| 06:01 | <annevk> | MikeSmith: someone downvoted your SO question... |
| 06:01 | annevk | added an upvote to put it back at 0 |
| 06:15 | <MikeSmith> | hahah |
| 06:15 | <MikeSmith> | ah, my SO reputation has increaed now! I guess that was you |
| 06:15 | <MikeSmith> | (high five) |
| 06:16 | <MikeSmith> | I may put a "bounty" on that question |
| 06:16 | <MikeSmith> | never done that on SO before |
| 06:16 | <MikeSmith> | I think the way it works, I have to "pay" some points, but I kinda care fuck all how much points I have, so that would still be all win |
| 06:20 | <annevk> | MikeSmith: I think some of those index issues are mentioned in either wattsi or html-build |
| 06:28 | <MikeSmith> | annevk: ah ok |
| 06:28 | <MikeSmith> | I will take a look at that |
| 06:28 | <MikeSmith> | there is some really clever stuff in those tools |
| 06:29 | <MikeSmith> | I mean Hixie's spec-production tools |
| 06:29 | <MikeSmith> | others could learn a lot from them *cough*Bikeshed*cought* |
| 06:30 | <MikeSmith> | like, that thing I accidentally discovered yesterday, which Hixie has in there to look for places where the source has "a" where it should have "an" (and vice versa) |
| 06:32 | <annevk> | Kind of cool that it discovered that even for stuff wrapped in markup |
| 06:41 | <MikeSmith> | indeed |
| 06:41 | <MikeSmith> | Hixie++ |
| 06:41 | <MikeSmith> | annevk: btw, speaking of SO bounties, http://stackoverflow.com/questions/32137010/how-to-allow-webworker-to-call-https-url/32296750?noredirect=1#32137010 |
| 06:42 | <MikeSmith> | or just http://stackoverflow.com/questions/32137010/how-to-allow-webworker-to-call-https-url/32296750 |
| 06:43 | <MikeSmith> | dunno if there's a good answer to that question but at least it seems like that OP hasn't gotten a good one yet |
| 06:44 | annevk | discovers https://www.w3.org/Bugs/Public/show_bug.cgi?id=27869 |
| 06:44 | <annevk> | The IETF actually changed the WebSocket specification to be incompatible with the API |
| 06:45 | <MikeSmith> | eh? |
| 06:45 | MikeSmith | reads |
| 06:46 | <MikeSmith> | "Could we just have the IETF spec fixed instead?" |
| 06:46 | <MikeSmith> | indeed |
| 06:46 | <annevk> | There's no way to actively invoke the IETF algorithm with a set of headers even |
| 06:47 | <MikeSmith> | I thought they had quit touching that stuff a long time ago |
| 06:47 | <MikeSmith> | somebody has still been fiddling with the protocol spec? |
| 06:47 | <MikeSmith> | kinda ironic |
| 06:49 | <MikeSmith> | given that most IETF specs are written and then never changed, and when you *want* somebody to make a change to an IETF spec, the responsible parties usually just shrug their shoulders and explain why it's not gonna happen |
| 06:50 | <MikeSmith> | hmm but looking at the link that Takeshi cites, that change was made a long time ago https://tools.ietf.org/html/draft-ietf-hybi-thewebsocketprotocol-10 |
| 06:50 | <annevk> | MikeSmith: that SO question is a bit vague, might just be that Safari has a bug |
| 06:50 | <MikeSmith> | I had thought he was saying they changed it after it went to RFC |
| 06:51 | <MikeSmith> | annevk: OK, well if it's not clear to what the problem is, then I reckon it's not a general problem |
| 06:51 | <MikeSmith> | maybe there's a Safari bug open for it |
| 06:52 | <MikeSmith> | hmm, yeah, now that I think of it, I have seen reports of other weird https-related issues that only seem to be reproducible in Safari |
| 07:01 | <annevk> | "Your message to hybi awaits moderator approval" |
| 07:01 | <annevk> | ugh |
| 07:44 | <annevk> | grep -ni 'and/or' source | perl -lpe 'print "\nOccurrences of making Ms2ger unhappy and/or annoyed:" if $. == 1' |
| 07:44 | <Ms2ger> | Ha |
| 07:48 | <mkwst> | annevk, MikeSmith: Thanks for going through the nonce patch with me. :) |
| 07:48 | <mkwst> | I'll have more for you, I'm sure. |
| 07:48 | <Ms2ger> | mkwst, wanna merge CSP into Fetch? |
| 07:49 | <mkwst> | Ms2ger: I want to give Fetch some hooks into CSP. Like http://mikewest.github.io/webappsec/specs/content-security-policy/#algorithms-fetch |
| 07:50 | <mkwst> | But CSP isn't exclusive to Fetch. See the `nonce` thing I just mentioned. :) |
| 07:52 | <MikeSmith> | mkwst: thanks for taking time to write up that patch. It hits the sweet spot (as far as having something that provides enough to make it wortwhile to add, without waiting to have anything at all in the HTML spec until the other bits (from your TODO editorial note) are also added) |
| 07:53 | <mkwst> | MikeSmith: I want to get rid of my monkey patches. *shrug* |
| 07:53 | <MikeSmith> | yeah |
| 07:53 | <MikeSmith> | anything that first at least just gets rid of monkey patch is a win |
| 07:54 | <mkwst> | MikeSmith: Of course, that means I need to get the patches into W3C's HTML as well. |
| 07:54 | <mkwst> | MikeSmith: Which looks like it's going to be a huge mess, honestly. :) |
| 07:54 | <gsnedders> | Is anyone planning on getting all the ruby stuff everyone implements into HTML? |
| 07:55 | <MikeSmith> | mkwst: also fwiw (and not to make you blush) I also admire your sense of judgment as an editor. You have a somewhat rare combination of being highly prolific while at the same time consistently hitting the right targets |
| 07:56 | <mkwst> | MikeSmith: Aw shucks... Thank you. :) |
| 07:56 | <MikeSmith> | heh |
| 07:56 | <mkwst> | MikeSmith: I'm working on too much at the moment. Not enough is actually getting done. :/ |
| 07:56 | <mkwst> | I'm good at starting things, but not yet good at finishing them. This credential management api quagmire, for example. |
| 07:57 | <MikeSmith> | mkwst: ah yeah the credential-management thing seems like something I'd like to stay way away from :) |
| 07:58 | <MikeSmith> | mkwst: but by my reckoning you and Brad have been handling that well (and patiently) |
| 07:58 | <annevk> | gsnedders: if you want to give it a go, it's yours |
| 07:58 | <mkwst> | MikeSmith: well, yes and no. It's taking too long for not enough benefit (either for me or the credentials CG). |
| 07:59 | <annevk> | mkwst: W3C HTML is not really a thing implementors use |
| 07:59 | <mkwst> | We're talking in circles, and I need to just suck it up and cut the knot. |
| 07:59 | <mkwst> | annevk: 1. I know. 2. It doesn't matter. |
| 07:59 | <gsnedders> | annevk: I don't massively want to. Rewriting all the spec text that berjon wrote doesn't seem worthwhile. It's just really awkward not having the parser as implemented defined anywhere. |
| 08:00 | <MikeSmith> | mkwst: about "Not enough is actually getting done." I guess it should be about quality and not quantity. One person can only do so much. We need more people writing security-related specs. More hands make light work, and all that. |
| 08:00 | <mkwst> | I only care about the W3C HTML spec insofar as I need to advance specs to REC due to patent idiocy. |
| 08:00 | <mkwst> | MikeSmith: Hey! I have some specs you could write! |
| 08:00 | <MikeSmith> | hahah |
| 08:00 | <MikeSmith> | mkwst: oh, speaking of that, I have something for you to please review |
| 08:00 | <MikeSmith> | not a spec, sadly |
| 08:00 | <mkwst> | sure |
| 08:00 | <MikeSmith> | but less work |
| 08:01 | <mkwst> | I like less work. |
| 08:01 | <MikeSmith> | hah. I would like to have a T-shirt that says that. |
| 08:02 | <MikeSmith> | mkwst: https://developer.mozilla.org/en-US/docs/Web/Security/Subresource_Integrity |
| 08:02 | <mkwst> | MikeSmith: Probably better to get one of the SRI editors to review that. It's the only spec I've been able to successfully delegate out. :) |
| 08:03 | <mkwst> | But I'll skim it. |
| 08:04 | <MikeSmith> | thanks yeah I was thinking about asking the editors, but... I've never communicated with any of them before. But anyway, I'm being lazy, I'll ping them too |
| 08:05 | <MikeSmith> | mkwst: when you skim it, feel free to just make any changes directly if you want. Or if you prefer just lemme know what if anything I should change (or add) |
| 08:05 | <mkwst> | Sure. I just sent Joel an email (CC'd you). He can loop in other folks if he wants. |
| 08:06 | <MikeSmith> | super |
| 08:06 | <MikeSmith> | thanks |
| 08:06 | <mkwst> | (Or you could just tweet at them, which is fairly low overhead. :) ) |
| 08:06 | <MikeSmith> | ah yeah that too |
| 08:07 | <mkwst> | Here: "@metromoxie @fmarier @frgx @freddyb: Hey! Mind taking a look at some SRI documentation I put on MDN? https://developer.mozilla.org/en-US/docs/Web/Security/Subresource_Integrity"; |
| 08:07 | <mkwst> | ;P |
| 08:07 | <MikeSmith> | gsnedders: about the ruby stuff maybe Robin can write up a patch himself. If there are issues with it as far not being up to snuff with the norms for the HTML spec itself, then people can make review comments and he can make fixes. |
| 08:08 | <gsnedders> | MikeSmith: yeah, probably more worthwhile |
| 08:08 | <gsnedders> | MikeSmith: it's just that I think all the arguments that have been made about it before don't stand when everyone implements it and the WHATWG spec is fiction. |
| 08:09 | <MikeSmith> | Robin has sorta been the one who cared most about getting it specced and implemented (I didn't) and he's the one who did the work of communicating with the implementors and seeing things through to getting the browser patches landed and all that |
| 08:09 | <gsnedders> | i know |
| 08:09 | <gsnedders> | but he isn't on IRC :P |
| 08:09 | <gsnedders> | and therefore isn't a real person |
| 08:10 | <MikeSmith> | gsnedders: I don't think anybody disagrees with that (what you said about the arguments). Not at this point at least. |
| 08:10 | <MikeSmith> | and it's far from being the worst thing that's ever been added to the platform. And some smart people in Japan really think it's good. |
| 08:11 | <MikeSmith> | Masuyuki (from Mozilla) tore into me pretty seriously when I criticized it |
| 08:12 | <annevk> | darobin hangs out here from time to time |
| 08:12 | <annevk> | gsnedders: it would help me if you could identify the patch against W3C HTML that added these elements and corresponding links into W3C HTML where these elements are defined |
| 08:13 | <annevk> | gsnedders: given that I might be able to give it a go soonish since the mismatches should be sort of a priority |
| 08:13 | <annevk> | gsnedders: I still need to get the build system running locally though |
| 08:14 | <MikeSmith> | gsnedders: s/Masuyuki/Masatoshi/ https://lists.mozilla.org/pipermail/dev-platform/2014-December/008136.html |
| 08:15 | <MikeSmith> | I think Robin is in another timezone now, btw |
| 08:16 | <MikeSmith> | annevk: I am working on a patch for the build system (or rather I'm really not yet but I should be and will try to get the patch written up today and PR'ed for review) |
| 08:20 | <annevk> | MikeSmith: for Mac, which of ftp://freepascal.stack.nl/pub/fpc/beta/3.0.0-rc1/i386-macosx/ should I get? |
| 08:21 | <annevk> | ftp://freepascal.stack.nl/pub/fpc/beta/3.0.0-rc1/i386-macosx/fpc-3.0.0rc1.intel-macosx.dmg ? |
| 08:22 | <annevk> | It seems Free Pascal will be one of the first to fall when mkwst removes ftp URL access |
| 08:23 | <MikeSmith> | annevk: not fpc-3.0.0rc1.intel-macosx.dmg |
| 08:24 | <MikeSmith> | oh |
| 08:24 | <MikeSmith> | no, actually, yeah, that one |
| 08:24 | <MikeSmith> | that's the same I grabbed for my macbook and it works |
| 08:25 | <MikeSmith> | hahahah :) > It seems Free Pascal will be one of the first to fall when mkwst removes ftp URL access |
| 08:26 | <annevk> | MikeSmith: so if we update the DEFINES in the build.sh it would fail on other platforms I guess? |
| 08:28 | <MikeSmith> | annevk: dunno |
| 08:29 | <MikeSmith> | you'd need to ask Hixie |
| 08:29 | <MikeSmith> | he's the one who told me it was Mac-specific |
| 08:29 | <annevk> | hmm I get complaints about Xcode |
| 08:29 | <MikeSmith> | you might have to update your XCode command-line tools |
| 08:29 | <MikeSmith> | gimme a minute to find the docs |
| 08:31 | <annevk> | Yeah, but Xcode claims I only have outstanding documentation updates |
| 08:31 | <MikeSmith> | annevk: just "xcode-select --install" I guess |
| 08:31 | <MikeSmith> | annevk: the command-line tools are separate thing, right? |
| 08:31 | <MikeSmith> | separate update |
| 08:31 | <annevk> | oh |
| 08:31 | <MikeSmith> | (why they do it that way I dunno) |
| 08:33 | <annevk> | MikeSmith: you said you made some changes to html-build too write that made things easier? |
| 08:34 | <MikeSmith> | I just made symlinks |
| 08:34 | <MikeSmith> | to everything in the html dir |
| 08:34 | <annevk> | MikeSmith: how do I go about that? |
| 08:34 | <MikeSmith> | ln ../html/source . |
| 08:35 | <MikeSmith> | in the html-build dir |
| 08:35 | <MikeSmith> | and same for other files |
| 08:35 | <MikeSmith> | or really if you want to do it quick and dirty, use wildcards |
| 08:35 | <MikeSmith> | ln ../html/* . |
| 08:35 | <MikeSmith> | and |
| 08:35 | <MikeSmith> | ln ../html/.* . |
| 08:37 | <MikeSmith> | or maybe just ../html/.multipage-404 is all you need (as far as the dot files) |
| 08:37 | <MikeSmith> | actually I guess you don't strictly even need that |
| 08:37 | <mkwst> | annevk: I won't be terribly sad. If you want to FTP, use the FTP app we'll make available in the Chrome Web Store. And lock youself into our platform forever. Muwahaha. /me destroys the web |
| 08:38 | <MikeSmith> | annevk: there's also a .htaccess file there but I don't think you need that (unless you want to test in Apache locally) |
| 08:40 | <annevk> | So the wattsi repo should maybe have a .gitignore file for bin/? |
| 08:46 | <annevk> | Hah, I don't have wget |
| 08:46 | <annevk> | Not sure if that was the only thing that made this fall apart but it's certainly part of it |
| 08:58 | <MikeSmith> | annevk: yeah probably there should be a .gitignore for bin/ |
| 08:59 | <annevk> | And html-build needs .gitignore .cldr-data |
| 08:59 | <MikeSmith> | yeah |
| 08:59 | <mkwst> | Can we simply integrate the three repositories? Is there a good reason for keeping the build scripts separate from the only source file that will ever use them? |
| 09:00 | <annevk> | We run the build scripts on the server |
| 09:00 | <annevk> | And wattsi is separate since it's fairly standalone |
| 09:00 | <mkwst> | But I assume you update the scripts from the repo before you run them, right? And you update the source file as well. Why call `git pull` twice when you could just call it once? |
| 09:00 | <MikeSmith> | yeah to that (about wattsi) |
| 09:00 | <mkwst> | wattsi as submodule? |
| 09:01 | <MikeSmith> | I hope we don't need to do submodules |
| 09:01 | <mkwst> | well, i hope we don't have to symlink or copy. :) |
| 09:01 | <annevk> | mkwst: file a bug against html-build? |
| 09:01 | <annevk> | mkwst: best to discuss it asynchronously since Domenic needs to partake |
| 09:01 | <mkwst> | oh.. is IRC not the bug-filing mechanism? :) |
| 09:01 | <MikeSmith> | for wattsi we should just make binaries available, really. I'm pretty sure that's what Domenic would like and what he wrote up somewhere already |
| 09:02 | <annevk> | mkwst: "depends" |
| 09:02 | <MikeSmith> | distribute three binaries: one linux binary, one mac binary, one windows |
| 09:03 | <annevk> | Yeah that would be nice |
| 09:03 | <MikeSmith> | mkwst: I do think we should consider moving the html-build contents to the html repo, at least |
| 09:04 | <MikeSmith> | yeah, distributing the binaries would not be hard to do. one person one time builds the mac binary, etc. |
| 09:04 | <MikeSmith> | it's not like we're going to be updating the wattsi sources all the time |
| 09:05 | <MikeSmith> | we could actually have the html-build just automatically download the appropriate wattsi binary |
| 09:06 | <MikeSmith> | the html-build already requires a network connection, and it downloads a shitload of... what? I don't know actually. (data from caniuse maybe?) |
| 09:07 | <mkwst> | MikeSmith: Yeah. It would be nice to separate "update dependencies" from "generate the spec". |
| 09:07 | <mkwst> | The build script would be better split into component steps, rather than the monolithic "do everything" it is today. |
| 09:10 | <mkwst> | https://github.com/whatwg/html-build/issues/5 |
| 09:10 | <annevk> | I think what Domenic wants to do is that when you submit a PR for "source", some script does the building for you and supplies debugging information on GitHub |
| 09:11 | <mkwst> | annevk: yeah. integrating with Travis seems like a clear win. |
| 09:11 | <annevk> | We should probably still simplify building locally though for if you want to make bigger changes |
| 09:11 | <MikeSmith> | yeah |
| 09:11 | <MikeSmith> | I'm sure that's the plan (full automatation via Travis) |
| 09:11 | <MikeSmith> | annevk: right |
| 09:11 | <MikeSmith> | we still need to be able to build locally as well |
| 09:12 | <MikeSmith> | you want to be able to check something before you push |
| 09:12 | <mkwst> | annevk: They go hand in hand, right? I don't think the current setup is going to be easy to get running on Travis in anything like a performant fashion. |
| 09:12 | <mkwst> | improving things for local builds is the same as improving things for Travis. |
| 09:13 | <MikeSmith> | aslo when reviewing, you want to be able to check the rendered output from code somebody's PR branch |
| 09:13 | <mkwst> | (as a side note, I'm so happy that this move has happened... submitting PRs is significantly simpler for me than sending an email into a black hole) |
| 09:14 | <MikeSmith> | hear hear |
| 09:14 | <darobin> | +1 |
| 09:15 | <MikeSmith> | hey darobin! |
| 09:15 | <MikeSmith> | darobin: gsnedders wants to talk with you about ruby (since for some reason he actually wants to have a spec to check tests against, or something) |
| 09:16 | <darobin> | that's madness |
| 09:16 | <MikeSmith> | heh |
| 09:16 | <darobin> | I'm going to be in and out of IRC this week; prepping last details for the trip and all |
| 09:16 | <gsnedders> | darobin: plz help merge the ruby stuff into the WHATWG spec |
| 09:16 | <darobin> | so email might work best |
| 09:17 | <MikeSmith> | darobin: well, I already volunteered you to (re)write a patch. so set all that other trip stuff aside for now |
| 09:17 | <darobin> | gsnedders: have you looked at the stuff and hit a bump, or are you just asking for the merge? |
| 09:17 | <darobin> | ah ok |
| 09:17 | <darobin> | I can do that |
| 09:17 | <darobin> | but it might have to wait until I'm in NY |
| 09:17 | <darobin> | lol |
| 09:17 | <MikeSmith> | darobin: http://krijnhoetmer.nl/irc-logs/whatwg/20150831#l-256 for the context |
| 09:17 | <gsnedders> | darobin: the big problem is the fact that it's W3C licensed so it can't just be merged into WHATWG, AFAIK |
| 09:18 | <MikeSmith> | darobin retains copyright on what he authored |
| 09:18 | <MikeSmith> | anyway, let's please not get into that, I think |
| 09:18 | <MikeSmith> | not right now |
| 09:18 | <darobin> | gsnedders: don't ask questions you don't want the response to |
| 09:18 | <MikeSmith> | heh |
| 09:18 | <darobin> | under French law I cannot cede my copyright to anyone, actually |
| 09:19 | <darobin> | so I can do whatever I want with the text I wrote |
| 09:19 | <MikeSmith> | you'll be in NYC |
| 09:19 | <MikeSmith> | you can get a gun! |
| 09:19 | <darobin> | hahaha |
| 09:19 | <darobin> | don't give me bad ideas :) |
| 09:19 | <MikeSmith> | "enforce" your copyright |
| 09:19 | <MikeSmith> | heh |
| 09:19 | <gsnedders> | yes, /you/ retain copyright, but the only license I can get what you wrote under is the one provided by the W3C |
| 09:19 | <darobin> | gsnedders: not to mention the fact that the ruby text was produced CC-BY |
| 09:20 | <MikeSmith> | darobin: seriously I don't think there's any rush on this. Except that after neglecting use for a long time, gsnedders is back and just wants everything his way from now on |
| 09:20 | <MikeSmith> | *us/me |
| 09:20 | <darobin> | gsnedders: honestly, don't worry your pretty head about copyright |
| 09:20 | <darobin> | I'll move the text; if there's a problem it'll all be my fault |
| 09:20 | <darobin> | we'll cross that bridge when we get there |
| 09:21 | <darobin> | (and I live within walking distance of a gun shop) |
| 09:21 | <darobin> | MikeSmith: yeah, I doubt there's a rush but I would still really like to get to a complete defork |
| 09:21 | <darobin> | ruby's one big chunk of that |
| 09:21 | <gsnedders> | yeah, I basically want everything done my way now I'm starting to pick all this stuff up again after years :P |
| 09:22 | <darobin> | good, good |
| 09:22 | <darobin> | I support that |
| 09:23 | hsivonen | wonders if non-UTF-8 encoders in browser need to be fast |
| 09:24 | <gsnedders> | hsivonen: encoders? for stuff like forms? |
| 09:27 | <hsivonen> | gsnedders: yes. also URLs |
| 09:28 | <hsivonen> | gsnedders: hmm. does XHR support non-UTF encoders still, too? |
| 09:28 | <hsivonen> | supporting serializing XML to anything other than UTF-8 is a really bad, but sadly common, idea |
| 09:30 | <hsivonen> | example: https://issues.apache.org/jira/browse/XALANJ-2419 Xalan's UTF-8 output is broken thanks to its attempt to support non-UTF-8 output |
| 09:30 | <hsivonen> | I filed that bug in January 2008 |
| 09:30 | <hsivonen> | still not fixed |
| 09:31 | <hsivonen> | meanwhile, I've written my own UTF-8-only XML serializer and moved on |
| 09:35 | <darobin> | hsivonen: is it still done a lot even in browsers? |
| 09:36 | <darobin> | it's been a while since I've seen that in any code tbh |
| 09:38 | <annevk> | hsivonen: only URLs and <form>, not XMLHttpRequest |
| 09:39 | <MikeSmith> | https://twitter.com/UzEE/status/638282373235384320 "Its 2015 and its sad that @w3c still hasn't come up with a standard way to allow #localStorage to be shared across domains/sub-domains." That could be read as saying, "...and it's good that they haven't" |
| 09:41 | <annevk> | hsivonen: parsing XML without the Encoding Standard is actually undefined territory since a lot of legacy encodings don't define errors... I mentioned this once to the XML folks, but they didn't care much |
| 09:41 | <philipj> | MikeSmith: regarding merge commits, I don't mind them, but I don't have the context of the discussion you were having |
| 09:47 | <hsivonen> | darobin: looking at the Gecko source, it seems that our XHR always uses UTF-8. Hooray. |
| 09:47 | hsivonen | has a vague recollection it wasn't always this way |
| 09:48 | <gsnedders> | Do we have any spec defining stuff around the DOM and XSLT? AFAICT not? |
| 09:48 | <darobin> | yeah, ISTR encoding issues with XHR but from a *long* time ago |
| 09:48 | <darobin> | gsnedders: there's a bug for that :) |
| 09:48 | <darobin> | but no, no one did the work for XSLT |
| 09:49 | <gsnedders> | This is, of course, hardly surprising. It's XSLT. :) |
| 09:51 | <annevk> | hsivonen: correct, we used to use inputEncoding or some such for serializing documents |
| 09:51 | <annevk> | hsivonen: I think per some research from hallvors we opted to remove that |
| 09:51 | <hsivonen> | annevk: cool |
| 09:51 | <hsivonen> | annevk: well, for now, I'll assume that encode needs to be fast |
| 09:52 | <hsivonen> | annevk: if we later find that it's OK for encode to be slower, we can rewrite code |
| 09:52 | <annevk> | hsivonen: I think it's okay for non-utf-8-encode to be slow personally |
| 09:52 | <annevk> | hsivonen: we already flag such pages in the console |
| 09:52 | <annevk> | hsivonen: and it'll only hit them for certain URLs and whenever they use <form> |
| 09:57 | <MikeSmith> | philipj: I wasn't advocating for merge commits. But anyway we already have a prohibition on doing them for this repo, so I think we're good. |
| 09:58 | <MikeSmith> | philipj: also, please review my preload patch ;) |
| 10:00 | <philipj> | MikeSmith: preload, huh? |
| 10:00 | <philipj> | I'll see if I can get to that tomorrow |
| 10:02 | <MikeSmith> | philipj: k |
| 10:03 | <MikeSmith> | (it's a minor change; shouldn't take much time to review, but no rush) |
| 10:04 | <hsivonen> | annevk: hmm. maybe I should save space after all, then |
| 10:04 | <philipj> | MikeSmith: ah, that fix |
| 10:04 | <philipj> | MikeSmith: I'll take a look now |
| 10:05 | <annevk> | philipj: are you planning on PR'ing all your own bugs? |
| 10:07 | <philipj> | annevk: I plan to spend some time looking at open bugs, sure, but my feelings will not be hurt if someone else fixes bugs I've filed |
| 10:07 | <philipj> | and at most I could probably spend an hour or two per day |
| 10:08 | <annevk> | Domenic: I use your merge script, but it's rather weird as sometimes it does the cleanest merge possible: https://github.com/whatwg/html/commit/882803c4cc8fba2fa5472b76f628d95cc82c421d and sometimes it does this: https://github.com/whatwg/html/commit/bbccfc976754def0c187ac8ce5891d2fb20dfc15 |
| 10:08 | <annevk> | Domenic: I committed both... |
| 10:12 | <annevk> | Domenic: I filed https://github.com/github/github-services/issues/1086 on shorter tweets |
| 10:17 | <hsivonen> | annevk: regarding https://encoding.spec.whatwg.org/#index-big5-pointer, are there other duplicate entries where the encoder should pick the *first* entry? |
| 10:17 | <annevk> | hsivonen: 4 others |
| 10:17 | <annevk> | iirc |
| 10:18 | <hsivonen> | annevk: ok. It would be nice to add a note that there are other duplicates where you want the first entry |
| 10:18 | <hsivonen> | annevk: to avoid an implementor just searching everything backwards |
| 10:20 | <annevk> | hsivonen: I was hoping the implication that it was special was obvious due to nothing else having this special case, but I'll add a note |
| 10:21 | <MikeSmith> | philipj: Thanks (for the review and push) |
| 10:22 | <hsivonen> | annevk: thanks |
| 10:23 | <annevk> | https://github.com/whatwg/encoding/commit/ce4e83d0df5b5efec0697fc76e66699737e033a3 |
| 10:31 | <annevk> | Domenic: oh, perhaps it only happens when the commit is based on the current master... |
| 10:41 | <annevk> | MikeSmith: when I ln source and then run ./build.sh it seems source ends up being copied somehow? |
| 10:44 | <annevk> | Now I get Parse Error: (61067,67) unexpected end tag |
| 10:44 | <annevk> | So something went wrong... |
| 10:44 | <MikeSmith> | annevk: maybe, but I'm on from my mobile atm, and can check when I get back home in ~45 minutes |
| 10:47 | <annevk> | I found it, mkwst made a typo somewhere |
| 10:47 | <mkwst> | Impossible! |
| 10:47 | <annevk> | And I didn't catch it because I wasn't yet building |
| 10:48 | <mkwst> | I'm still not building. So, my fault. |
| 10:50 | <MikeSmith> | that typo... doesn't sound like our mkwst |
| 10:51 | <MikeSmith> | this one must be an imposter |
| 10:51 | <mkwst> | no, I typo all the time. vim's spellchecker is horrible so I don't use it. |
| 10:51 | <MikeSmith> | check his gpg key |
| 10:51 | <mkwst> | If you get a GPG key, you _know_ it's not me. |
| 10:51 | <MikeSmith> | there's a way to hook in better spell checking in vim |
| 10:51 | <MikeSmith> | heh |
| 10:52 | <MikeSmith> | I'm pretty sure I got my vim set up with decent spell checking |
| 10:53 | <MikeSmith> | z= zg |
| 10:53 | <MikeSmith> | The suggestions I get from z= are usually ok |
| 10:54 | <MikeSmith> | and I don't get tons of false positives |
| 10:54 | <annevk> | Both PRs contained a typo |
| 10:55 | <mkwst> | annevk: Sorry. :/ |
| 10:56 | <annevk> | No worries, glad I found out now |
| 10:56 | <annevk> | And it's my fault since I merged them |
| 10:57 | <MikeSmith> | yup |
| 10:57 | <mkwst> | CC me on the PR? Just so I know what I broke? |
| 10:59 | <annevk> | mkwst: sure, I'll clean it up in an hour |
| 10:59 | annevk | has to go for a bit |
| 10:59 | <mkwst> | ok. tell me what I broke, I'll send you a PR? |
| 10:59 | <botie> | will do |
| 11:11 | <MikeSmith> | annevk: http://stackoverflow.com/questions/32309156/activate-browser-tab-using-html5-notification |
| 11:11 | <MikeSmith> | beverloo: http://stackoverflow.com/questions/32309156/activate-browser-tab-using-html5-notification |
| 11:35 | <annevk> | MikeSmith: mentioned something, haven't actually tested it through |
| 11:35 | <annevk> | though* |
| 11:37 | <smaug____> | has there be any work on standardizing popup blocking |
| 11:37 | <smaug____> | like, when can one open a new window and so |
| 11:38 | <annevk> | smaug____: yeah |
| 11:41 | <smaug____> | s/be/been/ |
| 11:44 | <MikeSmith> | annevk: thanks |
| 11:45 | <annevk> | smaug____: you want https://html.spec.whatwg.org/multipage/#allowed-to-show-a-popup |
| 11:46 | <MikeSmith> | annevk: I'm back home now, and can help you with build problems if you're still having anyy |
| 11:46 | <annevk> | MikeSmith: if I create a ln to another file, it seems after running the build script it's no longer an ln, but an actual file |
| 11:47 | <annevk> | MikeSmith: that's somewhat annoying since then I can't run the script multiple times without first removing the file and creating a new link |
| 11:47 | MikeSmith | looks in his own working directory |
| 11:47 | <smaug____> | ah, thanks |
| 11:49 | <MikeSmith> | annevk: I don't see that behavior in my environment. I have symlinks set up and they don't get overwritten no matter how many times I build |
| 11:49 | <MikeSmith> | anyway I will write up a patch for this tonight hopefully |
| 11:50 | <MikeSmith> | then you can please help test that |
| 11:51 | <annevk> | https://github.com/whatwg/html-build/issues/6 issue with caniuse integration |
| 11:52 | <annevk> | it seems the downloaded caniuse.json is input for wattsi, so maybe I should have filed the bug there... |
| 11:52 | <annevk> | MikeSmith: where you talking about palpable elements the other day? |
| 11:52 | <MikeSmith> | yeah |
| 11:53 | annevk | just found <!-- XXX this index doesn't list the palpable elements --> |
| 11:53 | <MikeSmith> | wrote a minor patch to add info to the spec |
| 11:53 | <MikeSmith> | yeah, saw that |
| 11:53 | <MikeSmith> | too |
| 11:54 | <MikeSmith> | I can add whatever's needed there, I guess (or try; I haven't looked yet at what it needs) |
| 12:21 | <annevk> | MikeSmith: does your html-build patch move the output elsewhere? |
| 12:21 | <annevk> | MikeSmith: or should I add unicode.xml, caniuse.json, cldr.inc, etc. to .gitignore too? |
| 12:24 | <MikeSmith> | annevk: I could make a patch that moves the output elsewhere, but I wasn't planning to move it. Was planning to just keep it where it is now |
| 12:24 | <annevk> | seems fine to me |
| 12:24 | <annevk> | I'll add more stuff to .gitignore |
| 12:24 | <MikeSmith> | yup |
| 12:24 | <MikeSmith> | sounds good |
| 12:24 | <zcorpan> | annevk: Domenic: philipj: is it intentional that your emails are not listed in html's acks? |
| 12:25 | <annevk> | zcorpan: dunno, ask Hixie |
| 12:25 | <philipj> | Domenic: I don't think it's intentional |
| 12:26 | <philipj> | It doesn't matter to me if it listed or not, but I also noticed the difference |
| 12:37 | <zcorpan> | Domenic: is d⊙dm your preferred email? and do you object to it being in the spec's acks? |
| 12:41 | <annevk> | zcorpan: see https://streams.spec.whatwg.org/#acks |
| 12:44 | <zcorpan> | thx. i'll just push this change without a PR |
| 13:05 | <mkwst> | Also, since I'm adding things to HTML: how do y'all feel about https://mikewest.github.io/credentialmanagement/writeonly/#examples-signin? |
| 13:05 | <mkwst> | Or is the answer "File a bug"? :) |
| 13:08 | <annevk> | mkwst: still not entirely clear to me whether that's worth the complexity, but yes, we can triage bugs |
| 13:09 | <mkwst> | annevk: I'm adding opaque FormData elements to CREDENTIAL based on the discussion at https://github.com/w3c/webappsec/issues/241. Seems like a short hop to writeonly fields. *shrug* I'll file a bug. |
| 13:09 | <annevk> | mkwst: perhaps I'm not understanding you |
| 13:09 | <annevk> | mkwst: what would be the alternative to an issue? |
| 13:11 | <mkwst> | Now I'm not understanding you. "alternative to an issue"? I'm happy to file a bug, that's a totally reasonable way to reboot the discussion. |
| 13:12 | <annevk> | mkwst: I wasn't sure what other answer you might be expecting |
| 13:12 | <mkwst> | annevk: "Yes! It's such a good idea, I'll grab the spec you wrote and backport the monkey patches myself right now!"? |
| 13:12 | <annevk> | mkwst: I see, hah |
| 13:16 | <mkwst> | So... that's not what you're saying, I guess? :) |
| 13:17 | <annevk> | mkwst: I think the main problem with the credential management API is lack of interest from other vendors, coupled with it not really solving federation well |
| 13:17 | <annevk> | mkwst: note that I'm not sure how to address either of those :-( |
| 13:18 | <mkwst> | annevk: Vendor interest will work itself out, one way or another. Federation is a hard problem. |
| 13:19 | <mkwst> | annevk: I don't know how to solve it completely and well. I do know how to solve a small piece of it that seems like it would have real value in terms of allowing users to avoid the "NASCAR"-style choosers on pages. |
| 13:19 | <mkwst> | annevk: But folks don't seem to like that proposal. |
| 13:20 | <mkwst> | annevk: Or, the three people who are engaging on the list don't like it. Not sure how well that scales to everyone; I know random folks on Google's identity team are standing on my neck about it. :) |
| 13:20 | <annevk> | I guess it would be a little bit more compelling if we had some content teams chime in |
| 13:20 | <annevk> | E.g., GitHub seems to say it doesn't work for them, but maybe Stack Exchange would adopt this immediately? |
| 13:21 | <annevk> | Also, do we have stats now on how often the cookie clearing happens affecting such scenarios? |
| 13:33 | <MikeSmith> | mkwst: your latest e-mail summarizes it all aptly |
| 13:33 | <MikeSmith> | will be interesting to see what responses you get to that |
| 13:34 | <mkwst> | annevk: I know that ~11% of Chrome users that opted into metrics clear cookies on a weekly basis. |
| 13:34 | <mkwst> | annevk: as I noted on the other thread, I don't have any more detail than that. Nor do I really know how I'd get it... |
| 13:35 | <MikeSmith> | ~11% is a relatively huge number |
| 13:36 | <mkwst> | 11.8% last week. |
| 13:36 | <mkwst> | I wonder if I can get historical information on this... my recollection is that it's been pretty constant. If users have learned anything at all about privacy on the web, it's "Clear cookies! All the time! Twice, even, just to be sure!" |
| 13:38 | <mkwst> | All I have is a raw counter. It's interesting, though: the counter is 5x the number of users. So this set of users is really clearing cookies _often_. |
| 13:40 | <mkwst> | MikeSmith: Dunno. I just want to ship the pieces that solve the problems I care about. I feel like doing that in an iterative fashion is reasonable. It's just not clear to me what the MVP actually is. |
| 13:40 | <annevk> | mkwst: I guess the other number that's interesting is how many have stored passwords and how often those are cleared (if ever) |
| 13:41 | <mkwst> | 3.04% of users check the "passwords" box when clearing browsing data. |
| 13:42 | <mkwst> | or, more accurately, 3.04% of users cleared passwords in the last week. |
| 13:42 | <mkwst> | (where "users" === "users who have opted into sharing statistics", etc) |
| 13:42 | <annevk> | so this is a feature for the remaining 8%? |
| 13:42 | <annevk> | I wonder what the numbers of Firefox are |
| 13:43 | <mkwst> | Cookies are fragile. *shrug* We lose them for all sorts of reasons unrelated to user intention. |
| 13:44 | <annevk> | Really? |
| 13:45 | <mkwst> | The login team has some numbers I don't actually remember about how often users log in without explicitly logging out. It was higher than I expected. |
| 14:33 | <Domenic> | MikeSmith: mkwst: Travis is a no-go because we need to maintain local caches/built copies of stuff |
| 14:33 | Domenic | is still reading scrollback |
| 14:34 | <mkwst> | Hrm? Travis can pull things down to do a clean build, right? |
| 14:36 | <wanderview> | mkwst: maybe they have the setting to clear cookies on browser close? |
| 14:37 | <mkwst> | wanderview: I think that's probably included in the counter, sure. But I'd be surprised if that was a significant percentage. |
| 14:37 | <mkwst> | wanderview: I wonder if we track that. Give me a minute. |
| 14:37 | <wanderview> | it was just a thought... or an addon to clear cookies on a regular interval |
| 14:41 | <wanderview> | I wish there was a way to follow significant changes to the html spec and easily hide smaller editorial changes |
| 14:41 | <wanderview> | it seems people have been waiting to change the html spec for a while |
| 14:42 | <mkwst> | wanderview: Something like 0.3% of users have the "session only" setting toggled. So it's probably a contributing factor, but not a huge one. |
| 14:42 | <Domenic> | mkwst: yes. That's the problem. It would do a clean build (~45 minutes) each time |
| 14:42 | <mkwst> | Domenic: ... |
| 14:42 | <Domenic> | mkwst: after we get the separate-directories thing straightened out complete builds should be ~2 minutes |
| 14:42 | <Domenic> | but it's incremental |
| 14:43 | <mkwst> | Domenic: I'd say that a ~45 minute build is the problem. |
| 14:43 | <Domenic> | it reuses caches and checkouts from the previous commits |
| 14:43 | <mkwst> | I'd say that a ~2 minute build is a problem. :) |
| 14:43 | <Domenic> | yes, well, it's 8 MB source |
| 14:43 | <Domenic> | maybe it's faster on a Z620 |
| 14:43 | <mkwst> | Chrome is a bit more than 8 MB, and it builds in a half hour on my mac mini at home. :) |
| 14:44 | <annevk> | Domenic: it only takes 5min to run the build script here |
| 14:44 | <mkwst> | maybe 45m. |
| 14:44 | <annevk> | Domenic: actually, prolly less |
| 14:44 | <Domenic> | annevk: right, that's an incremental build. No need to svn checkout, rebuild all the caches, etc. |
| 14:44 | <annevk> | Domenic: ah |
| 14:44 | <Domenic> | Also the demo CURLing has spiked the time a bit |
| 14:44 | <Domenic> | https://github.com/whatwg/html/issues/30 |
| 14:45 | <mkwst> | Domenic: have you considered caching some of the dependencies directly in the repo? I mean, caniuse doesn't change that often, right? |
| 14:45 | <Sebmaster> | annevk: is url's path reset somewhere when the parser encounters a windows drive letter? |
| 14:45 | <mkwst> | Ah, yeah. Same as that bug. :) |
| 14:46 | <Domenic> | mkwst: caniuse changes surprisingly often I think. But yeah I think for people besides the official build server we should have more cache type things |
| 14:46 | <annevk> | Sebmaster: for override you mean? |
| 14:46 | <Domenic> | mkwst: part of your "don't do everything every time" issue in some ways |
| 14:46 | <annevk> | Sebmaster: that might not be covered |
| 14:46 | <annevk> | Sebmaster: never mind, it is |
| 14:47 | <Sebmaster> | annevk: no i mean if you have if you have a file:C:\test parsing against a base of file:///tmp/var/ |
| 14:47 | <annevk> | Sebmaster: it should simply not copy base's path |
| 14:47 | <Sebmaster> | ah i see |
| 14:47 | <Sebmaster> | I'll need to find where i do that then |
| 14:57 | <Sebmaster> | sweet, got it, thanks annevk |
| 15:05 | <Domenic> | OK, caught up on email, time to eat breakfast... then a day to spend on the build script. |
| 15:25 | <wanderview> | annevk_: is it reasonable for a 404 html response page to register a service worker? |
| 15:27 | <annevk> | wanderview: yeah |
| 15:27 | <wanderview> | interesting, ok |
| 15:51 | <zcorpan> | annevk: so the spec changes that were done in ResponsiveImagesCG is cc0. the comment in the source saying so is not preserved |
| 15:52 | <annevk> | zcorpan: I'm not sure how you could tell what are changes and what are not |
| 15:53 | <zcorpan> | annevk: yeah dunno |
| 15:53 | <annevk> | zcorpan: that would be helped by the commit log |
| 15:54 | <annevk> | zcorpan: we should maybe address that in the Acknowledgments section? |
| 15:54 | <annevk> | zcorpan: that might also be a better place for the history of that section |
| 15:56 | <zcorpan> | annevk: yeah. we can work on that as a separate change |
| 15:58 | <annevk> | speaking of which, it would be nice if any new contributions to the HTML Standard were CC0 |
| 16:26 | <annevk> | Domenic: apparently you can also turn an issue into a PR |
| 16:26 | <annevk> | Domenic: but you need GitHub command line tools for that |
| 16:26 | <Domenic> | annevk: yeah, I find that a bit meh too; I like having PRs close issues. *shrug* |
| 16:27 | <annevk> | I see |
| 16:27 | <annevk> | I thought it was rather cool when someone did that to me |
| 16:27 | <Domenic> | heh |
| 16:27 | <Domenic> | OK so I'm going to try the picture stuff |
| 16:27 | <annevk> | Domenic: zcorpan is trying to merge |
| 16:28 | <Domenic> | oh ok then |
| 16:28 | <Domenic> | let me know when it's time ot update the build script |
| 16:28 | <annevk> | Domenic: he hasn't done the clean merging yet so seems like a good opportunity |
| 16:28 | <zcorpan> | yeah i've only push left to do, assuming i got the rest right |
| 16:28 | <annevk> | Domenic: apparently git push upstream doesn't work |
| 16:28 | <annevk> | Domenic: should that just be "git push" |
| 16:28 | <Domenic> | annevk: it depends what you named your remote. Mike named his upstream I think |
| 16:29 | <zcorpan> | i'm doing this from whatwg/html, not from a fork |
| 16:29 | <Domenic> | then yeah git push === git push origin should work |
| 16:29 | <zcorpan> | k |
| 16:29 | <Domenic> | (origin is the default remote name) |
| 16:29 | <annevk> | Domenic: you merge on your remote? interesting |
| 16:29 | <annevk> | so much to learn |
| 16:30 | <Domenic> | oh no i just push to it |
| 16:30 | <Domenic> | I think |
| 16:30 | <zcorpan> | there we are |
| 16:30 | <Domenic> | I pretty extensively use a GUI to visualize the tree tbh |
| 16:30 | <Domenic> | nice zcorpan |
| 16:30 | <Domenic> | I'll do the build merge |
| 16:31 | <annevk> | it seems it did lose the PR connection |
| 16:31 | <zcorpan> | thanks for the pr thing Domenic |
| 16:31 | <Domenic> | annevk: yeah gotta force-push to the appropriate branch after the rebase. Which is actually kind of hairy when someone else created the branch so you don't have a local copy yet. Probably not worth the trouble. |
| 16:32 | <tantek> | ^^^ lol git |
| 16:32 | <zcorpan> | -_- |
| 16:32 | <Domenic> | just read the man pages http://git-man-page-generator.lokaltog.net/ |
| 16:32 | <tantek> | "just" :) |
| 16:32 | <Domenic> | tantek: click the link. and refresh a few times ;) |
| 16:33 | <zcorpan> | i'll go running instead. that i know how to do |
| 16:33 | <tantek> | zcorpan++ I'm right there with you. |
| 16:33 | <tantek> | Domenic: ah, a js;dr page I see ;) |
| 16:34 | <Domenic> | tantek: stop trying to make js;dr happen. it's not going to happen. |
| 16:34 | <tantek> | ok that's pretty amazing |
| 16:37 | <tantek> | zcorpan: next f2f meeting, bring your running shoes and come running with Rossen and me. |
| 16:38 | <zcorpan> | tantek: sure. if i forget the shoes i can run barefoot :-) |
| 16:40 | <zcorpan> | i've run 10k on tarmac/gravel on 52 mins (but my feet needed some recovery after that) |
| 16:41 | <tantek> | ouch. you must have some serious callouses! |
| 16:41 | <Domenic> | annevk: going to just merge https://github.com/whatwg/html/pull/7 mk? |
| 16:41 | <tantek> | you'll be good competition for Rossen then, I'll catch up eventually. |
| 16:42 | <Domenic> | I think we should prioritize load time over issue filing. |
| 16:42 | <annevk> | Domenic: you mean the other way around? |
| 16:42 | <Domenic> | annevk: defer prioritizes load time |
| 16:42 | <Domenic> | annevk: and I think we should keep it that way |
| 16:43 | <Domenic> | Relatedly, we should do this https://www.w3.org/Bugs/Public/show_bug.cgi?id=25943 |
| 16:43 | <annevk> | Domenic: anyway, it's fine |
| 16:43 | <zcorpan> | tantek: i have thick skin from walking barefoot on gravel on a daily basis |
| 16:43 | <annevk> | Domenic: I'd like to audit our scripts at some point, but this is not the time |
| 16:43 | <Domenic> | annevk: agreed |
| 16:44 | <Domenic> | Oh, I need to set up the PDF stuff today |
| 16:44 | <annevk> | Ah yeah, would be good to reply to them |
| 16:45 | <Domenic> | I am glad we're keeping number of open PRs low |
| 16:45 | <Domenic> | I think a good goal for any project is zero open PRs even if many open issues. |
| 16:47 | <Domenic> | Hmm https://html.spec.whatwg.org/#dependencies-2 seems kind of out of place |
| 16:48 | <annevk> | Domenic: yeah, <picture> is not done |
| 16:48 | <annevk> | Domenic: that first commit was just doing what the build script does |
| 16:48 | <Domenic> | Maybe add a checklist to https://github.com/whatwg/html/issues/52 |
| 16:48 | <annevk> | Domenic: or did I miss something the build script does? |
| 16:48 | <annevk> | sure |
| 16:48 | <Domenic> | annevk: nah I think that was it |
| 16:49 | <MikeSmith> | Domenic: fwiw after reading https://github.com/whatwg/html/pull/58#issuecomment-136397127 I agree with you completely |
| 16:50 | <Domenic> | cool :) |
| 16:50 | <MikeSmith> | until I read that, I just didn't know that's what y'all had been doing |
| 19:59 | <Domenic> | annevk: thanks for catching me on acks. I will try to do better in the future! |
| 20:07 | <annevk> | Domenic: no problem |
| 20:08 | <annevk> | Domenic: any reason we can't start using utf-8 in the source? I guess Hixie had some reason for keeping it ASCII-clean |
| 20:08 | <Domenic> | annevk: yeah no idea, I just stuck with existing convention |
| 20:08 | <gsnedders> | annevk: dev.w3.org at least used to be a problem, and getting it to reliably send any charset was hard |
| 20:09 | <gsnedders> | annevk: that's the reason why it stayed ASCII-onyl |
| 20:09 | <annevk> | hmm, if that's the only reason we can move, but we should maybe ask Hixie to be sure |
| 20:14 | <zcorpan> | i just realized that it was about 10 years ago i sent my first email to whatwg. https://lists.w3.org/Archives/Public/public-whatwg-archive/2005Jun/0099.html |
| 20:25 | <zcorpan> | next 10 years, guys! |
| 20:34 | <gsnedders> | annevk: certainly in my days developing Anolis that was it |
| 20:39 | <jamesr___> | zcorpan: https://lists.w3.org/Archives/Public/public-whatwg-archive/2005Jun/0109.html is a pretty good idea |
| 20:40 | <zcorpan> | jamesr___: one part of the proposal was adopted :-) |
| 20:41 | <jamesr___> | the better part |
| 23:25 | <Jasper> | Do any browsers implement the XHTML syntax as specified in the HTML spec? |
| 23:42 | <MikeSmith> | Jasper: yes |
| 23:42 | <Jasper> | Hm. |
| 23:42 | <Jasper> | When I tried locally here, Chromium and Firefox both allowed document.write(); in XHTML documents, even though the spec said that was disallowed. |
| 23:43 | <MikeSmith> | well if you mean do all browsers completely conform to the spec requirements, then the answer is, probably not |
| 23:44 | <MikeSmith> | Jasper: to be clear, by "XHTML syntax", you mean the document is served with a XML media type, right? |
| 23:44 | <MikeSmith> | instead of as text/html |
| 23:44 | <Jasper> | MikeSmith, yes. |
| 23:45 | <Jasper> | MikeSmith, application/html+xml, I believe it is |
| 23:45 | <Jasper> | Although https://html.spec.whatwg.org/multipage/xhtml.html does not exactly say. |
| 23:45 | <Jasper> | I do like the wording of "At the time of writing, no such rules actually exist." |
| 23:45 | <MikeSmith> | yeah, because it doesn't have to stricly be application/html+xml |
| 23:45 | <MikeSmith> | there is lots of great wording in the spec like that |
| 23:46 | <MikeSmith> | anyway, for one thing, I'd suggest you submit some actual test cases |
| 23:46 | <MikeSmith> | to https://github.com/w3c/web-platform-tests |
| 23:46 | <MikeSmith> | or check there and see if there are already some test cases |
| 23:47 | <MikeSmith> | under https://github.com/w3c/web-platform-tests/tree/master/html |
| 23:47 | <MikeSmith> | the subdirs there correspond to sections in the spec |
| 23:47 | <MikeSmith> | and the subdirs of those, to subsections, etc. |
| 23:48 | <MikeSmith> | and you can run the tests in browsers from http://w3c-test.org/html/ |
| 23:48 | <MikeSmith> | that's just a mirror of the current state of that github repo |
| 23:49 | <MikeSmith> | as far as the spec goes, if tests show that UAs don't conform on this, then that might merit putting a note or annotation in the spec so that people can know that |
| 23:49 | <Jasper> | MikeSmith, I'm more curious because I thought the WHATWG gave up on XHTML, and was surprised to see a half-finished spec about it. |
| 23:50 | <MikeSmith> | or it might even merit changing the requirements in teh spec to match what most UAs do |
| 23:50 | <MikeSmith> | it's not a half-finished spec |
| 23:50 | <MikeSmith> | for XHTML |
| 23:50 | <Jasper> | I'm part of a team that wants to perhaps use schemas and XSLT and such to verify and transform XHTML, and when I mentioned it was dead, they pointed me there. |
| 23:50 | <MikeSmith> | well |
| 23:51 | <MikeSmith> | the HTML spec is not something that any reasonable person would see as being strong support somehow for using schemas and XSLT |
| 23:52 | <MikeSmith> | and if you're in a team that wants to use schemas and (client-side?) XSLT then you have my sympathy |
| 23:52 | <MikeSmith> | because that will probably end badly |
| 23:52 | <MikeSmith> | and I say that as somebody who has done a lot of work in XSLT, and knows quite a lot about schemas and still does a lot of work in schemas |
| 23:52 | <Jasper> | MikeSmith, I agreed. |
| 23:53 | <MikeSmith> | OK |
| 23:53 | <Jasper> | MikeSmith, we hired a ~*~ contractor ~*~ who loves XML and XSLT and recommended buying an enterprise XSLT server license. |
| 23:53 | <MikeSmith> | anyway, the spec is in this regard mostly just documenting what's actually implemented in UAs |
| 23:53 | <Jasper> | It's not client-side XSLT -- it's going to be server-side. |
| 23:53 | <Jasper> | OK. |
| 23:53 | <MikeSmith> | christ |
| 23:53 | <MikeSmith> | so get rid of that contracter |
| 23:54 | <Jasper> | I'm pushing back as hard as I can. |
| 23:54 | <MikeSmith> | yeah you should |
| 23:54 | <Jasper> | We already have a contingency plan to get rid of XSLT, since we already believe they'll deliver us something completely unmanageable. |
| 23:54 | <MikeSmith> | well I don't think server-side XSLT is the worst thing in the world |
| 23:54 | <MikeSmith> | it's just that there are many better things |
| 23:55 | <Jasper> | OK, fair enough. I don't have any experience with XSLT. |
| 23:55 | <Jasper> | Currently we're using manual DOM manipulations, which isn't great either. If you have any recommendations, please let me know. |
| 23:55 | <MikeSmith> | ok, hire me as a consultant :-) |
| 23:56 | <Jasper> | MikeSmith, what are your rates? |
| 23:57 | <Jasper> | MikeSmith, it would be nice to know what you think about HTML, XML, transformations, and such -- we're far from experts on the stuff, we just have half-baked opinions based on what we've written. |
| 23:57 | <Jasper> | But this channel probably isn't the time or place for it. |
| 23:58 | <MikeSmith> | let's see, last time I did consulting full-time I think my employer billed 1800 USD per day for my time. And that was a few years go. So something in that range. |
| 23:59 | <MikeSmith> | Jasper: yeah I don't know where would be the best place to ask |
| 23:59 | <Jasper> | Do you mind if we talk a bit in PM? Don't want to mooch or anything. |