| 06:51 | <hsivonen> | annevk: if I put the fonts on resources.whatwg.org, should I use * or https://encoding.spec.whatwg.org in Access-Control-Allow-Origin? |
| 06:52 | <hsivonen> | annevk: I'm guessing the server doesn't support dynamic headers with rules like "any .whatwg.org" |
| 06:55 | <hsivonen> | hmm. is there a particular reason for the WHATWG deployment scripts to use Oracle Java instead of OpenJDK? |
| 08:15 | <MikeSmith> | hsivonen: if you mean the part that calls the HTML checker, no reason to use Oracle other than I knew the package name to get it to set the Oracle java as the default |
| 08:16 | <hsivonen> | MikeSmith: I see |
| 08:16 | <MikeSmith> | but I can figure out the OpenJDK equivalent |
| 08:17 | <hsivonen> | I didn't mean to suggest that there was a need to use OpenJDK. I was just curious. |
| 08:18 | <MikeSmith> | ok |
| 08:21 | <annevk> | hsivonen: just * I suppose |
| 08:22 | <annevk> | hsivonen: it's Apache, we could run a script, up to you |
| 08:25 | <zcorpan> | should module scripts only be available in secure contexts? |
| 08:25 | <annevk> | lol |
| 08:27 | <hsivonen> | annevk: ok |
| 08:29 | <hsivonen> | one would hope that Apache had a mod_cors by now |
| 08:30 | <ondras> | that would be cool |
| 08:31 | <zcorpan> | annevk: ? |
| 08:32 | <annevk> | zcorpan: just that CORS is already super controversial |
| 08:34 | <zcorpan> | annevk: heh yes. we had a discussion last week about requiring secure context for worklets, which had some opposition... but it seems to me the arguments against will be moot in 1-2 years |
| 08:47 | <annevk> | zcorpan: just do it |
| 08:48 | <annevk> | zcorpan: my main fault here is getting involved in the weekend |
| 08:48 | <annevk> | zcorpan: anxiety levels have steadily been rising, though hike through the snow this morning helped |
| 09:42 | <annevk> | hsivonen: even if Apache did, DreamHost having it installed would be a whole other thing |
| 09:47 | <hsivonen> | annevk: I'm quite hesitant to represent the visual information twice visually. It's cluttered enough as-is. I think I'll go with aria-label. |
| 09:48 | <hsivonen> | annevk: I'm now using a color scheme that's supposed to work for different forms of color blindness. |
| 09:55 | <annevk> | hsivonen: sweet |
| 10:43 | <annevk> | hsivonen: it's rather nice to have a compelling ARIA example |
| 11:18 | <hsivonen> | annevk: unclear how compelling this is as an ARIA example. It's unclear to me if the tables are actually useful for screen reader users. |
| 11:18 | <annevk> | Fair, I guess we still need to know if it actually works |
| 11:18 | <hsivonen> | annevk: so if the HTML files are generated by Travis, I guess we don't need a subdirectory to avoid directory clutter and could instead go for shorter URLs |
| 11:19 | <annevk> | hsivonen: okay |
| 11:19 | <annevk> | I'm happy with flat |
| 11:19 | <hsivonen> | annevk: ok. I submitted a PR with the fonts |
| 12:15 | <annevk> | hsivonen: https://bugzilla.mozilla.org/show_bug.cgi?id=1331351 was filed on top-level data URLs |
| 12:26 | <hsivonen> | annevk: nice |
| 12:27 | <hsivonen> | annevk: thanks for merging the fonts |
| 12:31 | <hsivonen> | annevk: how do I tell bikeshed to <link> to an extra css file? |
| 12:33 | <annevk> | hsivonen: not sure |
| 12:34 | <annevk> | hsivonen: why does the main resource need to link to it? |
| 12:35 | <hsivonen> | annevk: to reuse the same CSS for the legend colors so that they don't get out of sync |
| 12:38 | <hsivonen> | Using <style>@import for now, but that's sad |
| 12:40 | <annevk> | hsivonen: hopefully TabAtkins can help with that |
| 12:41 | <annevk> | I don't see anything in https://tabatkins.github.io/bikeshed/ that suggests it's possible |
| 12:53 | <tobie> | annevk: should "abort these steps" and "abort these sub-steps" end-up in the Infra standard? |
| 12:54 | <tobie> | annevk: the first one is obvious, the latter isn't super explicit as to whether control is returned to the super-steps or to the caller. |
| 12:55 | <tobie> | annevk: actually, "abort these steps" used in a sub-step isn't super explicit either. |
| 12:59 | <tobie> | annevk: should I file an issue? |
| 12:59 | <annevk> | tobie: please |
| 12:59 | <annevk> | tobie: or if you want to figure it out and provide a PR... |
| 12:59 | <annevk> | tobie: that's all part of the general English-algorithm-control flow question |
| 13:02 | <tobie> | annevk: no time for a PR this week. Sorry. :-/ |
| 13:02 | <annevk> | tobie: no worries, having it tracked is good |
| 14:50 | <annevk> | botie: thank zcorpan for the reviews |
| 14:50 | <botie> | annevk: what? |
| 14:51 | <annevk> | botie: tell zcorpan thanks for the reviews |
| 14:51 | <botie> | will do |
| 15:13 | <TabAtkins> | hsivonen: Just put the <link> in. Bikeshed will shift it into the head for you automatically. |
| 15:14 | <botie> | zcorpan, at 2017-01-16 14:51 UTC, annevk said: thanks for the reviews |
| 15:28 | <annevk> | hsivonen: I'll probably review your second PR tomorrow |
| 15:54 | <jgraham> | Manishearth: non ascii in irssi you say? That sounds like deep magic. |
| 16:46 | <Manishearth> | jgraham: it sometimes works |
| 16:46 | <Manishearth> | old setup used to work. now it doesn't again |
| 17:50 | <MikeSmith> | jgraham:π΄ π π¦ |
| 17:51 | <MikeSmith> | irssi |
| 17:55 | <jgraham> | MikeSmith: Three question marks in diamonds? |
| 18:03 | <MikeSmith> | jgraham: https://gist.githubusercontent.com/sideshowbarker/8284404/raw/2c2bc535919e64fe2023ace0e451ea20f38a619d/irssi.png |
| 18:03 | <annevk> | jgraham: π |
| 18:05 | <jgraham> | MikeSmith: Seems implausible |
| 18:07 | <jgraham> | MikeSmith: Is that using screen? How did you get it working? |
| 18:09 | <MikeSmith> | tmux |
| 18:09 | <MikeSmith> | dunno if it works in screen |
| 18:11 | <MikeSmith> | I donβt think I have anything special set in my irssi config |
| 18:12 | <MikeSmith> | set -g default-terminal "screen-256color" |
| 18:12 | <MikeSmith> | in my tmux config |
| 18:14 | <jgraham> | Hmm, OK. nice to know that there is some setup that works |
| 18:15 | <jgraham> | Even if it's not the setup I have :( |
| 18:24 | <Domenic> | hsivonen: annevk: isn't ACAO: * set on a higher level .htaccess already? Or maybe that's only for *.spec.whatwg.org |
| 18:24 | <annevk> | Domenic: could be, yes, file an issue and I'll sort it out |
| 19:00 | <annevk> | Manishearth: so yeah, https://github.com/tc39/proposal-intl-segmenter does suggest grapheme clusters depend on language, which doesn't make them ideal as a unit |
| 19:00 | <annevk> | Manishearth: I wonder how Swift works around that then |
| 19:03 | <Manishearth> | annevk: weeeelll, UAX 29 EGCs do not depend on language |
| 19:03 | <Manishearth> | the spec explicitly says that you will want to tailor the algorithm to your use case |
| 19:04 | <Manishearth> | in case you need something more specific |
| 19:04 | <Manishearth> | annevk: swift implements UAX 29. period |
| 19:05 | <Manishearth> | annevk: so for example, consonant clusters in my language are 2 graphemes by UAX 29, even though they're logically one, because depending on the font they may render as two (there are equivalent ways of rendering them, fonts usually render as one). I put in a proposal to change that though. |
| 19:05 | <Manishearth> | word segmentation has similar issues, UAX 29 doesn't try to handle Thai |
| 19:06 | gsnedders | would love to see a spec for that |
| 19:06 | <Manishearth> | for thai? |
| 19:06 | <gsnedders> | y'know, with citations like [thai-dictionary] |
| 19:06 | <Manishearth> | yeah |
| 19:06 | <gsnedders> | so yeah, for Thai |
| 19:14 | <annevk> | Manishearth: makes me wonder what the JS proposal is doing |
| 19:15 | <annevk> | Something for tomorrow I suppose, thanks for the answer |
| 19:40 | <Manishearth> | annevk: they're trying to support all the tailorings |
| 19:40 | <Manishearth> | an interesting idea |
| 19:40 | <Manishearth> | this will not end well |
| 19:40 | <Manishearth> | but an interesting idea |
| 19:40 | <Manishearth> | something I wish we had good libraries for. but it will be hell making those libraries |
| 19:50 | <annevk> | Manishearth: maybe you can give them feedback |
| 19:51 | <annevk> | Manishearth: happy to help if that gets difficult |
| 20:16 | <Manishearth> | annevk: "graphemes" isn't the right word to use there |
| 20:16 | <Manishearth> | for granularity |
| 20:27 | Manishearth | will file an issue |
| 20:32 | <annevk> | π |
| 20:32 | <Manishearth> | yes, unicode lego to you too |
| 20:32 | <Manishearth> | ;p |
| 21:01 | <annevk> | Unicode Lego would be great |
| 21:52 | <smaug____> | annevk: so URLSearchParams in browsers have ctor taking URLSearchParams as param |
| 21:52 | <smaug____> | was that at some point in the spec |
| 21:52 | <smaug____> | and is it not needed if URLSearchParams' stringifier is implemented? |
| 21:53 | <smaug____> | annevk: feel free to answer in https://bugzilla.mozilla.org/show_bug.cgi?id=1330678 |
| 22:37 | <tobie> | Is there a shortcut to say that an interface must inherit from A or from an interface that itself inherits from A? |
| 22:39 | <tobie> | C must be one of A's inherited interfaces? |
| 22:40 | <tobie> | I find that painful to read. |
| 22:46 | <tobie> | Obviously, it's the minute I right this down here that I realize I've been reading the definition of inherited interfaces incorrectly all along. :-/ |
| 22:46 | <tobie> | * I write |
| 22:46 | tobie | needs to hit the sack, apparently. |
