01:27
<a-ja>
MikeSmith: ping
01:30
<a-ja>
MikeSmith: dunno if you noticed, but styled scope got back out of blink...due to being buggy and no progress in 6+ months. maybe need a warning for <style scoped> again?
01:30
<MikeSmith>
a-ja: dunno maybe
01:30
<MikeSmith>
I thought it was implemented elsewhere
01:31
<MikeSmith>
Firefox
01:31
<a-ja>
MikeSmith: i think it's in parsers....but, beyond that???
01:31
<MikeSmith>
it's implemented at least one other engine
01:32
<MikeSmith>
I think gecko
01:33
<MikeSmith>
plus I think there's also some polyfill for it
01:33
<a-ja>
think you're right
01:34
<MikeSmith>
so my current opninion is that we shouldn't annoy devs with warning messages if they're using polyfills
01:35
<MikeSmith>
at the time when Henri first added those this-is-unimplemented warnings, I think we didn't really have polyfills
01:35
<MikeSmith>
they weren't common at least, and maybe not even the word polyfill had been coined yet
01:35
<MikeSmith>
but the situation's different now
01:39
<a-ja>
i take it's way too early to consider <style nonce> and <script nonce> at this point
01:39
<a-ja>
...as called for by CSP 1.1
01:46
<MikeSmith>
a-ja: I guess the WebAppSec WG would need to file HTML spec bugs to get that added
01:48
<caitp>
do you ever feel like maybe having 900 working groups and 1800 specs isn't the best approach
01:49
<MikeSmith>
a-ja: but anyway @nonce is also already implemented and shipped
01:50
<MikeSmith>
caitp: there aren't that many WGs that matter. Most of them can be safely ignored
02:19
<Hixie>
MikeSmith: for the record, we designed many parts of HTML specifically so that they could be "polyfilled", so at the time we called it "shimmed"
02:19
<Hixie>
s/so/though/
02:31
<roc>
<style scoped> is shipping in Firefox I think
02:32
<a-ja>
roc: specificity probs ever resolved?
02:32
<roc>
dunno, ask heycam|away
02:33
<roc>
yeah, it's actually been shipping for quite some time
02:33
a-ja
needs to search bugs for @scope at-rule support, too
02:43
<MikeSmith>
Hixie: yeah but as far as the validatotor warnings it's just that nobody had actually started to do did much shimming of new features at the time we added those warnings
02:44
<MikeSmith>
Hixie: btw I thought Henri added those initially based on some part of the spec stating guidance for conformance checkers that they should warn about unimplemented features
02:44
<MikeSmith>
but maybe that was my imagination
03:30
<Hixie>
MikeSmith: dunno of such a section, but who knows
05:35
<JakeA>
TabAtkins: I was thinking about https://twitter.com/tabatkins/status/474209390254620672 - what were you thinking there?
05:36
<JakeA>
TabAtkins: I was thinking about https://twitter.com/tabatkins/status/474209390254620672 - what were you thinking there?
07:18
<annevk>
krit: what I'd like is some information as to why my input will now be used to some effect
07:18
<annevk>
krit: is there an active author/editor?
07:21
<zcorpan>
is there a spec for how drag-and-drop within contenteditable should work?
07:29
<krit>
annevk: The behavior ought to be specified in https://svgwg.org/specs/integration/ at some point. The reason this came up again at all was a request from WebAppSec WG. I hope that his gives new impulses to continue working on the spec + me slowly rolling out of other spec work like masking. I believe to find more time next month. However, if the Wiki
07:29
<krit>
progresses faster and the future structure of the document gets more clear, I am happy to do edits earlier. Editors are Cameron and me. You are right that many parts of the discussions haven't been added yet. I will search through the mailing lists next week. If you want to wait to see more progress on the wiki, I can understand it. But you are more than
07:29
<krit>
welcome to add notes to the Wiki sooner.
07:30
<krit>
annevk: the wiki is meant as a white board to collect ideas and thoughts. No need to write spec text there
07:30
<krit>
annevk: if you have any short notes in your mind, just add it
07:30
<krit>
annevk: links are fine as well
07:59
<annevk>
Domenic: "in prose" works if you're Allen or are keeping closely in touch with him, it doesn't really scale beyond that
08:00
<annevk>
krit: that link does not load
08:00
<annevk>
krit: "integration" seems kind of weird, HTML has no such chapter
08:01
<annevk>
krit: and things such as fetching and security are part of the core processing model, you can't just bolt them on somehow
08:01
<krit>
annevk: the purpose of the spec is covering eBooks and other languages including subsets of SVG as well
08:01
<krit>
annevk: :) Do not disagree
08:02
<annevk>
The fact that you guys can't even come up with a sane organization does not inspire confidence
08:02
<annevk>
And suggests nothing much has changed in the intervening years
08:03
<annevk>
These questions are not something you look at separately typically. You should come across these as you define SVG as a language
08:14
<krit>
annevk: SVG as a language has existed for more than 14 years. It is huge, it is complex. That members of the WG don't have as much time as the language deserves is more than unfortunate. But SVG is not the only specification suffering years of stagnation.
08:15
<krit>
Exmples are CSS OM for instance which has had multiple editors as well as far as I can remember, you were one of them.
08:15
<annevk>
That does suggest there's nobody to take on this issue
08:16
<krit>
It doesn't help to blame people instead we should look forward.
08:16
<annevk>
CSSOM has made steady progress. A great many things that was once undefined was defined over time. Not much new features, but the overall model was made much more clear and unambiguous.
08:17
<krit>
annevk: another point is probably the lack of experience in technologies like fetching... at least from the specification point of view.
08:17
<annevk>
As in, we went from HTML4-era level of detail to HTML5-era level of detail. That's pretty great I think.
08:18
<krit>
annevk: and we hope yo reach the same with SVG eventually
08:19
<annevk>
krit: I'm not trying to assign blame, I'm trying to figure out if there's anyone actively working on this and writing things down. We're not going to solve this by jotting notes on a wiki page and having teleconferences.
08:19
<annevk>
krit: This requires someone to put a few months of effort into this.
08:19
<krit>
annevk: honestly, I can not guarantee it. I can just promise that I do my best to keep it rolling.
08:20
<krit>
annevk: I do not have the possibility to work full time on this subject... all that I can offer is to help as much as I can.. even with editing
08:22
<krit>
annevk: I am not sure if heycam|away can, but he has a lot on his plate too. We can not just hand it over to anyone and force a different WG to take care of it. As you said, it is fundamental for SVG. But we can try to find help.
08:24
<zcorpan>
MikeSmith: https://github.com/ResponsiveImagesCG/picture-element/pull/204
08:31
<annevk>
krit: I guess I'm ready to help once there is someone actively working on it. This is not too different from how it was a couple of years ago.
08:31
<annevk>
(I hope heycam|away starts working on IDL. There's a ton of issues there :-()
10:01
<hsivonen>
Hixie, MikeSmith: validator.nu is back.
10:01
<hsivonen>
somehow, the VM's kernel modules have become broken enough that iptables-restore can't do its thing and the port mappings were missing
10:02
<hsivonen>
as a band-aid, I installed nginx to do proxying from one port to another
10:02
<hsivonen>
please let me know if POST or something like that broke
10:02
<hsivonen>
my other VMs don't have this problem
10:05
<hsivonen>
I'm guessing the reason for brokenness is an incomplete kernel update relative to Xen modules
11:50
<MikeSmith>
hsivonen: thanks I'll test and see if I notice any problems
11:50
<MikeSmith>
hsivonen: glad it wasn't my fault at least :-)
12:15
<annevk>
krit: how much time can you get from your manager?
12:15
<annevk>
krit: I'm happy to match your time and see how far we can get together
12:38
<krit>
annevk: I'll check this evening. I am actually close to your timezone CET
12:42
<jabbok>
hi all
12:43
<jabbok>
anybody got tips on using html5 notification api cross browser/platform?
12:46
<annevk>
krit: we're the same timezone these days
12:47
<krit>
annevk: did you move?
12:47
<annevk>
krit: yeah, I'm based in Zürich
12:47
<krit>
annevk: oh, didn't know
12:47
<jabbok>
annevk maybe?
12:48
<krit>
annevk: why did you move to Zurich? London needed the space for Servo?
12:48
<annevk>
jabbok: what's the issue?
12:48
<annevk>
krit: personal reasons, we don't really have an office here
12:48
<jabbok>
well whe’re trying to find a cross browser/platform notification method for ower webapp...
12:49
<krit>
annevk: ok :)
12:49
<jabbok>
maybe using the html5 notification api
12:49
<darobin>
the couldn't stand having him in the office so they sent him to the closest large city with no official presence
12:49
<krit>
annevk: with not real office, do you actually mean no office space at all?
12:50
<jabbok>
is there maybe some kind of javascript wrapper to make this work between browsers?
12:51
<darobin>
jabbok: check https://developer.mozilla.org/en/docs/Web/API/notification for support information and such
12:51
<jabbok>
I’ve been there, and in firefox this works all fine and dandy
12:52
<annevk>
krit: yeah I work from home
12:52
<jabbok>
but if i want to trigger a message with an event (dom ready for example) chrome doesn’t do that so it’s an api issue i think
12:53
<jabbok>
i thought maybe you guys would have suggestions
12:53
<krit>
annevk: well then meeting in Zurich is more difficult... IRC or video conference works too though
12:53
<annevk>
In Europe Mozilla has offices in Berlin, London, and Paris
12:53
<annevk>
krit: ooh, I was thinking meeting over IRC, but just having fulltime availability
12:54
<annevk>
krit: if you want an actual meeting maybe we can get TabAtkins and the Google office involved somehow :p
12:54
<krit>
annevk: k. Will check this evening without budget involved it is always easier to get some time
12:55
<annevk>
jabbok: hmm not sure
12:55
<annevk>
jabbok: I'd suggest stackoverflow.com
12:55
<jabbok>
those issues are all dated
12:56
<annevk>
jabbok: I don't get much further then trying to make sure the spec is okay and then moving to the next undefined thing
12:56
<annevk>
jabbok: but you can ask questions there, no?
12:56
<jabbok>
i’ll try that. Thanks for your time!
12:56
<jabbok>
and good luck on the standards ;)
13:20
<annevk>
JakeA: it seems that if we want SW to just modify the request and then return it (should change from null in that case I think), Fetch needs to restart from the top as the mode might have changed and such, no?
13:23
<annevk>
JakeA: e.g. if mode is same-origin and you change the url of the request, we're in trouble
13:29
<JakeA>
annevk: are we in trouble, or would it just fail?
13:30
<annevk>
JakeA: since step 6 of http://fetch.spec.whatwg.org/#concept-fetch is not currently I think we're in trouble
13:33
<annevk>
JakeA: it would actually help to have a set of use cases for modifying headers and mode
13:34
<annevk>
JakeA: as making sure "no CORS" fetches are <img>/<script>-compatible is hard
13:35
<JakeA>
annevk: if we go read-only, it's not going to stop us opening it up later is it?
13:35
<annevk>
JakeA: no
13:36
<annevk>
JakeA: it would probably still require answering the above question
13:36
<JakeA>
annevk: Let's shut it down. *looks into camera* shut it all down.
13:36
<annevk>
JakeA: e.g. I do think we your header use case to work
13:36
<annevk>
want*
13:37
<JakeA>
hah, can't remember it
13:37
<annevk>
JakeA: https://gist.github.com/jakearchibald/9f556ce0eafc47ba3d6e
13:38
<JakeA>
annevk: I've heard use-cases around tagging requests with priorities. Being able to polyfill headers like clienthints would be nice too.
13:39
<JakeA>
annevk: But you could do that with respondWith(fetch(event.request, modifications))
13:39
<JakeA>
annevk: except for the redirect issues
13:40
<JakeA>
annevk: For priorities we need to see how they'll be represented on the request object
13:41
<annevk>
JakeA: yeah, maybe that is better
13:42
<annevk>
JakeA: the modify existing request and then return it is kind of weird
13:42
<annevk>
JakeA: it'd be a new "default"
13:46
<JakeA>
annevk: I think it feels intuitive to modify the request object like that, but agree it's too complex in terms of the implementation for a first-pass
13:47
<annevk>
JakeA: so even if we drop that, if you do new Request() should you be able to set .mode?
13:47
<annevk>
JakeA: it's kind of unrelated whether we take https://gist.github.com/jakearchibald/9f556ce0eafc47ba3d6e into account
13:47
<JakeA>
annevk: Oh yeah, definately. Else how do I cache the JS on my CDN?
13:48
<annevk>
JakeA: well should be able to set Request.prototype.mode, of course you can set it through new Request()
13:48
<JakeA>
annevk: or do you mean set it *after* construction?
13:48
<annevk>
Yup
13:49
<JakeA>
annevk: I say lock it all down. We can open stuff up as we work it out & get good usecases
13:49
<annevk>
Okay
14:50
<annevk>
People who reply to GitHub via email, grr
15:10
<caitp>
whens the last time you tried to reply to github from your phone any other way? it's basically impossible :(
19:03
<zewt>
it should probably be possible to just reply to github notification emails and have it get translated into a reply (but I've never tried it)
20:07
<Hixie>
why does an argument with a default have to be explicitly marked as optional. isn't it obvious it's optional if it has a default? sigh.
20:11
<gsnedders>
Hmm, Google+ photo upload gave no choice of circles as far as I could see, and can't change it after upload? That's pretty screwed up.
20:20
<Hixie>
"Google+ photo upload"?
20:20
<Hixie>
you mean the autobackup feature?
20:20
<Hixie>
that just backs up to your account privately, it doesn't share it
20:20
<gsnedders>
nah, from the Android gallery app
20:20
<gsnedders>
the Google+ share, which is provided by the Google+ app, no?
20:20
<Hixie>
Gallery -> Share -> Google+ should give you an ACL editor
20:20
<gsnedders>
I couldn't see it anywhere, eh
20:21
<Hixie>
at the very top of the share dialog, there's an icon, a circle name, and a drop-down
20:21
<gsnedders>
Not here?
20:22
<Hixie>
show a screenshot?
20:23
<gsnedders>
It simply says "Geoffrey Sneddon\nTo Your Circles" with the dropdown, with reshare/comment options. Seemingly, though it looks like a mere statement, you can change who its shared to be pressing on the "to x" bit.
20:23
<gsnedders>
Which wasn't obvious to me at all
20:23
<Hixie>
ah well i'm not going to make any arguments about how intuitive it is
20:23
<Hixie>
:-)
20:24
<Hixie>
you might be on a different version than i have
20:24
<gsnedders>
oh well, then I guess only people in my circles can read that wonderful user manual.
20:25
<gsnedders>
heh, hadn't even noticed the "Please strongly recommend reading the MANUAL before using the product". Who am I meant to recommend this to?
20:27
<jgraham>
gsnedders: If it helps you can recommend that I read the manual. Then you can use the product and I can ignore your recommendations, like normal :)
20:58
<gsnedders>
jgraham: as the manual says, "You can do everything u like as using a disk driver"
21:04
<Hixie>
yeah it's not like this was even just a matter of a bad automatic translation
21:04
<Hixie>
it's more that they hired a high school student as an intern to write the documentation
21:05
<Hixie>
and did no review
21:06
<gsnedders>
Interestingly, the manufacturer, CiT, appears to have no website.
21:07
<gsnedders>
On the other hand, it cost almost nothing and has good reviews