02:14
<GPHemsley>
SimonSapin: Maps everything to the sad-face emoticon
03:21
<MikeSmith>
SimonSapin: for those they're all using code points in the "user-defined" ranges of Shift-JIS
03:22
<MikeSmith>
SimonSapin: that is, the ranges in Shift-JIS that were intentionally set aside for private use, and not for interchange
03:26
<MikeSmith>
SimonSapin: so if the Encoding spec already says anything specific about how to process Shift-JIS code points in those "user defined" ranges, then I'd think the spec already has them covered
03:27
<MikeSmith>
SimonSapin: I wouldn't think the Encoding spec should do anything much beyond that -- certainly it shouldn't specify that they must be mapped to something else
03:29
<MikeSmith>
SimonSapin: because, e.g., UAs running within some system that consumes them need for them to stay just as what that system expects
03:30
<MikeSmith>
SimonSapin: e.g., at least for e-mail, some user sending a message within a Docomo network to another Docomo user needs for them to not be changed in between
03:32
<MikeSmith>
SimonSapin: it's a different thing of course for Web documents -- but authors should not be using those Shift-JIS-variant emoji in Shift-JIS Web documents anyway -- because they already can't expect any kind of interchange if they do
03:37
<MikeSmith>
SimonSapin: anyway, I think that at this point attempting to change any of the Web UA handling of those characters will just risk breaking existing (legacy) content
06:22
<MikeSmith>
the styling of https://wiki.whatwg.org/wiki/HTML_derivatives seems broken
06:23
<MikeSmith>
e.g., not login link at the top
06:23
<MikeSmith>
oh I guess it's the whole wiki that's borked styling
06:35
<tantek>
MikeSmith, FYI: http://lists.w3.org/Archives/Public/public-html/2014Sep/0061.html
06:55
<annevk>
SimonSapin: yup, it has the same PUA mapping they require
06:55
<annevk>
Hixie: I thought the idea was to wait with that bug for a couple of months
08:18
<foolip>
Hixie: sorry I'm slow to react to IRC, I've commented on the two bugs you pinged about
08:18
<MikeSmith>
tantek: thanks
08:19
<annevk>
JakeA: response.clone() in Chrome, it copies each field of the response and sets body to a tee?
08:19
<annevk>
JakeA: I haven't really figured out what the best way to define it is, without enumerating all request/response fields, which is going to be a maintenance hassle
08:20
<JakeA>
annevk: I haven't dug into the implementation, but in terms of the body, doesn't the original request also need to become a tee?
08:21
<annevk>
JakeA: request's body, yes
08:21
<JakeA>
annevk: gotcha. Not sure what we're doing with the other properties, I assume they're copies
08:23
<MikeSmith>
annevk: is there a plan to un-bork the mixed-content situation with the wiki styling?
08:24
<MikeSmith>
whatwg wiki
08:27
<annevk>
MikeSmith: GPHemsley should know
08:27
<annevk>
MikeSmith: I didn't notice a problem for the TLS page
08:27
<annevk>
MikeSmith: is it only cached pages?
08:27
<MikeSmith>
seems to be everything
08:27
<MikeSmith>
ah wait
08:27
<annevk>
I don't get a mixed content warning for that page
08:28
<MikeSmith>
https://wiki.whatwg.org/wiki/HTML_derivatives
08:28
<MikeSmith>
annevk: what about that one
08:28
<annevk>
MikeSmith: when I'm logged out I do
08:29
<MikeSmith>
ah yeah
08:29
<MikeSmith>
I'm logged out
08:29
<MikeSmith>
weird that it matters whether you're logged in or not
08:30
<annevk>
MikeSmith: old pages are cached with the wrong link for the style sheet and such
08:30
<annevk>
MikeSmith: but only cached for users that are logged out
08:30
<MikeSmith>
ah
08:30
<annevk>
E.g. FAQ had this problem, but now https://wiki.whatwg.org/wiki/FAQ renders fine, even when logged out, because we updated it recently
08:31
<MikeSmith>
I see
08:31
<annevk>
I don't know why GPHemsley can't clear the cache
09:54
<smaug____>
++foolip
09:54
<smaug____>
"sync XHR one could artificially make it slower to the point where people will notice."
09:54
<smaug____>
mean, but might work
10:03
<zcorpan>
increase the delay with half a millisecond per day
10:03
<annevk>
JakeA: beverloo: ETA on notifications in service workers email?
10:04
<GPHemsley>
annevk: Because I'm having trouble logging in. Once I can do that, I'll clear the cache and upgrade and whatnot.
10:04
JakeA
reads
10:05
<annevk>
GPHemsley: logging in to the server?
10:05
<GPHemsley>
yes
10:05
<annevk>
GPHemsley: you can ask Hixie to reset your credentials
10:05
<GPHemsley>
I have
10:06
<annevk>
GPHemsley: and you're using the correct username? (always trips me up with SSH)
10:06
<GPHemsley>
As far as we've been able to determine.
10:07
<annevk>
odd
10:13
<MikeSmith>
anybody know if there is yet a web-based feed reader that's actually usable in mobile browsers
10:14
<MikeSmith>
unlike, say, the web version of feedly, which is pretty much unusable on mobile
10:19
<annevk>
I'm using newsblur.com at the moment, but I haven't checked out the mobile version since I read enough stuff on mobile as it is
10:20
<jgraham>
MikeSmith: "feed reader"? Is that like some way to check your Facebook, or something?
10:27
<annevk>
jgraham: it's how you Twitter on G+
10:27
<annevk>
jgraham: you wouldn't understand
10:28
<annevk>
zcorpan: thanks for reviewing my patch to web-platform-tests!
10:28
<zcorpan>
annevk: welcome
12:37
<zcorpan>
annevk: pls review https://github.com/whatwg/web-apps-tracker/commit/66f7dfa4e32623157aa7c47618e09d621a5be008
12:37
<annevk>
please don't review newlines at end of files
12:38
<annevk>
zcorpan: not sure I'm competent enough to review that regular expression
12:38
<MikeSmith>
jgraham: roc isn't on twitter so I otherwise dunno how to fine out when he posts to his blog
12:38
<zcorpan>
oh my editor does that for me. can remove the newline though
12:39
<zcorpan>
annevk: i tested the regexp
12:39
<annevk>
zcorpan: oh you added a newline?
12:39
<MikeSmith>
I vote for please leave the newline in
12:39
<annevk>
zcorpan: that's fine, my editor works the same, thought it was removed
12:40
<zcorpan>
annevk: yeah i added it
12:41
<zcorpan>
MikeSmith: ended up making a new browser icon
12:41
<MikeSmith>
oh boy
12:42
<MikeSmith>
does it have a skull?
12:42
<zcorpan>
no, but it has a back button
12:43
<MikeSmith>
oh that's actually pretty clever
12:43
<zcorpan>
that is 1x1 CSS px. that has a radial gradient. :-)
12:44
<MikeSmith>
man that just looks like a box
12:44
<MikeSmith>
it's too tiny
12:44
<MikeSmith>
it looks like a placeholder thing, img missing
12:45
<MikeSmith>
but with a white bar
12:46
<annevk>
zcorpan: turns out browsers.svg needed to be without extension
12:46
<annevk>
zcorpan: just committed that
12:46
<zcorpan>
annevk: oops
12:47
<annevk>
zcorpan: some SVG files specify width/height, others don't
12:47
<zcorpan>
MikeSmith: hmm, maybe i can make the toolbar bigger
12:47
<MikeSmith>
christ we need an actual designer in this club
12:49
<annevk>
zcorpan: so is this to reduce the amount of icons?
12:50
<zcorpan>
annevk: Hixie always lumps the browsers together in commit messages so it seemed like it makes sense to use one icon
12:50
<annevk>
maybe we should ask Unicode if they can mint some emojis for us
12:51
<MikeSmith>
zcorpan: maybe could just use the reload icon?
12:56
<zcorpan>
MikeSmith: i don't associate that with a browser really
12:57
<zcorpan>
annevk: pushed an update
13:02
<zcorpan>
Hixie: you can use "b" instead of "giow" in commit messages now
13:17
<MikeSmith>
zcorpan: http://www.famfamfam.com/lab/icons/silk/ might provide some further inspiration
13:17
<MikeSmith>
"Creative Commons Attribution 2.5 License"
13:17
<MikeSmith>
whatever that is
13:18
<MikeSmith>
not cc0
13:19
<zcorpan>
MikeSmith: i've made the toolbar bigger now
13:19
<MikeSmith>
ah ok
13:19
MikeSmith
looks
13:20
<zcorpan>
annevk: can you deploy again? :-)
13:20
<MikeSmith>
ah that's better yeah
13:20
<zcorpan>
MikeSmith: great
13:22
<MikeSmith>
hmm but when I look at https://html5.org/tools/icons/browsers it doesn't look different
13:22
<MikeSmith>
not that it really matters what it looks like big
13:23
<zcorpan>
MikeSmith: that's the old version
13:23
<MikeSmith>
oh ok
13:27
<MikeSmith>
anyway for future reference as far as other symbol ideas for representing a browser, http://www.famfamfam.com/lab/icons/silk/icons/application_view_tile.png kind of looks like speed dial (or whatever other browsers call the similar thign they have)
14:56
<Domenic>
JakeA: annevk: +1 for keeping close for the use case JakeA outlined.
15:50
<hsivonen>
I like the way https://w3c.github.io/webappsec/specs/subresourceintegrity/ defines "MIME type".
15:51
<hsivonen>
I wonder how many opposing comments the editors end up getting over an unorthodox definition like that.
15:53
<darobin>
hsivonen: lol, that's brilliant
16:32
<Domenic>
We need to come up with a living dead standard to publish on halloween
16:32
<Domenic>
something about zombie service workers
16:33
<Domenic>
maybe resurrecting finalized objects or something
16:58
<Hixie>
annevk: wait with which bug?
17:19
<tantek>
for any future dated WHATWG snapshots, I propose two things in combination: 1) a shorter title, without reference to any specific constituents. AND 2) for the first sentence to make it clear Implementers should reference the living specification.
17:21
<Ms2ger>
For any such snapshots, I propose a small script that makes all necessary alterations
17:21
<tantek>
Ms2ger - as long as it's server-side. As people do View Source without scripts running on the client.
17:22
<tantek>
And yeah, automating that on the server via a script is a good suggestion
17:22
<Ms2ger>
I was thinking generate in python and upload the result
17:22
<Ms2ger>
But any way that doesn't involve me doing anything :)
17:22
<tantek>
SGTM
17:23
<tantek>
e.g. right now the dated URL spec snapshot has the first sentence "For the current version of the URL Living Standard, including significant errata to the contents of this specification, please see: http://url.spec.whatwg.org/";
17:23
<tantek>
whereas I'd suggest something more like this for the first sentence: "Implementors should reference the URL Living Standard, including significant errata to the contents of this specification: http://url.spec.whatwg.org/";
17:23
<tantek>
so there's a concrete example of a transform
17:29
<Domenic>
I like the existing sentence
17:29
<Domenic>
As ideally *everyone* should reference the living standard
17:31
<MikeSmith>
annevk: you saw http://lists.w3.org/Archives/Public/uri/2014Sep/0002.html ?
17:31
<MikeSmith>
about file:
17:32
<Domenic>
I have been bugging annevk about file:, and specifically file:///, for probably a year now :)
17:35
<Hixie>
yeah the only people who should reference the snapshots are lawyers, for patent purposes, since the snapshot is what the patent license covers, and government officials who for reasons beyond my understanding are required to reference unchanging documents in their references, and where it would be illegal to do otherwise.
17:36
<Hixie>
everyone else should just use the living standard
17:36
<annevk>
Hixie: the events bug
17:36
<Hixie>
using the snapshot for a reference e.g. from a w3c spec, is a fantastically bad idea.
17:36
<annevk>
zcorpan: will deploy now
17:37
<Hixie>
annevk: what did I do instead of waiting?
17:37
<annevk>
Hixie: you pinged me
17:38
<Hixie>
wait, which events bug. https://www.w3.org/Bugs/Public/show_bug.cgi?id=12230 ?
17:38
<MikeSmith>
Domenic: looks like that guy is actually working On that i-d https://tools.ietf.org/html/draft-kerwin-file-scheme-11
17:38
<annevk>
zcorpan: deployed: https://html5.org/tools/icons/browsers
17:39
<annevk>
zcorpan: I don't get why you would put height/width in the icon
17:39
<annevk>
zcorpan: that seems to defeat the purpose of it being scalable
17:39
<annevk>
Hixie: no, the IDL one
17:40
<annevk>
Domenic: MikeSmith: browser vendors have not been interested in revamping their URL parsers thus far, much less in revamping the file URL part of it I suspect
17:42
<Hixie>
annevk: oh... why are we waiting again? sorry, i'm apparently confused about some of these bugs.
17:42
<annevk>
Hixie: mostly because it was not high priority for you and I have other things to shift through
17:43
<annevk>
Hixie: there's also a thing with event objects that we don't really define the internal slots that well, so we could fix them but it would be a larger effort
17:44
<Hixie>
annevk: k. i'll punt it to 2015 on my side. feel free to bring it closer if you have time for it.
17:45
<Hixie>
can't find the bug
17:45
<Hixie>
do you remember what it's called?
17:46
<annevk>
Hixie: https://www.w3.org/Bugs/Public/show_bug.cgi?id=25099
17:47
<Hixie>
thanks
17:57
<zcorpan_>
annevk: i put it there so they would have the same intrinsic dimensions as the pngs in case someone uses them without giving width height. but i guess it doesn't really matter
18:05
<annevk>
Hixie: the only change to https://whatwg.org/specs/url/2014-07-30/ that would be good would be to change http to https for both <img> references
18:05
<annevk>
Hixie: that would also ensure the images keep loading into the future
18:07
<Hixie>
i'm not changing it. it's not important enough to violate the principle of the snapshot.
18:13
<tantek>
in order to keep with the principle of the snapshot, may I then suggest a new snapshot with the fixes I proposed above, and changing http to https for the <img> references
18:14
<Domenic>
TabAtkins: lol @ bikeshed + git bisect :D
18:15
<TabAtkins>
Why lol?
18:16
<Domenic>
your enthusiasm was funny :)
18:16
<Hixie>
tantek: given that we only got 5 patent commitments to the last one, it seems a bit premature to already be making another patent snapshot. Also, see above regarding why I disagree with the change you're proposing.
18:16
<TabAtkins>
Domenic: Ah, kk. ^_^ I've been meaning to use it anyway, and it looks cool.
18:31
<TabAtkins>
Yay, bisect is great!
18:31
<TabAtkins>
Confirmed that the problem was in the commit I thought it was, so I can look more closely at what's going wrong here.
18:33
<tantek>
Hixie, I'd like to iterate on snapshots to get *more* patent commitments. Including making changes like a shorter title that omits references to specific audiences.
18:35
<Hixie>
tantek: why do you think changing the title would get more companies on board the FSA?
18:35
<tantek>
Hixie, because in-person several representatives (AC even) of companies reacted poorly *just to the title*
18:36
<tantek>
whereas they seemed to otherwise like or be ok with the whole CG way of getting patent commitments etc.
18:36
<Hixie>
who?
18:37
<tantek>
unfortunately it was in the context of a private w3c meeting :( so I don't think I can say who.
18:37
<tantek>
you may accept that as weak data if you wish
18:37
<Hixie>
and did they react poorly because they thought "stale drafts should be referenced by more than lawyers!" or "we would never sign a license to commit patents to a draft made for that purpose!" or what?
18:38
<tantek>
in short it was a purely emotional reaction AFAIK, without any additional actual rational thought
18:38
<tantek>
I realize that may be a human bug but there you have it.
18:38
<Hixie>
that's the entire point of the title
18:38
<Hixie>
to make people not want to read that version of the spec
18:38
<Hixie>
so, working as intended
18:39
<Hixie>
one would assume that a company's lawyers don't, however, base what they will sign or not sign on emotional reactions
18:40
<tantek>
not want to read and not want to patent commit are not the same thing.
18:41
<Hixie>
[24~exactly
18:41
<tantek>
but to get to a company's laywers, typically there are intermediate humans that make decisions
18:41
<tantek>
fortunately not all decisions are made by lawyers :P
18:41
<tantek>
or rather, if that's what you're looking for, you may find exactly that in the "government officials" you refer to ;)
18:42
<Hixie>
if you can find me a company who would commit to signing the license based on nothing more than a title change, let me know.
18:42
<Hixie>
until then, this is just a distraction from more important issues, like actual bug fixing.
18:50
<tantek>
Apologies for the distraction. I'm hoping others can help with silly things like static snapshots for external referencers so you can keep focusing on actual more important issues like bug fixing (agreed about importance).
19:20
<annevk>
tantek: it would help if we had on black and white that it is what is blocking progress I think
19:21
<annevk>
tantek: I'm willing to help out create another snapshot, but to be fair we already published one, it seems kind of weird to put rules on titles after the fact
19:22
<jsbell>
annevk: how close are you to yanking the Errors stuff out of dom? i.e. is IDL covering it all yet? Someone was bugging me that DOM's DOMException IDL fragment doesn't have `name`
19:22
<annevk>
jsbell: that's because it's an exception and not an interface...
19:22
<annevk>
jsbell: I can do it today, but it would help if IDL actually covered it
19:22
<jsbell>
Not urgent at all.
19:23
<annevk>
jsbell: so far nobody updated IDL... there's open bugs and all, but we don't have a person writing IDL full time
19:23
<annevk>
jsbell: if you happen to have a spare six months or so
19:23
<annevk>
(or anyone else in this channel for that matter)
19:24
<jsbell>
I know how much fun others have working on the IDL spec. So tempting. ;-) It looked like there was at least some blather in IDL about DOMException and general exception attributes.
19:24
<jsbell>
Anyway... thanks (as always)
19:51
<tantek>
annevk - no rules, just suggestions for increasing the chance of getting more support, and fewer emotionally negative reactions which aren't helpful to anybody.
19:51
<tantek>
never going to get black and white about this kind of thing - it's a process of iteration. we try something, see what breaks, fix it, iterate.
19:51
<tantek>
the usual
19:52
<TabAtkins>
Domenic: When I said "should be fixed", I of course meant "When I push, which won't happen yet because I just found a bug".
19:53
<annevk>
tantek: understood, but I've the feeling that if we did iterate the same people would find something else
19:54
<tantek>
annevk, that's not my impression. rather, there are *some* who would see iterations as good faith listening / response and be supportive. and unfortunately there are likely others who would not be satisfied by any changes.
19:54
<tantek>
important to distinguish those two sets of people. I have specific examples in mind for each set but have no desire to call out specific names in public (yet)
19:55
<annevk>
no need
19:55
<tantek>
in addition, by addressing the concerns of set 1, they often see the stubbornness of set 2 as *bad faith* and will turn right around and call them out for it
19:58
<annevk>
tantek: if these are in-person meetings btw, did you try telling them they could ask us directly?
19:58
<annevk>
tantek: I don't think either Hixie or I heard about this until it broke on public-html
19:59
<annevk>
(well a bit before, but from other people than proxies I mean)
20:01
<annevk>
tantek: comment-by-proxy is a huge pain in the ass
20:01
<tantek>
annevk - am I not part of "us" in the "ask us directly"? ;)
20:02
<annevk>
tantek: well you didn't publish this doc
20:02
<tantek>
ah, ask the editors/publishers directly
20:02
<tantek>
that's a very reasonable request
20:02
<tantek>
I'll keep asking for such direct communication over comment-by-proxy
20:02
<annevk>
ta
20:02
<tantek>
but I'll continue facilitiating until the communication improves
20:03
<tantek>
more than that I'll probably start a wiki page collecting every little minor complaint anyone has about any such external normative references
20:03
<tantek>
for future guidance for such static snapshots
20:04
<tantek>
I'm giving public-html another 48-72 hours to collect reactions
20:04
<tantek>
treating it kind of like snail-mail, seems to be a reasonable model.
20:06
<annevk>
tantek: I'm happy to make another snapshot, not a big deal to me
20:06
<annevk>
not sure how Hixie feels about it though
20:10
<tantek>
annevk - sounds like Hixie has more important things to focus on, which I'm cool with.
20:11
<annevk>
http://www.economist.com/news/international/21620221-translating-technological-terms-throws-up-some-peculiar-challenges-cookies-caches-and-cows is rather beautiful
20:11
<annevk>
'Ibrahima Sarr, a Senegalese coder, led the translation of Firefox into Fulah, which is spoken by 20m people from Senegal to Nigeria. “Crash” became hookii (a cow falling over but not dying);'
20:11
<Ms2ger>
Nice comparison
20:13
<tantek>
thus causing a crash is cow-tipping?
20:17
<tantek>
annevk - cool - re: happy to make another snapshot, not a big deal to me
20:18
<tantek>
let's give folks some time to dump their current set(s) of objections publicly, I'll collate and document on a wiki and we can try to fix a bunch in a new snapshot.
20:18
<tantek>
I don't want to spend a lot of your or Hixie's time on this. So, fewer iterations with more things fixed in each.
20:19
<annevk>
I wonder if HSTS can be abused for tracking
20:20
<annevk>
E.g. if we offered the network error on redirect feature, you could see whether someone visited a HSTS site before
20:20
<annevk>
Not a huge deal now perhaps, but might soon be
20:29
<Ms2ger>
"CVE-2014-6271 looks like a variant of a problem we fixed ~30 years ago in Research UNIX sh(1)."
20:29
<Ms2ger>
Now if only they'd written tests
20:40
<tantek>
famous last words
20:43
<TabAtkins>
Domenic: Okay, *now* fixed and pushed.
20:53
<annevk>
Hixie: I just cleared cache in a browser and still can't open e.g. https://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-March/018888.html
20:53
<annevk>
that I copied that prolly means the cache is not cleared completely
20:54
<Hixie>
try http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-March/018888.html
20:55
<Hixie>
tantek: if you have a concrete commitment from someone that their patent-rich company will sign the FSA for a snapshot with reasonable changes, then i'm interested.
20:56
<tantek>
Hixie, noted.
20:56
<Hixie>
("reasonable" being the main vague word in that sentence, of course)
20:56
<tantek>
I don't expect to see "concrete" commitments about anything usually. There's more room to make progress than that I think.
20:57
<Hixie>
we had a concrete commitment from three companies before we published the last one, fwiw.
20:57
<Hixie>
(and they followed through)
20:57
<tantek>
good! hopefully no changes to put any of that into jeopardy then.
20:57
<Hixie>
well those commitments are done, so it's moot :-)
21:34
<annevk>
zcorpan++ for https://html.spec.whatwg.org/multipage/introduction.html#abstract
21:34
<annevk>
Hixie: yeah made it work
21:35
<Ms2ger>
zcorpan++
21:36
<Ms2ger>
in general
21:46
<Hixie>
made what work? what?
21:46
<Hixie>
and yeah, zcorpan++ in general
21:47
<Hixie>
oh, the link
21:48
<tantek>
that picture is amazing.
22:30
<Hixie>
hsivonen: ping https://www.w3.org/Bugs/Public/show_bug.cgi?id=26804
22:48
<jwalden>
Hixie: I see the URL in source, but does https://html5.org/r/8821 probably want some sort of inline attribution link, and perhaps an addition to acknowledgments or so?
22:48
<Hixie>
it's already in the acknowledgements, no?
22:48
<jwalden>
oh, I probably missed that
22:49
<Hixie>
nothing changed except that it became a jpeg+svg instead of a png
22:49
<Hixie>
(and a few words changed)
22:49
<jwalden>
ah, right
22:49
<jwalden>
carry on!
22:49
jwalden
figured there was some stupid explanation like that :-)
22:51
<Hixie>
:-)