Because the initialization of attributes in non-deterministic, we need to code 
defensively and not just assume all attributes have valid values any time any 
attribute is set.  In this case, Firefox is surely suspecting that if you try 
to draw a shape with a missing dimension, you may have made a mistake?

On 2011-02-12, at 17:20, Henry Minsky wrote:

> The cause seems to be if the width is zero in this code
> 
>           if (this.seekablefillcolor != null) {
>               console.log('drawshape', context,0,h,seekwidth,h * 2);
>                this.drawshape(context,0,h,seekwidth,h * 2);
>                context.fillStyle =
> this.seekablefillcolor.toCanvasColor(context, this._colorcache);
>                context.fill();
>            }
> 
> Apparently the drawshape code barfs in Firefox if the width arg is zero.
> 
> Now, the question is why would that be?
> 
> 
> 
> On Sat, Feb 12, 2011 at 5:15 PM, Henry Minsky 
> <[email protected]>wrote:
> 
>> 
>> 
>> I'm working on a subclass of slider which displays a colored bar for the
>> current playhead position, and a second bar to show load progress.
>> 
>> It's in  components/demos/av/test-progress.lzx
>> 
>> It gets an error when it tries to call context.fill to draw the bars. The
>> two bars are specified like this, one is a color , the other a gradient
>> 
>>        progress_slider {
>>           progress-fill-color: top, #D2D1D1 0%,#165E7D 100%;
>>           seekable-fill-color: #d2d1d1;
>>        }
>> 
>> But when instantiated, either fill gets this error in firefox
>> 
>> 
>> An invalid or illegal string was specified" code: "12
>> this.context.moveTo($4[1],$4[2]);break;
>> test-p...(false) (line 232)
>> An invalid or illegal string was specified" code: "12
>> this.context.moveTo($4[1],$4[2]);break;
>> 
>> 
>> Can someone help me figure out why this fails in Firefox, but works in
>> Safari, Opera, Chrome, IE 8?
>> 
>> 
>> --
>> Henry Minsky
>> Software Architect
>> [email protected]
>> 
>> 
>> 
> 
> 
> -- 
> Henry Minsky
> Software Architect
> [email protected]
> _______________________________________________
> Components mailing list
> [email protected]
> https://hedwig.laszlosystems.com/mailman/listinfo/components


Reply via email to