IRC log of dig on 2012-01-20

Timestamps are in UTC.

00:07:54 [Pipian__]
Pipian__ has quit (Quit: Pipian__)
01:03:19 [presbrey]
hmm rszeno it seems we already do isExtension check for XPConnect
01:06:18 [presbrey]
everywhere I see
01:06:19 [presbrey]
if (!(typeof tabulator != 'undefined' && tabulator.isExtension))
01:06:29 [presbrey]
$rdf.Util.enablePrivilege("UniversalXPConnect")
01:11:12 [rszeno]
in 3.5 was some changes and seems this doesn't work
01:11:56 [rszeno]
i make it work by changeing settings in prefs.js
01:16:58 [rszeno]
maybe is useful http://pastebin.com/8WhnaxUv
01:18:33 [rszeno]
in first link is a solution given by alien51, last two talk about user_pref in prefs.js
01:51:36 [tlr]
tlr has quit (Quit: tlr)
05:39:43 [presbrey]
presbrey has quit (Quit: Terminated with extreme prejudice - dircproxy 1.2.0)
08:11:02 [bblfish]
bblfish (~bblfish@AAubervilliers-651-1-342-31.w83-114.abo.wanadoo.fr) has joined #dig
08:45:03 [bblfish]
bblfish has quit (Quit: bblfish)
08:46:18 [bblfish]
bblfish (~bblfish@ALagny-551-1-33-93.w90-3.abo.wanadoo.fr) has joined #dig
09:06:59 [timbl]
timbl (~timbl@host109-156-61-106.range109-156.btcentralplus.com) has joined #dig
09:39:01 [timbl]
timbl has quit (Quit: timbl)
09:54:47 [melvster]
melvster (~melvin@p4FF97358.dip.t-dialin.net) has joined #dig
10:22:51 [danbri]
danbri has quit (Ping timeout: 252 seconds)
10:39:27 [tlr]
tlr (~tlr@ip-88-207-233-25.dyn.luxdsl.pt.lu) has joined #dig
10:43:55 [danbri]
danbri (~danbri@cable-146-255-156-245.dynamic.telemach.ba) has joined #dig
11:13:26 [bblfish]
ok, so there is one good reason to use pure HTTP Proxy though
11:13:53 [bblfish]
and that is that the semantics of a pure HTTP Proxy are much clearer
11:14:10 [bblfish]
ie: I don't have to change the URLs of the document
11:14:27 [bblfish]
or parse it
12:24:48 [Ralph]
Ralph (Ralph@30-7-118.wireless.csail.mit.edu) has joined #dig
12:25:23 [Ralph]
Ralph is now known as RalphS
13:33:54 [timbl]
timbl (~timbl@host109-156-61-106.range109-156.btcentralplus.com) has joined #dig
13:39:46 [bblfish]
timbl, presbrey was saying you had written some code for tabulator to go through a proxy
13:39:51 [bblfish]
to avoid CORS
13:43:10 [bblfish]
I am just starting to need that for https://github.com/bblfish/rdflib.js/commit/2c0ee65e30cca014849a8065d513e3cf8fcaccb6
13:43:13 [bblfish]
oops
13:43:22 [bblfish]
I meant for http://bblfish.github.com/rdflib.js/example/people/social_book.html
13:46:21 [timbl]
ummm.... yes .. I Was building that and I found the problem was recognizing when in fact he cross-site scripting blaock had happened.
13:46:42 [bblfish]
ah yes, I thought that might be difficult to work out
13:47:55 [bblfish]
I was thinking of setting up a pure HTTP Proxy service because that should be very efficient to set up
13:48:11 [bblfish]
it would just add cors
13:48:12 [bblfish]
a
13:48:29 [bblfish]
and it would be clear that my site was not making the statements.
13:48:56 [bblfish]
(efficient, because there would be no need to parse the rdf)
13:49:35 [bblfish]
But I don't understand quite what security CORS brings to the table
13:52:04 [timbl]
The code (untested and buggy) is in rdflib.js/web.js
13:52:12 [timbl]
grep crossSiteProxyTemplate
13:52:53 [timbl]
should be newURI = this.crossSiteProxyTemplate.replace('{uri}', encodeURIComponent(uri));
13:58:40 [bblfish]
ok thanks
13:59:48 [bblfish]
timbl, btw we have a new WebID Community Group http://www.w3.org/community/webid/
14:08:21 [timbl]
Great
14:31:26 [bblfish]
mho mind you I suppose that trying to connect through a proxy, might not work if the user is already behind a proxy.
14:31:47 [bblfish]
so, the pure proxy idea does not work that well….
14:53:58 [melvster1]
melvster1 (~melvin@p4FF97F1C.dip.t-dialin.net) has joined #dig
14:55:18 [melvster]
melvster has quit (Ping timeout: 272 seconds)
15:11:58 [timbl]
timbl has quit (Read error: Connection reset by peer)
15:12:10 [timbl]
timbl (~timbl@host109-156-61-106.range109-156.btcentralplus.com) has joined #dig
15:42:32 [bblfish]
I think it should be possible to find out why it did not work, because you can call xhr.getAllResponseHeaders()
15:47:21 [bblfish]
yep, it's working. probably going through the proxy
16:22:07 [timbl]
-- xhr.getAllResponseHeaders() how will that show that the cal was blcoked for CORS?
16:32:59 [bblfish]
trying to work out what is going on here. I thought it was working for a while, but ...
16:42:14 [bblfish]
I don't think it ever goes into state 0
16:42:21 [bblfish]
it always first starts with 1
16:42:31 [bblfish]
ie. OPENED
16:43:14 [bblfish]
then it goes over to 4
16:43:28 [bblfish]
at least you can see that in debug mode with http://bblfish.github.com/rdflib.js/example/people/social_book.html
16:50:44 [Pipian_]
Pipian_ (~pipian@31-33-50.wireless.csail.mit.edu) has joined #dig
16:52:28 [bblfish]
ah found a way to get a message in chrome
16:54:04 [bblfish]
it throws an exception in the case 4 handle response
16:54:30 [bblfish]
well there it calls the onerror method
17:31:27 [bblfish]
oh stack overflow has something http://stackoverflow.com/questions/5018566/catching-xmlhttprequest-cross-domain-errors
17:47:33 [bblfish]
I still completely fail to see what the point of CORS is
17:53:15 [amy]
amy has quit (Ping timeout: 260 seconds)
17:58:32 [bblfish]
yes, so the only way to know it failed seems to be to catch an error. (Pretty inefficient really)
18:12:03 [presbrey]
presbrey (~presbrey@2001:4830:2446:b5:aede:48ff:fe00:2001) has joined #dig
18:12:36 [presbrey]
DIGlogger, pointer
18:12:36 [presbrey]
See http://dig.csail.mit.edu/irc/dig/2012-01-20#T18-12-36
18:27:14 [timbl]
bblfish, from http://stackoverflow.com/questions/5018566/catching-xmlhttprequest-cross-domain-errors :- 1) On CORS requests, the browser's console.log will display an error message. However, that error message is not available to your JavaScript code (I think this is done for security reasons, I asked this question once before: Is it possible to trap CORS errors?).
18:27:14 [timbl]
2) The xhr.status and xhr.statusText aren't set in the onError handler, so you don't really have any useful information as to why the CORS request failed. You only know that it failed.
18:27:27 [bblfish]
yes
18:27:48 [timbl]
but onError() takes a exception param surely
18:28:22 [bblfish]
I have just rewritten the code to move it into onError
18:28:29 [bblfish]
in a second or two it will be live
18:28:35 [timbl]
I have alreday: xhr.onerror = function(event) {
18:28:35 [timbl]
sf.failFetch(xhr, "XHR Error: "+event)
18:28:35 [timbl]
}
18:28:57 [timbl]
failFetch will leave a status in the tabulator metadata
18:29:39 [timbl]
I don't remember seeing it
18:31:09 [bblfish]
well it takes a bit more than an error
18:31:12 [bblfish]
a second
18:33:45 [timbl]
Ahhh .. with safari I see it: [18:32:27.732] XHR Error: [object XMLHttpRequestProgressEvent]
18:34:04 [bblfish]
arg
18:34:15 [timbl]
So a XMLHttpRequestProgressEvent does't have a toString
18:34:17 [bblfish]
I can't get my caches to update to the latest
18:34:32 [bblfish]
version
18:34:45 [bblfish]
mhh
18:34:56 [timbl]
and see what it is
18:35:15 [bblfish]
silly me
18:36:45 [RalphS]
RalphS has quit (Ping timeout: 252 seconds)
18:36:54 [bblfish]
ok it's updated
18:44:28 [Ralph]
Ralph (Ralph@30-7-118.wireless.csail.mit.edu) has joined #dig
18:45:04 [timbl]
http://opensource.apple.com/source/WebCore/WebCore-737.5/xml/XMLHttpRequestProgressEvent.h
18:57:07 [amy_]
amy_ (~amy@30-6-207.wireless.csail.mit.edu) has joined #dig
19:09:35 [bblfish]
well the changes are up there in rdflib.js
19:09:36 [bblfish]
view-source:http://bblfish.github.com/rdflib.js/example/people/social_book.html
19:09:41 [bblfish]
but it does not quite do it
19:10:29 [bblfish]
I moved the code to onerror
19:10:53 [bblfish]
http://bblfish.github.com/rdflib.js/example/people/social_book.html
19:12:15 [bblfish]
ie here: https://github.com/bblfish/rdflib.js/blob/master/web.js
19:12:28 [bblfish]
but I need it to run the next xhr2
19:19:44 [bblfish]
ah it does not reach there
19:24:57 [presbrey]
I think we shouuld drop the XPConnect exceptions
19:25:20 [presbrey]
hardly worth breaking the dataskin on firefox
19:35:27 [bblfish]
I am not sure what is meant to happen with this xhr2 https://github.com/bblfish/rdflib.js/blob/master/web.js#L665
19:40:33 [bblfish]
mho perhaps it works, but no longer calls my callback
19:42:57 [rszeno]
simplest way to check is to use a callback like this function(){ document.body.style.backgroundColor = '#000'; }
19:43:50 [bblfish]
mhh yes
19:43:51 [bblfish]
thanks
19:44:14 [rszeno]
you are welcome, :)
19:44:20 [melvster1]
presbrey: http://webcredits.org/ :)
19:45:28 [rszeno]
in firebug there is a net panel where you can watch xhr.
19:52:12 [bblfish]
well yes. IT does fetch it. It's just I need to have xhr2 have the same handlers as xhr initial
19:52:27 [bblfish]
perhaps if I just copy the handlers over?
19:55:47 [rszeno]
with initial xhr, open method will reuse it as it is
19:56:39 [rszeno]
but results are tricky, if is async you will never know which open respond
19:57:13 [rszeno]
if is sync, second will not run untill first is finished, so second is useless
20:00:08 [bblfish]
I call it like this
20:00:08 [bblfish]
fetch.nowOrWhenFetched(docURI, undefined, function() {
20:00:09 [bblfish]
card(person,kb)
20:00:10 [bblfish]
friends(person,kb,col+1)
20:00:11 [bblfish]
});
20:03:53 [rszeno]
is nothing wrong with it imo
20:04:47 [rszeno]
if card, friends do what you expect and if kb is populated i think should work
20:06:16 [rszeno]
you are far before me, last thing i tried was to populate kb, but i want to do this in ffox, :)
20:07:30 [bblfish]
™mhh
20:07:31 [bblfish]
ok
20:08:08 [bblfish]
it's the definition of https://github.com/bblfish/rdflib.js/blob/master/web.js#L545
20:08:12 [bblfish]
nowOrWhenFetched
20:08:26 [bblfish]
this.addCallback('done', function(uri2) {
20:08:26 [bblfish]
if (uri2 == uri) callback();
20:08:27 [bblfish]
return (uri2 != uri); // Call me again?
20:08:28 [bblfish]
});
20:08:46 [bblfish]
of course if we use the proxy, then uri2 !- uri
20:09:28 [rszeno]
first call will answer with 'unrequested' and will triger the last if
20:10:07 [rszeno]
then probably will fire again
20:10:52 [rszeno]
sta = this.getState(uri), first line :)
20:11:16 [rszeno]
sta will be 'unrequested'
20:11:38 [github]
github (~github@sh1-ext.rs.github.com) has joined #dig
20:11:38 [github]
[rdflib.js] presbrey pushed 2 new commits to master: http://git.io/pZu97g
20:11:38 [github]
[rdflib.js/master] web.js: continue without XPConnect - Joe Presbrey
20:11:38 [github]
[rdflib.js/master] jquery.xmlns.js: add upstream fixes - Joe Presbrey
20:11:38 [github]
github has left #dig
20:12:01 [github]
github (~github@sh1-ext.rs.github.com) has joined #dig
20:12:01 [github]
[rdflib.js] presbrey force-pushed gh-pages from 81020e7 to 8b15d85: http://git.io/rfmeFA
20:12:02 [github]
[rdflib.js/gh-pages] gh-pages: add dist - Joe Presbrey
20:12:02 [github]
github has left #dig
20:12:09 [bblfish]
why without XPConnect?
20:12:28 [bblfish]
is that not something that works on IE?
20:12:46 [presbrey]
no thats a Firefox thing
20:12:52 [presbrey]
and CORS can provide access without it
20:13:19 [rszeno]
doesn't work in ffox
20:13:25 [bblfish]
ah it's XDomainRequest that's IE
20:14:05 [bblfish]
rszeno, no, it never calls my "done" handler because of that if statement
20:14:25 [bblfish]
perhaps if I remove the if?
20:14:41 [bblfish]
I wonder why it is there though
20:14:57 [rszeno]
it manage the state
20:15:03 [bblfish]
why does the callback check the URIs
20:15:16 [rszeno]
one of 'fetched, done or unrequested
20:15:35 [bblfish]
in nowOrWhenFetched
20:15:41 [bblfish]
https://github.com/bblfish/rdflib.js/blob/master/web.js#L545
20:15:58 [rszeno]
yes, lines 545 to 555
20:16:59 [bblfish]
yes. How do I overcome that if we change the URIs to be using proxy?
20:17:15 [bblfish]
perhaps just do a proxy text too?
20:17:34 [bblfish]
(that feels quite hawkish)
20:17:42 [bblfish]
hackish
20:18:04 [rszeno]
uri2 != uri in proxy case?
20:18:17 [bblfish]
yes, currently
20:18:42 [bblfish]
see the xhr2 area of code
20:18:53 [bblfish]
627->
20:19:44 [bblfish]
oh I know. I'll set the URI back then and then call the done method again?
20:20:07 [bblfish]
at least it puts the code together
20:20:40 [melvster1]
bblfish: does this help? http://hacks.mozilla.org/2009/07/cross-site-xmlhttprequest-with-cors/
20:20:46 [rszeno]
i guess this will run only in case of error
20:20:57 [melvster1]
request.withCredentials
20:22:14 [bblfish]
mho, I think my problem is just now with the code web.js that I am hacking
20:22:19 [bblfish]
but that could be useful
20:22:45 [melvster1]
a bit on xhr2 and preflighted requests too
20:23:41 [bblfish]
no xhr2 is a variable in the web.js code
20:23:45 [bblfish]
it's not a standard
20:26:19 [amy_]
amy_ is now known as amy
20:26:36 [presbrey]
look here in latest Firefox
20:26:38 [presbrey]
http://data.fm/proxy?uri=http://bblfish.net/people/henry/card
20:26:52 [bblfish]
yes?
20:27:00 [presbrey]
got it past the XPConnect but now its not seeing content-type or any other headers
20:28:30 [melvster1]
crashed my ff :)
20:30:16 [presbrey]
arg sorry
20:30:22 [melvster1]
lol np
20:30:36 [rszeno]
3.6 didn't crash but give an error
20:31:12 [rszeno]
"HTTP failed unusually. (no handler set) (cross-site violation?)"
20:31:48 [rszeno]
and work in rest, :)
20:57:54 [kennyluck_]
kennyluck_ (~kennyluck@114-43-116-53.dynamic.hinet.net) has joined #dig
21:01:07 [kennyluck]
kennyluck has quit (Ping timeout: 276 seconds)
21:01:08 [kennyluck_]
kennyluck_ is now known as kennyluck
21:14:38 [Pipian__]
Pipian__ (~pipian@30-9-3.wireless.csail.mit.edu) has joined #dig
21:16:36 [Pipian_]
Pipian_ has quit (Ping timeout: 240 seconds)
21:18:43 [Pipian__]
Pipian__ has quit (Client Quit)
21:19:28 [Ralph]
Ralph has quit ()
21:28:41 [melvster1]
melvster1 has quit (Quit: Leaving.)
21:39:20 [bblfish]
ok done http://bblfish.github.com/rdflib.js/example/people/social_book.html
21:39:39 [bblfish]
it now uses presbrey's proxy
21:39:58 [bblfish]
not sure the code is the best
21:40:10 [bblfish]
well very likely it is not
21:40:45 [bblfish]
ran into lots of bugs because of there being no typing in javascript
21:42:15 [bblfish]
mho let me check the code in
21:47:42 [bblfish]
mho, well now it really does not work in Firefox
21:47:53 [bblfish]
but Chrome and Safari do work
21:48:50 [bblfish]
opera neither
21:52:37 [bblfish]
I have a feeling that one should not really be using CORS at all
21:52:42 [bblfish]
just slows things down
21:52:50 [bblfish]
proxies should work well enough
21:55:40 [bblfish]
1. use core, and fall back on proxy
21:55:52 [bblfish]
2. transform all urls trhough proxy directly
22:05:10 [bblfish]
good night folks
22:18:27 [melvster]
melvster (~melvin@p4FF97F1C.dip.t-dialin.net) has joined #dig
22:41:41 [presbrey]
bblfish, " please check that this.proxyUsed is the right " ?
22:45:48 [presbrey]
why did you move from state=0 to onerror?
22:50:51 [github]
github (~github@sh1-ext.rs.github.com) has joined #dig
22:50:51 [github]
[rdflib.js] presbrey pushed 5 new commits to master: http://git.io/6HbTBA
22:50:51 [github]
[rdflib.js/master] getting the template to work - Henry Story
22:50:51 [github]
[rdflib.js/master] was missing full path for $rdf.Fetcher.crossSiteProxy - Henry Story
22:50:51 [github]
[rdflib.js/master] was not using the right uri. xhr.uri is the one. - Henry Story
22:50:51 [github]
github has left #dig
22:51:00 [github]
github (~github@sh1-ext.rs.github.com) has joined #dig
22:51:00 [github]
[rdflib.js] presbrey force-pushed gh-pages from 8b15d85 to a753c20: http://git.io/rfmeFA
22:51:00 [github]
[rdflib.js/gh-pages] gh-pages: add dist - Joe Presbrey
22:51:00 [github]
github has left #dig
23:04:50 [github]
github (~github@sh1-ext.rs.github.com) has joined #dig
23:04:50 [github]
[tabulator] presbrey pushed 2 new commits to master: http://git.io/zBdZ0A
23:04:50 [github]
[tabulator/master] mashlib: testing proxy support - Joe Presbrey
23:04:50 [github]
[tabulator/master] Makefile: add PHONY targets - Joe Presbrey
23:04:50 [github]
github has left #dig
23:24:05 [timbl]
PHONY ?
23:28:06 [presbrey]
tells make not to look for a real file called 'update' or 'all'
23:28:14 [presbrey]
designates them as phony/fake targets
23:46:19 [bblfish]
hi
23:47:43 [bblfish]
presbrey: state 0 was never called
23:47:53 [bblfish]
in Safari and chrome at least
23:49:23 [timbl]
state 4 is the only one you can reply on
23:49:53 [timbl]
the other states get called 1 after open, 2 ehen headers, 3 as reading data, etc but not reliably especially if thise things don't happen
23:52:01 [bblfish]
please review my code in detail if you can. I am new so there must be improvements to be made
23:52:13 [bblfish]
I am just trying to get things working