IRC log of dig on 2010-08-05

Timestamps are in UTC.

00:47:38 [nunnun_away]
nunnun_away is now known as nunnun
02:06:19 [timbl]
timbl (~timbl@pool-96-237-236-72.bstnma.fios.verizon.net) has joined #dig
03:57:05 [kennyluck]
kennyluck (~kennyluck@EM114-48-10-67.pool.e-mobile.ne.jp) has joined #dig
04:23:35 [kennyluck]
kennyluck has quit (Ping timeout: 265 seconds)
05:18:09 [melvster]
melvster (~melvster@p5797FDCC.dip.t-dialin.net) has joined #dig
05:38:24 [kennyluck]
kennyluck (~kennyluck@EM111-188-24-54.pool.e-mobile.ne.jp) has joined #dig
06:07:27 [oshani]
oshani (~oshani@c-67-161-2-233.hsd1.ca.comcast.net) has joined #dig
06:08:24 [oshani]
oshani has left #dig
06:13:18 [kennyluck]
kennyluck has quit (Ping timeout: 265 seconds)
09:29:58 [knappy]
knappy has quit (Quit: knappy)
10:49:10 [kennyluck]
kennyluck (~kennyluck@EM111-188-1-159.pool.e-mobile.ne.jp) has joined #dig
11:06:49 [RalphS]
RalphS (~swick@30-7-139.wireless.csail.mit.edu) has joined #dig
11:11:54 [kennyluck]
kennyluck has quit (Quit: kennyluck)
11:12:26 [kennyluck]
kennyluck (~kennyluck@EM111-188-1-159.pool.e-mobile.ne.jp) has joined #dig
11:35:32 [kennyluck]
kennyluck has quit (Quit: kennyluck)
12:01:34 [kennyluck]
kennyluck (~kennyluck@EM111-188-13-105.pool.e-mobile.ne.jp) has joined #dig
12:25:42 [kennyluck]
kennyluck has quit (Quit: kennyluck)
12:26:26 [kennyluck]
kennyluck (~kennyluck@EM111-188-13-105.pool.e-mobile.ne.jp) has joined #dig
12:31:16 [kennyluck]
kennyluck has quit (Ping timeout: 258 seconds)
12:39:19 [Yudai]
Yudai has quit (Ping timeout: 258 seconds)
12:49:13 [Yudai]
Yudai (~Yudai@p034a6d.kngwnt01.ap.so-net.ne.jp) has joined #dig
13:20:40 [timbl]
timbl has quit (Ping timeout: 240 seconds)
13:51:39 [Yudai]
Yudai has quit (Ping timeout: 264 seconds)
13:52:19 [Yudai]
Yudai (~Yudai@p034a6d.kngwnt01.ap.so-net.ne.jp) has joined #dig
14:23:26 [timbl]
timbl (~timbl@31-35-252.wireless.csail.mit.edu) has joined #dig
14:39:18 [Yudai]
Yudai has quit (Ping timeout: 258 seconds)
14:41:13 [Yudai]
Yudai (~Yudai@7c29565d.i-revonet.jp) has joined #dig
14:48:03 [oshani]
oshani (~oshani@c-67-161-2-233.hsd1.ca.comcast.net) has joined #dig
14:52:21 [mcherian]
mcherian (~mathew@c-24-61-8-94.hsd1.ma.comcast.net) has joined #dig
14:56:37 [nunnun]
nunnun is now known as nunnun_away
15:24:13 [mcherian]
mcherian has quit (Ping timeout: 265 seconds)
15:32:43 [nunnun_away]
nunnun_away is now known as nunnun
16:17:18 [knappy]
knappy (~haoqi@dhcp-18-111-36-78.dyn.mit.edu) has joined #dig
16:36:54 [lkagal]
lkagal has quit (Quit: lkagal)
16:49:41 [mcherian]
mcherian (~mathew@30-6-25.wireless.csail.mit.edu) has joined #dig
16:59:14 [presbrey]
ericP, you around?
17:02:28 [ericP]
presbrey, here
17:18:24 [ericP]
hmm, no hints in preprocessor output
17:18:36 [ericP]
i'd'a figured va_list was some crazy macro
17:20:54 [oshani]
oshani has quit (Quit: Mama nidi!)
17:21:36 [presbrey]
I have a fix for Makefile echo on that server as well
17:27:11 [ericP]
rock on!
17:29:14 [ericP]
[[
17:29:16 [ericP]
Yes;swig in some versions has a bug in that it violates C. The way
17:29:16 [ericP]
va_args are done on AMD64 is not conpatablewith swig. Indeed I have
17:29:16 [ericP]
complained about this on this list,and to teh swig developers. I alwats
17:29:16 [ericP]
had to hand-edit the output from swig to get iyt to compile. Since updating operating system this went away -- not sure exactly
17:29:19 [ericP]
what/what but probably the change from Suse10.2->Suse10.3 If you wat I can did out the necessary changes.
17:29:22 [ericP]
]] — http://csound.1045644.n5.nabble.com/Problem-compiling-csound-on-amd64-td1132920.html#postCount1132921
17:30:25 [presbrey]
I'm not positive now what you meant yesterday about echo '-e' flag
17:30:34 [presbrey]
since my fix is simply adding it
17:31:28 [ericP]
i haven't found a recipe that works on ubuntu and e.g. mac yet
17:31:57 [ericP]
the echo in gnu make on the mac didn't not support -e
17:32:10 [lkagal]
lkagal (~lkagal@30-6-179.wireless.csail.mit.edu) has joined #dig
17:33:33 [presbrey]
oh
17:33:48 [presbrey]
ok its regular make works fine
17:33:55 [presbrey]
regular echo
17:33:55 [presbrey]
primac:~ presbrey$ uname -a
17:33:56 [presbrey]
Darwin primac.local 10.2.0 Darwin Kernel Version 10.2.0: Tue Nov 3 10:37:10 PST 2009; root:xnu-1486.2.11~1/RELEASE_I386 i386
17:33:56 [presbrey]
primac:~ presbrey$ echo -e 'a\nb'
17:33:57 [presbrey]
a
17:33:59 [presbrey]
b
17:34:26 [presbrey]
how about a 'configure' script then
17:34:37 [ericP]
yeah, i guess i'll have to go there
17:34:48 [ericP]
oof, those things alway bloat terribly
17:35:24 [presbrey]
Makefile is 527 lines :P
17:35:25 [ericP]
i feel like there should be a site that has a database of recipes. you click the features you need and it gives you a configure
17:35:37 [ericP]
fair point
17:37:48 [presbrey]
CONFIG_DEFS seems fairly isolated at the top there, shouldn't be too hard
17:39:49 [presbrey]
can you get a fix/patch for the cxx?
17:40:09 [presbrey]
I'm going to fix apache and all the fastcgi stuff
17:40:11 [ericP]
nope -- till dinking with va_lists of va_lists
17:40:29 [presbrey]
almost everything broke when I upgraded yesterday
17:40:49 [presbrey]
eg. rdfns.com
17:42:12 [ericP]
pita
17:42:46 [presbrey]
whats the current bug with data wiki that adding a timeout fixes?
17:44:35 [presbrey]
timbl
17:44:48 [ericP]
i think i know this one
17:45:00 [ericP]
given two successive updates, the first was lost
17:45:20 [presbrey]
that doesn't sound like a bug with algae
17:45:26 [presbrey]
?
17:45:46 [ericP]
i don't believe it is, just need a mutex around the calls
17:46:13 [ericP]
though it could be argued that it will be easier to add threading to SWObjects than to algae
17:46:30 [presbrey]
ahh ok
17:47:04 [ericP]
for now, just a mutex in the python would do the job
17:47:29 [presbrey]
depends how many pythons run eh? maybe flock would be easier
17:47:36 [presbrey]
we could get that in php already
17:47:38 [ericP]
yeah, i was just getting there
17:50:17 [ericP]
what the hell, this takes a long time to compile
17:50:20 [presbrey]
cool apparently g++ uses >400M swap to compile the 3M SWObjects swig cxx
17:50:27 [presbrey]
ha
17:50:39 [presbrey]
so thats why
17:52:06 [ericP]
well, i commented out the functions which called c++ with varargs
17:52:14 [ericP]
and one that calls them
17:52:42 [ericP]
i was hoping to play around with a real fix, but i don't want to compete for resources
17:52:57 [presbrey]
I'm looking into more ram
17:53:31 [ericP]
yeah, .5G isn't much today
17:53:55 [presbrey]
no its fine with 4 on my worky
17:54:54 [presbrey]
whenever you're done shut it down and I'll try increasing it
17:55:39 [ericP]
hmm, i wonder how long this compile will take
17:56:04 [ericP]
could kill it and restart, with more ram, could take advantage of all the work it's done so far
17:56:28 [ericP]
let's kill it in 4 mins
18:03:37 [presbrey]
ok?
18:03:49 [presbrey]
kill it 3mins ago?
18:04:42 [ericP]
kill kill!
18:07:56 [presbrey]
I think I'm all out of ram quota for guests
18:08:08 [presbrey]
do you have an athena cert you can try modifying it with?
18:14:27 [ericP]
presbrey, i do have an athena cert
18:14:40 [ericP]
can you give me some guidance?
18:16:42 [knappy]
knappy has quit (Quit: knappy)
18:17:58 [ericP]
presbrey, i started my all-consuming compile again
18:20:10 [presbrey]
its https://xvm.mit.edu/machine/1430/
18:21:14 [presbrey]
whats your athena name?
18:33:04 [nunnun]
nunnun is now known as nunnun_away
18:34:56 [nunnun_away]
nunnun_away is now known as nunnun
18:42:33 [knappy]
knappy (~haoqi@dhcp-18-111-36-78.dyn.mit.edu) has joined #dig
18:55:32 [ericP]
presbrey, eric_ath, i think
18:56:33 [ericP]
make -k python-test worky
18:56:35 [presbrey]
were you able to access that url?
18:57:05 [ericP]
yep
18:57:16 [ericP]
i should shut it down first?
18:57:30 [presbrey]
yes it won't let you change the ram while its powered
18:57:55 [presbrey]
anytime is fine if the make finished
18:58:15 [ericP]
yup
18:58:17 [ericP]
$G?
18:58:19 [ericP]
4G?
18:58:31 [presbrey]
yeah def
18:59:07 [ericP]
Maximum 512 MiB for w3
18:59:07 [presbrey]
is that your max?
18:59:26 [presbrey]
is that the only machine listed?
18:59:36 [presbrey]
for you on https://xvm.mit.edu/list ?
19:00:09 [ericP]
i don't see a machine listing
19:00:32 [presbrey]
`List` is in the top-left
19:01:03 [ericP]
You don't currently control any VMs.
19:01:14 [ericP]
maybe this one is out of control
19:01:55 [presbrey]
you can still see https://xvm.mit.edu/machine/1430/ ?
19:02:09 [presbrey]
try changing the owner to yourself there
19:02:21 [presbrey]
I think we may just end up emailing xvm@mit.edu
19:04:10 [ericP]
you mean, change the administrator to eric@csail.mit.edu ?
19:04:20 [presbrey]
no, owner
19:04:40 [presbrey]
I dont think it supports csail stuff, just athena
19:05:02 [ericP]
ahh, where it says the owner is w3
19:07:03 [ericP]
but still a max of 512
19:07:53 [ericP]
i changed it back to w3 and powered it on
19:08:04 [ericP]
the python stuff should work
19:08:48 [ericP]
i can dink arouund i a bit for a less hacky way to fix the va_list stuff
19:09:06 [ericP]
(by emulating in a tiny file and compiling that)
19:16:18 [mcherian]
mcherian has quit (Ping timeout: 265 seconds)
19:19:18 [knappy]
knappy has quit (Quit: knappy)
19:21:21 [ericP]
presbrey, /usr/local/src/rdfwiki works
19:21:36 [ericP]
well, it's just the unit tests copied to another dir
19:21:52 [ericP]
two ideas come to mind for how this could work
19:25:49 [ericP]
g = open(form["default-graph-uri"])
19:26:34 [ericP]
fcntl.flock(g, LOCK_EX)
19:27:14 [ericP]
one, a RdfDB which is file backed -- each
19:27:16 [ericP]
oops
19:27:18 [ericP]
tparser = SWObjects.TurtleSDriver("", F)
19:27:32 [ericP]
tparser.setGraph(referenceDB.ensureGraph(SWObjects.cvar.DefaultGraph))
19:28:07 [ericP]
tparser.parse(SWObjects.IStreamContext(form["default-graph-uri"], SWObjects.StreamContextIstream.STRING))
19:28:15 [ericP]
... execute the query
19:28:20 [ericP]
,,, close g
19:29:37 [ericP]
s/SWObjects.IStreamContext(form["default-graph-uri"]/SWObjects.IStreamContext(turtleStr/ # where turtleStr is read from g
19:29:59 [ericP]
what would be cooler is to have an RdfDB backed by a filesystem
19:31:47 [presbrey]
test_update in the unittests looks helpful to me
19:32:25 [knappy]
knappy (~haoqi@dhcp-18-111-36-78.dyn.mit.edu) has joined #dig
19:32:41 [ericP]
INSERT { GRAPH <http://dig.xvm.mit.edu/rdfdocs/path/and/file> { <s> <p1> <o1> ; <p2> <o2> } } would synchronize with $(ROOT)/path/and/file
19:32:45 [presbrey]
oh, my friends TurtleSDriver and RdfDB.assureGraph!
19:33:13 [ericP]
yeah, i've been meaning to change assureGraph for ages -- makes me blush
19:33:23 [ericP]
momentary typo gone viral
19:34:10 [presbrey]
so I don't recall the way data wiki works right having anything to do with a GRAPH <> pattern/expression(?)
19:34:19 [presbrey]
right now*
19:35:35 [ericP]
maybe it doesn't (or at least, it's users don't) use GRAPH{} directives -- just functions like "INSERT { <s> <p1> <o1> ; <p2> <o2> }" on the default graph
19:35:52 [ericP]
which could be set by form["default-graph-uri"]
19:37:31 [ericP]
another option is [[
19:37:48 [ericP]
g = open("theDB.trig")
19:37:49 [presbrey]
sprintf("$bin_algae -b %s -d $data_file -f $tmp_sparql_input --sClass rdfxml --dump-default >$tmp_data_output 2>$tmp_error_output")
19:38:05 [ericP]
tparser = SWObjects.TrigSDriver("", F)
19:38:23 [presbrey]
so, cwm is also getting called in the data wiki now. thats news to me. timbl?
19:38:37 [ericP]
...]], which would have only one file containing the whole db
19:39:20 [presbrey]
if the request subpath is /rss/ ?... kennyluck could be helpful here as well
19:39:34 [ericP]
so currently you have some way outside the SPARQL query to control $data_file
19:40:09 [ericP]
and presumably, updates from SPARQL queries with nested GRAPH<x>{}s are simply lost
19:40:35 [presbrey]
how are they serialized ?
19:40:57 [ericP]
which "they"?
19:41:01 [presbrey]
it seems only the default graph is dumped by algae in this mode
19:41:05 [ericP]
you mean in the current algae?
19:41:11 [presbrey]
meaning only updates to the default graph will get saved
19:41:19 [presbrey]
the data_file is based on REQUEST_URI
19:41:31 [ericP]
--sClass rdfxml --dump-default tells me that the dfault graph gets dumped as rdfxml
19:41:56 [presbrey]
the model here is opposite the favorite /sparql?query= 1 sparul endpoint to rule them all
19:42:37 [ericP]
not sure it's favorite, but yeah, it doesn't seem to honor GRAPH updates
19:42:44 [presbrey]
you can sparul any uri under /data/ and the updates get saved to the file behind the uri
19:42:45 [ericP]
or queries for that matter
19:42:59 [ericP]
ahh, i get it
19:43:12 [presbrey]
the file behind the uri is sourced as the default-graph, and its request-uri as the default-graph-uri, as well
19:43:12 [ericP]
that is sort of natural
19:44:12 [ericP]
http://dig.xvm.mit.edu/rdfdocs/path/and/file?query=INSERT { GRAPH </some/other/file> { <s> <p> <o> } } would be odd to honor
19:44:19 [presbrey]
i think this mode even provides for your 'what would be cooler is to have an RdfDB backed by a filesystem' desire
19:44:37 [presbrey]
yes, that is awkward
19:44:49 [presbrey]
?query= isn't supported. sparqul is only taken via POST
19:45:23 [ericP]
gotcha, but the issue of selecting one resource to update another would still be weird
19:46:08 [presbrey]
yes, that is very awkward in this model
19:46:19 [presbrey]
I don't think it should be supported
19:46:56 [presbrey]
it doesn't make sense to me
19:47:05 [presbrey]
esp. for nix-based, file-backed RDF sparul'ing
19:48:15 [ericP]
but how am i going to attack /etc/passwd without it?
19:48:21 [presbrey]
ha
19:48:50 [presbrey]
keeping sparqul/graph=>http/uri tightly coupled also lets us easily slip in other cool #swig stuff based on URI like WebID authn/authz etc.
19:49:35 [ericP]
so [[ g = open(fpath(REQUEST_URI)) ]] it is
19:50:11 [presbrey]
where is your foaf, anyway ericP, I don't seem to foaf:knows you
19:50:49 [ericP]
http://www.w3.org/People/Eric/ericP-foaf
19:51:06 [ericP]
i'm pretty lame about my foaf
19:51:22 [ericP]
but i do know one person
19:51:32 [presbrey]
you're missing out on the webid craze?
19:51:48 [ericP]
(and i have a DTD for my foaf, which is a trifle unusual)
19:52:01 [ericP]
so far, yes, really really missing out
19:54:40 [presbrey]
wow this is going to be so much better
19:54:54 [presbrey]
almost exactly a year ago
19:55:05 [presbrey]
[root@dig conf.d]# cat sparul.conf
19:55:11 [presbrey]
# mod_sparul
19:55:12 [presbrey]
#LoadFile /usr/local/src/swobjects/trunk/lib/libSWObjects.a
19:55:12 [presbrey]
LoadFile /usr/lib64/libstdc++.so.6
19:55:12 [presbrey]
LoadModule sparul_module modules/mod_sparul.so
19:56:42 [timbl]
presbrey, cwm being called I suspect is a think Keenyy added
19:56:58 [timbl]
he was plaing with a scheme for uploading proofs of data to the wiki
19:57:03 [timbl]
playing
19:57:22 [timbl]
I think that should probably be removed.
19:57:38 [presbrey]
ok thx. I suspected
19:59:41 [timbl]
ericp, "ericP: maybe it doesn't (or at least, it's users don't) use GRAPH{} directives -- just functions like "INSERT { <s> <p1> <o1> ; <p2> <o2> }" on the default graph"
20:00:08 [timbl]
The read-write data web by doing sparql updates to the URI of the document you want modified.
20:00:31 [timbl]
The default graph is the graph of that doc.
20:02:13 [presbrey]
the doc of the request uri
20:03:07 [presbrey]
its also important that the request-uri gets used as the default-graph-uri so serializing after an update doesn't produce file:/// references from relative paths
20:03:13 [ericP]
so it would be polite to fail noisily when encountering GRAPH<x> | x != REQUEST_URI
20:04:58 [presbrey]
probably
20:05:16 [ericP]
presbrey, does it seem to do the right thing no wrt base uri?
20:05:24 [ericP]
i haven't thought about it at all
20:05:32 [presbrey]
I haven't checked swobjects, algae did
20:05:48 [ericP]
knobs to turn: the base uri supplied to the parsers
20:06:50 [ericP]
sparser = SWObjects.SPARQLfedDriver($(REQUEST_URI), F)
20:18:00 [timbl]
The protocol should be outlines in http://www.w3.org/DesignIssues/ReadWriteLinkedData.html
20:18:11 [timbl]
(It is also in a tabulator paper)
20:18:51 [timbl]
The base URI supplied to parsers is clearly the URI of the document, right?
20:18:59 [presbrey]
yes, the request uri
20:19:22 [RalphS]
RalphS has quit (Quit: leaving ...)
20:19:22 [presbrey]
currently gets sent as '-b' argument to algae
20:20:11 [timbl]
(I wonder, a nice feature could be that there is a .rdfe added to the URI to get the filename, hidden from the URI completely, but allowing people to have wiki files in their normal filespace and just declare them editable)
20:20:38 [presbrey]
why .rdf and not .n3 ?
20:20:40 [presbrey]
or .ttl ?
20:20:49 [timbl]
I'd prefer N3
20:20:54 [presbrey]
so would I
20:21:04 [presbrey]
but I don't call my N3 .rdf, do you?
20:21:09 [timbl]
Let's swicth it to N3 then
20:21:14 [ericP]
i don't have an n3 parser
20:21:19 [timbl]
No, I'd call it .n3
20:21:23 [presbrey]
ericP, apt-get install rapper
20:21:24 [ericP]
i do have a parser for the turtle subset
20:21:35 [presbrey]
oh right blah
20:21:35 [timbl]
You don't have an n3 parser in which language, ericp?
20:21:41 [timbl]
C?
20:21:57 [ericP]
any, i guess
20:22:06 [ericP]
i never got an lalr(1) grammar to work for n3
20:22:11 [timbl]
Well, we have them in python and js and php etc
20:22:44 [presbrey]
timbl, lets keep all the Access-Control-Allow that you added?
20:23:15 [timbl]
Access-Control-Allow ?
20:23:24 [timbl]
This system currently does access control?
20:23:42 [presbrey]
I'm not sure, it certainly looks like you added something about it r29061
20:23:56 [timbl]
I edited the server code?
20:24:32 [presbrey]
r29061 | timbl | 2010-05-05 13:37:41 -0400 (Wed, 05 May 2010) | 1 line
20:24:33 [presbrey]
Add headers for Croos-site acces
20:26:00 [timbl]
ericp, did you try http://www.w3.org/2000/10/swap/grammar/n3-selectors.n3-yacc.y
20:26:35 [presbrey]
there's also some meta/proof stuff in the GET handler for data wiki. I'll discard that for now as well, timbl?
20:27:26 [timbl]
Ah, headers for cross-site access is not access ocntrol, it defeating the javascript cross-site-scriting machinery which by defaults blocks a browser from alloiwng js code to acess data on random sites. Th ffox code does this protocol where by those headers our side says "anyone can access this dta"
20:27:50 [timbl]
meta/proof stuff yes should go, it was an experiment but we can't have it in the running wiki
20:28:16 [presbrey]
yes I've seen it do its OPTIONS requests
20:28:23 [presbrey]
sad that has to be an extra round trip
20:28:24 [timbl]
email it to kenny so he can run his own verion if he wants. I htink his prtocol over the wire for sending the proofs was kinda random
20:28:40 [timbl]
Needed a little revision by some p[rotcol designers
20:29:13 [timbl]
round trip - for kennys stuff or xss?
20:29:22 [presbrey]
looking for the headers
20:31:00 [presbrey]
ah thanks for keeping a sparul curl example handy in designissues
20:31:36 [presbrey]
funny that example.edu really exists
20:31:45 [presbrey]
running an 'example' apache I suppose
20:31:53 [timbl]
It should jsut say "this is an example"
20:32:53 [timbl]
You have reached this web page by typing "example.com", "example.net","example.org" or "example.edu" into your web browser.\nThese domain names are reserved for use in documentation and are not available for registration. See RFC 2606, Section 3.
20:33:05 [timbl]
I wonder how many easter eggs there are at other URLs :)
20:33:27 [presbrey]
ha yeah thats what I was seeing as well
20:34:59 [timbl]
Do you think I should leave dig.csail in the paper?
20:35:50 [presbrey]
a working example server is nice
20:36:14 [presbrey]
I don't see google/bots copying curl commands out of web pages and executing them in the near future...
20:45:29 [presbrey]
ericP, ready for SWObjects in python...
20:45:32 [presbrey]
curl -d 'INSERT { <http://example.edu/2007/wiki/people/JoeLambda#JL> <http://xmlns.com/foaf/0.1/age> 66 }' -H Content-type:application/sparql-query http://dig.xvm.mit.edu/wiki/people/JoeLambda
20:58:09 [ericP]
timbl, http://www.w3.org/2005/01/yacker/uploads/n3_B?lang=perl&text={%3Fs%3Fp%3Fo}%3D%3E{%3Fs%3Fp%3Fo}.&action=validate+text
20:58:37 [ericP]
though more informative names than _g{0,1,2,3,4} would be cool
20:59:00 [ericP]
also, note 1 reduce/reduce conflict
20:59:12 [ericP]
http://www.w3.org/2005/01/yacker/uploads/n3_B/n3_B.output?lang=perl
21:08:09 [mcherian]
mcherian (~mathew@209.117.47.253) has joined #dig
21:13:36 [presbrey]
oh you found the MS note for MS-Author-Via timbl ?
21:13:48 [presbrey]
perhaps we should update that then to MS-Author-Via: SPARQL, DAV
21:17:23 [presbrey]
turtle/n3 PUT makes sense for data wiki
21:18:05 [presbrey]
I even put an DAV PUT mechanism in tabulator
21:18:25 [presbrey]
can't remember if the tabulator n3 serializer worked with it
21:22:39 [ericP]
presbrey, do you have some test suite to throw at http://www.w3.org/2005/01/yacker/uploads/n3_B/n3_B.output?lang=perl ?
21:23:32 [presbrey]
not that I know of
21:23:52 [ericP]
i'm wrestling with one reduce/reduce error
21:29:44 [timbl]
Can you tell from that output which it is?
21:31:34 [timbl]
you want an N3 parser test suite?
21:32:47 [timbl]
Yosi built a pretty complete one
21:36:37 [ericP]
timbl, it's a trivial one...
21:36:51 [ericP]
n3_formulacontent: | n3_declaration* n3_universal* n3_existential* n3_statementlist ;
21:36:57 [ericP]
stupid tabls
21:37:01 [ericP]
tabs
21:37:15 [ericP]
n3_formulacontent: | n3_declaration* n3_universal* n3_existential* n3_statementlist ;
21:37:25 [ericP]
oi, what's with my paste \ns?
21:37:59 [ericP]
anyways, n3_formulacontent: could be empty or n3_declaration* could be empty, so which rule to reduce on?
21:41:20 [timbl]
Hmm
21:48:43 [ericP]
exactly
21:50:13 [timbl]
Which production do you mean?
21:53:06 [timbl]
By the way folks remember when the client asks for a random unknown URI to give back a valid data file, either a little rdf/xml file or empty N3 fiel but not empty rdf/xml file.
21:55:11 [ericP]
i speak of n3_formulacontent
21:55:44 [timbl]
formulacontent cfg:mustBeOneSequence (
21:55:44 [timbl]
( statementlist )
21:55:44 [timbl]
).
21:57:10 [ericP]
yeah, i can't really correlate http://www.w3.org/2000/10/swap/grammar/n3-selectors.n3-yacc.y tothat
21:58:36 [ericP]
i guess i know how to read n3.n3
21:58:48 [knappy]
knappy has quit (Quit: knappy)
21:58:48 [timbl]
It is just BNF
21:58:50 [timbl]
in n3
21:59:02 [timbl]
as nested lists
21:59:46 [ericP]
so is this allowed [[ { forsome ?x. ?x <p> <o> } ]] ?
22:00:21 [timbl]
yes
22:00:52 [timbl]
In n3.n3 it seesm though that statements can just be declarations or data and they can be in any order
22:01:09 [timbl]
statement cfg:mustBeOneSequence (
22:01:09 [timbl]
(declaration)
22:01:09 [timbl]
(universal)
22:01:09 [timbl]
(existential)
22:01:10 [timbl]
(simpleStatement)
22:01:10 [timbl]
).
22:01:53 [timbl]
Which is different from the yacc, in whoch there are lists in order of declation of each type then data.
22:02:01 [ericP]
hmm, trying to rectify that against http://www.w3.org/2005/01/yacker?name=n3_B&replace=1&lang=perl#prod-n3_B-n3_formulacontent
22:02:25 [ericP]
how about [[ { ?x <p> <o> . forsome ?x . } ]] ?
22:03:20 [timbl]
Well, you don'yt use the ?x when you are using explicit quantification
22:03:56 [timbl]
{ <x> <p> <o> . @forSome <x>. }
22:04:23 [ericP]
it's certainly easier if i can just model this as n3_formulacontent:= (n3_declaration | n3_universal | n3_existential |n3_statement)*
22:04:44 [timbl]
echo ' <x> <p> <o> . @forSome <x>. ' | cwm
22:04:46 [timbl]
works
22:04:50 [timbl]
cwm is forgivingthere
22:05:10 [timbl]
Yes, model it that way
22:05:24 [ericP]
...
22:05:46 [timbl]
Funny that the yacc and the n3.n3 are not in sync
22:08:10 [timbl]
Maybe the yacc is old .. no .. make: Nothing to be done for `n3-selectors.n3-yacc.y'.
22:10:07 [timbl]
Maybe the first version o fthe should just use rdf/xml
22:10:44 [ericP]
turtle would be the easiest for me -- already have a turtle serializer
22:11:18 [ericP]
i'm confused about how trailing '.'s are supposed to be allowed
22:12:21 [ericP]
n3_statementtail seems to always have a statement after a '.'
22:13:32 [timbl]
No, it always has a statement list after a . but th estatementlist can be empty
22:14:29 [ericP]
i thought this should be concatonatable
22:14:44 [ericP]
which would imply '.'s at the end in all cases
22:14:44 [timbl]
Or you could take the javascript versiuon of the parser and convert it to C
22:15:06 [ericP]
i'm not sure that would get me closer to the definition of n3
22:15:12 [timbl]
No, statementtail can be empty too.
22:15:22 [timbl]
So it can stop after either
22:15:55 [ericP]
would you like concatonatability?
22:16:05 [ericP]
(i'd say that's a strong feature)
22:19:16 [timbl]
Let's not design it now.
22:19:31 [timbl]
Yes I agree concatability is a strong feature.
22:19:34 [timbl]
Cwm's input is
22:19:41 [ericP]
i'll live with a reduce/reduce error
22:19:46 [ericP]
gotta go cook something out
22:19:59 [timbl]
but the n3 language as specced you are supposed to put your decs first not after a bunch of statements
22:20:06 [ericP]
(as in, there's a cookout, and i expect i'm expected to perform)
22:20:18 [timbl]
You have to cook, then I will rid ehom as the sun has come out
22:20:35 [ericP]
feel free to play with yacker
22:20:49 [timbl]
I suggest that we just use rdf/xml for now
22:21:11 [ericP]
booooooorrrinnnnng
22:21:13 [timbl]
to get the data wiki up so that knappy can develope stuff for it without having to add ajax timeous
22:21:37 [timbl]
and then just move it to n3 after we have it working
22:21:37 [ericP]
i imagine an ajax timeout of 1s is sufficient and not a big pain for her
22:21:43 [ericP]
okidoke
22:21:56 [ericP]
well, i still have to write a serializer if we use rdf/xml
22:22:17 [ericP]
turtle and trig are written and tested
22:22:28 [timbl]
It really messes up he code -- she is suppose to do be doing great user interface code useing simple widgets and teh read-write data space and coding around server bugs is much more boooooooring!
22:22:30 [ericP]
i can read rdfxml, but have no serializer
22:22:34 [ericP]
gotcha
22:22:49 [timbl]
Suppose you write it all in js
22:23:06 [timbl]
You can read rdf/xml and serialize turtle??
22:23:17 [timbl]
seriously you can't round trip either?
22:24:19 [timbl]
If you can read and write turtle, do that
22:29:07 [ericP]
can do turtle
22:29:20 [timbl]
timbl has quit (Quit: timbl)
22:42:19 [nunnun]
nunnun is now known as nunnun_away
22:57:03 [mcherian]
mcherian has quit (Ping timeout: 260 seconds)
23:34:13 [timbl]
timbl (~timbl@pool-96-237-236-72.bstnma.fios.verizon.net) has joined #dig
23:47:47 [kennyluck]
kennyluck (~kennyluck@EM111-188-46-115.pool.e-mobile.ne.jp) has joined #dig
23:48:30 [lkagal]
lkagal has quit (Quit: lkagal)
23:53:19 [nunnun_away]
nunnun_away is now known as nunnun
23:56:02 [mcherian]
mcherian (~mathew@209.117.47.253) has joined #dig