poop i missed ole neilt nickr (nick@63.163.68.113) left irc: Ping timeout for nickr[63.163.68.113] nickr (nick@63.163.68.113) joined #gnuenterprise. you get the DSL back? nickr nickr (nick@63.163.68.113) left irc: Killed (NickServ (This nick is reserved by another user)) guess not :) heheh hiya derek sacha_ (sacha@chime-a-187.conceptual.net.au) joined #gnuenterprise. hey chillywilly ello ra3vat (ds@195.239.64.223) left irc: [x]chat Mr_You (rouzer@mankind.boredom.org) left irc: night reinhard (rm@N802P023.adsl.highway.telekom.at) joined #gnuenterprise. hi morning sorry for slow response every morning i start xchat then i switch to another virtual screen and read my mails aaaaah ra3vat (ds@195.239.66.12) joined #gnuenterprise. ra3vat: looking at your error now hello reinhard: you use X? yep chillywilly: he puts the X in Sex wait thats an old KISS song neverming reinhard: my error - can you comment a little? derek: hehe reinhard: what are you running inX? chillywilly: gnome of course :) you never fail me :) reinhard: how would be right to solve it? ra3vat: i think the rtrim function is not needed any more it was needed once because andrewm used char instead of varchar in postgres reinhard: ok ra3vat: sorry for misunderstanding with "your error" i did not mean that it was your fault i meant the error message you got chillywilly (danielb@d132.as13.nwbl0.wi.voyager.net) left irc: night thanks for clarifying, now i can stop to blame myself :) ra3vat: should be fixed now reinhard: thanks, i can try it only at the evening reinhard: what files have you touched? ra3vat: reinhard did not have 'sexual relations' with those files oh.. you mean a different kind of 'touching'... i realize that was so american centric humor it probably makes no sense to you non americans (unless you watched a lot american news) Nick change: derek -> pissed_off derek: ah, is it another song? about 'touching' Nick change: pissed_off -> derek its bed time reinhard: ok, found it now bbl ra3vat (ds@195.239.66.12) left irc: Client Exiting Rafterman (tim@lister.sesgroup.net) joined #gnuenterprise. derek: are ya ok buddy? :) sacha_ (sacha@chime-a-187.conceptual.net.au) left irc: Read error to sacha_[chime-a-187.conceptual.net.au]: Connection reset by peer neilt (neilt@dialup-63.208.167.246.Dial1.Washington2.Level3.net) joined #gnuenterprise. #gnuenterprise: mode change '+o neilt' by ChanServ!s@ChanServ neilt (neilt@dialup-63.208.167.246.Dial1.Washington2.Level3.net) left irc: Ping timeout for neilt[dialup-63.208.167.246.Dial1.Washington2.Level3.net] neilt (neilt@user-2ivekde.dialup.mindspring.com) joined #gnuenterprise. #gnuenterprise: mode change '+o neilt' by ChanServ!s@ChanServ Rafterman (tim@lister.sesgroup.net) left irc: [x]chat hello all hello neilt reinhard: hello say your changes thanks we needed to remove that code did you gat all of the db switches taken care of no actually it was to make oql code db independent s/gat/get/ but to fix another bug ra3vat found out i was thinking about that very thing last night after the iRC with derek the remaining db dependencies outside objectstore are absolutely minor we should scan the code and see how many db switches are left jamest_ (jamest@hobbes.math.ksu.edu) joined #gnuenterprise. 31 21 of them in oql ToyMan (stuq@c5300-1-ip123.albany.thebiz.net) joined #gnuenterprise. most of them are useless actually will i did not put them there, and i thought it was my job to add useless mostly working functionality to geas. So how has been stealing my thunder? s/how/who/ :) :) dare i ask what you're talking about jamest_: not sure if it's good when you know it ;) there is still some code that is db dependent and should not be but most of them is very easily fixable that's it in a nutshell neilt (neilt@user-2ivekde.dialup.mindspring.com) left irc: Ping timeout for neilt[user-2ivekde.dialup.mindspring.com] :) neilt (neilt@user-2ivel2k.dialup.mindspring.com) joined #gnuenterprise. #gnuenterprise: mode change '+o neilt' by ChanServ!s@ChanServ jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) joined #gnuenterprise. morning, y'all howdy he lives, he breathes, (well maybe not he is AI after all) lol jcater: the question is, "Are you completely machine, or are you borg?" ra3vat (ds@195.239.66.38) joined #gnuenterprise. bigbrother joined #gnuenterprise. jamest (jamest@fh-dialup-202033.flinthills.com) joined #gnuenterprise. bigbrother and jamest : the :REAL: dynamic duo bye bye work time derek (derek@cpe-24-221-112-50.az.sprintbbd.net) left #gnuenterprise. work? neilt (neilt@user-2ivel2k.dialup.mindspring.com) left irc: Ping timeout for neilt[user-2ivel2k.dialup.mindspring.com] dyfet (dyfet@63.124.177.154) joined #gnuenterprise. dyfet (dyfet@63.124.177.154) left irc: Client Exiting jcater? is that you? yip it's me ra3vat (ds@195.239.66.38) left irc: Ping timeout for ra3vat[195.239.66.38] ra3vat (ds@195.239.66.10) joined #gnuenterprise. neilt (neilt@user-2ivelk3.dialup.mindspring.com) joined #gnuenterprise. #gnuenterprise: mode change '+o neilt' by ChanServ!s@ChanServ reinhard: you here? yews yes did you see the email, we need to change sys_ to geas_ in order to work with oracle aaaaarrrggh!! I like geas_ better, but I don't think you have to make the change specifically for oracle larry wall will have to pay for that !! (well i hope that was his name, wasn't it?) just saw it :) will do not today or tomorrow but will ok, thanks Please send me a copy of the Tuesday gnue-geas digest. I inadvertently deleted it while cleaning up a bunch of junk emails. Thanks in advance, Jim who can do that? i just responded usually the mail list can but the user will have to requuest it directly from the mail list jamest: you here - neilt (neilt@user-2ivelk3.dialup.mindspring.com) left irc: Read error to neilt[user-2ivelk3.dialup.mindspring.com]: Connection reset by peer here dneighbo (dneighbo@mail.libertydistribution.com) joined #gnuenterprise. just read mail and i dont get it neilt is asking for data to make oracle and db2 drivers for geas? i guess that is a RESOUNDING *NO* about trying to get driver reusage? Action: dneighbo is very frustrated Action: jcater just responded im not saying we must use what is in gnue-common currently but we should be sharing drivers btw jcater did you see the error in the quick fix and work shouldnt really be put into one or the other in any major way until we decide what we are going to do is there a chance of patch in the next 24 hours or so Action: dneighbo is still failing to see how we can not call this a 'divide' or can you point me at the area needing attention jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) left irc: Read error to jcater[w202.z065105010.mem-tn.dsl.cnc.net]: Connection reset by peer jamest_ i guess thats a hell no :) lol jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) joined #gnuenterprise. who is the gnue-geas moderator? jamest: I will look now btw jcater did you see the error in the quick fix is there a chance of patch in the next 24 hours or so jamest: I will look now or can you point me at the area needing attention yes I'll make a quick fix to my quick fix until a good fix rears its head jcater: i just approved it I hate to see divides :( :( jamest: I just committed a fix... I am not in a position to test it, though but if it doesn't work, you can tell from the commit email where to look (it should work, though) cool very cool boy, had that been C it'd taken me a while to trace and fix :) ra3vat (ds@195.239.66.10) left irc: Ping timeout for ra3vat[195.239.66.10] Mr_You (rouzer@mankind.boredom.org) joined #gnuenterprise. ra3vat (ds@195.239.66.34) joined #gnuenterprise. neilt (neilt@user-2ivelcd.dialup.mindspring.com) joined #gnuenterprise. #gnuenterprise: mode change '+o neilt' by ChanServ!s@ChanServ hello all its not a reounding NO resounding bbl.. lunch Nick change: jcater -> jcLunch dneighbo: the decisions have already been made Action: neilt has been frustrated for some time and the lack of integration between 2 tier and n tier what decision? jamest: i sent you an mail about the email listserver help not working ra3vat: geas and 2 tier will have separate drivers for SQL server interface will have? or for a long time had? ra3vat: your right, for a long time have had but design requirements for drivers in geas and forms are different from a performance and orientation standpoint when the drivers were being rewritten for forms i dont remember anyone asking how do we support geas anil (anil@host-64-110-96-159.interpacket.net) joined #gnuenterprise. or getting into that discussion i'd like to see geas support in preference btw, for the comments here the email was about looking at options, not about making decisions If so, please let me know so I can look at the complexity of writing the drivers for GEAS. i estimate that it will take a day for oracle and a day for db2 i can in not way fathom that we could do C->python->driver API->sql ->driver api ->python ->c in two days bbl ra3vat (ds@195.239.66.34) left irc: Client Exiting pheeww.. lots of mozilla bugs :-(((( oh well, guess no one is around to discuss when i am, thats part of the problem later all neilt: actually i was here but i am not sure if it makes sense when i discuss with you as i basically share your view :) not when i am, that is, but that we dont all get together at the same time reinhard: thanks im here now two things the reason that i dont think it is a divide is that a. when the drivers were rewritten i asked at that time to look into making something be usable for geas and forms it was ignored and i was equally frustrated by that there are real technical issues that the 2 tier team is not aware of this is not philosophical as to 'performance' i think james was willing to benchmark the difference this weekend and i though i heard them say they were willing to move to C if there was a performance issue as to the wrapper from jamest and jcater point of view i think they are thinking it could be a a day or two of code my issue is more work put into EITHER system will only make it harder to be united as no one likes to ditch something put time into so more time put into either will only make it harder for the teams to agree on one or the other or something new i dont think there is technical reasons this cant be done i.e. driver sharing as i dont see anything in forms that wouldnt be used by geas i do see some geas stuff that forms wouldnt use but im not hearing the forms team say these additional geas (ntier) apis are a concern to them in sense of bloat or performance there is no reason for anything in geas that forms does not use neilt: bad wording, ntier vs 2 tier hello !!! but you are correct even some of those things could e 'reused' that currently are not being used dneighbo: here is my question anil: hi hi anil in geas we can search 1000 records in 3 to 5 seconds how to install geas development libraries ??? cvs or tarball? that is toooo slow is there any flag in the Makefile like make devel to install ? neilt: currently in forms its much quicker than that now we are talking about inserting a c->python and python->c abstraction i have geas-0.0.6.tar.gz but we are not going via an abstraction layer of objects that will only slow things down neilt: i dont see it as abstraction ok call it an interface then i use c applications with vb activex bundled in them all the time the performance is not hideous or really even noticeable it will still be 4 to 10 calls per routine to go both ways anil: can you please tell me what exactly you are trying to do anil: and what you are missing? neilt: maybe im misunderstanding how python and C are able to integrate the only way we can tell is by prototype i am trying to develop geas clients in C so we either a. create all new C drivers that are reusable and get forms to work with them dneighbo: i dont know either but i assume a bunch of wrappers i am trying to translate the java code given in the geas/examples to an equivalent C code b. prototype reusing python drivers that in place in gnue-common anil: may i ask why? not be a wise guy, but basically the power of geas is turning things into objects C is predominately not equiped to use objects with out lots of effort? dneighbo: ??? brb therefore it doesnt seem like a really good language choice for a 'client' to geas i dont know your needs but there is a forms client that uses geas (for data storage now, very soon in full glory) dneighbo: i dis-agree but only slightly so you have to do no work geas does not have to do objects it just handles objects and converts them to SQL short of creating the gui (with a RAD tool we have) and creating your gcd's neilt : i agree, my point being if you didnt use the objects why not just use sql :) but tool choice and usage is up to you because a single sql server does not scale well without a middleware server just trying to understand, also curious as to what using geas for with C geas is about speed if you dont need to scale and you dont need speed and you have 10 clients you should use 2 tier 2 tier will not scale well into multi server distributed databases espeically multivendor databases so 2 tier works well now, for small demos and small business apps but it wont work for large scale stuff neilt: i coud NOT agree more i am a huge fan of n-tier if that architecture worked in large scale stuff we would not see middleware in the marketplace i agree you arent seeing me argue that my point was if using geas and not using objects why bother using geas :) dneighbo: agree who said anything about not using objects me :) i told anil that C was not OBJECT friendly so its not a 'good' 'client' language for geas IMHO sorry, i gotmixed up thought you were talking to me something like java, python, C++ or that supports objects better is much more suited to writting client applications for geas dneighbo: agree ok :) thought you might have been thinking i was meaning somthign else :) you know im very pro n-tier and why it drives me nuts that there is an app server with skeleton objects that i NEED now today for paying work but cant use the only reason for C could be non-gui uses without creating my own custom clients(which i have no urging to do) all we need is geas driver if you have immediate need we could also talk with cw and maybe we could start on methods and give him some help i would not want to step on his toes, but he has been tied up with other activities neilt im not worried about methods i have short term solution there shared 'triggers' are advanced enough that i can just have biz logic on 'client' its ugly cool but once geas triggers are made it should be easy to cut and paste to gcd method and do minor altering the real issue is i cant use the gcd data structs based on what i posted the other day anil: did you get your questions answered as objects are subclassed and referenced etc anil there are not examples in cvs you might drop i am on my way neilt andrewm@gnue.org an email and ask for a sample he might have one dneighbo: there are examples in cvs for java and python gnue/geas/examples/ i am trying to convert the java code to C back anil: sorry i was not here :( dneighbo: did you tell me to mail andrew ? anil: the client has to talk to CORBA using a language binding anil: yes what neilt said so you actually need a corba client library i am using C as the binding language not a geas client library so you will have to have an ORBit to C binding to make a C cleint anil: just use orbit-idl to generate geas-stubs.c and geas.h those 2 files are your client library ok Action: dneighbo is away workign The client application needs the stubs to contact the server so i need the idl file which the geas server uses !!! anil: yes they are in the idl/ subdirectory of geas tarball ok there are 13 idl's yes use geas.idl it includes the other 12 so i have to copy all the 13 to my client app dir hmmm i think the correct way would be that geas installs the idl's when it is installed but of course geas is installed on the server machine and the client on the client machine so actually yes you need all 13 files neilt (neilt@user-2ivelcd.dialup.mindspring.com) left irc: Ping timeout for neilt[user-2ivelcd.dialup.mindspring.com] IIRC the new gnuef driver specs were put on-line and a URL posted to various lists for feedback before they were written but that's besides the point GEAS.ConnectionFactory factory = GEAS.ConnectionFactoryHelper.narrow(obj); //Java code how can i do the same in C I've started looking at the new geas driver over the weekend and looking at C/Python integration i think i need geas development libraries to do the same i have found the above functions in geas.h thus my question yesterday about using omniORB for Python anil: yes but i dunno how to link ? all functions should be in geas.h as for why things get done in gnuef when they do anil: you just have to link with geas-stubs.c it's all driven by need of the developers and with the ORBit libraries I _want_ to get geas working if i use geas.h there should be some geas.so.0 to link !!!! I needed (as part of my job) something that worked months ago (when geas still had issues that now seem to be getting resolved) today I need to support that stuff and fix issues in it anil: maybe you will want to check ORBit documentation to understand how to write a client with orbit or maybe you will want to work with another orb for your client jcater is replacing a Oracle SQL*Forms implementation so his needs are very similar to my own i got it reinhard (rm@N802P023.adsl.highway.telekom.at) left irc: Those who say it can't be done should never stop those who are actually doing it. Nick change: jcLunch -> jcater ping pong reinhard (rm@62.47.44.55) joined #gnuenterprise. back Nick change: jcater -> jcater_ Nick change: jcater_ -> jcater i am able to write a geas client in C i just rewrote the java code in the examples dir to C anil: cool thnx for ur help reinhard no prob this will be a big boost to me you didn't need much of it anyway :) ok anil (anil@host-64-110-96-159.interpacket.net) left irc: Client Exiting ra3vat (ds@195.239.66.3) joined #gnuenterprise. ra3vat (ds@195.239.66.3) left irc: Ping timeout for ra3vat[195.239.66.3] jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) left irc: ra3vat (ds@195.239.64.96) joined #gnuenterprise. neilt (neilt@user-2ivem0o.dialup.mindspring.com) joined #gnuenterprise. #gnuenterprise: mode change '+o neilt' by ChanServ!s@ChanServ jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) joined #gnuenterprise. reinhard: do we know if anil is a contributer we should add the c code to cvs neilt: not yet jamest: say your comments, about why and how things get done cool, i am not unhappy with effort or any people just with progress with geas i understand reinhard: thanks been talking w/ jcater and reinhard (after you'd left) jcater and I are willing to do what we can to make code sharing possible so i have to vent sometimes and its way too much lately and I'll redouble my efforts on the geas driver 0 * 2 is 0 right ? j/k i am extremely interested to verify that code sharing is the right solution if code sharing means geas uses python drivers i dont understand how you can benchmark the drivers in a weekend because you will need to compare apples and apples well none of the people (jcater or myself) working in gnue-common demand python code, we are both open to C code and I'm going to take the time to write a C wrapper over next few days to the gnue-common drivers then pump N number of inserts, updates, deletes into a simple table (minus the gnuef overhead) then attempt to do the same thing with the base geas drivers (minus the geas overhead) jamest_: then we have to pump the same stuff thru geas without the geas overhead yes just raw inserts, updates, deletes we'd use the same test file against the same backend jamest_: yes may i add the same inserts, updates and delets to the raw database? yes, both directly to the database via the drivers not via the gnuef or geas overhead and do that we will need a test server using just the objectstore.c code so we will have to build a test server we can use my home server if you like jamest_: i think neilt is talking about server program not server machine cool i meant that we would have to write some modified geas code to make a test yes, that is what I planned a C program that uses geas drivers stripped jamest_: i'll just send you a binary and you install on your home machine, ok :) and same program modified to use gnue-common drivers striped neilt: sure, you look honest in your pic just let me know if I have to run as root :) I can't stress enough that we not against C code we're for code reuse we = those of us in gnue-common and we're willing to switch to C where it makes sense and use python where it makes sense :) you might, i'm sure as i have not figured out all that permission stuff and it would save time on my part huh? jamest_: i would write in cobol if i thought it solved the requirements i dont care my only concern is performance and we already have performance problems in geas yes imho a fixable one i think we should compare 3 numbers andrewm spent 30 days on performance to get it from 15-20 seconds/1000 records down to 3 to 5 seconds testprog -> objectstore -> postgres testprog -> gnue-common -> postgres and neilt: are we sure that's in the dbdrivers or the geas overhead? testprog -> postgres so i would be surprised if performance was an easy fix jamest_: that's exactly the reason i think a good portion of it is in the idl i wonder if we can graft geas onto GApp.py it contains a per function profiler run gnuef -p form.gfd there are many C profilers out and when you exit you get a per functoin breakdown of cpu time usage that was what andrewm was doing yes ah, cool manually :) my point is that we currently dont have any performance to spare hehe we will probably never have :) if GEAS get a reputation as a dog it does not matter how we improve it we are dead ye s so i think this discussion really doesn't make any more sense until we get the figure s actually as all c we can say we know and are working on it and people will understand we say Oh we are doing all drivers in python and they say "really" why? reinhard: agree and if the figures say python is faster, i'll be the first to switch opinion i think the 3 numbers are important to prog->object->postgres prog->common->postgres prog->postgres as it gives us data on diff between current way of geas and common way but the last test lets us know how much bottle neck is in drivers at all as stated earlier the issues is not object vs non-object but python drivers vs c drivers object will allways be a little slower im just retyping what was giving when i said object i meant geas driver he meant "objectstore" you give that up for power and flexibility does anyone have profiler nubmers on current geas no all in all i agree on server we need to be very performance minded but will add people use java regularly in enterprises no which says somethign about what htey are weilling to accept ;) btw: before flaming that was suupposed to be a funny haha how hard are they to generate? reinhard: now it complains about upper(data) java will compile to native code python will not ra3vat: crap seems like date and time was not tested very well let me try to fix hmmm my impression was such a thing existed for python as well maybe that was just byte code compile though neilt1 (neilt@user-2ivek1g.dialup.mindspring.com) joined #gnuenterprise. jamest: by the way is anyone in 2 tier group using parse_load_sql.py i wrote to load SQL into database that does not rely on geas neilt1 : i may have used in teh past neilt (neilt@user-2ivem0o.dialup.mindspring.com) left irc: Ping timeout for neilt[user-2ivem0o.dialup.mindspring.com] is that the one that reads a gcd and builds sql file for creation/insert? we should be talking about strategies for merging 2 tier and n tier data yes thats the one the one that substitutes geas uid's via libuid? if so i have used it in the past yes it now also creates creation date and user and modification date and user fields but stopped when hit brick wall with forms geas driver yes via libuued uuid it would be PAINFUL to do conversions i looked into this long ago neilt1: i have not used it originally i thought the problem was only the uuid issue the same data should run which is now solved boith ways at least the basic data sounds like idea inclusion for gnue-common the problem is that relational tables are built MUCH differently than way geas builds them i am not sure geas is right here geas should hide the differences so if you built a two tier system using gcd to start from you could do more than likely would just be kind of ugly is a way that is compatible but to go the other way would hurt you could certainly DO it just you would have ugly looking objects :) i dont think many (at least me) we can build gcd's that are compatible for a lot of things is going to go from 2 tier to n tier using my relational structures :) if im gonna go via pain of upgrade im gonna go to gnue-config :) i.e. base gnue w/ mods i need but i see value in making gcds and having them build 2 tier tables but if we are about business applications then we are going to build two business frameworks so someone wanting 2 tier today wouldnt have the pain in upgrading later :) personally i see two tier as a tool not a application if that makes sense im not sure we need to make sure all of gnue 'applications' work with two tier chillywilly (danielb@d153.as9.nwbl0.wi.voyager.net) joined #gnuenterprise. so you are saying that we wont build 2 tier sales system and 2 tier supply chain system we just need gnue tools to work in two tier mode for people wishing to develop thier OWN 2 tier solution neilt1 : im saying i would lean towards that but im not saying its teh DEFINITIVE or CORRECT answer Nick change: neilt1 -> neilt i do NOT want to create TWO of all business world i.e. one table script for two tier and one gcd for n-tier so if two tier runs the applications i think it will have to be able to make its stuff off of n tier stuff as business rules become issues too if they are all on server in n tier sorry quick question you would have to shove them into the forms for 2 tier we can do it are we aware that in geas all queries are case insensitive? and do we want it that way? not sure how it works but i say both if you have a query function of some sort you could tell it whether case sensitive or not reinhard: i thought most SQL was case insensitive are you saying for the SQL or for the data part as in stuff im doing i often need both for the values neilt i assume he means data not for the field names yes data select name from contact where name="mike' wouldnt find Mike I think data should be case sensitive select name from contact where upper(name)='MIKE' would find Mike jamest_ i think it should be case sensitive reinhard: i agree with jamest_ as if we stored mime encoded data but i like idea of passing flag to tell it to ignore case case sensitive on the query i do this as my users would die ok thanks on the mike example its stored Mike but i let them in a particular area searc for mike [15:10] MSG541 and it will come back as a hit on any of them anyhow i dont know the geas code so dont know if its functions are suited for this ok thanks and i don't know 100% if that was actually correct what i said about how geas does it now edmz (nobody@dial-148-240-68-195.zone-2.dial.net.mx) joined #gnuenterprise. anyone alive ? have to run to meeting later all define alive? you bet :) neilt: later what is gnue being coded in ? (i just found about it...) neilt (neilt@user-2ivek1g.dialup.mindspring.com) left irc: bb in 3 hours a state of chaos edmz: well what would you like? python :P i guess we can offer a piece of code in every language you like ;) edmz: you lucky bastard ! :) :) python and C is the answer hehe we've been debating on merging pieces of both into one unnatual union today nice.. :) you interested in helping out? geas is only the framework, right ? it's our application server I don't think framework quite describes it jamest (jamest@fh-dialup-202033.flinthills.com) left irc: Ping timeout for jamest[fh-dialup-202033.flinthills.com] ToyMan (stuq@c5300-1-ip123.albany.thebiz.net) left irc: Ping timeout for ToyMan[c5300-1-ip123.albany.thebiz.net] my strength is definetly more on the python side than on C.... but it's very much the heart of our system framework is all tools together Action: jamest_ is trying to earn brownie points with the geas developers :) :) geas will be the heart of things soon though is there a roadmap somewhere ? some developers docs ? jamest (jamest@fh-dialup-202033.flinthills.com) joined #gnuenterprise. developer docs yes ToyMan (stuq@c5300-1-ip123.albany.thebiz.net) joined #gnuenterprise. roadmap i tried to beat it out of folks in SF it didnt happen what folks ? :) we were too quick for him ra3vat (ds@195.239.64.96) left irc: Ping timeout for ra3vat[195.239.64.96] gotta lay off the doughnuts and 64oz Dr Peppers there Mr Neighbors folks = jcater and myself others that should be beaten include (but are not limited to) neilt and reinhard hehe ah and chillywilly but they didn't make it to SF to the LWCE ? ra3vat (ds@195.239.64.96) joined #gnuenterprise. yip yes beatings are definitely global (so reinhard dont think you are opting out by default) ra3vat: which method are you using to query the database did you guys gave a speech or something like that ? getObjects? reinhard: hold on i do newQuery ah cool then you do addField? edmz : unfortunately we butchered a speech yes :) hehe reinhard: yes good the struct Field has a member is there another place where to get orbit-python other than the url from the INSTALL ? casesensitive you have to set that to TRUE for date values edmz by that we gave a speech just not on a topic we were comfy with :) this is of course a quick fix and not the final solution ;) actually you have to set it to true for everything except char and varchar cant get orbit-python :/ edmz: hold on edmz: we gave a "gee, thanks for letting us know the day of the speech that we have to give a speech" speech edmz: did you try sourceforge? reinhard: ah, that easy :) jamest: didnt know it was there....... let me check.... http://sourceforge.net/projects/orbit-python edmz sorry INSTALL is not up to date :( thanks reinhard: that syntax for f=GEAS.Query.Field makes me ill can orbit-python work with 2.1 ? you mean can orbit-python work with python 2.1 or can geas work with orbit-python 2.1 ? first one.. :) should work imho reinhard: seems geas puts upper() regardles of casesensitive ra3vat: really? you have set it to TRUE? reinhard: yes, only if TRUE should be casesensitive itself? True true TRUe ? whats a good --prefix for compiling geas ? /usr ? edmz: for now it's better to not install geas we all run geas from source directory ouch ahh.. so the prefix is not of much interest ok.. glad I asked :) ra3vat: hmmm.... ra3vat: not sure ra3vat: try 1 should worke likewise do I need to download any other packages or is only geas enough ? edmz: depends on what you want to do reinhard: try the whole thing :) edmz: i do ./autogen.sh --with-postgresql --enable-debug edmz: trying the whole thing could become hard because we currently have front end (form generator / report generator) and middleware (application server) both running for themselves but as we speak we are working on combining them so you have to choose one of the two for now reinhard: what do you suggest I choose ? as you like python i would rather suggest the front end gnuef and reports they are written in python geas is written in c reinhard thats not true forms works with geas currently as does reports just not in a 'fully functional' way dneighbo: oh sorry are the independent ? (do i need geas to run gnuef) i thought this was broken with forms driver update edmz: they are independent you can run gnuef in 2 tier mode which means gnuef directly accesses the database edmz: fully independent in the long run we want 3 tier mode which means gnuef accesses geas bbl: meeting and geas accesses the database then I should get GNUe-Forms-0.0.9.tar.gz right ? maybe or maybe you may want to get cvs which is the bleeding edge of current development edmz: i always prefer cvs version of a whole project edmz: he knows what he talks about :) whats the cvsroot ? :) what should i check out ? (please) edmz: http://savannah.gnu.org/projects/gnue contains a cvs section with the info you need i guess I need to checkout the whole gnue...... :) edmz: for forms and reports you need gnue/gnue-common gnue/gnuef gnue/reports iirc thanks cvs -d :pserver:anoncvs@subversions.gnu.org:/cvs co gnue l8r guys bb in about 2 hours jamest_ (jamest@hobbes.math.ksu.edu) left irc: [x]chat could somone please approve my message? Action: chillywilly needs to resubscribe with his FD email I wish mailman would have like an account for you and you could just change your email can a list damin change my subscriptions? admin seems like wxWindows is a prerequisite too edmz yip and if you wish to use forms with geas you will have an python-orbit dependency too i solved that one already.... jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) left irc: later edmz: you said you just heard about gnue? mind me asking how? well, really, i found about it like 5 months ago...... but I didnt see much activity going on...... today, i just suddenly rememberd about it, checked the site and i saw more going on... :) i really dont remember how I found about it 5 months ago... i think a friend told me dneighbo: you have access to the mailing lists? yeah but i dont have time just unsubscribe then resubscribe i cna look at it tonight there's too may friggin lists ugh I am not doing all of that shit again...mailman sucks it wouldn't be so bad of there was a place you could see all your subscriptions at instead of needing those tupid mail with URLs in them grrr nite all bb in about 9 hours ;) it might be nice to admnistrate, but the user side of it sucks big time bye reinhard (rm@62.47.44.55) left irc: Those who say it can't be done should never stop those who are actually doing it. Action: chillywilly should hack mailman Wednesday 31 October 2001, 16:30 until 19:00 Max Palevsky Cinema, Ida Noyes Hall 1212 East 59th Street Chicago, IL 60637. RMS is gonna speak there Action: chillywilly should go to that one what is bayonne for ? bayonne is telephony application server used to write pbx applications for call centers and the likes among other thigns ultimately it will be gnue aware and take advantage there of ohh dneighbo: it is used for like phone system in building and stuff right? systems, buildings neilt (neilt@user-2ivekii.dialup.mindspring.com) joined #gnuenterprise. #gnuenterprise: mode change '+o neilt' by ChanServ!s@ChanServ hello all neilt (neilt@user-2ivekii.dialup.mindspring.com) left irc: Ping timeout for neilt[user-2ivekii.dialup.mindspring.com] can anyone point me to a form that works ? mhm GFClient is crashing on me... :( i think i dont have something configured correctly.. DB000: AttributeError: 'None' object has no attribute 'setLoginHandler' is jamest or dneighbo here? they are some forms ppl Action: chillywilly hasn't played with forms enough yet what are you doing exactly? just trying to make it run.... with a form? yes which one? one in the samples dir? yes, helloworld DB000: Traceback (most recent call last): DB000: File "/home/edmz/code/gnue/gnuef/.cvsdevelbase/gnue/forms/GFClient.py", line 160, in ? DB000: GFClient().run() DB000: File "/home/edmz/code/gnue/gnuef/.cvsdevelbase/gnue/forms/GFClient.py", line 118, in run DB000: self.runForm(formfile, self.disableSplash) DB000: File "/home/edmz/code/gnue/gnuef/.cvsdevelbase/gnue/forms/GFClient.py", line 135, in runForm DB000: self.getConnectionManager().setLoginHandler(self._ui.UILoginHandler()) DB000: AttributeError: 'None' object has no attribute 'setLoginHandler does it work for you ? umm lemme try ah I gotta grab dustutils distutils Action: chillywilly does apt-get install python-distutils http://www.mjq.net/misc/kettle.jpg this is funny man my connection to my debian mirror has sucked lately yay um you using cvs? installing distutils now how did you insteall forms? did you create a gnue.conf file? Action: dneighbo needs more data probably not doesn't it make one for you? depending on if jamest fixed cvs you gotta install common first right ? non dataaware forms might be broken you need to get gnue-common and gnuef the ./setup.py devel (if you want to run cvs version) this will create simlinks in /usr/local/bin for gfcvs it doesn't liek devel and create files in /usr/local/gnue/etc/ a sample.gnue.conf you have to copy that to gnue.conf in same dir this should give you a working forms why is th emodule called gnuef instead of gfclient I don't think clean works anymore ./setup.py clean mdean (mdean@arc10x100.kcnet.com) joined #gnuenterprise. neilt (neilt@p199.usnyc3.stsn.com) joined #gnuenterprise. #gnuenterprise: mode change '+o neilt' by ChanServ!s@ChanServ chillywilly it should be called forms os you would have gnue/forms gnue/common gnue/config gnue/designer etc neilt any word back from teh publisher fellow? :( why even use a gnue dir? chillywilly: so i can go to one dir and do cvs update -Pd and not have to type it 5 times what doe sthat connections file do anyway? how do you build common from cvs? ./setup.py build? mdean (mdean@arc10x100.kcnet.com) left irc: Ping timeout for mdean[arc10x100.kcnet.com] mdean (mdean@arc11x111.kcnet.com) joined #gnuenterprise. chillywilly you dont at least you dont if you are running from cvs just build forms by doing ./setup.py devel neilt: amen to that :) chillywilly connections files lets you define a database connection 1 time and reuse over and over and over in forms so each form you dont have to define all the database info mdean (mdean@arc11x111.kcnet.com) left irc: Ping timeout for mdean[arc11x111.kcnet.com] you just define the name of the record in connections.conf er pg.connections or whatever its callled Action: dneighbo has sloppy memory right now mdean (mdean@arc11x103.kcnet.com) joined #gnuenterprise. be fore warned i believe its 'case sensitive' amen nothing you can setup cvs to have a module just include others like gnue &forms &designer.... mdean (mdean@arc11x103.kcnet.com) left irc: Ping timeout for mdean[arc11x103.kcnet.com] if you don't build common then why doe sit got a setup.py? bah I need wxpython :P it has setup.py for those INSTALLING It not running cvs ok btw: you are evil no I am not i HATE projects that dont do cvs properly i.e. i would be pissed if i di d so all of gnoem should be in one dir? co gnue-forms and gnome it dumped it all in root area i a was in bah, you musyt not ever use branches often then chillywilly i dont use them often but i wouldnt point to gnome for how to do 'anything' cept argue piss moan and fight amongst yourself while your product turns to shite (as the aussies would say) ;) ok so how about dotgnu then dont know not a cvs wizard but I KNOW we all agree gnue/gnue-something is pretty redudant! oh crud 12MB er redundant even whatever what do you mean whaterver what call it gnue-forms if its in GNUE cvs? Building Dependency Tree... Done The following NEW packages will be installed: libwxgtk2.2 libwxgtk2.2-contrib libwxgtk2.2-contrib-dev libwxgtk2.2-dbg libwxgtk2.2-dev libwxgtk2.2-python wxwin2.2-doc wxwin2.2-examples wxwin2.2-headers wxwin2.2-i18n 0 packages upgraded, 10 newly installed, 0 to remove and 199 not upgraded. Need to get 12.3MB of archives. After unpacking 42.0MB will be used. no I agree you shouldn't call it gnue-forms I just no that this is pointless anyway arguing with you ahhh, let me weigh in here we arent arguing we are discussing and its not pointless gnue-config is there on purpose no, this conversatio is over im saying, simply gnue/gnue-somethign is stupid neilt= i told you to do that as the other items were doing that nickr (nick@63.163.68.113) joined #gnuenterprise. i was told it would be changed later and it will :) but after 2 days no-one came up with better way of doing it but i am interested in hearing to something that makes things easier gnue/something nickr (nick@63.163.68.113) left irc: Killed (NickServ (This nick is reserved by another user)) makes sense to me hierarcharally so how do we designate something that is not a tool and applies to all of gnue but it it makes cvs a bitch then maybe there is better way like forms designer config etc and they are aliased to a module called gnue so if do co gnue it gets all the modules if we just called it config i have never seen this only the reverse so that you could have then we would all be asking the question config what? gnue/forms gnue/designer etc and then make alias so you coudl do co forms and it woudl get gnue/forms neilt i think we were going to call it base instead of config but there were other issues we have a base as far as im concerned this stuff is not done well I like aliasing it like this in CVSROOT/modules have gnue &gfclient &geas &designer &..... to diff between applications adn tools and once we get to applications we need to call it something that means GNUe Business Applications Stuff the whole issue of templates re surfaces so i am happy to change it but that was too long GEBAS lol templates? wassat? yes we need to change it but much more pressing issues on the horizon we need tools we can build apps with then we cna fuss on such things :) along with that a smart update tool what are templates? chillywilly please dont go there :) things that specify gcds, gfds, grds, etc.? I wann know wanna its not defined plz tell me sacha (sacha@chime-a-96.conceptual.net.au) joined #gnuenterprise. in a nutshell a way to customize applications what do you see them as then? nickr (nick@63.163.68.113) joined #gnuenterprise. Nick change: sacha -> Sacha dneighbo: it is defined, RTFM cw so in a module you might have gfd,gcd,grd etc neilt: thanks, you're so helpful i see them as ways to do vertical industry stuff your welcome but module vs package is kind of muddy and where you could apply a template is kind of muddy chillywilly: http://www.gnuenterprise.org/~neilt/templates/config_templates/t1.html i.e. at package or module level we have discussed and have fairly fleshed out but not until implementation time i think will the real 'a ha' happen ok where we find 'damn we are dumb why didnt we do X' ;) edmz: are you still getting errors? jamest: he is using forms what do you think ;) hardy har har har um thats DB0000: hardy DB0001: har [20:08] MSG541 to you :) heheh dneighbo: you running newest forms cvs dneighbo? nope wize man might tonight if its not broken it was edmz who was getting the errors and i get a hey try this i REALLY need postgres VPN like yesterday VPN? Virtual Private Network? VPN = more work for jamest that would e 'MWJ' oh no he lives :) Action: nickr shamles about is anyone using orbit-python 0.3.0 not for you they areb't aren't bye all dneighbo (dneighbo@mail.libertydistribution.com) left irc: BitchX: the cootiless client Action: neilt brushes off the cold shoulder neilt: i'm not i really want to get rid of the orbit-python on geas driver eh? Sacha (sacha@chime-a-96.conceptual.net.au) left irc: Ping timeout for Sacha[chime-a-96.conceptual.net.au] as IIRC it's a mojor PITA for gnuef/geas communicatoin on win32 Action: nickr places ice on neilt's shoulder Traceback (most recent call last): File "", line 3, in ? AttributeError: 'CORBA' module has no attribute 'load_idl' was goign to work on it tonight until had a few machines at clients melt down on me was trying this try: import CORBA CORBA.load_idl("../../idl/geas.idl") will try and work on in a few minutes that looks very wrong the ../../idl/geas.idl should be gone now Action: jamest dives into cvs tree argh! sandbar! what code are you running? nickr: thanks gnue/geas/examples/python ah, ok test1.py thought it was new geas driver which was way wrong try replacing load_idl with _load_idl neilt: didn't you RTFM! god what's wrong with you! is that even in the manual? doesn't matter just RTFM you luser load_idl is 0.1.0 orb-py, 0.2.0+ use _load_idl IIRC yep which you would've found out if you just RTFM cool that worked chillywilly: there is no manual for orbit-python jezzzzzz bah, you just don't knwo how to RTFS then source == manual sheesh : :P nope, i even looked at the header files did a search one them etc. so your just outa line well ig you were reading gnuecommon src there's a not e in there heh, you've been outta line i count on me being outa line whats your excuse? well I was tryint o make a point which you obviously did not get :) so i'm stuborn now also just don't appreciate ppl saying oh go RTFM as if asking about something requires you to for everything what's GNUe oh go RTFM what are you some kinda idoit yes except that i tried to run the test.py without geas or postgres running and now its hung up Action: chillywilly sighs IIRC the error handling in that thing is um lacking IIRC the andrew wrote a gnuef screen to do geas user maintenance that works well for a demo you might have to go to the premerge tag though to run it jamest: thanks, i cant run gui oh yeah sorry because exPython still does not work they making progress? i am working on python test scripts to excercise geas yes they are just use a real OS like GNU/Linux wxGtk and wxMac both compile or windows, it works on windows, yeah that's the ticket a nice fresh XP install chillywilly: so i have to take away more time from GNue to support the OS rofl neilt: GNU/Linux is only hard for lusers who don't RTFM :P with a mac you dont need the fsck'ng manual it works the way it should neilt: you just to be apple's little bitch so why spend time reading when you could be coding Action: ajmitch wonders if he can buy a copy of windows xp in the shop yet well I am not being serious as a matter of fact why even think, just code sheesh you should go RTFM one of the funniest things I ever did to a mac user lemme drag my floppy to the trash so it will eject now ok? i was the support person for the macs at an old job of mine and I hated them jamest: i understand no manuals the marketing dept people like to give me crap about it as I was the PC/Vax man working on their macs one day I was looking thru an old mac users guide with one of the users apple finally got wise and built something based on Free software, but Mach sucks ass....just try running the Hurd it is a slow POS the manual used two pages with illustrations to describe the process of ejecting the disk chillywilly: that is so much harder than trying to figure out how to umount -x &*^*&^$# hehe she said "2 pages to explain how to eject a disk. What is up with that?" i said "Well, they are dealing with mac users" not if you undertsand the unix way of mounting file systems in the first place I still get a laugh from her reaction to that one heheh btw - I was joking w/ her. I still support macs, and I don't hate them as much as i did lots of improvements over the last few years the Hurd does it even cooler though you just set a passive translator onde and you don't worry about it a file system translator will start up when chaning or accessing things in that dir just count the number of mac and pc and then divide your support hours translator on the noee node you'll find that on a % basis more time is spent with windows than macs every time neilt: I'll agree with that but how much do you spend ona GNU/Linux system once you get it up and configured? you can lock it ina friggin closet if you want chillywilly: it depends provided it is run as a server not as a desktop machine I've working on a linux system now that's made me it's bitch it's a desktop machine decided that it no longer wants to run staroffice try running full color graphic edit, sound movies etc and you have a different reaction now this machine is the one machine that _MUST_ run staroffice fresh install doesn't help, blowing away the user config file doesn't help sacraficing small animals to the staroffice gods doesn't help and ximian gnome has cost me hours of bullshit error tracing which is why I'm purging it from my systems with pure gnome 1.4 and kde installs GNU/Linux is weak on the desktop everyone knows Redmond owns that market...UNIX has always traditionally been a server OS...I am not gonna delude poeple into thinking that a GNU/Linux desktop is ready for prime time ximian sucks in more ways then one debian gnome? jamest: i would not wish staroffice on anyone staroffice is sloooooow you get a free ride for the rest of the day actally KDE isn't too bad for a beginner or someone who wants to use GNU/Linux on the desktop jamest: are you home yes back snack run good idea ;) neilt: I know a guy who was a total mac guy too...even had one of those newtons and worked for LinuxPPC too how do i find out if i am behind a firewall hmmmm um good question kinda hard to tell I'd look for ports that seem to not work ya but I know of no sure fire way of telling wssat? ok, thanks what makes you think you are? i am using hotel high speed connection and wondering if you can scan me that can be done or if i am vunerable if i leave my computer on neilt: you paranoid? always just a little less than jamest though :) Action: chillywilly doesn't even have a firewall up anymore mdean (mdean@arc11x104.kcnet.com) joined #gnuenterprise. jamest: you paranoid? yes so I can h4x0r your box then? :P chillywilly: just ask jamest how long after the new gnue server was up before getting scanned at the new and unknown address mdean (mdean@arc11x104.kcnet.com) left irc: Ping timeout for mdean[arc11x104.kcnet.com] i dont' recall i think it was less than 24 hours and you ranting and raving about the logs however I did recentely open an ftp port for a client wow they get scanned about every other day on that ftp port showing all kinds of activity math used to get hit several times a day now I block it all at the firewall fun neilt: so that hotel has its own LAN then? odd definition of fun there chilly :) yep was sarcastic dude i know....thus the :) at the end hotel lan is rape yes they wanted $10 a day where we were staying ethernet cable runs at 75k yep same here but its also 0.50 per phone call and after 10 call i still could not get an analog connection so the company pays for it its not really a lan, if you go by throughput speed it just has an ethernet interface that runs like a fast modem is there a command that does through put analysis on tcp OS X eh? Action: neilt stop that cw ouch i used to use ethertop puch ouch not found oh you got it so it will beep? yes hehe sorry is there any other way yam turn it off :P so i dont know you are probing me i don't know other program a ctcp version is not "probing" I'm sure they exist I just don't use them now traceroute will show latency IIRC but not throughput isn't there a web site that measures transfer speed IIRC derek uses something to monitor his connection ya there is Nick change: neilt -> neil-away ok dude I built gnuef how do you build it to run form cvs? cool two things unfortunatly ./setup.py build? or something else? ok you manually create /usr/local/gnue/etc/gnue.conf and connections.conf or you install gnue-common k then in gnuef directory you do ./setup.py devel derek (derek@cpe-24-221-112-50.az.sprintbbd.net) joined #gnuenterprise. the only things you need in /usr/local/gnue/etc/ are the gnue.conf and connections.conf IIRC there are samples of each? yip gnue/gnue-common/etc you may also need the images from the images dir k adn those go where? and the easiest thing is to install gnue-common jamest: thats my bad i told him no need to install but its been so long since i 'installed' it :) well he shouldn't need to but I don't think we fixed this issue I just did install it anyway cool the classes it installs won't be referenced then switch to the /usr/local/gnue/etc dir I don't see an examples conf for connections and make the gnue.conf and connection.conf files from the samples it's either sample.pg or sample.mysql oh badly named this is all 2-tier though? and really redundent as you can mix and match db's in the same connections.conf file Action: chillywilly is confused ah even reference the geas stuff in there wouldn't it be better than to have one file then with examples? yes it would shall I make one? fine with me they are in gnue-common/etc right? yip jcater (while spending whole day at airport) started the new install system I need to copy images? which will (hipefully) make these files go away the install should have done that ok, and how are you doing that? k black python voodo magick the new setup system allows us to build config files for install bah you generate them then? then when updates/installs are done the gnue.conf is updated/created as needed or you gonna make a gui config thingy? and if the file is removed then the system will fall back to the install defaults it's not planned as it's a developer thing kinda like adjusting our parser what's wrong with the parser? I am not sure I follow gnuebot (eggdrop@mail.libertydistribution.com) left irc: Ping timeout for gnuebot[mail.libertydistribution.com] gnuebot (eggdrop@mail.libertydistribution.com) joined #gnuenterprise. #gnuenterprise: mode change '+o gnuebot' by ChanServ!s@ChanServ sorry back the setup system is for developers forms runs fine here so we don't really need a gui just ran helloworld2.gfd jamest: i think eventually we will want a gui cool but certainly it should not be 'mandatory' your splash screen is behind the window though I have to give it focus in order to see it chillywilly: that is by design yip on slower systems the splash comes up first splash screen should be in front then go away followed by the login box hmmm faster system pop them up together or very close to it doesn't really matter ai guess it is just eye candy yip chillywilly: splashscreens do that by design isn't it very 2-tier oriented in the menu selections....well I suppose those are all 2-tier examples they were created for applications that took long time to load this way user got feedback app was starting yea even though no form was painted well I got a PIII 450 in here kind of funny now a days but back in 486 win 3.1 days pretty necessary slow = P133 it runs acceptably on a Sun 167MHz system often times the splashscreen hits so quick i dont even see it :) at least enough to read it suggestions for the menu system are welcome a problem i gladly would have though is there a disable splashscreen option in the client? how is menu 2 tier? chillywilly: yes -d er not -d -s i think commit, rollback, very database oriented but i dunno if that matter um those would be same for n tier maybe labeled save cancel instead of commit / rollback a menu could be just a collection of widgets that enable forms within the same window to be more business less geek wouldn't it depend on what you are doing though? yead i think you are mis understanding a bit this is data navagation to make things easy save would be more non db-admin-like true menu etc is something different part of what stan is looking to do for business procs well I don't know wnough abot forms which would be all objects so feel free to ignore me no good questions its been a long time since we saw forms for 'first' time that's a cool picture though the splash screen so its cool to hear first time users comments well I have run it before it was ajusta looong time ago chillywilly: we sent jamest to jamaica to get that splashscreen foto lol pretty nice of us huh actually my camera failed walking down to those falls that was a pic pulled from the net of said falls you took that pic in ajmaca? no jamaca even jamest: tis cause your zoom lens isnt expensive enough ;) aw darn jamaica whatever i took 1 nice camera, one cheep camera, and 2 disposables to jamaica the nice one batteries died thnik I really care about spelling in IRC :P the cheap one broke walking down to those waterfalls chilly chill and most my pics ended up comming from the disposables :( Mr_You: I am chilled you can really tell the difference in the pics too :( guess I'll have to go back! :) btw - i agree on commit/rollback save and abort or cancel make more sense I thought commit/rollback is mainly for multiple changes.. ie. you make 10+ changes to a db old oracle sql*forms terminology you made all your changes to all the records in the form then commited them forms works along the same lines I meant changes to like 10-300 records not a single record you can load 30000 records into forms and change any number of them it's also transactional in that if any changes in the form abort gnuef (via the datasources) can rollback the transactions the back end has to support this though cool we had hooks for row level locks in there too in the old drivers not sure they survived the updates is forms more like a GUI db amdin thing or do you see it being more like a real application eventually? forms doesn't do any admin of db it's purely end user well it displays it all as records anc such right? yes it provides an interface to the data.. Mr_You: exactly DB adminstration usually requires more however I have no intention of making glade or VB does it do MDI yet? gnuef is supposed to make interfacing to data lightning fast 90% of the time easily you can make it dance via it's triggers but got to know your kerap for the other 10% mdi should be possible IIRC someone submitted patches is EOM an acronym for something however I must stress that gnuef isn't about mimicing VB or glade end of message maybe bleah jamest: how do you handle displaying business objects...or don't you? did the geas driver ever work? sigh, yeas the geas driver worked what are the issues? objects are just sources of data thus a datasource but wit methods too yip you can make them do stuff what I wanted to do was query a geas server for a list of objects so ideally triggers should call methods right? (this I couldn't do) hmmm then the designer would let you paint the form with them the methods would map to trigger namespace you should use som search criteria shouldn't you? datasource.method() type setup in the trigger I am just tryin to picture how they would work together you're talking instances of objects I think oooooh I'm talking object classes you want introspection type stuff so you can define the UI to the objects via designer? bingo I see deos designer work? when the geas drivers first were created I don't think geas could do this can I make some forms? now i can (right?) with it ummmm you can use designer I am nnot sure what interface sis there however it's a 2 part process I think there's one that wil let you I'd have to look at the idl files ok I'm all ears the wizards are not working in designer nor the introspectoin of databases I just wanna build an app that'll let me manage a TRIO program ;) so you can either start w/ an empty form or use the script in the designer directory GFDesigner-old.py how does that work? which will create a basic .gfd file from a database hmmmm after you supply a talbe name but no forms form business objects? yet no, we have no way to get the data we needed at the time IIRC brb after you have the base gfd file you can do all the work in gfdes edit triggers, add remove widgets, etc, etc it's a pretty cool system you can go from a table to a working app in about 90 seconds (with practice) back derek is a master at this k what is missing for me are widgets and menus but I can't take a business object and build a UI to it? right now anyway he did demos on the LWE floor with people kewl and altered the system on the fly and they were amazed as how quick things could be built altered how/ ? added fields to the table oooh with designer popped open designer, tacked on the fields, adjusted layout you can manipulate the db then eh? no db manip that's planned just adding fileds that were not shown then? well you need to manipulate business objects too right? alter table add (foo varchar(10)) then into designer add the field save it run the app and you're there and for object you would use the introspection interface yes, long term is to be able to build biz objs in designer objects yes btw - in gnuef terms the instances of the objects become records well that's how they are stored too right instead of use sql you get to play with objects in the form instead right? most of the work is in place in one form or another to make these additions fairly easy forms builders don't need to know sql lemme build desgner and object designers will eventually be able to build the gcd in designer /setup.py devel??? ./ but that is no where close to ready yes ./setup.py devel in designer should work I know I'm getting so frustrated i wanted to spend tonight on gnuef/geas am I taking up your time? and install I get mad goat raped by star office, a faulty proxy system, and a new gnuef bug chillywilly: you are not wasting my time ah I'm happy to help get forms working for you do just run it stand alone here or what? gfdes is all you need to type or gfdes formname.gfd with my theme the about dialog isn't big enough to see the whole title themes are kicking our ass I am using th eeazel sawfish theme and the mozilla modern gnome theme it doesn't matter too much they all kick our ass heheh gnuef looks at everything in terms of x,y coords at the character size level heh this way you don't have to deal with pixels at all makes life fun? you say 5 spaces over, 2 lines down ok so help me out how can I make a form the wxpython conversion routines bite I think I should install postgres here if I don't have it installed already I forget if I do mysql would work too or others if it'll save you an install Exception exceptions.TypeError'call of non-function (type None)' when I clicked exit known bug with wxpython oh ok :) bleeding edge rocks! yea! I wanna use bleeding edge glib....dunno if derek will bye it GObject would make life easier I'm sure it would there are some purple mushrooms in our yard does glib 1.2 and 1.3 coexist? don't eat them I think so just install it in a other prefix any timeline for 1.3 dunno gstreamer is using it afaik is 1.3 at all stable? I dunno I guess they aren't using it is derek here omega wants to talk to him he says not tonight but soon 2:15:27] chillywilly: you happen to know if masta is still recording his church services? [22:15:34] I'd like to have a nice object system [22:15:38] afaik, yes [22:15:44] I've finally got hardware+software to do it, would like to talk with him before Sunday if he's around [22:16:18] not tonight though, I'm almost out of here to do some other stuff [22:16:24] oh, ok Action: jamest performs the summoning ritual (plays the tape of a coder being flogged) oh oh oh heheh im here i heard some flogging noises you see that? he finally has hadware and software to record church services but would like your help he's about to split though nickr (nick@63.163.68.113) left irc: Ping timeout for nickr[63.163.68.113] chillywilly: how/where does geas store object instance attribute values (say that 5 times fast :) ummmm in the database :P what exactly do you mean? in memory lemme explain cached objects in gnuef we treat data-aware fields and non-dataaware fields as part of the same record I don't care for our system I think we need one with multiple caches so we can handle transactions I have a nice design here in my design book you RTFM!?!?!?!?!?! that's just wrong! :) uh? or is that RTFDB? there's a whole persistant framework it si very nice design book maps object to a database yip I even took a course CS-4xx something software engineering design er, persistent i.e., the objects could be stored in file too just by extending certain classes files nickr (nick@63.163.68.113) joined #gnuenterprise. althought what I'd really like to get my hands on are the ODMG standards argh! as they have solutions for almost all of what we are doing in GEAS I've read their stds you got their book? that's the CORBA group right? nooo nickr (nick@63.163.68.113) left irc: Killed (NickServ (This nick is reserved by another user)) the other one ODMG object database management group not object management group ah, ok omg is the croba ppl corba whihc you know odmg has standards for middleware technology and OO to relation mappings I swear omg would have a 350 page std for wiping your butt heheh it'd be a very robust, stable butt wiping but by the time you fully understood the process you'd have crapped yourself nickr (nick@63.163.68.113) joined #gnuenterprise. lol rofl jamest: :) :) dude I am still laughing :) aaaaah roflmao ok, that was some nice hardy chuckling jamest: that was a GOOD analogy next time a guy comes in here screaming standards of ebXML, j2ee or the likes (.net) etc i will just say Action: derek prepares to crap himself and you will know what i mean :) why does neilt insit on contering everything I post to the list? Action: chillywilly pokes neil-away onery mood i suppose :) fine I am not posting anymore :P I usually don't anyway gnue reports will work like business fellow thinks real hard about report he wants and reports auto creates it no human intervention right? oh while you are all here is it easy for an admin to change my emila address for the lists or should i go through the painful process of unsubscribing and resubscribing to all of them? derek: hehe derek: we need a telepathic interface jamest: you can whip one up right? only if we abstract it out first k so I can do more than one type of telepathy ok ra3vat (ds@195.239.64.96) left irc: [x]chat jamest: can you hack the mailing lists? that I can't do why not? they still haven't been moved to ash but you asn login via mailman can't you? they are on goats s/asn/can i dunno ya but the web interface for each one is still there right? I got to the admin apges i've never offically got the password and not sure if I recall what it is I don't knwo the passwords aw crud where is goats? AZ ash is your box right? yip hmmm, jade setup goats? yip ok guess I'll have to suffer and do it manually at least I have all the filters how many lists you on all of them I want to know all oh crap ;) well I remember the password it seems kewl is it easy for you to do this? not in this version of mailman I mean do you have alist of subscribers that you can modify I have to unsubscribe you then resubscribe you chillywilly: yes and no its a pain in the ass but probably easier than you doing the alpha version of mailman now has an easy way to do this well I don't want you to have to do it if it is not easy jamest: kewl can jade do it? he's probably not around right? bah I don't wanna cause too much headache maybe I should just make mutt use a hook for gnue lists and use my msoe.edu address if that is possible although the reply-to fiels is not correct field it always wat to send it to the list poster Action: jamest needs to get off ass and finish migration to ash hehe what domains are you handling? gnue.org an gnuenterpruse.org? gnuenterprise gnue.org is NZ k gnuenterprise.org is KS they should both be KS mmhmm with AZ becomming a mirror argh argh argh gnue.org is AZ NZ? you mean AZ right not NZ my name is derek not andrew I'z been searching for da white whale my entire life lol sorry AZ crap crap crap not here plz either place is full of odd people though i hate that not until 7.1 did postgres support joins properly Action: chillywilly hands dereka shovel Action: chillywilly has never used a joid join our little app was pretty simple well I am not going to do this other thing maybe I should do some gnue hacking ;) ya think? of course you should heheh gnue is life, life is gnue I wish maybe some day someone will pay us all to do GNUe jamest: would you hack GNUe full time if you could? depends on the day today...no why? heheh I see seriously though i think so i don't know 100% sure that'd I'd want it to be my only job but don't you get that programmer's high when you fix those bugs? as it sits now if I get burnt out I can take a break when it earns your pay you can't do that so you take on another project too ;) I got some for ya http://sourceforge.net/projects/scrappy, http://savannah.gnu.org/projects/hurd ;P :) chillywilly (danielb@d153.as9.nwbl0.wi.voyager.net) left irc: Ping timeout for chillywilly[d153.as9.nwbl0.wi.voyager.net] chillywilly (danielb@d72.as29.nwbl0.wi.voyager.net) joined #gnuenterprise. well that was fun I don't think there is any introspection stuff :( well hold up I might be wrong connection.ild er connection.idl // Schema related methods/attributes /** * A list of the names of classes currently visible to the user on * the server. */ readonly attribute classnames classes; /** * Full information on all classes available to the user. * (Not yet implemented.) */ readonly attribute classlist schema; you cann learn its methods tough or attributes maybe I should fix that yes you can IIRC says it is not implemented there is a function that you can pass a classname too ah, crap I should make some UML diagrams on some of these internal classes that would help oh yea here's something * \brief Class information. * * This method retrieves information on fields in a particular class. * The list does not include inherited fields, but does name the * parent classes. * \sa GEAS::Connection::getFullClassDefinition */ ClassDefinition getClassDefinition( in string name ) // excludes inherited fields raises(UnknownClass,ServerError); i think you want getFullClassDefinition from the same file IIRC it is all in the connection object as it includes the parent object fields too ya right under that one includes inherited fields there ya go that's all we needed IIRC yip /* did I say you should use this? no, I didn't. so don't. */ hehe ? a comment in the code connection.h ah ObjectData *xyz; /* sensible variable names? this is already a nasty hack, so why start being sensible now? */ this is just entertaining who knew reading code could be so fun? nite jamest (jamest@fh-dialup-202033.flinthills.com) left irc: [x]chat