read cx->fp->script, given JSOP_STOP, so we can just rip that part out.  If so, that's an easy fix, I'll post tomorrow.

Comment 1

[Brendan Eich [:brendan]]

JSOP_STOP FTW!

Sorry, exactly what new code do you mean here?

> At this precise moment, it's unclear to me that the new code there is correct,

/be

Comment 2

[Jason Orendorff [:jorendorff]]

Detecting does need cx->fp->script to read atoms, to check for `name undefined`.  So this is not an easy fix.  (But Detecting does not need to calculate pcend, so the code could be a bit simpler.)

Comment 3

[Brendan Eich [:brendan]]

Maybe it's time to pull DETECTING back from the API and make the hack be based on name advice: 'all' and only all in a detecting context (the compiler can see these). I know this was mentioned in another bug, but I can't find it now. Anyhoo if we can fix both this bug and improve detecting not to need cx->fp so much the better.

/be

Comment 4

[Jason Orendorff [:jorendorff]]

Attachment: v1

Comment 5

[Jason Orendorff [:jorendorff]]

http://hg.mozilla.org/tracemonkey/rev/ad3cd0759495

Comment 6

[Robert Sayre]

http://hg.mozilla.org/releases/mozilla-1.9.1/rev/2d121da4f43e

Comment 7

[Robert Sayre]

http://hg.mozilla.org/mozilla-central/rev/ad3cd0759495

Comment 8

[Bob Clary [:bc] (inactive)]

 v 1.9.1 - the stack analysis on tracemonkey is green.