352084 - Decompiler omits parens needed for comma to be seen as "sequential evaluation"

Status: VERIFIED FIXED

(Core :: JavaScript Engine, defect)

Description

[Jesse Ruderman]

The decompiler omits parens around sequential-evaluation expressions in at least these two contexts.  In both contexts, this results in the comma being interpreted as something other than sequential-evaluation.

js> function() { h = {x:5, y:(a,b)}}  
function () {
    h = {x:5, y:a, b};
}

js> function() { x(a, (b,c)) }
function () {
    x(a, b, c);
}

Comment 1

[Jesse Ruderman]

This is a regression (perhaps from bug 352022?).

Comment 2

[Jesse Ruderman]

Two more:

js> function() { return [(x, y) for each (z in [])] }
function () {
    return [x, y for each (z in [])];
}

js> function() { return [(a, b), c]; }
function () {
    return [a, b, c];
}

Comment 3

[Brendan Eich [:brendan]]

(In reply to comment #1)
> This is a regression (perhaps from bug 352022?).

No, this is from the big re-do of opcode precedence, bug 350531.

It'll be fixed by the same patch that fixes bug 352026.  The virtue of using opcode precedence is that we don't need to add lots of special cases in code.  The vice (temporary) is that you'll have to bear with followup bugs to bug 350531, but there's light at the end of this tunnel.

/be

Comment 4

[Brendan Eich [:brendan]]

Fixed by patch for bug 352026.

/be

Comment 5

[Bob Clary [:bc] (inactive)]

Checking in regress-352084.js;
/cvsroot/mozilla/js/tests/js1_6/Regress/regress-352084.js,v  <--  regress-352084.js
initial revision: 1.1
done

Comment 6

[Bob Clary [:bc] (inactive)]

verified fixed 1.9 20060914 windows/mac*/linux
bug not evident on 1.8...

Comment 7

[Bob Clary [:bc] (inactive)]

fixed by Bug 352026
verified fixed 20061203 windows/linux/mac*