p01 at opera.com
Mon Jun 4 05:37:20 PDT 2007
Maciej Stachowiak wrote:
> On Jun 1, 2007, at 6:03 PM, Ian Hickson wrote:
>> On Fri, 23 Mar 2007, Anne van Kesteren wrote:
>>> Wouldn't it be better if no INDEX_SIZE_ERR was raised but instead the
>>> previous value was retained? For consistency with
>>> CanvasRenderingContext2D.globalAlpha for instance. It's not really
>>> important, but I think that some consistency between the various APIs
>>> would be nice.
>> In general, actually, raising INDEX_SIZE_ERR is what the APIs do. So for
>> consistency, volume is correct.
>> globalAlpha, though, is not. What do people think? Should we change the
>> canvas globalAlpha attribute to raise an exception for out-of-range
>> values? Any browser vendors have an opinion?
What about INDEX_RANGE_ERR ?
This name is explicit, and consistent with INDEX_SIZE_ERR.
> In general I prefer silent failure to exceptions, since JS exceptions
> tend to lead to catastrophic failure of a whole script and thus tend to
> turn slight implementation differences (possibly misreading the spec in
> some cases) into serious compatibility issues.
Although I also like silent failure, I tend to think it leads to more
flamboyant failures of scripts. Having the spec. push exceptions
Mathieu 'p01' HENRI
More information about the whatwg