16832 2012-04-24 07:56:31 +0000 Don't get the length property off platform objects with indexed properties 2012-06-20 05:36:46 +0000 1 1 1 Unclassified WebAppsWG WebIDL unspecified All All RESOLVED FIXED P2 normal --- 1 Ms2ger cam bzbarsky mike public-script-coord zcorpan public-webapps-bugzilla oldest_to_newest 66959 0 Ms2ger 2012-04-24 07:56:31 +0000 See discussion starting at <http://krijnhoetmer.nl/irc-logs/developers/20120424#l-1470>. In particular, those objects need not have a length property 66961 1 zcorpan 2012-04-24 08:24:06 +0000 http://krijnhoetmer.nl/irc-logs/whatwg/20120424#l-372 69271 2 cam 2012-06-20 04:20:08 +0000 OK, so two things to fix here: 1. We shouldn't look up a "length" property on a platform object that has indexed properties when we want to know how many there are, since that property from the prototype could be changed. Instead, we should just know internally how many indexed properties there are. 2. We should always expose a "length" property on objects that have indexed properties, and not have to define an IDL attribute for it. If we do this, then we should require indexed properties not to be sparse (which I think they currently can be). Boris mentioned that it might be better to have "length" exposed as a data property rather than an accessor property (and if we do that, then it'll need to be on the object itself rather than on the prototype). Is anyone opposed to doing that? 69272 3 bzbarsky 2012-06-20 04:30:19 +0000 I'm no longer entirely sure it's a good idea, for what it's worth... 69275 4 cam 2012-06-20 05:18:18 +0000 One disadvantage to having "length" automatically be on the prototype is that there's no way in the IDL to indicate whether this is writable. It's clear if you have to explicitly include: attribute unsigned long length; 69276 5 cam 2012-06-20 05:36:46 +0000 Oh, #2 is bug 16833 actually. I did #1: http://dev.w3.org/cvsweb/2006/webapi/WebIDL/Overview.xml.diff?r1=1.530;r2=1.531;f=h so I'll mark this bug fixed.