06:23
<annevk>
zcorpan: that all image fetching logic needs to be duplicated is...
06:23
<zcorpan>
yeah :-(
06:26
<zcorpan>
filed an issue
07:43
<zcorpan>
are people going to tpac?
07:43
<annevk>
zcorpan: I am
07:43
<annevk>
Might be the last time, we'll see
07:44
<zcorpan>
ok
07:44
<annevk>
I think MikeSmith is also going
07:44
<botie>
okay, annevk.
07:45
<annevk>
botie: yes
07:45
<botie>
annevk: excuse me?
08:05
<zcorpan>
annevk: what are you attending on thursday and friday? :-)
08:07
<zcorpan>
hmm gsnedders got a room for Improving browser interoperability brainstorm
08:08
<annevk>
zcorpan: Web Platform maybe, likely WebAppSec
08:08
<annevk>
maybe Web Performance
08:09
<annevk>
but I also hope that we can figure out some unconf track
08:19
<zcorpan>
i think i'll select "other" on thursday and friday and hang around in gsnedders' room, and maybe pop in some meeting
08:22
<annevk>
maybe I should try bring my PS4 for gsnedders' room
08:23
<annevk>
also, we should take a (half) day somewhere to walk around Lisbon
08:23
<annevk>
if there's going to be a lot of #whatwg folks that'd be a nice way to socialize for a bit
08:35
<zcorpan>
yes certainly
08:37
<zcorpan>
i'll also most likely be running at least once, in case somebody here is reading this and also likes to run
09:22
<MikeSmith>
yeah I’ll be at TPAC for sure
09:22
<MikeSmith>
glad to know zcorpan is going
10:44
<zcorpan>
hehe "... we find the size of the response as 14600 − 529 − 26 − 708 = 13337 bytes." http://papers.mathyvanhoef.com/blackhat2016.pdf
10:44
<jgraham>
I am also going to TPAC for some reason. although maybe the unofficial testing meeting will be useful
10:44
<botie>
okay, jgraham.
10:45
<jgraham>
Did botie suddenly become very sarcastic?
10:45
<zcorpan>
maybe is botie taking notes of who are going to TPAC
11:26
<MikeSmith>
jgraham too
11:26
<MikeSmith>
getting the band back together
11:29
<MikeSmith>
caitp: fun V8 bug that nobody has claimed yet https://bugs.chromium.org/p/chromium/issues/detail?id=627309&desc=2
11:40
<caitp>
MikeSmith: I suggest CCing toon or jochen, I don't have editbugs on the main bug tracker, can't do it myself
11:41
<caitp>
I imagine they would have views on changing the contract of api methods in v8, though
11:42
<MikeSmith>
caitp: I don’t have editbugs either, but jochen__ is around here when online so will see this cc yhirano_
11:42
<caitp>
ah, right
13:09
<zcorpan>
MikeSmith: about https://github.com/whatwg/html/issues/709 my hunch is that it's marginal benefit to be checking these characters, and non-trivial to get it right. should we drop it?
13:14
MikeSmith
looks
13:15
<MikeSmith>
zcorpan: yes
13:15
<MikeSmith>
I looked at that a lot and talked to Richard and yeah it seems very hard to get right
13:16
<MikeSmith>
and I do not look forward to needing to implement it :)
13:22
<zcorpan>
ok
13:32
<zcorpan>
PR at https://github.com/whatwg/html/pull/1645
14:00
<annevk>
Sebmaster: potential URL parsing bug in whatwg-url: https://github.com/whatwg/url/issues/139
14:01
<annevk>
Sebmaster: either that or I'm missing something
14:01
<Sebmaster>
yeah I saw, he already created an issue in our repo
14:01
<annevk>
which is likely
14:01
<annevk>
okay
14:01
<Sebmaster>
i got a bit of time, ill check it out
14:07
<Sebmaster>
hm, works for me
16:27
<smaug____>
anyone with Edge?
16:34
<smaug____>
nm
17:35
<TabAtkins>
Aaaaaaah we might finally have a path to getting rid of the "try layout without scrollbars, and if it overflows, restart with scrollbars" behavior https://github.com/w3c/csswg-drafts/issues/92#issuecomment-238940665
17:38
<Domenic>
I guess I don't understand why what's being discussed there is different than overflow: scroll
18:17
<wanderview>
Mek: do you know if chrome canary keeps the service worker alive after I close the last tab with a document of that origin? does it run it in a separate process?
18:19
<wanderview>
Mek: I'm running a test where I repeatedly open a tab, measure its load time, and then close a tab... in chrome release this seems to incur a SW startup penalty... in canary it seems like the SW startup penalty is not there
18:20
<Domenic>
botie: ask rniwa if he has any opinions on https://github.com/w3c/webcomponents/issues/537 ?
18:20
<botie>
will do
18:43
<TabAtkins>
Domenic: overflow:scroll always shows the scrollbar, regardless of whether the content overflows. This is ugly and people don't like it, which is the entire reason overflow:auto exists (and has *way* higher usage).
18:43
<Domenic>
TabAtkins: what would value1 or value2 do differently?
18:43
<TabAtkins>
Domenic: But overflow:auto changes the element's internal geometry based on whether it overflows, which is bad for a bunch of reasons. This proposal is preventing that. You still only *see* a scrollbar when things overflow, but you'll always have that space reserved.
18:44
<TabAtkins>
Domenic: ...exactly what it says? Like, value2 is literally a diff on value1, it describes precisely what its difference is.
18:44
<Domenic>
right, I guess I don't understand the difference between "space reserved" and "shows a scrollbar"
18:44
<Domenic>
reserved for the scrollbar, I assume?
18:44
<TabAtkins>
Yeah.
18:44
<TabAtkins>
You don't paint a scrollbar, but still reserve the geometry for it.
18:45
<Domenic>
What do you paint instead?
18:45
<TabAtkins>
Nothing.
18:45
<TabAtkins>
It's just empty space; you'll see whatever's underneath it.
18:45
<Domenic>
I guess that makes sense
18:45
<TabAtkins>
(Sorry, I'm still really confused about what's confusing you, so I don't know how well I can explain.)
18:46
<Domenic>
I think that was what I was missing
18:48
<Mek>
wanderview: closing all documents is at least nog guaranteed to kill the SW, afaik. Not sure why there would be a difference between stable and canary though.
18:48
<annevk>
Mac-style scrollbars still have this issue?
18:58
<TabAtkins>
annevk: Those are "overlay" scrollbars, and no, they're perfect. But (1) it's an *option* in Mac, and Apple is pretty fervent in maintaining it as such, and strongly opposed to a value that just unilaterally turns on overlay scrollbars.
18:58
<TabAtkins>
And (2) at least on Chrome, internal sentiment is still cool on switching to overlays on non-Mac platforms (especially Windows), as it's a foreign UI affordance.
18:59
<TabAtkins>
So we're trying to get the benefit we want (stable layout size regardless of whether the scrollbar shows up) without necessarily requiring overlay scrollbars.
19:00
<annevk>
I see
19:00
<annevk>
Thanks for the update
19:00
<TabAtkins>
And I'm gonna start some internal conversation about switching our "overflow:auto" behavior to just being the "value1" behavior, as we really hate the complexity of the "jiggle" on our internal code, and it would let us drop a big chunk of complexity from Houdini Layout.
19:01
<TabAtkins>
(Author-facing complexity, as right now they have to always be checking to see if they've overflowed the bounds and need to ask their parent to restart layout for them with scrollbars inserted.)
22:20
<botie>
rniwa, at 2016-08-10 18:20 UTC, Domenic said: if he has any opinions on https://github.com/w3c/webcomponents/issues/537 ?
22:21
<rniwa>
botie: my opinion is that the change is not web compatible
22:21
<botie>
OK, rniwa.
22:21
<rniwa>
botie: so unless someone ships with that change, i don't think we should change the spec
22:21
<botie>
rniwa: huh?
22:22
<rniwa>
botie: oh, sorry, I mistook with another one related to doument.write
22:22
<botie>
rniwa: excuse me?
22:22
<rniwa>
botie: I think throwing is okay
22:22
<botie>
OK, rniwa.
23:01
<Domenic>
Anyone know how to disable Chrome devtools SSL checks while running service worker web platform tests?
23:43
<MikeSmith>
Domenic: --disable-web-security doesn’t do it?
23:44
<Domenic>
MikeSmith: yeah I found the magic incantation in https://bugs.chromium.org/p/chromium/issues/detail?id=561820#c29
23:44
MikeSmith
looks
23:44
<MikeSmith>
heh
23:45
<MikeSmith>
gotta love that name
23:45
<MikeSmith>
--unsafely-treat-insecure-origin-as-secure
23:45
<Domenic>
I kind of thought this was supposed to work out of the box
23:45
<MikeSmith>
how so?
23:45
<Domenic>
I thought localhost was treated as secure out of the box in Chrome
23:46
<Domenic>
Maybe only for http://localhost:80 instead of https://localhost:80443
23:46
<MikeSmith>
ah
23:46
<MikeSmith>
hmm yeah maybe so
23:46
<Domenic>
*8443
23:46
<MikeSmith>
maybe only only for default ports
23:47
<MikeSmith>
--ignore-certificate-errors should have a better name. The similar flag that hsivonen coined for the HTML checker is called --promiscuous-ssl
23:48
<MikeSmith>
Domenic: nice to see the SW tests rolling into wpt
23:50
<Domenic>
They told me the only way I could continue my anti-entry/incumbent crusade was to write WPTs :P. So I did!!