746973 - crash in DebuggerScript_setBreakpoint

Status: RESOLVED DUPLICATE of bug 738479

(Core :: JavaScript Engine, defect)

Description

[Panos Astithas (he/him) [:past] (please ni?)]

This bug was filed from the Socorro interface and is 
report bp-a65f13a0-e5e4-4cfa-bedd-322282120419 .
============================================================= 

STR:

1) Open http://htmlpad.org/debugger/
2) Open the script debugger
3) Click the click me button
4) Press resume
5) Then at the next pause add a breakpoint on line 15
6) Firefox crashes

Comment 1

[Victor Porof [:vporof][:vp]]

Not tied to DebuggerScript_setBreakpoint, but possibly related.
Same thing happens if I just try to debug a WebGL page. For example, go to http://mrdoob.github.com/three.js/examples/misc_lookat.html and just start the debugger => Segmentation fault.

Comment 2

[Victor Porof [:vporof][:vp]]

(In reply to Victor Porof from comment #1)
Not related, there's bug 725733 for that.

Comment 3

[Jason Orendorff [:jorendorff]]

In a debug build, we crash in 

(gdb) p script->getGlobalObjectOrNull()
$3 = ('js::GlobalObject' *) 0x0
(gdb) p script->compileAndGo
$4 = true

Comment 4

[Jason Orendorff [:jorendorff]]

In a debug build we crash in ScriptGlobal, called from setBreakpoint:

#0  0x0000000104556bb4 in ScriptGlobal (cx=0x10032ba00, script=0x10d0719c0, scriptGlobal=0x0) at Debugger.cpp:279
#1  0x00000001045603bf in DebuggerScript_setBreakpoint (cx=0x10032ba00, argc=2, vp=0x10c900688) at Debugger.cpp:2843

The stack is like bug 738479, but this is a different bug; it happens for an eval script that is not on the stack. Trying to make a shell testcase.

Comment 5

[Jason Orendorff [:jorendorff]]

var g = newGlobal('new-compartment');
var dbg = Debugger(g);
g.eval("function f() { return eval('2+2'); }");
var s;
dbg.onNewScript = function (script) { s = script; };
g.f();
var hits = 0;
for (var offset of s.getLineOffsets(s.startLine))
    s.setBreakpoint(offset, {hit: function () { hits++; }});
assertEq(g.f(), 4);
assertEq(hits, 1);

Comment 6

[David Mandelin [:dmandelin]]

Jim just brought this to my attention for K9O. IIUC, it's a crash in a fairly basic script debugger use case, so +ing.

Comment 7

[Jason Orendorff [:jorendorff]]

The patch in bug 738479 fixes this after all. I added the test case from comment 5 to that patch.

I don't think we need to worry about tracking this; we have the fix and it'll land this week, assuming the tree opens.