[00:09] Last message repeated 1 time(s). mdean (mdean@arc11x105.kcnet.com) left irc: Ping timeout for mdean[arc11x105.kcnet.com] ra3vat (ds@195.239.66.45) left irc: [x]chat TuxSleep (taft@as2-dial23.esjd.mi.freeway.net) joined #gnuenterprise. TuxSleep (taft@as2-dial23.esjd.mi.freeway.net) left #gnuenterprise. uday (uday@host-64-110-96-159.interpacket.net) left irc: Client Exiting reinhard (rm@62.47.44.46) joined #gnuenterprise. ra3vat (ds@195.239.66.41) joined #gnuenterprise. reinhard: you here? yes reinhard: had no luck with neither libpython2.0.so no libpython2.0.a where is libpython2.0.a for you? uday (uday@host-64-110-96-159.interpacket.net) joined #gnuenterprise. SachaS (sacha@chime-a-199.conceptual.net.au) joined #gnuenterprise. reinhard: initially it was in /usr/lib/python2.0/config/libpython2.0.a ah and now? i was doing it on another machine at home i was trying to add directory in ld.conf and rebuild it ok can you try to do LDFLAGS=-L/usr/lib/python2.0/config ./configure --your-other-parameters and then make? ok, any other instructions before i switch off from irc i can try all this only later today thanks bbl ra3vat (ds@195.239.66.41) left irc: Client Exiting uday (uday@host-64-110-96-159.interpacket.net) left irc: Read error to uday[host-64-110-96-159.interpacket.net]: EOF from client SachaS (sacha@chime-a-199.conceptual.net.au) left irc: Ping timeout for SachaS[chime-a-199.conceptual.net.au] perlh (a@202.9.144.242) joined #gnuenterprise. hi guys i am new to this like i am trying to download the crm package but i am not able to find it where can i get it well then is there some one who knows in depth about bayonee trying to install it is some one here hello perlh: you trying to install bayonne? yeah reinhard like i also need more info i can help you with info i dont know the links from where i can download i think i can't help you with bayonne oh ok fine i dont know the links from where i can download for bayonne you should find on www.gnu.org/software/bayonne GNUe Packages Accounting Customer Relations Forecasting Human Resources Manufacturing Project Management Sales Supply Chain i dont want the accounting package but rest i need to know from where i can donwlaod can you tell me whom should i contact to get bayonee related info r u there ri r u there reinhard sorry phone call :( oh ok fine r u back still on the phone oh ok fine pls msg as soon as your back pls ok thanx will david sugar come online are u back reinhard now back david sugar sometimes comes online here but not regularly under the name "dyfet" wrt GNUe Packages there is not really much you can download yet we are still working on the basics oh ok cool like the form generator or the application server itself oh i sit well so like when can expect it to be up all the packages are -- umm.. --- let's say in conceptual stage or i mean at downloadable stage normally we say in 4 months but we have been saying that for 2 years now ;) seriosly we _are_ making progress atleast how about Customer Relations Project Management Supply Chain pakacges how long do you think it will take more but it's all volunteer spare time work so it depends on how many volunteers we find to work on it actually yeah i agree and more over one more thiong i tried to hunted around to learn what the dif relase means that is what a milestone release means ? and what are the other types of release and what do they mean are u there reinhard well reinhard can u tell me on what package in gnue are you working in reinhard are u there? a milestone release is simply (sorry had a phone again) oh ok no probs i am sorry to distrub u a milestone release is simply yeah tell me a release where you reach a specific goal oh ok ... so if you have a release 0.2 and for the first time feature x really works and what are the other types of releases it is a milestone oh got it and if you release 0.3 and it contains only bugfixes it's not a milestone oh ok and what are the other types of releases ... in most cases milestone releases have numbers like x.y.0 oh ok cool and bugfix or maintenance releases are x.y.z where z>0 i currently maintain GEAS (the application server) together with neil tiffin (neilt) i also want to contribute thats why i am trying to get the info what are your major interests business stuff well for me or hardcore c code hacking c and c++ i am perl programmer basically um i guessed that much from your nick ;) and i am influenced by the ivr and stuff like the only part of gnue that is coded in c is geas so probably you may want to get geas from cvs and look at the code, try to understand it find out what you don't like about it and change it 100% sure you _will_ find something ;) oh ok fine www.gnuenterprise.org contains instructions how to access cvs if you have problems with compiling geas please tell us we are working on making it very portable yeah sure will do that but my interest lies in and around bayonee first step should be to make it compile and to run the examples ah ok oh ok ... bayonne is still a very seperate project so i need like to contribute for it and for geas but it is listed under geas we will just integrate it someday but it's a completely seperate team afaik oh ok at least i don't know anything about bayonne like can i do this for them . We are also looking for a someone who might be brave enough to volunteer to operate, or that would be willing to contribute to support deployment of, a public access Bayonne driven XML telephony voice portal. maybe derek why no body here speaks is there any reason probably you should contact david sugar by email all of them sleep oh ok fine :) most are in the u.s. and just don't log out over night Action: perlh thanx for reinhard being awake must leave now sorry :( hope you get in contact with david ok sure so shall i hang around here well so i will donwload geas today in 2-3 hours others will wake up probably ok and try installing it cool i am online most of the time, too thanx a lot reinhard oh cool that is daytime europe then it would be great l8r reinhard (rm@62.47.44.46) left irc: Those who say it can't be done should never stop those who are actually doing it. what is that 18r ? uday (wirc2@host-64-110-96-159.interpacket.net) joined #gnuenterprise. hi uday perlh hi well uday are u from india perlh yes oh cool where in india uday then what is that of gnue do u look into perlh I am a newbie here, I wish to develop a package like supply chain or .. perlh how about you well i am into bayonee and trying to get a hang of it which part of india are you from brb perlh I am from hyd hi what's bayonee bayonee is a telephonic software where does that fit into gnue? it is now trying to find a place there as of now it is only a seperate software da like it replaces all your telephonic needs like of having a seperate call center filled with human so where are you from ( I guess in India) yeah i have been to india i am from bdesh :P why :P just like that well what do u do uday I am into my final year undergrad oh ok cool engineering ? yes computer science oh ok cool which college how about you how long have you been with gnue? well for over 6months which college are you from from IIIT, you work on linux or yeah i work on linux :) you busy? yeah kind of tell me so like were you born and bought up in hydi yes oh ok neilt (neilt@dialup-63.208.68.139.Dial1.Chicago1.Level3.net) joined #gnuenterprise. #gnuenterprise: mode change '+o neilt' by ChanServ!s@ChanServ uday (wirc2@host-64-110-96-159.interpacket.net) left irc: You could be a WiRC-warrior too - WiRCirc.cjb.net derek (derek@cpe-24-221-112-50.az.sprintbbd.net) left irc: Ping timeout for derek[cpe-24-221-112-50.az.sprintbbd.net] derek (derek@cpe-24-221-112-50.az.sprintbbd.net) joined #gnuenterprise. hi derek perlh: hello i dont suspect derek is up yet hi neilt how do you do i am interested with bayonee fine and you and like to contribute for it cool so like i was speaking with reinhard he asked me to chat wiht derek yep, derek is the right person well neilt what gnue app do u take care of or r u involved in [07:13] Last message repeated 1 time(s). i do geas and the business applications supply chain, acocunting, sales etc oh cool all these are developed in what language neilt i know perl python can i contribute in any way GEAS is written in C GEAS is our multi tier applications server ok cool our business rules are written in python right now and all these packages are written in phyton right by sometime next year we should have a multi loanguage interface well will you be writing them with other langauges as well like perl php and other so you can use java c perl or whatever coz i would like to contribute oh so right now geas the application server can onlky understand phtyon right so in near future it will also support other langiuages right we had some interest about integrating with "double choco latte" which is a php web front end that could access all our business objects derek is the one to talk about that also oh ok cool so why dont you guys use perl at all :( near future = next year oh ok so why dont you guys use perl at all license problem i htink think what license problem like perl is also open source right then ... we are NOT open source we are free software fsf yeah i know that uday (wirc2@host-64-110-96-159.interpacket.net) joined #gnuenterprise. well neilt ... but again, this does not matter to me who is taking care of crm thing but it does to someone oh ok the issue is that we use clean free software for our base develment so right now you develop the business rules only in phyton or what ? but interface to other languages for example we also interface with java oh i got the logic but it is not our core oh o language right now we can use python or C for business rules we have had some interest in crm but no work that I know of we would need a proposal for crm package chillywilly is currently working on a general language abstraction layer proposal as such what do u mean we have had some interest in crm cant undersatnd neilt hold on we have descriptions started for Sales, Base, Supply Chain, and Accounting we dont have a description of crm these are at http://www.gnuenterprise.org/~neilt/sc.html the first step for a crm package would be to create a specification oh ok cool similar to the above yeah i get it ok just a minute i will take a look at it and come well how do i also get involved in the gnue ? you contact derek by sending email to info@gnue.org and let him know your interest area he will send you a FSF assigment oh ok cool and you start looking at docs and asking questions here shall i mail him today then you contribute ok fine sure well suggest me an area no ? so that it would be easy for me as well coz i am interested in being part of it we like to use people in the area they are interested in we need help every where so if you like crm that would be one area if you want to hack C then geas would be another area well i guess i can get into crm and ... human resource if you want to learn python we have a lot of GUI code that you could help with HR is good oh ok cool we dont have a start there either yeah i think it would make a lot of sense to :) ok cool if you want to create a perl GUI driver you could do that also although that would not be our first choice i think i prefer the perl way though i dont mind using python oh ok :( perl is pretty good na ok so tell me what shall i mail derek like yes, i like perl also oh cool have you worked in perl sure or do u work on python right now i work in python here oh ok because it was the language chosen like do u work some where else too so like now can it be made that the crm and hr are done in perl ;) its possible within our GEAS design specifications but it depends on getting multi loanguage support in GEAS which is a year away oh ok yeah i get it so lik enow it must be in python right ok no probs cool for the most part yes like do u work some where else too neilt converting the specification into a working package involves writing mainly the gcd files and then implementing the business methods and finally writing the clients, anything more? uday: exactly plus writting report definitions perlh: i support myself as a management consultant neilt so the problem of language choice is part of only the business method and client code oh ok cool neilt ok so tell me what shall i mail derek like so my involvement is mostly from a business standpoint, not technical oh ok cool :)0 uday: the language choice depends on an interface from GEAS to the language system currently we only have C or python and even then you have to chose at GEAS compile time which language you are going to use the client code is not effected neilt the geas server has to 'read' the methods and make them available to the client ToyMan (stuq@c5300-1-ip165.albany.thebiz.net) joined #gnuenterprise. because all of the clients, in any language, java perl, python c communicate to GEAS via CORBA idl oh ok neilt client is no problem then, even the platform does not matter yeha that is rieght uday uday: almost, it has to locate and register the code and dispactch to the code when the method is executed well why dont you use soap or xml rpc instead for the calls my understanding is that soap and xml rpc were not mature 1 1/2 years ago when the current server was started but there could be other reasons and that part is in a complete rewrite right now so it may use soap or xml rpc chillywilly is doing the rewrite oh ok cool i think then it would be very very easy rather than corba yeah i think that time soap and xml rpc were not that matured "rather than corba" is big statement, unfortunately GEAS is structured around CORBA as the only solution, the next major project with GEAS is to restructure it and remove the CORBA requirement and make CORBA an options, so other technologies can be plugged in yeah this would be great too neilt: when do you plan to start the development of packages? uday: we have already started but we go it bursts because we need some functionality in GEAS completed before we know how to create the package so we have some basic packages defined in cvs at gnue/gnue-config and the specs at http://www.gnuenterprise.org/~neilt/sc.html in fact the specs have all of the business rules, gcd files and default data in them but the package specs need a lot of work up until a couple of days ago methods were broke neilt: but most of them are being modified and edited at present so we have been trying to get them working again uday: yes, they are a work in progress our approach to packages has been to provide a place to capture ideas and as people contribute we build and expand on the modules so if you want to review any modules and provide comments or suggestions, then we will discuss via email or irc and when reinhard and I agree we put it into cvs ok that does not mean that we always agree before it goes into cvs for example in accounting we have two different structures in cvs and we are working on finding the best approach neilt: I guess it might take some time before finalising the proposals jamest (jamest@hobbes.math.ksu.edu) joined #gnuenterprise. #gnuenterprise: mode change '+o jamest' by ChanServ!s@ChanServ yes cool neilt but not much we know the structure of the modules from the last year of talk what we need now is to stabilize geas so we can finish developing them and a big part of that is that forms does not currenly work with geas, although jamest has checked in some of the changes required in the last few days and that should be fixed in the next 2 or 3 weeks, hopefully it's all I'm working on for the moment so it should be next week (after my test) I think we will make great progress in the next 2 months on packages we have a callable methods system in geas now? hi jamest jamest: cool jamest: yes we do cool hi perlh neilt so tell me what shall i mail derek hi shall i tell him that i am interested in crm and hr packages the working example is at gnue/geas/examples/python/neil_methods_test.py perlh: you going to start working on gnue? perlh: that seems to make sense, yes neilt: It needs some extra classes in addition to test and addressbook, they were not included in the cvs yesterday yeha neilt jamest: i think the query code in geas is flakly yeah jamest perlh: we'll need copyright disclaimer before we can check stuff into cvs uday: not sure what you mean is it jamest neilt: I'm real good at segfaulting geas now w/ forms driver jamest copy right disclaimer cool neilt: real, real good we have never segfaulted GEAS in our tests, so your new nick is mini-masta i think I can segfault it w/ a call to executeQuery and getFullClassDefinition if you grab the latest cvs and run the user.gfd sample we have asserts and g_asserts( whatever they are called all over the place) she'll segfault 10 times out of 10 neilt tell me what else do i need to mail derek? its possible that you are hitting our stuff with a bat :) perlh: please include your background and skills oh ok cool perlh: if you just need copyright disclaimer I can send it to you i think perlh: info@gnue.org goes to a bunch of us, that way we know what you can doo yeah jamest it would be very useful the copy right disclaimer ok sure perlh i will compose a mail tonight and send it s/perlh/neilt jamest/ jamest can you send me the copy right disclaimer yes but please send the stuff to derek to as we do try and track things uday: neil_test_methods.py was meant to work with gnue-config not the normal test dir i'll need a mail address perlh: send to info@gnue.org oh ok cool ok sure will send it there can u just summaries what and all i need to send sure email address skills experience (brief) area you would like to work on ok cool name and location (country, city) i just need the mail address right now as I've got the form up and ready to mail it will make sure i funrnish all this stuff :) :) perlh: jamest would like your email now and he will send you some stuff neilt: ok I get it - with gnue-config oh ok sorry did not listen it just a sec perlhin2k@yahoo.com sorry just now created this mail i thought may i can use this for gnue uday (wirc2@host-64-110-96-159.interpacket.net) left irc: Now, with power steering! WiRCirc.cjb.net perlh: the mail is sent btw - anyone see the change to the gnome.org site at least I think it's a change perlh: you should also subscribe to the gnue-discuss mail list thanx jamest well neilt can u tell me the email address so that i can subsrice http://www.gnuenterprise.org/ you should see a selection for mail lists yeah i am there already ok cool yeah got it do i need to subsribe only to the thing which i am interested or like i mean do i need to sub only to crm and hr perlh: right now most stuff goes to gnue-discuss if you look at the logs you not see anything i hr or crm i hope what happens is when we start gettting a lot of traffic we move from gnue-discuss to the more detailed list oh ok so i better subrsice to gnue-discuss right and we post an email so people know right ok cool jamest: you will happy to know that in the last hour i have executed 6 querys i have a timing loop going to time querys that is one query every 10 minutes what is that for you r doing that neilt to test geas we are working on some improvements in the query code in geas 1 every 10 ? with all due respect the the query is on the zip code database yes 1 query every 10 minutes oh cool chillywilly (danielb@d62.as29.nwbl0.wi.voyager.net) joined #gnuenterprise. is that what's it currently running full speed and the returns are random i hope not :) or are you timing it like that 8902 0 [08:25] Last message repeated 2 time(s). 15638 0 [08:25] Last message repeated 1 time(s). 6064 that is the result count so i am doing some really big queries cool from a big database what's the avg return time not timing that ah you know that you can compile GEAS with the profiler just the overall run time ok using a gcc flag plus andrewm added profiling code when it is done it will give me the average time per single object returned sunno why he did it that way chillywilly: we are remoming the profiling code so we can do it the right way it is stupid because the compiler will do it for you exactly jamest: i also have full debug reporting on so my terminal window for geas is running wild but i am not getting any seg faults for some pretty large queries jamest: have you checked in your code? yes Action: neilt makes his first forray into gnue-common it could very well be me misusing a method but it should segfault look in gnue-common/src/dbdrivers/geas/DBdriver.py give me a sec and I'll find the line have to get new cvs i updated this morning is that new enough yip btw - this driver is a mix of several pieces of gnue-common, and the old geas drivers IOW - it's not purdy :) yes er, yet ok, search for getSchemaList goit you want the function definition not the call to it gotit that call to self._dataConnection.getFullClassDefinition segfaults every time the reason i do a str(classname) is that in the query system if I passed a INT as a field value geas would segfault let me get this puppy running here so I can give debug output it'll take a bit :) are you using a "guest" user wehn are we gonna have some UML diagrams of the packages? ;) never great, just great... unless, someone creates a UML to gcd tool you are lazy then we will move all package development to UML chillywilly: yes i am and proud of it yes I am using guest point is ppl can know how things are supposed to work with reading all those damn gcds Action: jamest isn't a huge fan of detailed diagramming except that everyone wanted to write field lengths and names bah and not write uml i think general diagrams are very nice you don't know what I really want to do to you ppl and if we write uml then we have to had translate to gcd http://www.objectsbydesign.com/books/larman_process.html then the uml or gcd will never agree neilt: that's the biggest problem with detailed diagrams they don't sync exactly and without writing use cases, and using other diagram techniques you cant document the design anyway bullshit so just UML is not the answer use cases is what you should start with what?!? obvously you know nothing about real design i was doing real design when you were in diapers plz so be careful there's more than just use cases and class design diagrams read the whole damn thing chillywilly: how many people are running software you have deisgned will there are over 30,000 people working on software i have deigned just the MSOE TRIO Porgrams (internally noe) now 30,000 users that is I don't see how you can attack that design process damn, logo apps scale that well? including systems with over 60 sites all over the damn world that's like shooting yourself int the head chillywilly: I don't think it's an attack on the design process it's saying that in the real world specs rapidly fall out of sync with the code chillywill: obvously you know nothing about real design no i do know real deisign it was an attack in that statement I know that it is an iterative process and it pays lots to keep the docs in sync even if it is a pain sometimes from chillywilly and so you're overloading yourself w/ someone needing to constantly keep a fine grained design doc in sync w/ the code i have used uml and a half dozen other development methodologies before it and in business world you have to justify this cost to a mgr that cares nothing about anything other than "why isn't it done yet" and in free software world you can't force a coder to update the docs to match his patches/changes so a more realistic approach is general design docs IMHO well, I am starting another prorject for Free CASE tool and developing the Free Software Design Process based heavily on Larman's Process chillywilly: thats cool chillywilly: if we had the tools that integrated with the software developemtn process i would be the first to use them once you have a design you can code it in anything even logo i have already hit my head against the desc several times wishing i had UML diagrams desk but its not worth the effort until the tools improve I still think using dia wouldn't hurt then again I usually do dia2code right after that althought I used it for a PHP app too which you can't generate code for oh, one other thing with geas i had an old geas install on the system at home when I updated to latest cvs jamest: yes i was getting errors about unable to insert record due to a null field insert into a not null marked field that is an old problem what seems to have happened is that the geas tables changed and were not updated we did a really ugly patch to fix it :) you need to remove all tables and let geas rebuild them cool, just a FYI from a "newbie" geas user geas can only make certain changes to tables will really geas can not make any changes to tables it can only build tables and add new fields it wont remove fields or change fields once the tables are built ah, limitation of the backends i think andrewm was scared that someone would lose data from a mistake so its built to not make those changes too bad it doesn't create tempfoo as select * from foo then drop foo, recreate foo and then insert into foo select stillExistingFields from tempfoo maybe as a seperate upgrade script ack! that sounds like work! its not as hard as it sounds and is a future refinement dres (dres@4.18.171.42) left irc: Read error to dres[4.18.171.42]: Connection reset by peer dres (dres@4.18.171.42) joined #gnuenterprise. but you could also do that from psql? yip Action: chillywilly whispers aomsething about using xml as data interchange format, but then again....it is only in the ODMG spec OIFML would that be the point of the "integrator"? what is intergrator supposed to be anyway? integrator is an xml conversion tool because it does not make sense to have xml in the main data repositories k it could also be an edi conversion tool or any other type of mapping tool between private data and public data this looks cool: http://www.objectsbydesign.com/projects/xmi_to_html.html well i dunno what edi is other than it stands for enterprise data interface :P it was xml before xml oh and it was not standardized xml isn't really either other than being structured...unless you make a DTD that everyone accepts oh, btw, someone once told me that it isn't very good if engineers aren't arguing with each other...you should be very worried if they aren't :P chillywilly: yep, you know i dont mind arguing hehe errrr, discussing yeah, that the politically correct way to say it :P perlh (a@202.9.144.242) left irc: Ping timeout for perlh[202.9.144.242] jamest: did you ever post the error report for geas geas won't start on solaris how come? :( Creating ConnectionFactory instance for new connections [function call] [geas-skeleton.c/2291] [impl_GEAS_ConnectionFactory__create] Initialising security system and it stops there heh, security system we have no stinking security system i never get the GEAS ready line i thought we had a security system that stinks :) neilt: that's the little password that you type in ahh the one that does not do anything yea btw i have finished my 7th query yet in my test program and its been an hour why are they taking that long? i think the numbers are getting worse does geas do the filtering itself? that's the firat time it tries to play with the ORB? excuuuuuuuuuuuse me, i am on the 9th query void init_security_system(void) { GEAS_ObjectList users = CORBA_OBJECT_NIL; CORBA_Environment ev; char *admin,*adminpw,*p; CORBA_exception_init( &ev ); ... ok python is not responding to control c C-d is there another key to make it stop th interprter? C-d well, gnome doesn't run on the sun box geas is on i ran it from the shell prompt is there something in orbit I need to have running in the background? nope jamest: well you need orbit wait, gnome isn't on my home machine either i have orbit jamest: no orbit runs inside geas euler$ orbit-config --version ORBit 0.5.8 didn't know if geas required a naming service or something same here no naming service yet nope you need to start a database then start geas thats it ya but it would've gotten farther and it would've core dunped dumped it is accessing a db on a remote host via my user id seems to be outputing select statements in the debug query will be select a.attname, t.typname from pg_class c, pg_attribute a, pg_type t where c.relname = 'rtest' and a.attnum > 0 and a.attrelid = c.oid and a.atttypid = t.oid query will be select a.attname, t.typname from pg_class c, pg_attribute a, pg_type t where c.relname = 'timetest' and a.attnum > 0 and a.attrelid = c.oid and a.atttypid = t.oid Initialising ORB Creating locator objects Creating ConnectionFactory instance for new connections [function call] [geas-skeleton.c/2291] [impl_GEAS_ConnectionFactory__create] Initialising security system i need to go to class bbl RMS is gonna be in Chicago on Ovt. 31st er Oct. you gonna go neilt? :P that's halloween isn't it? chillywilly: how do i get gnu/linux to tell what type of system i am on from the shell prompt probably now no interest no time probably not jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) joined #gnuenterprise. yes, it's me :) hi howdy howdy, its awful early for you howdy neilt: uname -a neilt: why would you not be nterested in listening to RMS speak? thanks because i dont have the time and his thoughts have no value to me right now dres (dres@4.18.171.42) got netsplit. ajmitch (ajmitch@p20-max7.dun.ihug.co.nz) got netsplit. derek (derek@cpe-24-221-112-50.az.sprintbbd.net) got netsplit. chillywilly (danielb@d62.as29.nwbl0.wi.voyager.net) got netsplit. jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) got netsplit. dres (dres@4.18.171.42) returned to #gnuenterprise. ajmitch (ajmitch@p20-max7.dun.ihug.co.nz) returned to #gnuenterprise. chillywilly (danielb@d62.as29.nwbl0.wi.voyager.net) returned to #gnuenterprise. derek (derek@cpe-24-221-112-50.az.sprintbbd.net) returned to #gnuenterprise. chillywilly: i pick rms up for a week of fun on saturday and what does he do for a week in phoenix jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) got lost in the net-split. not much he will speak sunday night and probably tuesday night and we set up a luncheon for him with local users group i am trying ot get him to meet with the city, state, county CIO's and Adminstrators then we are shipping him up north aaaah to play in sedona speak 2 days and NAU working an angle eh? and visit the grand canyon neilt: RMS thoughts have much value for all, imho we only had 2 week notice he was coming so our turn outs will be small wehn we loose the "right to read" because of DMCA, SSSCA, they're going to be more valuable than ever I think his little fictional story is becoming true right before our eyes or it will if we don;t do something about it chillywilly: is there a way i can make money from his talks? oh god :) that's sounds very cold, shallow, and selfish why neilt: actually you could making money is bad i guess making money isn't the end all in a way it is uuuh, no if you dont, then you live on the streets in a cardboard box or in your office at the MIT lab if you are at mit you make more money than most it's the whole what's in it for me attitude that makes me fucking sick well RMS used to live in his office at MIT what is his family situation, they got money derek: how so you keep me in suspense - jamest this is for you neil_methods_test.py Fri Sep 21 14:44:59 2001 UTC loop count = 50 time to create object and set one field = 153.498 milliseconds. time to create object, set one field, and call method = 177.788 milliseconds. time to call a method only = 24.290 milliseconds. query found 2981 results from zip codes 20534 to 10854, in 1.470 minutes. query found 12142 results from zip codes 60651 to 24310, in 16.695 minutes. Do nothing out of selfish ambition or vain conceit, but in humility consider others better than yourselves.4 Each of you should look not only to your own interests, but also to the interests of others.5 Your attitude should be the same as that of Christ Jesus:6 Who, being in very nature God, did not consider equality with God something to be grasped,7 but made himself nothing, taking the very nature of a servant, being made in human likeness.8 And being found in appearance as a man, he humbled himself and became obedient to death-- even death on a cross!9 Therefore God exalted him to the highest place and gave him the name that is above every name,10 that at the name of Jesus every knee should bow, in heaven and on earth and under the earth,11 and every tongue confess that Jesus Christ is Lord, to the glory of God the Father. jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) joined #gnuenterprise. and what does this have to do with making money what you said just made my sick to my stomach before as all it sounded like was being selfish...GNU is about sharing ya know the first couple verses put this into perspective I don't expect you to understand anyway i understand, but you cant feed a family and take care of your elders if you dont take care of your finances family is very important and more important than sharing something with someone you dont even know no one is saying anything to the contrary we are all family in th eeyes of god....but yes my immediate family is th emost important to me neilt: what kind of search is that doing? is it loading all results then filtering in geas? dont know how it does it yet that seems slow i just checked in the python test program cool seems slot to me also bb in a few slow let me test something here informal gnuef will pull 3624 records and load into memory in approx 2 seconds from zipcode database with a zip = 7* or 6* aw crud I forgot to attach my resume dih! doh! jamest: that sounds about right for a small server running a lot of other stuff now this is old db code it's on our multiuser machine w/ 14 people running about 14% idle jamest: my code executes a single query for zip between two random numbers ah ok i can't simulate that easily until new forms takes advantage of the new Gconditional system and returns the results in a list i can do a select like all the town starting w/ M% in a zip of day 4% it returned 269 records faster than my watch ticked to next second i know it's not an apples to apples comparison but it does seem geas is a bit slow yep its apple to linux :) lol we'll get the speed up there! as i have been saying for months now and we have the data to prove it neilt: saying what? geas is too slow ah umm, I thought we all knew that already but we dont even have a todo to speed it up so we have talked about it but not done anyhting about it I don't think it is the right time to profile the code oprtimize it i think the problem is not code, but is structure because i think we will have to revise the idl if we are going to change the idl now is exactly the time before we get a lot of people using it - i also think that optimizing code for 10% improvement, now is not the right time but we need 200% improvement in queries maybe 2000% I think the objectcache sucks, but how do we know until it is properly profiled, what you are getting is execution times we need to know where it spends the most time chillywilly: exactly so we need to profile but not run around and change a bunch of code we need some analysis Action: jcater would like to note that the gnue-common has not been profiled either so we can decide what is going on uday (uday@host-64-110-96-159.interpacket.net) joined #gnuenterprise. from a geas newbie i think the IDL needs some work I think it sucks finally someone says it I have thought it sucks for a while the building of a query is amazingly difficult IMHO jamest: if you tell us what you think it needs, we'll get to work doing it but without forms working we dont know what it needs jamest: do you have an example of building one? in the examples dir gnue/geas/examples/python/neil_methods_test.py provided you have cvs from 20 minutes ago oh ya :P if not try query.py querytest.py in the same dir what is killing me in forms is the executeQuery thing Action: neilt open up the TODO file it requires an orderby for one god why does everyone have to use stinking shockwave flash oh great i cannot even visit this site death to flash! so if i don't pass it an order by the sql it creates is bogus blah that's lame it should be optional not to start another debate uh oh run! argh! just a sex er sec hmm. freudian slip Action: chillywilly runs and trip, hit his head on a rock and dies Action: neilt runs to turn Rush Limbaugh on chillywilly: you listening to rush today ok, now not to start a debate dneighbo (dneighbo@mail.libertydistribution.com) joined #gnuenterprise. nope why woudl I want to do that ? will it make me money? :P chillywilly: yes it will how? a lot of it too shea right in gnuef we do the fields via what amounts to a building a dictionary of values this is similar to what I think geas does but things like the orderby are optional and default to sane values sane as can be via the state of gnue coders mind this it canot be jamest: yes neilt: how does Rush help you make money neilt: you;re avoiding the question uday (uday@host-64-110-96-159.interpacket.net) left irc: [x]chat chillywilly: attitude neilt: that's a weak argument Action: jamest needs to think about the query system more before talking how so we should be using OQL for GEAS stuff but you can;t get an OQL spec without paying for it i'm not sure oql is the thing to do sure it is, these are objects as working with geas it's not truly OO i mean accessing it is similar to access relational data well, you don;t have to worry about it that's our job unless I'm missing something it doesn't seem to nest objects within each other again I'm speaking as someone diving into fairly fresh jamest: ahh it does well I dunno wtf oql entails only that it is a spec fopr objects but in fact we are removing that capability to make it simpler to make other things work better you can get a grammar file for parsing oql for free not much help though what sucks is th espec is not free, but yet ppl buy books on omg specs even though they don't have to so it makes no fucking sense to make ppl buy it, they'll do it anyway if it is any good maybe that's a sign? Action: chillywilly takes hs trout over to odmg.org hehe maybe jcater: you here? yam why should anyone buy a book on a aspec they can't even look at that makes no sense how do you even know if it is any good jamest: back i started to code the getSchema* for geas late last night you have schemaTypes of 'view' and 'table' defined for postgresql but I didn't see them used other than internal to the driver they are not currently used BUT in designer figured maybe it's a designer lol anyway i added 'object' to the list for the geas driver that'd be correct :) that's why it wasn't hardcoded ok basically that specifies what types of sources can be retreived from that provider just wanted to touch base gotcha you'll see the TODO there even mentioned you by name :) neilt: did you happen to see any obvious error in that DBdriver.py file Action: chillywilly thwaps neilt with a trout just because he can neilt: on the benchmarks - do have any any index's on the zipcode table? yes its indexed by zip code damn jamest: did not see anything from inspection inspectoin of? but i dont know if indexes work hold on issues an explain from pgsql er psql how is that explain select * from zipcode where zipcode between foo and bar; i don't recall the between syntax off top of head it'll dump the query optimizer analysys ntiffin=> \d location__post_code Table "location__post_code" Attribute | Type | Modifier --------------+--------------------------+---------- sys_id | character(32) | not null sys_cre_user | character varying(16) | sys_cre_time | timestamp with time zone | sys_mod_user | character varying(16) | sys_mod_time | timestamp with time zone | zip_code | character varying(5) | city | character varying(15) | state | character varying(2) | latitude | double precision | longitude | double precision | Index: location__post_code_pkey it does not indicate an index on zip_code at the bottom Index: location__post_code_pkey explain select * from location__post_code where zip_code between '10000' and '40000'; ntiffin=> explain select * from location__post_code where zip_code between '10000' and '40000'; NOTICE: QUERY PLAN: Seq Scan on location__post_code (cost=0.00..25.00 rows=10 width=104) EXPLAIN so indexs are not being used which makes sense fwiw : they are not use here either for my query return results devel=> explain select * from zipcode where city like 'M%' and zip like '6%'; NOTICE: QUERY PLAN: Seq Scan on zipcode (cost=0.00..1.03 rows=1 width=36) EXPLAIN the gcd file is module location { class post_code { char zip_code<5>; char city<15>; char state<2>; float latitude; float longitude; UNIQUE INDEX(zip_code); btw if you ned to see what your index is based upon }; \d location__post_code_pkey you don't think geas does the following do you queries a record see if it meet the criteria if matches keeps repeat looks like maybe it is that brain dead ntiffin=> \d location__post_code_pkey Index "location__post_code_pkey" Attribute | Type -----------+--------------- sys_id | character(32) unique btree (primary key) ouch ouch ouch ouch that is perfectly fine but isn't sys_id the 128bit uuid of an object yes ok i can safely say indexes didn't help the query on my end either so i didn't get an unfair advantage there but we may want to have some way of telling via the gcd fields that should be indexed we do neilt: UNIQUE INDEX(zip_code); as without indexing on fields we're going to take a huge hit but that index didn't end up being applied to the DB backend there is no other reason for it to be specified if not for the backend but do a \d on the table then a \d on the index if they are as pasted then the index was never created that is correct geas seems to ingnore the index keyword it's a minor issue for now Mussi (x@200.167.235.34) joined #gnuenterprise. as indexes are not going to solve this problem how does an index help, sorry my db skills are rusty why is my explain 1..25 and yours is 0..1.05 it has to do with the vacuuming of the db i don't recall the specifics can you do same explain at home I could and see what it reports geas won't run here :( no i mean just do this explain select * from location__post_code where zip_code between '10000' and '40000'; but put in your column names and table names neilt you may have a way of telling them in gcd but the backend isnt listening btw: i noticed this a month ago and was bad and didnt tell anyone :) prod=> explain select * from zipcode where zipcode between '10000' and '40000'; NOTICE: QUERY PLAN: Seq Scan on zipcode (cost=0.00..696.31 rows=8940 width=36) EXPLAIN note: i used the wrong db for my previous explain try that query and your db will come to a halt my responded in < 1 sec :( with first results 9063 rows vacuum analyze do that if you haven't vacuum'd in a while it'd be nice if this was a db issue and not a geas issue ok i did that try the query again how hard would it be to get your test program to query somthing like pull all entries for cities starting w/ 'M' in zipcodes starting with '6' pull all entries for cities starting w/ 'M' in zipcodes starting with '4' pull all entries for cities starting w/ 'M' in zipcodes starting with '5' as I can't do it here w/o geas and I can't do betweens in the current gnuef Mussi (x@200.167.235.34) left irc: jamest: ok doing now Action: chillywilly is away: lunch Nick change: chillywilly -> cw-lunch Nick change: jcater -> jcLunch neilt: cool does geas require orbit-python? it's in the geas.conf can I have multiple datastores in gnue? like to segragate various groups of compontants via 2-tier or geas? only if you want to use python methods jamest: neilt: thanks [localhost:geas/examples/python] ntiffin% ./neil_methods_test.py Connecting Got GEAS.Connection object neil_methods_test.py Fri Sep 21 16:48:15 2001 UTC loop count = 50 time to create object and set one field = 198.770 milliseconds. This query test requires the zip code database be loaded into GEAS. Email neilt@gnue.org if you do not have a copy of it. time to create object, set one field, and call method = 231.879 milliseconds. time to call a method only = 33.109 milliseconds. query found 53 results between zip codes M - 6, in 10.477 seconds. query found 34 results between zip codes M - 4, in 6.828 seconds. query found 37 results between zip codes M - 5, in 9.824 seconds. time to query zip file = 0.221 seconds. number of items returned = 124 - i did not change the text but the logic is as you ask for much, much better er is this w/ the zipcode data file i had? as the result count is low ra3vat (ds@195.239.66.47) joined #gnuenterprise. yes hello all neilt: no sorting of the data either? query found 53 results between zip codes M - 6, in 5.940 seconds. query found 34 results between zip codes M - 4, in 5.918 seconds. query found 37 results between zip codes M - 5, in 5.256 seconds. but it isn't sorting the result set is it? i am playing with the orderby clause ok, i don't have a timer the second set is with orderby = sys_id which has an index but gnuef pulls all records (creating a record object for each, and displays in < 2sec) and the new code bombs on that query :) joy hrmmm Action: cw-lunch is back (gone 00:14:57) Nick change: cw-lunch -> chillywilly why aren't the various classes in the people module used for person? jamest: cool oohh nm. when i issue the query in psql i get (269 rows) returned geas only gives 53 you're smoking crack if you expect middleware to be as fasta regular database but ti shouldn't suck as much as geas neilt that's what I'm seeing here chillywilly they arent comparing psql to middleware they are comparing gnue-common data access to middleware er geas yeah i guess only getting 53 out of 269 rows is ok too i get 271 entries they only went to psql level to make sure they were comparing as close to apple as apples as possible but I've added to the db some :) what is your query dneighbo: which is praticallt database vs. middleware pratically i.e. indexes same, structure same, vaccum done recently chillywilly yes and no wtf is a vaccum? chillywilly: its a hoover city starts with M, zip with 6 < 1 sec inregard to a database btw: we might be crazy to think they could become 'equal' Action: chillywilly thwaps neilt :) that's till it's displayed in the ui system but 2 seconds vs 6 seconds is 3 TIMES as slow so there is some overhead in the internal building of the UI and entry objects lunch I am gonna boogie or if its 1 to 5 thats 5 times as my woman is driving me up a wall while i expect geas to be slower than native db my woman jeeezze thats an attitude by magintude of 5 is not ok (imho) and i think that neilt would agree :) or he wouldnt be writing test programs to prove it :) :) so this isnt a pick on geas thing its a lets fix geas thing this is the query i used select * from location__post_code where city like 'M%' and zip_code like '6%'; for the log neilt: but attitude is what makes you money not that attitude with a woman s/attitude/bad attitude/ she's the one with the attitude today i can see ehy why therefore I am going home to retain my sanity later cw neilt: wtf do you know? are you here? i know women that attitude wont get you far unless you want to be by yourself :) heh e have been to gether for 6 years thats cool well then, she's just cranky today you beat the average and I cannot work when she's yelling at me :P you can't ever make a woman completely happy they always find something to complain about lets not go there doesn't mean I don't love her, but I need to print something and there's no printer here bye chillywilly (danielb@d62.as29.nwbl0.wi.voyager.net) left irc: Nick change: jcLunch -> jcater ra3vat (ds@195.239.66.47) left irc: Ping timeout for ra3vat[195.239.66.47] cranky women? I don't buy that. jcater / jamest you just got some lovely rant mail from me im curious as to your opinion I'll have to remove my NO DEREK RANTING filter dont worry not a rant about gnue have been causing trouble on gnucash list chillywilly (danielb@d20.as15.nwbl0.wi.voyager.net) joined #gnuenterprise. actually not causing trouble having good discussion on things that are relevant to us when are you not causeing trouble dneighbo? :P i'll have to remove my .procmail anything from derek filter Mr_You (rouzer@mankind.boredom.org) joined #gnuenterprise. so did anyone see all that I was babbling about in reards to GComm the other day? yesterday chillywilly: gcomm? hi Mr_You Action: jcater missed that rtfl :P j/k which logs? umm, you know th ecomminucations abstraction thingy i saw some of it Action: jcater will look at it as I am antsy for gcomm hi can we agree to put all note for this in a gnue-common/doc text file for now (gcomm) even if it doesn't end up in there basically build a libodl and then use remote proxy pattern so that you use eith GObject or some C represetation of the object and python objects like they are normal objects and Gcomm treats them like remote objects and takes care of using the proper mechanism postgres' authentication thing is always a struggle. chillywilly: you've got to take it a step further anyone know th eremote proxy pattern? rpc isn't OO that I knew neither is EDI IIRC CORBA and SOAP are OO RPC chillywilly: you just gave me a massive headache Action: jcater goes to get ibuprofin oops CORBA and SOAP are yes well we are gonna let them use whatever right? but there are more than 2 methods of accessing remote system some of which are not OO but they just configure GComm and difne the object using odl (gcd) and Gcomm takes care of doing the work of remote calls if needed jamest: how are you guys gonna use it? maybe that's what I need to know i wanted GComm to be used in the gnuef geas dirver to access geas you know how idl works right at least for ORBit is processes the idls I also wanted it in geas to provide the interface to clients well here we prcess odl and setup the correct remote mechnism whatever they want i wanted GComm to be used in the report server to export interfaces to clients... the clients would also use GComm to access the report server which parallels jamest's needs so it must IMHO have the following features yea, the interface in forms then would just use regular old python objects and not have to muck with anything special complete client/server functionality it is transparent and done by the Gcomm complete abstration away from any communication protocol yep no importing orbs or anything just using the pything bindings to Gcomm it must come in blue or if ti si in python just the python objects jamest: of course :P I'm not overly worried about a perfect python representation of objects as I think language specific interfaces to GComm could behave differently it should be generic enough that even something like bonobo could use it and have support for more than just CORBA, but I'm not too worried about that right now or DotGNU ;) ergh I might see what the DotGNU folks think nickr: eh? I can't figure out why postgres won't let me login as a new user I created. I think I'm just dumb. oh ra3vat (ds@195.239.66.37) joined #gnuenterprise. in fact I should spam the dotgnu listm when I have time anywho is that something useful or am I smoking crack Ia m gonna send an email to gnue-discuss and dotgnu developers list and see what ppl say quickly reading log chilly we can not put implementation specif things in our RPC example on of the first RPC's i want is Sockets I don't want to do that masta this is a way to get any mechnism you would like even your own home grown tcp/ip yes it's gonnabe a big project I think but jamests' point was basically not everything is OOp then why are we using python> ? im not saying im ANTI OOP OO is good and you can even do it in C doesn't matter though nevermind we can expose a fnctional interface i was simply commenting on jamest comment i will read the log later so i can argue specifics well you can make some functions to do it too just another interface to the library or Gcomm, whatever jamest / jcater no comment on my mail s/or/er Action: dneighbo is suprised i figured you may have agreed at least :) dneighbo: I'd agree with it not much to say... it was well written what mail? jcater: you think it would be worth making it a short essay for linux today? yes what did you write? i.e. a calling gnome out on the carpt to get thier shit together challenge of sorts? why didn't I geta copy? it needs to be said over and over and over [13:16] Last message repeated 1 time(s). :'( masta can I have a copy? pleeeeeease does chillywilly@gnue.org work? I dunno well what does work? danielb@freedevelopers.net does dres (dres@4.18.171.42) left irc: Ping timeout for dres[4.18.171.42] I think chillywilly@gnue.org works too it should point to baumannd@msoe.edu and over [13:18] Last message repeated 1 time(s). I think they got their net connection back up jcater :) ok i emailed it they were having problems chillywilly we need to get a dan@gnue.org that other email is a bitch to type sure go ahead Action: dneighbo thinks perhaps gnulinux@gnue.org might be better suited ;) hehe Nick change: chillywilly -> gnulinux oo a message gnl@gnue.org yea, that's even better less chars although I could get on a rant about that whole GNL thing based on derek's email and jaemst can be ngl@gnue.rog but I won't Nick change: gnulinux -> gnl 'naughty goat love; heheh Nick change: gnl -> ngl wooohoo lol biggoat@gnue.org, mastagoat@gnue.org, killergoat@gnue.org roflmao dude did you see that movie with the duece bigalow guy? gos what's that called i feel sorry for first bill here the animal wait we had a bill rofl billy goat he was feeeling on that goat (bill hamilton) and it kicked him you know what I am talking about? yeah I saw that movie in the theater that's GNUe lovin' right there ;) Nick change: jcater -> chillywilly Action: chillywilly slaps ngl with a trout Nick change: chillywilly -> jcater ouch Nick change: ngl -> chillywilly no one has commented on my www.gnome.org comment Both of these packages are complete clusterfscks wrt to release schedules. lol about the second sentence on their website clusterfscks! rock on! GNOME is included in pretty much every BSD and GNU/Linux distribution and works on many other Unix systems. that one? or this one: GNOME is part of the GNU project, and is free software (some times referred to as open source software.) that one they are a GNU project I think you're looking too far into it jamest its just a reference/link The FSF asks us not to use the term open source IIRC maybe I am dude yea, I agree they should mention it but they did say sometimes I hope theres a good reason for it, because that sounds completely rediculous hold on a sec Mr_You http://www.gnu.org/philosophy/free-software-for-freedom.html ;P funny how freedom becomes so restrictive ok I see there's only one restriction (all derivative works and chages go back to the comminty) in the GPL compared to the many that proprietary licenses put on you...prop. license == copyright + restrictions GPL == copyright - restrictions, imho of course jamest: iirc it has been there for at least 6 months HOWEVER the LINK is new no.. the philosophy has its restrictions what I meant and iirc its ok to say opensource (if in the right context) on a gnu site but not to link to things like that so to say no, it doesn't if you look at the freedoms on the GNU page and the DFSG and the OSI page, they all say the same thing well I gotta start workin sometime gnu enterprise is free software (sometimes confused with opnesource software) is acceptable saying i am sorry but a movement that has an enemy, is exploting violence, there is not enemy in software gnu enteprise is free sotware (also called opensource software) is bad bad bad while we are discussing unrealistic idealism, I'd like for us to remove a line from our license headings jcater i will ask RMS on this when he is here reinhard (rm@62.47.44.46) joined #gnuenterprise. as you know i agree with you on it ok what line? neilt: ha, you don't think M$ wants to enslave the world, that everyone will have to go to M$ and only M$ for software...do you see anyone toting guns no, we create and that is how we combat things the line that says covered by GPL 2.0 or greater as what if the GPL 2.1 is shit and we dont like it microsoft is not my enemy or the 3.0 or the 4.0 actually it says, GPL 2.0 or, at your option, any later version all our code is then releasable under that proprietary software ism I just used them as an example I am not comfortable with that instead we just say GPL2.0 if 2.1 comes out or 3.0 and we like it dneighbo: you don't trust the FSF? then we just change our files to say GPL 3.0 neighbo: you can use any version.. if that makes sense proprietary software is not my enemy Mr_You no you can use 2.0 or greater just skip 2.1 for gnu er gnue some older gnome stuff and emacs stuff says 1.0 or greater you choose to distribute which license i think the point is rather than 'retract' (which we cant do) As long as we leave the "Stuff as much money as you can in an envelope and mail to GNUe Hackers Need Porches Too Fund Po Box...." lets fix it you don;t have to say or any later version, in fact Linus doe sht esame thing with the kernel so we say you can ONLY release under the 2.0 then if 3.0 comes out we change that line and rerelease if we dont liek teh 3.0 we do nothing dneighbo: I see that as not trusting the FSF this protects us as developers btw: i dont trust FSF completely as I don't think Linus is a good person to model after i dont think they woudl make things worse as far as give my code to prop companies under the GPL dneighbo: who can you trust then? but i fear they could become MORE radical what matters what version of the license/philosophy we distribute.. versus latest FSF version chillywilly: yourself adn say something ok you can only view my GPL application with a GPL browser IM NOT OK WITH THAT jcater: well then I feel sorry for you ppl as I trust the lord more than myself :P and i assume the rest of the core team would agree with that reinhard: query found 13807 results between zip codes 57820 - 48127, in 31.843 minutes. Mr_You: it was giving the end user the option of what version they wanted to use so in a nutshell its always good to be prudent i think changing the line to specifically state which GPL does no harm jcater: yeah I understand.. rip it out neilt: ugh it is simply a 'prudent' move I don't like it it only causes a minor amount of work for us when a new GPL is released chillywilly why? cause I trust the FSF because its not 'free' enough?" i mean its still the friggin GPL :) that's a half-assed Linus Opne Source dickhead move ToyMan (stuq@c5300-1-ip165.albany.thebiz.net) left irc: Client Exiting imho chillywilly: I saw the FSF in action at LWE, and I don't trust them w/my work... but that's a different story all-together chillywilly: why does it matter what version of the license/philosophy that we distribute.. at the heart of it, its still GPL neilt: can you tell me what platforms this is on sigh oh boy cause what if they fix some legal loophole then you are changing it again if they fix a loophole chillywilly: its a precautionary decision mac os x we change teh license to new license re release that simple I'm running gnuef on a Sparc 167MHz multiuser machine accessing a DB on a p2-300 over a 100Mbit line if there was a loophole in 2.0 and 3.0 fixed it when you distribute saying 2.0 or greater jamest: cool teh bastards could still about the 2.0 loophole as it doesnt say whatever the current gpl is dneighbo: uuuh, that's not what I thought you meant so your argument is dead there I thought you wanted to get rid of any later version part very least, there's no reason to have a wait and see time period when a new license is released i do want to get rid of any later version part and HARDCODE the version so today it woudl say liscensed under the GNU GPL 2.0 if the 3.0 came out and it looked good we sed the files to say 3.0 instead of 2.0 Action: Mr_You doesn't have blind faith, completely opposite ;-) and rerelease if it looked like crap we leave it at GPL 2.0 I don't think it would be a nice PR move i think RMS put that clause in only to make it EASY for maintainers as then they dont have to worry about it they just release probably iirc its not part of teh 'license' so we HAVE THE ABILITY AND AUTHORITY TO CHANGE IT chilly: I don't see any problem, unless GNU REQUIRES you to use the latest GPL does gaes take advantage of SMP? its not like we are changing the LICENSE itself you think 3.0 is gonna be that mental that they are gonna say you can only use this with Free software and only Free software, how the hell can they say it is free software then you are going way too farand I would have to tell RMS to got o hell if he did that Mr_You GNU can not REQUIRE you to use the GPL the problem is that others could redistribute under a worse license if GPL license got worse chillywilly no i dont ok, well I'm for being picky nickr: nah but you can never be sure people said gnome and miguel were fsf hardcores look what ximian has done to thier 'ideals' rms isnt young you know they aren't though chilly: for all we know it will be a sane license and this conversation won't matter if he were to pass away who is to say that the new fsf leadership would take the same stands he does? well I don't care as long as someone doesn't try and twist it around and say we don't trust the FSF its a foundation, what if he was voted out? etc etc etc this is a practical move I don't blindly trust them, but I have to believe in something we wont do anything till i talk to rms i want to find out why he put in there what his feeling is on removing it then we can discuss as a core group and make a decision at this point i lean towards hardcoding it for sake of being prudent i think jcater agrees it may be just for convienence i can speak for no one else like you said I will agree jcater doesn't like the FSF anyway chillywilly: I have issues w/them didn't say I didn't like them exactly I have issues w/dneighno didn't say I didn't like them chillywilly: let me put it this way err, him how many GNU projects are there? a LOT how many do you think got _any_ mention to anyone at LWE? well doesn't mena someone can't do aomething about that and ask them to do less philosphical bantering and more software promoting ;) best of luck they are nice enough but really don't market anything of theirs they have no money to do that do they? it wouldn't have cost the people on the floor anything to talk to people findout their needs and suggest GNU systems to fill those need Action: jcater guesses that GNU's Not Linux stand wasn't cheap :) instead they very abbrasively promoted GNU's Not Linux does thos have anything to do with license clauses? to the exclusion of anything else no, it has to do with why some of us have issues jcater doesn't have to dislike GNU http://www.gnu.org/licenses/gpl-faq.html#VersionTwoOrLater oh, nevermind this has to do with "issues" he can dislike their methods and I wouldn't accuse him of not likeing the FSF anyway please stop sure! no jcater has donated more code than most to the FSF so to say he doesnt like them or is anti FSF is LUDICROUS argh! no breaks! jamest (jamest@hobbes.math.ksu.edu) left #gnuenterprise. dneighbo: no it isn't, ppl have a various reasons for being here jamest (jamest@hobbes.math.ksu.edu) joined #gnuenterprise. #gnuenterprise: mode change '+o jamest' by ChanServ!s@ChanServ ok so officially > This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. it looks the 'any later version' is at our option BUT they want GNU Projects to use it so they can easily create distrubtion CD's and easily do a mass license change i will definitely discuss this with RMS it is neither in the license or the preamble dneighbo: like I said earlier, I have no problems w/the intentions... but that clause can be taken many ways jcater kudos for catching this and bringing up the issue I don't understand what the license has to do with easily creating distro CDs jcater i think thier point is good on the GPL FAQ i posted here BUT ther is no issue there imho it is all under assumption that GPL will always get better (in sense of freedom) and not worse which is probably true, but that leaves you software out in the wind Mr_You because if every software product on teh CD was hard coded to 2.0 and they wanted to relsae gpl3.0 they would have to get the permission of every software maintainer/copyright holder and before they could chagne thier release CD each software has its own license, what does that have to do with GPL v3? now the ironic thing jcater is we require assignment :) so technically FSF is a copyright holder and can change license thus nullifying to a degree this argument :) btw danielb@obfuscation:~/src/gsim/src$ cat *.cpp *.h | wc -l 2753 Mr_You never mind its a philosphy thing all lines GPL'd if you dont natively get what richard is wanting in it and soon to be assigned FSF you probably wont :) Action: Mr_You doesn't understand why it matters what VERSION of the PHILOSOPHY.. with any later version clause dneighbo: sounds wacked after I rewrite a few things really this is like gnome having these arguments :P i will talk to richard get his view and then we will discuss and decide all this sounds a bit anal to me oh wait that wasn't recursive it might be anal but i find it funny that jason brought the point up because he didnt want to some day find his hard work in a microsoft product that wasnt free yet he is the one that is not 'pro free software' Action: dneighbo wonders about that even if jcater does have issues with me ;) yeah, I don't believe in commiting the fate of my software based on future circumstance, but others may not care so much damnit how can I make it go recursive and count all the lines of code? reinhard: you here? Action: chillywilly bangs head on wall chillywilly: what are you trying to do? count all linkes in .h and .cpp files recursively in the srcdir src dir use find I was thinking of that but how do you make it doa regex? find ./ -name "*.h" -o -name "*.cpp" -exec (lc) {} \; ah is it lc ? wc -l wc -l {} \; -exec wc -l {} \; I use find + grep all the time, very useful umm how can I make it total it? tho I'll tell you my -o syntax might be wrong it worked just no total 80 ./controller.cpp 297 ./gsim.cpp 62 ./priorityqueue.cpp 98 ./queue.cpp 156 ./resource.cpp 156 ./scheduler.cpp 46 ./simentity.cpp 262 ./simevent.cpp 52 ./simulation.cpp 61 ./examples/mm1/customer.cpp 34 ./examples/mm1/mm1.cpp 54 ./examples/mm1/mm1simulation.cpp 41 ./gslcpp/gsldistribution.cpp 226 ./gslcpp/gslhistogram.cpp 55 ./gslcpp/gslhistogrampdf.cpp 59 ./gslcpp/gsllognormal.cpp 51 ./gslcpp/gslnormal.cpp two ways I guess stop 108 ./gslcpp/gslrandom.cpp 52 ./gslcpp/gsluniform.cpp 60 ./tests/distributions/customer.cpp 35 ./tests/distributions/main.cpp 60 ./tests/priorityqueue/customer.cpp 53 ./tests/priorityqueue/main.cpp floood! um, It is stopped chillywilly: just copy it into excel :) roflmao did someone say something? nah anyway actually.. I think the only way might be to use bc I wish cat would do it recursively I could wirte a script grep -c " " `find . -name '*.cpp'` or some wacky hack like that jcater wow you really ahve grep on things ;) ra3vat: yes i am here dneighbo: I hadn't thought about the copyright assignment... you're kinda scaring me here thatonly does each individual file i tried to read the logs but they gave me no chance to catch up ;) cat `find . -name '*.cpp'`|grep -c " " reinhard: PgUp ;) chillywilly: i still have no chance to catch up when you type faster than i read ;) reinhard: ok, i did .so as said at http://mail.python.org/pipermail/python-list/2001-May/043271.html ra3vat: and did it work? let me guess: no? reinhard: compiling goes further and then chillywilly: here it is: cat `find . -name '*.cpp'`|awk '{print length($1)}'|grep -c "1" there ya go /usr/lib/libpython2.0.so: undefined reference to `dbopen' err, cat `find . -name '*.cpp'`|awk '{print length($1)>=1}'|grep -c "1" /usr/lib/libpython2.0.so: undefined reference to `openpty' /usr/lib/libpython2.0.so: undefined reference to `forkpty' Action: jcater loves grep/awk combos throw in sort and I get all excited it didn't liek ti awk: line 1: syntax error at or near = 0 ra3vat: i still believe that you don't need .so and that .a should do it reinhard: and what you suggested earlier does not work too especially because i don't have .so either reinhard: let me try again hmm you paste that as a complete line? works for me but I am on slackware :) I got 3.1.0 gawk version it doesn't liek that >=1 thingy GNU Awk 3.0.6 bah I win not if they broke relational tests well,.... well, that should give you ideas for a scripts anywho cat `find . -name '*.cpp' -o -name '*.h'`|awk '{print length($1)>=1}'|grep -c "1" that should work? yes worked on my machine at least chillywilly: sometimes the most primitive things are the best try wc *.cpp */*.cpp */*/*.cpp */*/*/*.cpp ;) hehe... well, whatever floats your boat :) :P Action: jcater gets off w/complex cat/find/awk/greps figuratively, not literally 2158 total jcater: what about sed? i like that very much too ;) yip speaking of floating boats Action: chillywilly used to be able to sring awk together with grep and do soem funky things too send is th hairest though geas has been initializing the security system here for at least 3.5 hours er sed any ideas? reinhard: if i do LDFLAGS=-L/usr/lib/python2.0/config ./configure --with-postgresql --enable-debug --enable-methods=python it stops with /usr/lib/python2.0/config/libpython2.0.a(posixmodule.o): In function `posix_tmpnam': also neilt did you happen to look at that DBdriver.py file and see if I was screwing something up that'd segfault geas? jamest: i looked at did not see anything but we need to get reinhard to check it /usr/src/RPM/BUILD/Python-2.0/Modules/./posixmodule.c:4106: the use of `tmpnam_r' is dangerous, better use `mkstemp' arrrgh please everybody pull a number ;) Action: jcater pulls #231 Action: neilt gets his number oh boy its 9456217 haha Action: neilt looks at reinhard as he starts to work on number 5 4337 total woah my little lib is getting big all GPL at 2.0 or later ra3vat: this is a error message about python code ?!? it does stop at that message? this is my reason for non gnue code flowing form me 4000+ lines for this other thingy reinhard: few other messages similar to the above crap oh sorry i nearly forgot that python dependency is a non issue ;) anybody else here uses python 2.0? i use 1.something i have it and it works for me neilt uses 2.1 and it works for him dneighbo: you think miguel is gonna be a "real" leader? jamest: you got geas to compile? yip well i guess so actually running too but this is where she stops query will be select a.attname, t.typname from pg_class c, pg_attribute a, pg_type t where c.relname = 'timetest' and a.attnum > 0 and a.attrelid = c.oid and a.atttypid = t.oid Initialising ORB Creating locator objects Creating ConnectionFactory instance for new connections you have libpython2.0.so or libpython2.0.a? [function call] [geas-skeleton.c/2291] [impl_GEAS_ConnectionFactory__create] Initialising security system who me or ra3vat? jamest: you ./lib/python2.0/config/libpython2.0.a and how did you make geas link with that? did you set LDFLAGS? nope ./configure --enable-debug --with-postgresql Action: reinhard will never understand how ld finds it's libraries chillywilly no i dont think he will be a real leader but i have been suprised in the past i never figured dubba would do as well as he has in the current situation jamest: please edit the file security.c below the line dubba? //* load all users * who's dubba? george w. bush /* load all users */ bofh versino of GEAS jamest: below this line is a call of GEAS_Connection... /* explode all users */ jamest: please insert a message (....) with a text of your choice jamest: just before and after this call i guess this is the call that hangs not 100% sure but i guess it's the first orbit call in the whole game -- we changed the db schema jamest: you need to remove all tables and let geas rebuild everyhting again i bet it will work if you do that neilt: i bet not what will we bet about? ;) why, what do you know that i dont :) if the table schema wasn't ok it would not simply hang imho ra3vat: i am still thinking about your problem i strongly believe we should try with .a as it works for the rest of us reinhard: that certainly sounds reasonable neilt: ok, then i'm wrong probably reinhard: just remember this is GEAS you are working on ;) neilt: yeah exactly um it didn't recreate the tables the new geas upgrade model reinhard: jamest also reported the not null problem She drop--he drop--a--we drop I drop--you drop--a--they drop Be drop--be drop--a--lu--she drop, I hope GEAS will understand She drop--he drop--a--we drop I drop--you drop--a--they drop Be drop--be drop--a--lu--she drop, Oo--oo--she--do--she drop--she drop where multiple not nulls caused geas to not work yeah we still need to fix that but we only can fix it with an idl change and that gcd index isn't being honored i thought you did a quick and dirty and removed not null from all internal classes which is something you agreed on iirc from todays logs and I can segfault geas a few different ways neilt: yes i did sorry bbl jamest: rock on Nick change: reinhard -> rm-away jamest did you read the geas.log and debug.log in the geas src directory its geas.log and geas.debug.log what is highest debug level let me play with geas tonight this may sound silly. How do I create a 'shadowpw' file? i will send my 'list' ;) Action: dneighbo should know that jamest and jcater are thankful geas is the victim tonight and not forms stuff jamest: -d10 actually, i don't wish that on anyone its a good thing we change the geas debug procedure that is we dont accept bug reports any more only patches no patch, no bug jamest -d10 does that run jamest at debugging level 10? if so, lindsay might like to know that might have an interview next week.. and I just applied for unemployment well, good luck Mr_You the job is right above where I worked for Lucent.. (re the interview, not the unemployment application) even some people I worked with at Lucent [2001-09-21 14:47:55 00000001 e: message u:] Debugging output logged to 'geas.debug.log' [2001-09-21 14:47:55 00000002 e: message u:] Loading initial list of reserved words [2001-09-21 14:47:55 00000003 e: message u:] Load business class definitions [2001-09-21 14:47:56 00000001 e: message u:] Initialising business class method handling. [2001-09-21 14:47:56 00000002 e: message u:] Initialising object store. [2001-09-21 14:47:56 00000003 e: message u:] Initialising object cache. [2001-09-21 14:47:56 00000004 e: message u:] Validating/updating database table definitions. this guy got like a 15% raise [2001-09-21 14:47:56 00000005 e: message u:] Initialising ORB [2001-09-21 14:47:57 00000001 e: message u:] Creating locator objects when he went there [2001-09-21 14:47:58 00000001 e: message u:] Creating ConnectionFactory instance for new connections [2001-09-21 14:47:58 00000002 e: message u:] Initialising security system houston i htink we have a problem no tables created in the geas db Action: neilt tries to open geas.debug.log gnue=> \d No relations found. GEAS debugging log ------------------ 02: Created 4096 oids 03: [function call] [geas-server.c/1668] [reserve_sql_words] [14:53] Last message repeated 1 time(s). 03: [function call] [objectcache.c/278] [create_object_cache] 10: writing header to change file 10: getting db handle 10: updating the tables 10: pg - updating the tables 10: query will be select a.attname, t.typname from pg_class c, pg_attribute a, pg_type t where c.relname = 'geas__listholder' and a.attnum > 0 and a.attrelid = c.oid and a.atttypid = t.oid 10: query will be select a.attname, t.typname from pg_class c, pg_attribute a, pg_type t where c.relname = 'geas__listitem' and a.attnum > 0 and a.attrelid = c.oid and a.atttypid = t.oid 10: query will be select a.attname, t.typname from pg_class c, pg_attribute a, pg_type t where c.relname = 'geas__user' and a.attnum > 0 and a.attrelid = c.oid and a.atttypid = t.oid 10: query will be select a.attname, t.typname from pg_class c, pg_attribute a, pg_type t where c.relname = 'geas__transaction' and a.attnum > 0 and a.attrelid = c.oid and a.atttypid = t.oid 10: query will be select a.attname, t.typname from pg_class c, pg_attribute a, pg_type t where c.relname = 'geas__searchcriteria' and a.attnum > 0 and a.attrelid = c.oid and a.atttypid = t.oid -rw-r--r-- 1 ntiffin unknown 116219882 Sep 21 14:52 geas.debug.log -rw-r--r-- 1 ntiffin unknown 135218 Sep 10 08:28 geas.h -rw-r--r-- 1 ntiffin unknown 2843638 Sep 21 12:42 geas.log 03: [function call] [geas-skeleton.c/2291] [impl_GEAS_ConnectionFactory__create] i need to rotate logs lol thats not the funny part i just tried to open the debug log in a text editor neilt hopefully not emacs emacs != liking big files does the word thrashing mean anything Nick change: rm-away -> reinhard here for some seconds any ideas? go jamest jamest: no tables created in the geas db neilt tries to open geas.debug.log jamest: gnue=> \d jamest: No relations found. what i would do oh geas doesn't create the tables? no or he is not connecting to the right database and geas db is set to gnue db in the geas.conf and user active true type PostgreSQL dbname gnue username jamest hostname owl port 5432 connections 1 I have a entry in geas/src/shadowpw geas has only been tested with localhost arrgh i bet it does not work with a real host test_postgresql: likeiwouldtellyou :) Action: jamest blinks Action: jamest thinks i only have remote db server at home and it works there i think cool then it has been tested maybe i don't recall my geas.conf settings but i don't intentionally run postgresql on my workstation at home if i only knew how geas works actually i need more time to look at this jamest: you will be here in 30 min? bb in 30 min Nick change: reinhard -> rm-away danielb (danielb@d20.as15.nwbl0.wi.voyager.net) joined #gnuenterprise. wil be here danielb (danielb@d20.as15.nwbl0.wi.voyager.net) left irc: danielb that was cool telnetting on port 6667 :P rfc 1459 is cool gnusnotlinux (a@d20.as15.nwbl0.wi.voyager.net) joined #gnuenterprise. woohoo! isn't one chillywilly enough? raw irc is how real men do it \msg someone dcc him while he's telnetted :) duh, it won't work because I don't know the command to accept it Action: jcater will stick to gnue clients, then :) :P eh? gnue clients? we don't have an IRC client Sure... I call it GNUe QuIRC :p ACTION slaps jcater with a trout bah almost worjed Action: chillywilly kicks gnusnotlinux action test b;lah maybe I shoudl read the rfc kicks chillywilly darn Nick change: rm-away -> reinhard Nick change: gnusnotlinux -> gnugod ok now the adults are gonna work serious stuff nope all children please leave the room reinhard: if we can get this to work i can send geas crach info j/k Action: chillywilly strangles reinhard as geas doesn't seem to like poor gnuef much jamest: first we have to make geas work :q oh worry wrong window um what did :q mean in emacs again? ;) "chastize old vi user" i think jamest: in geas.conf C-x C-c databasechangefile /dev/null # don't care about it just now :) can you give a real filename there please? gnugod (a@d20.as15.nwbl0.wi.voyager.net) left irc: gnugod this file should contain then all database changes that are done by geas you ppl commit too much so your saying we work too hard? ya and monkeys fly out of my butt actaully I am dying to play with python I should redo gsim in python it has threads right? right k jsut kidding I dun feel like redoing 4000 lines in python it would be like 20, i hear doubt it ahhh no, thats perl sam amount of classes jamest: you still here? back but it would debug itself in python it'd be about 23.5 lines in python reinhard: another error Initialising business class method handling. 'import site' failed; use -v for traceback error: [methods_python.c/491] [load_python_binding] Could not load ORBit-Python module at: error: [methods_python.c/492] [load_python_binding] /usr/lib/python1.5/site-packages/CORBAmodule.so ra3vat: hey does that mean you got geas to compile? i added a new config option to gnue.conf /usr/lib/python1.5/site-packages/CORBAmodule.so hardcoded in methods_python.c to fix that ra3vat: you need orbit-python ah it only looks harcoded lol wtf? added that line and it created the tables ra3vat: you need to have this in gnue.conf # Where to find orbit module # defaults to /usr/lib/python1.5/site-packages/CORBAmodule.so if not specified ORBit-python /usr/lib/python1.5/site-packages/CORBAmodule.so of course if your location is different then you change the location neilt: thanks jamest: huh? the line in geas.conf? yes that was only change I made and it spit out table creation info to stdout and in the dbchangelog and I see the tables in the gnue database but [2001-09-21 15:24:26 00000008 e: message u:] Initialising ORB [2001-09-21 15:24:28 00000001 e: message u:] Creating locator objects [2001-09-21 15:24:29 00000001 e: message u:] Creating ConnectionFactory instance for new connections [2001-09-21 15:24:29 00000002 e: message u:] Initialising security system and there I sit these are the pigstys of augias i guess ? jamest: did you add those message lines where i said? not yet pigstys of augias are from the hercules story how can I make X use 2 different mouse protocols? say when my wireless jobby plugged in I want it to use ImPS/2 jamest: then we would know if orbit is a problem on solaris or not and when is it not plugged in I want it to use the touchpad PS/2 anyone know that? its called a switch er? you can buy at any local computer store :) switch for what? how's that gonna make x recognize things? did you want to switch between two mice you plug both into switch and then change back and forth jamest: geas reads an internal table to get the list of known geas users jamest: and guess how it does it phases of the moon jamest: it calls connection.loadAll through the corba interface brail on the back of a sponge lol so geas goes through corba itself why? i guess for the same reason as it uses 37 sql statements to do a select you don't know the pigstys of augias? cause its not fsck'ing complicated enough if it doesn't :) they were full of dung hercules was forced to clean them up and whenever he was finished on one end it was already filled up again on the other end this is how i start to feel w/ geas good thing we never complete one end neilt: lol chillywilly_ (danielb@d20.as15.nwbl0.wi.voyager.net) joined #gnuenterprise. neilt: seriously neilt: i beg you on my knees jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) got netsplit. Mr_You (rouzer@mankind.boredom.org) got netsplit. chillywilly (danielb@d20.as15.nwbl0.wi.voyager.net) got netsplit. ra3vat (ds@195.239.66.37) got netsplit. Action: neilt watches his back please look at my code in src/classdef and give me your comments sigh ok geas will not compile today Action: jamest bangs head on pile of crap on desk jamest: what's your problem? no more free deskspace to use arrrgh geas made me it's bitch you need more detail ? reinhard: is there a secret to compiling it secret switch or something neilt: no ra3vat (ds@195.239.66.37) returned to #gnuenterprise. Action: dneighbo dcc's jamest his geas' bitch tshirt neilt: dozens of switches ;) I'm rebuilding from clean download ./configure --enable-debug --with-postgresql jamest: your on thin ice here we have never done a build from clean download Nick change: chillywilly_ -> chillywilly Possible future nick collision: chillywilly lol you need the method switch jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) returned to #gnuenterprise. i do? yes, because it is gay and doesn't use plugins (yet) --enable-methods=python config's w/o errors jamest: yes it will build with c method support instead of python how would you guys break geas down into components? that's why it was compiling for you jamest: chain saw jamest: with a really big hammer its made of foo, foo, and foo seriously ok, not seriously make a methods sever with that lib that lets you do mutli-lang striggers/methods jamest: please use --enable-methods=python er, triggers jamest: then try to compile the one that doesn't exist reinhard: I'm trying that jamest: and then please tell me again that python dependencies are not an issue ;) Mr_You (rouzer@mankind.boredom.org) got lost in the net-split. i think cause it doesn't have a plug in methods system :) (neither does gnuef :) jamest: re: components of geas i would say 1. class definition handling Action: chillywilly picks up his guitar 2. database backend interface 3. method handling you guys want to know about compile warning? warnings? 4. interface to the front end (CORBA and friends/warnings) jamest: yes as long as they are in our code ;) reinhard: btw jamest: 3 or 4 warnings are in yacc/lex generated code configuration.c: In function `alloc_config_item': configuration.c:183: warning: subscript has type `char' configuration.c: In function `read_shadow_password': configuration.c:295: warning: subscript has type `char' configuration.c:313: warning: subscript has type `char' configuration.c:315: warning: subscript has type `char' configuration.c:318: warning: subscript has type `char' reinhard: derek is going to work on geas tonight configuration.c:320: warning: subscript has type `char' configuration.c: In function `read_line': configuration.c:413: warning: subscript has type `char' configuration.c:417: warning: subscript has type `char' configuration.c:432: warning: subscript has type `char' reinhard: i made a new rule configuration.c: In function `read_main_configuration': configuration.c:493: warning: subscript has type `char' geas-server.c: In function `sigsegv_handler': geas-server.c:147: warning: function declaration isn't a prototype geas-server.c: In function `main': geas-server.c:405: warning: function declaration isn't a prototype geas-server.c: In function `readOptions': geas-server.c:1088: warning: subscript has type `char' geas-server.c: In function `reserve_sql_words': reinhard: we dont accept bug reports without patch files geas-server.c:1691: warning: subscript has type `char' geas-server.c:1694: warning: subscript has type `char' geas-server.c: In function `loadTableAliases': geas-server.c:1871: warning: subscript has type `char' neilt: good geas-server.c:1879: warning: subscript has type `char' geas-server.c:1885: warning: subscript has type `char' geas-server.c:1903: warning: subscript has type `char' geas-server.c:1923: warning: subscript has type `char' geas-server.c:1933: warning: subscript has type `char' geas-server.c:1939: warning: subscript has type `char' geas-server.c:1950: warning: subscript has type `char' geas-server.c:1956: warning: subscript has type `char' geas-server.c:1978: warning: subscript has type `char' geas-server.c:1986: warning: subscript has type `char' jamest: i don't get those subscript has type char warnings at all me either is your system posix i guess it depends on gcc version fwiw geas is biulding again for some reason :) it had nothing to do with a missing ; nope not at all :) jamest: i didn't know it is possible to redirect the output of "make" to xchat. how did you do that? make | xchat ? ? i cut n paste copy and paste j/k you asked for warning, you got warnings that's called "mouse piping" message("Word!"); that function call message("Up!"); results in Initialising security system Word! jamest: "gcc --version | xchat" please /exec -o command works too jamest: ok i would bet it is an orbit problem :((( 2.95.2 jamest: i have exactly the same you defined CFLAGS or something like that in environment? orbit 0.5.8 sorry was incorrect no jamest: ok i would bet it is an orbit <> problem yea slowaris sucks Action: chillywilly runs ToyMan (stuq@c5300-2-ip206.albany.thebiz.net) joined #gnuenterprise. ra3vat: you need to have this in gnue.conf neilt: gnue.conf or geas.conf? ODMG 2.0 ODL Metadata and OML Interfaces for Java (JDK 1.2) Free download sorry geas.conf think that would be worth looking at? neilt: ok you should be able to copy it from geas.conf.sample now i have Initialising business class method handling. 'import site' failed; use -v for traceback Traceback (most recent call last): File "", line 1, in ? ImportError: No module named CORBA CORBA._load_idl("../idl/geas.idl") Traceback (most recent call last): File "", line 1, in ? NameError: There is no variable named 'CORBA' Traceback (most recent call last): File "", line 1, in ? ImportError: No module named GEAS chillywilly: java? jamest: lol about the warnings jamest: ya but the implementation might gives us some clues as to the spec without having to buy the damn thing :P ah not a bad idea :) ttp://www.odmg.org/library/readingroom/Article%20-%20Components%20Strategies%20-%20July98.html er http://www.odmg.org/library/readingroom/Article%20-%20Components%20Strategies%20-%20July98.html make -j 2 on orbit jamest: can you look at the definition of "isspace" in your /usr/include/ctype.h ? extern int isspace(int); ra3vat: it did not find the corba????.so now if they had UML diagrams then you wouldn't need to look at their friggin code jamest: hmmm... it is not a macro? but all examples work and import CORBA works there all those char index warnings point to calls to isspace like if (!isspace (*p)) and i don't see where there is a subscript unless isspace was a macro euler$ grep isspace /usr/include/ctype.h extern int isspace(int); #define isspace(c) ((__ctype + 1)[c] & _S) #define isspace(c) ((_ctype + 1)[c] & _S) neilt: what is corba????.so? where you have yours? jamest: ah jamest: this causes the warnings I'm happy to ignore it jamest: please send a patch to solaris ;) # orbit location ORBit-python /Volumes/AnagadaII/sw/lib/python2.1/site-packages/CORBAmodule.so reinhard: you want a patch? no was just kidding ah ok actually i regard this as a bug in the header file not in geas ah 'import site' failed; use -v for traceback jamest: re: our other problem i have not seen that and dont know the code well enough you have any information about whether orbit runs on solaris or not? reinhard: i have ximian gnome running on a sun box except that we just found out that it appearently doesn't oh so it should actually orbit builds on it and runs on a different box here build != runs ok i'm rebuilding and will make test if it support s it btw: was looking inot omniORB it does solve problems for forms :) (yeah) BUT no C bindings so cant use with geas unless someone wants to convert the c++ or python bindings to C btw: i think someone should do this :) omniORB looks very nice c++ is nice....however no one else here likes it do they? C++ is ok v4.0 is in preview now C++ has none of the benefits of C (speed portability) it had issues with a fluxuating ABI IIRC yet carries most of its baggage thus if you code in C++ I ask why not java or python? its C++ w/o C's bagagge C++ is bit better in the performance area then python and java C++ is a cluster fsck results in code bloat and is slow but the main reason someone uses C++ over C is to have objects if so use python java or something that was DESIGNED for objects I wanna learn objC one of these days not just has objects 'bolted' on although it is still faster than java, dont know about python it is fater than python c++ is faster than pthon python but in reality on todays machines speed diff between native compile and interpretted is minimal c++ is no more bloated then any other OO language in most htings the problem is that objects in c++ bloat the code and make it really slow fwiw - orbit passed all tests during a make check un, no c++ is even used in embeddd systems embedded if you use any templates you get 20 meg exec's jamest: Action: reinhard sighs yes, templates are heavy Action: reinhard doesn't know what to say no one uses them in an embedded system Nick change: neilt -> neilt-away jamest: i believe you now know exactly as much as me reinhard: I'd suggest you say "Just use linux" however, kde is all c++ and form what you guys say it is a lot better so what does that say? jamest: no i know somehting better as that's the typical responce to problems these days jamest: works for me ;) chillywilly hell java is used in embedded systems :) rofl KDE developers have m4d sk1lz ? dneighbo: only by mental ppl, for this wireless medical application we did on a little PIC you are not going to fit a JVM on there jamest: i _really_ care about portability c++ can be portable jamest: i currently have like 50 sco linux machines out in the field mozilla developers guide has some nice portability rules jamest: and you probably know i care more and more seriously about portability than probably any other of this project I really don't care will i care so this will never be the answer i will give you reinhard: it should please you to know that our install is SCO as well I'd like to use GObject in the newest glib for objects, that's probably not going to happen either jcater: not that i like SCO it is there for historical reasons Bayonne is all C++ and GNU Common C++ is a nice lib....;) anyway... it's not gonna happen reinhard: believe me, I understand I have a SCO and a Solaris machine to support (for historical reasons) I have solaris and BSD machines yep, C will be more sane on those machines and I know I'm not an exception :) (cause they fit the need) c++ compilers are known to be broken doe skde run on SCO? chillywilly: have no desire to test but I sure know gnome doesn't they don;t use templates or any crazy shit liek that I dunno it was an oracle server gnome sucks for portability a telnet + oracle server they did it all back asswards if you pay ximian I am sure they'll port it....bbwaahahhahahaha reinhard: are the interfaces between the 4 pieces cleanly defined? derek (derek@cpe-24-221-112-50.az.sprintbbd.net) got netsplit. reinhard (rm@62.47.44.46) got netsplit. jamest: for the most popart yes part jamest: you talking about geas modules? reinhard (rm@62.47.44.46) returned to #gnuenterprise. am i back now? something like that farmer and benford split 1. class definition handling yop 2. database backend interface 3. method handling yop [16:22] Last message repeated 1 time(s). 4. interface to the front end (CORBA and friends/warnings) yop where the four that were listed are the interfaces clean? they are all separated pretty miuch yea wouldn't you say reinhard? i would say mostly clean and well defined still not perfect of course derek (derek@cpe-24-221-112-50.az.sprintbbd.net) got lost in the net-split. it's very much like 1, 2 and 3 are separate and 4 sits on top of all and connects them it is not 100% true but it comes near ya hopefully 4 will move out into Gcomm or something :P at least they are so clear that i would dare to let somebody work on any of 1 to 3 without knowing the other stuff in detail right? is 1 where object instances are kept alive no the objectcahce and object store make up 2 1 is where class definitions are handeled objectcache keeps instances alive and do the objects exist in ram and are flushed to persistant storage during slow periors reinhard: can you explain this: Initialising business class method handling. 'import site' failed; use -v for traceback i.e. *.gcd are read and a memory structure is held so that other modules can ask which object exist and which fields and methods they have ra3vat: this is a python error i guess not sure what it means so 1 is the gcd parser jamest: objects exist in ram and are flushed to db with a not yet very sophisticated system and if 3 clients ask for a handle to an object jamest: yes 1 is the gcd parser plus the code around it that keeps the parsed information alive while geas runs they each get a copy don't know for sure about that I'm not even sure how well it does with multiple clients stbain (gunandgiz@cx726805-a.roanoke1.va.home.com) joined #gnuenterprise. i haven't looked at the caching code in detail there's no locking or anything reinhard: what do you think about concurrent access? geas would handle it or would it fall on its face? the current code chillywilly: what everybody would think about concurrent access at 11:30 p.m. and after 15 hour of work nothing currently :) seriously this is an issue we haven't tested by now I didn't think it would hold water either reinhard: you think it would be worth looking at ODMG's Java implementation of their 2.0 spec? everything should be at least worth looking at if somebody has the time derek (derek@cpe-24-221-112-50.az.sprintbbd.net) joined #gnuenterprise. wb masta masta ignores me :( Action: chillywilly thwaps masta with a 25lb salmon that'll teach ya derek (derek@cpe-24-221-112-50.az.sprintbbd.net) got netsplit. dneighbo (dneighbo@mail.libertydistribution.com) got netsplit. ajmitch (ajmitch@p20-max7.dun.ihug.co.nz) got netsplit. Topic changed on #gnuenterprise by ChanServ!s@ChanServ: GNU Enterprise : http://www.gnue.org [If no one is home email info@gnue.org] #gnuenterprise: mode change '-o jamest' by ChanServ!s@ChanServ #gnuenterprise: mode change '-o gnuebot' by ChanServ!s@ChanServ #gnuenterprise: mode change '-o neilt-away' by ChanServ!s@ChanServ #gnuenterprise: mode change '+o gnuebot' by ChanServ!s@ChanServ derek (derek@cpe-24-221-112-50.az.sprintbbd.net) returned to #gnuenterprise. dneighbo (dneighbo@mail.libertydistribution.com) returned to #gnuenterprise. ajmitch (ajmitch@p20-max7.dun.ihug.co.nz) returned to #gnuenterprise. weeeee seems like all these friggin Java based standards you have to buy...those ODMG ppl are Java freaks l8r jamest (jamest@hobbes.math.ksu.edu) left #gnuenterprise. bah, their code isn't even complete chillywilly: is ours complete? :) no, but I mena they only give like the interfaces and no code whatsoever i.e., you cannot reverse engineer th standard bastards :P ah you mean like this their code _would_be_ complete but they don't give it to you public interface Type extends MetaObject { public DSet getCollections(); public void setCollections(DSet newCollections); static final public FeatureDescriptor collectionsInverse = Relationship.Inverse.getDescriptor(Collection.class, "subtype"); public DSet getSpecifiers(); public void setSpecifiers(DSet newSpecifiers); static final public FeatureDescriptor specifiersInverse = Relationship.Inverse.getDescriptor(Specifier.class, "type"); public DSet getUnions(); public void setUnions(DSet newUnions); static final public FeatureDescriptor unionsInverse = Relationship.Inverse.getDescriptor(Union.class, "switchType"); public DSet getOperations(); public void setOperations(DSet newOperations); static final public FeatureDescriptor operationsInverse = Relationship.Inverse.getDescriptor(Operation.class, "result"); public DSet getProperties(); public void setProperties(DSet newProperties); static final public FeatureDescriptor propertiesInverse = Relationship.Inverse.getDescriptor(Property.class, "type"); public DSet getConstants(); public void setConstants(DSet newConstants); static final public FeatureDescriptor constantsInverse = Relationship.Inverse.getDescriptor(Constant.class, "type"); public DSet getTypeDefs(); public void setTypeDefs(DSet newTypeDefs); ya these methods are not defined almost all of the classes are like that oh wait that's an interface anyway duh Initialising business class method handling. 'import site' failed; use -v for traceback Traceback (most recent call last): File "", line 1, in ? ImportError: No module named CORBA CORBA._load_idl("../idl/geas.idl") Traceback (most recent call last): File "", line 1, in ? NameError: There is no variable named 'CORBA' Traceback (most recent call last): File "", line 1, in ? ImportError: No module named GEAS method dir name: '../examples/businessobjects/methods' Loading file ../examples/businessobjects/methods/test.py Loading file ../examples/businessobjects/methods/neil_test.py Loading file ../examples/businessobjects/methods/addrbook.py warning: [methods_python.c/315] [python_load_method_handlers] Attempt to register method: root__test_callme Traceback (most recent call last): File "", line 1, in ? NameError: There is no variable named 'root__test_callme' warning: [methods_python.c/315] [python_load_method_handlers] Attempt to register method: root__test_dosomething Traceback (most recent call last): File "", line 1, in ? NameError: There is no variable named 'root__test_dosomething' ** CRITICAL **: file classdata.c: line 529 (odl_namelist_free): assertion `list != NULL' failed. Initialising object store. Initialising object cache. Validating/updating database table definitions. Initialising ORB is i tok that it seek methods as 'root__test_dosomething' with root__ before? don't think so who is trying to do 'import site'? this sounds like a python statement isn't there a site-python directory or something like that? maybe it tries to import that directory or the modules that are there or whatever Action: reinhard really does not know much about python this is a fles grammar file for OQL er, flex huh? Object Query Language flex does not define grammar you mean bison? whatever options { language="Cpp"; } class OQLLexer extends Lexer ; options { k=2; charVocabulary = '\3'..'\377'; testLiterals=false; // don't automatically test for literals caseSensitive=true; caseSensitiveLiterals=true; } is that Bison syntax? ugh it is java lol oh no cpp duh reinhard: why embedded into geas interpreter can't import that site-packages? jamest (jamest@fh-dialup-201168.flinthills.com) joined #gnuenterprise. #gnuenterprise: mode change '+o jamest' by ChanServ!s@ChanServ reinhard: it complains about "ImportError: No module named CORBA" that is it tries to import the corba module python-orbit um orbit-python or whatever reinhard: and than loading file ../examples/businessobjects/methods/test.py reinhard: http://www.odmg.org/files/oqlg.zip ra3vat: that's ok chillywilly: thanks reinhard: that have all the same import statement in there reinhard: and all examples work fine for me reinhard: it's only behaviour of embedded interpreter hmm..... reinhard: after compiling geas --with-methods=python i have no clue why embedded interpreter does not load site package which normal python interpreter load by default i think? reinhard: http://obfuscation.dyndns.org/odmg/ nice transaction interface we can steal :P reinhard: btw, that my web server that's hehe obfuscation :) good name ra3vat: i have no clue either ra3vat: sorry reinhard: I thought it was a cool name obfuscation.dyndns.org output form hostname -f :P I wish they ha d afreely downloadable C++ implementation as this Java one is just greek to me bah is there any way to unjar something without having jar? chillywilly: try unzip it k aaah kewl it worked Nick change: neilt-away -> neilt #gnuenterprise: mode change '+o neilt' by ChanServ!s@ChanServ kernel segmentation fault: root file system corrupted. system terminating. Just kidding :) Remember: Detailed bug reports are a valuable tool, if you want the problem fixed. :) A full stack trace from 'gdb' in particular is useful for identifying exactly where and why an error occured. Segmentation fault jamest: looks familiar but didn't get that for quite some time :) masta would be so proud what is a segfault again :) seriously neilt: it' s called debugging methodology :) ahhh ok i'm too tired today to track that, if you don't find it within the next 12 hours i will be here again then i must be number one programmer because i dont any segfaults anymore define hook-stop handle SEGFAULT nopass end define hook-run handle SEGFAULT nopass end define hook-continue handle SEGFAULT nopass end that'll fix it yes it will :) mdean (mdean@arc11x115.kcnet.com) joined #gnuenterprise. reinhard: print sys.path from normal python gives: ['', '/usr/lib/python2.0', '/usr/lib/python2.0/plat-linux-i586', '/usr/lib/python2.0/lib-tk', '/usr/lib/python2.0/lib-dynload', '/usr/lib/python2.0/site-packages', '/usr/lib/python2.0/site-packages/PIL'] from embedded: ['/usr/lib/python2.0/', '/usr/lib/python2.0/plat-linux2', '/usr/lib/python2.0/lib-tk', '/usr/lib/python2.0/lib-dynload'] after that import site does not work in embedded ra3vat: i am now 100% sure i can't answer this this is too python specific maybe you can find something in the python faq or mailing lists or something like that must be 3:30 a.m. for you? now? ok, may be tomorrow :) jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) left irc: home nite all reinhard (rm@62.47.44.46) left irc: Those who say it can't be done should never stop those who are actually doing it. night later all ra3vat: i will think about your problems also neilt (neilt@dialup-63.208.68.139.Dial1.Chicago1.Level3.net) left irc: ok, guys slowly they fall one by one muuuwahahahahahaha ******************* for neilt/reinhard ******************* just changed methods_python.c to extend sys.path PyRun_SimpleString ("import sys"); /* ra3vat */ PyRun_SimpleString ("print sys.path.append('/usr/lib/python2.0/site-packages')"); PyRun_SimpleString ("print sys.path"); PyRun_SimpleString ("print sys.path"); it solved previous problems. neil_methods_test.py works. still don't understand where emb interpreter initializes sys.path from ************************************************************ ToyMan (stuq@c5300-2-ip206.albany.thebiz.net) left irc: [x]chat dneighbo (dneighbo@mail.libertydistribution.com) left #gnuenterprise. l8r ra3vat (ds@195.239.66.37) left irc: [x]chat jcater (jason@1Cust10.tnt3.memphis3.tn.da.uu.net) joined #gnuenterprise. howdy chillywilly_ (danielb@d123.as7.nwbl0.wi.voyager.net) joined #gnuenterprise. chillywilly (danielb@d20.as15.nwbl0.wi.voyager.net) left irc: Ping timeout for chillywilly[d20.as15.nwbl0.wi.voyager.net] Nick change: chillywilly_ -> chillywilly howdy dude this telethon thing is cutting into Friday Prime on scifi rofl not to mention it is very high on the cheese o' meter Action: jcater is away: dinner can't they read something without screwing up :P man I am really annoyed I wanted to watch farscape neil young was cool though neil young is always cool just a figglet of your application hehe Action: jcater is back (gone 00:16:59) he's back! yes he is put him back we don't want him Action: jcater slaps chillywilly with a catfish Action: chillywilly devours the catfish and jcater arm jcater's oooh Action: jamest slaps chillywilly with a bucket of catfish turds dont' abuse the coder! we're not done exploiting him then you can abuse him why si everyone always against me? Action: chillywilly takes a big shit on all chillywilly: no one's against you the world is against me but I don't like ppl shitting on me umm, too bad :P let the gnue ops ban me then Nick change: derek -> dnWifey poor dnWifey mdean (mdean@arc11x115.kcnet.com) left irc: Ping timeout for mdean[arc11x115.kcnet.com] chillywilly (danielb@d123.as7.nwbl0.wi.voyager.net) left irc: jcater (jason@1Cust10.tnt3.memphis3.tn.da.uu.net) left irc: nite jamest (jamest@fh-dialup-201168.flinthills.com) left irc: [x]chat dnWifey (derek@cpe-24-221-112-50.az.sprintbbd.net) left irc: Ping timeout for dnWifey[cpe-24-221-112-50.az.sprintbbd.net] uday (uday@host-64-110-96-159.interpacket.net) joined #gnuenterprise. nickr (nick@sdsl-64-32-227-138.dsl.nyc.megapath.net) left irc: Read error to nickr[sdsl-64-32-227-138.dsl.nyc.megapath.net]: EOF from client nickr (nick@sdsl-64-32-227-138.dsl.nyc.megapath.net) joined #gnuenterprise. --- Sat Sep 22 2001