| 00:19 | <jarek> | Hi |
| 00:21 | <jarek> | Are there any plans to allow real physical units or density-independent units (like Android's dp) when specifying dimensions? |
| 00:23 | <jarek> | It looks like there is currently no easy way to e.g. define button height to match the average human pointing finger size |
| 05:12 | <frivoal> | Does respec have an equivalent of this bikeshed feature: https://tabatkins.github.io/bikeshed/#autolink-shortcuts |
| 05:13 | <frivoal> | (automatic links to definitions form other specs, preferably being smart about the type of things you're trying to link to) |
| 09:31 | <zcorpan> | botie: tell jarek https://discourse.wicg.io/t/one-inch-is-not-an-inch/1228 |
| 09:31 | <botie> | will do |
| 11:01 | jgraham | realises that the argument isn't called "behaviour" |
| 12:05 | <smaug> | jgraham: has there been any talk to have cross-browser tests for IME handling or how to even write such tests? |
| 12:08 | <smaug> | though, that is very platform specific |
| 12:08 | <jgraham> | smaug: There was a discussion at TPAC for mostly unrelated reasons. I think the consensus was some kind of fake test-only IME with known behaviour |
| 12:09 | <jgraham> | Because interacting with the actual platform IME isn't a stable target (not just between platforms but also because it might e.g. learn which characters you type and move them about in the display) |
| 12:09 | <smaug> | that too |
| 12:09 | <smaug> | and different IMEs change over time |
| 12:09 | <jgraham> | (apparently Apple had spellchecker tests fail because the system learnt to autocorrect the misspellings in the tests :) |
| 14:00 | <smaug> | annevk: does any spec define what should happen to open network connections when a window object from which the connection is created isn't anymore the current window. (say new somewindow.EventSource() or new somewindow.WebSocket() and then somewindow becomes non-current) |
| 14:04 | <jgraham> | On a not-intentionally related, but similar, note, the fact that https://pastebin.mozilla.org/8929276 "works" in Chrome is a bug, right/ |
| 14:04 | <smaug> | right |
| 14:27 | <annevk> | smaug: what is "non-current"? |
| 14:28 | <annevk> | smaug: the section in HTML titled "Aborting a document load" does go into this a bit, but not as detailed as I want |
| 14:29 | <annevk> | smaug: it's related to lifetime of these objects not being well-defined in general |
| 16:03 | <smaug> | annevk: non-current is the window which isn't attached to the currently active document |
| 16:03 | <smaug> | another question, in the specs, Notes normative? |
| 16:04 | <annevk> | smaug: notes are never normative |
| 16:05 | <smaug> | that is what I thought |
| 16:17 | <smaug> | igrigorik: where does the spec say https://bugs.chromium.org/p/chromium/issues/detail?id=662697#c2 ? |
| 16:31 | <Ms2ger> | \o farre |
| 16:36 | <farre> | Ms2ger: hi |
| 16:36 | <botie> | privet, farre |
| 18:00 | <smaug> | BroadcastChannel's constructor shouldn't ever throw, right https://html.spec.whatwg.org/multipage/comms.html#broadcasting-to-other-browsing-contexts |
| 18:25 | <TabAtkins> | botie, tell frivoal No, respec doesn't have cross-spec linking (and can't, without XHRing a multi-MB file, or hitting up a web service every time someone looks at a spec). One of the many reasons nobody should be using it. ^_^ |
| 18:25 | <botie> | will do |
| 18:47 | <annevk> | smaug: it can throw in the binding layer |
| 18:47 | <annevk> | smaug: it doesn't throw in the specification layer since it doesn't say throw anywhere in the description |
| 18:58 | <smaug> | annevk: I was thinking about the origin |
| 18:58 | <smaug> | when one is using data: url |
| 18:58 | <smaug> | or such |
| 18:59 | <annevk> | smaug: I see, would still work in theory seems like |
| 18:59 | <smaug> | yeah, I think it should work within such data- url page |
| 18:59 | <smaug> | but doesn't seem to in blink nor in gecko. But there really isn't any reason for such inconsistency, so I filed a Gecko bug |
| 19:00 | <annevk> | 👍 |
| 19:00 | <smaug> | ...and I assume baku will fix it :) |
| 23:46 | <botie> | frivoal, at 2016-11-18 18:25 UTC, TabAtkins said: No, respec doesn't have cross-spec linking (and can't, without XHRing a multi-MB file, or hitting up a web service every time someone looks at a spec). One of the many reasons nobody should be using it. ^_^ |