[Laszlo-dev] /test/lfc/data/alldata.lzx is failing in dhtml

[P T Withington]

Ha!  I forgot we already talked about this.  Thanks for the reminder.   
Yes, it would be better if we could not look at _instanceAttrs at all.

On 2008-04-03, at 07:51 EDT, André Bargull wrote:
> Also see this thread on the laszlo-dev mailing-list: http://www.openlaszlo.org/pipermail/laszlo-dev/2008-March/013631.html
>
> On 4/3/2008 1:13 PM, P T Withington wrote:
>> Yes my change to create instance-classes has resulted in  
>> _instanceAttrs being null some times.  I think that was possible  
>> even before my changes, but never occurred.
>>
>> OTOH, I think this is really my bug, because attributes like `name`  
>> have to come as _instanceAttrs so that they will invoke the setter  
>> when the instance is created.  I am working on a fix for this and  
>> should have it early this morning.
>>
>> There is no harm in verifying _instanceAttrs is non-null (it is  
>> declared as so), so if you are being blocked by this, feel free to  
>> add that check.
>>
>> On 2008-04-03, at 03:39 EDT, André Bargull wrote:
>>> I've already talked to Tucker about that issue some days ago, I  
>>> can send a changeset for the LzDataset special-naming thing.
>>> [Most likey, that will be at 14pm or 15pm (CET).]
>>>
>>> - André
>>>
>>>> Hi Tucker,
>>>>
>>>> I think there is an error coming from your 4/1 changest. There is  
>>>> some code in LzDataset which is failing when /test/lfc/data/ 
>>>> alldata.lzx is run in dhtml:
>>>>
>>>> this._instanceAttrs has no properties
>>>>
>>>> The code is:
>>>>
>>>>    if (this._instanceAttrs['name'] == null) {
>>>>        this._instanceAttrs['name'] = 'localdata';
>>>>    }
>>>>
>>>> Should I add a check that this._instanceAttrs exists? I'm debugging
>>
>>