IRC log of dig on 2012-03-22

Timestamps are in UTC.

00:00:52 [Pipian_]
Pipian_ has quit (Quit: Pipian_)
00:19:55 [timbl]
timbl (~timbl@c-24-62-225-11.hsd1.ma.comcast.net) has joined #dig
00:51:54 [tlr]
tlr has quit (Quit: tlr)
01:57:09 [nunnun]
nunnun is now known as nunnun_away
08:34:08 [Deiu]
Deiu (~deiu@157.159.103.120) has joined #dig
08:34:08 [Deiu]
Deiu has quit (Changing host)
08:34:08 [Deiu]
Deiu (~deiu@unaffiliated/deiu) has joined #dig
08:35:52 [bblfish]
bblfish has quit (Remote host closed the connection)
08:39:17 [tlr]
tlr (~tlr@88.207.162.149) has joined #dig
08:40:09 [bblfish]
bblfish (~bblfish@AAubervilliers-651-1-91-92.w83-112.abo.wanadoo.fr) has joined #dig
09:32:13 [cheater_]
cheater_ (~cheater@p4FD0E774.dip.t-dialin.net) has joined #dig
09:35:25 [timbl]
timbl has quit (Ping timeout: 244 seconds)
09:35:27 [cheater]
cheater has quit (Ping timeout: 265 seconds)
09:43:24 [rszeno]
rszeno has quit (Quit: Leaving.)
10:01:44 [Deiu]
'morning
10:02:22 [Deiu]
I've been looking into my problems with rdflib.js from yesterday and I still have no answer to why it doesn't work.
10:03:33 [Deiu]
if any of you guys are working on rdflib.js, I'd love to have a chat you :)
10:04:02 [Deiu]
Deiu has left #dig
10:04:03 [Deiu]
Deiu (~deiu@unaffiliated/deiu) has joined #dig
10:14:06 [melvster]
melvster (~melvin@p4FF977CF.dip.t-dialin.net) has joined #dig
11:17:37 [RalphS]
RalphS (Ralph@30-7-118.wireless.csail.mit.edu) has joined #dig
11:34:54 [cheater_]
cheater_ has quit (Ping timeout: 272 seconds)
11:41:53 [cheater_]
cheater_ (~cheater@p4FD0FD6D.dip.t-dialin.net) has joined #dig
12:05:14 [betehess_]
betehess_ (~betehess@c-24-61-44-145.hsd1.ma.comcast.net) has joined #dig
12:05:25 [betehess_]
betehess_ has quit (Client Quit)
12:25:51 [cheater__]
cheater__ (~cheater@p4FD0F059.dip.t-dialin.net) has joined #dig
12:26:29 [cheater_]
cheater_ has quit (Read error: Operation timed out)
12:29:59 [cheater_]
cheater_ (~cheater@p508954AB.dip.t-dialin.net) has joined #dig
12:32:50 [cheater__]
cheater__ has quit (Ping timeout: 260 seconds)
13:22:23 [bblfish]
Deiu, most people on this channel are in Cambridge USA, so 11am is a bit early
13:22:41 [Deiu]
No worries, I've done it without rdflib.js now. :)
13:22:55 [Deiu]
Which is better I think, since rdflib.js only works in Chrome.
13:23:23 [bblfish]
betehess: I fixed the play library I think for fetching resources WS.get https://groups.google.com/d/topic/play-framework/AKqQAQGkHjU/discussion
13:23:39 [timbl]
timbl (~timbl@31-34-149.wireless.csail.mit.edu) has joined #dig
13:24:16 [amy]
amy has quit (Ping timeout: 276 seconds)
13:24:23 [bblfish]
betehess: I did not end up needing the akka future to play promise future transform
13:25:49 [bblfish]
so those were two bugs in Play 2.0 framework which I fixed. https://play.lighthouseapp.com/projects/82401-play-20/tickets/269-wsexecutestream-always-returns-null
13:25:58 [bblfish]
I wonder if they'll accept them
13:26:10 [bblfish]
Deiu: rdflib works everywhere I think
13:26:24 [bblfish]
it was designed to work in Firefox
13:26:25 [Deiu]
It didn't work for me in FF.
13:30:24 [bblfish]
well I like your new solution. Getting rdflib to work, will just make it even more powerful.
13:30:57 [bblfish]
In any case rdflib does work in FF. So if you have an issue it's something that should not be a deep problem.
13:34:25 [amy]
amy (~amy@30-6-207.wireless.csail.mit.edu) has joined #dig
14:42:18 [Pipian_]
Pipian_ (~pipian@31-33-168.wireless.csail.mit.edu) has joined #dig
14:43:40 [mattl]
mattl (mattl@fsf/staff/mattl) has joined #dig
14:50:14 [mhausenblas]
mhausenblas (~mhausenbl@wlan-nat.fwgal01.deri.ie) has joined #dig
16:03:06 [bblfish]
how do I set the base for cwm?
16:03:11 [bblfish]
cwm --base=http://lifeshare.org/ posts/sioc_post_container.n3 --ntriples
16:03:14 [bblfish]
does not work
16:03:25 [bblfish]
it tries to fetch that document remotely
16:24:21 [bblfish]
mhh put a @base in the doc
16:30:30 [presbrey]
Deiu, what ffox version are you using? do you have an example url?
16:30:57 [Deiu]
hey presbrey
16:31:01 [Deiu]
I was using FF 11.0
16:31:13 [Deiu]
on ubuntu 10.04
16:31:17 [presbrey]
we have just moved to github; sorry its unstable
16:31:36 [Deiu]
No worries, I'm not using rdflib anymore, though I might use it in the future.
16:31:45 [Deiu]
(hence no example URL)
16:32:01 [presbrey]
but it used to work in firefox and I'd be happy for it to again
16:32:19 [presbrey]
if you can pastebin the example code, would be great
16:32:50 [presbrey]
did you find a better suited data api?
16:33:53 [Deiu]
I'm calling a php script which does the data parsing and transform, similar to a proxy
16:34:32 [Deiu]
you can probably view the script by looking at the source of http://webid.fcns.eu/connections/friends.php
16:34:35 [presbrey]
transform to html or json?
16:34:47 [Deiu]
html
16:37:39 [presbrey]
I thought timbl used rdflib.js on ffox 11
16:37:51 [presbrey]
eg. this app uses it: http://data.fm/proxy?uri=http://www.bbc.co.uk/music/artists/9fdaa16b-a6c4-4831-b87c-bc9ca8ce7eaa.rdf
16:38:45 [presbrey]
is that working in your copy of ffox? sorry I only have 10.0.1.
16:38:46 [Deiu]
might have been my system then
16:39:10 [Deiu]
it does
16:40:22 [presbrey]
that uses mashlib.js, currently part of tabulator to do the transform
16:45:34 [presbrey]
mashlib concats rdflib.js, tabulator files, and sometimes jquery for rdfa support
16:46:56 [presbrey]
ofc it can use a proxy too normalizing to turtle to access other formats too
16:47:30 [presbrey]
but if you have to go to html, you might as well use php :o
16:48:58 [Deiu]
Yeah, also since not many hosts support CORS.
16:49:14 [Deiu]
s/support/have enabled
16:49:55 [presbrey]
you can use rdflib.js with a cors proxy
16:50:26 [presbrey]
eg. data.fm/proxy configures it for itself
16:52:29 [sandro]
btw, presbrey, I like your suggestion of adding "S"....
16:53:46 [presbrey]
PQRS? its a natural progression :)
16:56:58 [sandro]
Yes. And I found during the talk that I kept having to make a point of the streaming nature.
16:58:20 [bblfish]
ok. I have the tools now to build a CORS proxy too...
16:58:25 [presbrey]
aw sounds fun; sorry I missed it!
16:58:44 [bblfish]
Just kind of weird how the Play 2.0 people never ever say thanks, even when you fix their bugs
16:59:00 [sandro]
I'm sure we'll have more chanced to talk about, presbrey
17:20:02 [presbrey]
indeed sandro
17:20:32 [presbrey]
Deiu, if you try it your browser again, refer here: https://github.com/linkeddata/rdflib.js/wiki/Howtos for setting up cross-origin proxy
17:21:05 [Deiu]
presbrey, will do! :)
17:21:08 [Deiu]
thanks btw
17:22:44 [presbrey]
np sorry there is more bugs to fix and docs to write ahead, thanks for trying!
17:30:58 [Deiu]
Deiu has quit (Ping timeout: 252 seconds)
17:40:50 [timbl]
presbrey, http://data.fm/proxy?uri=http://www.bbc.co.uk/music/artists/9fdaa16b-a6c4-4831-b87c-bc9ca8ce7eaa.rdf with my firefox gives me an rdf tree because I hav tabulator and it has broken recently
17:41:27 [timbl]
that is, i have just notioced it is broken
17:45:17 [timbl]
what? "is incompatible with firefox 11 - has been disabled"
17:45:18 [timbl]
<em:maxVersion>11.*</em:maxVersion>
17:45:27 [presbrey]
you need 11.*.*
17:45:49 [presbrey]
pull latest, I think I fixed that already
17:46:23 [timbl]
thanks
17:46:57 [presbrey]
I'm surprised you got the rdf tree if tabulator was off, it should have given you the skin
17:47:34 [timbl]
maybe the accept header though is a personal setting in about:config
17:48:47 [timbl]
Works now, thanks
17:52:12 [timbl]
ReferenceError: jQuery is not defined
17:52:12 [timbl]
chrome://tabulator/content/js/panes/internalPane.js:50
17:52:16 [timbl]
of course
17:52:25 [presbrey]
yes perhaps np
18:11:54 [github]
github (~github@sh2.rs.github.com) has joined #dig
18:11:54 [github]
[tabulator] timbl pushed 1 new commit to master: http://git.io/-TtA_g
18:11:54 [github]
[tabulator/master] bug fixes - Tim Berners-Lee
18:11:54 [github]
github has left #dig
18:14:16 [danbri]
danbri (~danbri@cpc6-aztw25-2-0-cust83.aztw.cable.virginmedia.com) has joined #dig
18:15:16 [mhausenblas]
mhausenblas has quit (Quit: brb)
18:18:10 [cheater_]
cheater_ has quit (Ping timeout: 260 seconds)
18:21:17 [timbl]
Grumble. FF doesn't make it ieasy .. is you are rujnning each version and they are called Firefox10.app etc on your machine, it will upgrdae Firefox10.app to actually be Firefox11.
18:22:08 [timbl]
Plus, if you move a new version of Firefox into /Applications, OS X will track where the old one went and still conectthe dock icon to it
18:28:50 [timbl]
This seems to work in a pane:
18:28:51 [timbl]
if (typeof jQuery == 'undefined') var jQuery = tabulator.jQuery;
18:28:51 [timbl]
if (doc_uri) jQuery(div, myDocument).append('<tr><td>Editable</td><td>'+tabulator.sparql.editable(doc_uri)+'</td></tr>');
18:30:10 [timbl]
in fact you don't seem to need th emyWindow either
18:30:25 [timbl]
just jQuery(div)
18:31:34 [timbl]
I guess from div the dom is defined and so the window is too
18:35:02 [presbrey]
great will it work to stash that snippet in the docHTML template?
18:36:16 [timbl]
where?
18:40:18 [github]
github (~github@sh2.rs.github.com) has joined #dig
18:40:18 [github]
[tabulator] timbl pushed 1 new commit to master: http://git.io/txPJvA
18:40:18 [github]
[tabulator/master] Pass a jQuery from tabulator.xul to a pane as tabulator.jQuery -- seems to work on the surface - Tim Berners-Lee
18:40:18 [github]
github has left #dig
18:41:07 [Pipian_]
Pipian_ has quit (Quit: Pipian_)
18:47:13 [presbrey]
cool thx timbl does the $ work now?
18:49:12 [presbrey]
ofc that will not work on mashlib or tabulator-chrome
18:49:26 [presbrey]
or, tabulator.jQuery will not be defined there
18:49:37 [presbrey]
since tabulator.xul does not run?
18:57:31 [cheater]
cheater (~cheater@p508954AB.dip.t-dialin.net) has joined #dig
19:02:45 [timbl]
yes. well, if jQuery is defined from anywhere else it won't look there in tabulator.jQuery
19:02:56 [timbl]
so the mashlib should work
19:04:08 [timbl]
have to investigate the pth from ythat file to my mashglib
19:05:29 [Deiu]
Deiu (~Deiu@unaffiliated/deiu) has joined #dig
19:05:52 [Deiu]
Hello
19:06:22 [Deiu]
bblfish just told me you found a bug in rdflib.js related for FF
19:06:30 [Deiu]
s/for/to
19:09:13 [bblfish]
I think. Not sure where the logs are published for this group
19:10:32 [bblfish]
Deiu, check the commit in the source repo on git
19:11:00 [Deiu]
Yeah, on it now.
19:13:34 [timbl]
What bug, Deiu?
19:14:59 [Deiu]
I was trying to use rdflib.js for an example and it didn't work in FF.
19:15:06 [Deiu]
It worked in Chrome though.
19:15:14 [Deiu]
I was FF11 btw.
19:15:43 [timbl]
i'm using FF11 -- did you isolate the bug?
19:15:59 [Deiu]
No, because I couldn't really put my finger on it.
19:16:14 [Deiu]
The browser console did not report any errors.
19:16:36 [timbl]
The javascript console didn't?
19:17:00 [Deiu]
Yes
19:22:04 [bblfish]
Deiu btw, has implemented WebID on https://webid.fcns.eu/
19:22:24 [bblfish]
( but Deiu said he was still new to JS earlier )
19:22:56 [bblfish]
Deiu: I think the best is to start really small and build up your application line by line. JS is a bit tricky if you copy and paste
19:23:04 [Deiu]
Actually, I wanted to use rdflib.js on http://webid.fcns.eu/connections where it would allow clients to load the RDF data on their machines...
19:23:16 [bblfish]
that is really cool
19:23:46 [bblfish]
I think it probably won't work as well for timbl, because his foaf:knows relations are in a different file
19:24:14 [Deiu]
I'm still working on that, and ACLs too.
19:26:48 [bblfish]
btw, I met someone who said recently he could use OCaml to create JS
19:26:58 [bblfish]
could be a way to create much more reliable JS
19:27:17 [bblfish]
with typing and full ML
19:33:32 [timbl]
What won't work for me?
19:34:03 [timbl]
... http://www.w3.org/People/Berners-Lee/card#i ?
19:36:43 [timbl]
i mean http://webid.fcns.eu/connections ?
19:39:20 [Deiu]
timbl, can you log into that website?
19:40:39 [bblfish]
I was able to authenticate
19:40:41 [bblfish]
:-)
19:40:43 [bblfish]
in FF
19:40:53 [bblfish]
11.0
19:41:01 [bblfish]
and also saw all my connections
19:41:08 [bblfish]
and I could ping them by clicking a button
19:41:40 [presbrey]
not for me, I get: "Authentication: Response from delegate IdP was outside of the allowed time window
19:41:40 [presbrey]
"
19:41:58 [bblfish]
you did a good job ! I can even see James Gosling's picture :-) And he did not have a foaf - I just added his dbpedia URL
19:42:01 [melvster]
Deiu: i was able to authenticate and http://webid.fcns.eu/connections/friends.php works for me!
19:42:14 [timbl]
timbl has quit (Ping timeout: 245 seconds)
19:42:22 [bblfish]
you have to use FF
19:42:30 [bblfish]
(cause that's what's being tested)
19:42:43 [Deiu]
It works just as fine in Chrome (for me)
19:42:54 [Deiu]
but you might have an issue with the server cert
19:42:57 [melvster]
works for me in FF11
19:42:59 [presbrey]
oy I'm using chrome and got the above response outside of the allowed time window
19:43:00 [bblfish]
In Chrome I usually have a bit of an issue with the certificte though
19:43:42 [Deiu]
presbrey, I think that's related to libAuthentication
19:43:59 [Deiu]
try to reload the page and authenticate again
19:44:31 [presbrey]
yes thats what I did first, perhaps your idp is returning a cached response?
19:45:00 [presbrey]
also my profile is public so I should not need to authenticate to see it :)
19:45:03 [bblfish]
in chrome I just got again:
19:45:03 [bblfish]
Warning: stream_socket_client() [function.stream-socket-client]: unable to connect to ssl://bblfish.net:8443 (Connection timed out) in /home/auth/public_html/auth/arc/ARC2_Reader.php on line 232
19:45:03 [bblfish]
Warning: Cannot modify header information - headers already sent by (output started at /home/auth/public_html/auth/arc/ARC2_Reader.php:232) in /home/auth/public_html/auth/WebIDauth.php on line 501
19:45:24 [Deiu]
ssl://bblfish.net:8443 ?
19:45:38 [timbl]
timbl (~timbl@31-34-149.wireless.csail.mit.edu) has joined #dig
19:45:39 [melvster]
Deiu: this is what I get http://ubuntuone.com/7jclpeVvUNUPqi0q5DS8Xr
19:45:40 [Deiu]
what cert are you using, bblfish ?
19:45:46 [melvster]
very nice!
19:46:21 [Deiu]
presbrey, yeah, but it uses the authenticated person as a source for the list of friends
19:46:55 [presbrey]
auth.fcns.eu is responding with:
19:46:56 [Deiu]
you should also be able to add ?webid=URI (where URI is any webid)
19:46:59 [presbrey]
Location:http://webid.fcns.eu/connections/index.php??error=noVerifiedWebId&sig=t-TrgyvzkdyX...
19:47:08 [Deiu]
I see
19:47:27 [presbrey]
I see the Request URL:https://auth.fcns.eu/auth/index.php?authreqissuer=http://webid.fcns.eu/connections/index.php?
19:47:43 [Deiu]
presbrey, could it be that you don't have a valid cert description in your webid?
19:47:52 [Deiu]
hold on...
19:48:06 [bblfish]
ah yes. I think my certificate must have expired chrome is no longer sending it
19:48:18 [bblfish]
s/sending it/showing it/
19:48:20 [Deiu]
presbrey, https://auth.fcns.eu/auth/index.php?verbose=on
19:48:29 [bblfish]
I need to make a new cert
19:48:34 [Deiu]
tell me at which step the process fails
19:48:37 [Deiu]
bblfish, ;)
19:48:43 [presbrey]
seems to parse here: http://data.fm/test/webid?uri=http://presbrey.mit.edu/foaf%23presbrey
19:49:23 [Deiu]
I doubt it's a parsing error
19:49:35 [presbrey]
it says testing modulus fails
19:49:40 [Deiu]
aha
19:49:51 [presbrey]
I have 6 there so it probably does not checking enough of them?
19:49:59 [Deiu]
stops after the first 3
19:50:17 [Deiu]
hmm, might just change that
19:50:44 [presbrey]
the modulus comparisons should be cheap relative to the http get / parse; can you unlimit that?
19:50:46 [bblfish]
presbrey: you have 6 SANs>
19:50:51 [bblfish]
ah no
19:50:55 [bblfish]
6 modulii
19:50:56 [Deiu]
I will do it now
19:51:04 [presbrey]
:) thx!
19:51:47 [Deiu]
try again
19:52:47 [presbrey]
no now it just checks multiple sans, like bblfish thought. multiple modulii are not checked still
19:55:17 [Deiu]
weird
19:56:08 [Deiu]
I was limiting the check at 3 SANs, not at 3 modulii
19:56:49 [presbrey]
http://65dc19cbbb10403d.paste.se/
19:57:30 [presbrey]
the cert I am using now is the 2nd modulus in the file
19:57:45 [presbrey]
b2ab...
19:58:00 [Deiu]
weird...you have 6 modulii for the same certificate?
19:58:13 [presbrey]
no 6 modulii for the same webid
19:58:34 [Deiu]
now I understand
19:59:04 [Deiu]
I was expecting one modulus for each certificate (key)
19:59:21 [bblfish]
Deiu, you are correct
19:59:27 [bblfish]
presbrey's file is wrong
19:59:38 [Deiu]
yep
19:59:47 [presbrey]
really? I have a different modulus for my phone, tv pc, desk pc, work pc, etc.
19:59:59 [presbrey]
I don't need separate URIs for each one
20:00:00 [bblfish]
presbrey: you need to wrap each of your mod,expo pairs with a <cert:RSAPublicKey>
20:00:06 [bblfish]
it should be
20:00:25 [Deiu]
yeah, but the canonical way is with one module & exponent for each certificate that you use
20:00:30 [bblfish]
me :key [ mod ...; exp...], [ mod ..; exp ...] .
20:00:46 [presbrey]
ofc they all have the same modulus
20:00:50 [presbrey]
er, exponent
20:00:59 [bblfish]
yes, but still it's semantically wrong
20:01:12 [bblfish]
for example you can't add info for when you made each cert
20:01:15 [bblfish]
which is very useful
20:01:28 [bblfish]
you can't distinguish your keys
20:01:56 [bblfish]
you can't give them different names
20:02:13 [bblfish]
or if we come round to it put expiry dates on them
20:02:18 [bblfish]
etc...
20:02:26 [presbrey]
"it used to work"
20:02:34 [bblfish]
well it's lick that it does
20:02:41 [bblfish]
but you see it does not with Deius
20:02:50 [bblfish]
it depends on how the query works
20:02:55 [Deiu]
not sure it will work on foafssl.org either btw
20:03:27 [bblfish]
well it probably will, because in a rewrite I removed some code
20:03:35 [bblfish]
which I need to put back
20:03:58 [presbrey]
either way should be fine in representation
20:04:26 [bblfish]
well not really. semantically you are saying something a bit nonsensical.
20:04:34 [bblfish]
you are saying you have 1 key with 7 modulii
20:04:38 [bblfish]
that does not exist
20:04:44 [presbrey]
just because x509 doesn't do that doesn't mean it doesn't exist
20:04:52 [bblfish]
no that is different
20:05:00 [bblfish]
you are confusing the key with the certificate
20:05:10 [bblfish]
the public key can only be one thing
20:05:23 [bblfish]
the certificate can have more than one key if you want in pgp perhaps
20:05:27 [Deiu]
semantically, in your foaf card you are describing a <cert:key>
20:05:42 [Deiu]
which can only have one modulus
20:05:46 [bblfish]
well a cert:RSAPublicKey
20:06:15 [bblfish]
those are specified modulus and exponent are functional
20:06:50 [melvster]
technically an RSA key has exactly one modulus and one exponent, what's called a GPG Key can actually be several keys (primary and subkeys) and has some meta data, and X.509 cert has a key and some meta data
20:07:11 [bblfish]
yes, that's why the ontology distinguishes keys and certs
20:07:26 [presbrey]
I didn't make this up randomly
20:07:29 [presbrey]
I set this up a couple years ago by following somone else's example
20:07:35 [presbrey]
so I won't be the last person with this problem
20:07:43 [bblfish]
that's ok
20:07:48 [presbrey]
it is your choice as server admin as to whether you want to be strict or lenient
20:07:51 [bblfish]
we are adding simple test suites
20:07:54 [presbrey]
my server will allow both
20:07:58 [Deiu]
actually, the WebID spec should be the main source of information :)
20:08:19 [bblfish]
well if we are so lenient that we don't care at all then we'll end up with tag soup
20:08:20 [presbrey]
everyone is welcome on my server! even people who copied timbl's multiple modulus example :)
20:08:27 [Deiu]
:))
20:08:44 [bblfish]
that's ok. But if you have a bank they will be more careful
20:09:03 [bblfish]
so then you end up giving webid a bad reputation because things will work for you and not for others
20:09:11 [bblfish]
and people just get confused
20:09:32 [Deiu]
I guess what any relying party should do is first to conform to the spec, and then decide to be lenient or not
20:10:05 [bblfish]
being lenient sounds cool, but it makes things more complicated later
20:10:41 [presbrey]
you can recommend what you want but if you are too strict then you are the one giving a bad reputation bblfish
20:11:16 [bblfish]
well you just had an issue right now because someone implemented it carefully
20:11:20 [bblfish]
the spec
20:11:38 [bblfish]
it's better to explain to people what the error is
20:12:13 [bblfish]
that's what test suites are for.
20:12:39 [bblfish]
you can bet that the more "lenient" you are the easier it will be for someone to break the system
20:12:59 [bblfish]
and then we'll have egg all over our faces, just like github did recently
20:13:11 [presbrey]
possibly for other issues, but I disagree regarding multiple modulus
20:13:28 [bblfish]
well the place to take that is to the webid mailing list
20:13:35 [bblfish]
and ask for change of spec
20:14:13 [bblfish]
in any case as a key implementer of webid it would be cool if you had a spec compliant webid :-)
20:14:18 [bblfish]
just helps us...
20:15:19 [presbrey]
yes I will "fix" my webid eventually
20:15:51 [presbrey]
but you seem to be suggesting my server is too lenient
20:16:00 [bblfish]
you will see there are a lot of other good reasons for doing it
20:16:10 [bblfish]
if you look at the spec you'll see a few
20:16:15 [bblfish]
http://webid.info/spec/
20:16:19 [bblfish]
in the diagrams
20:16:28 [bblfish]
and there will be others
20:16:59 [presbrey]
so I would like to fix our server too; can you check specifically what is wrong?
20:17:04 [melvster]
presbrey: i think that's the way to go ... consider at another time when someone implements encryption ... they will want to use your public key to encrypt but then they have the challenge to select a modulus from a list ... again it's possible to decrypt with a range of keys, but I think it's cleaner to have the one key and one exp/mod pattern
20:17:42 [presbrey]
yes if you scroll up you will see I will fix my key; I would also like to fix our server... so?
20:17:45 [bblfish]
presbrey your server is probably ok
20:17:50 [melvster]
i didnt phrase that very well ...
20:17:50 [presbrey]
https://github.com/linkeddata/data.fm/blob/master/www/inc/webid.lib.php
20:18:03 [bblfish]
it's when you start doing reasoning that things will start going weird
20:18:43 [bblfish]
yep, your query is ok
20:19:19 [bblfish]
it's if you had a reasoning engine there it would notice
20:19:23 [presbrey]
thx #L33. so where do we go wrong? too lenient?
20:19:37 [bblfish]
no your foaf is wrong :-)
20:20:33 [bblfish]
deiu just parsed it differently
20:20:42 [bblfish]
or he is doing a bit of reasoning
20:20:59 [bblfish]
perhaps he is looking for all cert:key relations
20:21:23 [presbrey]
does the spec say to do that?
20:21:35 [bblfish]
then he just does one check, if the mod does not succeed he moves to the next
20:21:46 [presbrey]
I thought I copied the sparql from the spec exactly
20:21:52 [presbrey]
is the fcns source available?
20:21:55 [bblfish]
yes your SPARQL is good
20:21:58 [bblfish]
your foaf is wrong
20:22:22 [bblfish]
and not everyone will notice your foaf is wrong
20:22:31 [RalphS]
RalphS has quit ()
20:22:42 [bblfish]
so you see we have lenience
20:22:57 [bblfish]
but not everyone is lenint the same way, hence it is important to be spec compliant in your foaf
20:23:39 [bblfish]
:-)
20:23:56 [bblfish]
well it's cool that we have more people joining
20:24:22 [bblfish]
I am pointing people to data.fm all the time
20:24:46 [bblfish]
(well not all the time, I am coding too)
20:25:16 [presbrey]
cool I will fix my foaf ofc; I am more worried about the server because it affects others
20:26:01 [presbrey]
in time, I am already 5 minutes late for an appt; bbl
20:26:26 [Deiu]
back
20:27:20 [Deiu]
what I'm doing is to take each SAN and look for a modulus/exp pair that matches it
20:28:53 [Deiu]
https://github.com/WebIDauth/WebIDauth/blob/master/WebIDauth.php
20:29:07 [Deiu]
line 374 & 391
20:34:16 [Deiu]
Taken from the WebID spec: Because one could not rely on the modulus having been normalized, one would have to start with the WebID - bob:me and find all it's cert:key relations to objects - which we know to be keys - and then iterate through each of those keys' modulus and exponent, and verify if the normalised version of the value of those relation is equal to the numbers found in the certificate.
20:43:08 [bblfish]
anyway the saying is "be lenient in what you accept, and strict in what you produce""
20:44:08 [bblfish]
you are strict because everyone is leninent in non overlapping ways
20:44:26 [bblfish]
or each server tends to be lenient differently
20:45:51 [Deiu]
yes
20:46:05 [timbl]
bblfish, take presbrey's file which works in some implementations and not in others and put it as a test case
20:46:33 [timbl]
It is asking for security attacks if you have a situation which is interpreteted differently by different types of sofwatre.
20:46:43 [timbl]
SO define what will work and wat won't and make sure
20:46:54 [timbl]
every implementaton does the same thing.
20:47:04 [timbl]
That is my suggestion.
20:47:33 [timbl]
" guess what any relying party should do is first to conform to the spec, and then decide to be lenient or not" <-- I don't thing "leniency" is a godo idea. What property must the graph have?
20:47:45 [timbl]
if it has it it successed, if it doesn't it doiesn't
20:47:55 [bblfish_]
bblfish_ (~bblfish@AAubervilliers-651-1-292-172.w83-114.abo.wanadoo.fr) has joined #dig
20:47:59 [Deiu]
timbl, presbrey's server conforms to the specs, afaik
20:48:51 [bblfish]
bblfish has quit (Ping timeout: 246 seconds)
20:48:51 [timbl]
Then your server doesn't conform
20:48:55 [timbl]
you mean.
20:48:59 [Deiu]
and I also agree that leniency allows security holes
20:49:16 [bblfish_]
timbl, presbrey's foaf is not spec compliant
20:49:33 [bblfish_]
it works because few people turn on owl inferencing
20:49:40 [bblfish_]
bblfish_ is now known as bblfish
20:49:55 [bblfish]
so the SPARQL query succeeds
20:50:01 [Deiu]
my server conforms too, but I'm doing the logic myself, instead of relying on sparql
20:50:10 [Deiu]
(as per specs)
20:50:41 [timbl]
the spec shopuld be abolutely precise about the protocol of what wone party produces and th eother party expects.
20:50:48 [Deiu]
I think the sparql query allows for presbrey's foaf to pass
20:50:51 [timbl]
eithe owl doe sor doe not get used
20:51:18 [Deiu]
it's quite an interesting fact
20:51:32 [bblfish]
well we will want to have owl used in future
20:51:51 [bblfish]
those who have rdf that is not inconsistent will not have a problem
20:52:59 [bblfish]
currently the definition of cert:key cert:modulus and cert:exponent are owl:functional
20:53:50 [presbrey]
ok I'm back sorry
20:54:10 [Deiu]
welcome back
20:54:37 [bblfish]
so there is an importance in consistency that is not always evident immediately
20:54:46 [Deiu]
I think the sparql ask query allows for presbrey's key interpretation to pass
20:55:25 [bblfish]
we don't want people to be forced to add owl inferencing initially and this is what creates some wiggle room
20:55:43 [presbrey]
yes its a spec bug that the sparql and non-sparql procedure describe different evaluation methods
20:55:59 [bblfish]
Deiu: not really. the semantics of cert:key is in definined at the cert namespace
20:56:29 [bblfish]
the evaluations methods will succeeds equally if the rdf is good
20:56:43 [Deiu]
indeed
20:57:17 [bblfish]
and the RDF is not defined by the SPARQL
20:57:48 [bblfish]
it's a query mechanism
20:57:59 [bblfish]
your query is leninent, to save cpu time
20:58:19 [bblfish]
because it probably does not matter as you said earlier if these bugs appear for the server
20:58:23 [presbrey]
you may have missed some good points timbl made about security
20:58:33 [bblfish]
but as you saw they do appear for you as a logger-into-something
20:58:40 [Deiu]
I wonder if I should switch to sparql or not, to be more lenient in the future
20:58:41 [presbrey]
while you were rejoining as bblfish_?
20:58:55 [bblfish]
ah can you point me to the logger?
20:59:06 [Deiu]
I'll paste it on skype
20:59:06 [presbrey]
DIGlogger, pointer?
20:59:06 [presbrey]
See http://dig.csail.mit.edu/irc/dig/2012-03-22#T20-59-06-1
20:59:34 [Deiu]
bblfish, 20:46
21:00:01 [timbl]
"my server conforms too, but I'm doing the logic myself, instead of relying on sparql" means there are two different algorithms being used, which mean the spec is ambiguous.
21:00:05 [bblfish]
ok. we will add presbrey's file as a test case
21:00:14 [timbl]
excellent
21:00:15 [bblfish]
in fact our test suites usually do have test cases like that
21:00:34 [Deiu]
timbl, actually, the logic I follow comes from the spec itself
21:00:37 [bblfish]
mine used to
21:01:06 [bblfish]
but as I have been rewriting my implementations so often I have lost my advanced test suite
21:01:30 [Deiu]
last paragraph before section "3.2.4.3 Authorization"
21:01:56 [Deiu]
Because one could not rely on the modulus having been normalized, one would have to start with the WebID - bob:me and find all it's cert:key relations to objects - which we know to be keys - and then iterate through each of those keys' modulus and exponent, and verify if the normalised version of the value of those relation is equal to the numbers found in the certificate.
21:01:56 [presbrey]
I'm working out an LD test suite for node.js using nodeunit I'll push on github soon
21:02:11 [presbrey]
it would be very beneficial to all of us to collaborate on test suite :)
21:02:28 [bblfish]
I agree.
21:02:34 [presbrey]
has anyone used nodeunit or something else better?
21:02:59 [bblfish]
bergos put a lot of work in a test suite for webid
21:03:20 [presbrey]
basic things like header checks and in-band checks like updating graph sets
21:03:31 [bblfish]
here http://www.w3.org/2005/Incubator/webid/wiki/Test_Suite
21:03:43 [presbrey]
oh I mean to target the HTTP interface
21:03:57 [presbrey]
authentication can be a suite of it
21:04:14 [presbrey]
but also authorization, storage, streaming, etc.
21:04:18 [bblfish]
yes. we even had an EARL vocab of WEbID
21:06:27 [bblfish]
yes. I think betehess is working on a test suite too :-)
21:06:36 [bblfish]
those are very important
21:10:01 [presbrey]
we need tests for rdflib.js too; node seems like a good target there too
21:10:20 [presbrey]
maybe even some of the HTTP graph manip tests can be checked using rdflib.js
21:11:07 [presbrey]
betehess, can you publish the list of tests?
21:11:25 [Deiu]
btw, the bug you found regarding FF, was it related only to tabulator? (I didn't see a new push for rdflib.js on github)
21:13:17 [presbrey]
the most recent bugfixes are related to missing jquery in firefox but not in rdflib.js used vanilla or skin and I see you load jquery already
21:13:40 [Deiu]
the current version doesn't use rdflib.js
21:13:42 [bergi]
bergi (~bergi@essen062.server4you.de) has joined #dig
21:13:55 [bergi]
hi!
21:13:59 [Deiu]
hi bergi
21:14:10 [presbrey]
welcome!
21:15:23 [presbrey]
Deiu, its best to load jquery before rdflib.js; its xhr factory is cleaner for cors in standalone use
21:15:53 [bblfish]
bergi is the person who did a lot of work on the test suite
21:16:09 [bblfish]
bergi there was a bit of talk on this http://dig.csail.mit.edu/irc/dig/2012-03-22#T20-59-06-1
21:16:34 [Deiu]
presbrey, sec, I might still have a previous version with the javascript code
21:16:51 [presbrey]
oh great; just curious for a link to an existing test suite of linked data over HTTP
21:18:34 [bblfish]
we should work together on it...
21:19:07 [bblfish]
bergi: at some point betehess here might pop up - he is missioned with test suiting the w3c
21:19:50 [timbl]
timbl has quit (Quit: timbl)
21:20:07 [bblfish]
We have just been building the blocks for some very efficient servers
21:21:42 [bblfish]
btw. this is a cool presentation "#Akka 2.0: Scaling Up & Out With Actors" presented at http://nescala.org/
21:24:17 [bblfish]
bergi also has an very good WebID implementation
21:32:22 [Deiu]
off-topic: any of you guys going to ESWC 2012?
21:34:02 [bblfish]
last time I went and presented I had to pay $500 - and I don't ahve the money. I'll be in Lyon WWW2012
21:37:29 [Deiu]
well, my uni refused to pay for it, so if I go, I'm paying out of my own pocket
21:41:57 [github]
github (~github@sh2.rs.github.com) has joined #dig
21:41:57 [github]
[rwt.js] presbrey created master (+1 new commit): http://git.io/Mj6tPw
21:41:57 [github]
[rwt.js/master] rwt.js: add OPTIONS test - Joe Presbrey
21:41:57 [github]
github has left #dig
21:42:17 [presbrey]
rwt is read-write tests
21:42:41 [presbrey]
let me know if you can help :)
21:46:02 [bblfish]
ah that's useful... will look at it as soon as I am back on track. which should be soon... Need to see if my Iteratees work now...
21:46:34 [presbrey]
haha you are too generous. theres only 1 test so far!
21:51:02 [bergi]
presbrey: any plans for a earl nodeunit reporter?
21:52:04 [github]
github (~github@sh1-ext.rs.github.com) has joined #dig
21:52:04 [github]
[rwt.js] presbrey force-pushed master from ffca207 to de7073d: http://git.io/NwfHjg
21:52:04 [github]
[rwt.js/master] rwt.js: add OPTIONS test - Joe Presbrey
21:52:04 [github]
github has left #dig
21:52:14 [presbrey]
whats that bergi?
21:53:22 [presbrey]
I am new to nodeunit so mostly I will be focusing on the LD tests and look to someone else for the node stuff until I am up to speed
21:56:21 [bergi]
for junit i've written an EarlJunitRunListener class which translates the junit results into earl. nodeunit seems to have reporters for different formats. that could be useful
21:58:42 [github]
github (~github@sh2.rs.github.com) has joined #dig
21:58:42 [github]
[rwt.js] presbrey pushed 1 new commit to master: http://git.io/k4Am3g
21:58:42 [github]
[rwt.js/master] add README.md - Joe Presbrey
21:58:42 [github]
github has left #dig
21:59:51 [bblfish]
ah my code seems to work!
22:00:38 [Deiu]
I have to go...I'll be back tomorrow morning.
22:00:59 [Deiu]
Deiu has quit (Quit: Leaving)
22:01:00 [presbrey]
l8r Deiu; bergi, we are happy to have you add reporters if you are interested?
22:10:23 [bergi]
presbrey: currently i have to many open task, but there is the will to learn/do node and scala coding at least till end of this year.
22:11:41 [bblfish]
presbrey: I think Deiu someimes has node.js students
22:12:21 [bblfish]
one of them wrote this really good report on webid http://magnetik.github.com/node-webid-report/
22:12:31 [bblfish]
really excellent
22:12:33 [bblfish]
one of the best
22:13:41 [bblfish]
betehess: here is my new Play 2.0 scala code for fetching graphs asynchronously when we have a parser, and synchronously otherwise https://github.com/bblfish/Play20/blob/webid/framework/src/webid/src/main/scala/webid/GraphCache.scala#L116
22:15:36 [bblfish]
that's essentially 2-3 months of work I think, from going from a non asyn code to async code