IRC log of dig on 2010-08-08

Timestamps are in UTC.

00:20:55 [melvster]
melvster has quit (Ping timeout: 260 seconds)
00:43:38 [presbrey]
do jambos rdfwidgets work on the data wiki
00:44:52 [presbrey]
anyone have a demo? i only see demos against jambo.xvm looks like it might be arc?
01:55:06 [timbl]
timbl (~timbl@pool-96-237-236-72.bstnma.fios.verizon.net) has joined #dig
02:32:16 [kennyluck_]
kennyluck_ (~kennyluck@EM114-48-139-210.pool.e-mobile.ne.jp) has joined #dig
02:34:15 [kennyluck]
kennyluck has quit (Ping timeout: 246 seconds)
02:34:16 [kennyluck_]
kennyluck_ is now known as kennyluck
03:00:20 [nunnun_away]
nunnun_away is now known as nunnun
03:08:29 [nunnun]
nunnun is now known as nunnun_away
03:09:55 [nunnun_away]
nunnun_away is now known as nunnun
03:56:17 [melvster]
melvster (~melvster@p5797FA45.dip.t-dialin.net) has joined #dig
04:41:10 [lkagal]
lkagal has quit (Quit: lkagal)
05:52:45 [kennyluck]
kennyluck has quit (Quit: kennyluck)
14:10:50 [timbl]
Hmmm .. the wiki doesn't seem t be saving anyting .. just pretending to
14:16:16 [timbl]
hg commit -m "Still have problem with multiline strings but issue tracker getting better"
14:16:16 [timbl]
$ hg push
14:19:31 [timbl]
Alway sthe same length now:
14:19:33 [timbl]
sparql: update complete for <http://dig.xvm.mit.edu/wiki/tabulator/tracker/state> status=200, text.length=9612
14:19:43 [timbl]
It never actually adds stuff
14:19:51 [timbl]
That is today, worked yesterday.
14:35:07 [nunnun]
nunnun is now known as nunnun_away
15:28:05 [kennyluck]
kennyluck (~kennyluck@114-25-243-107.dynamic.hinet.net) has joined #dig
15:36:52 [lkagal]
lkagal (~lkagal@pool-96-237-240-136.bstnma.fios.verizon.net) has joined #dig
17:10:32 [mcherian]
mcherian (~mathew@bne75-8-88-161-125-97.fbx.proxad.net) has joined #dig
17:17:19 [mcherian]
mcherian has quit (Ping timeout: 265 seconds)
20:01:36 [presbrey]
should be working again now timbl, email has more info
20:06:30 [timbl]
Oh, great - thatnks -- i'll try it
20:06:49 [timbl]
If there are errors in future wil the thing respond 200 still?
20:08:33 [presbrey]
i hope not. cpp segfaults are 500s. pythons seem to be 200 weirdly. error handling is definitely a work in progress
20:09:46 [timbl]
It said something in the returned page about hte formatting becasue the soemthingorother flag had ben turned on -- mauybe turn that off and the 200 will go to 4xxx
20:13:16 [timbl]
Currently I have silly problem on client that it sends multi-line strings without escaping, I wondered whethe rthat crashed the thing.
20:14:13 [timbl]
"he above is a description of an error in a Python program, formatted
20:14:13 [timbl]
for a Web browser because the 'cgitb' module was enabled. "
20:14:55 [timbl]
It is of course nice to get the stack dump but not as 200, and I'd forego the fancy formatting too
20:15:21 [presbrey]
it is forgone in <1-- comment at the bottom
20:16:23 [timbl]
Yes. I mean if one could get a 500 delivered by turing odd the fancy formatting and just get 500 and the straight plain text that would be better
20:16:51 [presbrey]
i prefer python stack traces emailed
20:17:24 [presbrey]
to the server dev rather than lost to the client
20:18:34 [presbrey]
im not sure how sparql parse errors are handling now... maybe segfaults... perhaps should be a 40x
20:19:18 [mcherian]
mcherian (~mathew@bne75-8-88-161-125-97.fbx.proxad.net) has joined #dig
20:19:24 [timbl]
I like them in the client as well, so the client can emil them to the dev :-)
20:19:50 [timbl]
Ooops -- now maybe from changing a selection box twice in quick successsion
20:20:09 [timbl]
500 Internal Server Error
20:21:43 [timbl]
Its down again.
20:22:34 [timbl]
(Is theer a way I can see the server logs?)
20:23:10 [presbrey]
i have inserted another age for joelamba 67
20:24:40 [presbrey]
do you want server access
20:25:12 [timbl]
IS it easy?
20:25:21 [timbl]
I was lookig at http://dig.xvm.mit.edu/wiki/tabulator/tracker/state#Iss1281298627621
20:25:59 [presbrey]
its pretty easy. eric and i are working there together now. send me a ssh pubkey
20:30:12 [timbl]
ssh-rsa AAAAB3NzaC1yc2EAAAABIwAAAQEAtfHbgYKgOauFiZK/l6JcYAUhTR5IzYRXSy+YFO8iyOYhKTXH35+XmbCkNPuG9wEVAZGW8E0zlQDB9rTeRzL9m4ilvBCgouXNI86nebjVvWqngz6ErBDzvxCgvkuaCfhVoSqHklebCwsYgTZV8jqA4PPZpaNolhgmGfg2f3jONjxUDX/MU+KvDJR4TIBlt95Ts28VexJOV/ZA2RUPc1Nl3ra4uYeBZTQ11QuwVOi1Am5qVjZZYnFUh5cMUAcweO7DHmpOZ6oTzA9GM2L16nlqDcZ0GSKNAao18DVSz+R7KUPiZQQfsvrV4zebh/Ez0z+ck27LcNk4qvxG2ov1+/klFw== timbl@cascade5.home
20:34:45 [timbl]
http://lists.w3.org/Archives/Public/www-archive/2010Aug/0011.html <-- client log
20:36:12 [timbl]
Inserts seem to work but not deletes
20:36:39 [timbl]
Yes, I can insert a comment but not delete it
20:37:41 [timbl]
The attempt to delete doe snot delete the thing from the file, and reeturns 500
20:38:13 [timbl]
(yes directory index pages wit ACL control would be great)
20:39:32 [presbrey]
seemed like something a few rdfwidgets would be good at. esp since there is a rdf acl 'coming soon' page on jambos site ;)
20:40:29 [timbl]
Well, he had a ffox sidebar for ACLs which is perfect for the ffox user but would also be good to have widget-based one for others
20:41:13 [presbrey]
yes i installed the acl sidebar. works and very cool
20:41:36 [presbrey]
there were some patches to the apache module for that as well
20:41:59 [timbl]
How did you install it? He has it as a add-on?
20:42:25 [timbl]
So the wiki at the moment has ACLs on it?
20:43:25 [presbrey]
webid authentication is working so acls may as well but i havent checked
20:43:56 [presbrey]
i added you on dig.xvm.mit.edu
20:44:36 [presbrey]
the wiki is in /srv/dig/www/wiki
20:48:26 [timbl]
(There is also a bit of code which charles2 wrote i PHP to get someone up to speed from zero to having a new account and an early FOAF file after which everything else takes off - a user bootstap)
20:48:32 [presbrey]
no acls are not on right now
20:48:40 [presbrey]
just webids
20:49:11 [timbl]
Hmm .. dig.xvm prompts me for a password after offering the key
20:49:47 [presbrey]
does it run a self-issued ca
20:49:51 [timbl]
called something like "keygen" rather than "user bootstrap"
20:50:05 [timbl]
yes it does, yo have to go through that hoop
20:52:43 [presbrey]
ok retry that server timbl
20:53:48 [timbl]
500
20:53:58 [presbrey]
i meant retry the ssh
20:54:20 [presbrey]
i cant do much with the 500s without eric. those are the swobjects methods segfaulting python.
20:54:36 [timbl]
Oh, sorry
20:54:39 [presbrey]
it would be nice to get the keygen stuff stable in browsers. it seems like that they keep changing. esp. the ie activex
20:55:19 [presbrey]
even chromes is wildly underdeveloped
20:55:24 [timbl]
works
20:55:32 [timbl]
Strange: "Last login: Sun Aug 8 16:52:28 2010 from c-65-96-162-214.hsd1.ma.comcast.net" on my first login
20:56:01 [presbrey]
oops you found me
20:56:12 [presbrey]
you can go in as root directly with ssh or sudo from timbl
20:56:28 [timbl]
Someone just mailed foaf-protocols to say they have a compete flash-based client for webid
20:56:42 [timbl]
Trouble is it stores the user keys in flash storage
20:56:44 [presbrey]
i saw i havent tried it yet
20:56:54 [presbrey]
there must be an import/export
20:57:49 [presbrey]
i am still looking at their javascript tls
20:58:24 [presbrey]
flash storage... wont work in my ibrowser
20:58:34 [timbl]
no :)
20:58:54 [timbl]
and in general what a place to put serious user data -- unless you can get at it easily
20:59:31 [timbl]
Where are the logs from the server -- do they have more info?
20:59:53 [presbrey]
/var/log/httpd/error...
21:00:09 [presbrey]
everything should stay default on that server except /srv/... stuff
21:01:35 [timbl]
I don't intend to change things round, just look and possibly restart something if it is strightforward
21:01:43 [presbrey]
sure restart and check about
21:02:07 [timbl]
about?
21:02:11 [presbrey]
let me know if you change something so we can track non-standard setup issues to document
21:02:20 [timbl]
Totally
21:02:37 [presbrey]
i am only using yum search.... yum install... to add packages
21:02:50 [presbrey]
and compiling swobjects, mod webids, so far...
21:03:43 [presbrey]
there is an unfortunate ruby error caused from google bots you'll need to skip past in apache error log
21:04:10 [timbl]
drwx------. 2 root root 4096 Aug 8 16:47 httpd
21:05:20 [presbrey]
sudo su -
21:07:16 [timbl]
yes
21:09:21 [timbl]
What is redmine?
21:10:11 [presbrey]
redmine is an issue tracker i was using webid with
21:10:18 [presbrey]
for the webid modules
21:10:45 [timbl]
Ah yes .. but soon you will be using the wiki and tabulator :)
21:12:23 [timbl]
When we have the wiki working agin you'll have to hg pull and see what the issue pane doesn't have which you need
21:12:51 [timbl]
I can't wait until we are eating our own dogfood with that
21:13:23 [presbrey]
rdfns.com also runs there. i need to commit that somewhere...
21:16:57 [timbl]
eh? [Thu Aug 05 20:51:57 2010] [warn] [client 38.111.20.228] mo in error_log-20100808 ? Dates off in server log?
21:18:11 [timbl]
So anyway.
21:18:44 [presbrey]
the filename discrepancy has to do with rotation
21:18:47 [timbl]
We need ericP to get the swobjects working before we can delete thins but in the meantime we can add things
21:18:57 [presbrey]
check the last filename before is 20100805
21:19:16 [presbrey]
so i can see how this log would be inclusive of that range
21:19:23 [timbl]
I see
21:20:46 [presbrey]
do you have your delete statement
21:21:52 [presbrey]
we should log the sparql somewhere
21:22:04 [presbrey]
cgi print should get into the error log
21:22:38 [presbrey]
swobjects segfaults python so only gdb or strace with the sparql can help here i think
21:24:53 [presbrey]
i disabled redmine for now
21:25:20 [presbrey]
should just 404 for google
21:27:27 [timbl]
We should definitely log the sparql and in fact we should log it in later in N3 or provid eit in N3 so that people can look at it the logs using tabulator's buuble-like display
21:28:28 [timbl]
Like [] date "2010-05-06"; by ... :deletion { .,.. } ; insertion { ... }.
21:31:00 [timbl]
but yes, making the logs available and maybe put them near the thing edited and check them in
21:32:02 [clochette]
clochette (~clochette@ANantes-552-1-20-225.w86-203.abo.wanadoo.fr) has joined #dig
21:33:36 [clochette]
clochette has quit (Client Quit)
21:33:54 [presbrey]
log the client and sparql into the svn message
21:34:55 [timbl]
Would that be too much bloat for svn?
21:35:55 [timbl]
We could make a virtual nearby resource like foo/bar,svn
21:36:16 [timbl]
which pulls the SVN logs for foo/bar
21:36:53 [timbl]
and returns them in turtle and is pointed to from a link header from foo bar
21:37:12 [timbl]
from foo/bar
21:38:28 [presbrey]
depends how much updating is going on
21:38:35 [presbrey]
no reason all document updates would have to go in the same repository
21:38:51 [presbrey]
could be one per folder or one per document
21:40:03 [presbrey]
as specified in the .meta.n3
21:41:06 [timbl]
yes
21:41:12 [presbrey]
whats being used to access different revisions here, svn anyway
21:42:28 [presbrey]
i wonder if the svn client will send an ssl certificate
21:43:13 [presbrey]
you can check on the sparql in error-log now
21:43:17 [presbrey]
send some deletes
21:43:36 [presbrey]
ah you did
21:43:46 [timbl]
Just now yes
21:46:12 [timbl]
The INSERTs show in the error log but not the DELETEs
21:46:25 [timbl]
But the inserts returned 200
21:46:30 [timbl]
and the deletes 500
21:46:39 [presbrey]
i see deletes in the error-log
21:46:59 [timbl]
scuse me yes
21:48:38 [presbrey]
so we'll see in a sec if swobjects is compiled with gdb symbols
21:55:45 [timbl]
What is the communication between the python and C - separate processes or one?
21:57:05 [presbrey]
one
21:57:47 [presbrey]
hence segfaults get reported as 500s internal server error
21:58:35 [presbrey]
i think it would be silly to fork swobjects calls
21:59:55 [presbrey]
we already did a multiprocess demo with algae
22:10:02 [presbrey]
here backtrace for your delete http://pastebin.com/08UVwRHv
22:11:14 [presbrey]
wiki]# gdb --args python ./test.py tabulator/tracker/state.n3 ./delete.sparql http://dig.xvm.mit.edu/wiki/tabulator/tracker/state
22:23:45 [timbl]
lib/RdfDB.cpp:30
22:45:03 [timbl]
I wonder where we should accumulate test data
22:45:30 [timbl]
Must look wherethe DERI folks put it
22:46:11 [timbl]
... to new $rdf.IndexedFormula();
22:46:19 [timbl]
which was a bit of a mouthful
22:57:13 [timbl]
Fixed the multiline thing hopefully
23:05:30 [presbrey]
i consolidated the python handlers
23:06:21 [presbrey]
and you can send an x-devel header w/ post to see the final turtle
23:06:47 [presbrey]
since i always do that next when testing update with curl
23:08:29 [presbrey]
timbl i disabled the cgitb pretty errors as well
23:09:17 [timbl]
That will help the bandwidth :-)
23:09:39 [timbl]
Did it fix the 200 code?
23:15:02 [presbrey]
inserts are 200
23:15:50 [presbrey]
should have fixed python exceptions as 200 yes
23:19:17 [timbl]
Great.
23:19:59 [timbl]
Is not good to think you are storing data away and have the tabulator fill up with dat and in fact it is all lost! Also really puts off real users
23:20:27 [timbl]
DanC's rule#n -- never lose my keystrokes
23:21:45 [timbl]
By the way I added Append access to the ontology
23:21:50 [timbl]
dunno if i mentioned
23:22:21 [timbl]
so we should be abl eto give public append to incoming message lists
23:25:29 [timbl]
BTW you can help me fix my kerberos maybe ... I get "kinit: Unable to create principal for 'timbl': Configuration file does not specify default realm"
23:25:45 [timbl]
It used to work, maybe on diff machine
23:26:58 [presbrey]
specify the realm with timbl-at-csail.mit.edu or timbl-at-athena.mit.edu
23:27:06 [presbrey]
realms are in all caps
23:27:40 [timbl]
aaaaah
23:27:56 [timbl]
and how do I make it know that about me
23:28:10 [presbrey]
it what
23:29:01 [timbl]
is ther e a .kerberosrc which I should put that in?
23:29:09 [timbl]
the realm
23:29:18 [presbrey]
you can set your default-realm in /etc/krb5.conf
23:29:45 [presbrey]
setup ssh with something like [~]$ cat .ssh/config
23:29:45 [presbrey]
Host *.mit.edu
23:29:45 [presbrey]
GSSAPIAuthentication yes
23:29:45 [presbrey]
GSSAPIDelegateCredentials yes
23:30:19 [presbrey]
you may not want delegate
23:31:39 [timbl]
delegate means get tokens from anotehr process?
23:31:46 [timbl]
lik ethe token manager thing?
23:32:20 [presbrey]
delegate means across ssh to the next server for chained use from there
23:32:36 [timbl]
ah
23:32:45 [presbrey]
if you turn it off, you need to kinit on other machines to continue using your tickets
23:32:55 [timbl]
I see
23:37:57 [presbrey]
you can send your tickets with curl very easily like; curl --negotiate http;//
23:38:37 [timbl]
where -negotiate means "figure out what auth thye do and do it"?
23:38:38 [presbrey]
need to find out how to get my webid into a curlrc
23:38:48 [timbl]
not just the kerberos?
23:39:30 [presbrey]
yes negotiate in curl refers to the that in gssapi
23:56:23 [ericP]
hi guys
23:56:25 [ericP]
what up?
23:56:49 [ericP]
trop tard