I can't seem to construct a test case for the WARNING lossage, but this seems
to be where JScript is dying if I run smokecheck with backtrace on:
> for (var v_$1=$4.lineno = 341, init_$0() in {}) {
Looks like the unparser didn't plan on that comma expression being there...
On 2011-02-08, at 10:39, P T Withington wrote:
> I looked at the LFCdhtml-backtrace.js output, and it looks fine. I bet
> JScript is just broken, and we'll need to compensate for it in our unparser.
> Rats!
>
> On 2011-02-08, at 10:36, André Bargull wrote:
>
>> I'm going to investigate this and respond later!
>>
>>
>> On 2/8/2011 4:21 PM, P T Withington wrote:
>>> On 2011-02-08, at 10:09, P T Withington wrote:
>>>
>>>> And with backtrace off I am getting:
>>>>
>>>>> WARNING: The dhtml runtime does not support dhtml. dhtml
>>>>> WARNING: The dhtml runtime does not support dhtml. dhtml
>>>>
>>>> Say what?
>>>
>>> I tracked down why this warning is reading somewhat bizarrely. This
>>> statement:
>>>
>>> var check = capabilityname == '' ? '' : 'Check "canvas.capabilities.' +
>>> capabilityname + '" to avoid this warning.';
>>>
>>> seems to be being parsed as:
>>>
>>> var check; (check = capabilityname) == '' ? '' : 'Check
>>> "canvas.capabilities.' + capabilityname + '" to avoid this warning.';
>>>
>>> Is this a result of your recent parser changes?
>>>
>>>
>
