Re: Canvas gradients should not be optional

On Fri, 14 Dec 2007, Philip Taylor wrote:
> 
> The spec says:
> 
>   "Support for actually painting gradients is optional. Instead of 
> painting the gradients, user agents may instead just paint the first 
> stop's color."
> 
> Firefox, Opera and Safari support gradients. Rhino Canvas does too. The 
> recent (unreleased) Konqueror with <canvas> also does gradients. 
> ExplorerCanvas attempts to support gradients, despite getting them 
> utterly wrong - presumably it was decided that buggy gradient support 
> was better than conformingly following the spec's option of no 
> gradients. I can't remember any other implementations now.
> 
> Optional features are harder to test, and a worse implementation (which 
> lacks the feature entirely) could claim better conformance than a better 
> implementation (which has the feature, and maybe fails in a few obscure 
> edge cases). Since every implementer has decided to implement gradients, 
> it seems there is no compelling reason to not implement it. So, the 
> "Support for actually painting gradients is optional" paragraph should 
> be removed.

Done.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'