On Fri, 29 Jan 2010, Dominique Hazael-Massieux wrote: > > I'm making progress on adding a layer of semantic checks to the WebIDL > checker that verifies some of the constraints defined in the WebIDL > specs; this includes verifying that the extended attributes are known, > the interfaces match defined names, etc. > > So far, I still haven't found any real problem in the HTML5 spec, > although the (not-online-yet) checker raises a few false positives: > • for the unknown Supplemental and ReplaceableNamedProperties extended > attributes (that aren't defined in WebIDL yet) > • for the references to WindowProxy type which isn't defined anywhere > > I know you've filed a request to add Supplemental to WebIDL; I'm not > sure if you have done so for ReplaceableNamedProperties; in both cases, > it would probably useful to document minimally the constraints > (accepting arguments or not, applicable to interfaces only or exception > as well, ...) you have in mind for these attributes to > public-script-coord?
[ReplaceableNamedProperties]: http://www.w3.org/Bugs/Public/show_bug.cgi?id=8241 Both are for interfaces only, both accept no arguments. > For WindowProxy, I've seen that the spec says that WindowProxy is kind > of magical, which I assumes explains why "interface WindowProxy" appears > nowhere; I guess I could hard-wire the checker with knowledge about this > mythical interface, but I was wondering if there wasn't a way to still > declare the interface somewhere in the spec somehow? I've added an IDL block that defines it to the version of the spec I send only to the validators: http://www.whatwg.org/specs/web-apps/current-work/.last-version/source-validation -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
