| 11:05 | <annevk> | TabAtkins: is there a reason for not marking up definitions of IDL members with <code>? |
| 12:04 | <GPHemsley> | annevk_: Is there a term for host + port? |
| 12:13 | <jgraham> | http://tantek.com/2011/238/b1/many-ways-slice-url-name-pieces |
| 12:27 | <yhirano_> | annevk: ping |
| 12:47 | <yhirano_> | annevk_: ping |
| 13:38 | <annevk_> | yhirano_: hi |
| 13:38 | <annevk> | GPHemsley: not in the spec |
| 13:48 | <annevk> | To this day I have trouble spelling Docs, thanks Opera |
| 13:48 | <annevk> | Or I guess, thanks Stig |
| 14:11 | <TabAtkins> | annevk: Since I wrap IDL *links* in <code>, no, it just looks like an oversight on my part. File an issue on me. |
| 14:31 | <annevk> | TabAtkins: do you think we should have a way to do it automatically? |
| 14:31 | <annevk> | TabAtkins: e.g. perhaps <dfn>{{Node/baseURI}}</dfn> should be a thing |
| 14:31 | <TabAtkins> | Yeah, makes sense to me. I can just check that it's not already in a <code> and wrap it. |
| 14:31 | <annevk> | TabAtkins: or that could be {{{Node/baseURI}}} :p |
| 14:32 | <TabAtkins> | Oh, yeah, there's an issue on me to figure that out, too. |
| 14:32 | <annevk> | I guess I can wait for that |
| 14:32 | <TabAtkins> | Why wait? Just file the issue. Wrapping a <code> is easier than figuring out what syntax I want for definitions. |
| 14:36 | <annevk> | done |
| 14:38 | <TabAtkins> | Thanks. ^_^ |
| 15:15 | <yhirano_> | annevk: I'd like to talk a bit about bodyUsed. |
| 15:16 | <yhirano_> | annevk: let res be a non-empty response. |
| 15:16 | <yhirano_> | annevk: var stream = res.body; assert_false(res.bodyUsed); is that OK? |
| 15:19 | <yhirano_> | oh, anne is away. |
| 15:27 | <Domenic> | timezones :( |
| 15:33 | <TabAtkins> | circular earth :( |
| 15:34 | <Ms2ger> | It's not like it's better on Discworld |
| 15:34 | <jgraham> | TabAtkins: Feeling flat today? |
| 15:35 | <TabAtkins> | I'm just saying, if there sun shined from a skybox at infinite distance, we wouldn't have these problems. |
| 15:38 | <jgraham> | I don't know what a skybox is, but a sun at infinite distance would have problems of its own |
| 15:39 | <TabAtkins> | Nah, you just have light move at infinite speed. Easy. |
| 15:40 | <annevk> | "Getting DOM to REC: last interop issues" lol |
| 15:41 | <TabAtkins> | A skybox is a programming technique. Easiest to talk about with raytracing - whenever a ray fails to hit anything and just goes off toward infinity, the skybox defines the color of that point-at-infinity. |
| 15:41 | <Ms2ger> | Ha |
| 15:42 | <annevk> | yhirano_: sorry, we've been having some connection issues |
| 15:43 | <annevk> | yhirano_: wanderview will be back from vacation next week or so I'm told and hopefully he can sort that through |
| 15:43 | <annevk> | yhirano_: I have to say I'm not super committed to any particular model, as long as it makes sense... |
| 15:43 | <Domenic> | I too have stopped having strong feelings about bodyUsed |
| 15:44 | <yhirano_> | ok, thanks. |
| 15:51 | <annevk> | yhirano_: I agree with you that if body is null, it would make sense for text() et al to not set it and always succeed |
| 15:51 | <annevk> | yhirano_: note that body being null is different from it being empty |
| 15:51 | <annevk> | (I said as much on the issue) |
| 15:57 | <Ms2ger> | TabAtkins, we'll be expecting you in #rust :) |
| 16:02 | <yhirano_> | annevk: interesting. I don't understand the difference between null body and empty body (and I merged them in my draft). can you tell me? |
| 16:02 | <TabAtkins> | Ms2ger: Get SimonSapin_ to pull the HTML parser + selector engine out of Servo, and we'll talk. |
| 16:02 | <jgraham> | TabAtkins: Done |
| 16:03 | <Ms2ger> | TabAtkins, that is done indeed |
| 16:03 | <SimonSapin> | kind of |
| 16:03 | <TabAtkins> | WHY DIDN'T SIMON TELL ME |
| 16:03 | <TabAtkins> | oh ok |
| 16:03 | <TabAtkins> | phew |
| 16:03 | <Ms2ger> | Well, they're not integrated |
| 16:03 | <SimonSapin> | We still need a nice tree API and plugging everything together |
| 16:03 | <Ms2ger> | But html5ever and rust-selectors exist |
| 16:03 | <jgraham> | https://github.com/kmcallister/html5ever https://github.com/SimonSapin/cssselect/issues/48 |
| 16:03 | <jgraham> | Er, not issue 48 specifically |
| 16:03 | <TabAtkins> | Bikeshed is just an HTML parser, a selector engine, and then some other random code. |
| 16:03 | <jgraham> | Silly search |
| 16:04 | <gsnedders> | TabAtkins: it's like why Anolis is so slow — looping over stuff is just too slow, silly Python |
| 16:04 | <jgraham> | Oh, not that project al all even |
| 16:04 | <jgraham> | Anyway, it's somewhere |
| 16:05 | <SimonSapin> | I have some experiments at https://github.com/SimonSapin/kuchiki , but it’d need someone to take the lead and make a proper library and maintain it |
| 16:08 | <SimonSapin> | TabAtkins: wanna make a tree library in Rust? :) |
| 16:08 | <TabAtkins> | Probably not! |
| 16:09 | <gsnedders> | Totally unrelated: is there any serialisation format that can serialise all DOMs? Like, can serialise both <!-- -- --> and stuff with arbitrary namespaces? |
| 16:10 | <SimonSapin> | Probably, somewhere in the infinity of all possible serialization formats |
| 16:15 | <TabAtkins> | Actually, a tree library sounds like an interesting learning project, with all the ownership semantics. Maybe. |
| 16:16 | <jgraham> | Certainly you are going to learn all the hard things rather fast if you make a tree with parent references |
| 16:17 | <gsnedders> | I tried doing that once. Then decided I was getting in too deep. |
| 16:19 | <TabAtkins> | Well, I know Bikeshed's tree-library needs, and they include walking up the ancestor list. |
| 16:19 | <annevk> | yhirano_: the difference is what happens with Content-Length |
| 16:19 | <annevk> | yhirano_: mostly |
| 16:19 | <annevk> | yhirano_: if there's no body, there's no Content-Length |
| 16:21 | <annevk> | gsnedders: didn't we develop a format like that once? |
| 16:21 | <yhirano_> | annevk: thanks. I will fix the draft. |
| 16:21 | <annevk> | gsnedders: for testing? |
| 16:23 | <TabAtkins> | annevk: Opinion - if SVG explicitly embraces a looser format for standalone, should we just use HTML (special-case when the first tag is <svg> to make it an SVG document rather than HTML) or use your XML5 or something similar? |
| 16:23 | <TabAtkins> | We dont' actually give a shit about namespaces. |
| 16:23 | <annevk> | TabAtkins: my opinion is that if we keep XML in browsers we should make it "XML5" |
| 16:24 | <annevk> | TabAtkins: and we can make namespaces nice |
| 16:24 | <Ms2ger> | There's even a rust implementation with tests |
| 16:24 | <annevk> | TabAtkins: e.g. <svg xmlns=svg> or some such |
| 16:24 | <annevk> | that autoexpands to the actual namespace |
| 16:24 | <TabAtkins> | Is XML5 as easy to use as svg-in-html? Similar affordances? |
| 16:24 | <gsnedders> | annevk: can't quite deal with arbtirary namespaces |
| 16:25 | <annevk> | TabAtkins: yeah, you can even use </> to close an open tag |
| 16:25 | <annevk> | gsnedders: XML5 can |
| 16:25 | <Ms2ger> | What is this, sgml? |
| 16:25 | <TabAtkins> | Nah, that's just use / |
| 16:25 | <annevk> | Ms2ger: you're too young to be making that joke :p |
| 16:25 | <TabAtkins> | s/that'd/ |
| 16:25 | <Ms2ger> | Get off my lawn |
| 16:26 | <Ms2ger> | And onto that beautiful mountain of yours |
| 16:26 | <gsnedders> | TabAtkins: SGML allows both depending on configurations for the application; HTML<5 per spec allows both NET (a/bar/) and whatever-it's-called-syntax (<a>foo</>) |
| 16:26 | <annevk> | I was about to say, get off my mountain |
| 16:26 | <gsnedders> | Wait, I'm younger than Ms2ger. Shit. |
| 16:26 | <TabAtkins> | Hahaha, all you babies! |
| 16:27 | TabAtkins | studiously hides that he's not 30 yet either. |
| 16:27 | <Ms2ger> | gsnedders, :D |
| 16:27 | gsnedders | hopes he isn't still the youngest guy active in this IRC channel |
| 16:27 | <gsnedders> | Because, like, I've been here almost a decade. |
| 16:28 | <gsnedders> | And that just leads to disturbing questions |
| 16:28 | <jgraham> | Well, if you will kill off all the newbies |
| 16:28 | <gsnedders> | Hey, I never killed anyone! |
| 16:56 | <caitp-> | ...but (and here's the key shock), to ever have a hope of doing that, we need to accept and embrace: we have to break the old broken web. <standing ovation from browser vendors and users who shed happy tears upon hearing this speech> |
| 16:58 | <TabAtkins> | caitp-: Why would *users* shed happy tears? They're the whole reason we can't break the old web in the first place. |
| 16:58 | <TabAtkins> | Authors don't care; they've abandoned the pages we want to break. |
| 17:00 | <caitp-> | oh, it's sarcasm |
| 17:00 | <TabAtkins> | Oh, ok. You quoting that from somewhere? |
| 17:01 | <caitp-> | kyle simpson |
| 17:07 | <TabAtkins> | Ah, getify. Ok. |
| 17:34 | <TabAtkins> | I can't tell if this "You update XMLHttpRequest has a bug" thread is spambot, or just a very confused and non-native English speaker. |
| 18:54 | <Ms2ger> | TabAtkins, I guessed "confused" :) |
| 19:00 | <caitp-> | but it makes his web music lose voice |
| 19:00 | <caitp-> | that's just terrible |