01:54
<variable>
Hixie, try "ifconfig flushall" from Terminal as root - it /may/ help
02:12
<Hixie>
variable: thanks, will try if it happens again. I ended up solving it by installing any pending updates and rebooting twice.
02:12
<variable>
Hixie, all that command does is forces the mac to release any current DHCP leases - I duno if it would help
02:12
<Hixie>
well i tried "renew dhcp lease" in the pref panel and that had no effect
02:13
<Hixie>
i could see in the tcpdump log that it was asking for a DHCP lease
02:13
<Hixie>
and it was getting a reply
02:13
<Hixie>
it was just not doing anything with it
02:13
<variable>
yeah - that is why I said to flush it first
02:13
<Hixie>
k
02:13
<variable>
another solution might be to statically set it to the reply that it got
02:14
<Hixie>
well that's how i got on IRC earlier and how I got online to download the updates
02:14
<Hixie>
but that's not a great solution ;-)
03:01
Hixie
gets to the volume/muted content attribute thing
03:10
<Hixie>
i don't see any way to have a muted="" content attribute that is reflected by the IDL attribute, since it would mean that if the user had requested a blanket mute of all audio, and you had a page today (with no muted="" attribute), it would magically sprout a muted="" attribute in the DOM each time the document was parsed
03:10
<Hixie>
that seems suboptimal
04:10
<Hixie>
i've no idea how to do this
04:12
<variable>
Hixie, do what?
04:18
<Hixie>
variable: the muted thing i mentioned above
04:18
<variable>
oh - woops
10:49
<annevk>
http://twitter.com/sideshowbarker/status/25201432241 o_O
10:59
<MikeSmith>
annevk: the word "faktisk" is a model of precision and economy compared to "daadwerkelijk"
11:01
<annevk>
it sounds more like the Dutch "feitelijk"
11:01
<annevk>
than "daadwerkelijk"
11:01
<MikeSmith>
oh
11:02
<MikeSmith>
so it is
11:02
<annevk>
I was mostly surprised you tweet in Dutch now :)
11:03
<annevk>
or something very close to it anyway ;p
11:04
<MikeSmith>
heh
11:04
MikeSmith
finds daadwerkelijk
11:04
<MikeSmith>
oops
11:04
MikeSmith
finds http://www.mijnwoordenboek.nl/synoniemen/daadwerkelijk
11:05
<MikeSmith>
I've quit tweeting in English forever
11:19
<hsivonen>
http://domscripting.com/blog/display/99
11:20
<hsivonen>
looks like adactio has been promoting a trick that doesn't work with HTML5-compliant innerHTML
11:21
<hsivonen>
correction: it's not adactio but one of the commenters who suggests using the trick for executing innerHTML-inserted scripts
11:22
<annevk>
it's from 2007
11:23
<hsivonen>
annevk: which means that potentially a lot of people have read it
11:52
<MikeSmith>
speaking of bad advice, the following seems questionable:
11:52
<MikeSmith>
http://dev.opera.com/articles/view/efficient-javascript/?page=2#implicitconversion
11:53
<MikeSmith>
annevk: btw, does Tarquin still work at Opera?
11:53
<MikeSmith>
that article is otherwise great, btw
12:00
<annevk>
MikeSmith, yup
12:00
<MikeSmith>
cool
12:00
<MikeSmith>
dude should write more
12:00
<MikeSmith>
write more articles I mean
12:00
<MikeSmith>
at dev.opera.com or his own site
12:00
<MikeSmith>
he seems to have gone quiet some time back
12:04
<jgraham>
MikeSmith: Clearly we work him too hard
12:05
<jgraham>
And yeah, that advice sucks
12:05
<jgraham>
But it is from 2006
12:05
<MikeSmith>
true
12:06
<jgraham>
That is like 1000BC (Before Carakan) in javascript years
12:06
<MikeSmith>
yeah
12:07
<MikeSmith>
but to some degree, it seems to have been intended as general advice (though I to realize he makes a point there of saying it applies more directly to the then-shipping Opera JS engine)
12:07
<jgraham>
We should write a new article that actually applies to the mordern world
12:07
<jgraham>
MikeSmith: Yeah, but 2006 is before fast javascript in general
12:07
<jgraham>
isn't it?
12:08
<jgraham>
Unless I am off by a couple of years at least
12:08
<jgraham>
Some stuff still applies of course
12:09
<jgraham>
"don't use try/catch in inner loops" for example
12:09
<jgraham>
And he doesn't appear to say "don't use the with statement, ever", but if he did that would still apply
12:09
<MikeSmith>
jgraham: no, you right (about the dates)
12:10
<jgraham>
(he says the rather too weak "avoid using with")
12:10
<Philip`>
Ideally all articles about JS performance would come with benchmarks so you could see how relevant they are in your current browser
12:10
<Philip`>
The benchmarks would probably be rubbish but at least they'd be better than nothing
12:11
<jgraham>
(which is just wrong. You should assume that using "with" is forbidden by law and carrys a prison term)
12:12
<annevk>
i like with
12:12
<annevk>
sue me
12:12
<annevk>
i'll make a point about using it more
12:13
<MikeSmith>
every article about JS performance should probably have an accompanying script with a timer that triggers after 1 year ot 2 years at most and causes the contents of the pages to be replaced with the single sentence: "The chances of this article still being relevant are very low. Find a more up-to-date article instead."
12:13
<MikeSmith>
make that, replaced with the _two_ single sentences
12:14
<Philip`>
Also they should all start with a section about how to profile your code to find bottlenecks, before wasting time micro-optimising irrelevant parts
12:14
<Philip`>
(Do good profiling tools for JS actually exist?)
12:15
<gsnedders>
Philip`: JSC has one that disables JIT, at lesat
12:15
<Philip`>
That sounds the opposite of "good"
12:15
<gsnedders>
Well, it's the best I know of :PP
12:16
<Philip`>
It's kind of useless getting profiling data over a version of your code that is not the version whose performance you actually care about
12:17
<gsnedders>
Well, yeah.
12:17
Philip`
wonders if authors would get better performance improvements if browser developers focused on decent profiling support to help authors optimise their code, rather than focusing on making the JS engine faster
12:17
<gsnedders>
That just kinda says how diabolical profiling is
12:18
<gsnedders>
Philip`: Harder to market, though
12:18
<Philip`>
Development tools like Firebug seem good marketing points (which forced other browser developers to implement something similar)
12:19
<Philip`>
at least for authors
12:20
<Philip`>
and for users you claim that more real web sites will be optimised for your browser since you're making it easy for authors to do that
13:11
<MikeSmith>
somebody please help me out here… where in the Ecmascript spec does it actually specify that you can get the character at a particular index in a string value by using square-bracket syntax?
13:12
<MikeSmith>
…or even where it states that you can use that to get retrieve a particular element from an array?
13:13
<miketaylr>
MikeSmith: section 15.5.5.2 of http://www.ecma-international.org/publications/standards/Ecma-262.htm
13:13
<miketaylr>
(for string)
13:13
<miketaylr>
i think that was non-standard until es5
13:14
MikeSmith
reads http://localhost:8080/index.html#x15.5.2
13:14
<MikeSmith>
oops
13:14
<paul_irish>
Ooh! /me bookmarks.
13:16
<MikeSmith>
:)
13:16
<MikeSmith>
http://sideshowbarker.github.com/es5-spec/#x15.5.2
13:16
<MikeSmith>
oops
13:16
<miketaylr>
ooOOooo html
13:17
MikeSmith
copies the section number correctly this time
13:17
<MikeSmith>
http://sideshowbarker.github.com/es5-spec/#x15.5.5.2
13:17
<MikeSmith>
ah, sweet
13:17
<miketaylr>
:)
13:19
<MikeSmith>
hey, I initially misread step 8 of that algorithm as "If resultStr be a String of length 1,"
13:19
<MikeSmith>
it would be great if specs were actually written with that kind of language
13:19
<MikeSmith>
"If foo be bar, then
13:21
<Dashiva>
Only one day a year
13:22
<MikeSmith>
heh
13:22
<MikeSmith>
I guess that day passed recently
13:22
<MikeSmith>
too bad
14:26
<jgraham>
annevk: I have no idea why you would like making your code look more like it was written in Visual Basic
14:26
<jgraham>
If God had intended the web to be programmed in VB, He would have given Microsoft a monopoly
14:35
<Philip`>
If JS used VB's with syntax (where you write ".Height = 100" etc so it's easily distinguishable from local/global variable references), would that have avoided the performance problems it causes?
14:41
<jgraham>
Maybe
14:42
jgraham
hasn't actually programmed any VB for >10 years and had forgotten the . thing
14:42
<MikeSmith>
anybody know if there is interactive help or a man page or doc of some kind for the OSX "jsc" command?
14:44
<jgraham>
MikeSmith: No idea. I doubt it. Do you have a question?
14:44
<MikeSmith>
nah, nothing specific
14:44
jgraham
is unlikely to be able to help, but nevertheless
14:45
<MikeSmith>
just running the jsc shell
14:45
<MikeSmith>
and wondering if there is any help on the shell at least
14:45
<MikeSmith>
node shell has a .help
14:46
<MikeSmith>
but that is actually not very useful at all
14:46
<MikeSmith>
at this point
14:46
<jgraham>
Does it do more than just take javascript expressions and evaluate the results?
14:46
<jgraham>
There is a print() function to write to stdout if that helps
14:47
<MikeSmith>
node shell has a .break and .clear
14:47
<MikeSmith>
which are actually useful in the case of node
14:47
<jgraham>
It's not like command-line repls have been the main focus of javascript
14:47
<MikeSmith>
true
14:48
<jgraham>
Or non-browser use cases in general really
14:48
<MikeSmith>
a thing that the node shell does that I find highly annoying is that if you hit ^C it drops you out of the shell entirely and exists back to your bash prompt
14:49
<MikeSmith>
I don't think ^C in a shell or shell-like thing should do that
14:49
<MikeSmith>
that's what ^D is for
14:49
<Philip`>
Perhaps you could run it in rlwrap to get a more standard shell-like interface
14:49
<MikeSmith>
so I guess this .break thing is what you are meant to use instead
14:49
<MikeSmith>
what's rlwrap?
14:50
<MikeSmith>
…readline wrapper
14:50
<MikeSmith>
I see
17:10
<karlcow>
in http://mozillalabs.com/zaphod/2010/09/16/6/ spotted a http-equiv=”Content-Script-Type”
17:10
<karlcow>
<meta http-equiv=”Content-Script-Type” content=”application/narcissus” />
17:11
<karlcow>
http://www.w3.org/TR/REC-html40/interact/scripts.html#h-18.2.2.1
17:11
<annevk>
ugh
17:11
<annevk>
"Doing so will cause Narcissus to execute the various on* listeners specified for different elements. (Unfortunately, SpiderMonkey will also execute them, which may or may not cause issues)."
17:12
<karlcow>
yes it seems SpiderMonkey it doesn't remove the other. Kind of strange
17:24
<hsivonen>
application/narcissus is like application/chromeframe, expect the former is only for experimentation while the latter is marketed for non-experimental deployment
17:28
<annevk>
could have used a more "there be dragons" identifier
17:28
<annevk>
but I guess it doesn't really matter
17:29
<karlcow>
hmm application/crouching-tiger
19:25
<jacobolus>
is there any way to add focus/blur events to custom objects?
19:25
<jacobolus>
for instance, to get a blur event on a div for someone clicking on some other form element?
19:28
<Hixie>
jacobolus: i don't follow
19:28
<Hixie>
can someone with IE let me know what IE does for http://software.hixie.ch/utilities/js/live-dom-viewer/saved/647 ?
19:28
<gsnedders>
jacobolus: var div = document.getElementsByTagName("div")[0], form = document.getElementsByTagName("form")[0]; form.addEventListener("blur", function(e){div.dispatchEvent(e);}, false);
19:29
<gsnedders>
jacobolus: Or something like that?
19:29
<Hixie>
so far i have three browsers with three different results
19:29
<jacobolus>
Hixie: in other words, currently only form elements or something containing them gets key presses and focus/blur events right?
19:29
<gsnedders>
jacobolus: Oh, that'll throw
19:30
<jacobolus>
can I get a div to accept focus & key events?
19:30
<Hixie>
jacobolus: per the spec, all kinds of elements can be focusable, see http://www.whatwg.org/specs/web-apps/current-work/complete/editing.html#focus-management
19:30
<jacobolus>
oh, okay
19:30
<Hixie>
jacobolus: plus you can be "specially focusable" http://www.whatwg.org/specs/web-apps/current-work/complete/editing.html#specially-focusable
19:31
<Hixie>
it strikes me that contenteditable isn't on either of those lists
19:31
<Hixie>
but that would also make something focusable
19:31
<Hixie>
probably
19:31
<jacobolus>
Hixie: so if I have some random div in the middle of my page, can I make it focusable?
19:31
<Hixie>
sure, just set tabindex=0 on it
19:31
<jacobolus>
(or perhaps an SVG shape that I want to be selectable)
19:32
<jacobolus>
neat
19:32
<jacobolus>
let me try that
19:32
<Hixie>
SVG probably has an attribute like fcFocusableElementActivationBehaviorMatrix="focusable-element"
19:32
<jacobolus>
:)
19:32
<Hixie>
but i wouldn't know what it was off-hand
19:33
<jacobolus>
I can't just set tabindex on it?
19:33
<jacobolus>
that's too bad
19:33
<Hixie>
dunno if SVG has tabindex
19:33
<Hixie>
maybe
19:33
<jacobolus>
aha! "focusable"
19:33
<jacobolus>
at least in svg tiny 1.2
19:33
<Hixie>
wow that's remarkably succint
19:34
<jacobolus>
and the associated event is apparently "DOMFocusIn"
19:36
<jacobolus>
I wonder if any browsers support that
19:44
<jacobolus>
sorry for the obvious questions, but what exactly is the relation of svg tiny 1.2 to svg 1.2, and what specs are browsers trying to implement?
19:49
<jacobolus>
or are browsers just shooting for svg 1.1?
19:49
<jacobolus>
(which doesn't include this keyboard stuff as far as I can tell)
19:58
<Hixie>
hsivonen: i'm completely confused by http://www.w3.org/Bugs/Public/show_bug.cgi?id=9843
20:08
<hsivonen>
Hixie: the patch that landed in Gecko makes document.written scripts not block on style sheets. seems to match IE
20:09
<Hixie>
hsivonen: i don't think we should be adding things that make document.write('foo') work differently than data:text/html,foo
20:10
<Hixie>
at least not without clear compat need