| 00:01 | <Hixie> | oh wow, this is awesome |
| 00:01 | <Hixie> | the aria thing actually breaks h1-h6 in ATs |
| 00:02 | <Hixie> | (it removes their nesting depth) |
| 00:02 | Hixie | cries a little inside |
| 00:10 | <michaeln> | LOL at programming-motherfucker :) |
| 00:13 | <Hixie> | (oh wow, it also removes role=heading for h1-h6) |
| 00:13 | <TabAtkins> | Wut. |
| 00:17 | <hober> | that makes no sense |
| 17:13 | <zewt> | what's the reasoning behind canvas.getContext returning null on unsupported context types instead of throwing an exception? it's lead to the terrible "webglcontextcreationerror" contortion in webgl, since there's no exception to report error details in |
| 17:14 | <zewt> | seems very unlike the rest of the platform |
| 17:15 | <Philip`> | zewt: I think Firefox still throws an exception, so you could probably argue that the spec should be changed to throw and it won't break compatibility |
| 17:16 | Philip` | doesn't remember seeing a reason for returning null, except that it's easier to catch than an exception |
| 17:16 | <zewt> | canvas.getContext("garbage") returns null |
| 17:17 | <zewt> | the important case is when a type is supported, but can't be created for some reason--in that case you want to be able to say why |
| 17:18 | <zewt> | i don't know if error handling in "return a new object for contextId" is actually defined by the spec |
| 18:12 | <zewt> | webgl using the phrase "Later, at the normal event delivery time" makes me sad |