04:30
mpilgrim
would like it noted for posterity that I am connecting to this channel from the Apple //e that my parents bought in 1983
08:38
<boogyman>
Can someone link me to the pages that highlights the differences between releases?
08:39
<annevk>
if you are talking about HTML5 there is http://html5.org/tools/web-apps-tracker
08:39
<annevk>
and also http://dev.w3.org/html5/html4-differences/#changelog
08:40
<boogyman>
I thought the purpose of this channel is for HTML5 discussion?
08:41
<annevk>
did you read the topic? ;)
08:41
<boogyman>
it mentions the organization and a log url, that's it
08:42
<annevk>
it also says "Please leave your sense of logic at the door, thanks!" at the end
10:48
<gsnedders>
annevk: happy birthday!
10:55
<micheil>
Hixie: on the websockets protocol; I like the look of just a single format for messages; although, the idea of "oh yeah, it's just always binary data" needs to be changed
10:55
<micheil>
annevk: now that gsnedders' has mentioned it, Happy Birthday!
11:15
<jgraham>
When people say Happy Birthday, it always puts "Unhappy Birthday" by The Smiths in my head
11:15
<jgraham>
annevk: But happy birthday irrespective of that
12:03
<gsnedders>
jgraham: https://code.google.com/p/html5lib/source/detail?r=5872b5724934b9dd4193758e1d11bbaa79239236
12:03
<gsnedders>
jgraham: Does that get a non-ugly prize?
12:03
<gsnedders>
(not quite what you described yesterday)
12:07
<jgraham>
gsnedders: Doesn't self.dataStream.read(chunkSize) just return the empty string if there is no more data?
12:07
<jgraham>
So you don't need to chech if data:
12:08
<Philip`>
Am I totally misunderstanding or shouldn't you be adding _bufferedCharacter onto the beginning of data, not onto the end?
12:08
<gsnedders>
jgraham: No, it returns None, I think
12:09
<gsnedders>
Philip`: Yes I should
12:09
<jgraham>
Also, where does _bufferedCharacter get reset?
12:10
<gsnedders>
jgraham: where is it missing getting reset?
12:10
<Philip`>
gsnedders: Sounds like it needs tests cases :-)
12:11
<jgraham>
gsnedders: Well I don't see the code that makes _bufferedCharacter be None or the empty string or whatever when there is no character to buffer
12:11
<jgraham>
if there was in the last chunk
12:11
<gsnedders>
jgraham: it is initially None, and it gets reset to None after it is used
12:11
<gsnedders>
line 354
12:12
<jgraham>
Oh
12:12
<jgraham>
OK
12:12
<jgraham>
Can you show me where dataStream.read returns None?
12:13
<gsnedders>
Well, actually, is it not dependent on the file object?
12:13
<jgraham>
Right, but a file always returns "" from .read when there is no more data, doesn't it?
12:13
<gsnedders>
Yeah, that appears to be true
12:14
<jgraham>
So that is what I would expect
12:14
<jgraham>
in file-like objects
12:15
<jgraham>
(also I am not quite convinced that detting bufferedCharacter to be "" when it is not needed and always doing bufferedCharacter + data should not be faster)
12:15
<jgraham>
(but I may be overestimating the optimisations in the python interpreter)
12:16
<gsnedders>
jgraham: premature optimization much?
12:17
<jgraham>
gsnedders: Well I like it because I think it makes the logic simpler
12:18
<gsnedders>
Bah
12:18
<Philip`>
jgraham: Seems like you should test that optimisation
12:19
<Philip`>
It doesn't seem a good idea if parsing a 5MB file results in an extra 20MB of memory-copies when you're prepending an empty string to each chunk
12:21
<jgraham>
That is indeed the worry
12:23
<Philip`>
Hmm, looks like it probably does do that optimisation, though it's not obvious exactly when it optimises strings
12:24
<jgraham>
Yeah, I get similar numbers for the two cases in a trivial testcase
12:24
<jgraham>
Which might be highly misleading
12:24
Philip`
was looking at the id()s of the strings
12:25
<jgraham>
That works too
20:05
<duryodhan>
hi ... what language's HTML5 parser in HTML5lib is the recommended one ?
20:06
<duryodhan>
(talking about http://code.google.com/p/html5lib/ )
20:07
<Philip`>
duryodhan: Only the Python version is being relatively actively maintained
20:07
<Philip`>
The others are obsolete and probably broken
20:07
<duryodhan>
thanks!
23:20
<annevk>
thanks for the b-day wishes!
23:20
annevk
should get some sleep
23:33
<AryehGregor>
Bleh, othermaciej isn't here.
23:33
<AryehGregor>
He mentioned something about how Apple backports fixes for significant standards issues in Safari, so I wanted to ask him if this could be backported. https://bugs.webkit.org/show_bug.cgi?id=40747
23:33
AryehGregor
needs to not forget about that.