00:15
zcorpan
wonders why http://hacks.mozilla.org/2010/02/firefox-46-features/ has stray </p>s in there
01:55
<gsnedders>
http://www.python.org/dev/peps/pep-0372/
01:55
<gsnedders>
html5lib mentioned in a PEP
01:55
<gsnedders>
(it's also mentioned in the Unladen Swallow PEP)
01:55
<gsnedders>
http://www.python.org/dev/peps/pep-3146/
06:16
<JonathanNeal>
It's really weird when Opera, Safari, and Firefox need a style, but Internet Explorer and Chrome do not.
06:27
<JonathanNeal>
Any folks around?
06:28
<Peter`>
that's just because of UA stylesheets
06:28
<Peter`>
also the reason reset-css files exist
06:33
<JonathanNeal>
Peter`, yes I don't want to spam this board, but that's why I'm labeling all the entries in my html5 reset.
06:33
<JonathanNeal>
I should also label when I'm setting the style to match ua and when i'm setting the style to match spec.
07:10
<JonathanNeal>
I'm sure it's a slippery slope, but there are some things I wish the spec didn't leave out and up to the ua.
07:15
<Peter`>
I concur, but I doubt it's going to change anytime soon
07:20
<JonathanNeal>
I didn't realize that type="a" and type="A" didn't work.
07:22
<JonathanNeal>
They work until you specify them in css ol[type="a"] {} ol[type="A"] {} both get eval'd.
07:22
<JonathanNeal>
for either a or A
07:23
<JonathanNeal>
Wow, in Chrome it even says the UA style is for ol[type=a] but when you style ol[type=A] it overwrites it anyway.
07:24
<JonathanNeal>
So ol[type=a] and ol[type=A] are different selectors within the UA's CSS, but the same outside.
07:48
<hsivonen>
JonathanNeal: I believe the dual selectors in the UA style sheet are there because the UA style sheet gets parsed in the case-sensitive mode to make it work right with XHTML
09:42
<annevk>
hsivonen, do you see validators as conversion help?
09:42
<annevk>
hsivonen, 'cause personally I would only start using the validator after I have made some attempt at following the rules
09:46
<hsivonen>
annevk: some people have already made fairly reasonable attempts to follow the previous spec
09:47
<hsivonen>
correction
09:47
<hsivonen>
the validator for the previous spec
09:47
<hsivonen>
which is Transitional
09:47
<hsivonen>
maybe I should have clarified that I'd like to loosen things based on usage data and that I wouldn't want to permit <font face> or <font size>
09:48
<hsivonen>
basically, I think the individual pre-existing features should be measured against usage on near-valid Transitional sites
09:48
<hsivonen>
and against Hixie's stated criteria
09:48
<annevk>
Can't you just group them all into one?
09:49
<hsivonen>
I think not
09:49
<annevk>
I think Hixie suggested that already
09:49
<annevk>
I guess schema-based validation does not give you the right messages for that...
10:07
<jgraham>
It never faiuls to amaze me how much people will pay to listen to some guy from the internet talk about stuff
10:08
<jgraham>
400 pounds for a one day workshop to learn about part of HTML 5 (or 200 if you're a student) is pretty insane
10:09
jgraham
hopes there are no students who can actually afford that
10:09
<Philip`>
Depends on the value of your time and how long it would take you to learn the same amount without that guidance
10:09
MikeSmith
wonders who the particular speaker is
10:10
<annevk>
Jeremy Keith I guess
10:10
<MikeSmith>
ah
10:10
<MikeSmith>
well, Jeremy's a great speaker
10:10
<annevk>
to his credit they organize a fairly cheap conference themselves in England
10:11
<jgraham>
I'm not disputing that he's a good speaker
10:11
<jgraham>
I'm just amazed that there are people who can afford it
10:11
<annevk>
I'd guess companies pay the bill
10:11
<jgraham>
Well sure.
11:09
<zcorpan>
JonathanNeal: http://www.w3.org/Bugs/Public/show_bug.cgi?id=9102
11:27
<annevk>
we should really remove those silly Selector rules
11:27
<annevk>
but I guess it is too late or something :/
11:30
<zcorpan>
hmm, i wonder if the spec is broken for <input type="COLOR"/> etc in xhtml
11:31
<annevk>
broken how?
11:32
<zcorpan>
the attribute value is case-insensitive in xhtml
11:32
<zcorpan>
but the selector matches case-sensitively
11:33
<annevk>
doesn't seem like a problem to me
11:34
<annevk>
well, maybe a problem with Selectors
11:34
<zcorpan>
it's a problem because writing valid xhtml would result in broken rendering
11:34
<zcorpan>
if the spec is implemented literally
11:34
<annevk>
oh, the spec
11:35
<annevk>
sorry, I see what you mean now
11:35
<annevk>
yeah, that seems broken
11:35
zcorpan
files a bug
11:38
<annevk>
i initially thought you meant for authors trying to style such inputs
12:17
<jgraham>
hsivonen: IO think it may be worth explaining why you use OpenOffice to create web pages if it involves a significant amount of manual post-processing work. At the moment it sounds a bit like the start of a "Doctor Doctor, it hurts when I do this" joke
12:18
<hsivonen>
jgraham: it lets me do paragraphs, headings and tables
12:18
<hsivonen>
jgraham: and Nvu wasn't better at that
12:19
<hsivonen>
jgraham: and getting a Dreamweaver license would be an overkill
12:19
<jgraham>
Tables I can somewhat understand, but it seems like for paragraphs + headigns you would be better off with a text editor
12:19
<hsivonen>
jgraham: I disagree
12:19
<hsivonen>
press return vs. typing all those tags
12:20
<jgraham>
Including the effort of later post-processing the output?
12:20
<hsivonen>
yes
12:20
<jgraham>
I'm pretty surprised
12:20
<jgraham>
(I guess tables also need post processing since they will not have proper headers or anything)
12:21
annevk
usually edits by hand
12:22
<hsivonen>
jgraham: tables from OOo do have <th>s
12:22
<hsivonen>
the problem with tables in that OOo wraps the cell contents in <p>
12:22
<hsivonen>
and puts style and align all over the place
12:22
<jgraham>
In the right places (and in how many scenarios)?
12:23
<hsivonen>
jgraham: I'm using OO.o Writer/Web. Are you talking about taking an OO.o Writer doc and exporting to HTML?
12:23
<hsivonen>
the "/Web" part is significant
12:23
<hsivonen>
uses HTML as the native format
12:24
<hsivonen>
and exposes semantic elements is the style palette
12:24
<annevk>
ooh great, I'm losing an hour sleep this Sunday?
12:24
<jgraham>
hsivonen: Ah. My OpenOffice install doesn't include anything that obviously looks like that
12:24
<MikeSmith>
hsivonen: so I pulled latest htmlparser source and updated my workspace and now trying to build... does ecj have the same command-line options as javac? can I just alias javac to ecj?
12:24
<annevk>
this trip to SFO is going to be sweet :/
12:24
<hsivonen>
MikeSmith: no idea. I'm working on fixing that right now (as soon as I stop typing on IRC)
12:25
<jgraham>
Oh, curious. There is no top level menu for it, but you can create a new HTML document which changes the application title
12:25
<jgraham>
s/top level menu/application menu item/
12:26
<jgraham>
The badness of the OO.org UI never fails to amaze me
12:26
<annevk>
creating a good editor is apparently a pretty hard problem
12:26
<hsivonen>
also, never ever open an existing HTML doc into OO.o
12:27
annevk
would love a good robust web-based editor
12:27
<hsivonen>
OO.o is write-only
12:27
<jgraham>
Totally. That doesn't excuse OO.org for having a lame UI in every respect, not just for creating clean HTML
12:30
<hsivonen>
Mozilla has made much better recovery progress from having a COM clone, an app-specific crossplatform UI toolkit, an app-specific cross-platform 2D graphics layer and a horrible GUI (Netscape 6 Preview Release) than OO.o
12:32
<jgraham>
OO.o has never recovered from the policy of making a pixel-by-pixel clone of Office 95
12:32
<jgraham>
(at least I assume it was a policy. It would be remarkable if it happened by accident)
12:39
<hsivonen>
MikeSmith: fix pushed to hg
12:46
<MikeSmith>
hsivonen: thanks
12:47
hsivonen
goes back to making the C++ version of that file compile nicely with GCC and MSVC
12:47
<hsivonen>
apperently, this area is too tough for 3 compilers
13:15
<MikeSmith>
hsivonen: btw, when I run jing from the command line in Terminal on OSX, I get "?" characters instead of curly quotes
13:15
<hsivonen>
MikeSmith: you need to set the terminal to UTF-8
13:15
<MikeSmith>
e.g., rror: Element ?base? from namespace ?http://www.w3.org/1999/xhtml? is missing one or more of the following attributes: [href, target]
13:15
<hsivonen>
MikeSmith: or if your terminal is in UTF-8, you need to set the System.out to UTF-8
13:16
<MikeSmith>
ah, OK
13:16
<hsivonen>
there's a magic java system property for that
13:16
<hsivonen>
it's very sad if Apple still keeps Java defaulting to MacRoman
13:17
<hsivonen>
the system property is file.encoding
13:17
<hsivonen>
-Dfile.encoding=UTF-8 should do the trick
13:17
<MikeSmith>
hsivonen: OK, trying that now
13:18
<MikeSmith>
indeed, yeah, that works
13:18
<MikeSmith>
thanks
13:27
jgraham
wonders if Sam actually read the IRC logs
14:11
<MikeSmith>
hsivonen: I'm trying to figure out where the w:string datatype is defined
14:12
<hsivonen>
MikeSmith: syntax/relaxng/datatype/java/src/org/whattf/datatype/AsciiCaseInsensitiveString.java
14:13
<MikeSmith>
ah
14:13
<MikeSmith>
thanks
15:29
<Philip`>
"I see value in being able to express notions such as "this page intentionally violates rule number 34"."
15:29
<Philip`>
I don't see how it's possible to violate http://tvtropes.org/pmwiki/pmwiki.php/Main/RuleThirtyFour
15:32
<Dashiva>
Perhaps by somehow providing content that cannot be made pornographic
15:34
<Philip`>
I don't see how it's possible
15:37
<Dashiva>
That's only because you lack faith
16:20
<JonathanNeal>
Thanks for the heads up on case insensative selectors, zcorpan.
16:27
MikeSmith
wonders what client do people use for jabber.ietf.org meetings
16:27
<annevk>
Empathy
16:27
<annevk>
it ships with Ubuntu
16:28
<annevk>
and does pretty much all messaging
16:29
jgraham
doesn't go to ietf meetings but uses bitlbee+irssi for jabber (donno if that has all the needed features for that usecase though)
16:29
<annevk>
somewhat tempted to install the new Ubuntu beta
16:29
<MikeSmith>
annevk: thanks
16:29
<MikeSmith>
wonder if there's anything that works on OSX
16:29
<MikeSmith>
jgraham: it uses some extension, XEP-0045
16:30
<annevk>
MikeSmith, http://www.pidgin.im/download/mac/ prolly works
16:30
<MikeSmith>
thanks
16:30
annevk
was using pidgin before
16:40
<lazni>
adiumx uses the pidgin lib
16:42
<annevk>
this is quite boring
16:42
<franksalim>
?
16:42
<annevk>
iri meeting
16:43
<franksalim>
i believe that
16:43
<jgraham>
Not just anne dissing the channel
16:43
<annevk>
mumbling on Unicode, UTF-8, octet sequences
16:44
<annevk>
just some meta discussion on the presentation of a IRI vs a IRI
16:52
<Dashiva>
IRI vs IRI?
16:53
<annevk>
dunno, we're now at LEIRI and HTML5 URL...
16:53
<annevk>
but it's going very slow
17:01
<Philip`>
The concept of competition between validators would be more convincing if more than one validator existed
17:02
AryehGregor
concurs.
17:02
AryehGregor
also isn't very inspired by how few web developers seem to use non-officially-endorsed tools like jslint, compared to Standards-Based Validators
17:03
<Philip`>
Also it would be more convincing if anybody used the HTML4 validators that compete with validator.w3.org
17:03
<Philip`>
(but as far as I can tell, everyone uses the latter)
17:04
<annevk>
i've seen people use validator.nu
17:04
<Philip`>
I tried using jslint once but then I realised it was a waste of time
17:04
<Dashiva>
^
17:05
<Philip`>
and I was making my code uglier and less readable just to satisfy its rule for semicolons and for not using the same name with var twice in one function
17:05
<Philip`>
s/rule/rules/
17:07
<AryehGregor>
Also, for compiled languages, where it's common to use warnings, everyone just uses their compiler's built-in warnings, so there's little to no competition on that score per se.
17:08
<AryehGregor>
(although I guess quality of warnings might be a contributing factor in the decision on which compiler to use, if there are multiple realistic choices for your platform)
17:15
<JonathanNeal>
Are there any other vendor stylesheets out there? Like, has someone compiled a stylesheet of how one browser does everything?
17:16
<AryehGregor>
Every browser has its own UA stylesheets.
17:17
<Philip`>
I thought MikeSmith collected some for his guide thingy
17:17
<JonathanNeal>
I would love to get ahold of those. I realized it might be a much more effective way to write mine.
17:17
<JonathanNeal>
I've been doing it through a process of laying out a pangram of every element in html5, then applying the styles from the html5 rendering guide.
17:18
<MikeSmith>
the h:tml doc uses the Webkit default UA stylesheet
17:18
<JonathanNeal>
Then applying additional styles in a vendor css file to catch whatever else was needed to normalize them.
17:18
<MikeSmith>
http://svn.webkit.org/repository/webkit/trunk/WebCore/css/html.css
17:23
<erlehmann>
the pecurialities are interesting. epiphany-webkit for example, does not have a default styling for <i>
17:24
<JonathanNeal>
Thanks MikeSmith, 0__qem sure is a strange value.
17:26
<MikeSmith>
JonathanNeal: I think that just gets interpreted as 0em
17:27
<MikeSmith>
I think for the build of the h:tml doc, I have it stripping out the __q part
17:27
<MikeSmith>
I don't remembering ever hearing what purpose the __q part serves
17:28
<MikeSmith>
yeah
17:29
<MikeSmith>
http://dev.w3.org/html5/markup/tools/css2xml
17:29
<MikeSmith>
(if you care to see all the ugliness)
17:30
<JonathanNeal>
I do care, I'm very interested in seeing how uas style and how people are using this info
17:34
<TabAtkins>
JonathanNeal: Poke around in your Firefox install to find it's stylesheet. I've modified mine for a few things. Relatively easy to get to.
17:34
<JonathanNeal>
Awesome!
17:34
<JonathanNeal>
Now just to get IE's
17:38
<MikeSmith>
annevk: is that Yngve speaking?
17:46
<annevk>
MikeSmith, I dropped out
17:46
<annevk>
MikeSmith, figured I should get some food
17:48
<AryehGregor>
http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/latest-trunk/ "This page is in Irish. Would you like to translate it? [Translate] [Nope]"
17:48
<annevk>
I got bored with it too. All the high-level theoretical thought trains. How about starting with where contemporary implementations are?
17:50
<JonathanNeal>
Interesting how Firefox tightens up their css file (e.g. "margin: 1em 0") but then does "0px" and "margin: 0.5em auto 0.5em auto;" in other places. It's nice to know these stylesheets are written by humans.
17:52
<Philip`>
JonathanNeal: Or it's worrying to know these stylesheets were written by robots that can pass the Turing test
17:53
<JonathanNeal>
Philip`, true dat, I am C.
17:54
<JonathanNeal>
If we find IE's, my assumption is that it would pass with flying colors.
17:55
<MikeSmith>
annevk: Larry is mentioning use of the term "URL" now
17:56
<annevk>
oh, maybe I should dial back in
18:00
<Philip`>
Does IE actually use a default stylesheet that's a real CSS file, rather than something more ad-hoc and legacyish?
18:04
<MikeSmith>
I think even for Mozilla and Webkit, the default styling behavior (logic?) is not totally captured in their UA stylesheets
18:04
<AryehGregor>
Philip`, I would guess so for IE >= 8.
18:05
<AryehGregor>
But yeah, some things aren't specifiable in CSS.
18:05
<JonathanNeal>
AryehGregor, we could write one if they don't already have one.
18:05
<AryehGregor>
They also all use vendor-specific extensions in their UA stylesheets, I assume.
18:05
<AryehGregor>
For anyone who hasn't seen it yet, this is interesting: http://themaninblue.com/writing/perspective/2010/03/22/
18:05
<MikeSmith>
AryehGregor: some things are, but they don't specify them in the UA stylesheet for some reason
18:06
<MikeSmith>
e.g., neither Gecko and Webkit UA stylesheets set bold on the th element
18:06
<MikeSmith>
they do it in program code instead
18:06
<AryehGregor>
Really? Why?
18:06
<MikeSmith>
AryehGregor: dunno
18:07
<MikeSmith>
there may not be many (or any) more instances of that then the th case
18:07
<MikeSmith>
but I know of the th case, at least, so I suspect there might be others
18:07
<annevk>
how long does this meeting go on for?
18:08
<MikeSmith>
I think it was supposed to be 2.5 hours
18:08
<JonathanNeal>
I'm cleaning up the stylesheets and placing them @ http://sandbox.thewikies.com/html5-browser-stylesheets/
18:09
<annevk>
MikeSmith, kk
18:09
<JonathanNeal>
I wasn't sure what the version was on that webkit stylesheet, it was from their trunk, so I was going to add the version numbers once I know.
18:09
<annevk>
now they're talking about vests vs t-shirts
18:09
<annevk>
clearly ninjas win
18:10
<MikeSmith>
JonathanNeal: note that those stylesheets get updated/changed somewhat regularly
18:10
<MikeSmith>
so if you make copies of them, you are going to need to track the upstream changes
18:12
<JonathanNeal>
MikeSmith, definitely the reason why I want to grab the versions. Yea, I can always tag it 56629
18:19
<JonathanNeal>
Tagged them as firefox-v3.6.2.css and webkit-r56629.css
18:29
<erlehmann>
JonathanNeal, better tag it gecko-1.9.xyz
18:29
<erlehmann>
after all its the rendering engines that change this
18:29
<JonathanNeal>
how would i verify whcih gecko it was?
18:30
<erlehmann>
doesn't ff report this ?
18:30
<JonathanNeal>
btw, from what I understand MS IE saves their stylesheet in an xml file.
18:30
<erlehmann>
check the about dialog.
18:31
<JonathanNeal>
1.9.2.2 for gecko erlehmann?
18:31
<erlehmann>
XML stylesheets, what a funny idea :D
18:31
<erlehmann>
JonathanNeal, dunno. but i have iceweasel 3.5.8, which has 1.9.1.6, so it seems plausible
18:34
<JonathanNeal>
I found IE6's stylesheet!
18:36
<Philip`>
Where?
18:36
<JonathanNeal>
Or I'm really dang close.
18:36
<JonathanNeal>
res://msxml.dll/defaultss.xsl
18:38
<Philip`>
That sounds like it's a stylesheet for XML files, not for HTML
18:38
<Philip`>
given that it's in msxml.dll
18:42
<JonathanNeal>
Yea, just need to figure out what the dll is for html
18:42
<JonathanNeal>
http://sandbox.thewikies.com/html5-browser-stylesheets/defaultss.xls.txt
18:42
<JonathanNeal>
that's the one for xml, at least.
18:47
<Philip`>
HTML is lots of DLLs
18:47
<Philip`>
like mshtml.dll
18:51
<JonathanNeal>
Yea ... it sure feels close though.
18:54
<JonathanNeal>
According to http://msdn.microsoft.com/en-us/library/aa741312(VS.85).aspx it might be in mshtml.dll
18:55
<Philip`>
There are no strings in that file that look like CSS to me
18:56
<AryehGregor>
Maybe they're ROT13ed.
18:56
<AryehGregor>
Or base-64.
18:56
<jgraham>
Philip`: I'm pretty sure that most people that use jsLint turn of many of the more opinionated warnings
18:56
<JonathanNeal>
Maybe they're in the registry?
18:56
<jgraham>
and I don't think it's really comparable to HTML since javascript doesn't have any conformance criteria
18:57
<AryehGregor>
Some people want that for HTML too, apparently. Or at least are open to considering it.
18:58
<jgraham>
(I guess theoretically one might like a porgram that told you if you were going to hit any features that change behaviour in ES5 strict mode. But that probably isn't statically determinable)
18:58
<jgraham>
AryehGregor: Yes some people want it. That doesn't mean it's a good idea
19:11
<AryehGregor>
Do all browsers support oninput?
19:12
<AryehGregor>
Meh, never mind.
19:12
<AryehGregor>
Doesn't matter.
19:14
<JonathanNeal>
I found a hacky way to grab 'em all.
19:47
<annevk>
i cannot access google.com
19:47
<annevk>
weird
19:51
<Dashiva>
Can you access this? http://www.schneier.com/blog/archives/2010/03/schneier_bloggi.html
19:52
<annevk>
sure
19:53
<annevk>
or google.nl etc.
19:53
<annevk>
it's very weird, my mother called me this morning and she had exactly this problem
19:54
<annevk>
maybe restarting the browser will fix it...
19:54
<annevk>
yup...
20:21
<jgraham>
Cookies?
20:22
<annevk>
i accept cookies
20:22
<annevk>
visit my site for address details
21:24
<annevk>
hmm, I wonder if the way Chrome (does not) implement(s) form validation will cause us problems going forward
21:24
<annevk>
i.e. required="" is "supported" but does not actually prevent form submission unless some JavaScript is used
21:25
<annevk>
not cool
21:27
<Hixie>
annevk: um yeah that's very not cool
21:27
<Hixie>
annevk: can you file a bug pointing out the very likely intrerop implications of this?
21:28
<TabAtkins>
Damn, anyone know where the hell image-fit is specified? (In CSS)
21:28
<Philip`>
TabAtkins: http://dev.w3.org/csswg/css3-page/#propdef-image-fit ?
21:29
<TabAtkins>
Danke, Philip`.
21:29
<Philip`>
Thank Google
21:29
<TabAtkins>
...I should have thought of that.
21:31
<annevk>
Hixie, would this be a bug on WebKit or Chrome?
21:31
<Hixie>
webkit i guess
21:36
<annevk>
https://bugs.webkit.org/show_bug.cgi?id=36677
21:36
<Hixie>
thanks
23:17
<Traveler>
hi
23:49
<JonathanNeal>
I set up a script to capture IE8's default styles.
23:49
<JonathanNeal>
http://sandbox.thewikies.com/html5-browser-stylesheets/ie8.css this is what it returned me.
23:56
<erlehmann>
JonathanNeal, next step: use the script to automate the capturing job for all browsers ? maybe ab-using browsershots.org and AJAX ?
23:56
<TabAtkins>
JonathanNeal: Very cool. I like it.
23:57
<erlehmann>
font-family: Courier New; — i suppose thats monospace ?
23:58
<TabAtkins>
Yeah, likely. It's a monospace font, at least, and I guess it's JonathanNeal's default.
23:58
<erlehmann>
i wonder if the „zoom: 1;“ often seen is a IE hack :D
23:58
<TabAtkins>
It triggers hasLayout!
23:59
<JonathanNeal>
yes, it triggers hasLayout.