| 06:12 | <annevk> | rbyers: one compat issue where it's been super hard to get feedback from Chrome folks is URLs |
| 06:12 | <annevk> | rbyers: I pinged Philip about it, but thought I'd let you know too |
| 06:21 | <annevk> | foolip: you had some way to view tests that were checked into a branch, how does that work? |
| 06:21 | <annevk> | jgraham: would be nice if a bot for WPT posted a comment with the test URLs for the branch |
| 07:27 | <MikeSmith> | annevk: should be easy to tweak the stability checker comments to add that |
| 07:27 | <MikeSmith> | you mean links to w3c-test.org/submission/NNNN right? |
| 07:34 | <annevk> | MikeSmith: ideally links to all the files that have changed |
| 07:34 | <annevk> | MikeSmith: otherwise it's still manual work of putting 2 and 2 together |
| 07:34 | <MikeSmith> | yeah |
| 07:35 | <MikeSmith> | OK will hack up a PR for it right now |
| 07:39 | <tobie> | What's the preferred option when it comes to algorithms with an if -> then -> return structure, e.g.: https://gist.github.com/tobie/56751420a4597ba25a7ed0e9aaac6f6b |
| 07:41 | <tobie> | Split the return into it's own substep or keep them both together? |
| 07:46 | <annevk> | Is there an easy way to remove local git branches that are not published? |
| 07:47 | <annevk> | tobie: that is pretty good, sometimes "Otherwise" is not used when the first clause returns anyway or the clauses are mutually exclusive |
| 07:47 | <annevk> | tobie: we should have some norms for this, I used to think they would belong in IDL, but maybe something even higher-level should describe the rules |
| 07:47 | <tobie> | git branch -D [BRANCH_NAME] |
| 07:48 | <annevk> | tobie: is there a way to do it without names? (there's a lot) |
| 07:49 | <tobie> | no idea. :( |
| 07:50 | <tobie> | wrt to algorithms, just a set of best practices would be amazing. |
| 07:50 | <tobie> | annevk: have you looked at both versions of the algo I linked to? |
| 08:02 | <jochen__> | annevk, I think the problem with tests that demonstrate compatible implementations is that they're not compatible |
| 08:08 | <MikeSmith> | annevk: about the git thing, to get just a list of those branches, try this: |
| 08:08 | <MikeSmith> | git branch -vv | cut -c 3- | awk '$3 !~/\[/ { print $1 }' |
| 08:13 | <MikeSmith> | hmm actually I think that may still list branches that do have remotes |
| 08:14 | <MikeSmith> | will try something else |
| 08:15 | <MikeSmith> | annevk: in the mean time, see the wpt-stability-bot comments in https://github.com/w3c/web-platform-tests/pull/4050 |
| 08:17 | <MikeSmith> | that uses http://w3c-test.org as the base but since there are some tests that need be run under TLS I guess I should make it emit https://w3c-test.org instead |
| 09:00 | <tobie> | MikeSmith: that's pretty cool! |
| 09:00 | <tobie> | MikeSmith: I <3 GH bots |
| 09:05 | <annevk> | So a lot of the advice in https://github.com/whatwg/html/blob/master/TEAM.md#handling-pull-requests is dated I think given that you can now push to a contributors branch (after cloning their repository) |
| 09:05 | <annevk> | That is, the approach MikeSmith added doesn't allow for pushing fixes to those branches, it just gives you something to look at which you can also do online |
| 09:06 | <annevk> | But also, I think we should now try to clone those repos (even though they're large) if we want to make fixup commits as that'll lead to purple PRs, which is something contributors like |
| 09:08 | <MikeSmith> | agreed |
| 09:08 | <MikeSmith> | so I can spend some time updating it |
| 09:09 | <MikeSmith> | especially if you can take time to raise an issue |
| 09:09 | <annevk> | MikeSmith: oh great, thanks |
| 09:09 | <annevk> | MikeSmith: I'll create an issue |
| 09:09 | <MikeSmith> | great |
| 09:10 | <MikeSmith> | tobie: jgraham points out that some PRs don’t get mirrored to w3c-test.org so those URLs will be 404 in those cases |
| 09:11 | <MikeSmith> | tobie: but I think for now it’s not (yet) worth trying to make it smarter about what to output on those cases |
| 09:11 | <annevk> | MikeSmith: is that when the PR doesn't use a local branch? |
| 09:12 | <annevk> | MikeSmith: if it's easy to detect might be nice to fail gracefully, but something is better than nothing |
| 09:12 | <jgraham> | When the submitter is trusted |
| 09:12 | <jgraham> | We can detect it, but it's a little bit of work |
| 09:15 | <annevk> | tobie: sorry for replying so late, definitely a over b |
| 09:15 | <annevk> | jochen__: I think WebKit and Blink will largely be the same (apart from that same-origin check) |
| 09:15 | <annevk> | jochen__: dunno about Edge |
| 09:16 | <annevk> | jochen__: but proving Edge doesn't create a new global would be good and maybe figuring out some of the other things it does |
| 09:17 | <tobie> | annevk: np. I see two tendencies these days wrt to algorithms. One favoring making it look as much as possible like regular prose, the other aiming for near pseudo-code. |
| 09:18 | <tobie> | annevk: is there reasoning behind this outside of personal taste? |
| 09:18 | <annevk> | tobie: I think I'm definitely in the latter camp, but if it's simple like this I don't see the need for verbosity |
| 09:19 | <tobie> | we definitely need a guide for this stuff |
| 09:19 | <annevk> | tobie: the verbosity might be needed if what to return is itself conditional or requires some variable allocation and such |
| 09:19 | <annevk> | tobie: but when it's just if X, then Y, I don't see the need for a nested list |
| 09:19 | <tobie> | annevk: makes sense. thanks |
| 09:22 | <annevk> | another pattern I've started using is with "one of X / Y / Z" but X / Y / Z are complicated; slowly been using lists for that where the list is basically part of the paragraph but cannot be nested in it due to silly HTML parsing rules |
| 09:25 | <Ms2ger> | Remember when Hixie_ invented <span><ul> for that? |
| 09:26 | <annevk> | Oh is that standardized? |
| 09:26 | <Ms2ger> | No, it was dropped at some point |
| 09:26 | <annevk> | Seems a little icky, but if it works I wouldn't necessarily mind using it |
| 09:26 | <annevk> | Oh okay |
| 09:27 | <annevk> | I wonder if this new <dl> <div> proposal will suffer the same faith down the line |
| 09:28 | Ms2ger | wonders if anyone's got a mapping {entry, incumbent, current, relevant} -> Gecko concepts |
| 09:30 | <annevk> | Ms2ger: I wouldn't quite start implementing those in Servo yet |
| 09:30 | <Ms2ger> | Still in flux? |
| 09:30 | <annevk> | Ms2ger: WebKit/Chrome don't have incumbent I believe, so this might well change still |
| 09:30 | <annevk> | Ms2ger: entry/current/relevant are mostly stable I'd say though |
| 10:30 | <MikeSmith> | annevk: if you haven’t already manually removed your local non-published git branches, try this to first get a list of those: |
| 10:30 | <MikeSmith> | for branch in $(git branch -vv | cut -c 3- | awk '$3 !~/\[/ { print $1 }'); do if [[ -z "$(git ls-remote --heads origin $branch)" ]]; then echo $branch; fi; done |
| 10:30 | <zcorpan> | <p><span><ul> doesn't work; the p is still closed |
| 10:31 | <MikeSmith> | annevk: oh maybe make sure ot run "get remote prune origin" before you run that |
| 10:31 | <zcorpan> | <p><applet><ul> works though :-D |
| 10:32 | <MikeSmith> | annevk: and if that lists the branches you expect it to llist, then replace the "echo $branch" with "git branch -D $branch" and run that |
| 11:43 | <smaug> | annevk: uh, I'm missing some news. How will WebSocket be obsolet? |
| 11:43 | <smaug> | e |
| 11:44 | <annevk> | smaug: you can do the same thing with Fetch + Streams and work over HTTP/2 |
| 11:44 | <annevk> | smaug: you even get some reduced latency due to lack of handshake |
| 11:44 | <smaug> | I didn't know you can write to the stream you get from Fetch |
| 11:45 | <annevk> | smaug: I'm not saying we'll ever get rid of WebSocket btw, just that going forward it's not a good idea |
| 11:45 | <annevk> | smaug: soonish, not landed just yet |
| 11:46 | <smaug> | annevk: so will fetch have somehow two modes, one for read-only, one read-write ? |
| 11:47 | <annevk> | smaug: you'd just pass request a readable stream as body |
| 11:47 | <annevk> | smaug: the other readable stream would be on the response |
| 11:48 | <smaug> | so you keep the connection open as long as what? |
| 11:48 | <smaug> | until the request stream is closed? |
| 11:48 | <smaug> | and response is closed too? |
| 11:48 | <annevk> | smaug: yeah, as long as either stream is open |
| 11:51 | <smaug> | annevk: how will progress events like things work with this setup? |
| 11:51 | <smaug> | (which fetch doesn't have yet, right?) |
| 11:53 | smaug | wishes he could read Stream spec |
| 11:53 | <annevk> | smaug: we don't have a good idea around progress events yet, other than actually using the stream directly (which is more CPU-intensive) |
| 11:54 | <annevk> | smaug: part of the discussion takes place here: https://github.com/yutakahirano/fetch-with-streams/issues |
| 11:55 | <annevk> | smaug: with the Streams Standard, if you see the [[internalSlots]] things as private C++ fields and the rest as JavaScript accessors around that, is it still hard? |
| 11:55 | <annevk> | smaug: it can't be just a lack of IDL that's the problem right? |
| 11:55 | <smaug> | idl eases readability a lot |
| 11:56 | <smaug> | usually one just needs to see the idl to get the big picture of how the API works |
| 11:56 | <annevk> | I don't disagree, but it has sketches that are fairly similar to that |
| 11:56 | <smaug> | with Streams...that doesn't happen |
| 11:56 | <annevk> | E.g., https://streams.spec.whatwg.org/#rs-class |
| 11:56 | <annevk> | What it calls "class definition" is basically IDL |
| 11:57 | <smaug> | that doesn't tell what is being returned for example |
| 11:57 | <smaug> | which is rather important to understand what happens |
| 11:57 | <annevk> | That is a good point |
| 11:57 | <smaug> | or what the params are |
| 11:57 | <annevk> | Domenic: ^^ |
| 11:58 | <smaug> | pipeTo. It takes some dest? Is that another stream or some reader or what? |
| 11:59 | <smaug> | (just as an example) |
| 12:04 | <smaug> | annevk: what does ? mean in "Let typeString be ? ToString(type)." |
| 12:05 | <annevk> | smaug: rethrow the exception |
| 12:06 | <annevk> | smaug: I tried to express your "class description" concerns in https://github.com/whatwg/streams/issues/45#issuecomment-255363252 |
| 12:07 | <smaug> | thanks |
| 12:08 | <smaug> | that ? stuff is such that webidl nicely hides from the reader, and from spec writer. |
| 12:08 | <annevk> | smaug: not really, we mostly have to write "Rethrow the exception" in a bunch of places |
| 12:09 | <annevk> | smaug: unless you're only talking about initial argument conversion |
| 12:09 | <smaug> | not when passing parameters |
| 12:09 | <smaug> | you get the type validation for free from webidl |
| 12:09 | <annevk> | sure, but that's not all ? is good for of course |
| 12:09 | <smaug> | sure. But how often do we need to explicitly throw stuff elsewhere? |
| 12:10 | <smaug> | comparing to how common it is to pass arguments |
| 12:12 | <annevk> | smaug: for some good news: https://github.com/whatwg/html/pull/1944 |
| 12:13 | <smaug> | annevk: btw, I like the idea using fetch for two way communication, but while designing that, please keep the progress tracking in mind |
| 12:14 | <smaug> | (I like the approach even if it makes the two way communication less trivial than what WebSocket API has. I guess we create something on top of fetch to make it less annoying to create all the needed objects for reading and writing) |
| 12:18 | <annevk> | Someone could literally write WebSocket over Fetch |
| 12:27 | <smaug> | well, the API part |
| 12:27 | <smaug> | what I'm hoping we provide good APIs in the platform too |
| 12:27 | <smaug> | so that one doesn't need to rely on random script libraries to get simple things done easily |
| 12:27 | <smaug> | in this case at least |
| 12:27 | <smaug> | in general providing low level stuff is ofc the right thing to do |
| 12:28 | <annevk> | I think the problem we've had so far is that we exposed seemingly simple things in simple ways and then it turned out folks actually wanted to do complicated things |
| 12:28 | <smaug> | (but whatever higher level API might be added, that could happen way later once we know how the fetch is being used) |
| 12:29 | <annevk> | So the new approach is to add APIs that enable complicated things and then maybe add simple APIs on top later once we're confident they hit the 80/20 mark |
| 12:29 | <annevk> | Although a whole lot of fetch() so far is managing to do both |
| 12:29 | <annevk> | Which is kinda nice |
| 12:33 | <smaug> | annevk: how will two-way fetch with streams provide messaging style communication? Or will the web app need to split the data from stream and detect what might form one particular message? |
| 12:35 | <annevk> | smaug: the latter |
| 12:36 | <annevk> | smaug: we just expose bytes going to the server and bytes coming from the server |
| 12:37 | <annevk> | smaug: and the initial set of bytes going to and coming from has some semantic wrapper that we call HTTP headers and are exposed as such |
| 12:37 | <annevk> | (ignoring the TLS wrapper here since there's no way we're going to expose that) |
| 12:51 | <annevk> | GitHub is less responsive today for anyone else? |
| 12:52 | <jgraham> | annevk: DDOS on DynDNS |
| 12:57 | <zcorpan> | MikeSmith: https://github.com/whatwg/html/pull/1945 -- i considered rewriting the rules about what to consider part of a group into an algorithm, it seems a bit hard to follow like this with div adding complexity |
| 13:01 | MikeSmith | looks at https://github.com/whatwg/html/pull/1945 |
| 13:06 | <MikeSmith> | zcorpan: on the one hand it seems to me like for stating document-conformance requirements with authors are the intended reader, we probably want to avoid specifying the requirements in the form of an algorithm |
| 13:06 | <MikeSmith> | but this case is maybe exceptional enough that it would actually be appropriate and maybe better |
| 13:06 | <zcorpan> | MikeSmith: the rules i'm referring to are not about the content model, but about "getting the name-value groups" |
| 13:07 | <MikeSmith> | ah yeah true |
| 13:07 | <MikeSmith> | yeah so authors don’t need to care about this anyway, right? |
| 13:07 | <MikeSmith> | or do they for CSS reasons or something? |
| 13:53 | <zcorpan> | nope, they don't. it's in <div w-nodev> |
| 13:54 | <zcorpan> | it's for tools getting the semantic juice out of an html document, should they exist |
| 13:54 | <zcorpan> | i'm also happy to leave it as is for now |
| 14:10 | <nox> | Would it be an improvement if I manage to express https://html.spec.whatwg.org/multipage/syntax.html#markup-declaration-open-state in terms more similar to the 78 other states? |
| 14:13 | <annevk> | nox: yeah, maybe |
| 14:13 | <annevk> | nox: hard to say upfront |
| 14:39 | <nox> | annevk: I don't like that it seems to be the only state that does some sort of look-ahead reading. |
| 16:47 | <rbyers> | annevk, foolip: what is the current status of the URL interop issues you mention. Do we even know who in blink owns this space? Is there a crbug? |
| 16:51 | <annevk> | rbyers: sleevi is helping some, not sure about Blink side |
| 16:51 | <tantek> | so, github |
| 16:51 | <annevk> | rbyers: mostly discussion still about what to do |
| 16:51 | <annevk> | tantek: yup |
| 16:53 | <rbyers> | annevk: Ok. Is there a github issue to refer to? I'm just trying to figure out which blink TL this would fall under (or if I need to do another: "this area appears to be unowned" push). Filing a crbug (eg: "URL is not interoperable" or whatever, with a link to the spec issue) would help because then we have our existing process of "what labels should this |
| 16:53 | <rbyers> | have, then what team is accountable for that label" :-) |
| 16:55 | <annevk> | rbyers: https://github.com/whatwg/url/issues/150 |
| 16:55 | <annevk> | rbyers: https://github.com/whatwg/url/issues/148 |
| 16:55 | <rbyers> | In general, the best way to get anything from the blink project (even if in the end there's no blink code changes necessary) is to file a blink bug. That invokes all our project management discipline (to the extent we have any anyway <grin>). |
| 16:56 | <annevk> | rbyers: and more in that repo |
| 16:56 | <annevk> | rbyers: I have filed URL issues before and the result wasn't great |
| 16:57 | <rbyers> | annevk: ok, sorry to hear that. point me to those issues and I can follow up internally using them as examples. |
| 16:58 | <annevk> | rbyers: will try to remember, supposed to be doing something else now 😊 |
| 16:58 | <rbyers> | (we're definitely not consistent across teams here yet, but it's now part of my job to try to improve that - eg. via better bug discipline, metrics we're tracking around bugs, etc.) |
| 16:58 | <annevk> | rbyers: if you can look up issues filed by me should be easy enough btw |
| 16:59 | <rbyers> | Some teams are very good at interop bug discipline already, so for some issues just filing a bug is likely to have the right result :-) |
| 16:59 | <rbyers> | Ok, I'll search for URL bugs |
| 16:59 | <rbyers> | filed by you |
| 17:41 | <Domenic> | OK, so, what work can I do that doesn't require GitHub... |
| 17:42 | <annevk> | GitHub works here |
| 17:42 | <KiChjang> | the west coast can't access GH |
| 17:42 | <annevk> | Heh |
| 17:43 | <annevk> | Domenic is east coast though |
| 17:43 | <Domenic> | I heard east coast got it worst |
| 17:43 | <annevk> | Domenic: DOM snapshots |
| 17:43 | <Domenic> | hmmm |
| 17:43 | <jgraham> | It was down here not long ago |
| 17:44 | <jgraham> | So it seems to be in and out everywhere |
| 17:44 | <annevk> | Domenic: not sure if that's feasible without GitHub though |
| 17:44 | <Domenic> | Might be. Hmm. |
| 17:44 | <tantek> | Domenic, you can review CSS WG editor's drafts especially those going to CR soon ;) https://drafts.csswg.org/ |
| 17:44 | <KiChjang> | so basically 'murrica is not doing fine |
| 17:44 | <KiChjang> | must be those russians |
| 17:45 | <annevk> | tantek: not a bad idea for any drafts containing APIs |
| 17:45 | <Domenic> | tantek: but where would i file issues |
| 17:46 | <annevk> | Domenic: post-its that you fax to TabAtkins |
| 17:46 | <tantek> | Domenic, on your own website presumably, with your server auto-POSSEing them to www-style and to github once it comes back up |
| 17:46 | <annevk> | Domenic: ooh, print them out marcosc_-style and use a marker |
| 17:47 | <TabAtkins> | It took me a while to realize exactly what level the snark that named me was operating on. |
| 17:47 | <annevk> | TabAtkins: level 3? |
| 17:48 | <annevk> | I guess y'all saw https://twitter.com/bcrypt/status/789510213237895168 already |
| 17:48 | <Domenic> | Ah Twitter is working again |
| 17:49 | <TabAtkins> | And broken for me again. :( |
| 17:49 | <TabAtkins> | goddam nation state attacks |
| 17:49 | <tantek> | TabAtkins, that's ok, you have your own website right? |
| 17:49 | <TabAtkins> | Doesn't help me *read other people's tweets*. |
| 17:50 | <TabAtkins> | Also: I can't tell how many people like my selfies on my own website. |
| 17:50 | <tantek> | TabAtkins: if you can't read them, chances are, they can't write them either |
| 17:50 | <Domenic> | Ooh interesting 192.30.253.113 http://github.com 151.101.44.133 http://assets-cdn.github.com |
| 17:50 | <tantek> | no https? |
| 17:50 | <TabAtkins> | I have no idea what point you're trying to make, Tantek. |
| 17:51 | <Domenic> | I actually typed github.com but IRCCloud adding the http:// |
| 17:51 | <Domenic> | Or, hmm, I copied from Twitter which added them. |
| 17:51 | <tantek> | TabAtkins, I can help you with that "how many people like my selfies on my own website" |
| 17:51 | <annevk> | Oh, exciting https://github.com/blog/2267-introducing-github-community-guidelines |
| 17:51 | <Domenic> | Oh but it does seem that those IPs don't match the certificate |
| 17:52 | <annevk> | GitHub basically helping out with CoCs to some extent |
| 17:52 | <KiChjang> | the raw IP of GH doesn't seem to be doing me any good |
| 18:03 | <TabAtkins> | KiChjang: I suspect the GH servers do enough talking between each other using DNS that they can't actually function at the moment. |
| 18:04 | <tantek> | TabAtkins the larger point is that we should all be posting such content on our own web sites first, and then syndicating copies to where "audiences" are like Twitter, GitHub etc. |
| 18:04 | <TabAtkins> | I know your larger point, I'm just confused about how what you're saying relates to my specific problem, which is that I wanted to read someone's else's tweet and can't right now. |
| 18:04 | <tantek> | that way when those silos go down due to Russian attacks, we can keep posting on our own servers, you know, like the internet was designed to be resilient to Russian attacks on silos ;) |
| 18:06 | <tantek> | ah, for multiple non-DNS availability of URLs, you'll have to look at archive.org, possibly Google's cache, and DHTs like IPFS if you want the bleeding edge |
| 18:08 | <gavinc> | I'm not sure how more servers helps with attacks on DNS infrastructure but otherwise generally agree |
| 18:14 | <jgraham> | GH seems to work well enough from here right now for me to create a PR. But not well enough for travis to run. |
| 19:22 | <KiChjang> | Domenic, has XHR been rewritten to make use of promises? |
| 19:22 | <Domenic> | KiChjang: yes, it's called fetch |
| 19:22 | <KiChjang> | oh, the fetch API |
| 19:23 | <KiChjang> | if i understand correctly, fetch is still a relatively new thing, and was only limited to the use case of XHR |
| 19:24 | <KiChjang> | but i think the idea is to use fetch everywhere in places where a network call is required |
| 19:25 | <KiChjang> | e.g. loading images, <script>, <link> and so on |
| 19:25 | <KiChjang> | even hyperlinks |
| 19:25 | <Domenic> | There's a difference between the the process of fetching and the fetch API |
| 19:25 | <Domenic> | loading images, <script>, <link>, XHR, and fetch() (the fetch API) all use the process of fetching |
| 19:26 | <KiChjang> | was fetching only spec'd for XHR before though? |
| 19:26 | <Domenic> | Not really |
| 19:26 | <Domenic> | It was specced for everything, spread throughout various places |
| 19:27 | <Domenic> | HTML had something vaguely fetch-like for a while |
| 19:27 | <Domenic> | Other places just waved their hands |
| 19:44 | <tobie> | Grr. Can't push to GH but navigate to the website. How can I share DNS caches between these? |
| 19:55 | <tantek> | hey TabAtkins re *read other people's tweets* - try looking up their URLs on Bing - worked for me with https://twitter.com/bcrypt/status/789510213237895168 |
| 19:55 | <tantek> | http://www.bing.com/search?q=https%3A%2F%2Ftwitter.com%2Fbcrypt%2Fstatus%2F789510213237895168&go=Submit+Query&qs=ds&form=QBLH |
| 19:56 | <KiChjang> | tobie, you might wanna add the IP for github to your hostfile |
| 19:57 | <KiChjang> | skip DNS lookup altogether |
| 19:58 | <KiChjang> | 192.30.253.113 github.com |
| 19:58 | <KiChjang> | 151.101.4.133 assets-cdn.github.com |
| 19:58 | <KiChjang> | but it should be fine by now |
| 19:58 | <KiChjang> | as in, the DNS service for GH should be back up |
| 20:03 | <tantek> | GH DNS still down here |
| 20:03 | <KiChjang> | i hear there's another DDoS attack on the west coast |
| 20:06 | <tantek> | we're on at least #3 for the day |
| 20:09 | <Domenic> | TabAtkins: do you know if <dfn data-export>auxiliary browsing contexts</dfn> will make it into shepherd as "auxiliary browsing context"? or just "auxiliary browsing contexts" (plural) |
| 20:10 | <TabAtkins> | The plural, tho Bikeshed's variation handling will let the singular link to that. (It doesn't handle the transitive closure of variations, tho, so it won't link "auxiliary browsing context's" to the plural.) |
| 20:10 | <Domenic> | so i should probably put data-lt="singular", basically |
| 20:11 | <TabAtkins> | Ideally, yes. |
| 20:11 | <TabAtkins> | (Or phrase the sentence such that the term appears in its base form.) |
| 20:19 | <tobie> | ty KiChjang. Still down here in Switzerland. |
| 20:22 | <annevk> | tobie: maybe close to Geneva 😛 |
| 20:22 | <annevk> | Has been working fine here |
| 20:22 | <tobie> | annevk: you're ok? |
| 20:22 | <annevk> | Bit slow now and then, mostly okay |
| 20:22 | <tobie> | annevk: could it be ISP related? |
| 20:23 | <annevk> | No idea |
| 20:35 | <tantek> | GH seems up now, here on the West coast at least |
| 21:19 | <TabAtkins> | annevk, Domenic: Periodic reminder that both HTML and DOM define case-(in)sensitive, and it would be great if only one of them did. |
| 21:20 | <tantek> | TabAtkins: and CSS too right? |
| 21:21 | <TabAtkins> | No, CSS now links it. |
| 21:21 | <tantek> | to which? |
| 21:21 | <TabAtkins> | Whichever one is arbitrarily chosen due to Python's dict ordering. |
| 21:21 | <TabAtkins> | At the moment, looks like HTML. |
| 21:22 | <tantek> | anyone here have any opinions on http://www.w3.org/TR/presentation-api/ or http://www.w3.org/TR/remote-playback/ ? |
| 21:23 | <Domenic> | TabAtkins: did you ever file an issue on that? |
| 21:23 | <TabAtkins> | Not yet, I don't think. |
| 21:32 | <jyasskin> | tantek: What sort of opinions are you looking for? I might be able to get Mark Foltz to appear. |
| 21:33 | <tantek> | jyasskin: very coarse. It seems like the work is proceeding reasonably and is reasonably minimally scoped. |
| 22:23 | <gsnedders> | TabAtkins: wait, BS just arbitrarily picks one when multiple definitions of a term exist? that sounds like it could end badly |
| 22:24 | <jyasskin> | We shouldn't define terms multiple times. |
| 22:25 | <jyasskin> | BS does warn when it finds multiple definitions. |
| 22:58 | <gsnedders> | jyasskin: yeah, I agree we shouldn't, but I think it's a touch too easy to accidentally end up with them |