Re: Trailing semi-colons (Re: Agenda, 11 September 2014 SVG WG telcon)

Hi Erik,

----- Original Message -----
> From: "Erik Dahlström" <ed@opera.com>
> To: www-svg@w3.org, "David Dailey" <ddailey@zoominternet.net>, "Brian Birtles" <bbirtles@mozilla.com>
> Sent: Thursday, 11 September, 2014 5:18:53 PM
> Subject: Re: Trailing semi-colons (Re: Agenda, 11 September 2014 SVG WG  telcon)
> 
> On Thu, 11 Sep 2014 01:18:59 +0200, Brian Birtles <bbirtles@mozilla.com>
...
> > This has been discussed before:
> >
> >    http://lists.w3.org/Archives/Public/www-svg/2011Nov/0136.html
> 
> Yes, it has been discussed, but for semicolon-separated lists only. I'd
> like to expand this topic to cover all list-of-Ts.

Ok, I didn't understand that.

> > There is a more thorough test case here:
> >
> >    http://brian.sol1.net/svg/interop/semicolon-test.svg
> >
> > It seems Chrome is inconsistent in where it allows trailing semi-colons
> > and where it does not.
> 
> Yes, there is one bug open where someone asks for ignoring the trailing
> semicolon in the keySplines attribute, that's the only one I'm aware of
> atm. AFAICT Chrome/Opera and old Opera (Presto) all behaved exactly the
> same on semicolon-test.svg. Unless I'm missing some subtle detail it
> looked consistent across all the browsers that support SMIL, except for
> how "red;green;;" is handled.

Yes, those are the two differences I observe.

> > I'd rather we just resolve this on the list if possible.
> 
> Is your opinion still the same as it was in the old thread? I'm willing to
> accept the proposal there, on the condition that we also go ahead with
> allowing trailing separators in list-of-Ts.

Not receiving any objection, I implemented the behavior proposed in semicolon-test.svg. So yes, I still think that behavior is preferred in light of the compatibility constraints we faced. The main reason we decided to make "a;b;" only produce 2 values was because of compatibility (specifically I think it was content generated by Ikivo Animator).

It seems reasonable to apply that behavior elsewhere if there are no compatibility constraints. 

Best regards,

Brian

Received on Thursday, 11 September 2014 10:44:19 UTC