00:27:23 rbm has quit 00:31:21 * AaronSw reconnects 01:17:51 blueseed has quit 01:31:47 Psychephylax has joined #openacs 01:31:57 well that wasn't very rewarding 01:34:46 john has quit 02:11:05 graszew has joined #openacs 02:24:26 rbm has joined #openacs 02:27:44 donb has joined #openacs 02:28:02 rant rave rant rave rant rave .... 02:28:39 Anyone around? Everyone napping? 02:28:49 hi don! 02:28:53 keep on ranting :-) 02:29:14 Hey, Adam ... boy you used to work for a company that delivered some really *shitty* code. 02:29:20 haha 02:29:32 which module are you looking at this time? 02:29:59 i think ben should set up a poll: which module is the biggest train wreck? 02:30:14 * davb_ is back 02:30:23 It's a good thing Helen graduated from Tufts and moved back here. If I were visiting her in Boston I'd probably go buy and Uzi and vent my feelings on what's left of the company. 02:30:45 haha 02:31:03 it's amazing that for all the arrogance and bluster, the actual code that was generated was so awful. 02:31:41 If I upload an image, say "/home/donb/alex.jpg", into wimpy point no one else can load a file whose path ends in "alex.jpg" into wimpy point. That's the path on one's *local* computer, and *every* user! Grrr ... 02:32:15 (I mean "is true for every user" - i.e. the filename on your computer has to be unique among ALL SITE USER'S UPLOADED FILES!) 02:32:28 Is that a crock of shit or what? 02:32:39 hey don 02:32:50 hi docwolf and dave :) 02:32:51 Hey there ... 02:33:02 oh my, that is insane. 02:33:22 Continuing my rant and rave ... the problem is somewhat systematic system-wide because it's due to a unique contraint on the "name" field of content_items. 02:33:35 * Psychephylax growls at the moon and begins his transformation into a.... 02:33:37 Psychephylax is now known as Spork 02:34:12 Fortunately some (not all and not only the wp idiot) were smart enough to use object ids and leave the name NULL or work around it some other way, I need to start digging into this. Yes, it's insane and unacceptable. 02:34:38 sounds like a major problem and i've only been here a few minutes =) 02:34:53 It *is* a major problem and it will be a tedious PITA to fix. 02:35:04 mmmm...pita bread 02:35:13 * Spork closes his eyes and thinks of a falofel 02:35:16 Of course in the limited case of wimpy-point it doesn't really matter because you can't display the images you upload anyway, again due to aD problems. 02:35:38 at least I can solve my problems =) 02:35:45 well...no...I can't lol 02:35:49 I bug dave and mark 02:36:08 So the problem is that someone made filename a unique key or something? 02:36:40 Someone made the "name" of a content_item a unique key and wimpy point uses the filename portion of the path of the file on your local machine the item's name ... 02:37:31 I think the CMS must use the name as its key into items and manage namespace differently (i.e. by tacking on folder names). But the CR is supposed to be much, much more general than that and we have these nice little integers called "object_id" to find objects with. 02:37:35 can't you (meaning someone) just add an indexing column and create a sequence and create a new table mapping user_id to picture_id 02:38:26 The *bleeping* code doesn't *need* the filename to start with! It could just as easily use the object_id just like any -1 modded Slashdot anonymous coward could've told the author. 02:38:51 lol 02:39:14 Don, I think you need one of these things I've been giving Mark all along 02:39:27 * Spork takes out a freshly made chocolate chip cookie and hands it to Don 02:39:41 Would you like some milk? 02:39:46 graszew has quit 02:39:54 And it will do that before much longer. I've already e-mailed JackP, the guy who ported it, to see if he has time to work on it. He's not a very experienced programmer but a little supervision and he can fix it (if he has time). 02:40:01 I want some beer, probably ... 02:40:28 Hey, Karl Goldstein lives in Berkeley, I could be down there by daylight and give him hell in person. 02:41:10 * Spork gives Don a bottle of Goldshlager 02:41:13 (he didn't write wimpy point but he's responsible for a bunch of other crap, like the templating API that uses its own argument parser instead of the very nice ad_proc and doesn't check for parameters that aren't wanted). 02:41:30 heh 02:41:47 Did you know you can do "template::element -this_is_garbage $garbage" without getting an error from element complaining that it doesn't want that parameter? 02:42:10 One little slip of a finger while touch-typing and you've entered bug city ... 02:42:48 The unbelievable thing is that ad_proc's so nice and was available to his crew ... grrr ... grrr ... grrr .... 02:42:52 no I did not...but I am also not a very competent OACS programmer 02:43:13 Infact....I think I suck at programming period. 02:43:21 aD has a job for you! 02:43:23 Maybe I chose the wrong profession -=) 02:43:39 no, cablevision has a job for me! But I don't know for how much longer! lol 02:44:05 Actually plenty of aD folks are good programmers and plenty of them weren't very happy at a lot of what was going into ACS 4 ... 02:44:27 actually I guess that's "ex-aD folks were" ... 02:44:38 I don't know....The biggest problem I have with aD and ACS is the documentation does not "flow" 02:44:47 I still don't fully grasp how the stuff works 02:45:16 Because I was forced to read sections of the documentation written by the people who did the stuff and assumed I know what they did 02:45:45 Don, you ever pick up a book in a subject you were clueless about and try reading the book? 02:45:56 It could use a good primer, that's for sure. One thing that makes it hard is that there are multiple paradigms mixed up (like the template proc arg handler vs. ad_proc). ACS 4 was supposed to fix this by providing a more rational framework engineered free of the pressure of client customization work. 02:46:11 Yeah ... programming :) 02:46:14 hah 02:46:15 me too! 02:46:47 I kinda want to get laid off right now...so after I graduate I can spend 6 months slacking off learning stuff I think would be of use to me 02:47:10 I figured, hey....I got 2 BS degrees in 3.5 years...I think my .5 of a year is well deserved =) 02:47:44 A half year off is always deserved, which is why I've not worked more than half-time since 1986. 02:48:03 heh 02:48:06 I need a job like that 02:48:43 my job is alright....The hours are good and the pay (at least for me) is decent 02:48:49 So, Adam, you're awfully quiet! 02:48:50 yeah! static-pages uses the filesystem relative path for name. 02:48:57 hey dave 02:49:09 * Spork jabs a Psychephylax at Dave 02:50:02 Hmmm...at least that will be unique within the server ... 02:50:40 heh 02:50:46 They were smart enough to leave the extension on, I assume! 02:50:52 * Spork feels proud of himself...At least I understand this DB talk =) 02:51:26 I impressed my professor today...I wanted to use triggers and stored procedures to do half of my project for databases 02:51:41 He's like..."If you want to excersize that" 02:51:48 * Spork heads over to the Sybase site 02:52:12 Triggers, RI constraints (and "on delete/on update"), stored procedures are all really great things to be using to keep things consistent and in order. 02:52:30 PG has better "on delete/on update" support that Oracle (8.1 at least, haven't checked 9i). 02:52:39 Damn "than Oracle" 02:52:53 lol 02:52:59 Well, we use Sybase at school 02:53:14 Oh, that's cool. A lot of people really like Sybase. 02:53:36 I don't know...I use Oracle syntax on tests =) 02:53:42 He still marks them right...hehe 02:54:12 If you use "varchar2" you should be flunked :) 02:54:22 Why?! 02:54:24 :P 02:54:30 back. sorry for my silence. nice flame ;-) 02:54:44 don: as you're well aware, the berkeley folks were essentially a separate company. 02:55:26 Because "varchar" is standard SQL92 and supported by everyone and most RDBMS's don't have user defined types. So you need to edit any Oracle datamodel using "varchar2" to port it when (thus far) "varchar" works just fine in Oracle ("but it may not someday", they keep saying). 02:55:44 heh 02:56:31 i'd actually love to know what karlg, etc. are working on now 02:56:31 Adam, yes I do know that. And Karl had to do *everything* differently than was done in the ACS. Sometimes his efforts were real improvements, sometimes they were really bad, but always different. The notion of code-sharing with the rest of the company just wasn't something he cared about. 02:56:55 don: right. there was very little cross-talk between cambridge & berkeley. 02:57:03 Finishing up ACS Java, which means commerce, KM, and uh I forget but it's not collaboration. 02:57:13 markd2 has joined #openacs 02:57:32 I am less convinced than ever about the appropriateness of java for the class of problems they are trying to solve. 02:57:35 That's unforgivable. Not only on management's part, Karl's a good engineer and should know better without being told. 02:57:36 Hey it's Oracle Genius mark 02:57:49 Hey Mark I'm ranting and raving. 02:58:00 Yeah 02:58:07 (over the last couple of days i've had a run-in with a few weblogic sites that defy the imagination) 02:58:16 Don went on this cool rampage dissing people hehe 02:58:19 Oh, yeah, like what? 02:58:25 http://www.wml.com 02:58:33 this is a site for doctors to do their continuing education online 02:58:34 Not dissing people - dissing their code. There's a big difference ... 02:58:42 hehe 02:58:44 Close enough 02:58:46 it essentially consists of a "users" module, and a way to stream real video 02:58:54 it's a no-brainer of a site. 02:58:55 Web logic is wayyyyyy too expensive with their licensing 02:59:05 but... implemented in java using weblogic & oracle. 02:59:12 hiya 02:59:18 even the registration crashes. This is a multi-million dollar site. 02:59:35 I sent them an enormous "you suck" letter for wasting my time. 02:59:44 I just went there, Adam. It *talked* to me. I don't like websites that speak before they're spoken too. 02:59:53 one of our projects was using weblogic but switched to persistence 03:00:14 haha. Yeah, they hit you with the annoying real video on the front page. most heinous. 03:00:37 the whole thing is a disaster. They could have implemented it in openacs and saved at least $500k. 03:00:59 again, what slays me is that the site isn't at all complicated, nor is it super high-volume. 03:01:17 (in their wildest fantasies they'd only get 100k users) 03:01:55 and the reason i know they're using weblogic & java is because of the heinous messages it was throwing back at me, the helpless end-user. 03:02:09 I wonder if the "dot-bomb" could've been partially avoided if people had used good tools reasonably priced and licensed rather than spend so much on piece-of-shit proprietary systems that were thrown together hastily? 03:02:30 I mean the ACS was largely thrown together hastily but the licensing was reasonable! 03:03:22 hehe. reasonably priced tools would have helped; stuff like this is just terrifying though.. realize that these nudnicks take out full-page ads in the Journal of the American Medical Association ($$$$$) 03:03:48 lol 03:03:50 they clearly have more money than brains. The idea is good, the implementation is awful. 03:03:58 Huh ... what's the registration fee? Or do they have another revenue model? 03:04:17 after the "free trial" it's $25 a month. 03:04:31 which is actually reasonable, b/c continuing education can be really expensive. 03:05:14 Yes, it is very reasonable. They'll need decent market penetration to make decent revenue at that rate, though. Maybe they feel that it's so low that no doctor who hears about it can justify not registering. 03:06:13 yeah. their value proposition is good -- cheap CME, and lectures by people who are actually respected in the field. 03:06:46 (as an example, if I don't do my education credits with these people, i've got to go to a 4 day seminar, which costs about $2000, when all is said & done) 03:06:48 Another words the people who built the site aren't lecturing :) 03:06:55 haha 03:07:01 Damn in other words, my mind's tired. 03:07:31 actually, i looked at the management staff -- the company was founded by jerry della famina (the advertising exec from NYC..) totally wacky. 03:08:10 Well ... he obviously has a good idea ... 03:09:17 it's just a shame, though. If he'd hired in a small, crack team with folks like yourself 03:09:31 he would have gotten a better result for less money. 03:10:11 Hmmm...I'd prefer the "better result for the same money" myself! 03:10:17 haha 03:10:29 I'm also not convinced that they needed to shell out bucks to Larry Ellison. 03:10:52 if you look at the site, there are almost no financial transactions... or transactions of any type, for that matter. 03:11:33 i wonder if he just brought in some dolts from Accenture who told him "exactly what they needed" 03:12:08 Something like that, I'm sure... 03:13:36 Well ... I'm going to run off and try to get a little more work done ... 03:13:45 See y'all later. 03:13:47 donb has quit 03:19:10 vinod has joined #openacs 03:23:26 arg 03:23:37 I can't do my Oracle statements in Sybase =( 03:23:45 Hey Vinod! 03:23:56 you missed Don, he was venting =) 03:25:15 now that web/db is almost dead, he has no good place to explode. 03:25:42 lol 03:26:08 heh 03:26:33 I figured docwolf would license a tradmarked DonB 'you suck' letter 03:26:41 ow ow ow ow ow 03:26:46 both my legs fell asleep with a cat on them 03:26:50 and now they're waking up 03:26:51 Ahahaha 03:26:59 * Spork pokes them with a psychephylax 03:27:03 ouch! 03:27:31 i have to remember how to program in Java 03:27:40 I want my rest :( 03:27:56 use lots of objects 03:28:04 have fun, I'll be back tomorrow! 03:28:07 heh 03:28:08 ok 03:28:19 I'll be back here tomorrow waiting ;) 03:28:28 * Spork wonders if Vinod is awake 03:28:33 davb_ has quit 03:30:49 what was donb ranting about? 03:31:16 * vinod is barely awake 03:31:17 poor code from aD 03:31:25 Oh about poor coding on Wimpy Point's part 03:31:38 They made something a key that wasn't supposed to be ;) 03:31:39 ahhh... (phew) ... i'm always worried it's gonna be about bad code from me :-) 03:31:45 nah 03:31:47 someone else though heeh 03:32:02 oy 03:32:05 I have so much crap to do 03:34:22 Vinod, are you going to be around tomorrow morning? 03:34:59 i dunno - my life is quite unpredictable ;-) 03:35:03 heh 03:35:13 he is a super-hero, after all 03:35:16 I gotta get that Oracle problem fixed asap or I won't make any progress 03:35:23 lol 03:35:35 Boss gets mad when I make no progress 03:35:43 markd2: no... docwolf is the superhero 03:35:51 that's right 03:35:57 you're the Trusty Side-Kick 03:36:15 Don't you mean Site-kick? 03:36:17 at your service! 03:36:25 heh 03:36:58 Vinod, so you have any ideas what it might be off the top of your head? 03:37:08 Spork: did you run that last query that i asked about before you left this afternoon? 03:37:19 nope ;( let me do it now... 03:37:49 let's hope Dave's box is on now 03:38:23 looks doubtful 03:38:32 Page cannot be displayed =-( 03:39:07 declare v_id integer; begin snmp_incident.new(mac_address => 'junk', problem_id => 1 ); end; 03:39:36 ERROR at line 1: 03:39:36 ORA-06550: line 1, column 29: 03:39:36 PLS-00221: 'NEW' is not a procedure or is undefined 03:39:36 ORA-06550: line 1, column 29: 03:39:36 PL/SQL: Statement ignored 03:39:53 sorry 03:40:10 stick a 'v_id := ' before snmp_incident 03:40:15 ok 03:40:16 but 03:40:22 there's 2 parameters I must pass on as well 03:40:26 for new to work 03:40:33 problem_id and comments 03:41:16 comments is default null (in the file you sent me) 03:41:25 oh 03:41:27 only mac_address and problem_id are needed 03:41:27 heh 03:41:53 aha 03:41:55 ERROR at line 1: 03:41:55 ORA-02291: integrity constraint (MYOPENACS.ACS_OBJECTS_OBJECT_TYPE_FK) violated 03:41:55 - parent key not found 03:41:55 ORA-06512: at "MYOPENACS.ACS_OBJECT", line 68 03:41:55 ORA-06512: at "MYOPENACS.SNMP_INCIDENT", line 39 03:41:56 ORA-06512: at line 1 03:42:01 that's the error alright =( 03:42:50 I am not sure what it's doing 03:42:56 markd2: is there a way to list a package function's code from sqlplus? 03:43:03 pointing at my object and failing or me pointing a diff 03:43:03 yeah 03:43:04 yeah 03:43:05 wait 03:43:08 let me try answering =) 03:43:18 user_source I think 03:43:21 select * from user_source 03:43:22 * markd2 fires up a shell window 03:43:23 heh 03:44:37 donb has joined #openacs 03:45:03 Well for those who are there I wasn't quite right about the CR's sitewide uniqueness constraint on items ... 03:45:20 hi Don 03:45:24 It's a uniqueness constraint for items with the same parent, i.e. in the folder ... 03:45:30 Don 03:45:34 Maybe you can help 03:45:34 Yeah ... 03:45:42 If it's quick ... 03:45:43 Have you ever seen a parent key constraint violation? 03:46:00 Nope ... 03:46:07 ORA-02291: integrity constraint (MYOPENACS.ACS_OBJECTS_OBJECT_TYPE_FK) 03:46:07 violated 03:46:11 sounds kinky 03:46:28 * vinod pulls markd2's head out of the gutter 03:46:33 i am not sure if it's my objects not introducing themselves correctly or if it's an openacs bug 03:47:24 vinod: SQL> select line from user_source where name='PA_PHOTO' order by line; 03:47:31 you'll probably want to do a 'set pagesize 5000' first 03:47:38 thanks markd2 03:47:50 It's telling you you're referencing an object type that doesn't exist. 03:47:57 right 03:48:15 but is it my object that doesn't exist or am I referencing something that doesn't exist 03:48:41 You're trying to create an object of type "foo" where there's no object_type "foo" defined. 03:48:43 because my snmp_incident object does exist 03:49:03 What's the "new" you're executing? 03:49:58 snmp_incident.new(params) 03:50:44 what table can I check to make sure my stuff is in there properly? 03:51:36 Sure, but what are the params, in particular the "object_type" param? 03:51:53 ummm 03:52:48 a problem_id (a number mapping to a problem description) a mac_address and a few parameters that have a default value generated so I haven't worried about those yet since the problem occurs when I pass 2 things in only 03:53:39 It supposedly defaults to "acs_objects", which of course exists. You might be passing in NULL as the constraint is also "not null" ... 03:53:57 what do you mean? 03:54:26 you mean I try to stick a null into a not null column? 03:54:54 Either that or your passing something in the wrong position if you're not using the "parameter_name => value" notation. 03:55:05 hmmm 03:55:12 lemme look at my pl/sql block 03:55:18 Right ... 03:55:30 Well I've got to run again, sorry, haven't eaten yet and it's 8 PM ... 03:55:35 it's ok 03:55:40 I'll cover those bases 03:55:40 seeya don 03:55:44 donb has quit 03:56:48 well, I am using parameter_name => value binds 03:56:52 so that's not the problem 03:57:13 so let's look for null insertions into NN constraints 03:57:38 Vinod, here's a question for you 03:57:50 You still got my create file somewhere? 03:58:44 * Spork pokes Vinod ;) 03:59:31 yep i got it. i'll be back in a sec 03:59:36 ok 03:59:45 I found one spot that's a null 03:59:55 but that should be taken care of acs_objects 04:10:27 back yet? 04:15:24 he's prolly out saving lives 04:15:35 Pshhh 04:15:43 I thought he went to do something important =) 04:17:32 i'm back. i told the nurse to keep the pt alive til i fix spork's problem 04:18:00 found yer problem 04:18:10 "Hold it there! Don't die on me! I've got a quick hack to make and I'll be right back" 04:18:29 lol - i can see the lawyers circling around me :-) 04:19:25 Yeah. Just like their sister species: vulturs 04:19:41 Spork: oracle apparently uses the param defaults that you supply in the PACKAGE definition, not the PACKAGE BODY definition 04:19:50 haha 04:19:58 that makes sense 04:20:37 you have object_type defaulting to 'incident' in the declaration, but 'snmp_incident' in the body 04:21:45 oh 04:21:57 markd2: what do you use for quick debug statements in plsql - is it dbms_output? 04:22:26 yeah 04:22:29 ok 04:22:32 so what I have to do is... 04:22:33 dbms_output.putln (or maybe putline) 04:22:46 put_line 04:22:51 ohhhhhh 04:22:52 set serveroutput on size 10000 04:23:02 i was trying dbms_output.put 04:23:35 and i guess oracle keeps adding to the line until you do a put_line 04:24:01 Vinod, I don't see where I have incident for my object 04:24:40 i have a create or replace package snmp_incidernt 04:24:41 as 04:24:45 function new ( 04:25:05 but I don't see where I'm delcaring object_type to be incident by default 04:25:34 the file you sent me has: object_type in acs_object_types.object_type%TYPE default 'incident', 04:26:47 aha 04:26:51 it was at the bottom 04:26:56 so I have to change that to snmp_incident 04:27:06 yep 04:28:39 heh 04:28:48 * Spork tries doing a modem lookup 04:29:24 * markd2 is up way past his bedtime 04:29:45 wooohooo! 04:29:47 ora8.c:3568:ora_tcl_command: error in `OCIStmtExecute ()': ORA-01400: cannot insert NULL into ("MYOPENACS"."SNMP_INCIDENTS"."CM_RCV_PWR") 04:29:47 ORA-06512: at "MYOPENACS.SNMP_INCIDENT", line 47 04:29:47 ORA-06512: at line 1 04:30:00 * vinod slips caffeine into markd2's cookies 04:30:06 rofl 04:30:15 * markd2 shakes, rattles, and rolls off the chair 04:30:18 You know...this isn't ACS related but I figured I'll tell ya all anyway 04:30:48 I might not like ACS, I might not love Oracle (but I do :)) or what I'm doing with it...but if I didn't have to do this I wouldn't have made friends here =) 04:31:19 awwwwww 04:31:27 * vinod sheds a tear 04:31:29 * markd2 takes the pins out of the Psychephalyx doll 04:31:30 hehehe 04:31:40 Wow, my headache is gone! 04:31:42 * vinod is waiting for spork to ask for money now 04:31:48 heh 04:31:52 :-) 04:31:55 it is the local classical station's beg week 04:31:55 I was expecting you to ask for money instead 04:31:56 lol 04:32:22 That oracle error was driving me nuts and I don't think I would have been able to figure it out on my own 04:32:33 you've just heard 2 minutes of mozart. give us money and we'll play another minute 04:32:33 at least not by the 9th of when my demo is due 04:32:54 Spork: here's how i woulda figured it out (maybe) 04:33:10 * Spork listens 04:33:20 we knew it was crapping out because it didn't find a proper acs_object_type 04:33:33 right 04:33:37 so, first select * from acs_object_type and see if my object_type is there 04:33:58 ok 04:34:01 and if it was 04:34:06 snmp_incident 04:34:06 Wow. I didn't know Amazon had switched to Linux 04:34:09 http://news.cnet.com/news/0-1003-200-7720536.html 04:34:12 Hey Roberto 04:34:20 if so, then see what you are sending to acs_object.new as object_type 04:34:33 i assumed 'snmp_incident' 04:34:43 markd2 has quit 04:34:51 but then i stuck in a debug line before the call and found out that it was 'incident' 04:35:02 ah 04:35:04 and that was the 'aha' moment :-) 04:35:21 :) 04:35:47 now I must figure out why it's trying to insert null into cm_rcv_pwr 04:35:55 i can figure that out...I think 04:36:06 but I think I will leave that for work 04:36:22 I can't do all my work at home..hehe.... 04:36:32 :-) 04:36:41 Vinod, how did you get it to print what it was sending though 04:37:36 i stuck this just before the call to acs_object.new 04:37:41 DBMS_OUTPUT.PUT_LINE('object_type='||object_type); 04:38:29 ah, well, see...I don't know that... 04:38:32 but I will write this down 04:38:50 me neither, that's why i looked it up in my BIG ORACLE BOOK :-) 04:38:55 lol 04:39:11 page 559 04:39:33 mine is in the car :( waiting for it's commute to work tomorrow morning 04:40:24 ahh - you can always check oradoc.photo.net in a pinch 04:40:51 wow 04:40:52 nice 04:40:58