| 07:56 | <XhmikosR> | MikeSmith: are you around? |
| 09:18 | <Ms2ger> | > If ! IsCallable(F) is false: |
| 09:18 | Ms2ger | wonders who invented that syntax |
| 09:24 | <annevk> | Ms2ger: TC39 |
| 09:25 | <annevk> | Ms2ger: they prolly got inspiration from CSS |
| 09:26 | Ms2ger | goes back to trying to find the point of the skip-when-determining-incumbent counter |
| 09:27 | <MikeSmith> | XhmikosR: here now |
| 10:21 | <annevk> | zcorpan: what is still broken about the wiki? |
| 10:21 | <annevk> | zcorpan: it seems to function here |
| 10:21 | <annevk> | zcorpan: both logged in and out |
| 10:21 | <annevk> | OhReally added by MikeSmith even modified a page |
| 10:27 | <zcorpan> | annevk: i did not mean that it is currently broken, but it was recently |
| 10:27 | <zcorpan> | MikeSmith: does email from the wiki work now? |
| 10:28 | <MikeSmith> | yes |
| 10:28 | <zcorpan> | ok great |
| 12:11 | <annevk> | Anyone know when mkruisselbrink returns? |
| 12:11 | <annevk> | I'd like to get https://github.com/whatwg/url/issues/127 sorted |
| 12:37 | <XhmikosR> | MikeSmith: I went ahead and opened https://github.com/whatwg/html/pull/2230 for rel=mask-icon |
| 12:39 | <XhmikosR> | but I'm not sure there should be a specification for this at all. I mean it's a proprietary attribute |
| 12:41 | <annevk> | XhmikosR: lots of standards start out that way |
| 12:42 | <XhmikosR> | annevk: true, but still, personally I simply want it as an exception at this point |
| 13:16 | <annevk> | zcorpan: not sure what a "control picture" is |
| 13:23 | <zcorpan> | annevk: https://en.wikipedia.org/wiki/Unicode_control_characters#Control_pictures |
| 13:28 | <annevk> | zcorpan: ah okay, the stuff I found |
| 13:28 | <annevk> | I think I'd prefer something like that as well, I've never really understood why the long form is helpful |
| 13:31 | <zcorpan> | i guess the redundancy can help catch mistakes; if the character and the code point aren't the same, the spec editor made a mistake. but that's something the tooling can point out early if we want to provide both when writing a spec |
| 13:34 | <zcorpan> | but i'm not sure that's very useful anyway. could as well write "<code>a</code>" (or something simpler still) in the source and have bikeshed generate the desired output |
| 15:12 | <MikeSmith> | XhmikosR: thanks for creating that PR. As annevk mentioned, there are already lots of things in the spec that started out as proprietary, that were for better or worse initiated unilaterally by one browser project or another. |
| 15:17 | <MikeSmith> | XhmikosR: and I agree with zcorpan that we should just go ahead and define the UA processing. If you have time to help write up the details for that, it would be very welcome. But if your level or interest in this doesn’t merit you putting that much time into, that’s OK too, because as zcorpan says there, we can push further commits to that PR branch to add those details, and land it all together in |
| 15:17 | <MikeSmith> | the spec after that’s done. |
| 16:56 | <mounir> | bzed, annevk: I see that the notifications api has a readonly attlibute FrozenArray<NotificationAction> but WebIDL says that Dictionaries can't be used as parameters -- is it okay because it is technically a FrozenArray or is this part of the spec changing? |
| 16:57 | <Ms2ger> | What |
| 16:57 | <mounir> | s/parameters/attributes/ |
| 16:58 | <Ms2ger> | Well, the type of the attribute is not a dictionary |
| 16:58 | <mounir> | isn't this mostly the same? |
| 16:58 | <Ms2ger> | So if the rule is "the type of an attribute must not be a dictionary", that API is technically correct |
| 16:58 | <mounir> | or maybe I misunderstand why there is this requirement |
| 16:59 | <Ms2ger> | (best kind of correct) |
| 16:59 | <mounir> | Ms2ger: "Dictionaries must not be used as the type of an attribute or constant." |
| 16:59 | <mounir> | so yes, but I'm not fully satisfied :) |
| 16:59 | <Ms2ger> | The reason for the rule is to ensure foo.x === foo.x is true |
| 16:59 | <Ms2ger> | If the type was a dictionary, those would be two different attributes |
| 16:59 | <Ms2ger> | s/attributes/objects/ |
| 17:02 | <mounir> | Ms2ger: thanks for the background :) |
| 17:04 | <XhmikosR> | MikeSmith: I don't mind spending time on it. It's just that I'm not familiar with the whole process. So perhaps it would be better if one more experienced wrote the UA spec |
| 17:07 | <MikeSmith> | XhmikosR: well we’re not in any hurry and can help get you familiar with the process if you want to put time into it. They real payoff for you is that the next time you want to get something added to the spec, since you will have had the experience already, you can just sail through it then :) |
| 17:07 | <MikeSmith> | and the next time and the next |
| 17:08 | <MikeSmith> | til you have contributed a lot of stuff :) |
| 17:08 | <XhmikosR> | I doubt I'll need to do it anytime soon but I see your point :p |
| 17:09 | <MikeSmith> | yeah you may be surprised how soon something else comes up again that you have an interest in seeing get added or changed in the spec |
| 17:10 | <MikeSmith> | once you know it is a real possibility to be able to change the spec, it tends to make you see other things in need of improvement in it |
| 17:11 | <MikeSmith> | whereas before then it just seems like something unapproachable that other people toil away at |
| 17:11 | <XhmikosR> | MikeSmith: all right, how should I proceed then? Perhaps it might be better to keep the code comments on GH for future reference |
| 17:12 | <MikeSmith> | OK well I guess what I would do in this case is look for the most recent time that somebody added something similar to the spec |
| 17:12 | <MikeSmith> | which for this is either a standard rel value but maybe also a standard meta name value |
| 17:12 | <MikeSmith> | like meta[mame=theme-color] |
| 17:13 | <MikeSmith> | because that one like this also has UA requirements |
| 17:14 | <MikeSmith> | so I would search git log output to find when that meta[name=theme-color] addition was made and take a look at the diff for that |
| 17:14 | <XhmikosR> | it requires some tests too, right? |
| 17:14 | <MikeSmith> | yeah but you do not have to write those |
| 17:15 | <MikeSmith> | you can if you want and you may find that you have to for yourself in order to be able to spec it |
| 17:15 | <XhmikosR> | I'll try to see a previous patch and see how it goes |
| 17:15 | <MikeSmith> | OK |
| 17:17 | <MikeSmith> | that particular meta[name=theme-color] change might not be the most relevant corollary to rel=mask-icon, I dunno |
| 17:18 | <MikeSmith> | but if it’s not, you can ask here or in that github issue to see if anybody else has any better ideas as far as something to look at as a starting point |
| 17:28 | <Domenic> | XhmikosR: in this particular case I would try to base the UA processing model off of rel=icon. And looking at that part of the spec, most of it is dealing with sizes (not applicable here) and /favicon.ico fallback (not applicable here). So it's going to be pretty short, just a paragraph or so. |
| 17:29 | <annevk> | mounir: I suspect dictionaries will be allowed as attributes values at some point, but who knows |
| 17:29 | <XhmikosR> | Domenic: thanks, it's indeed a lot simpler. I'm just trying to set up the build tools locally for testing |
| 17:29 | <Domenic> | XhmikosR: ah yeah. Let us know if you need help with that. |
| 17:30 | <XhmikosR> | I'm a Windows user but shouldn't be hard since I have MSYS around |
| 17:33 | <gsnedders> | Is there any reason why HTML doesn't move to Bikeshed? |
| 17:33 | <gsnedders> | Apart from, well, the work involved? |
| 17:33 | <bzed> | mounir: I doubt you wanted to ask me that :) |
| 17:35 | <annevk> | gsnedders: speed, though perhaps that's less of an issue now |
| 17:36 | <XhmikosR> | Domenic: has anyone used the build tools on Windows? I'm getting tons of lint errors |
| 17:36 | <XhmikosR> | does the linter check the line endings? If so that would explain it |
| 17:36 | <XhmikosR> | bbs |
| 17:37 | <Domenic> | XhmikosR: ah yeah it probably does |
| 17:37 | <Domenic> | XhmikosR: I thought we set up Git to only use LF via .gitattributes, interesting |
| 17:38 | <Domenic> | XhmikosR: I use Windows myself :) |
| 17:57 | <XhmikosR> | Domenic: I don't see a .gitattributes file in the repo |
| 17:57 | <Domenic> | XhmikosR: there is one in whatwg/html; maybe not in whatwg/html-build |
| 17:57 | <XhmikosR> | yeah I mean in html-build |
| 17:57 | <Domenic> | XhmikosR: but yeah we haven't set the eol=lf setting anywhere, I guess we should |
| 17:58 | <XhmikosR> | Domenic: definitely if the linter expects LF |
| 17:58 | <XhmikosR> | BTW this is what I get if I remove the linter from build.sh |
| 17:58 | <XhmikosR> | Downloading list of W3C bugzilla bugs... |
| 17:58 | <XhmikosR> | Pre-processing the source... |
| 17:58 | <XhmikosR> | missed end of <noscript> intro. |
| 17:58 | <Domenic> | Yeah I imagine a lot of the toolchain gets thrown off by CRLFs |
| 17:59 | <XhmikosR> | also, the gitattributes in html repo isn't enforcing LF either |
| 17:59 | <XhmikosR> | source text diff=html linguist-language=HTML |
| 17:59 | <Domenic> | yeah, as I said, we should add that |
| 17:59 | <Domenic> | But in the meantime you can just do a clone with those settings |
| 17:59 | <Domenic> | I still can't believe Git has such a bad default on windows... |
| 17:59 | <XhmikosR> | how about having AppVeyor for Windows CI |
| 17:59 | <XhmikosR> | I can just add a gitattributes file and make a PR |
| 17:59 | <Domenic> | sure |
| 17:59 | <Domenic> | bbiab |
| 18:01 | <nox> | Domenic: Well, AFAIK notepad doesn't handle LF. |
| 18:01 | <nox> | Or did that finally get fixed? |
| 18:01 | <XhmikosR> | if one is using notepad for any kind of development, they shouldn't be called devs |
| 18:05 | <nox> | That doesn't sound very inclusive. |
| 18:06 | <XhmikosR> | nox: it's a fact. It's just a text viewer, nothing more. Doing any development with Windows notepad is asking for trouble |
| 18:07 | <nox> | Same as using nano from command line or textedit on OS X, that you shouldn't do it on a day by day basis doesn't mean it's never legit. |
| 18:08 | <nox> | Also, VS uses CRLF by default too, |
| 18:08 | <nox> | so I don't see how that is a bad default for Git. |
| 18:11 | <XhmikosR> | I have no interest in debating this. Use whatever you want yourself |
| 18:11 | <XhmikosR> | Domenic: the linter fails with LF too |
| 18:12 | <annevk> | XhmikosR: and you made no changes? |
| 18:15 | <annevk> | Is it normal for WPT Travis to hang for a long time? |
| 18:15 | <annevk> | Been waiting on https://github.com/w3c/web-platform-tests/pull/4413 for an hour now apparently |
| 18:23 | <jgraham_> | annevk: So I broke travis, sort of |
| 18:23 | <jgraham> | Upstreamed ~100 commits from m-c and that caused many failed travis builds |
| 18:24 | <jgraham> | It seems like a problem that could be fixed |
| 18:28 | <XhmikosR> | annevk: no changes, only LF for my local repos |
| 18:28 | <XhmikosR> | it's this line in lint.sh that fails on Windows https://github.com/whatwg/html-build/blob/master/lint.sh#L17 |
| 18:28 | <XhmikosR> | I made PRs for you guys to look at |
| 18:31 | <Domenic> | XhmikosR: something must still be going wrong; on my Windows it never fails. |
| 18:31 | <Domenic> | PR looks pretty good, can merge in a bit after a doctor's appointment |
| 18:32 | <XhmikosR> | Domenic: have you set autocrlf to false somewhere? |
| 18:32 | <Domenic> | XhmikosR: yes, globally |
| 18:32 | <XhmikosR> | that must be why |
| 18:32 | <Domenic> | I guess |
| 18:32 | <XhmikosR> | I use the default setting |
| 18:33 | <XhmikosR> | but since it doesn't clone the html repo it shouldn't matter. My local branch has LF |
| 18:35 | <XhmikosR> | so I believe it must be the grep binary. What does `which grep && grep --version` output for you? |
| 18:35 | <XhmikosR> | $ which grep && grep --version |
| 18:35 | <XhmikosR> | GNU grep 2.5.4 |
| 18:36 | <XhmikosR> | err |
| 18:36 | <XhmikosR> | $ which grep && grep --version |
| 18:36 | <XhmikosR> | /bin/grep.exe |
| 18:36 | <XhmikosR> | GNU grep 2.5.4 |
| 18:36 | <Domenic> | /usr/bin/grep |
| 18:36 | <Domenic> | grep (GNU grep) 2.26 |
| 18:37 | <XhmikosR> | Domenic: is that from msysgit? |
| 18:38 | <Domenic> | I think so |
| 18:38 | <XhmikosR> | yup I confirmed it |
| 18:38 | <XhmikosR> | ok so it's due to the old grep MSYS has |
| 20:32 | <XhmikosR> | Domenic: any idea why the output html files have tons of trialing whitespace? |
| 20:34 | <Domenic> | XhmikosR: hmm I've never checked, let me see if they do locally... |
| 20:36 | <Domenic> | XhmikosR: I don't see any locally just scrolling through... any line numbers I should check out? |
| 20:37 | <Domenic> | Ah I see some blank lines with trailing whitespace |
| 20:37 | <XhmikosR> | you won't see them in the resulting HTML |
| 20:37 | <XhmikosR> | but they are there |
| 20:37 | <Domenic> | Right I'm looking at the index.html output |
| 20:37 | <XhmikosR> | if your editor has an option to trim trailing spaces just do that |
| 20:38 | <Domenic> | Yeah I never edit the index.html directly |
| 20:38 | <XhmikosR> | ok so it's in the source expanded files too |
| 20:38 | <Domenic> | It's probably some artifact of the build process |
| 20:38 | <Domenic> | It is? The `source` file, at least, has none |
| 20:39 | <XhmikosR> | yeah `source` has nothing |
| 20:39 | <XhmikosR> | so since they are in temp/*preprocessed something in .pre-process-main.pl must do that |
| 20:44 | <XhmikosR> | it's 6149 whitespace characters in the all in one index.html file |
| 20:46 | <XhmikosR> | one I think is when a block is skipped like `w-nohtml`; that is being replaced with 3 spaces |
| 20:48 | <Domenic> | Makes sense. Not really something to be worried about. |
| 21:09 | <XhmikosR> | the html.spec.whatwg.org server is slow and the content is just static / |
| 21:10 | <XhmikosR> | it takes 0.6 s to load the html file from London |
| 21:15 | <XhmikosR> | also, doesn't seem the docs to comply with wcag2aa |