721497 - Assertion failure: compartment mismatched, at ../jscntxtinlines.h:153 with dumpStack

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

[Christian Holler (:decoder)]

The following code asserts on mozilla-central revision edf8075b0333 (options -n -m -a):


var g = newGlobal('new-compartment');
var dbg = new Debugger(g);
dbg.onDebuggerStatement = function (frame) {
    assertEq(dumpStack(), true);
};
g.eval("function f(a, [b, c], {x: [y], z: w}, {q}) { debugger; }");
g.eval("f(1, [2, 3], {x: [4], z: 5}, {q: 6});");


Not sure if this is really a bug, or just a problem because of "dumpStack". There are way too many shell builtins floating around in global namespace :(

Comment 1

[Chris Leary [:cdleary] (not checking bugmail)]

(In reply to Christian Holler (:decoder) from comment #0)
> There are way too many shell builtins floating around in global namespace :(

Has anybody proposed making a Shell object (akin to the Math object)?

Comment 2

[Jason Orendorff [:jorendorff]]

Reduced:
  var g = newGlobal('new-compartment');
  g.f = function () { dumpStack(); };
  g.eval("f();");

The comment on DumpStack in the source code says:
/*
 * This shell function is temporary (used by testStackIter.js) and should be
 * removed once JSD2 lands wholly subsumes the functionality here.
 */

JSD2 clearly isn't there yet, so I'll fix the bug. 2-line patch coming.

Comment 3

[Jason Orendorff [:jorendorff]]

Attachment: v1

Comment 4

[Luke Wagner [:luke]]

Comment on attachment 622037 [details] [diff] [review]
v1

wrap all the values!

Comment 5

[Jason Orendorff [:jorendorff]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/d8f95bbf161b

Comment 6

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/d8f95bbf161b

Should this have a test?

Comment 8

[Christian Holler (:decoder)]

Automatically extracted testcase for this bug was committed:

https://hg.mozilla.org/mozilla-central/rev/efaf8960a929