- From: Chris Lilley <chris@w3.org>
- Date: Wed, 13 Nov 2002 11:45:51 +0100
- To: www-svg@w3.org, "Sigurd Lerstad" <sigler@bredband.no>
Hello Sigurd, www-svg, You wrote http://lists.w3.org/Archives/Public/www-svg/2002Jul/0027.html > in the SVG spec, the interface SVGNumber is > specified, and SVGNumberList contain pointers to these > object types. And it's even stated that if an object is already in another > list, it should be removed from that list first before being inserted in a > new list. Yes. That behavior is correct. It can't be in two places at once, even for a short while. > I'm just making sure that this is correct (which precludes me from just > using double in C++ COM , but that I have to wrap it inside an > object) > > Is that correct? No, its not correct. The list behavior is the important thing. Whether you choose to wrap it as an object or not in a given implementation is up to you. > But what about SVGStringList, which uses DOMString as it's object > (but DOMString is not defined as an interface, but as > sequence<unsigned short>, which would map to BSTR in C++ COM. Yes, that would be a good mapping. DOM Strings are sequences of 16bit quantities, which (for basic multilingual plane characters) are characters and (for plane one to 16 characters) are one of a surrogate pair. > And there's also no way to create a DOMString, how can you then > implement SVGStringList correctly, maybe there should be a SVGString > that wraps a DOMString ? All that is required is that the List behaviour is correct. If the implementation wraps objects or uses pointers, that is an implementation choice. -- Chris mailto:chris@w3.org
Received on Wednesday, 13 November 2002 05:45:52 UTC