| 11:44 | <annevk> | emilio: so what's the viewport of new Document()? The viewport of its relevant global object or some such? |
| 11:47 | <emilio> | annevk: I think viewports of documents which have no rendering are generally 0x0 |
| 11:47 | <emilio> | annevk: display: none iframes are another example |
| 11:48 | <annevk> | I see what you mean with the matchMedia() remark now. Okay, so code exists, unclear if it's defined |
| 11:49 | <annevk> | And abstracting/duplicating that logic for workers is not something you see the value in |
| 11:50 | <annevk> | Seems we might need something for letterSpacing too |
| 11:51 | <annevk> | And wordSpacing |
| 11:52 | <emilio> | yeah, I looked into it and it was rather annoying, in the sense that you either get all the viewport sizes upfront and pass them around (which slows down absolute length resolution), or make a bunch of that stuff virtual, which is also not-great, and adds another abstraction layer just for this |
| 11:52 | <emilio> | (same for stuff like font-metrics, inherited font family, etc) |
| 11:52 | <emilio> | So it is either a bunch of stuff to pack upfront, or a silly abstraction layer to have (imo) |
| 11:53 | <emilio> | Specially since it's not like it does something useful, like on the main-thread case |
| 11:53 | <emilio> | It's just fixing a 0x0 viewport with a 10px sans font |
| 11:53 | <annevk> | Yeah, even if it makes some code more portable it doesn't seem entirely likely that you'd always expect a 10px font or a 0x0 viewport for that matter |
| 11:54 | <annevk> | And even if there was some magical future with elements in workers they wouldn't become reference points for OffscreenCanvas |
| 15:49 | <annevk> | Media elements are so weird |
| 15:50 | <annevk> | Changing state synchronously, but then dispatching the corresponding events in different tasks |
| 16:03 | <annevk> | https://github.com/whatwg/html/pull/7855 |
| 20:06 | <smaug> | What non-empty <SomeDictionary> might mean? |
| 20:06 | smaug | is trying to interpret https://w3c.github.io/user-timing/#measure-method step 1 |
| 20:07 | <smaug> | Yoav Weiss: you might know ^ |
| 21:38 | <TabAtkins> | I presume what's intended is "the object doesn't contain any of the defined dictionary keys" |
| 21:39 | <TabAtkins> | I don't understand why that spec text is adding that condition, tho, since all the text under that step just explicitly tests for key presence. |
| 21:39 | <TabAtkins> | So I suspect in context the meaning is "mu" |