00:02
<zewt>
gah
00:02
<zewt>
alright, can anyone give me a plausible explanation for this which has driven me insane for years: "if your download doesn't start in 15 seconds..." and a redirect to download links, instead of plain standard old download links, heh
00:03
<zewt>
so broken and it's hard to think up even an implausible explanation
00:05
<TabAtkins>
Those types of things typically use an iframe pointing to the file with appropriate headers to trigger an automatic download prompt.
00:05
<zewt>
which is the same thing you'll get with a plain download link
00:05
<TabAtkins>
Not necessarily, if it's a type of file that the browser thinks it can handle itself.
00:05
<TabAtkins>
Also: more branding!
00:05
<zewt>
that's what content-disposition: attachment is for (iirc)
00:06
<zewt>
and it's so broken--breaks browser copy-link, and in the case of the Firefox upgrade I just did, it happily pops up another download prompt after I restart the session (since the download page is still open)
00:08
<TabAtkins>
Yup, it really shouldn't be necessary anymore.
01:31
<zewt>
missing messages in the archive? http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/029492.html should be http://www.mail-archive.com/whatwg⊙lwo/msg24505.html
01:48
<Hixie>
zewt: weird
02:06
<zewt>
looks like the list indexes are broken when the trailing slash is missing from the URL
02:06
<zewt>
http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2011-February vs http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2011-February/
02:06
<Hixie>
not much i can do about it, it's a dreamhost thing
06:52
<abarth>
Hixie: is there an easy what to see which rev you removed <device> in?
06:52
<abarth>
i don't see it on http://twitter.com/whatwg
06:53
<zcorpan>
it was http://html5.org/r/5945
06:55
<abarth>
yhx
07:12
<zcorpan>
Hixie: <p><b>.... <p><b>... <p><b>... won't nest the <p>s! don't you know the parsing algorithm by heart? :)
08:49
<hsivonen>
I guess I should do community service and draw a flowchart now that IE9 is final...
09:24
<hsivonen>
Am I reading the CSS 2.1 spec under /TR/ wrong or is its "Features at risk" out of sync with the actual spec requirements as far as the HTML body background rule goes?
09:25
<hsivonen>
it says the there's the risk that the exception might be extended to XHTML, but AFAICT, it has already been extended in the spec
09:25
<hsivonen>
and implemented across the top 4 engines, too
09:25
<zcorpan>
yeah, it's out of sync
09:25
<hsivonen>
are there now any CSS differences between XML and HTML Standards Mode?
09:25
<zcorpan>
i pointed that out when the spec was changed a few years ago
09:26
<zcorpan>
i think only in Selectors case-sensitivity
09:26
<hsivonen>
zcorpan: thanks.
10:26
<hsivonen>
Findings so far: IE9 sticks to the IE8 notion of what's quirky. Doesn't align with HTML5 and other browsers.
10:31
<zcorpan>
also for comment before doctype?
10:34
<hsivonen>
zcorpan: didn't test that yet
10:51
smaug____
doesn't understand how the new Stream + video is supposed to work. In what format is the Blob created from Stream? Do all the browsers need to support that?
11:22
<Lachy>
smaug____, the spec says "Generate a file that containing data in a format supported by the user agent for use in audio and video elements."
11:23
<Lachy>
so it seems that browsers could save the file as WebM (VP8/Vorbis) or MP4 (h.264/AAC) or whatever they choose
11:24
<smaug____>
so the whole PeerConnection would work well only it both ends support same codecs
11:24
<Lachy>
that just means that if the website wants to accept those file uploads, then the servers are going to need to have support for both formats
11:24
<Lachy>
the P2P stuff is more complicated, as you will only be able to stream between browsers that support the same formats
11:24
<Lachy>
right
11:25
<smaug____>
That is bad
11:25
<smaug____>
leaving a major point of the feature unspecified
11:25
<Lachy>
yeah, it sucks big time. We just need to make sure we can make WebM successful for that purpose
11:25
<smaug____>
(which is ofc the case with <video> too)
11:26
<jgraham>
smaug____: Yeah, it does suck. Not sure what the alternative is though :(
11:26
<Lachy>
but we haven't implemented that stuff yet. Only what we announced yesterday, which was initial support for <device> and the ability to play that in a <video> element. We should be able to change the new API easily though.
11:27
<Lachy>
I mean, change *to* the new API easily
11:27
<smaug____>
I'm not sure I like the PeerConnection anyway
11:28
<smaug____>
the whole thing reminds a lot like WebSockets (with support for Streams), but with different protocols
11:30
<jgraham>
Isn't being like websockets a good thing?
11:30
<jgraham>
Or are you saying that we should reuse websockets?
11:31
<smaug____>
we could perhaps have just one API
11:31
<smaug____>
with support for different protocols
11:33
<jgraham>
I wonder if Hixie has looked at the design of 0mq
11:33
<jgraham>
It does multiple protocols / connection topologies with more or less a single API
11:34
<jgraham>
AFAIK
11:34
jgraham
has never used it but has read the manual
11:34
<jgraham>
Or parts thereof
11:51
<smaug____>
hmm, so implementation would need to encode video stream to webm and then show it in <video> ?
11:51
<smaug____>
though, perhap using the URL.createObjectURL doesn't require any encoding
11:52
<smaug____>
erm, but that is done for Stream, and Stream needs to be encoded...
11:55
<smaug____>
Lachy: how does that work with <device>? Do you first encode the video and then decode using <video> ?
11:57
<Lachy>
there is some magic in the back end that somehow renders the video directly from the camera without encoding as webm first.
11:58
<smaug____>
actually, <device> allows that quite easily, but with Streams that is not so easy
11:59
<Lachy>
the new API will also allow that quite easily, at least as far as rendering it locally is concerned. The recording or P2P streaming features will require webm encoding though.
12:41
<benschwarz>
Hi all
12:43
<zcorpan>
hi benschwarz
12:50
<benschwarz>
zcorpan he did it
12:51
<benschwarz>
(your tweet)
12:52
<zcorpan>
yeah
12:52
<zcorpan>
though i guess it still has default phpbb3 styles
12:53
<benschwarz>
zcorpan: its a start
12:53
<hsivonen>
Are "Compatibility Mode Settings" gone in IE9 or do I just have trouble finding them?
13:07
<hsivonen>
found it
13:08
<hsivonen>
required menu bar
13:08
<hsivonen>
well hidden!
13:24
<smaug____>
does Opera implement xhr.timeout?
13:35
<zcorpan>
smaug____: no
13:35
<smaug____>
ok
13:35
<smaug____>
then it is still easy to change it, since no one has implemented it
13:36
<smaug____>
zcorpan: thanks
13:37
<zcorpan>
how do you want to change it?
13:37
<smaug____>
zcorpan: it should be possible to cancel the timeout
13:37
<smaug____>
and perhaps also change it after send()
13:38
<smaug____>
about to send an email about this to webapps
13:38
<zcorpan>
maybe also set in UNSENT state?
13:39
<smaug____>
why does it need to be limited at all
13:39
<smaug____>
well, at all after opened
15:42
<jgraham>
Yay! Working group decision on ascii-ref means we get to waste even more time on arguing about the decision
16:59
<dylanw>
I'm trying to understand why header/footer are inline and not block elements. Could someone point me in the direction of that discussion?
17:00
<Ms2ger>
There's no such distinction
17:00
<Ms2ger>
But I think you're misinterpreting something, because they are what some would call block elements
17:01
<TabAtkins>
They're not inline, they're flow elements. The default UA stylesheet sets them to display:block.
17:01
<TabAtkins>
http://www.whatwg.org/specs/web-apps/current-work/complete/rendering.html#display-types
17:01
<dylanw>
FF3.6 by default has them display:inline
17:02
<dylanw>
browser bug?
17:02
<TabAtkins>
That's because it doesn't know what the elements are. Unknown elements are display:inline by default.
17:04
<dylanw>
So then FF3.6 doesn't actually support HTML?
17:05
<TabAtkins>
It doesn't understand several of the new HTML elements, no.
17:05
<TabAtkins>
3.6 is relatively old at this point.
17:07
<dylanw>
OK, I understand. I misinterpreted flow. And apparently so did the Moz crew.
17:07
<dylanw>
Thx
17:07
<TabAtkins>
No, Moz did not misinterpret it. 3.6 doesn't implement those elements at all.
17:08
<jgraham>
dylanw: To be fair all Mozilla did wrong was to have a long release cycle
17:08
<TabAtkins>
(Most flow elements are display:block, actually.)
17:08
<TabAtkins>
(Maybe all?)
17:09
<dylanw>
anchors are listed as flow
17:10
<dylanw>
yeah, the current flow list is a mix of classical block (div) and inline (abbr)
17:10
<TabAtkins>
Hmm, I'm surprised that abbr is flow. That sounds like it should be phrasing content.
17:11
<dylanw>
It's both.
17:11
<TabAtkins>
Indeed. I wonder why it's flow at all, though?
17:11
<dylanw>
It seems, and I'm just doing this prima facie
17:11
<dylanw>
that flow only means you're probably a block
17:11
<dylanw>
flow+phrasing you're probably inline
17:12
<dylanw>
though p is the obvious exception
17:12
<dylanw>
wait, no, p isn't phrasing
17:12
<dylanw>
duh. has its own section
17:12
<TabAtkins>
Ah, never mind, lots of elements are flow+phrasing.
17:13
<dylanw>
HTML standard never really specifies whether something is a "block"
17:14
<TabAtkins>
Yes it does, in the rendering section.
17:14
<TabAtkins>
Which I linked. ^_^
17:15
<dylanw>
OK, I see now
17:17
<dylanw>
I'll just use that as a reset stylesheet
17:18
<dylanw>
thx
17:18
<TabAtkins>
Nah, it's not quite good enough for a reset stylesheet, since it uses some things that aren't supported in CSS. Use html5boilerplate.com
17:21
<dylanw>
Hadn't seen Paul's stuff yet.
17:21
<dylanw>
HTML makes me feel like I'm back in 2000 again screaming at Netscape 4's DOM.
17:26
<AryehGregor>
abbr is flow content, but can only be used where phrasing content is expected . . . so what does it mean that it's flow content?
17:26
<AryehGregor>
Oh, of course.
17:26
<AryehGregor>
It means you can do <div><abbr>Foo</abbr></div> or such, right?
17:26
<AryehGregor>
Wait, no.
17:26
<AryehGregor>
You can't.
17:26
AryehGregor
is confused now
17:29
<jgraham>
"Flow COntent" doesn't mean much
17:29
<jgraham>
s/O/o/
17:30
<jgraham>
Just that the element represents part of the document rather than part of the metadata, or something
17:31
<jgraham>
Not even that
17:32
<jgraham>
Just "stuff that appears in the body"
17:33
<jgraham>
So for abbr the important requirement is that it can only be used where phrasing content is expected
17:34
<jgraham>
Hixie: You around yet?
18:08
<TabAtkins>
Wow, once you remove the filters and fonts from SVG, it's actually a pretty small and manageable spec.
18:22
<TabAtkins>
Hm, how appropriate is it to break up a table into sub-tables purely by using some rows as new headers? I want multiple related tables that all size their cells the same.
18:24
<hsivonen>
TabAtkins: even with <use> in place?
18:24
<TabAtkins>
Huh?
18:24
<TabAtkins>
Oh, SVG.
18:24
<TabAtkins>
One sec, breaking for lunch.
19:00
<Hixie>
jgraham: here
19:01
<Hixie>
smaug____: yes, PeerConnection would work well only it both ends support same codecs. I suppose I could put in the spec "user agents must implement the codecs that other user agents implement" but I doubt it would help...
19:01
<Ms2ger>
Hixie, Objection!
19:02
<Hixie>
to what?
19:02
<Hixie>
my being here? :-)
19:02
<smaug____>
Hixie: that would at least make it clear that PeerConnection works only in some cases
19:02
<Ms2ger>
To a requirement to implement codecs other UAs implement
19:03
<Hixie>
if the goal is to point out that PeerConnection's media streaming features are only going to interoperate between implementations that are interoperable, I think there'd be better ways to do it
19:03
<Hixie>
but it seems pretty self-evident to me
19:03
<Ms2ger>
Well, most things are only going to interoperate between implementations that are interoperable, no?
19:03
<Hixie>
precisely
19:04
<Ms2ger>
The question is how much interoperability the spec requires/can require
19:04
<Hixie>
for codecs, currently, none, as far as i can tell
19:05
<smaug____>
Could there be some hint which codecs are supported and then the other end wouldn't even try to connect if the right codecs weren't supported?
19:05
<Hixie>
there is
19:05
<Hixie>
it uses SDP offer/answer
19:05
<Hixie>
same as SIP
19:06
<smaug____>
but PeerConnection isn't about playing any stream
19:06
<smaug____>
One could just save the stream
19:06
smaug____
needs to re-read PeerConnection
19:06
<Hixie>
you're only allowed to save the stream if you can write it out in a format you support, currently
19:07
<Hixie>
(the goal is to enable video conferencing, not remote stream saving, so i think it makes sense to natively support the codec negotiation use case and probably doesn't make sense to allow arbitrary unsupported streams to be sent)
19:12
<smaug____>
Hixie: why you're allowed to save only stream you support?
19:12
<smaug____>
it would make sense to save the stream and perhaps use some other program to open it
19:12
<Hixie>
the actual requirement is that you be able to render streams you generate
19:12
<smaug____>
that way you couldn't really "stream" anything, but at least send videos
19:12
<Hixie>
so that you can always view local video
19:13
<smaug____>
sure, that is a different thing
19:13
<smaug____>
but if you get some other format from the other end
19:16
<Hixie>
well the idea is that you can't get an unsupported stream, so... :-)
19:16
<smaug____>
:)
19:17
<smaug____>
the whole Streaming needs to be extended a bit
19:18
<smaug____>
so that sending random files becomes possible etc.
19:18
<smaug____>
I also wonder if there could be just one API for streaming-like communication, not two WebSockets and PeerConnection
19:22
<zewt>
they're different enough that trying to merge them would probably end up more complex than the sum of the two separate APIs
19:25
<smaug____>
they are not that different
19:25
<smaug____>
though both APIs will change so we don't know what they will look like
19:26
<zewt>
maybe
19:26
<zewt>
always the tricky issue--between trying to generalize things, and overgeneralizing
19:27
<smaug____>
true
19:27
<Hixie>
smaug____: what's the use case?
19:28
<smaug____>
atm it would be still rather easy to change both APIs since neither of them are being used
19:28
<zewt>
particularly early in design where, as you said, where it's going hasn't solidified yet
19:28
<smaug____>
Hixie: use case? to keep web platform easier for developers by not adding too many APIs
19:28
<smaug____>
note, I'm not sure whether merging the APIs would be good, but something to think about
19:29
<Hixie>
no i mean for streaming unsupported media
19:29
<smaug____>
oh
19:29
<Hixie>
i agree that we should keep the APIs consistent
19:29
<Hixie>
i've got a bit more work to do on that shortly
19:30
<smaug____>
streaming unsupported message... say, someone wants to send me some video message, and I can't use my browser to show it on live, but I could just save it and use some other tool to watch it
19:30
<Hixie>
that seems like an extremely rare case that could just as easily be supported (possibly more easily be supported) by having the file be relayed through the server
19:31
<smaug____>
true, which is a reason why I said sending files/blobs should be supports
19:31
<smaug____>
supported
19:31
<Hixie>
you're saying we _should_ support it because it's rare and not needed? i'm confused.
19:32
<smaug____>
well, it would be natural to support sending all kinds of streams
19:33
<smaug____>
but if that case can be handled by just sending blobs, that is ok too
19:33
<Hixie>
i don't understand why it's natural to want to do P2P transfer of data that needs to be transmitted reliably and for which latency is not an issue.
19:33
<TabAtkins>
I need to poke around the issue with our idea for just ripping streams off of <input type=file accept>, so we have a nice consistent API living in the DOM.
19:33
<Hixie>
if anything it seems highly _un_natural to me
19:35
<smaug____>
Hixie: well, currently sending files isn't really supported by PeerConnection
19:35
<Hixie>
right, because it's rare and unnecessary
19:35
<smaug____>
ph
19:35
<smaug____>
er
19:35
<smaug____>
oh
19:35
<smaug____>
:)
19:35
<smaug____>
Hixie: how so?
19:35
<smaug____>
is the idea to send files using some other mechanism?
19:35
<Hixie>
yeah. XHR to the server.
19:36
<Hixie>
peer-to-peer transfer is hugely complicated but has one redeeming feature: low latency.
19:36
<smaug____>
ah
19:36
<Hixie>
if you don't need low latency, then really peer-to-peer transfer is far more pain than it's worth.
19:37
<smaug____>
hmm, not sure I like that a server needs to store some huge file before someone else can download it from there
19:37
<Hixie>
doesn't have to be stored, it can just be relayed.
19:38
<Hixie>
(if you use websockets(
19:38
<Hixie>
))
19:39
<smaug____>
but that causes still quite a bit network traffic to the server
19:39
<smaug____>
I wonder how skype works when sending files
19:39
<zewt>
proprietarily :)
19:39
<zewt>
(also, very poorly, in my experience)
19:39
<Hixie>
if users really transfer files to and from each other a lot, then it's a different matter
19:40
<Hixie>
but seriously, who does that anymore?
19:40
<zewt>
these days they're called "torrents"
19:40
<zewt>
things no javascript networking API is going to be thorough enough to support: torrents :)
19:41
<zewt>
(thankfully)
19:41
<smaug____>
anyone who needs to send huge audio file packages for mixing for example
19:41
<smaug____>
it is quite handy to use Skype to send a new mix of some song
19:41
<Hixie>
zewt: actually there's not really any reason you couldn't build torrents on PeerConnection as far as I can tell
19:41
<zewt>
one use case (not claiming it's a common one): web IRC clients, with DCC transfers or equivalent
19:41
<Hixie>
smaug____: you wait for the other person to be online to do that? that seems unlikely.
19:42
<zewt>
(it's common for IRC, though usually for bad reasons)
19:42
<Philip`>
People occasionally try to send me files (text files, images, etc) over MSN
19:42
<zewt>
same thing for sending files over IM clients in general, I guess
19:42
<Philip`>
(which fails because I use Bitlbee)
19:42
<zewt>
(let's say the same thing simultaneously, ready go)
19:43
<AryehGregor>
Hixie, people commonly spend all day on these messaging thingies.
19:43
<smaug____>
exactly
19:43
<AryehGregor>
I just idle on MSN all the time, except that when I reboot I forget and don't restart it because I never use it.
19:43
<Hixie>
for sending small files, relaying is fine
19:43
<zewt>
i'd argue not
19:43
<AryehGregor>
But I idle on IRC all the time, which is the same sort of idea.
19:44
<zewt>
when you have millions of users on an IM network, relaying every file transfer through centralized servers is a pretty hefty server requirement when it can often be completely elimianted
19:44
<zewt>
eliminated
19:45
<Hixie>
as the only person here who actually works for a company who runs an IM network as far as I know, I'm not at all convinced. :-)
19:45
<zewt>
i'd say typical size for files over IM clients is a few megs, at least (eg. larger than text files, but typically not CD images)
19:45
<Hixie>
trying to do it peer-to-peer probably means far more support costs than it would cost just to do it via relay
19:45
<TabAtkins>
Hixie: to be fair, the chat network is well-integrated into the email client, so there's little need to directly transfer files through the chat.
19:45
<smaug____>
Hixie: does anyone actually use Google's IM ;)
19:46
<Hixie>
TabAtkins: well yeah, exactly
19:46
<zewt>
Hixie: i think your company has more bandwidth available than most :)
19:46
<TabAtkins>
smaug____: ...yes? It's the only IM client I use, largely because I have gmail up at all times.
19:46
smaug____
has stopped using gmail web client
19:46
<zewt>
most IM clients are entirely peer-to-peer for file transfers, I *think*
19:47
<zewt>
or at the very least, most of the time with some kind of fallback (Skype feeds files through other users when it can't make a direct connection, which doesn't work very well--too slow)
19:48
<smaug____>
but usually Skype's file transfer works really well
19:48
<Philip`>
MSN apparently has 350 million active users per month (as of 2009), which is apparently about double Gmail's users
19:48
<zewt>
(not saying this is necessarily a use case that this API should try to handle--just that it is a real case)
19:48
<Hixie>
anyway. we can always add features later. we can't remove features. so let's start small.
19:48
<zewt>
(not familiar enough with the API to argue the former)
19:48
<Hixie>
especially when it's not clear we need the features :-)
19:50
smaug____
goes jogging and will think about a monster API which combines WebSockets and PeerConnection and which supports file transfer and any type of video streams :)
19:57
<hsivonen>
wasn't it part of the plan to be able to filter the bytes that come from Web Socket / Peer Connection before pushing them to a video decoder?
19:58
<Hixie>
there was a plan?
19:58
<hsivonen>
so that sites that implement their own obfuscations schemes in JS and stop whining about the lack of built-in DRM
19:58
<Hixie>
oh, no, i have no plan to support that kind of thing.
19:58
<Hixie>
the current model would support it though, fwiw. We'd just have to create a descendant of Stream that takes a Stream as input and invokes a callback regularly or some such.
19:59
<hsivonen>
Hixie: so if the current model would support it, you are planning on enabling it
19:59
<hsivonen>
even if not deliberately
19:59
<Hixie>
no
19:59
<Hixie>
the model could be extended to support it, it doesn't support it currently.
20:00
<hsivonen>
Hixie: oh
20:00
<Hixie>
and i'm not planning on enabling it, though it could be enabled as a side-effect of the design.
20:00
<Hixie>
a hallmark of good design is that it enables all kinds of things that weren't planned for :-)
20:02
<tw2113>
i think after this side of Pearl Jam's album Ten is done spinning, i should hit up the post office quick
20:08
<tw2113>
oops, wrong room :D
20:55
<jgraham>
Hixie: Would it be possible for you to look at the Mozilla history API changes soon?
20:55
<Hixie>
how soon is soon?
20:56
<jgraham>
Well the sooner the better, really
20:56
<jgraham>
But obviously it is your schedule :)
20:56
<Hixie>
i have to do something about issue 152, i'm already late with issue 147, and i have to send this feedback on p2p stuff
20:57
<Hixie>
also have to update the websocket api spec
20:57
<Hixie>
oh and issue 131 is coming up
20:57
<Ms2ger>
jgraham, obviously you should touch all older bugs :)
20:58
<Hixie>
oh man, 155 and 144 are coming up too
20:59
<jgraham>
Ms2ger: Or stop the HTMLWG chairs from cluttering up Hixie's life with fixed deadline issues that don't correspond to things I care about
20:59
<jgraham>
Not sure how to do that though
20:59
<Hixie>
well one way is to write the CCPs for those, that would really help
20:59
<Ms2ger>
That works too
21:00
<Hixie>
but i can understand if you're not inclined to do that :-)
21:00
<Hixie>
jgraham: is there a specific thread for the history stuff already in whatwg.org/issues?
21:00
<Hixie>
or bug#?
21:00
<Ms2ger>
There's a bug
21:00
<Ms2ger>
I think you filed it
21:01
<Hixie>
k
21:01
<jgraham>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=12277
21:02
<jgraham>
I think making me write CCPs for the multimedia stuff would not be a winning strategy :)
21:08
<Hixie>
could you write one for 147?
21:08
<Hixie>
anyway, i think i should be able to get to the history stuff in the near future
21:08
<Hixie>
(as in within the next month)
21:08
<Hixie>
keep bugging me about it
21:14
<jgraham>
Hixie: OK, thanks
21:31
<twisted>
hi, question, in Chrome background-size works if I put it on 80%. In Firefox it doesn't. I've checked and now also have the -moz-background-size set but still nothing in Firefox...
21:31
<twisted>
what am I doing wrong?
21:43
<TabAtkins>
Does the version of FF you're testing against support background-size?
21:47
<aho>
https://developer.mozilla.org/en/css/-moz-background-size
21:47
<aho>
3.6 requires a prefix, 4.0 doesnt anymore
21:53
<twisted>
using 4.0
21:53
<twisted>
http://twixted.hatkeintuv.de:8888/all4wheels/ that's the site in question
21:54
<twisted>
if you inspect the element #main-menu li a
21:54
<twisted>
it does show the css it should use but somehow it doesn't do anything with it
21:58
<TabAtkins>
Try changing the 'background' declaration on line 7 to just background-image.
21:58
<TabAtkins>
That *shouldn't* affect anything, but FF may be buggy and reseting the -size property when you use the background shorthand.
21:59
<twisted>
it now dissapeared completely
21:59
<twisted>
even went *poof* in chrome
21:59
<twisted>
lemme check
21:59
<TabAtkins>
Oh, right. The no-repeat value there would then need to be pulled out into background-repeat.
22:00
<twisted>
whoo
22:00
<twisted>
works
22:00
<TabAtkins>
If that worked, then FF is buggy. File a bug, please?
22:00
<twisted>
firefox actually does the center center more nicely then chrome
22:00
<twisted>
uhm, where?
22:01
<TabAtkins>
bugzilla.mozilla.org
22:01
<twisted>
cool I will
22:03
<twisted>
funny since I fixed it in Firefox now Chrome suddenly also does it properly centered
22:04
<TabAtkins>
Same deal - background-size was being reset by the 'background' rule, because 'background' was more specific than all the background-* rules you had set generally on the links.
22:05
<TabAtkins>
Why are you doing this with backgrounds, though? Why not just use an <img> with @alt, rather than putting in real text, hiding it, then putting in an image via the background property?
22:16
<twisted>
TabAtkins: because drupal #^@#%& sucks
22:17
<twisted>
already had to make a custom function to get an extra span
22:17
<TabAtkins>
Ah, indeed it does. I recommend not using it.
22:17
<twisted>
well considering the budget for this website
22:17
<twisted>
it's probably the cheapest solution
22:17
<twisted>
joomla really sucks
22:17
<twisted>
wordpress is for blogs
22:18
<TabAtkins>
I've been reasonably happy with Django, if you really need a CMS.
22:22
<twisted>
I unfortuanly spend waaay too much time already on Drupal atm to switch
22:27
<twisted>
still didn't get my bugzilla email :p
22:27
<twisted>
ah spam folder lol
22:30
AryehGregor
suspects he could construct a minimal test-case for the really annoying Web Inspector bug in Chrome, with some effort
22:30
<AryehGregor>
Unfortunately, I lack the time to expend effort.
22:31
<AryehGregor>
I think that when scripts are reloaded and change, it still partially caches the old version of the script sometimes, which means it only lets you set breakpoints at lines that you could have set breakpoints at in the previous version of the file.
22:31
<TabAtkins>
Your margins is too narrow?
22:35
<twisted>
TabAtkins: https://bugzilla.mozilla.org/show_bug.cgi?id=641977
22:52
<zot-->
dreaming of the day we can have some consistency in keyboard shortcuts for web apps :|
22:53
<zewt>
chrome seems really good at it, not eating many alt- keys; but everyone else...
22:54
<twisted>
TabAtkins: happy to have been able to file a bug!
22:55
<TabAtkins>
Thanks for doing so!
22:56
<zewt>
can't even use , and . as hotkeys--Opera users get irate
22:57
<TabAtkins>
We need a new hyper key which is explicitly reserved for use by apps.
22:58
<zewt>
heh
22:58
<zewt>
every time a new encapsulation layer is added to computers, a new key is needed
22:58
<zewt>
windows key, control key, web key
23:07
<TabAtkins>
Just do what I do and say that you get the web key when you press Alt+Ctrl+Windows at the same time.
23:07
<TabAtkins>
(I use this combo to move between desktops.)
23:09
<zewt>
i'll avoid using key combinations that many keyboards physically can't register. heh
23:14
<zewt>
heh, i'm half-inclined to just stomp over browser hotkeys (since they'll let you, generally) for web apps, so long as they don't give a better option :|
23:27
<hober>
TabAtkins: we just need to get everyone to use the space cadet keyboard; then we'll have several modifiers for use by web apps
23:31
<TabAtkins>
hober: Yup, that'd work.
23:32
<zewt>
browsers just need to let go of alt/meta :|
23:33
<zewt>
was easier for chrome to do (from the beginning) than for old browsers (with users and user habits), but I think it's just something to get over
23:44
<TabAtkins>
twisted: Sorry for misinforming you - I somehow missed the 'background-size' in the list of properties that the 'background' shorthand reset.
23:44
<TabAtkins>
So Firefox's behavior is correct, Chrome's isnt.
23:46
<twisted>
huh, but isn't the idea of CSS that it overrides based on cascading?
23:47
<twisted>
so my latter definition of the size should still affect it
23:47
<twisted>
and not ignore?
23:49
<TabAtkins>
twisted: the selector setting the plain 'background' property has higher specificity.
23:58
<twisted>
then it still doesn't get overidden?
23:59
<TabAtkins>
No, because the 'background' shorthand acts like all the component properties were being set instead. Thus, the "background-size:initial" that is implied by your rule overrides the "background-size:80%" that you actually have.
23:59
<TabAtkins>
Same with the background-position.