00:03:09 hahahaha 00:03:21 * Spork takes notes on how to do "that" 00:04:12 first thing I'm doing on monday is knocking down the cubicle wall 00:04:19 :) 00:06:43 hmm 00:07:05 Must go fishing tomorrow and catch some fluke to fillet on my status report 00:08:55 hahahaha 00:08:58 Mission statement 00:15:33 hahahaha. they stole the printer 00:15:56 and kicked it's ass 00:16:03 Yeah. 00:16:10 that's great! 00:30:40 davb has joined #OpenACS 00:30:51 hey dave 00:31:01 hi 00:31:17 * Spork is watching Office Space 00:31:24 cool! 00:31:34 I think on Monday I will steal a printer and smash it with a baseball bat 00:31:48 Oh you have watched quite a bit. 00:31:57 Almost done 00:34:29 this movie is great 00:34:47 "Pound me in the ass prison" 00:34:48 rofl 00:43:52 dave, I got a question for you 00:44:03 ok 00:44:15 in the movie, how much is left after they show that guy talking about mai tai and getting a pina colada? 00:44:24 Milton in the Resort 00:44:38 not much... 00:44:47 what else do they show? 00:44:53 My movie ended there =) 00:45:00 That might be it. 00:45:37 ok 00:45:38 hehe 00:45:47 fly has joined #openacs 00:45:58 hello fly 00:46:09 hey davb 00:47:21 fly has left #openacs 00:48:00 Well, he talks alot. 00:48:54 Does anyone know what port rexec uses? 00:49:10 I think Roberto was talking about remote X apps 00:49:24 Oh yeah. SSH tunneling? 00:49:31 yeah 00:49:43 hmmmm 00:50:25 I don't know about rexec tho 00:50:34 You shouldn't be running rexec. Ever. 00:50:39 why's that? 00:50:40 hold on 00:50:59 rexec was for the time when there were 5 nodes on the whole internet, and everybody knew each other. 00:51:13 related to rlogin? 00:51:33 you should not use any "r" service, with the exception of rsync 00:51:50 ssh is a secure replacement for all r services. 00:51:58 ok well here's my scenario 00:52:04 Maybe you can help me figure it out 00:52:13 i have 4 computers on my internal network 00:52:29 My Win2K box and my freebsd box get used the most 00:52:44 Say I want to run some X apps from my linux box on my win2k box 00:52:49 How do I go about doing this 00:52:55 I have installed XWin32 00:53:19 But, I also have a router so it blocks replies unless I forward 6000 to one computer 00:53:28 You can run the X app on the Linux box to be _displayed_ on the win2k box 00:53:31 I would prefer to use triggers but I don't know what port it triggers on 00:53:35 right 00:53:39 that's what I meant 00:54:08 It's been a long time I've messed with XWin32. I just use SSH these days. 00:54:27 lol 00:54:34 s/SSH/VNC/ 00:54:38 any of you listen to wallops? 00:55:39 Spork: Or, since you have XWin32 running, you can use TeraTerm Pro to forward the X port through the ssh tunnel 00:55:52 Spork: Any reason why you can't just use VNC? 00:56:17 It's too much of a pain to start the server because it doesn't keep my session 00:56:28 VNC does keep your session. 00:56:31 i don't really need a full blown desktop 00:56:45 yeah I know, but I mean between a local and remote 00:57:08 I don't understand what you're saying then. 00:57:13 :/ 00:57:31 What's your SSH client on Windows? 00:57:40 SSH Secure Shell from www.ssh.com 00:58:00 yes, it allows for X tunneling through SSH 00:58:41 I hope you're using it on an educational environment. I think it's not free for commercial use. 00:58:52 yeah educational of course 00:59:17 So, you forward the X port on SSH, then what happens? 00:59:23 This is what I did 00:59:27 I have xwin running 00:59:34 My router messes things up though 00:59:52 brb 01:00:11 brb2 01:00:26 back 01:01:41 Yeah! To play shoutcast streams I have to save the .pls file and then load it into XMMS. 01:01:56 is that good or bad? 01:03:25 Well it works, but I can't do it automagically from the browser. 01:03:35 Could you do that with Freeamp? 01:03:44 No. 01:03:56 Freeamp crashes too much even playing CDs 01:04:07 heh 01:05:37 i like XMMS 01:06:17 I think all older winamp skins and maybe even new ones work with XMMS? 01:06:37 Yeah I think so. 01:07:53 Ok, howcome windows cd buring software sucks so bad? 01:08:04 you have to be more specific 01:08:11 which one are you talking about 01:08:12 I am buring a CD at 8x, installing openacs datamodel, listening to XMMS 01:08:17 etc... 01:08:20 no problem. 01:08:33 If the screensaver comes on in windows, the recording fails EVERY time. 01:08:48 odd 01:08:52 I had that problem with Adaptec and WinOnCD. Don't know about Nero, I turned off the screensaver. 01:08:53 What software, what burner 01:09:03 What burner? 01:09:23 I have a Yahama 4416 SCSI, now I have a 8424 IDE. 01:09:33 That's odd I have 2 burners and I have almost never had any problems (when everything works correctly) 01:09:39 Cool. 01:10:09 my old mitsumi 4x burns without a problem when I multitask, I can surf the web and do a bunch of other things and never had a problem 01:10:16 Cool. 01:10:23 Must be me! 01:10:27 and the new plextor has Burn-proof so it doesn't crap out at all 01:10:40 Yeah, they are good. 01:10:59 yeah 01:12:20 Ok, so just forget what I said :) 01:12:21 I don't understand what's Adaptecs problem though 01:12:28 Uh, they suck? 01:12:31 they still dont' support overburning 01:12:39 Anyway, they are "Roxio" now. 01:13:02 i know 01:13:14 new great name...same crappy product 01:13:24 oh! 01:13:25 heh. They support 80 min CDRs now? 01:13:28 I will play some max Payne 01:13:30 Yeah 01:13:34 Well its a start. 01:13:38 heh 01:13:49 yeah but I had a 707mb file 01:14:04 I thought it would overburn if you just kept putting crap on the CD in the window. 01:14:13 Anyway you have Nero. 01:14:54 no 01:14:58 oh 01:15:02 it will reject the cd asking you to insert a bigger one 01:15:48 oh, that is no fun. 01:16:13 exactli 01:17:05 i'm gonna play some max payne 01:17:08 be back later 01:17:11 ok 01:41:44 Hmmmm. 01:42:02 Installing postgresql-contrib, installs the source for intarray, but doesn't build it. 01:42:17 arg 01:42:20 it frozeth 01:42:21 It puts the source in /usr/lib/postgresql/lib/contrib where it can't be built. 01:42:23 heh 01:42:43 I am wrestling with the debian package system. 01:44:45 its just about time to compile from source. oh well... 01:54:19 heh 01:54:25 BSD ports rule 01:54:33 Now I have 2 copies of postgresql. 01:54:55 Hah. Gilbert had to play musical symlinks with that to get openfts search working also! :) 01:55:09 This is a very picky package. 01:57:46 All it needs is a debian section in the configure script 02:06:41 SUCESS! 02:06:48 except in speleng 02:19:04 This openFTS is amazing! 02:21:08 jerryasher has joined #openacs 02:21:36 Bueller? 02:22:23 Ferris? 02:22:51 Hi Dave. So I have these blogs, and I have this tool to gen RSS .91 and 1.0. 02:22:58 great! 02:23:01 (from a list of items). 02:23:14 (not just from blogs) 02:23:23 BTW, the OpenFTS search in openacs4 is really incredibly amazing. 02:23:28 Oh cool. 02:23:38 aggregator.userland.com tells me that my .91 feeds are valid 02:23:44 excellent. 02:23:51 though I don't know about the 1.0 feeds. 02:24:14 So I was wondering if you or Aaron knew of a place to validate 1.0 feeds 02:24:28 Oh I bet there is one. I'll look. 02:24:30 and in general of a place I can test/publish my lil old blog with. 02:24:35 I can take a look at one. 02:24:41 syndic8.com is good for that. 02:24:45 Or AaronSw can look at it. 02:24:46 Oh! Hi Aaron! 02:24:57 * AaronSw heard his name 02:25:02 and sped over 02:25:22 As a demo, I've been thinking of creating an RSS feed of the last 15 OpenACS posts 02:25:36 and letting my.userland.com and webblogs.com carry them. 02:26:03 that'd be quit neat 02:26:34 There are times I perceive ACS as so stodgy...! 02:26:37 Hopefully we can get it in OpenAcs 4 so that openacs.org can do it itself. 02:26:57 The routine I wrote is completely general. It may only need to be updated to use 02:27:02 a new version of ad_proc. 02:27:23 Cool. 02:27:24 Although RSS 1.0 support is very primitive, only the .91 version so to speak. 02:27:37 No dublin core or anything else I didn't need. 02:27:55 Anyway, I would like to see OpenACS do it for themselves. After I get this working a bit 02:28:00 I was going to email talli. 02:28:33 Okay then, dinner is about ready. If you can find a 1.0 validator, I would appreciate it, 02:28:46 and if not, I will be back in a bit with a link that you can use to look at the xml yourselves. 02:29:17 Again, cool. (vocabulary is limited tonight... :) 02:29:31 After that, it's the Jules Dassin classic, Rififi http://www.imdb.com/Title?0048021 02:29:41 Say my name when you get back. 02:29:46 So i'll come check it out 02:29:50 Will do! See ya later. 02:31:22 c'ya 02:38:03 anyone know why photo.net is dead? 02:40:05 philip's site is toast, too. 02:40:25 ack! 02:40:30 They were talking about moving it. 02:41:20 hmm 02:41:28 looks like a multi-day outage 02:41:38 how long could it take to transfer it over to another machine? 02:41:41 weirdness 02:41:45 Philip.greenspun.com was up yesterday 02:41:59 right, but the little note says it's not coming back up until monday 02:41:59 They wanted to switch from solaris to linux, or at least were thinking about it. 02:42:01 (they think) 02:42:06 aha. 02:42:10 didn't see that. 02:42:12 you'd think if they had a migration plan 02:42:16 it would be rapid 02:42:31 They were asking on the bboard for advice. 02:42:35 !! 02:42:36 really?! 02:42:39 Yes. 02:42:54 you've gotta be kidding 02:43:21 http://openacs.org/bboard/q-and-a-fetch-msg.tcl?msg_id=00030u&topic_id=11&topic=OpenACS 02:43:21 F: http://openacs.org/bboard/q-and-a-fetch-msg.tcl?msg_id=00030u&topic_id=11&topic=OpenACS from davb 02:43:39 F:|Advice for moving photo.net off a solaris system 02:43:40 titled item F 02:43:45 F: openacs.org bboard thread 02:43:45 commented item F 02:43:57 wow 02:43:59 that's bizzaro 02:44:14 Not really. they were only on Solaris because aD was married to it. 02:44:31 i know, but photo.net 02:44:35 as far as i understood 02:44:38 was "business". 02:44:47 when i was at aD, I actually had a desk 02:44:55 that was intermingled in the photo.net space. cool bunch of guys. 02:44:58 Oh right. 02:45:04 Weird. 02:45:10 but the way I understood it, they had advertisers, etc.. 02:45:22 Yes. They were at least trying to make money. 02:45:23 so unlike a hobby site, "going down" would seem to be a big deal 02:45:31 But Sun boxes and Oracle are not cheap. 02:45:44 right. i understand the motivation for the switch 02:45:55 i just don't understand the execution -- killing the service for 2+ days 02:46:16 Yeah. That is strange. You;d think they would have a better plan. 02:46:29 wild stuff. oh well. the photo.net people were some of the coolest folk at 80 prospect st. 02:46:44 Interesting. 02:47:09 I guess they really liked what they were doing. 02:47:28 yup. they were all a little bit older, more experienced and i think worked at photo.net 'cause it was fun 02:48:00 Aha. Well hopefully it can support them. That has got to be one of the longest running independent web sites ever. 02:48:23 yeah. it would be a shame for it to totally die. 02:48:35 but if somethingawful.com can run on donations 02:48:40 i have a feeling photo.net could too 02:48:56 (somethingawful has serious bandwidth requirements...) 02:49:42 Hmm, Rajeev was pretty cool, I agree. 02:49:53 hi aaron 02:49:53 Don't know many of the others, I suppose. 02:50:03 yeah, hanging out on the 2nd floor was always a nice respite 02:50:04 Hi there, docwolf... just reading backlogs for a bit. 02:50:23 rajeev tried to get me to implement features for him, heh heh 02:50:24 I think I'll download Philip's photography book also. 02:50:36 That is funny. They should upgrade to OpenACS 4. 02:50:54 You could just have billed him $150/hour 02:50:59 :) 02:51:12 Heh! 02:51:23 I doubt they had the money to pay. 02:51:36 When I was over there Rajeev cheered every time they got a new Paypal payment... which was once. 02:52:23 that is really too bad. its a great site. 03:15:03 davb has quit 03:38:08 vinod has joined #openacs 03:39:45 hi vinod 03:40:31 hey docwolf, what's up? 04:14:01 * Spork pokes vinod in the eye with abbaJ the size of a docwolf 04:21:56 vinod has quit 07:10:39 docwolf has quit 08:39:00 jerryasher has quit 11:56:25 davb has joined #OpenACS 13:04:56 davb has quit 14:14:29 graszew has joined #openacs 14:21:34 davb has joined #openacs 14:28:48 rbm has quit 15:34:46 heh 15:34:50 My clocks changed 15:53:27 * Spork needs more sleep 16:00:49 davb has quit 16:09:55 davb has joined #OpenACS 16:25:00 graszew has quit 16:26:13 rbm has joined #openacs 16:33:27 til has joined #openacs 17:05:55 adler has joined #openacs 17:09:53 talli has joined #openacs 17:10:01 hey guys 17:10:57 hey talli 17:11:03 hey 17:11:11 i just hopped on myself 17:12:59 welcome. you at the NYPL? 17:14:24 i just hopped on myself Brooklyn 17:14:31 errss.... 17:14:42 fuckig mIRC 17:15:04 the nypl is closed on Sundays. So I"m on dialup from home in Brookln 17:15:49 i need to bend over and try once again to get DSL service 17:16:06 i'm about to do that myself 17:16:13 panix was with northpoint before 17:16:18 so i got hosed then 17:16:27 they moved to focal who is apparently offering decent service 17:16:46 christ. who can you trust in this filthy broadband market? 17:17:16 i do not know. verizon? 17:17:35 you can trust them to violate fair competition. 17:18:32 while they may simplify the process of installing a line, there IP-oriented service is sub par 17:18:42 so I've heard 17:25:13 no, they are horrible in every possible way. 17:25:18 they will not simplify anything 17:25:35 verizon is a soviet conspiracy to overtake our country. 17:26:17 you may have thought that communism was done, while in fact it has only gone underground bu infiltrating our great corporations and turning them into sloth. 17:27:47 adler has quit 17:31:29 adler has joined #openacs 17:32:09 i love verizon 17:33:45 wow everyone is in a great mood today 17:33:49 :) 17:34:16 i was walking down the street in Park Slope and I sense a palpable, general discontent 17:45:36 talli: are you listening to the radio? all the public radio stations are asking for money -- yet another conspiracy 17:45:53 chirst almighty 17:47:13 the nerve 17:49:02 talli has quit 17:51:39 graszew has joined #openacs 18:00:18 adler has quit 18:51:09 * Spork is freeeeeeezing 18:51:26 * Spork pokes davb with a til the size of rbm 18:53:45 heh 18:53:52 :) 18:53:54 WHat? Its nice! 18:53:59 Nice? 18:54:04 The weather is nice 18:54:13 but my toes are about to fall off and run to take a hot shower 18:54:13 Sure must be 50 degrees outside. Well we have sun up here anyway. 18:54:19 :) 18:54:38 I hate having multiboot machines 18:54:55 Somehow you ALWAYS wind up being 2-3 hours off when the time rolls over 19:03:29 I am always 5 hours off somehow. 19:03:43 Linux and windows like to play games with how they interpretr the time. 19:08:37 abbaJ has quit 19:19:12 k2pts has joined #openacs 19:19:18 hi guys 19:19:32 davb: got your email. i will get back to you after dinner. ok? 19:20:21 Sure. no problem! 19:20:29 bbl 19:20:33 k2pts has left #openacs 19:34:15 bbl gotta go get my cell phone 19:40:26 hazmat has joined #openacs 19:58:24 hi hazmat 19:58:37 hi davb 19:59:01 davb: i was curious where are the chump archives for this channel? 20:12:22 k2pts has joined #openacs 20:13:26 hi everybody 20:13:38 hazmat: http://www.thedesignexperience.org/openacs/ircblog 20:19:24 Hi k2pts 20:19:31 hi davb 20:20:13 the "content" field should include the absolute path to your file 20:20:22 when storage_type='file' 20:20:28 Ok. I fixed it so that the datasource proc figures it out. 20:20:39 yes, that's the right thing to do 20:20:41 The indexer seems to be doing alot of work, but I can't find my pages. 20:20:57 sounds strange 20:21:25 Is there a table I can look in to see what has been indexed? 20:21:37 yes, select * from txt; 20:21:42 ok. 20:21:55 i think i know what's your problem. i am trying to remember what causes it 20:24:12 k2pts: search question, is the search system an all or nothing, ie something is either searchable or not. can we define restrictions on the search results to pass in to the search query 20:24:46 elaborate 20:25:03 I am using mime type='text/html' BTW 20:25:20 davb: it should work 20:25:40 say i only want to search the bboard? or messages of the bboard that are within a certain forum, or that were posted after a certain date? 20:25:49 some of the pages did get indexed. But most did not. I am going to compare them. 20:26:22 hazmat: we left that for the next release cycle due to lack of time. it should be too hard to implement though (i hope) 20:27:30 Weird. A little tiny html file with almost no content has quite a few items in fts_index. A big page has none. 20:28:30 hazmat: i thought about it a bit but never got the chance to implement it. we can pass the mime type while indexing and it will get stored into the index table. then, we can provide an extra parameter for mime type (this will add an "and" clause to the sql query which performs the search) 20:29:15 davb: how big is "big" 20:29:21 checking... 20:29:53 2273 bytes. the small one was 258 bytes. I don't think that is it. 20:30:03 no, that's not it 20:30:06 bah 20:30:08 All of the pages I took over from my other site for testing are not indexed. 20:30:08 stupid Sprint 20:30:13 No good beatniks 20:30:16 The ones i created by hand work ok. 20:30:33 point me to one of your pages 20:30:40 they are indexed, but there isn't anything in fts_index. 20:31:09 which version of openfts have you downloaded? 20:31:15 never mind. It looks like they are adp pages with an html extension... 20:31:16 0.1 or 0.2? 20:31:17 0.2 20:31:55 strange 20:31:57 no not all of them are adps, just the one. 20:32:46 hazmat: I will try to implement this as soon as i'm done with acs-wf 20:32:55 All the titles are the same on almost all of them. 20:33:40 hmmm, that link I gave you was not indexed. 20:33:51 never mind... 20:34:56 ok, your html page starts with >. that might cause the parser to ignore stuff (i'm not sure though). try adding the tag and reindex. let me know what happened 20:35:45 hazmat: I meant it should *not* be too difficult to implement 20:35:51 Yes I noticted that. 20:35:57 I fixed it let me see what happens. 20:37:01 hazmat... 20:37:45 Does not seem to have helped. 20:38:11 check your logs. is there any helpful message 20:39:57 davb: ok, here's what I have for you: 20:40:18 a. open a terminal window and connect to pgsql 20:40:37 b. select * from search_observer_queue 20:41:02 o rows 20:41:03 c. try to index a page (new index -- rename one of those that don't work) 20:41:22 d. you should be really fast here and repeat the select statement given in (b) 20:41:31 k2pts: trying to understand looking at the wf query... 20:41:37 e. post the select results 20:41:57 right. 20:42:09 object_id | date | event 20:42:09 -----------+------------------------+-------- 20:42:09 2503 | 2001-10-28 15:52:24-05 | UPDATE 20:42:09 2505 | 2001-10-28 15:52:24-05 | UPDATE 20:42:09 2535 | 2001-10-28 15:52:24-05 | UPDATE 20:42:10 2537 | 2001-10-28 15:52:25-05 | UPDATE 20:42:12 2539 | 2001-10-28 15:52:25-05 | UPDATE 20:42:14 2541 | 2001-10-28 15:52:25-05 | UPDATE 20:42:15 2543 | 2001-10-28 15:52:25-05 | UPDATE 20:42:18 2545 | 2001-10-28 15:52:26-05 | UPDATE 20:42:19 2509 | 2001-10-28 15:52:26-05 | UPDATE 20:42:21 2511 | 2001-10-28 15:52:26-05 | UPDATE 20:42:23 2513 | 2001-10-28 15:52:26-05 | UPDATE 20:42:25 2515 | 2001-10-28 15:52:26-05 | UPDATE 20:42:28 2517 | 2001-10-28 15:52:26-05 | UPDATE 20:42:30 2519 | 2001-10-28 15:52:27-05 | UPDATE 20:42:32 2521 | 2001-10-28 15:52:27-05 | UPDATE 20:42:34 2523 | 2001-10-28 15:52:27-05 | UPDATE 20:42:36 2525 | 2001-10-28 15:52:27-05 | UPDATE 20:42:38 2527 | 2001-10-28 15:52:27-05 | UPDATE 20:42:40 2547 | 2001-10-28 15:52:27-05 | INSERT 20:42:41 2547 | 2001-10-28 15:52:27-05 | UPDATE 20:42:43 2547 | 2001-10-28 15:52:28-05 | UPDATE 20:42:45 2529 | 2001-10-28 15:52:28-05 | UPDATE 20:42:47 (22 rows) 20:42:51 2547 is the new page/ 20:43:39 ack 20:43:47 syn 20:43:54 res 20:44:10 davb: how many pages have tried to index altogether and how long ago? 20:44:34 is astiphylax written with a - or space? 20:44:58 I have 20 static pages. 20:45:15 I set them up before I sent that email. 20:45:47 Spork: "astiphylax" is right 20:45:58 i know, I can't find it on google though 20:46:43 davb: i'm confused, please explain. the 22 rows that you have posted were the results of step (b) or step (d)? 20:46:48 Ok. I just deleted all the static pages. 20:46:50 d 20:47:08 davb: can't you index one page at a time? 20:47:49 It is setup to scan the filesystem. I could have just one html page. 20:48:08 Spork: astiphylax is only used in greek (perhaps is an old word). probably, you couldn't find it because it translates to policeman (or something similar) 20:59:04 order keeper! 20:59:05 ;) 20:59:12 right 20:59:14 just like 20:59:17 Spork is now known as Psychephylax 20:59:32 is Soul Keeper =) 20:59:39 muahahaha 20:59:47 * Psychephylax sneaks behind dave and steals his soul 21:00:11 * Psychephylax feels guilty and stuffs it back into dave 21:00:31 davb: are you ok? 21:00:42 heh 21:00:43 hehehe...ok, if phylax is keeper what's _____ stealer ? 21:01:04 or thief I guess 21:01:04 Ok I made a little page and it doesn't seem to have worked. 21:01:11 kleftis 21:01:14 listis 21:01:15 cool 21:01:25 Psychekleftis 21:01:34 lol 21:01:50 davb: post link or send page 21:02:16 Interesting. 21:02:30 what? 21:03:13 well I can search for notes. so that is working. 21:03:14 I got one more 21:03:19 what's borrower? 21:06:14 ? 21:06:25 daneistis 21:06:30 cool 21:06:39 xrewstis 21:06:45 xreos=dept 21:06:48 daneio=loan 21:06:58 hehehe 21:07:22 so borrower is daneistis or xrewstis? 21:07:31 or are they synonimous 21:08:10 one is debtor and one is loaner 21:08:25 it looks like to me 21:08:25 AaronSw you know greek? 21:08:38 no... but that was k2pts said, soulboy 21:08:38 i guess not =) 21:08:50 oh 21:08:52 makes sense 21:08:52 :-) 21:08:59 * Psychephylax sucks AaronSw's soul through a straw 21:09:06 mmmmmmmmm....yummy! 21:09:07 aaah! 21:09:27 Lay off Pina Coladas dude :) 21:09:34 * Psychephylax chuckles 21:16:24 Opps. here is the problem... Database operation "dml" failed 21:16:37 Cannot insert a duplicate key into unique index index13_key 21:16:38 ): ' 21:17:00 are you sure you have the latest version 21:17:02 ? 21:17:04 Before it does that it deletes from index1 - index 12 21:17:09 this is the bug I mention earlier 21:17:21 Oh. Where is the bug? 21:17:24 openfts-driver? 21:17:27 no 21:17:29 openfts 21:17:31 just point me in the right direction and I'll steal it's soul! 21:17:41 it is missing an "=" 21:17:41 Hm. I downloaded openfts-0.2. 21:18:04 Ok. that is the aolserver module? 21:18:15 ok 21:18:39 not your fault. i just realize that we didn't create a new release when we found that bug 21:18:51 goto your openfts src directory and edit fts_index 21:19:00 s/fts_index/fts_index.tcl 21:19:19 ack 21:20:07 or you can cvs co 21:20:31 I will edit the file for now... what do I need to fix? 21:20:46 change line 364: 21:20:56 That is incredible!! 21:21:02 from:for {set i 1} {$i < $self(ING)} {incr i} { 21:21:06 My cursor just happened to be on line 364 21:21:16 to: for {set i 1} {$i <= $self(ING)} {incr i} { 21:21:21 notice the = 21:21:23 except that stuff is not on that line. 21:21:51 i am looking at the cvs diff 21:21:58 find that line 21:22:09 its close 367. got it. 21:22:59 you need to compile for this to take effect (no need to compile the aolserver module). just make 21:23:00 make install 21:23:34 ok 21:25:44 Yeah! 21:25:50 I found a bug in my code! 21:26:21 good 21:26:34 have you finished compiling? 21:27:15 yes. It found the page, but my url contract had a typo. 21:27:40 is it working now? 21:27:54 nope, my query is wrong. Easy to fix. 21:28:00 ok 21:29:04 yippee!! 21:29:43 congrats 21:29:49 thanks. 21:30:08 k2pts: does search use the keywords? 21:30:23 not yet 21:30:38 Ok. then I won't worry about extracting META keywords from the HTML yet. 21:31:06 cool. I added all my pages and they are in there now and searchable. 21:31:28 Now I have to upgrade my web site to openacs 4 so I can search the chump 21:31:36 well done davb. i feel very good to see the search package in a live site. 21:32:51 well not quite. Hopefully soon. I don't need to migrate my database, so it shouldn't be too bad. 21:33:12 Now I must do some work in the yard. bbl Thanks again. 21:33:20 cheers 22:10:38 talli has joined #openacs 22:17:51 hi talli 22:18:00 hey neophytos 22:18:04 how's cyprus today? 22:18:09 how are we doing today? 22:18:10 good 22:21:24 Hi talli 22:21:28 * davb is back 22:21:33 hey dave 22:21:55 hi davb 22:22:09 talli: davb installed openfts on his site 22:22:20 killer! 22:22:26 how's it work davb? 22:22:31 davb: what's the address again? 22:23:11 not much in there. 22:23:21 its on my development machine, 22:23:39 wait, I am reinstalling it for a test... 22:26:43 weird. Oops. I forgot to do load.sql. 22:28:16 my good friend ph4gdj84dx just send me some information about merhcant accounts. Anyone interested? :) 22:29:19 damn! 22:29:31 too bad your friend wasn't offering information on low-interest loans 22:29:46 sorry, that was last week. 22:29:46 i'm really hurting for one of those from a strange spammer 22:30:58 i bet i didn't act fast enough to not miss that opportunity. damn. 22:31:09 i spent my time checking out the hot young chicks instead. 22:31:30 heh 22:35:37 aha, its fixed. 22:35:47 I put my queries for the service contract in an xql file. 22:36:22 oops. 22:36:41 a template bug? can't read "offset_next": no such variable 22:36:41 while executing 22:36:41 "append __adp_output " 22:36:41 Next 22:36:41 "" 22:36:43 invoked from within 22:36:45 "if {"${from_result_page}" < "${to_result_page}"} { 22:36:47 append __adp_output " 22:36:49
22:37:46 it works for me 22:38:04 that URL? 22:38:06 oops sorry 22:38:33 I think its because there are more than 10 results. 22:41:27 it worked when I test it. probably changed something stupid. it's been a while. I will let you know how to fix it in a moment 22:44:02 this is probably something I changed and forgot to commit, can't find the fixed page. davb: change the whole ref to @url_previous@ and @url_next@ and let me know if it works 22:45:56 ack 22:46:37 davb?? 22:49:10 graszew has quit 22:52:30 i am back. 22:52:58 btw, neophytos, thanks for taking on the search problem 22:53:06 it's great that we have this solution. 22:53:12 What file is that in? 22:53:12 talli:np, davb:edit search/www/search-results-bottom.adp 22:53:41 have you done any benchmarks of openfts vs. ht:dig or intermedia? 22:53:50 do you have any sense of what the comparison might be like? 22:54:09 openfts is better at indexing 22:54:16 at online indexing 22:54:47 intermedia is using reverse indexing which means that they have to update the whole index (index for all documents) 22:55:05 aha, first page work. the "next" page is broken still: 22:55:06 We had some problems processing your entry: 22:55:06 22:55:06 offset is not an integer 22:55:06 You've supplied two values for 'offset' 22:55:09 openfts index one document at a time (all the rest remain intact -- no full-reindexing) 22:55:28 talli: also, it works :) 22:56:03 ah, yes. the little "it actually works" argument 22:56:04 should it be like this? &offset=@url_next@ 22:56:10 always rather compelling 22:56:30 davb follow me: change search?q=@urlencoded_query@&offset=@offset_previous@ to @url_previous@, similarly for next 22:57:05 talli: yes lots of features. not to mention flexibility (you can add as many dictionaries as you like) and close integration with pgsql 22:57:23 nice! 22:57:32 what version are you guys on right now? 22:57:32 i am currently working on a unicode parser but it is moving rather slowly. 22:57:45 oleg and teodor do most of the work 22:58:03 they have released tsearch which is very similar to openfts 22:58:17 openfts will make use of tsearch in the future 22:59:25 tsearch (is analogous to intermedia). you can ask queries like within pgsql rather than externally 22:59:49 tsearch is available as a contrib/module for pgsql-7.2 23:00:08 davb: it works now 23:02:50 davb are you with us? 23:04:06 so tsearch is a replacement of openfts? 23:04:23 no 23:04:27 and can openfts index files like .doc and .pdf? 23:04:36 (i though openfts was like intermedia) 23:04:49 ok give me a sec to explain 23:04:58 np. take your time :) 23:05:58 oleg and teodor also develop the intarray for pgsql. 23:06:09 intarray provide set operations for array of integers 23:06:57 so with intarray you can query a field (which is an intarray) if it is a superset, subset, or if it overlaps with another set. 23:07:27 but in a search engine we are interested to search for words, not integers. 23:08:21 so, openfts was built on top of intarray to parse text documents and assign a unique integer to each word. openfts also makes use of several dictionaries which are responsible for stemming and some ranking algorithms 23:09:14 during september, teodor improved the intarray package with extra operators (e.g. not) 23:10:29 he also replicate/copied some of the functionality (parser -- parsing documents to generate integer ids) into the tsearch modules. 23:10:34 s/modules/module 23:11:57 this enables the developer to write queries like "select * from foo where titleidx @@ '(the|this)&!we';" 23:12:05 i get the sense that oleg and teodor are really fucking smart 23:12:31 they are. @@=contains, !=or, &=and, !=not 23:13:23 tsearch translates this query to something like "select * from foo where titleidx @@ '(1234|455436)&!843;" 23:14:14 the openfts (which we use in our project) is doing this externally (out of pgsql) 23:14:28 in perl or tcl, yes? 23:14:35 both. 23:15:05 depending on whether you use the perl version or the tcl version, right? 23:15:10 in the future, openfts will contain the dictionaries, ranking algorithms but parsing and the rest of the stuff that I described will be included in tsearch 23:15:22 depending?? 23:15:59 brb dinner 23:16:23 davb: ok, we 'll talk tomorrow. 23:16:28 ok. enjoy 23:16:37 so tsearch is more similar to intermedia because it can be used inside pgsql 23:16:55 but the ideas are the same for tsearch and openfts 23:16:58 fyi, i meant that there are two versions, the tcl and the perl versions. each are different systems. but are mirrored in functionality 23:17:06 exactly 23:17:25 cool. so tsearch enjoys better performance because it is written in a language embedded in PG, yes? 23:17:32 yes 23:17:35 cool. 23:17:51 any idea what kind of performance benefit it gets? 23:17:58 is it in the orders of magnitude? 23:18:08 check this link: http://openfts.sourceforge.net/tsearch/README.tsearch 23:18:26 cool! thanks 23:18:41 towards the end 23:18:56 377905 titles 23:20:25 i must admit i can't read it. i don't know what the differences mean 23:20:30 but you can explain it to me later 23:20:39 sure, it's very fast 23:20:45 cool 23:21:24 tsearch also includes a morph dictionary which it's very interesting in it's own respect 23:21:53 stemming is used for the english language. when you follow english grammar (it cuts -ing from testing etc) 23:22:47 cool 23:22:52 now, think of astronomical terms, the names don't necessarily follow any grammar. so you cut the endings with respect to the morphology of a list of words that used to built your dictionary. 23:23:28 oh, that's cool 23:23:43 so you can create your own dictionary according to your particular need... 23:23:47 that's neat 23:24:06 yes 23:24:52 you provide a list of words. there is some utility (which more or less finds the endings). while searching the search engine will cut the longest common subsequence of this endings 23:25:08 cool! 23:25:12 you know what's the funny thing 23:25:17 can you create the dictionaries dynamically? 23:25:30 yes this is what I meant with morphology 23:25:35 ah, cool 23:25:47 know what's the funny thing? 23:25:51 what? 23:26:02 I still can't search in greek 23:26:19 but I did learned a lot and soon I will have the unicode parser built 23:26:33 so everything will turn out well. 23:26:46 nice 23:26:47 :) 23:27:04 k2pts: should it be offset=@url_next@ etc? It is saying offiset is not an integer 23:27:09 no 23:27:27 check the logs, i have posted what you should do 23:27:39 change search?q=@urlencoded_query@&offset=@offset_previous@ to @url_previous@, similarly for next 23:29:39 so I have now: h?q=@urlencoded_query@&offset=@url_previous@ 23:29:52 You guys need girlfriends =) 23:29:52 no 23:29:54 or hobbies 23:30:06 aside from OpenACS 23:30:08 Ok. Sorry I am just not getting it :) 23:30:19 change it to href=@url_previous@ 23:30:45 similarly for url_next 23:30:54 ack 23:30:55 aha. 23:31:35 * Psychephylax goes to order Dave a "mail order bride" 23:31:44 ..."don just in case you are reading the logs I am stuck with two outer joing in the wf package. waiting for Dan to reply"... :) 23:31:49 rofl 23:32:08 s/joing/joins 23:32:21 Dave, I'll try to find you one that's also into OpenACS but I think that's "extra" 23:32:28 rofl 23:32:32 lol 23:32:43 Psychephylax: you're here too, you know... 23:32:48 I know 23:32:54 but I'm watching TV 23:33:05 not discussing openACS ;) 23:33:07 my tv broke so you will see a lot of me around from now 23:33:10 maybe we can find *you* a bride that knows oracle syntax? 23:33:18 :) 23:33:19 lol 23:33:25 We sure can try 23:33:41 Try to find me that knows everything 23:33:45 oh...wait... 23:33:49 That'll be anyone... 23:34:01 Or at least they'll think they know everything....*women* 23:37:09 It works. Thanks! 23:37:29 The first page is over 1MB though. 23:37:43 davb: send me the file to commit it :) 23:37:47 Ok. 23:37:54 wow 23:37:57 very google like 23:38:02 yes, thanx 23:38:17 davb: disable the permissions to make it faster 23:38:29 go to the search package parameters 23:38:34 no 23:38:39 openfts-driver parameters 23:38:47 Ok. I see, it checks permissions for each link before sending it? 23:38:53 yes 23:39:12 i haven't done any optimizations yet. i have something in mind about that (next release) 23:39:58 talli: about doc and pdf: I have designed the search package in such a way that it is to index pdf, doc or other format files. i will automate it in the next release 23:40:10 s/it is/it is easy 23:40:22 k2pts: I send it with DCC, or should I email it? 23:40:28 you will need converters but they are available 23:40:51 can't receive it. would pls send it with email? 23:40:55 sure. 23:41:17 nkd: cool! 23:41:22 that's good to know. thanks 23:41:40 you are welcome 23:42:23 sent. 23:42:52 Gotta go perform pumpkin lobotomies :) 23:42:56 haha 23:43:01 davb: thanx. would you pls disable the permission checks so that people here can see how fast openfts is (the permissions are delaying) 23:43:18 Just make sure to stay above the pumpkin waist 23:43:33 ok 23:44:19 * Psychephylax wanders off to read some Electrical Engineering mumbo jumbp 23:44:27 s/jumbp/jumbo 23:44:41 should i restart? 23:44:55 no, thanx 23:45:03 could be my server making it slow... 23:45:10 0.11 sec 23:45:24 ok. 23:46:08 let me explain something about this time. openfts is a lot faster but we are also producing summaries (which means an extra db query for each result) 23:46:40 the summary function is a bit slow 23:52:39 need to go off to get some sleep. 23:52:51 buy guys 23:52:58 k2pts has left #openacs 23:53:58 Cool. 23:54:30 Take advantage of that while you can. I'll be taking the server down soon. I will be migrating my personal site this week over to OpenACS 4 so it will be back. 23:54:45 cool 23:54:52 you gonna use etp? :) 23:55:15 yes. I have to figure out how to get my old crappy weblog into that database. 23:56:27 cool 23:57:18 i'm going to write an email to you about some help we need for openacs.museatech. it will probably be tomorrow becuasa ei ahve a monster proposal that needs to go out tomorrow morning 23:58:13 Ok. 23:58:22 * davb is going to the dark side 23:58:36 (search page will temporarily disappear) 23:58:43 davb has quit