17:42:14 [DanC]yes, RuntimeError is like assert(). assert() is _not_ something you use to flag bad input data.
17:42:39 [timbl]so, what about toXML.py: raise RuntimeError("this graph cannot be serialized in RDF/XML")
17:43:01 [DanC]should be ValueError or some such, IMO
17:43:49 [DanC]using RuntimeError says, to me, that you can't serialize it as RDF/XML this week, but next week you're hoping to write some clever code to do it.
17:44:32 [DanC]eikeon, do you have any thoughts on what formula the proof generation should actually prove?
17:44:48 [DanC]oh... and did you see that I changed the request from a bnode to a mid: URI?
17:45:19 [DanC](had to fix a bug in uripath.py while I was at it.
17:45:21 [DanC])
17:45:27 [timbl]SpecError(spec, line, lastCallcomment)
17:45:52 [eikeon]DanC: no re: what formula -- Don't have my head far enough around the proof generation bits.
17:46:13 [eikeon]Yeah re: request id
17:46:33 [DanC]ok, some questions you might be prepared to answer: should this proof generation stuff work for PUT/POST/DELETE requests? or is it only suitable for GET requests?
17:47:54 [DanC]if it's just for GET requests, then something like { alice http:can-get <doc1> } is a good target. But for vague reasons, I prefer something that doesn't mention the method. Just { <mid:lkjlwkejr> a rein:ValidRequest }.
17:48:47 [DanC]currently, if it proves { sky color blue} , then engine.py is happy ;-)
17:49:24 [eikeon]Yeah... that might need to be changed ;)
17:51:10 [alerer]DanC, can you send me the URL of the website with the list of SPARQL servers?
17:51:46 [DanC]http://esw.w3.org/topic/SparqlEndpoints ?
17:52:34 [DanC]that one, alerer ?
17:53:04 [DanC]ok, now that engine.py is sorta working, I suppose I should undo some of the damage I did along the way.
17:53:18 [eikeon]What version of swap do I need?
17:53:31 [DanC]recent enough to get why.py 1.39
17:54:45 [DanC]is the rein engine still in flux? It seems to still use log:conclusion/log:includes, where I thought it should use log:supports. yosi? lkagal?
17:54:51 [eikeon]DocumentAccessError: Unable to access document <file:/home/connolly/paw/www/2005/demo/rfilt.n3>, because:
17:54:53 [eikeon]:)
17:55:13 [DanC]yes, that's part of the damaga
17:55:15 [DanC]damage
17:55:15 [alerer]yes that one DanC thanks
17:55:57 [DanC]for some reason, my paw/www commits weren't showing up in http space
18:05:01 [eikeon]Anyone know where we stand re: having lean version of all the girl scout use cases?
18:05:54 [DanC]not I; I'm not even sure I understand the question. I'd like a pointer from the code to the prose description of the use cases.
18:06:57 [eikeon]I'll add a pointer as soon as I can find where the prose live.
18:07:33 [DanC]er... how did you find the test data?
18:12:22 [eikeon]Looks like we must be going off of http://www.policyawareweb.org/2005/paw-use-cases-20050726/index.html -- last log message was "Updated PSD use case following Jan 2006 ftf discussions."
18:13:58 [eikeon]hum.
18:16:40 [DanC]and is there a path from http://www.policyawareweb.org/ to http://www.policyawareweb.org/2005/demo/paw ? http://www.policyawareweb.org/2005/demo/README doesn't seem up-to-date
18:18:44 [DanC]hmm... lkagal, in http://dig.csail.mit.edu/2005/09/rein/network.n3 , it looks like :ans is the mechanism for saying whether a request is valid
18:39:20 [DanC]beware, eikeon , you'll need to update swap soon. Do you use the delayed repository?
18:39:47 [eikeon]Yes, I'm still on the delayed repo
18:43:42 [DanC]http://www.policyawareweb.org/2005/demo/rfilt.n3 isn't updating real-time. anything you can do about that, eikeon? (speaking of delays)
18:44:41 [eikeon]I'll check with Ron... he recently upgraded the OS on the box that policyaware runs.
18:44:43 [DanC]hmm... we load stuff from the web in ReinEngine.__init__ ... __init__ isn't supposed to raise exceptions. (at least, not per Modula-3 norms)
18:45:27 [DanC] self.user = "unknown" <- what's that for?
18:46:43 [DanC]session:id rein:test <- rein:test should be the user name, no?
18:50:57 [avivo]avivo (n=Aviv@31-35-52.wireless.csail.mit.edu) has joined #dig
18:51:42 [eikeon]Think so... what version of bengine, filter and the policies are we moving forward with?
18:55:28 [eikeon]lkagal: Are you working on the policies for the GS demo? Looks like we're getting to the point where we can plug them into engine.py unit test code and hopefully have them work.
18:58:45 [DanC]I see you checked in a fix to notation3.py
19:04:27 [DanC]is "alice-passwd" hardcoded somewhere?
19:06:10 [DanC]in the bob test, I'm getting an "alice-passwd" in the proof somehow.
19:06:39 [eikeon]DanC: Yes, the passwords are hard coded as username-passwd
19:07:05 [DanC]but I'm getting an alice/bob mixup
19:07:20 [eikeon]Wonder if there's still a request coded in one of the files you're loading
19:07:51 [sheets]Does anyone have clue regarding uberprinter? I'm attempting to print from the StataCenter wifi using Ubuntu 6.06. I'm not quite sure what settings I need for the printer but when I use ipp://abyss/uberprinter I get a barrage of SMB authentication requests.
19:08:26 [eikeon]Engine's test shouldn't be getting them mixed up... unless everything is accumulating in a store maybe?
19:08:55 [eikeon]In engine.py's __main__: result = test_rein(resource, policy, '%s-passwd' % user)
19:09:11 [sheets]If only I had Project Oxygen...
19:11:06 [DanC]project oxygen would be one more layer of complexity which can only compound the bugs in the lower layers
19:11:50 [DanC]the printingn world is full of silent failures (at least the linux printing world)
19:12:30 [DanC]ah
19:12:52 [DanC]yosi_s, is the rein engine supposed to migrate to log:supports soon? the log:concusion steps are a bother
19:14:04 [yosi_s]the last time I tried to modify rein to use log:supports, I was getting errors I have not diagnosed
19:15:33 [yosi_s]I tested removing the log:conclusion from lean/bengine.n3, and turning the log:includes in bfilter.n3 into a log:supports --- that worked
19:15:50 [yosi_s]I'm not modifying those files --- if you want to do that, you can
19:16:19 [DanC]eikeon, engine.py calls test_rein for every test case... so it makes a new engine for every case, right? hmm... an engine can be shared accross requests.
19:22:18 [DanC]RuntimeError: No tracker. This may happen if the formula is validly empty. <- yosi, I'm still getting that. I need it to go away. I'm gonna whack it unless you say otherwise soon.
19:23:29 [yosi_s]DanC, you would only get that if you are trying to prove an empty formula
19:23:50 [yosi_s]if you are hitting that, the error has already happened
19:23:53 [DanC]is that something I should prevent?
19:24:07 [yosi_s]why are you trying to prove True?
19:24:20 [DanC]cuz that's what the rein filter yielded
19:25:00 [DanC]how does one test for the empty formula? len(f) == 0?
19:28:33 [yosi_s]if the rein filter yielded an empty formula, then it has told you nothing
19:28:39 [yosi_s]there is no reason to prove true
19:28:59 [yosi_s]further, proving true will not help you gain access to any resource
19:29:07 [yosi_s]simply if not f:
19:29:20 [yosi_s]will work, len(f) == 0 means that f is false
19:29:49 [DanC]when f is a formula, not f harks of the logical negation of f; I think I'll stick with len(f) == 0
19:31:20 [DanC]re "Check will not work with diag == 1 nor will check._s2f", did you write that, yosi_s ? why won't it work?
19:31:49 [DanC]ok, len(f) == 0 is working: [[
19:31:51 [DanC]DEBUG can't find proof.
19:31:51 [DanC]INFO Got expected result for http://demo.policyawareweb.org/images/group.jpg, http://www.policyawareweb.org/2006/lean/troop42-policy.n3, elle
19:31:52 [DanC]]]
19:34:46 [DanC]whoa... http://www.policyawareweb.org/2006/lean/troop42-policy.n3 doesn't make any sense to me. what the heck is going on in there?
19:37:36 [yosi_s]DanC, you can't pin that one on me
19:41:17 [DanC] 221 eikeon # Check will not work with diag == 1 nor will check._s2f
19:41:26 [alerer]alerer has quit ()
19:42:00 [tlr]tlr (n=roessler@132.219.18.172) has joined #dig
19:42:32 [avivo]avivo has quit (Read error: 110 (Connection timed out))
19:47:50 [DanC]engine.py r250
19:50:34 [sheets]For the interested, the uberprinter answer is: it speaks HP JetDirect on port 9100 at host uberprint.w3.org
19:50:50 [eikeon]DanC: Think I put that comment in there.
19:51:02 [DanC]so says svn
19:51:34 [eikeon]It was true at some point in the past... not sure if it's still needed.
19:54:13 [DanC]can you run engine.py r250 , eikeon ?
19:54:48 [eikeon]Just finished updating paw and swap... running...
19:55:29 [eikeon]DEBUG irrelevant conjecture: {"alic...sswd" http:can-access group.jpg}
19:57:19 [eikeon]DanC: Is that what you're getting?
19:57:42 [DanC]no... oops; gotta check in rfilt...
19:58:04 [DanC]er... I checked it in; can you svn up?
19:58:27 [DanC]that's the one that's called /home/connolly/etc in the code (I'm not sure how you're working around that)
19:58:52 [tlr]tlr has quit ("good luck")
19:59:18 [eikeon]Ah... I changed it to point at the version at www.paw... it looked up to date. Checking to see if it's up-to-date
20:04:18 [eikeon]Tracked down the www version being out of sync... there's a web cache running in front of all the sites... and it was caching it (improperly I'm assuming).
20:04:26 [eikeon]Have been told it's fixed.
20:05:41 [eikeon]INFO Got expected result for http://demo.policyawareweb.org/images/group.jpg, http://www.policyawareweb.org/2006/lean/troop42-policy.n3, alice
20:05:59 [DanC]:)
20:06:06 [eikeon]Indeed ;)
20:07:03 [DanC]yosi, have you fixed all the paw-demo-related problems you know of? or are you still hunting?
20:07:07 [eikeon]Now to get the rest of the policies written up
20:07:11 [yosi_s]still hunting
20:07:28 [DanC]ok... so you don't expect the non-lean stuff will work yet?
20:08:12 [yosi_s][[
20:08:16 [yosi_s]swap.check.LogicalFallacy: pr:Inference: Calculated formula: {5} is { @forSome bfi:_g_L23C8.
20:08:18 [yosi_s] bfi:_g_L23C8 rdf:type rein:Request .
20:08:20 [yosi_s] bfi:_g_L23C8 rein:access http:can-get .
20:08:21 [yosi_s] bfi:_g_L23C8 rein:ans rein:Invalid .
20:08:23 [yosi_s] bfi:_g_L23C8 rein:requester 1230701428 .
20:08:24 [yosi_s] bfi:_g_L23C8 rein:resource group.jpg .}.
20:08:26 [yosi_s]does not match given: {5} is { @forSome run:u_1_L3922C15.
20:08:28 [yosi_s] run:u_1_L3922C15 rdf:type rein:Request .
20:08:30 [yosi_s] run:u_1_L3922C15 rein:access http:can-get .
20:08:32 [yosi_s] run:u_1_L3922C15 rein:ans rein:Invalid .
20:08:33 [yosi_s] run:u_1_L3922C15 rein:requester eng:WHO .
20:08:35 [yosi_s] run:u_1_L3922C15 rein:resource group.jpg .}.
20:08:36 [yosi_s]]]
20:12:15 [DanC]ah.. http://dig.csail.mit.edu/2005/09/rein/examples/ has prose. That's what I'd like linked from the engine.py code, perhaps
20:16:50 [yosi_s]eureka!
20:17:13 [DanC]oh?
20:17:19 [eikeon]Ah yes... those user names line up with the current ones on the engine test cases. It'd be nice to pull them into 2005/demo -- or at least know which version we're trying to get working... that one or lean etc.
20:17:54 [yosi_s]http://groups.csail.mit.edu/dig/2005/09/rein/bfilter.n3 , in the failing case, has a {} => {[] rein:requester ?Who} .
20:18:20 [yosi_s]that is breaking the proof --- it's not putting the @forAll into the proof for some odd reason
20:19:14 [yosi_s]that's ok --- I have no clue what that is doing in that rule
20:20:26 [DanC]you have it working?
20:20:29 [eikeon]On which bit ;)
20:22:05 [RalphS]RalphS has left #dig
20:25:54 [ryanlee]ryanlee has left #dig
20:26:35 [yosi_s]I have determined that somebody else broke rein
20:27:23 [yosi_s]with NORMAL=true, it would work without that minor issue
20:28:01 [DanC]do you see how to fix it? or would you please nominate somebody to take the ball?
20:30:59 [claudiof]claudiof has quit (Read error: 110 (Connection timed out))
20:32:50 [sheets]sheets has quit (Read error: 110 (Connection timed out))
20:32:59 [eikeon]Committed revision 251 of engine.py -- now pointing to rfilt.n3 that's on the web; added pointer to Rein Examples since that's where we're currently getting our GS use cases.
20:35:46 [yosi_s]DanC, I just asked lkagal to fix it.
20:36:10 [DanC]ah. covert channel. (ftf) ;-)
20:37:36 [eikeon]would work... or might work... has it been tested with the minor issue fixed yet?
