01:39
<erlehmann>
“The message's content type was not explicitly allowed”
01:44
<erlehmann>
It would help to include which content types would be allowed. And no, I was not trying to send text/html.
02:03
<erlehmann>
seems multipart/mixed is not allowed. so much for PGP signatures.
08:28
<matjas>
As per HTML 4.01, the first occurrence of the character sequence </ (ETAGO or end-tag open delimiter) is treated as terminating the end of the <script> element’s content: http://www.w3.org/TR/html4/types.html#h-6.2
08:29
<matjas>
Has this changed since?
08:29
<Ms2ger>
That has never been true
08:29
<Ms2ger>
(Like most of HTML4)
08:29
<matjas>
It may have since no browser seems to enforce it; the de facto standard seems to be </script: http://kangax.github.com/jstests/etago_delimiter_test/
08:30
<matjas>
Ms2ger: Do you know if it was excluded from HTML5^H?
08:31
<Ms2ger>
script parsing is an awful mess
08:32
<Ms2ger>
Try <script><!--<script></script>
08:34
<matjas>
Mind = blown
08:49
<hsivonen>
matjas: if you are trying to sanitize html, you need an html parser
08:49
<hsivonen>
the tree builder is needed to tell apart html and svg scripts
08:49
<hsivonen>
which tokenize differently
08:50
<hsivonen>
Ms2ger: script parsing is a well-defined mess now!
08:51
<Ms2ger>
hsivonen, still an awful mess
09:41
<matjas>
hsivonen: I’m not trying to do anything like that; just curious what’s specced and what isn’t
09:41
<matjas>
so, the </ isn’t treated as terminating the end of the <script> element’s content in any browser
09:42
<matjas>
but "</script" is
09:42
<matjas>
perhaps that should be standardized / included in the spec then
09:47
<Ms2ger>
It's all in the spec
09:47
<Peter->
http://www.whatwg.org/specs/web-apps/current-work/multipage/tokenization.html#script-data-end-tag-open-state
09:48
<matjas>
That’s what I was looking for, thanks!
09:51
<hsivonen>
matjas: note that the string </script inside a script doesn't always go through that state
09:52
<hsivonen>
matjas: http://www.whatwg.org/specs/web-apps/current-work/multipage/tokenization.html#script-data-escaped-end-tag-open-state
11:31
<karlcow>
>Quoting standards is just stupid, when there's
11:31
<karlcow>
>two simple choices: "it works" or "it doesn't work because bugs happen".
11:31
<karlcow>
>Standards are paper. I use paper to wipe my butt every day. That's how much
11:31
<karlcow>
>that paper is worth.
11:31
<karlcow>
>Reality is what matters.
11:31
<karlcow>
— Linus Torvalds, https://bugzilla.redhat.com/show_bug.cgi?id=638477#c129
12:04
<ktoss>
any one use flexiejs here?
14:39
<MikeSmith>
anybody know any details about the Web browser that Baidu has supposedly developed?
14:40
<Ms2ger>
Not me, but I'm sure Fx is better :)
14:44
<MikeSmith>
what's Fx?
14:44
<MikeSmith>
Firefox?
14:44
MikeSmith
is listening to Invasion of the Reverb Snatchers by Bambi Molesters from Intensity! (✮✮✮✮✮)
14:51
<karlcow>
and for a depressing moment http://vimeo.com/21197696
15:01
<MikeSmith>
karlcow: I feel impressed rather then depressed
15:01
MikeSmith
wonders if gsnedders is around
15:04
<MikeSmith>
oh
15:04
karlcow
wonders if gsnedders still wears his make-up
15:04
<MikeSmith>
I got to the more depressing part
15:05
MikeSmith
is listening to Lands End by Laika & The Cosmonauts from Cosmopolis (✮✮✮✮✮)
15:10
<MikeSmith>
http://hacks.mozilla.org/2011/03/nocomply/ is nice
15:13
<erlehmann>
MikeSmith, Sputnik 2 was not designed to be retrievable, and Laika had always been intended to die. Depressed now?
15:19
<MikeSmith>
Laika died a hero and will remembered forever
15:22
MikeSmith
reaches the point in his macports reinstall where it's rebuilding atlas… whatever the hell atlas is… decides to check back in a few days to see if it's completed yet
15:37
<erlehmann>
MikeSmith, so you are saying … Laika died to become a meme? 4chan is not amused.
15:37
<MikeSmith>
I'm sorry for 4chan
15:39
<MikeSmith>
ok, this is pretty interesting: http://www.sciencedaily.com/releases/2011/03/110325102008.htm
15:39
<MikeSmith>
"The many hundreds of scales, however, seem to possess a deeper commonality: if their tones are compared in a two- or three-dimensional way by means of a coordinate system, they form convex or star-convex structures."
15:41
<MikeSmith>
http://www.english.uva.nl/news/news.cfm/9DEDC76E-E021-4F6E-AE9F0A55E899FC22
15:41
<MikeSmith>
http://staff.science.uva.nl/~ahoningh/publicaties/convex_scales.pdf
15:42
<MikeSmith>
"investigating the humanities from a computational perspective"
15:45
<MikeSmith>
100% of traditional musical scales form a star-convex pattern
15:47
<MikeSmith>
it would be interesting to do the reverse -- construct something in a star-convex pattern and generate a scale from it and see what it sounds like
15:51
<karlcow>
about laika, well like millions of cows, chicken that we raise to die :) still yummy
15:56
<MikeSmith>
dogs taste pretty good too
15:56
<MikeSmith>
whales also
15:58
<karlcow>
dogs I do not know
15:59
<karlcow>
or maybe I do not know if I had already one
16:03
<karlcow>
http://pmuellr.github.com/weinre/ remote debugging coming to webkit like Opera Dragonfly
16:05
<MikeSmith>
yeah, Patrick been working on that for a while now, I think
16:08
<karlcow>
I wish scope was standardized and implemented everywhere, so you could use any tools for debugging any devices
16:24
<karlcow>
http://www.facebook.com/note.php?note_id=10150103900258920
16:24
<karlcow>
>To accomplish this, we had to engineer a system that could process over 20 billion events per day (200,000 events per second) with a lag of less than 30 seconds.
16:50
<matjas>
Okay, I’m having some trouble understanding http://www.whatwg.org/specs/web-apps/current-work/multipage/tokenization.html#script-data-end-tag-open-state and http://www.whatwg.org/specs/web-apps/current-work/multipage/tokenization.html#script-data-escaped-end-tag-open-state
16:50
<matjas>
Can somebody translate it for me into human language? :)
16:50
<Ms2ger>
matjas, that's a sign of sanity
16:51
<matjas>
Specifically, I’m wondering what happens when you use <script>document.write('<foo></foo>')</script> — is that alright as per that algorithm or not?
16:51
<Ms2ger>
What is alright?
16:51
<matjas>
The “</foo” part
16:52
<Ms2ger>
What do you mean by alright?
16:52
<matjas>
Does it not trigger the end of the <script> element?
16:52
<matjas>
(I’d assume it would not)
16:52
<zewt>
on a quick read, it's ignored because it's not an "appropriate end tag token", right?
16:53
<Ms2ger>
You can test in Fx4, latest Chrome or Opera's test build
16:53
<matjas>
Well it works fine there
16:53
<Ms2ger>
Well, they do what the spec requires
16:53
<matjas>
Their HTML5^H parsers follow the spec perfectly? :)
16:53
<matjas>
Good to know
16:53
<matjas>
HTML 4.01 said the character sequence “</” is treated as terminating the end of the <script> element
16:54
<matjas>
I know that was never really enforced/implemented by browsers
16:54
<Ms2ger>
HTML4 is like a cake
16:54
<Ms2ger>
It's a lie
16:54
<matjas>
:)
16:54
<matjas>
Okay, so it seems HTML5^H rectifies this (although I couldn’t understand that from reading the spec)
16:55
<Ms2ger>
Our bus factor for the HTML parser is pretty low
16:57
<matjas>
So <script>document.write('<script></script>');alert(1)</script> shouldn’t alert(1) as per HTML5^5
16:57
<matjas>
→ Same in HTML 4.01
16:58
<matjas>
But <script>document.write('<div></div>');alert(42)</script> should alert(42) just fine
16:58
<matjas>
Even though HTML “Cake” 4.01 said it shouldn’t
16:58
<matjas>
(Tested and verified using Fx 4’s wonderful HTML5 parser)
16:58
<zewt>
why do people keep calling FF4 "Fx4"? heh
16:59
<zewt>
firefox = FF, it's waaaay too late to try to change that abbreviation
16:59
<matjas>
I’ve always used Fx
16:59
<zewt>
i've only seen it in like the last week
16:59
<matjas>
FF started to sound lame ever since the invention of Follow Friday
16:59
<matjas>
I know krijnhoetmer always uses Fx
16:59
matjas
blames him
17:00
<zewt>
"fx" is applied to so many countless things it's a terrible abbreviation, too
17:00
<zewt>
"FF", not nearly so much
17:00
<matjas>
The capital F makes it stand out IMHO
17:05
<Ms2ger>
Fx has always been the official abbreviation, afaik
17:06
<zewt>
nobody really gets to "pick" abbreviations; ultimately the language decides, and I've only seen FF, for many many years
17:07
<Ms2ger>
Not anymore :)
17:07
<zewt>
just was curious why suddenly I was seeing a few people going "Fx" in the last week or so--it's just confusing
17:08
<tw2113>
i think the version number helped out
17:08
<tw2113>
Fx4
17:17
<erlehmann>
“64-bit Cubic Flash debunks 32-bit AS WITCHCRAFT!”
17:55
<gsnedders>
karlcow: Well, apart from nail varnish, no.
18:50
<Yuhong>
"I know that was never really enforced/implemented by browsers"
18:51
<Yuhong>
FYI, this is another artifact of HTML being based on SGML.
18:51
<Yuhong>
HTML4, that is.
18:52
<Yuhong>
In fact, this artifact applies to all elements that was declared in HTML4 as being CDATA, including for example the XMP element too.
18:57
<Yuhong>
The history: while the earliest libwww did not parse HTML as SGML, DanC were going to transition HTML to being based on SGML in 1992.
18:58
<Yuhong>
But then came Mosaic which didn't parse HTML as SGML, and it got so popular that...
19:03
jgraham
wonders why gsnedders is still banging on about nail varnish
19:47
<gsnedders>
jgraham: because karlcow mentioned it!
21:44
<GPHemsley>
Are there any live use cases of @hidden? I'm having trouble understanding why the spec isn't self-contradictory about @hidden. What's the intended usage and what isn't?
21:44
<GPHemsley>
And what about the fact that browser implement @hidden simply by setting 'display: none;"?
21:56
<jgraham>
GPHemsley: The intended use case is for cases where you want something to be removed from all renderings of the page (inc. audio, etc.) because it is not part of the currently relevant content
21:57
<GPHemsley>
but what is the definition of "relevant content" that allows this while disallowing, e.g., a tabbed interface?
22:00
<jgraham>
The idea is that a tabbed interaface could be linearized so that it was all visible at once
22:02
<jgraham>
I'm not clear if this is really a useful distinction or not
22:08
<zewt>
are you saying that !hidden also implies audio elements are inactive? that's not clear to me from the spec--it says that hidden elements are still active
22:08
<jgraham>
No, I meant an audio presentation of the webpage e.g. via some AT
22:09
<zewt>
i've never been able to find any purpose to that attribute--that is, any reason to ever use it instead of display: none
22:14
<jgraham>
The theory is that you can always disable CSS without changing the meaning, but you can't remove @hidden without changing the meaning
22:14
<jgraham>
Like I say this may be a purely theoretical concern
22:14
<zewt>
that sounds like a theory that doesn't hold up in real world use
22:15
<jgraham>
It certainly isn't universially true that removing CSS leaves sites functional
22:15
<zewt>
eg. most nontrivial pages have display: none's that you can't simply disable--things tied to scripts and so on
22:16
<jgraham>
On the other hand one can argue that we should be trying to make the web more media-independent, not less
22:18
<zewt>
personally, rather than trying to make CSS purely presentational, which I think is entirely academic, it seems more useful to distinguish (at an authoring POV) between functional and presentational rules
22:19
<zewt>
just as my elements very often have functional class names (for scripts to detect) next to presentational ones (for CSS)
22:25
<GPHemsley>
so... would a multi-step installation process (e.g.) be a valid use case for @hidden?
22:26
<GPHemsley>
where all steps other than the one you're currently on are marked with @hidden
22:26
<Ms2ger>
I'd say yes
22:28
<GPHemsley>
interesting
22:32
<zewt>
i suppose the main use case is making pages not break in no-style mode, which is probably one of the few cases where functional CSS actually causes a problem
22:32
<GPHemsley>
ah; hmm
22:32
<zewt>
of course, the only time anyone does that is when a page's style is really, really bad, heh