[00:18] Last message repeated 1 time(s).
Mr_You (rouzer@mankind.boredom.org) left irc: night
<chillywilly> anyone home?
<chillywilly> how can I make an entry field contain the data of the field?
Action: dtm debates answering chillywilly because dtm knows he can't help chillywilly with his thingy
Action: dtm delays making his presence known out of cowardice
rek2 (BBK@66.92.76.239) joined #gnuenterprise.
<rek2> hIL
<rek2> HI
<chillywilly> hello
<rek2> what is this chat room about?
<rek2> GNU?
<rek2> what is gnue?
<chillywilly> GNU Enterprise
<chillywilly> http://www.gnuenterprise.org
<chillywilly> an enterprise application framework
<rek2> -but what is about? is a company? a project?
<chillywilly> a project
<rek2> cool
<chillywilly> go check out our site
<chillywilly> ;)
<rek2> and it have any relation with the real gnu?
<chillywilly> yep
<chillywilly> it is part fo the GNU project
<chillywilly> s/fo/of
<rek2> ohh wait I see is an application
<rek2> I get it
<rek2> now
<rek2> sorry
<chillywilly> we are actaully a meta project
<rek2> I saw GNU and thought it was a room about GNU
<chillywilly> #gnu is here
Action: chillywilly occupies that channel
<chillywilly> ;)
<rek2> what is a meta project(sorry I am not a english speaker)
<chillywilly> just a project that encompasses all sorts of software projects
<chillywilly> liek GNOME
<chillywilly> like
<chillywilly> we specialize in enterprise tools and applications
<rek2> I nsee
<rek2> well I will go better to the GNU channel :D
rek2 (BBK@66.92.76.239) left #gnuenterprise (lilo: KVIrc).
<derek> chillywilly: if you are trying to post null data to a field that is 'not null'
<derek> you have one of two problems
<derek> a. silly user
<derek> b. poor table design
<chillywilly> well why does it not populta eit with th efield data?
<derek> if you really want to store blanks then make the field allow nulls
<chillywilly> it should have the book title in there
<derek> if you dont want blanks
<chillywilly> it is not enpty
<chillywilly> so what did I gooof?
<derek> you said course1
<derek> or something how is that a book title
<chillywilly> oh yea
<chillywilly> nevermind
<chillywilly> still
<chillywilly> my book title isn't in there
<chillywilly> btw
<derek> ?
<chillywilly> the compand.gfd does not work
<chillywilly> weh thwe form starts up
<derek> what happens on the company.gfd
<chillywilly> uuuuuuh
<chillywilly> hold on
<chillywilly> I do have states and zips in there too
<chillywilly> so it should auto-fill the city when the trigger fires
<chillywilly> zipcode 12345
<chillywilly> exceptions.AttributeError in trigger code, value: createResultSet
<chillywilly> and to prove I have data
<derek> i believe you
<chillywilly> gnue=# select * from zipcode;
<chillywilly>   zip  |   city    | state_code 
<chillywilly> -------+-----------+------------
<chillywilly>  12345 | ALABAMA   | AL
<chillywilly>  12435 | ALASKA    | AK
<chillywilly>  53218 | WISCONSIN | WI
<chillywilly>  98745 | ARIZONA   | AZ
<chillywilly> (4 rows)
<derek> dont know if company.gfd is current to work with 'new triggers'
<chillywilly> this is from 0.0.9 release
<chillywilly> cvs is too broken to use right now
<chillywilly> imho
<derek> ah
<derek> ok with 0.0.9 it works
<derek> your doing somethign wrong i think :)
<chillywilly> well why dies it not populate my city?
<chillywilly> entry field
<derek> maybe its on strike?
<chillywilly> it should pull albama in there right?
<chillywilly> I think it is pre-focus-out
<chillywilly>             <trigger type ="Pre-FocusOut">
<chillywilly> print "zipcode",self._value
<chillywilly> #ziplookup.clear()
<chillywilly> #ziplookup.setField(0,"zipcode",self.value,TRUE)
<chillywilly> rs = ziplookup.createResultSet({"zipcode":self.value})
<chillywilly> rs.getFirstRecord()
<chillywilly> state = rs.current.getField("state")
<chillywilly> city  = rs.current.getField("city")
<chillywilly> blckcompany.fields.ntrycity = city
<chillywilly> blckcompany.fields.ntrystate = state
<derek> um
<chillywilly>       </trigger>
<chillywilly>       <label text="Zip" x="31" y="6"/>
<chillywilly>       <entry name="ntryzip" field="zip" x="31" y="7" width="10" max_length="10" uppercase="">
<chillywilly>             <options>
<chillywilly>               <tip>Please enter the value for zip code.</tip>
<derek> your table is wrong for starters
<chillywilly>             </options>
<chillywilly>             <trigger type ="Pre-FocusOut">
<chillywilly> print "zipcode",self._value
<chillywilly> #ziplookup.clear()
<chillywilly> #ziplookup.setField(0,"zipcode",self.value,TRUE)
<chillywilly> rs = ziplookup.createResultSet({"zipcode":self.value})
<chillywilly> rs.getFirstRecord()
<chillywilly> state = rs.current.getField("state")
<chillywilly> city  = rs.current.getField("city")
<chillywilly> blckcompany.fields.ntrycity = city
<chillywilly> blckcompany.fields.ntrystate = state
<chillywilly>       </trigger>
<chillywilly>           </entry>
<chillywilly> why?
<derek> )
<derek> <chillywilly> state = rs.current.getField("state")
<derek> <chillywilly>   zip  |   city    | state_code 
<derek> see a problem with that?
<chillywilly> yep
<derek> the code is looking for state
<chillywilly> but that's th eforms fault
<chillywilly> not mine
<chillywilly> :P
<derek> your field is state code
<chillywilly> that's what your zpicode example sql did
<derek> you didnt use the SQL script
<derek> in cvs
<chillywilly> yes I did
<derek> create table zipcode 
<derek> (
<derek>   zipcode       varchar(5)      PRIMARY KEY,
<derek>   city          varchar(30)     NOT NULL,
<derek>   state         char(2)         NOT NULL REFERENCES state,
<derek>   longitude float4,
<derek>   latitude  float4
<derek> );
<derek> you used wrong one :)
<derek> use 
<chillywilly> create table zipcode
<chillywilly> (
<chillywilly>   zip char(5)      PRIMARY KEY,
<chillywilly>   city varchar(30) NOT NULL,
<chillywilly>   state_code char(2)    NOT NULL REFERENCES state
<chillywilly> );
<derek> gnue/gnuef/samples/location/sql/create_locale.sql
<chillywilly> bah
<derek> i assume you used on in zipcode dir :)
<derek> not location dir
<derek> yes the samples are not well organized and silly
<chillywilly> darn
<derek> i know i know
<chillywilly> my bad
<derek> dont beat me please
<chillywilly> I used the one in zipcode
<chillywilly> but README says the right onwe
<chillywilly> I just skimmed it
<chillywilly> aw fooey
<derek> see at least i documented it :)
<derek> but i agree its misleading if you dont read the docs
<chillywilly> woah
<chillywilly> ok
<chillywilly> lemme try again
<derek> and if you go to http://goats.gnue.org/~jamest/
<derek> there is a flat file you can load into postgres
<derek> that loads like 50k zipcodes
<derek> to make the demo really kick arse
<chillywilly> k
<chillywilly> bah what's the table name for that?
<chillywilly> there are no zipcodes in there?
<derek> huh
<chillywilly> ok
<chillywilly> I am downloading the file
<chillywilly> lemme set it all up right ;)
<chillywilly> then I can marvel at it or something
<chillywilly> :P
<chillywilly> hwo do you make one form pop up another?
<chillywilly> say I want them to be able to search for something
<chillywilly> hmmm I could just put that all one one
<chillywilly> another thing, why would the "title" field not be filled in on a form of mine if the data source is correct?
<chillywilly> <?xml version="1.0"?>
<chillywilly> <!--  GNU Enterprise Designer (0.0.1)
<chillywilly>       Form saved on: 2001-09-08 12:28:10  -->
<chillywilly> <form title="Test">
<chillywilly>   <options>
<chillywilly>     <option name="title">Untitled Form</option>
<chillywilly>     <option name="height">12</option>
<chillywilly>     <option name="width">40</option>
<chillywilly>   </options>
<chillywilly>   <datasource database="trioweb" name="trioweb" table="t_books"/>
<chillywilly>   <page name="Page_1">
<chillywilly>     <block name="Block_1" datasource="trioweb">
<chillywilly>       <label x="1" y="1" width="11" name="Label_1" text="Student ID:"/>
<chillywilly>       <entry x="14" y="1" width="10" field="student_id" name="Entry_1"/>
<chillywilly>       <label x="1" y="3" width="10" name="Label_2" text="Title"/>
[01:37] Last message repeated 1 time(s).
<chillywilly>       <entry x="14" y="3" width="10" field="title" name="Entry_2"/>
<chillywilly>     </block>
<chillywilly>   </page>
<chillywilly> </form>
<chillywilly> it doesn't pull the titles of the books
<chillywilly> trioweb=# select * from t_books;
<chillywilly>  book_id |           title           | student_id | course_number1 | course_number2 | course_number3 | course_number4 | course_number5 | reserved | checked_out | deleted 
<chillywilly> ---------+---------------------------+------------+----------------+----------------+----------------+----------------+----------------+----------+-------------+---------
<chillywilly>        1 | Big Book                  |          0 | CS421          |                |                |                |                | f        | f           | f
<chillywilly>        2 | Big Book                  |          0 | CS421          |                |                |                |                | f        | f           | f
<chillywilly>        3 | Big Book                  |          0 | CS421          |                |                |                |                | f        | f           | f
<chillywilly>        4 | Small Book                |          0 | CS400          |                |                |       
<chillywilly> i realize it will complain about the other fields but when it is displayed it should pull the title into the entery field shouldn't it?
<chillywilly> another thing...is there a way to have a secure connection?
<chillywilly> like say if you were letting people access this form anywhere?
<chillywilly> s/?//
<chillywilly> from
<chillywilly> would you have to hack SSL support into forms or would this be a db driver thing that was supported?
<derek> lookup dialogs are on the TODO list
<derek> until then hit the prepare query button
<derek> then type in the field you wish to search (or multiple field) the string you wish to search for
<derek> and smack the execute query button
<derek> and voila it only brings back records making that query
<chillywilly> and normal ppl are supposed to figure that out?
<derek> so say for title you wanted Smurf's are cool
<chillywilly> ;)
<derek> you can do
<derek> Smur%
<derek> and enter
<derek> um you would be amazed at the power in it
<chillywilly> what is that "%" for?
<derek> its not 'microsloth' but its powerful
<derek> wildcard
<derek> it might be *
<derek> im not sure
<derek> we will have lookup dialogs
<derek> to make it more microslothy
<chillywilly> yu can't program a lookup field that fills the others in?
<chillywilly> or how about a form that pops up another form that has the results?
<derek> not really
<derek> well you could but you would have to know forms internals pretty well
<derek> lookup dialogs will do this
<derek> they just arent done
<chillywilly> bummer
reinhard (rm@62.47.44.46) joined #gnuenterprise.
<chillywilly> see cause one of the services is for TRIO program participants is to allow them to serach through the library of course books and reserve books
<chillywilly> which was pretty easy to do with php
<derek> you can do that
<derek> just not with a popup dialog
<derek> and its easy really
<chillywilly> ok
<derek> wish i coudl demo this for you
<derek> as irc LOOSES a lot
<chillywilly> so how does this "query" thing work?
<derek> its a real query
<chillywilly> you type in the field?
<derek> yes
<chillywilly> and that waht it uses?
<derek> and you can do it in more than one field
<chillywilly> does it replace the filds with the result?
<derek> yes
<chillywilly> uhuh
<derek> and you can get more than one result
<chillywilly> what if you have mor ethen one result?
<derek> so say you went to other
<derek> author
<chillywilly> how does that work/
<derek> and typed in 
<derek> stephen king
<derek> and hit query
<derek> it would bring back all records for stephen king
<derek> and then you use the navigation to roll through them
<chillywilly> k
<chillywilly> you can go back and forth?
<derek> yip
<chillywilly> k
<derek> you get a recordset back
<derek> that is full editable etc
<chillywilly> it is just not dumbed dow though eh?
<chillywilly> down
<derek> its just like a normal recordset
<chillywilly> mmhmm
<derek> it just doesnt work how a microsoft gui user would expect it
<chillywilly> I see
<chillywilly> that inventory demo is pretty functional then?
<derek> anyone who used commandline biz apps for long time wouldnt have much issue with it :)
<chillywilly> will that show me cool stuff?
<derek> no inventory isnt bound to data iirc
<derek> it was a mockup
<chillywilly> but it works for me with real data
<chillywilly> what good are those sql scripts then?
Action: chillywilly is confused
<chillywilly> brb
<derek> only the company.gfd works
<derek> some of the other inventories might work
<derek> been a LONG time
<chillywilly> ok
<chillywilly> [01:52:53] <-- myrddian has quit (BitchX: it adds that extra uNF to your sex life!)
<chillywilly> hehe
<chillywilly> nothin' like a BitchX quit message
<derek> did you guys know that geas is officially in debian?
<derek> and they actively are taking bugs against it?
<derek> anyone know who this is?
<derek> Maurizio Boriani
<LuftHans> He's baux@debian.org, duh ;-)
<derek> rofl
<derek> yes he is
<derek> do you know who that is?
<derek> http://bugs.debian.org/cgi-bin/bugreport.cgi?archive=no\&bug=101883
<derek> is his request for wish list to add geas
<derek> and looks like he did make 0.0.5 and patched it
Action: derek loves when we get stuff done unbeknownst to us
<derek> like the day ibm sent in the db2 driver
<derek> kind of like, kool :)
<derek> http://www.de.debian.org/devel/wnpp/prospective
<derek> woo hoo
<derek> all of the gnu enterprise stuff is in the debian to build queue
<LuftHans> cool, I can start playing with it :)
<LuftHans> maybe I'll even figure out what it does...
<chillywilly> wow
<chillywilly> IBM gave you guys a db2 driver once?
<chillywilly> interesting
<chillywilly> I know baux
<chillywilly> sorta
<LuftHans> derek: Jim is now shooting for Mon or Tue for RMS
<LuftHans> can you double-check with him that he
<ajmitch> well well well...
<LuftHans> 1) will do a large talk for us in Phx
<chillywilly> wow so it is in sid
<LuftHans> 2) is available either Sun, Mon or Tue night
<LuftHans> ?
<chillywilly> fucking awesome
<chillywilly> :P
<derek> chillywilly: did you confirm its in sid?
<derek> you know i dont do sid, so couldnt confirm
<chillywilly> well
<derek> LuftHans: size of audience is not an issue
<chillywilly> I am assuming
<chillywilly> lemme update
<derek> LuftHans: i woudl rather get date and time to propose to him before asking on the other
<derek> i assume he will speak any day he is here
<LuftHans> derek: look at info for /etc/apt/preferences
<derek> huh?
<LuftHans> derek: man apt_preferences, provided your running apt >= 0.5.1, e.g. woody
<ajmitch> [root @ gw ajmitch] apt-cache search enterprise
<ajmitch> geas - GNU Enterprise Application Server
<ajmitch> geas-doc - GNU Enterprise Application Server Docs
<derek> ajmitch: ROCK!
<LuftHans> derek: preferences allows you to update sid, but not actually have any of it be available for install
<derek> how about forms and designer
<derek> LuftHans: cool
<derek> im not too worried
<chillywilly> who uploaded it?
<derek> ajmitch: and chillywilly are sick enuff for sid :)
<derek> so i just ask them :)
<LuftHans> derek: let's double-check that he can speak any of those nights
<chillywilly> heheh
<derek> LuftHans: i will double check after we get clearance for them
<LuftHans> derek: I'd hate to have 500 of our closest friends show up and then find out the host couldn't make it...
Action: ajmitch doesn't see any other gnue stuff in sid yet
<derek> as i dont want to send him multiple mails on similar thing
<derek> im assuming we will know by tomorrow what is available to us?
<LuftHans> derek: OK. Yeah, I'm hoping to have word tomorrow.
<derek> if we know early enough i will call the office and talk with him to confirm so we dont have to wait for email delay
<LuftHans> derek: I'm pushing for ASU on Sun or a CC during the week.
<LuftHans> derek: parking at ASU on a weekday is insane.
<chillywilly> what are you talking about?
<derek> chillywilly: i suckered RMS into coming out to visit me 
<chillywilly> heheh
<chillywilly> how'd you manage that?
<derek> he will be here in two weeks
<LuftHans> cw: we're planning a protest to see if we can get RMS to shave :)
<derek> chillywilly: it has been in the works for about 5 months
<chillywilly> no way...I like hom all hairy and intimidating like a real GNU
<chillywilly> ;)
<chillywilly> what is the occasion?
<LuftHans> cw: but it'll give me beard envy cuz I just chopped about 8 inches last week :)
Action: ajmitch cannot imagine RMS with short hair & no beard
<chillywilly> hehe
<ajmitch> derek: think you could sucker RMS into coming to NZ for a holiday? ;)
<chillywilly> well ajmitch you first need to invent some kinda conference thing so he has an excuse to goto NZ
<chillywilly> ;)
<LuftHans> ajmitch: I might be able to sucker Dave Mandala ( former CTO of LinuxCare and current CTO where I work )
<ajmitch> he can just come for  holiday... ;)
<ajmitch> LuftHans: heh
<LuftHans> ajmitch: He'll be in .au in a few weeks.
<chillywilly> RMS is a cheapo though he won't go there for nothing
<chillywilly> ;)
<derek> LuftHans: not the beard?
<derek> LuftHans: you really chopped it?  tell me you're not in love or something!
<LuftHans> derek: yup, it's pretty short now
<chillywilly> where do you work LuftHans?
<LuftHans> derek: donated it to a soup kitchen. They fed 50 people with the stuff in it ;-)
<LuftHans> cw: DevelopOnline
<derek> chillywilly: he works in the desert
<chillywilly> hmmmmm
<chillywilly> I dunno who that is
<derek> www.developonline.com
<LuftHans> derek: actually, it's a concrete jungle, we gotta go 10 miles to hit a decent desert, 20 miles for something worthwhile :(
Action: derek smacks chillywilly with a trout
<chillywilly> I;m there already
<derek> you got the net boy, use it
<chillywilly> so he's in AZ with you?
<chillywilly> or what?
<LuftHans> cw: we make boards available for development online
<LuftHans> cw: yup, .az.us
<derek> LuftHans: i still want to know which genius decided that black asphalt parking lots everywhere was a good idea?
<ajmitch> derek: you're showing your age, calling chillywilly 'boy' ;)
<derek> guess the guy who was selling land
<derek> ajmitch: go pick on someone old like jamest :)
<ajmitch> heheh
<LuftHans> cw: the goal is to give engineers the ability to do whatever they'd've done if they actually had the board
Action: chillywilly loves embedded systems
<ajmitch> derek: is jamest old enough to be my grandfather, i wonder?
<LuftHans> derek: black asphalt in Phx was probably the idea of whoever made car A/Cs
<chillywilly> I worked at L. S. Research for a while doing wireless embedded stuff
<chillywilly> ajmitch: he's not that old
<chillywilly> he's in his 30s
<LuftHans> derek: that was what the alt fuel bill was trying to model itself after :)
<LuftHans> cw: oooh, cool. Wireless *and* embedded. Swe-eet.
<chillywilly> so you can like develop without having to buy th hardware
<ajmitch> chillywilly: who is the oldest then?
<chillywilly> that's actually kinda cool
<chillywilly> of our GNUe clan?
<ajmitch> chillywilly: yes
<derek> LuftHans: would never work for me
<chillywilly> gotta be jamest...I dunno how old reinhard is though
<chillywilly> ask da masta
<derek> i like to smash things i cant get to work
<chillywilly> he know all
<chillywilly> ;)
<chillywilly> derek: rofl
<derek> so unless they built a 'virtual' trash the hardware simulator
<derek> it jsut wouldnt be the same
Action: ajmitch wonders who is youngest of GNUe slaves
<LuftHans> cw: yup, it's awesome if the hardware is impossible to get
<derek> someday i will post a picture of the nintendo that didnt 'co-operate' as planned
<chillywilly> LuftHans: does it actually use the real board?
<chillywilly> heheh
<chillywilly> man I am gonna kick fitzix when I see him
<LuftHans> derek: junkyard wars :)
<LuftHans> derek: or battlebots :)
<chillywilly> I cannot be cut off this long
<ajmitch> hehe
<chillywilly> w/o email
<ajmitch> chillywilly: just relax
<chillywilly> only things I can get are GNUe mails
<ajmitch> chillywilly: lay back & hack ;)
<LuftHans> cw: yeah, you get access to a controller board with setup for develling for the board you want to work on
<LuftHans> cw: then you have full, unfettered access to the target
<LuftHans> cw: we do cross-compiler setup, etc. if necessary
<chillywilly> aw crud
<chillywilly> a java plugin
Action: chillywilly don't have the jvm installed atm
Action: LuftHans has to start doing java in browser stuff tomorrow :(
<chillywilly> blech
<LuftHans> at least I get an ultra 10 at my desk now :)
<ajmitch> hehe
Action: ajmitch is a mere student
<chillywilly> so is this Free software or what?
Action: derek feels for LuftHans where should i send the funeral arrangment?
Action: derek believes spiritually a developer dies when has to use java
<LuftHans> cw: we would prefer all free software, but that doesn't always work out :(
Action: ajmitch has to use java at uni
Action: chillywilly took Java as an elective a loooong time ago
<chillywilly> bah
<chillywilly> 186MB to upgrade sid
Action: ajmitch would prefer C, or even C++
<chillywilly> plz shoot me now
<ajmitch> my sid install is up to date
<LuftHans> cw: for GNU/Linux[1] we're debian for all the x86 and hard hat for PPC
<LuftHans> [1] see Derek, I'm practicing ;-)
<ajmitch> hehe
<chillywilly> heheh
Action: chillywilly is also a member of the GNU/Linux police
<chillywilly> danielb@obfuscation:~/src/gnue/gnuef/samples/location$ apt-cache search geas
<chillywilly> geas - GNU Enterprise Application Server
<chillywilly> geas-doc - GNU Enterprise Application Server Docs
<LuftHans> cw: I'm not religious about it, but it goes with the discussion about slashcode from this morning
<chillywilly> woohoo
<chillywilly> about what?
<chillywilly> Free software?
<chillywilly> the naming thing?
<LuftHans> cw: RMS and GNU deserve credit for what they've made possible
<ajmitch> chillywilly: bah, i pasted that ages ago
<chillywilly> so what
<LuftHans> cw: also RedHat Linux Hurd 7.5 sounds stupid ;-)
<chillywilly> it is now confirmed
<derek> that and hans could have RMS as a stay in guest for nearly a week
<derek> so he needs the practice
<chillywilly> heheh
<ajmitch> chillywilly: you're 20min slower than me :P
<ajmitch> haha
<chillywilly> dude I updated first to make sure
<chillywilly> so sue me
<derek> chillywilly: can you do a search for apt-cache gnue-designer
<chillywilly> yep
<derek> and forms
<chillywilly> not there
Action: ajmitch now goes to update his other computer
<derek> btw: real men do apt-cache | grep foo
<chillywilly> no forms either
<LuftHans> derek: apt-cache search gnue-designer, then we can cut and paste :)
<chillywilly> wwrong
<chillywilly> real men do dpkg -l *gnue*
<derek> ok you win
<chillywilly> yay!
Action: chillywilly does back flips
Action: chillywilly back is now broken
<ajmitch> chillywilly: cos your evil plans are coming along?
<LuftHans> cw: grep gnue /var/lib/apt/lists/* # ;-)
<reinhard> hey what is this?
<reinhard> gnue is in debian?
<ajmitch> chillywilly: 'apt-cache search scrappy'  shows nothing :P
<chillywilly> grrrr
<chillywilly> no thanks to you
<chillywilly> reinhard: geas 0.0.5 is in sid
<chillywilly> geas-doc too
<chillywilly> we're special or something
<reinhard> man
<reinhard> this is getting out of control
<reinhard> ;)
<chillywilly> lol
<reinhard> btw
<reinhard> iirc i am older than jamest
<chillywilly> hehe
<reinhard> like 3 months or so
<chillywilly> bah
<chillywilly> that doesn't count
<ajmitch> hehe
<chillywilly> 3 months is nothin
<reinhard> and i think neilt is older than both of us
<chillywilly> you're both old men
<chillywilly> great
<reinhard> jamest and me are both 31
<ajmitch> who is youngest?
<chillywilly> now I can really give him shit
<reinhard> and neilt has a grandson
<reinhard> so he is probably older than 31
Action: chillywilly hides
<reinhard> youngest?
<LuftHans> reinhard: or his family reproduces quickly :)
<reinhard> i guess bigbrother
<reinhard> ;)
<derek> reinhard: yes you are older than jamest but younger than neilt
<reinhard> LuftHans: :)
<chillywilly> hey my mom was a grandma at 39 or so
<chillywilly> ohw ait
<chillywilly> 38
Action: ajmitch feels young now...
<chillywilly> she wasa  grandma at 38
<derek> LuftHans: reinhard is austrian and speaks german fairly fluently :)
<derek> if you ever feel the urge to keep your german up
<ajmitch> i'm just a young'un compared to you oldies ;)
<derek> we have ruined him though, as he now gets american pop culture humor better than most of us now
Action: ajmitch will bbl
<derek> reinhard: i still remember the first day you came to irc and i think jade told jamest he was smoking crack
<ajmitch> cya
<LuftHans> cw: that's not bad. At least one generation was old enough to reach high school first :)
<derek> and you were not so sure about that :)
<chillywilly> sheesh I should go to bed or something
<chillywilly> heheh
<LuftHans> derek: ami pop culture? Ouch. Hurts the brain to bend over that far...
<chillywilly> crack is soo good
<LuftHans> cw: why did you have to say that right after I mentioned bending over?
Action: chillywilly passes the crack pipe around
<chillywilly> lol
<derek> LuftHans: roflmao
<chillywilly> oooops
<chillywilly> nah I dun wanna see your crack
<chillywilly> :P
<derek> well its 1am so i suppose i shoudl get some sleep
<derek> see you all manana
Nick change: derek -> dnZzzz
<chillywilly> manana
<chillywilly> wtf is manana?!?
<chillywilly> that  must be something old ppl say
<chillywilly> that explains a lot though
<chillywilly> neilt probably needs to take more naps
<chillywilly> being a granpa and all
<chillywilly> ;o
<LuftHans> cw: nana is grandma, so I guess it's "my grandma"
<chillywilly> I guess...
<chillywilly> hehe
<LuftHans> derek: I'll keep sending you info as I get it
<chillywilly> sheesh wazz wrong witcha reinhard didn't you rtfm
<reinhard> chillywilly: huh?
<chillywilly> you remember what neilt told me the other day?
<chillywilly> then I was messing with him for the last couple minutes asking him why he didn't rtfm
<chillywilly> eh, nevermind....it was just my lame attempt at being funny
<chillywilly> granpa neil
<chillywilly> heh
Action: chillywilly should go play medievia now
<chillywilly> or whatever that MUD is called
Action: LuftHans muss ins Bett
<chillywilly> eh?
<reinhard> schlaf gut!
<chillywilly> yea what he said
<LuftHans> rh: danke
LuftHans (lufthans@cpe-24-221-1-115.az.sprintbbd.net) left irc: der.hans schlaeft ein
chillywilly (danielb@d48.as15.nwbl0.wi.voyager.net) left irc: Ping timeout for chillywilly[d48.as15.nwbl0.wi.voyager.net]
chillywilly (danielb@d48.as15.nwbl0.wi.voyager.net) joined #gnuenterprise.
chillywilly (danielb@d48.as15.nwbl0.wi.voyager.net) left irc: Read error to chillywilly[d48.as15.nwbl0.wi.voyager.net]: EOF from client
ajmitch (ajmitch@p3-max1.dun.ihug.co.nz) left irc: Ping timeout for ajmitch[p3-max1.dun.ihug.co.nz]
ajmitch (ajmitch@p15-max7.dun.ihug.co.nz) joined #gnuenterprise.
neilt (neilt@dialup-63.208.66.199.Dial1.Chicago1.Level3.net) joined #gnuenterprise.
#gnuenterprise: mode change '+o neilt' by ChanServ!s@ChanServ
<neilt> hello all
<reinhard> goooood mornin'
<neilt> reinhard: i started playing with the query idl
<neilt> and it appears to be really slow
<neilt> i made two routines for reading the help text in schema.py
<neilt> if you switch between them in the code you will notice a considerable difference in the response time
<neilt> that is bad
<reinhard> how do they differ?
Action: reinhard thinks "rtfc"
<neilt> one uses a query object
<neilt> and one just does
<neilt> loadSingleObject()
<neilt> they are right at the top of the code
<reinhard> oh ah
Action: reinhard is dumb
<reinhard> maybe i should update cvs before looking at the code :)
<neilt> in running this program i notice the difference in response time betweent the two
<neilt> yes, i checked in last night
<reinhard>  # this function is at least one order of magnitude slower
<reinhard> does this mean
<reinhard> factor 10 or so?
<neilt> i would say 100 to 1
<neilt> if i notice it as a user
Action: reinhard is speechless
<reinhard> this is sick
<neilt> i was being kind in the comments
<neilt> if you change
<neilt> def show_help_info( name, field):
<neilt>    obj = get_help_object1( name, field )
<neilt> between get_help_object1 and get_help_object2 and run the program you will see
<reinhard> neilt: which one was the class you added help
<reinhard> ?
<neilt> of course, the reason for my comments is that i may have something screwed up
<neilt> reinhard: dont understand your question
<reinhard> me neither :)
<reinhard> sorry
<reinhard> i have to load the help table into the db first before i try it of course :)
<neilt>     ./parse_load_sql.py ../../gnue-config
<reinhard> just a sec
<neilt> from gnue/geas/tools i believe
<reinhard> found something els
<reinhard> have played too much with inactive code (src/classdef) instead of active code
<reinhard> must rebuild geas first :)
<neilt> :)
sacha (sacha@chime-a-105.conceptual.net.au) joined #gnuenterprise.
<reinhard> neilt: which class do you use for test of schema.py
<reinhard> ?
sacha (sacha@chime-a-105.conceptual.net.au) left irc: Killed (NickServ (This nick is reserved by another user))
sacha_ (sacha@chime-a-105.conceptual.net.au) joined #gnuenterprise.
Nick change: sacha_ -> SachaS
<neilt> back
<neilt> 17 person::comm_type
ajmitch (ajmitch@p15-max7.dun.ihug.co.nz) left irc: Ping timeout for ajmitch[p15-max7.dun.ihug.co.nz]
<neilt> reinhard: did  you change the make file
<neilt> configure is broken
<neilt> for me
<neilt> checking local Python configuration... ./configure: no such file or directory: /usr/local/bin/python [6023]
<reinhard> neilt: shouldn't have changed
<reinhard> you switched to automake 1.5 recently?
<neilt> not that i know of
<neilt> how do i check
SachaS (sacha@chime-a-105.conceptual.net.au) left irc: Ping timeout for SachaS[chime-a-105.conceptual.net.au]
<reinhard> or updated your python install?
<reinhard> automake --version
<neilt> automake (GNU automake) 1.4-p2
<reinhard> you have python in /usr/bin or in /usr/local/bin?
<neilt> and i have not update python recently
<neilt> its at /Volumes/AnagadaII/sw/bin/python
<neilt> where it has always been
<neilt>  and i configure with ./autogen.sh --with-postgresql --enable-debug --libdir=/Volumes/AnagadaII/sw/lib --includedir=/Volumes/AnagadaII/sw/include |& tee autogen.logfile
<reinhard> neilt: what was the last line before the error?
<neilt> checking for python... (cached) /usr/local/bin/python
<neilt> checking local Python configuration... ./configure: no such file or directory: /usr/local/bin/python [6023]
<reinhard> neilt: ok
<reinhard> please remove config.cache
<reinhard> not sure where you got that file from
<reinhard> it is not in cvs
<reinhard> in this file configure caches several test results
<reinhard> it has "remembered" in this file that python is in /usr/local/bin/python on your machine
<reinhard> which is incorrect, so not sure how this info got into that file.
SachaS (sacha@chime-a-85.conceptual.net.au) joined #gnuenterprise.
<neilt> ?? i am real confused because that worked
<neilt> i normally remove config.cache before i run autogen, but forgot this time
<neilt> it works thanks
<reinhard> ok
<reinhard> neilt: parse_load_sql.py only makes sql scripts, yes?
<reinhard> and i must still load them into database manually?
<neilt> no it loads them also
<reinhard> into which database?
<neilt> the default
<reinhard> ah
<neilt> whatever happens when you type "psql" with no args
<reinhard> ok
ajmitch (ajmitch@p40-max2.dun.ihug.co.nz) joined #gnuenterprise.
<neilt> if that is not right we can modify the script to connect ot the correct db
<reinhard> nevermind
<reinhard> i can change the database for now :)
<neilt> its real simple to change
<neilt> one line
<neilt> the script uses psql to load the data
<neilt> so you just change the command from "psql" to 'psql -U name' or whatever
<reinhard> neilt: it's more than 1:100
<reinhard> about 1:200 from what i can tell
<reinhard> and i am quite sure
<reinhard> a. it's not ORBit's fault
<reinhard> b. it's not because postgres is so slow
<reinhard> i would say the query is unusable in it's current state
<neilt> :(
<neilt> the idl seems to marshall a lot of objects?
<neilt> each field is an object?
<reinhard> i would have to look what eats the time
<neilt> ok, at least we have an example
jamest (jamest@hobbes.math.ksu.edu) joined #gnuenterprise.
#gnuenterprise: mode change '+o jamest' by ChanServ!s@ChanServ
<reinhard> neilt: a good one
<neilt> i think the query api is crap
<neilt> so i am up to changing it when we know what is slowing it down
<reinhard> neilt: i think parts of the slowness are
<reinhard> the temporary tables geas creates in the db during a query
<reinhard> geas::searchcriteria
<reinhard> like the "where" clause could be so big it don't fit in memory...
<neilt> ouchs
<neilt> does not fit im memory
<reinhard> sorry
<jamest> i thought geas constructed a normal sql statement for query?
<reinhard> yes
<reinhard> but the fields of the condition
<reinhard> and the compare value etc
<reinhard> are written in a db table temporarly
<reinhard> while building the sql statement
<jamest> why?
Action: jamest is just trying to learn
<reinhard> maybe andrewm was afraid there could be so many conditions
<reinhard> that they won't fit in memory :)
<jamest> Sheboing!
<jamest> That'd be one huge sql statement :)
<reinhard> jamest: seriously i don't understand
<jamest> not be mention what that'd mean to the DB server :)
<reinhard> lol
<jamest> well, I still plan on diging into the db python thing 
<reinhard> i think we all agree we must change this
<jamest> tonight if all goes well
<reinhard> neilt: we can't argue against python anymore with the argument of performance when we do such nonsense on the other hand :)
<jamest> lol
<jamest> man, get this
<reinhard> i am seriously considering a bottom up rewrite of geas
<jamest> rms agreed to do an interview w/ a magazine named "Computer Magazine" in april
<jamest> the report promised that he'd use GNU/Linux in the article which is why rms agreed to it
<jamest> report = reporter
<jamest> the magazine editor stripped the GNU/Linux prior to print and replaced with Linux
<jamest> typically I'm not one to pick nits
<jamest> but in this case I can see why rms is upset
<jamest> :)
<jamest> reinhard: are you serious?
<reinhard> jamest: btw may i guess whether you compiled geas over the weekend or not?
<neilt> reinhard: :)
<jamest> reinhard: I did not my work on ash took all my time, something is seriously wrong with dns system on it
<reinhard> jamest: nevermind, just teasing :)
<jamest> however the solaris build I'm doing today 
<jamest> the geas driver is tonight
<jamest> so what is making you want to rewrite geas from scratch?  I thought this was discussed in the past and peope thought there was enough good in there that clean up was all that was needed?
<reinhard> people == reinhard
<reinhard> people sometimes change their minds
<reinhard> :)
<jamest> lol
Action: neilt keeps quiet
<reinhard> current geas has 31651 lines of code
<reinhard> if i counted correctly ;)
<reinhard> i am rewriting gcd parser and classdef library as it is really necessary
<jamest> where is most of the code utilized?
<reinhard> sorry phone call
<reinhard> back
<reinhard> we have agreed that we need new method handling
<reinhard> we already talk about new/changed/integrated/better/etc. database interface
<reinhard> these 3 make up 15600 lines
<reinhard> which is just about 50% anyway
<reinhard> the rest of the code is highly orbit dependent
<jamest> gack!
<reinhard> and if we want to make geas multi-orb
<jamest> you're starting to talk like those forms developers
<jamest> always rewriting stuff
<jamest> :)
<reinhard> it's necessary to rewrite that rest anyway...
<reinhard> the good thing we have (and we should use) is
<reinhard> 1. we can use the "old" geas parts as a test bed for the rewritten modules
<reinhard> 2. we can of course use the old code for a source of good ideas
<reinhard> (as well as for a source of hints how to not do it ;) 
<reinhard> well that's my personal opinion anyway
<reinhard> not sure what neilt thinks about this
<reinhard> or if you are already /msg'ing with each other about me ;)
<neilt> agree 200%
<reinhard> but if we do that
<reinhard> arrg phone again
<jamest> neilt: agree with ?
<neilt> agree with reinhard
<jamest> about a rewrite?
<neilt> about the need
<neilt> for rewrite
<neilt> but GEAS functions
<neilt> so we can still start using it
<jamest> soudns like forms again
<jamest> it works but needs work :)
<jamest> argh!
<jamest> libuuid doesn't install by default on e2fsprogs
<neilt> mine did
<jamest> how did you ./configure ?
<neilt> yes, your right
<neilt> you have todo make install :)
<neilt> yes i had to install from source
<neilt> configure - make - make install
<reinhard> back
<reinhard> we need to provide libuuid source with geas
Action: jamest needs to read instructions
<neilt> iirc i just did ./configure
<reinhard> what i wanted to say is
<neilt> no args
<jamest> make install-libs
<reinhard> if we do rewrite
<jamest> to install the libs :)
<reinhard> we must not make the same mistake as before
<reinhard> we _need_ to look at the source w/ more than 2 eyes
ajmitch (ajmitch@p40-max2.dun.ihug.co.nz) left irc: Ping timeout for ajmitch[p40-max2.dun.ihug.co.nz]
<reinhard> look closely i mean
<reinhard> not just skimming over and think "looks good"
<neilt> thats why we need test suite
<jamest> checking whether to enable sigsegv handler... yes
<jamest> checking Method system to use... ./configure: test: unknown operator ==
<neilt> should build that first
<reinhard> jamest: lol
<reinhard> your shell doesn't support ==?
<jamest> it's tcsh
<neilt> jamest: that could be a problem :)
<neilt> i use tcsh
<neilt> and it worked
<neilt> bb 30 minutes
Nick change: neilt -> neil-away
<reinhard> jamest: if all else fails try sh -c ./configure
<jamest> same thing
<jamest> so how close is geas to a new release?
<reinhard> jamest: i would sy
<reinhard> say
<reinhard> we can release after each part we have rewritten
<reinhard> like next release is
<reinhard> "gcd parser and class definition library is in it's final state"
<reinhard> or something like that
<jamest> what I was wondering is if things have changed since 0.0.6....
<jamest>   we release geas 0.1.0, forms 0.1.0, designer 0.1.0
<jamest>   then make next versions targeted at 0.3.0
<jamest>   then 0.5.0 all the way up to 1.0
<reinhard> jamest: if we want to do linux kernel scheme
<reinhard> even numbers would be stable :)
<jamest> yip
<jamest> thus my avoidance of them :)
<reinhard> and i'm not sure if it makes sense to do even-odd before 1.0
<jamest> ok
<jamest> same error building geas
<reinhard> ah
<reinhard> == error?
<jamest> euler$ ./autogen.sh --with-postgresql     
<jamest> <snip>
<reinhard> can you update cvs please? i just fixed it
<jamest> updating
jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) joined #gnuenterprise.
<jamest> looks like it made it
<jamest> it = the configure script not jcater :)
<jamest> wow 128MB DIMMS for $12.88
<jcater> :)
<jcater> morning
[09:00] Last message repeated 1 time(s).
<jamest> jcater an issue
<jcater> I am not
<jcater> I am a resource
<jamest> postgresql drivers cast the datatypes of the fields based upon db field types
<jamest> the old gnuef drivers recast as strings
ajmitch (ajmitch@p13-max6.dun.ihug.co.nz) joined #gnuenterprise.
<jamest> i added a quick hack to the new postgresql dirver to cast ints as str to verify this was the issue
<jamest> now I want to change it to do cast everything to str
<jamest> but I didn't and still don't care much for this solution
<jcater> that's fine w/me although I would like to add a todo at that point reminding us to come up w/something better
<jamest> IIRC I already have the TODO in the code :)
<jcater> ah
<jcater> you use dcl?
<jamest> nope
<jamest> i keep hearing derek talk about it
<jamest> reinhard....
<jamest> gcc -DHAVE_CONFIG_H -I. -I. -I.. -I/usr/local/encap/glib-1.2.10/include/glib-1.2 -I/usr/local/encap/glib-1.2.10/lib/glib/include -I../lib/classdefs    -g -O2 -Wall -Wmissing-prototypes -Wstrict-prototypes -c geas-oidgen.c
<jamest> /bin/sh ../libtool --mode=link gcc  -g -O2 -Wall -Wmissing-prototypes -Wstrict-prototypes  -o geas-oidgen  geas-oidgen.o -luuid
<jamest> gcc -g -O2 -Wall -Wmissing-prototypes -Wstrict-prototypes -o geas-oidgen geas-oidgen.o -luuid
<jamest> Undefined                       first referenced
<jamest>  symbol                             in file
<jamest> socket                              /usr/local/lib/libuuid.a(gen_uuid.o)
<jamest> ld: fatal: Symbol referencing errors. No output written to geas-oidgen
<jamest> collect2: ld returned 1 exit status
<jamest> make[2]: *** [geas-oidgen] Error 1
<jamest> make[2]: Leaving directory `/home/jamest/cvs/gnue/geas/tools'
<jamest> make[1]: *** [all-recursive] Error 1
<jamest> make[1]: Leaving directory `/home/jamest/cvs/gnue/geas'
<jamest> make: *** [all-recursive-am] Error 2
<reinhard> socket
<reinhard> hmmm...
<reinhard> oh
<reinhard> socket is a bsd extension
<reinhard> cryp
<reinhard> crap
Nick change: neil-away -> neilt
<reinhard> jamest: src/ directory compiled ok??
<jamest> um
<jamest> reinhard: can I send you output of make command?
<reinhard> surely
<jamest> where?
<reinhard> reinhard@gnue.org
<jamest> ok
<jamest> sent
ToyMan (stuq@c5300-1-ip123.albany.thebiz.net) joined #gnuenterprise.
<ToyMan> hi all
<jamest> hey
<ToyMan> starting in on setting up my new BSD box as file/db server today
<ToyMan> so I'll bew ready to do the Gnue thang
<reinhard> jamest: ok you have an own library libsocket
<reinhard> for the socket function which is otherwise in libc
<reinhard> i mean in glibc
<reinhard> just for now please change the line in tools/Makefile.am
<reinhard> geas_oidgen_LDADD = -luuid
<reinhard> to
<reinhard> geas_oidgen_LDADD = -luuid -lsocket
<reinhard> we will need to add some configuration magic to this
SachaS (sacha@chime-a-85.conceptual.net.au) left irc: Ping timeout for SachaS[chime-a-85.conceptual.net.au]
<dnZzzz> jamest did you seen geas is in sid :)
<dnZzzz> i must go
dnZzzz (derek@cpe-24-221-112-50.az.sprintbbd.net) left irc: [x]chat
<neilt> reinhard: you here
<reinhard> yes
<neilt> you should feel free to make comments on the article
<reinhard> yes
<reinhard> i just saw the commit mail
<reinhard> will do this evening
<neilt> cool, thanks
<neilt> its still very rough
<reinhard> my wife goes to cinema this evening without me
<reinhard> so i will have time for gnue :)
<reinhard> neilt: on gcd syntax
<neilt> but i am interested if any major ideas need to be included that are not corrently
<neilt> reinhard: yes
<neilt> i have about 30 minutes
<reinhard> do we want the explicit LIST/REFERENCE/LOOKUP syntax?
<reinhard> yes or no? (1 minute) ;)
<neilt> hmmmmm
<neilt> i think that in order to make objects plug and play
<neilt> we will be using this syntax and not implicit one for some objects
<neilt> only because i can specify a link field that is data driven and not a made up field driven
SachaS (sacha@chime-a-181.conceptual.net.au) joined #gnuenterprise.
<reinhard> ok
<neilt> reinhard: so yes i think we need it
<reinhard> will we keep the old syntax or figure out a new (better rememberable) one?
<reinhard> and do we really need LOOKUP?
<neilt> what does LOOKUP do?
Nick change: neilt -> neil-phone
<reinhard> ok i explain while you phone :)
<reinhard> class customer has a field custno and a filed name
<reinhard> class invoice has a field custno
<reinhard> and can have a field
<reinhard> LOOKUP cust_name : customer (custno).name = custno;
<reinhard> and then cust_name is a virtual field of the invoice object that holds the name of the customer
<reinhard> --
Nick change: neil-phone -> neilt
<neilt> is this a preformance enhancment?
<neilt> i guess my question is why?
<neilt>  where would it be used
<neilt> why not use a reference
<reinhard> neilt: exactly
<neilt> and get the whole record
<reinhard> yes
<reinhard> not sure if it makes sense to talk of performance enhancements for geas ;)
<reinhard> well actually it could be one
<reinhard> get the whole record eats more bandwidth than get only a single field
<reinhard> plus geas could make a join
<neilt> based on what i have seen, we better start talking about performance enhancements :)
<reinhard> neilt: was j/k
<neilt> :)
<reinhard> i don't think i would be _very_ helpful
<neilt> but if i have a reference, dont i do getField() to get a single field in the client or does it always send the whole record to the client?
<reinhard> well
<reinhard> if the interface was sane
<reinhard> then you would tell geas in advance what fields you will want
<reinhard> so geas has 2 possibilities:
<reinhard> a) create a SELECT with all fields and cache them in memory and give you what you ask for bit by bit
<reinhard> or
<reinhard> b) create a new SELECT for every new field you request
<reinhard> both is a performance killer
<reinhard> imho
<reinhard> and if we define what fields we want before we issue the query
<reinhard> we can as well define which references we want, and what fields of those references
<reinhard> which again means we don't need LOOKUP
<neilt> ok, no lookup
<neilt> but its scary to here you use the word "sane"
<neilt> s/here/hear/
<reinhard> huh? why?
<neilt> GNUe != same :)
<neilt> sane
<reinhard> ah ok
<reinhard> lol
<neilt> have to go for 1 hour
<reinhard> just to make sure i looked up "sane" in my dictionary
<neilt> bbl
<neilt> and
<reinhard> and i found out what "SAP" means in american slang :)
<reinhard> bbl
<neilt> ok later
Nick change: neilt -> neil-away
<neil-away> SAP sure fits, lol
<jamest> neil-away: ????
<jamest> SAP?
<reinhard> oh
<reinhard> from e2fs configure.in
<reinhard> dnl
<reinhard> dnl Check to see if -lsocket is required (solaris) to make something
<reinhard> dnl that uses socket() to compile; this is needed for the UUID library
<reinhard> dnl
<reinhard> :)
<jamest> does that mean a similar check needs added to geas?
<reinhard> jamest: i will
<jamest> cool
<jamest> you want me to manually add that switch
<jamest> or wait to test the new configure?
<reinhard> jamest: if you have the time
<reinhard> both :)
<reinhard> first manually add, then when i fixed configure.in test :)
<reinhard> btw geas-server already built
<reinhard> it's just a secondary program that is missing
<reinhard> you can also try to run geas-server
<reinhard> you will probably have to edit geas.conf first
<reinhard> btw jamest would you agree that uuid and orbit are the main issues when building geas?
Mr_You (rouzer@mankind.boredom.org) joined #gnuenterprise.
<jamest> reinhard: i honestly don't recall
Action: reinhard is honestly trying to make geas install-friendly
<jamest> what an odd goal
<jamest> :)
SachaS (sacha@chime-a-181.conceptual.net.au) left #gnuenterprise (Client Exiting).
SachaS (sacha@chime-a-181.conceptual.net.au) joined #gnuenterprise.
ra3vat (ds@195.239.66.46) joined #gnuenterprise.
<ra3vat> hi all
<jamest> hi
<jamest> how goes the i18n?
<ra3vat> did not do any new things last days
<ra3vat> reinstalled machine at home
<jcater> ra3vat: you doing i18n for forms?
Action: jcater has been out of touch
<ra3vat> need to complete and get forms working again
<jcater> rock on
<jamest> he was extending the support in UIwxpython 
<jamest> i believe
<jcater> ah
<ra3vat> jcater: i did few minor fix
<ra3vat> for i18n
<jamest> ra3vat: did you get them checked in?
Nick change: jcater -> jcLunch
<ra3vat> jamest: not yet
dneighbo (dneighbo@mail.libertydistribution.com) joined #gnuenterprise.
<jamest> oooo
<jamest> http://computerworld.com/nlt/1%2C3590%2CNAV47_STO63683_NLTAM%2C00.html   
<jamest> Peoplesoft is getting sued
<jamest> i can't get to the article to find out the details
<dtm> jamest: hmm 
<dtm> i live down the street from their HQ, so i could throw some tomatoes at em
<dtm> while they're down and all
<Mr_You> haha
<Mr_You> their getting sued for making it to costly to program their software
<Mr_You> j/k ;-)
Nick change: neil-away -> neilt
<Mr_You> "you said it would cost $_,___,___ and it cost $__,___,___"
<neilt> reinhard: thanks for the comments
<neilt> i will write more this afternoon on the plane
<neilt> dneighbo: you here?
<dneighbo> yes
<neilt> dneighbo: can you give me feedback on article
<neilt> i am going to do a final push on the plane today
<reinhard> dneighbo: I am in the process of including the uuid source in GEAS so that somebody installing GEAS does not even know it depends on libuuid
<reinhard> libuuid is GPL
<reinhard> what must I do to give proper credit etc?
chillywilly (danielb@d141.as9.nwbl0.wi.voyager.net) joined #gnuenterprise.
<ra3vat> bbl
<chillywilly> wassup dudes?
ra3vat (ds@195.239.66.46) left irc: Client Exiting
<dneighbo> reinhard its always nice to contact the author
<dneighbo> and let them know what you are doing and why
<chillywilly> whadddup g?
<dneighbo> give them credit in the AUTHORS file
<dneighbo> and include thier copyright statement
<dneighbo> in the files where thier code exists
<reinhard> dneighbo: it's seperate files in a seperate subdirectory
<dneighbo> and make note that the non fsf copyright belongs to what segments so its clear
<dneighbo> reinhard copyright exists on every file so that shoudlnt matter
<dneighbo> btw why are we donig this?
<dneighbo> i would rather have the dependency
<reinhard> yeah
<dneighbo> than pull the code into our code
<dneighbo> as then we have to 'maintain' code someone else is willing to maintain
<chillywilly> whatcha talkin about?
<reinhard> i mean
<reinhard> <dneighbo> and include thier copyright statement
<reinhard> <dneighbo> in the files where thier code exists
<reinhard> doesn't apply then, does it?
<dneighbo> ?
<dneighbo> reinhard if you put thier code in foo.c
<dneighbo> then you need to include thier copyright
<reinhard> it's not putting their code into our files
<dneighbo> ok
<reinhard> it's just to include libuuid as a seperate directory
<dneighbo> then it doesnt apply :)
<reinhard> so when somebody gets gnue
<reinhard> he gets libuuid automatically
<reinhard> and with some configure magic
<dneighbo> reinhard just preserve thier copyright in thier own files then :)
<reinhard> think about like
<dneighbo> and that makes sense :)
<chillywilly> why do we not link to libuuid?
<reinhard> gettext is included in many projects in the intl subdirectory
<dneighbo> will you or someone be keeping this up to date?
<reinhard> chillywilly: we link to libuuid
<dneighbo> as libuuid relesaes new versions
<reinhard> but of all build problems i heard of till now
<dneighbo> chillywilly reinhard is taking good approach
<chillywilly> /usr/lib/libuuid.a
<reinhard> over 50% was "where do i get libuuid from?"
<dneighbo> of including source of libuuid along with geas
<dneighbo> to avoid confusion
<chillywilly> apt-get install uuid-dev
<reinhard> chillywilly: yes
<chillywilly> or, rom equivalent
<reinhard> but we want to build on
<chillywilly> rpm
<reinhard> rpm
<reinhard> as well as on
<chillywilly> /usr/lib/libuuid.so
<reinhard> mac osx
<reinhard> solaris
<reinhard> aix
<reinhard> hp-ux
<reinhard> ...
<reinhard> windows
Action: reinhard hides :)
<dneighbo> neilt i agree with most of reinhard s comments on article
<neilt> cool
<dneighbo> specifically using free software instead open
<neilt> that was a mistake
<chillywilly> we ahouls include every dependency in there for convienience and just make one big static binary right?
<reinhard> SCNR :)
<neilt> eventually i'll learn
<dneighbo> i agree with joe's comments as wells
<chillywilly> where do i get glib?
<dneighbo> er well
<chillywilly> where do i get ORBit
<dneighbo> we should hype gnue more
<neilt> yep thats why i am rewritting
<dneighbo> i want to make sure we make it clear that there are benefits to free software (which gnue is) but not dwell too hard on it
<neilt> this afternoon
<dneighbo> i think that would be a good separate article
<reinhard> neilt: btw i agree with the editor's point
<dneighbo> 'why free software is good for business' or something
<reinhard> i did not think about that the whole issue is about free software
<chillywilly> neilt: you get feed back from the magazine ppl?
Nick change: jcLunch -> jcater
<neilt> yes
<chillywilly> ok
<chillywilly> should i read the log?
<neilt> its in private email
<neilt> i will put in cvs
<chillywilly> ok
<jcater> dneighbo: you get my uploads last night?
<neilt> chillywilly: its there
<chillywilly> k
<chillywilly> tnx
Action: chillywilly is so nosy
<chillywilly> danielb@obfuscation:~/src/gnue$ cvs up
<chillywilly> Secure connection to subversions.gnu.org refused.
<chillywilly> cvs [update aborted]: end of file from server (consult above messages if any)
<chillywilly> bah
<chillywilly> wassup with this thing?
<jcater> apparently lsh keeps going down on subversions
<chillywilly> bugger all
<dneighbo> jcater yes thank you
<chillywilly> you guys here about this SSSCA legislation?
<Mr_You> I haven't read it
<jcater> yip. depressing, eh?
<Mr_You> seemed so
<Mr_You> thats the software regulations?
<chillywilly> oh yes USA Inc. shows how to do it man!
<Mr_You> for security?
<chillywilly> long like the united corporation of america
<chillywilly> live
<neilt> reinhard: did you see the bug i added to the TODO list
<chillywilly> neilt: you can use cvs?
<neilt> yes
<neilt> i can
<chillywilly> I mean youc access subversion
<chillywilly> swith ssh?
<neilt> i learned it years ago :)
<chillywilly> heheh
<reinhard> neilt: yes i saw
<chillywilly> neilt: are you using pserver?
<neilt> reinhard: did that make sense
<neilt> chillywilly: yes
<chillywilly> gah
<reinhard> neilt: i did not check if it is true what you state
<reinhard> but if it is true it makes sense to change it :)
<neilt> reinhard: the point was that the use of root:: should be symetrical
<neilt> always used or never used
<neilt> ok, thanks
<reinhard> yes
<reinhard> neilt: actually i think
<reinhard> as we decided to have everything in exactly 1 module
<reinhard> we don't need root:: at all
<reinhard> andrewm needed root:: for the things that are outside of all modules
<neilt> reinhard: cool
<reinhard> neilt: would you agree to include uuid source w/ geas
<neilt> i agree its a continuing problem
<reinhard> and make config.in use it only if no installed libuuid is available on the system
<neilt> because source is not available on its own
MrChumple (mrchumple@216.250.143.132) joined #gnuenterprise.
<neilt> i would rather see us make a distribution of only uuid available on ash
<neilt> and not necessarily put it in geas
<neilt> that way it is easily available
<neilt> for those that dont do atp-get
<reinhard> the whole uuid source is only 59k
<chillywilly> they have no rpm packages?
<reinhard> chillywilly: it's a part of e2fsprogs
<reinhard> so for example if you use reiser
<reinhard> you don't have it installed
<dneighbo> reinhard i think i might be with neilt
<chillywilly> but you can install it
<dneighbo> maybe not include it in our source
<dneighbo> but put in multiformats on ash
<chillywilly> it is not gonna blow away your reiser stuff
<chillywilly> by installing it
<dneighbo> and im in process of building CDs for gnue
<dneighbo> that have all dependencies
<chillywilly> make an rpm
<neilt> and a source tarball
<chillywilly> ya source is good too
<MrChumple> why is uuid needed (i've not read the code enough to know)
<chillywilly> to create univerally unique ids
<chillywilly> universally even
<chillywilly> for objects ids
<chillywilly> object
<MrChumple> universally unique ids - ah...  
<MrChumple> does it run as a daemon or something?
<chillywilly> you know it is the hwole OO to relation mapping thing
<chillywilly> whole
<MrChumple> or just a library?  does it store the last ID generated in some secret hiding place
<chillywilly> no, just a  lib to link to
<MrChumple> does it create the IDs as text readable strings?  is that why keys are varchars?
<chillywilly> I dunno how it work internally only that it gives you a unique id every time ;)
<neilt> it relys on hardward ethernet hardware address if you configure it that way
<neilt> so it has some privacy implications
<MrChumple> privacy--as in objects that only you can see?
<neilt> because it is possible to determine which hardware was running to create object ids
Action: neilt waits for cw reponse
<chillywilly> I dunno anything about how it works
<reinhard> dneighbo: i am trying to find the best way
<reinhard> of helping people that want to install geas
<neilt> i meant about the privacy implications
<chillywilly> dunno if that is a problem or not...just depends on what it does with the adress
<neilt> encodes in the object id
<chillywilly> yeah but it does the straight adress or it pump it through some hard  to crack algorithm?
<chillywilly> use the
<reinhard> dneighbo: "Likewise, if your program uses small GNU software packages like regex, getopt, obstack, or termcap, include them in the distribution file. Leaving them out would make the distribution file a little smaller at the expense of possible inconvenience to a user who doesn't know what other files to get.
<neilt> i understand it is possible to figure out the ethernet mac address from the resulting uuid
<reinhard> "
<reinhard> dneighbo: this is from the gnu coding standards
<reinhard> dneighbo: my understanding is that the like applies to libuuid
<chillywilly> although with proper security does it matter if you have the mac address? I if CORBA references are not forgable (but they are now) does it matter?
<chillywilly> I can sniff a packet and find out this same info but if they have a firewall it is useless
<neilt> chillywilly: ?
<MrChumple> i don't know how possible it would be do determine mac
<chillywilly> it is in the header
<MrChumple> because they use mac + time + (random key)
<MrChumple> something like that... -- or does it expose the mac entirely as the first part of the number?
<reinhard> if mac can be determined, would this really be an issue?
<chillywilly> I don't think so
<reinhard> when otoh we record the username of the user who created the record?
<reinhard> :)
<chillywilly> what are they gonna do forge packetS?
<Mr_You> umm..
<MrChumple> like a shift mac << 32 + date << 16 + rand
<Mr_You> you want to provide security to the hardware level?
<chillywilly> nah
<Mr_You> if your using the mac adddress you are.. 
<Mr_You> maybe I worded that wrong
<chillywilly> just do not configure it to use the mac address
<Mr_You> GNUe doesn't need to worry about that layer IMO
<chillywilly> and make sure they don't do it either
<Mr_You> just use IPsec if you need security
<Mr_You> s/security/encryption
<chillywilly> yea
<chillywilly> that is an issue though
<chillywilly> hmmm
<chillywilly> thw mac stuff I mean
<Mr_You> 3com has an IPsec chipped ethernet card now!! ;-))
<chillywilly> The uuid_generate function creates a new  universally  unique  identifier
<chillywilly>        (UUID).  The uuid will be generated based on high-quality randomness from
<chillywilly>        /dev/urandom, if available.  If it is not available,  then  uuid_generate
<chillywilly>        will  use an alternative algorithm which uses the current time, the local
<chillywilly>        ethernet MAC address (if available), and random data  generated  using  a
<chillywilly>        pseudo-random generator.
<Mr_You> doesn't use the CPU
<chillywilly> is it really possible to get the mac addy out of all of this?
<Mr_You> I believe most systems provide a /dev/random
<chillywilly> wwell lets say they don't....wouldn't the pseudo random generator be good enough assuming they are using a good algorithm (like taus or someting)
<Mr_You> I have used "random" formulas for "secure" CGI.. using cookie.. 
<Mr_You> yes
<chillywilly>  The  uuid_generate_time  function forces the use of the alternative algo­
<chillywilly>        rithm which uses the current time and the local ethernet MAC address  (if
<chillywilly>        available).   This  algorithm used to be the default one used to generate
<chillywilly>        UUID, but because of the use of the ethernet MAC  address,  it  can  leak
<chillywilly>        information  about when and where the UUID was generated.  This can cause
<chillywilly>        privacy problems in some applications, so the uuid_generate function only
<chillywilly>        uses  this algorithm if a high-quality source of randomness is not avail­
<chillywilly>        able.
<Mr_You> I feel its sufficient and if you want true security go with IPsec
<chillywilly> caution is you so not have /dev/random or /dev/urandom this can be a major security risk
<Mr_You> but most modern OSs have /dev/random
<Mr_You> I wouldn't consider it MAJOR
<chillywilly> yea which can be setup tranparently without geas even worry about it....but we have no real security anyway as object references can be forged right now
<Mr_You> I would consider it a weakness
<chillywilly> sure but how far do we go? Isn't their security their business?
<Mr_You> for a number of reasons, one being you could isolate the production servers if your machine didn't have "/dev/random" to eliminate mac address lookups
<dneighbo> reinhard i wouldnt see anything inheritly wrong with including it
<chillywilly> (our you business if they area client)
<chillywilly> ;)
<dneighbo> just in this day of easy access i question if really necessary
<Mr_You> chillywilly: there is risk involved in everything and the risk involved with new hole in sendmail/bind/etc is higher than someone attempting spoofed connections..
<dneighbo> maybe it is
<dneighbo> im all for easier user experience
<Mr_You> chillywilly: and if they cared about risk then they would be running and OS with /dev/random
<chillywilly> if you are extremely paranoid wouldn't you be using SE Linux?
<Mr_You> the MAC address formula should be the minimum
<Mr_You> chillywilly: no, standard kernel and userland is great
<reinhard> dneighbo: ok
<chillywilly> I think as long as we inform them
<chillywilly> then we are ok
<Mr_You> chillywilly: I've worked for companies that DID NOT want to patch their 11+ Sun machines
<dneighbo> chillywilly if thier warez are GPL it doesnt matter
<dneighbo> we should out of courtesy do so
<Mr_You> chillywilly: yeah
<dneighbo> but they couldnt prevent us from including in gnue
<chillywilly> dneighbo: yea
<chillywilly> just seems sully to include a library in the code, but I have seen that done before
<chillywilly> I don't care
<chillywilly> we could use GSL and make our own uuid generator
<chillywilly> GNU Scientific Library
<Mr_You> I'm all for redistributing libraries for the sake of easy install
<chillywilly> has some excellent random number generators
<chillywilly> ;)
<chillywilly> doesn't make a difference to me
<Mr_You> another thing, is look and see how many OSs in production don't have a /dev/random
<chillywilly> the kernel one has good entryopy
<Mr_You> very few I imagine.. anyone sane who wants a GNUe setup will want it on a future oriented platform
<chillywilly> like GNU/Linux on big iron ;)
<chillywilly> or S390
<Mr_You> we had really old NCR and HP boxes at Lucent and any org like that would love to have a low priced quad/dual server running a support OS
<Mr_You> hehe
<Mr_You> sure, I imagine that has /dev/random ;-)
<chillywilly> GEAS doesn't run on m$ yet
<chillywilly> dunno if any will bother porting it soon
<neilt> but it runs on Macintosh
<neilt> hehe
<chillywilly> OS X right?
<neilt> right
<chillywilly> that's a derver-like OS
<chillywilly> er, server
<chillywilly> UNIX underneath
<neilt> chillywilly: you any closer to revising methods?
<neilt> yeah, netbsd with mach kernal
<chillywilly> I got an imcomplete diagram
<chillywilly> I will work on it tonight
<chillywilly> ;)
<neilt> cool
<Mr_You> neilt: freebsd with mach kernel ;-)
<chillywilly> I actually want to makew a lib that will allow for langauge plugina that can be shared with forms and geas
<neilt> oh yeah, freebsd is correct
<chillywilly> they want triggers in many langauges and we want methods in many langauges
<chillywilly> I thought it would be nice to make some lib to stick in gnue-common
<neilt> cool
<neilt> thats a great idea
<chillywilly> dunno though then we get a new dependency
<MrChumple> I was thinking about the same thing over the weekend
<MrChumple> Doesn't it complicate development for Forms to make it have separate DB drivers and treat data like SQL data vs just object data?
<chillywilly> I would have to make a python binding for it
<MrChumple> If you had some intermediate driver when working in 2-tier that made it
<MrChumple> so you didn't have to go straight at the database...
<neilt> MrChumple: hehe
<neilt> the intermediate driver is GEAS is it not
<MrChumple> But why have a 2-tier version then?
<neilt> exactly
<MrChumple> I'm worried that Forms has to be overly complicated
<MrChumple> You guys said last time I was on that you would always have 2-tier and n-tier around
<MrChumple> in forms code
<Mr_You> 2-tier version is sorta temporary right?
<jcater> because forms is a separate product that can work with GEAS
<chillywilly> flexibility
<jcater> and 2-tier is still a requirement at some places
<neilt> Mr_You: not temporary
<jcater> soon forms will focus on n-tier, but 2-tier will always be there
<neilt> 2 tier is for rapid prototyping
<neilt> and also small projects, small businesses
<Mr_You> oh
<MrChumple> see - not temporary - so forms code always has to talk 2 languages?
<MrChumple> against sql dbs and against geas?
<MrChumple> why not a driver for running in 2-tier that acts to forms like geas does
<MrChumple> then your forms code _only_ has to talk objects
<chillywilly> build theor own object abstraction?
<chillywilly> then forms becomes pretty fat
<chillywilly> in that case
<jcater> MrChumple: how would this help... either way you have multiple code bases to support
<Mr_You> can't have two forms clients?  GEAS and SQL?  if the code does become complicated... you could build one or the other.. or both.. hmm ... just ramblin heh
<jcater> just curious
<chillywilly> I dunno, they could do it
<MrChumple> It just seems like the geas object stuff can do more
<MrChumple> than 2 tier
<chillywilly> forms is thinner in our case
<MrChumple> Can the 2-tier do fields that are derived from selects on other tables?
<chillywilly> bnothing stopping them from making their own object abstractions
<MrChumple> like the reference fields
<chillywilly> if they felt inclined to do so
<Mr_You> MrChumple: I believe you can accomplish that with nested selects in the DB table
<MrChumple> I would argue to drop 2-tier altogether --
<MrChumple> I know it seems radical --
<Mr_You> I would like to see the designer be the "rapid" prototyping/dev app
<chillywilly> hey i don't care as long as they support us
<chillywilly> ;)
<Mr_You> they who?
<chillywilly> forms guys
<jcater> 2-tier is more than rapid prototyping... it is a functionality that does have advantages over n-tier
<MrChumple> But it seems it would be a whole lot easier if forms only had to talk objects
<chillywilly> jcater, jamest, dneighbo
<jcater> just as n-tier has advantages over n-tier
<dneighbo> reinhard : on that note, the libuid stuff should go in gnue common
<jcater> err, over 2-tier
<dneighbo> as its possible something besides geas may need to use it
<jcater> it's all about choices
<chillywilly> dneighbo: I agree with that
<neilt> dneighbo: do you have a test database we can use for geas
<chillywilly> GNUe - It's About Choices
<chillywilly> hehe
<neilt> something like 10,000 names and address
<MrChumple> I'm sorry for bringing up an issue that didn't need to be
<chillywilly> justa  round about way to talk about Freedom ;)
<chillywilly> maybe the suits can eat that one up
<reinhard> dneighbo: i agree we should do that as soon as geas depends on gnue-common anyway
<MrChumple> It just seemed that geas wouldn't be much more overhead anyway
<MrChumple> no matter how small you go
<MrChumple> and It would seem to me that geas-server objects would be faster
<MrChumple> to implement in forms
<reinhard> but removing the dependency on libuuid and for that adding the dependency on gnue-common is not wise imho
<chillywilly> but libgecco or whatever will be in there ;)
<dneighbo> have like 50k zipcodes we can use
<neilt> i have that one
<MrChumple> Hey--I have a different question/topic of concern::
<dneighbo> as to dropping two tier support
<chillywilly> does anyone remeber the 3 areas we agreed were shared?
<dneighbo> i think the full gnue ERP stuff wouldnt be in two tier ever
<MrChumple> What is it that guarantees transactional integrity--suppose 3 ppl edit same record at same time
<chillywilly> we got gedi, scripting, and something else
<chillywilly> bah
<chillywilly> should've written this down
<dneighbo> but at beginning of this project we stated we would have all tools work together or independently
<MrChumple> That sounds good to me
<dneighbo> i think you maybe hung up on 'forms'
<dneighbo> BUT
<dneighbo> think reports
<MrChumple> 2-tier for folks not worried about triggers and methods?
<neilt> dneighbo: that statement is not logical
<dneighbo> if we make a decent report engine
<chillywilly> tiggers are 2-tier methods
<chillywilly> in a way
<dneighbo> why should only people using GNUe ERP be entitled to its benefits?
<MrChumple> Or would 2-tier have methods, triggers, everything
<dneighbo> believe me i like n tier
<dneighbo> i support n tier
<dneighbo> i could live w/o 2 tier
<dneighbo> BUT
<dneighbo> i still advocate at this point we support it
<dneighbo> neilt why is that statement not logical?
<chillywilly> as it allows forms to be indepedent
<dneighbo> gnue is not an ERP
<dneighbo> its not CRM
<dneighbo> its tools
<neilt> but at beginning of this project we stated we would have all tools work together or independently
<dneighbo> tools you could use to build those things
<neilt> what are the alternatives
<neilt> that is a non statement
<chillywilly> is it an "or" or an "and"
<neilt> :)
<dneighbo> ok substitute and/or
<dneighbo> for or
<MrChumple> I can see that 2-tier is fine...  It just seemed to me that it would lack the niceties offered by using it with the geas-server
<MrChumple> Can 2-tier to triggers, methods, constraints?
<dneighbo> MrChumple that would be correct
<dneighbo> it can do those things currently yes
<dneighbo> just tehy become client side or db side
<neilt> you could just put a period after ". . . work."
<chillywilly> triggers us a forms concept
<dneighbo> which is not way to write 'enterprise apps'
<chillywilly> s/us/is
<neilt> :)
<jcater> are we talking about advantages of n-tier over 2-tier, or advantages of object vs relational?
<dneighbo> i think what jcater is getting two
<dneighbo> is no one is saying 2 tier is better or worse
<dneighbo> they are DIFFERENT
<chillywilly> we are not at war here
<MrChumple> no
<neilt> both have different uses
<chillywilly> it is not 2-tier vs. n-tier
<dneighbo> there are advantages to each and disadvantages to each
<chillywilly> I say let the forms ppl support what they want as long as they also support geas I could care less :P
<dneighbo> chillywilly i think they agree with you
<MrChumple> I'm just trying to figure out from you guys if 2-tier and n-tier will be able to offer enough of the same advantages
<dneighbo> but it shouldnt be the 'forms people' and the 'geas people' it should be the 'gnue people'
<MrChumple> but does making 2-tier do all triggers, methods, etc cause it to be as complex as geas-server itself?
<dneighbo> part of the problem was solutions where needed a while back and we didnt have working middle ware
<MrChumple> plus the fact that it has to support multiple databases...
<dneighbo> there was functional 2 tier
<dneighbo> stuff got written in it
<chillywilly> well until my python knowledge is better I cannot make sense of any forms code anyway :P
<dneighbo> we might find as middleware becomes more usable and such that maybe 2 tier fades away
<dneighbo> but until then we need to support both
<dneighbo> MrChumple : no i dont think so
<MrChumple> If 2-tier would force forms to become a second geas-server, then I would wonder why the need...
<dneighbo> its not a 'second geas-server'
<dneighbo> believe me
<MrChumple> ok
<dneighbo> it treats things exactly like a sql server
<dneighbo> if we must have the argument
<MrChumple> no argument intended...
<MrChumple> i'm trying to get a feel for the project overall
<dneighbo> argue in the sense that the forms code will get mess as it will have to make the difference between objects and sql
<dneighbo> ;)
<MrChumple> that's exactly it
<MrChumple> that's why i said why not a middle-man library between forms and sql for 2-tier
<MrChumple> so forms could only act like it was playing with objects
<jamest> MrChumple: if you can find a copy of old Oracle SQL*Forms manual then you'll see some of the driver behind 2-tier in forms
<jamest> driver = drive
<MrChumple> that way your complexity of 2-tier is never causing a support split between 2-tier and n-tier
<MrChumple> in the forms code itself
<MrChumple> I'm only vaguely familiar with Oracle's SQL*Forms...
<dneighbo> MrChumple you were more saying the drivers would be complex
<dneighbo> this is not the case :)
<dneighbo> anyhow, the division MUST stop
<MrChumple> The drivers would have to be the more complex of the two, but it would offer the ability for Forms not to have difficulty supporting two different modes of operation
<jamest> forms treats objects and tables the same
<chillywilly> there's really not much difference conceptually expcept for methods objects hae fields and data
<chillywilly> have
<dneighbo> i dont think supporting both will be that ugly or impossible
<jamest> and treats app servers and database servers the same
<dneighbo> its just its not done yet so there are some 'issues'
<dneighbo> as to division on the gnue TEAM
<dneighbo> ;)
<dneighbo> thats all
<dneighbo> i think a good working example will solve the issue :)
<dneighbo> neilt the stuff you hvae been doing with python is big help
<dneighbo> and i think jamest is committed to looking at the driver 
<jamest> besides now that geas is going to be rewritten I assume it'll all be in python and thus perfection incarnate
Action: jamest dives for cover!
<MrChumple> :)
<dneighbo> um i talked to rms
<dneighbo> hehas demanded we rewrite all of gnue in lisp
<dneighbo> ;)
<MrChumple> Let's get crackin
<chillywilly> ah yea
<MrChumple> lol
<dtm> dneighbo: great!  as an elisp module for emacs?
<dneighbo> dtm: not really
<chillywilly> goooood wrooooooote iiiiin lisp coooooode
<jcater> we need an UIemacs.py driver
<MrChumple> GNU-Emacs-Enterprise
<jamest> dtm: that is far closer to the truth than I care to admit
<dneighbo> but emacs virtual lisp machine is a dependency
<dtm> jamest: h0h0
<MrChumple> The ERP/CRM/etc editor of choice
<dtm> MrChumple: with Tetris
<dneighbo> dtm: its not funny
<jamest> dtm: we've been asked to make it work in emacs
<dneighbo> that is actually partially underway
<MrChumple> Enterprise Tetris
<MrChumple> uh oh
<chillywilly> heheh
<dneighbo> i.e. we are workign to put gnue in place at FSF
<chillywilly> and emacs forms client?
<MrChumple> GNUe emacs-tier
<chillywilly> er frontend
<jamest> chillywilly: yip
<dneighbo> first goal is an xml db driver
<dtm> dude.  why doesn't emacs get device drivers?
<chillywilly> rock on!
<dneighbo> so they can just edit the xml in emacs
<jamest> and a flat file database driver
<dneighbo> and the gui folks can use forms
<dtm> is emacs going to be ported into a hurd kernel server?
<dneighbo> the second step is emacs hacker to make
<dneighbo> gnue-forms (emacs)
<chillywilly> dtm: nah
<dneighbo> i.e. a forms painter in emacs
<dneighbo> while i would like to say im kidding
<dneighbo> im not
<MrChumple> Doh
<MrChumple> Let RMS write it
<dneighbo> the fsf will write the emacs forms client
<jamest> maybe
<dneighbo> we will write the xml/flatfile driver
<chillywilly> RMS doesn't use GUI afaik
<jamest> again maybe :)
<dneighbo> as we need to provide those anyhow for integrator
<chillywilly> he needs emacs
<chillywilly> forntend
<MrChumple> The GNU/Emacs operating system
<neilt> from what i heard, the problem with lisp is that it is not python.  you see python code almost writes its self.
<jamest> that driver is further down the todo list than geas driver updates
<chillywilly> hold on
<jamest> and it's taking me forever to get that done 
<MrChumple> Cut Linux out altogether by replacing the Kernel with EMACS
<dneighbo> jamest: absolutely
<chillywilly> if we have the language plugin system we can support elisp
<chillywilly> ;)
<jamest> neilt: exactly
<chillywilly> triggers can be done in elisp
<dneighbo> chillywilly doubt it
<jamest> neilt: anything non-python is just silly to even think about
<dneighbo> triggers would be scheme probably (At least in geas)
<jamest> except of course emacs
<MrChumple> (To make sure I'm hearing correctly:)  What I'm hearing is that forms is just as useable in 2-tier as n-tier...
<jcater> well, currently, more so :)
<dneighbo> MrChumple eventually that iwll be the case
<MrChumple> OK
<neilt> jamest: exactly, just yesterday i was thinking about a python program and the debugger found my error before i wrote the code
<MrChumple> That's all I guess I was getting at...
<MrChumple> lol
<dneighbo> currently the driver for n-tier isnt working so its not functional as a whole
<MrChumple> OK, but that's the goal...
<dneighbo> ie geas works and forms works just not together for n-tier
<dneighbo> hopefully by end of month that will be fixed
<MrChumple> Ya... once they're together it'll be easier to start writing modules like CRM, Inventory, etc...
<MrChumple> right?
<dneighbo> yes
<jamest> neilt: yip, and as long as you use gnue-common to start your day and pass it a -d 100 level your entire life will become much easier to deal with
<dneighbo> but thats where the REAL issue lies for me
<neilt> MrChumple: we can write modules not
<neilt> now
<neilt> you just cant test them
<dneighbo> while ntier and 2tier can do the SAME things
<neilt> or use them
<dneighbo> they do them differently
<dneighbo> so the real debate is
<chillywilly> hey they have the mt orbit in sid
<chillywilly> heheh
<chillywilly> wonder if I cna break my gnome with it
<dneighbo> if you write a CRM module do we do so in a fashion that you could implment it in either
<dneighbo> or just one or the other?
<MrChumple> Ya?
<jamest> chillywilly: you go boy!
<dneighbo> so to me this is the REAL issue
<dneighbo> not forms supporting both :)
<MrChumple> Would it be both? or would they only be modules for geas-server?
<jamest> :)
<dneighbo> but do applications developers support both :)
<jamest> dneighbo: i think there is room for both
<dneighbo> im leaning towards modules only for geas
<MrChumple> Then do some modules only come out for 2-tier?
<jamest> as when people say 2-tier isn't enterprise level stuff
<dneighbo> and let people doing smaller things develop own
<dneighbo> BUT
<jamest> they neglect the fact that a ton of enterprise level apps are already 2-tier
<dneighbo> realistically would could amek it both with tools
<jamest> it's just not the enterprise solution of the day
<MrChumple> So both could be done likely with the same modules?
<dneighbo> but thats a different argument :)
<dneighbo> no
<MrChumple> One module coded - and something installs it for 2-tier, and another for n-tier?
<dneighbo> they couldnt use the same modules without work
<dneighbo> or tools
<dneighbo> to do conversions
<dneighbo> MrChumple we could do that with some work
<MrChumple> What kinds of things would be different?
<dneighbo> a lot would depend on if they share same trigger/bizrule syntax
<jamest> server side methods
<MrChumple> What currently would be the difference between running it 2-tier and through geas-server
<dneighbo> as if thats different then its hard to do
<dneighbo> if its same then wont be too bad
<MrChumple> True
<jamest> i _want_ to share trigger/methods code
<MrChumple> So the triggers/methods/etc need to be supported between both
<chillywilly> configure.ac:11: warning: AC_PROG_LEX invoked multiple times
<dneighbo> also geas builds ugly tables
<jamest> chillywilly: did you look at the doc in gnue-common/docs
<MrChumple> I like that idea most
<dneighbo> for two tier programming
<jamest> on triggerSpecs
<chillywilly> jamest: no
<chillywilly> jamest: whatis there?
<dneighbo> but i think that could be overcome
<jamest> the start of a method(trigger) spec
<chillywilly> ok
<jamest> dealing w/ layout, features, etc
<chillywilly> layout?
<dneighbo> MrChumple triggers will be supported in both
<dneighbo> the question is will the syntax be the same
<MrChumple> Methods too
<dneighbo> as if in two tier i say
<jamest> layout = how the method system will be implemented (Triggers, a trigger manager system, etc,etc)
<dneighbo> obj.fields.property = foo
<dneighbo> but in geas i say
<MrChumple> What will be the main advantage to n-tier?
<dneighbo> sys.obj.property = foo
<dneighbo> or something
<chillywilly> jamest: i.e., design?
<dneighbo> then i cant use same 'biz rule code' for both
<MrChumple> What will n-tier give me in the end?
<jamest> chillywilly: yip
<dneighbo> without a conversion
<neilt> MrChumple: scalability
<jamest> MrChumple: a buzzword compliant system
<neilt> better plug and play
Action: jamest again dives for cover
<jcater> rofl
<dneighbo> MrChumple scalablity AND methods not tied to the client
<chillywilly> jamest: well the thing that I have been tryint to geta  grasp on lately is forms so I know enought to make it fit for both and ell i wanna port this one app to a 2-tier solution for now
<neilt> multiple SQL db support at the same time
<chillywilly> s/ell/well
<neilt> non-SQL db support
<dneighbo> specifically distributed processing
<dneighbo> in two tier
<neilt> load leveling
<dneighbo> if you execute a rule it gets executed on teh client
<dneighbo> even if stored on a server
<neilt> distributed objects
<dneighbo> in geas world
<dneighbo> the code gets executed on server
<dneighbo> and result returned to client
<dneighbo> does that make sense?
<MrChumple> So n-tier for much larger installations?
<dneighbo> that is the FUNDAMENTAL difference
<dneighbo> between ntier and 2tier in my mind anyhow
<MrChumple> I have one concern that applies to both::
<MrChumple> Have you addressed the issue of multiple simultaneous updates?
<MrChumple> Two people pull up my contact information from the DB
<dneighbo> dont use mysql
<dneighbo> ;)
<MrChumple> one changes my address
<MrChumple> the other changes my phone
<dneighbo> we have discussed
<MrChumple> the second guy overwrote my address unless you have some checking in there
<dneighbo> basically i think there are a few ways to do it and it should be configurable
<jamest> i expect you'll see row level locking in forms driver
<dneighbo> one way is first post wins
<dneighbo> one is first grab wins
<dneighbo> one is merge records
<dneighbo> one is let user decide
<jamest> based upon if the backend supports it
<dneighbo> most apps do first grabs wins
<MrChumple> postgres can lock one row right?
<jamest> I'd _think_ the geas system could drop that to field level
<jamest> MrChumple: yes
<dneighbo> i.e. if i get a record i lock it so its readonly for everyone else
<dneighbo> first post is also common
<dneighbo> it works like cvs
<dneighbo> both checkout
<MrChumple> Hmm...
<dneighbo> first one in updates
<dneighbo> second update gets told sorry dude your record is out of date
<jcater> dneighbo: that's how forms will work, IIRC
<MrChumple> Then you force the second one to do a select --
<dneighbo> the merge is most powerful
<dneighbo> so if you changed name and i changed address
<dneighbo> it would merge
<MrChumple> True
<dneighbo> and only puke if we both changed same field
<dneighbo> saying record out of date
<MrChumple> Would there need to be code common to geas-server and 2-tier forms to handle that
<dneighbo> most powerful is merge with resolve
<dneighbo> so it merges but what it cant resolve it asks you what to do
<dneighbo> i.e. choose who is right
<MrChumple> And how does it know what fields were changed w/o revisioning changes?
<MrChumple> (in 2-tier mode)
<MrChumple> in n-tier mode I can see somehow maintaining change information for an object--
<chillywilly> is geas borked now in cvs?
<chillywilly> \does the old parser compile and work?
<MrChumple> Is there anything currently doing locking?
<chillywilly> GEAS needs security and locking
<MrChumple> I visited the Wyatt-ERP guys' installation last Friday, and their stuff looks great, but ...
<MrChumple> it doesn't address locking...
<MrChumple> They are different from GNUe in a sense because they are using postgres for all the functionality they can squeeze out of it
<nickr> that is a horrible pun
<MrChumple> They're using triggers, rules, views, relational integrity constraints
<dneighbo> yike
<dneighbo> old school two tier
<MrChumple> And that's got a certain value...
<jamest> MrChumple: we won't do that
<dneighbo> fake ntier by using db as 3rd tier
<dneighbo> which is horrid vendor lockin
<MrChumple> I can see how geas-server is going to be the one managing triggers, etc...
<jcater> well, that's one way to look at it
<jcater> there are advantages
<MrChumple> I know, horrid vendor-lock, supreme db usage
<dneighbo> not to mention utter hell upgrading your db
<dneighbo> without rewritting code
<jcater> to doing at db level vs middle-ware
<chillywilly> ServerError: Class 'language::help' not found
<dneighbo> jcater can surely tellyou the evils of this in oracle world? ;)
<MrChumple> middle-ware has a true advantage
<chillywilly> does schema.py just print everything out or what
<dneighbo> jcater: if its all in house
<dneighbo> i agree doing in a particular db may have advantages
<MrChumple> You end up using your RDBMS as a very intelligent hard-drive
<jcater> dneighbo: yes, but it has also saved my butt before over a middle-ware solution
<MrChumple> IN which case, you shouldn't use Oracle for sure
<dneighbo> if you are writing software to sell or give to others putting in a single DB is suicide unless you are sadistic
<chillywilly> schema.py is just looping forever
<jcater> when I do something at DB level I KNOW the necessary triggers are happening, regardless of how the data got into the DB
<MrChumple> True
<jcater> (i.e., if a developer forgot to go thru middle-ware and just issued a psql command)
<MrChumple> Triggers at the DB level are good if you get your fingers into the DB directly
<dneighbo> jcater: yip but one shouldnt be going directly against db
<dneighbo> but really its not worth debating
<jcater> should is an interesting word :)
<MrChumple> Is there any possibility of adding constraint checking into databases that can handle it?
<dneighbo> i dont think anyone here would advocate doing stuff for gnue that is database specific
<MrChumple> So the database does it anyway?... or too much work
<jcater> I'm not defending DB-level, but just pointing out that there's more to it than "vendor-locking"
<MrChumple> No, no no - not db specific
<dneighbo> jcater: right like all things there are trade offs
<jcater> dneighbo: I agree
<jcater> just keeping us honest :)
<dneighbo> i wasnt saying they do it to 'lock you in' (Though they might)
<dneighbo> that is just a side affect of it
<MrChumple> Can gnue take advantage of things like 'Views'?
<dneighbo> just like if we use python or C or foo
<dneighbo> we are tied to it
<dneighbo> and switching and/or upgrading is difficult w/o redoing our code possibly
<MrChumple> yes, and I appreciate the advantage of not being language-locked
<MrChumple> not being db-locked is valuable too
<MrChumple> when forms is run in 2-tier it will never go against the database that the n-tier is running with?
<MrChumple> right?
<dneighbo> MrChumple that is up to the institution using it
<MrChumple> Both is not an option...?  I can't create a quick 2-tier form to get at the db geas-server is using can I?
<dneighbo> like all things we cant and shouldnt dictate how they use the tools
<dneighbo> MrChumple sure you could
<MrChumple> Wouldn't that throw off the cacheing, 
<dneighbo> it might
<dneighbo> but you could do it
<dneighbo> just like crashig your car into a wall hurts your car
<dneighbo> you could do so :)
<MrChumple> So it's a 'could do it' situation, but not a 'should do it'
<dneighbo> right
<dneighbo> just like i could go into postgres and do
<reinhard> just for the record i will neither include libuuid in geas nor make a seperate tarball at this time
<dneighbo> UPDATE footable set foofield='foovalue' where foolookup='myfoo';
<dneighbo> and if i have permission i can do this
<reinhard> libuuid is highly system dependent and needs tons of configure options which i don't want to hack out now
<neilt> chillywilly: you get schema.py working?
<reinhard> maybe when i'm bored sometime i will get back on this
<reinhard> --
<chillywilly> neilt: no it loops forever spitting the menu out
<chillywilly> over and over
<neilt> without you typeing anything
<chillywilly> yes
<dneighbo> neilt remember forever for chillywilly is length of time it takes to down a pespi
<neilt> odd
<dneighbo> so if this is a script that takes a while... 
<dneighbo> ;)
<chillywilly> dneighbo: eh?
<chillywilly> dneighbo: hehe
<neilt> it should not take anytime
<chillywilly> it should stop so I cna enter somthing right?
<neilt> it should put up a list with 2 columes
<chillywilly> neilt: I set it to use the examples objects
<chillywilly> should I make it use gnue-config?
<neilt> and out 10 or 12 lines
<neilt> then ask for imput
<chillywilly> or ddoes that matter?
<neilt> i never tried it with examples objects
<chillywilly> well it goes for qwuite some time
<neilt> i just used it with gnue-config
<chillywilly> ah ok
<chillywilly> I'l change my geas.conf
<neilt> but it should work with whatever geas is running with
<reinhard> neilt: no
<neilt> ahhh, but what if there are not 20 objects
<chillywilly> that works fine
<chillywilly> gnue-config does
<reinhard> it depends on the language object being there probably
<neilt> i thought python though about all those other conditions
<chillywilly> there's nor restarting geas function is there?
<chillywilly> we need to make it more daemon like
<neilt> did not know i had to worrry about it
<reinhard> neilt: lol
<dneighbo> oh on that note about benchmarking and testing
<chillywilly> we need an init script
<chillywilly> ;)
<dneighbo> if we make scripts etc and plans
<dneighbo> we are cleared to use the hardware at osdl
<dneighbo> so we can test geas on 
<dneighbo> multi processor IA64 machines
<dneighbo> and the such :)
<chillywilly> ooooh
<chillywilly> so is OSDN lab gonna go bye bye?
<chillywilly> now that VA is having problems
<chillywilly> wow
<chillywilly> I got it to crash
<chillywilly> woohoo
<dneighbo> its osdl
<dneighbo> not
<dneighbo> osdn
<chillywilly> Select a class number to display [q,return] [0-78] 1
<chillywilly>  
<chillywilly> GLib-CRITICAL **: file gutils.c: line 265 (g_basename): assertion `file_name != NULL' failed.
<chillywilly> kernel segmentation fault: root file system corrupted. system terminating.
<chillywilly> Just kidding  :)
<chillywilly> Remember: Detailed bug reports are a valuable tool, if you want the problem fixed.
<dneighbo> its by hp, ibm and su
<chillywilly>   :)
<chillywilly> A full stack trace from 'gdb' in particular is useful for identifying
<chillywilly> exactly where and why an error occured.
<chillywilly> Traceback (innermost last):
<chillywilly>   File "./schema.py", line 205, in ?
<chillywilly>     c = con.getFullClassDefinition( classnames[idx-1] )
<chillywilly>   File "<string>", line 1, in getFullClassDefinition
<chillywilly> CORBA.COMM_FAILURE: Minor: 0, Completed: CORBA_COMPLETED_NO.
<chillywilly> [1]+  Segmentation fault      (core dumped) ./geas-server  (wd: ~/src/gnue/geas/src)
<chillywilly> (wd now: ~/src/gnue/geas/examples/python)
<neilt> cool, wish i knew what all that meant :)
<neilt> if probably a configuration error, just reboot and see if it happens again
<neilt> opps
<dneighbo> roflmao
<neilt> sorry, thats my windows playbook
<chillywilly> where do we use g_basename
<chillywilly> 3 works fine
<neilt> in the corba binding code to convert the filename
<neilt> it means that a filename = null crahses
<neilt> so we need to add a check
<neilt> i'll do that now
<chillywilly> but one makes things crash
<chillywilly> "1"
<chillywilly> the user class
<chillywilly> canyou dcc a patch when you are done?
<chillywilly> eh maybe i should checkout with pserver
<chillywilly> ;)
<chillywilly> what is th epoint of "root" classes that do nto have a module?
<neilt> ok, temporary fix committed
<jcater> ah, my favorite set of works: "temporary fix"
<jcater> use it alot myself :)
<jcater> works=words
<chillywilly> heheh
<chillywilly> neilt: so the gcd is not there?
<chillywilly> is that why it dies?
<chillywilly> I should get the commit mail I think
<chillywilly> mails
<chillywilly> if I didn't sign up under my FD email which I cannot fucking access right now
<chillywilly> ack
<neilt> some classes that are created by the server do no have a gcd filename so it is null
<chillywilly> ah
<chillywilly> like the internal ones?
<neilt>   /* class name */
<neilt>   retval->name = CORBA_string_dup (odl_class_get_full_name (c));
<neilt>   if ( odl_class_get_gcd_filename (c) )
<neilt>     {
<neilt>     retval->gcdfilename = CORBA_string_dup (g_basename(odl_class_get_gcd_filename (c)));
<neilt>     }
<neilt> the patch is the if statement
<neilt> in schema.c
<chillywilly> I see
<chillywilly> duoping a null is baaad
<chillywilly> duping
<chillywilly> ;)
<neilt> dneighbo: what do we have to do to use the test machines
<neilt> at osdn
<neilt> i am planning on testing some volume data
<neilt> with current geas
<neilt> to get benchmarks
<dneighbo> we have to submit a test plan etc
<dneighbo> i told them we werent quite ready yet
<dneighbo> when we are close enough i will get the paperwork and we can put in our request
<dneighbo> they will prepare the machine for us and give us accounts
<neilt> ok, let me get tests done for local machine
<neilt> then i try to take down ash
<neilt> then we will try osdn
<neilt> :)
<dneighbo> test on DEV16
<dneighbo> its a meager 16 processor 550hmz with 16Gig of RAM with 612 gig of storage
<chillywilly> wooooah
<neilt> just the machine for our benchmarks
<chillywilly> that be a big beastie
<dneighbo> neilt: i would hope that dev16 would produce favorable geas speeds
<dneighbo> if not we are in trouble
<neilt> i can see it now
<neilt> requirements: 12 to 16 processors
<neilt> minimum 400 Mhz
<chillywilly> lol
<neilt> minimum 12 gid
<chillywilly> woohoo!
<neilt> and you have to be able to processes 1 billion transactions per day
<neilt> then geas is deisgned for you!
<dneighbo> roflmao
<dneighbo> well isnt that SAP requirements?
<chillywilly> hehe
<dneighbo> at city of phoenix they bought a HP Unix machine with quad processors
<dneighbo> and like 8gig ram
<neilt> no sap requires 100 support programmers at $100,000 plus per person
<dneighbo> for informix
<dneighbo> or something
<dneighbo> neilt: yeah they bought those too (The programmers)
<neilt> but it will run on 100 Mhz 486
<dneighbo> neilt i recall that now
<dneighbo> SAP Hardware Requirements: See Consulting Section 2.4b
<chillywilly> heheh
<reinhard> jamest: solaris patch committed :)
<reinhard> wrt -lsocket
<dneighbo> SAP Consulting Section 2.4b - The size of machine SAP runs on is indirectly proportional to the amount you pay for consulting fees. This is because the better consultants you hire the more 'experienced' they are at squeezing out performance.  A 320 hour consultation is needed to assess what hadware and consultants you will need.  
<chillywilly> is that serious?!?
<dneighbo> i think these guys (sap, peoplesoft etc) are the only ones that actually 'make' a profit during an RFP process :)
<chillywilly> you really got that form their manual?
<nickr> I like your ideas and would like to subscibe to your newsletter.
<dneighbo> chillywilly i made it up, but i wouldnt be amazed to find something similar
<chillywilly> hehe
<neilt> ok, later all, i have to run to plane
<reinhard> l8r neilt
<dtm> dneighbo: haha good call
<dtm> neilt: later
neilt (neilt@dialup-63.208.66.199.Dial1.Chicago1.Level3.net) left irc: bb tomorrow
<MrChumple> Hey guys, I've got a question::  I've loaded up a few of the forms demos, but haven't seen them all, and...
<MrChumple> I was wondering if forms can preview a list of table data in a scrollable columned format?
<MrChumple> Where you can resize each column... -
<MrChumple> Say I want to see data from the 'PERSON' table, and want to order columns in this order 'FIRSTNAME', 'LASTNAME', 'TITLE'
<MrChumple> Can forms pull up the contents or partial contents of a table and show it to me in a list that I can select from?
<MrChumple> Is anyone around?
<reinhard> nobody who dares to answer this question ;)
<Mr_You> hehe
<MrChumple> :) hehe.  everyone hates MrChumple now because he started a 2-tier/n-tier argument
<Mr_You> I'm not sure of the extent of Forms functionality yet
<reinhard> MrChumple: i agree with you that this is standard functionality
<reinhard> but i think forms does not have this now
<MrChumple> It'd be a cool think for me to try to add in--
<Mr_You> yeah, need lots of Forms widgets
<dneighbo> MrChumple gnue isnt ms access
<dneighbo> is the easy answer to taht
<MrChumple> Then the ability to tie in actions so you can either double click an item in the list, or right click it to do stuff to it
<MrChumple> I've never seen ms access :)
<dneighbo> im not sure what you are askgin for
<Mr_You> dneighbo: you should be able to create a dynamic list of items and have functionality when an item is selected
<Mr_You> actually!!
<dneighbo> things that sound similar
<dneighbo> popup search dialogs
<dneighbo> grids
<Mr_You> how bout this: can you create a Form that browses through a list showing a specific amount of number each time?  
<MrChumple> Just a pane that is a list of data --the entire or partial contents of a table
<MrChumple> like this:
<MrChumple> Person table:
<MrChumple> FIRST NAME, LAST NAME
<MrChumple> Bob           Jones
<dneighbo> yes you can do that
<MrChumple> Harry          THomas
<MrChumple> etc...
<dneighbo> yes it does that now
<dneighbo> you would create two labels
<MrChumple> Awesome!
<dneighbo> one for first name one for last name
<dneighbo> then two entry widgets
<MrChumple> And they come up in a list?
<dneighbo> one for first name
<dneighbo> one for last name
<MrChumple> I'm not sure if you're thinking the same thing --
<dneighbo> then you define how many to display at one time
<Mr_You> yeah 
<dneighbo> rows="5"
<dneighbo> and that would let you see 5 at a time
<MrChumple> Can you resize columns?
<dneighbo> of firstname lastname
<dneighbo> no
<MrChumple> OK, I think it's kinda like that
<dneighbo> you will be able to when we support grids
<MrChumple> Where if you select the second line, the whole row is highlighted
<MrChumple> And be able to scroll through the list...
<dneighbo> you can scroll through the list
<MrChumple> Most similar to the 'Find files' utility in all the Windows 
<dneighbo> it doesnt 'highlight' the row
<MrChumple> A multi-function 'preview pane'
<dneighbo> but it puts the focus on that record
<dneighbo> and you can edit it right there
<dneighbo> the zipcodes sample should have one like this
<dneighbo> if you are talking find 
<MrChumple> I saw the zipcodes example...  that's not exactly what i'm aiming at
<dneighbo> im thinking you are wanting popup search boxes
<dneighbo> where say i have a field on a form called client number
<dneighbo> then next to it is a button to search
<dneighbo> i click that
<dneighbo> and it opens up a dialog
<dneighbo> that has firstname lastname
<MrChumple> Ya, where you can resize columns, and scroll up and down...
<MrChumple> Not necessarily only for popup
<dneighbo> i can search by first or lastname to 'narrow down' the choices
<MrChumple> Only using 'find' as an example of an app that has that kind of interface
<dneighbo> when i find the record i want i click on it
<dneighbo> and it populates the client number field
<MrChumple> Say you want to scroll through all of the data in a table quickly, locate the one row you wanted to edit, and then double-clicked it
<dneighbo> we have what you are describing now today w/o resizing
<dneighbo> we have that
<MrChumple> Not really a drop-down, or a search... but those are necessary too...
<dneighbo> only you dont need to double click it
<MrChumple> you have that -
<dneighbo> you can edit it right there
<MrChumple> Hmm...
<dneighbo> you running cvs?
<MrChumple> Sure...
<dneighbo> ok
<dneighbo> when did you get it?
<dneighbo> and when did you last up date it?
<MrChumple> Oh, oops, not the cvs version - I thought you were asking if I had cvs and knew how to run it --
<dneighbo> as this functionality was broken until like a few days ago in cvs for some of the samples
<MrChumple> I have 0.0.8 I think
<MrChumple> or 0.8.0 - I can't remember which it is - of forms
<dneighbo> ok that functionality should be ok in 0.0.9
<dneighbo> er or 0.0.8
<MrChumple> So you do have preview panes that don't force you to edit right there?
<MrChumple> Ones you can just scroll through...
<dneighbo> http://www.gnuenterprise.org/modules.php?op=modload&name=NS-My_eGallery&file=index&do=showpic&pid=16
<dneighbo> is a sample form
<dneighbo> that has this
<dneighbo> sku, description, oldqty, newqty, sold
ToyMan (stuq@c5300-1-ip123.albany.thebiz.net) left irc: [x]chat
<dneighbo> can be scrolled thorugh and contain unlimited records
<MrChumple> Hmmm...
<MrChumple> That's not quite the interface I was looking for though -- 
<dneighbo> you want a grid
<MrChumple> Maybe I should paint up an example... Or take a screenshot of another --
<dneighbo> as you want it resizable
<MrChumple> If I get a screenshot, is there a way to send it in...?
<dneighbo> that will come soon enough
<dneighbo> note you can add a 'scrollbar' to the right of this
<dneighbo> as well
<dneighbo> to make a visual connect
<dneighbo> i think that is probably what is confusing you
<MrChumple> Yes, I can imagine a scroller on the right of those, but it's not the UI I'm shooting for still...
<MrChumple> I'll get an example --
<dneighbo> MrChumple please note forms is about being quick and simple
<dneighbo> to rapidly build applications
<dneighbo> we dont plan on support a multitude of widget sets
<dneighbo> remember we have to from same code work on curses and html
<dneighbo> as well as mac, gtk, windows
<MrChumple> The same basic Idea could still easily be coded into curses or html...
<dneighbo> not saying it couldnt
<dneighbo> but im obviously not understanding what you are asking for
<MrChumple> I guess I'm asking for an edit field at top, and the possible fields to edit below
<dneighbo> or you arent asking for what you are wanting :)
<MrChumple> you scroll through the people in the company, hit enter, or dblclick on the one you want to edit
<MrChumple> the info pops up in the edit fields above
<MrChumple> you change it, hit update and you're done editing that one
<MrChumple> or...
<dneighbo> why?
<dneighbo> if you can edit where searchign why make them double click to edit it?
<MrChumple> you scroll down to the person, hit enter, and edit his address info - adding a shipping address
<dneighbo> what you could od
<dneighbo> say you had
<dneighbo> firstname lastnam
<MrChumple> because editing and searching are sometimes not good together like that
<dneighbo> and a list of people
<MrChumple> like a list of people - whose addresses i want to edit
<dneighbo> on double click of that box it could pop up a form
<dneighbo> that has all the fields to edit for that person
<MrChumple> you have to clunk up the UI a lot in order to show first-name last-name, and all the address fields on one line
<dneighbo> or the reverse of this is the popup search form
<dneighbo> this would be much easier to demo 
<MrChumple> dang -- gotta go to a meeting for a sec -- i'll get a screenshot of what i'm talking about -- and how it has advantages...
<dneighbo> this is why irc sucks :)
<MrChumple> hehe
<dneighbo> i think i know now
<dneighbo> and you can do that
<dneighbo> currently
<dneighbo> i personally would tackle in one of two other ways for cleaner ui
<dneighbo> but what youare wanting should be doable today
<MrChumple> :) sorry for the confusion initially
<MrChumple> ...
Action: MrChumple out for about 40 minutes
<reinhard> dneighbo: i need a native english speaker
<reinhard> if i define a type
<reinhard> type money
<reinhard> {
<reinhard>   int64 amount;
<reinhard>   int16 num_decimals;
<reinhard>   currency *currency;
<reinhard> }
<reinhard> could you call that a "compound type"
<reinhard> ?
<nickr> looks like a compound type to me
<reinhard> thanks
davi (davi@204-LASP-X11.libre.retevision.es) joined #gnuenterprise.
<dneighbo> to anyone that cares i releases geas to freshmeat and *apps.com
<dneighbo> for 0.0.6
<dneighbo> i just noticed that it hadnt been done yet over the weekend
<reinhard> dneighbo: ok
<dneighbo> hopefully there will be 0.1.0 or 0.2.0 release here soon :)
nile (Nile@216-210-171-2.atgi.net) joined #gnuenterprise.
<jamest> bbl
jamest (jamest@hobbes.math.ksu.edu) left irc: [x]chat
nile (Nile@216-210-171-2.atgi.net) left #gnuenterprise (Client Exiting).
davi (davi@204-LASP-X11.libre.retevision.es) left irc: Client Exiting
ToyMan (stuq@c5300-3-ip155.albany.thebiz.net) joined #gnuenterprise.
<chillywilly> reinhard: or an aggregate
<reinhard> chillywilly: can you live with compound?
<reinhard> i already spread it all over my source ;)
<chillywilly> sure
<reinhard> ok :)
<chillywilly> 1,$s/compund/aggregate/g
<chillywilly> :P
<chillywilly> j/k I can live with it
<jcater> :g/compound/s//aggregate/g
<jcater> :)
<reinhard> hey
<reinhard> you wanna know how you can fix 40 compile errors in a c program
<chillywilly> you don't think I forgot my VI did you?
<reinhard> with a single keystroke?
<chillywilly> how?
<jcater> power button?
<reinhard> remove the ";" between function header and body :)
<reinhard> jcater: lol
<chillywilly> you copied it didn't you?
<reinhard> chillywilly: i always do this :)
<chillywilly> ;)
<chillywilly> that's one thing I dug about M$'s VC++ 6 (yes I had to use it once for school) it would make a entry for the function implementation and throw you at it if you want also it would pop a dropdownlist when you were gonna call a method....it reduced typo error a lot
<chillywilly> errors
<reinhard> yeah
<chillywilly> you could do object. and then it would show a little list to pick from
<reinhard> seems like people that used it never learn to type properly :P
<chillywilly> heh
<reinhard> scnr
<chillywilly> I only used it once
<chillywilly> so you can't blame it on that
<chillywilly> ;)
<chillywilly> haha
jcater (jason@w202.z065105010.mem-tn.dsl.cnc.net) left irc: home
<MrChumple> dneighbo: I have an example of the scrollable 'preview pane' I was talking about earlier...
<dneighbo> ok
<MrChumple> I can send it over IRC if you want to check it out --
<dneighbo> upload it to web and give me a url
<MrChumple> ok...
<dneighbo> thats easiest
<dneighbo> only because i do magic to be on irc
<dneighbo> i.e. i go through multiple machines on ssh on non ssh ports with masqerauding etc
<dneighbo> im probably one of few on planet who requires specilaly compiled kernel to irc
<dneighbo> ;)
<MrChumple> ok...  i'm new to this IRC stuff -- 
<MrChumple> hehe
<chillywilly> dneighbo: just do port forwarding
<chillywilly> that's the sources.list line?
<chillywilly> woops
<chillywilly> wrong channel
<MrChumple> ok...
<dneighbo> chillywilly to do pf needed kernel recompile
<MrChumple> got the file up now -- 
<MrChumple> http://216.250.140.227/lists_edits.png
<chillywilly> dneighbo: so?
<chillywilly> dneighbo: you can do it
<chillywilly> ;)
<dneighbo> its done
<dneighbo> was just telling MrChumple why i cant accept dcc files
<dneighbo> what system is this from MrChumple 
<dneighbo> its png and i dont have laptop and m$ dont like png too mcuh
<dneighbo> booting laptop now
<reinhard> 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.
<chillywilly> why you using M$
Action: chillywilly sneers
<dneighbo> um i have to at work for some things
<dneighbo> my switch box isnt here still :(
<dneighbo> so too lazy to bounce back and form
<chillywilly> to support the  lusers?
<dneighbo> forth
<dneighbo> i use linux all day
<dneighbo> just from putty windows
<dneighbo> this is graphic so i need gui
<chillywilly> yea
<MrChumple> sorry - i was afk
<chillywilly> I just like picking on you masta
<chillywilly> ;)
<dneighbo> i hope to when switch box comes
<dneighbo> be in linux gui all time
<dneighbo> with vnc session to my windows boxen
<chillywilly> switch box?
<chillywilly> whatcha mean switch box?
<MrChumple> i'll put it into jpg
<MrChumple> is that ok?
<MrChumple> sorry - just catching up
<dneighbo> no its ok
<dneighbo> im in gnu/linux now
<MrChumple> sorry 'bout that
<dneighbo> nothing to be sorry about
<dneighbo> i would have complained if were not png :)
<dneighbo> as png is the 'correct' format for gnu :)
<MrChumple> hehe -- that's why i used png
<MrChumple> cuz chillywilly would've attacked ;)
<chillywilly> man tcl/tk is fugly
<MrChumple> i know
<MrChumple> BUT...
<MrChumple> these guys built an interface that was actually pretty intuitive...
<MrChumple> You double click ona  person
<MrChumple> and their info comes up above to where you can edit it...
<MrChumple> It also updates the other two panes (address & contact info) with lists of info
<dneighbo> ok you could do something like this
<dneighbo> currently
<MrChumple> So you go to 'John Doe' and it pulls up his address, contacts, etc...
<MrChumple> But not forcing them necessarily into popups of their own... 
<dneighbo> i dont think i agree with the gui design 
<dneighbo> but you could do this
<MrChumple> Granted the TCL/TK doesn't allow sizing of each field (something I'd like too)
<MrChumple> Really?
<dneighbo> you would do what i said before
<chillywilly> I dunno why our current forms isn't just an xml secription of GTK+ wxwindows or just wraps up some of the stuff
<dneighbo> example in your contact_editor section
<chillywilly> so we can make soem better GUI
<chillywilly> s
<dneighbo> you would make the bottom part like i said
<dneighbo> then whne you are on a record you could make it auto pop to the item above
<chillywilly> that screen is too busy
<dneighbo> via master detail relationship
<dneighbo> just masetr and detail would be same tables
<dneighbo> chillywilly i agree
<dneighbo> i dont see the need to do the lists in this application
<MrChumple> It is busy...
<MrChumple> It is clunky...
<dneighbo> a popup box solves all these problems
<MrChumple> It is fugly...
<dneighbo> i.e. you would just have the entry protions
<dneighbo> not the lookup portions
<MrChumple> But I actually saw secretary types using it and navigating it pretty easily...
<dneighbo> and if you wanted to do a lookup outside of scrolling through rcorts
<dneighbo> you simple use a search dialog
<MrChumple> Yes...
<dneighbo> that is same thing as they are using the lists for here
<dneighbo> so the lists are on 'demand' only
<MrChumple> Their app allows partially filling that scroll box by doing a 'LoadBy'
<dneighbo> if that makes sense
<MrChumple> where you can choose something to search by...
<dneighbo> that would be same in our search dialog
<dneighbo> where you could filter down
<MrChumple> With a large enough database, you wouldn't want to allow the entire thing to be displayed...
<dneighbo> as you go
<MrChumple> But their loadby thing is actually a filter-- and it's powerful to be able to search kinda like that...
<MrChumple> And you guys have that too -- 
<dneighbo> yip we would have that in popup search dialog
<dneighbo> its not implemented teh popup search
<dneighbo> but it could be
<dneighbo> and is on the TODO
<dneighbo> its my biggest gripe currently
<MrChumple> But it could be done in something that didn't necessarily pop up right?
<dneighbo> you can do it in a non pop up right now
<dneighbo> you can do prequery=""
<MrChumple> The part that says 'How do you want to search' pops up, but the results show up in the parent window's list box after you click 'Find'
<dneighbo> to default the filter to something at startup
<chillywilly> doww wxwinders have an xml GUI description?
<dneighbo> then you can query in teh form as well
<dneighbo> chillywilly originally they didnt now they might
<chillywilly> hmmmm
<dneighbo> MrChumple i would say in future you could do that
<dneighbo> but i think its bad design
<MrChumple> which part being bad design?
<dneighbo> and would recommend using the pop search box
<dneighbo> having that list in the main window
<dneighbo> as after you do the search it serves no real purpose
<dneighbo> so why take up the screen realestate for it?
<MrChumple> You might need to browse through them for a while...
<dneighbo> i could hear an argument for contact
<MrChumple> Not just contact...
<dneighbo> not much of one address/communications
<dneighbo> its a preference thing
<dneighbo> i wont debate preference
<dneighbo> you should be able to do these things in gnue
<MrChumple> Imagine a shipment/fulfillment app that you have to go through... one item at a time
<dneighbo> right now you can do someof them
<dneighbo> MrChumple i have written one
<MrChumple> and move each item in the Invoice into a 'shipped' status...
<dneighbo> and its lightinging fast we never display the items
<dneighbo> it all by itemid
<dneighbo> with search box
<dneighbo> wihtin a week the entry people have codes memorized
<dneighbo> they enter itemid and qty
<dneighbo> everything else is auto populated
<dneighbo> if they cant remember an id
<dneighbo> they pop open teh search
<dneighbo> and type description of item
<dneighbo> it filters as they type
<dneighbo> then tehy hit enter or click on item
<dneighbo> and it populates teh itemid
<dneighbo> can you imagine a screen where someone had to scroll a whole list to pick an item every time?
<dneighbo> to do fullfillment?
<dneighbo> think grocery store
<MrChumple> No - I'm talking something different
<dneighbo> why do you think the clerks memoriz codes for produce?
<dneighbo> instead of look at a list
<MrChumple> Sometimes it's one way, sometimes the other
<dneighbo> when i worked at safeway long ago they would get fired if they could pass the produce test
<dneighbo> MrChumple i agree
<MrChumple> When you want a list, you shouldn't have to memorize codes
<dneighbo> point is you should be able to do it how you want
<MrChumple> Yes
<dneighbo> so lets not debate it anylonger
<MrChumple> I'm just wanting to know if I could do it how I want :)
<dneighbo> right now today you can do it hwo you want
<dneighbo> you just cant have the 'find' option how you want it
<dneighbo> though you cna filter
<dneighbo> you be here for a few minutes?
<MrChumple> Scrolling through too much data is definitely an obstacle to productivity
<dneighbo> maybe i will whip up a demo
<MrChumple> For another 15 minutes or so...  Then I've gotta run until tomorrow...
<dneighbo> you have forms installed right?
<MrChumple> Yip
<MrChumple> Do you guys use Python 1.5?
<MrChumple> when developing?
<dneighbo> i do but moving to 2.x soon
<MrChumple> When I went to install forms, I downloaded the most recent Python (2.1.1 or something like that) and I think I had compatibility issues
<MrChumple> with the MySQL db driver on a couple lines of code
<MrChumple> once I changed the lines, it was cleared up...
<MrChumple> I don't know how Py1.5 and 2.1 are different, so I can't say for sure what the problem was...
<MrChumple> or if it would have worked with 1.5
<MrChumple> What's the Python guru's most compelling argument against using Java instead of Python? (i'm new to Python)
<dneighbo> no one tested mysql for a bit
<dneighbo> so coudl be a driver issue not a python one
<MrChumple> I just changed a line, and added an import and it ran fine
<MrChumple> in DBDriver.py I think
<dneighbo> you still here
<MrChumple> ya
<dneighbo> think i may have sample
<MrChumple> just trying to learn more about pgsql
<MrChumple> i have it installed and working...  was working on making inventory work --
<MrChumple> awesome!
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.
<MrChumple> (inventory the sample app)
<dneighbo> sigh maybe not
<dneighbo> the damn drivers are borked
<dneighbo> forgot runnign cvs on this machine
<MrChumple> the DBDrivers?
<dneighbo> give me antoher minute see if i can workaround it
<dneighbo> yeah dbdrivers
<MrChumple> :) dbdrivers for pgsql or mysql?
<MrChumple> I may have a fix if they're the mysql ones...
<dneighbo> my bad
<dneighbo> i think i have done somethign stupid :)
<dneighbo> like forgot to define which table in my datasource
<dneighbo> much better
<MrChumple> :) 
<dneighbo> still no dice though for this because of different driver issue im afraid
<dneighbo> let me try one other
<MrChumple> :( sorry i'm of little help
<dneighbo> had a sequence in postgres wrong
<dneighbo> teach me to cut and paste and change only have
<MrChumple> doh
<MrChumple> postgres is pretty cool
<MrChumple> pgaccess nifty
<dneighbo> for some reason the m/d not working 
<MrChumple> m/d?
<dneighbo> doh
<dneighbo> m/d = master detail
<MrChumple> ok
Nick change: dres__ -> dres
<MrChumple> Which Python PGSQL driver do I need to have installed for it to work with the forms code?
<MrChumple> It fails on import pg for me, but I'm not sure which driver I need to download - there's a few - PyGreSQL
<MrChumple> and a few others
<dneighbo> well i think maybe its screwed
<dneighbo> looks like you cant have a m/d using same table
<MrChumple> i'm sorry for running you on a goose chase
<dneighbo> i believe this used to wrok
<dneighbo> no its a good thing really
<dneighbo> one more thing to try
<MrChumple> Do you know which Python Postgres module it is that provides 'pg'
<MrChumple> My gfclient inventory.gfd fails and says it's failing in DBdriver.py for postgresql on the 'import pg' line
<MrChumple> 'gfclient  inventory.gfd' was actually 'gfclient company.gfd' in the samples/inventory/forms directory of
<MrChumple> Forms 0.0.9
<MrChumple> Where do I need to go to get PostgreSQL drivers for Python?  I'm at some place called parnassus?...
<MrChumple> gg - i'll get back on tomorrow -
<MrChumple> Thank dheighbo for your help -
Action: MrChumple will be back tomorrow...
MrChumple (mrchumple@216.250.143.132) left irc: Client Exiting
<dneighbo> sigh no jamest or jcater when needed
<dneighbo> i think they have successfully hosed the drivers :)
<chillywilly> oh yea
<chillywilly> cvs is nice and borked
<chillywilly> imho
jcater (jason@1Cust93.tnt5.memphis3.tn.da.uu.net) joined #gnuenterprise.
<dneighbo> pounce
<dneighbo> im having a bitch of a time with new datasources help
<dneighbo> help
[18:34] Last message repeated 2 time(s).
<chillywilly> use the stable release and quit your bieotching
<chillywilly> :P
<jcater> what's the problem?
<dneighbo> chillywilly thats coming from mr use sid or die?
<dneighbo> jcater cant seem to get masterdetail on same table to work
<dneighbo> at one point i remember jamest doing a demo
<dneighbo> where he had the same table be the master and teh detail
<dneighbo> and i thought how dumb
<chillywilly> sid isn't that bad only a few packages do cvs snapshots
<dneighbo> now some user asked if forms could do something
<dneighbo> i said it could but was under guise that this worked
<dneighbo> i have tried like 10 different ways to do this with no luck
<jcater> I see no reason forms would care (i.e., know) that a master and detail reference the same db table
<jcater> what are you trying?
<dneighbo> sigh i hate irc for explainign thigns
<dneighbo> took me half day to get what this fellow wanted :)
<dneighbo> in a nutshell
<dneighbo> say i have
<dneighbo> feilds
<dneighbo> foo bar stud
<dneighbo> in a table
<chillywilly> use the phone
<dneighbo> i want an entry form to have
<dneighbo> foo
<dneighbo> bar
<chillywilly> sheesh
<dneighbo> stud
<dneighbo> er
<dneighbo> foo:
<dneighbo> er
<dneighbo> foo: ____________
<dneighbo> bar: _________
Action: chillywilly should write a virtual helpdeak program
<dneighbo> stud: _________
<chillywilly> helpdesk
<dneighbo> then below that one of our fake grids
<dneighbo> foo     bar
<dneighbo> _______ ___________
[18:40] Last message repeated 4 time(s).
<dneighbo> the idea is as you roll through the 'read only' bottom grid
<dneighbo> the top entries change to the entry you are on
<dneighbo> then you would hop up to edit it in its entirty
<dneighbo> to me this is a poor way to do record searching
<dneighbo> and i would do this via popup search box we discussed but thats just me
<dneighbo> http://216.250.140.227/lists_edits.png
<dneighbo> is a sample application that gets the point accross visually
<dneighbo> i built a simple contact
<dneighbo> that has 
<dneighbo> first name
<dneighbo> last name
<dneighbo> title
<dneighbo> company
<dneighbo> for the top part
<dneighbo> and displays 
<dneighbo> firstname
<dneighbo> lastname
<dneighbo> in the 'grid' part
<dneighbo> BUT i cant get master detail to work at all with them
<jcater> food is here
<jcater> um, master/detail would only work in this case
<jcater> if no wildcards were used
<dneighbo> what do you mean wildcards?
<jcater> err
<jcater> nevermind
<jcater> I was thinking backwards
<jcater> master/detail should work
<dneighbo> what was REALLY odd
<jcater> with the "grid" being the master
<dneighbo> right
<dneighbo> thats how i had it setup
<jcater> and the top being the detail
<dneighbo> and it didnt like it
<jcater> and what is happening?
<dneighbo> didnt barf
<dneighbo> but didnt work
<jcater> ah
<jcater> did you look at select statements being generated?
<dneighbo> then i put the tag stuff on the detail (where it shuld be)
<dneighbo> and it pissed itself saying current not valid attribute
<jcater> I will test in half hour if you can wait
<jcater> food is here :)
<dneighbo> attributeError: 'none' object has no attribute 'addDetailDataObject'
<dneighbo> ok
<jcater> this **should** work the way you expect
<dneighbo> i know thats why im freaking
<dneighbo> well not freaking pounding head in wall :)
<dneighbo> going this should work
<dneighbo> go eat
mdean (mdean@arc10x26.kcnet.com) joined #gnuenterprise.
<dneighbo> sup mdean 
<dneighbo> i think i might have converted jason (jcater) to dcl
<mdean> howdy
<dneighbo> neither of us like web
<dneighbo> so you might see a gnue front end soon than later :)
<mdean> cool
jcater (jason@1Cust93.tnt5.memphis3.tn.da.uu.net) left irc: Ping timeout for jcater[1Cust93.tnt5.memphis3.tn.da.uu.net]
jcater (jason@1Cust186.tnt4.memphis3.tn.da.uu.net) joined #gnuenterprise.
jamest (jamest@fh-dialup-202033.flinthills.com) joined #gnuenterprise.
<dneighbo> another victim
<dneighbo> jamest you really here/
<dneighbo> i hav been slaughter the lamb again
Nick change: jamest -> johndoe
<johndoe> hi.  I'm new to GNUe
<johndoe> someone told me you like goats
Nick change: johndoe -> jamest
<jamest> sigh, so how would you like to abuse me tonight
<dneighbo> oh let me count the ways
<dneighbo> can you read the irc log
<dneighbo> the last 30 lines or so with me talking to jcater
<jamest> depends
<dneighbo> rather than me trying to repeat it
<dneighbo> i think he is willing to take blame
<dneighbo> when he gets back from dinner
<dneighbo> so you are fairly safe
<dneighbo> ;)
<jamest> http://www.gamespy.com/comics/dorktower/
jcater (jason@1Cust186.tnt4.memphis3.tn.da.uu.net) left irc: Ping timeout for jcater[1Cust186.tnt4.memphis3.tn.da.uu.net]
<jamest> go jcater!  you can make it!
jcater (jason@1Cust186.tnt4.memphis3.tn.da.uu.net) joined #gnuenterprise.
<jamest> damn!
<jamest> better luck next time jason
<jamest> was it the fence or the dogs that stopped you
<dneighbo> jamest: he is beta testing the new neck collar
<dneighbo> like in running man
<dneighbo> you go 300 yards from pc it shocks teh hell out of you
<dneighbo> 500 yards it blows your head off from neck up
Action: dneighbo hopes he ordered take out
<dneighbo> jamest so no ideas ?
<jamest> or worse...cuts off your caffine supply
<jamest> i'm still dl'ing logs
<jamest> don't ask
<dneighbo> ok
<dneighbo> wont ask
jcater (jason@1Cust186.tnt4.memphis3.tn.da.uu.net) left irc: Ping timeout for jcater[1Cust186.tnt4.memphis3.tn.da.uu.net]
jcater (jason@1Cust12.tnt7.memphis3.tn.da.uu.net) joined #gnuenterprise.
<jcater> damn msn!!!
<jamest> ok, this used to work
<jamest> but I agree w/ jcater that it should still work
Action: jcater thinks he's testing us
<jamest> dneighbo: do me a favor
<jamest> step about 30 yards away from pc and have someone else run the form
<dneighbo> it dont like GFDatasource.py line69
<jamest> see if it works then
<jamest> holy crap
<dneighbo> it is telling me that self.name and self.master are correct values
<jamest> whew
<jamest> i just looked at the png
<dneighbo> but addDetailDataObject doesnt seem to be valid for self.master
<jamest> wondered what the !@#!@# you were doing with our baby
<dneighbo> so many nested objects here getting dizzy so calling the pros :)
<jamest> jcater you going to mess with this or should I
<jcater> um
<dneighbo> i can send you SIMPLE sql file
<dneighbo> and the form
<dneighbo> to test it
<jcater> send it, baby... send it
<jcater> oops, wrong channel
<jamest> if you're going to mess with it I wanna start the new geas driver now
<jcater> icky
<jamest> and working on GComm.py
<jcater> have fun
<jcater> what's GComm?
<jamest> an orb/rpc abstraction layer (working title)
<jcater> ah
<jcater> I was contemplating that last night
<jamest> we need a non-communication specific place to add drainDereksCheckingAccount functions
<dneighbo> omniORB, omniORB, omniORB
<jcater> (did you see my split of GClientApp into GClientApp, GBaseApp, and GServerApp
<jamest> that way whatever communication system he uses
<jamest> we'll still get paid
<jamest> jcater: yes
<jamest> have not looked into the code yet
<jcater> so your GComm fits in w/my TODO's in there :)
<jamest> why did you split?
<jcater> schizophrenic
<jcater> Report server will use GServerApp
<dneighbo> http://goats.gnue.org/~dneighbo/gnue/sample.gfd
<dneighbo> http://goats.gnue.org/~dneighbo/gnue/sample.sql
<dneighbo> are the two files
<jamest> but what are major diffs? or is this doc'd somewhere?
<jcater> there are none yet
<jcater> hadn't gotten that far
<dneighbo> btw my cvs might be a few days stale
<jamest> dneighbo: update now
<dneighbo> so if one of you can run that gfd after creating the db and let me know
<dneighbo> jamest : i cant
<dneighbo> im at work, or i would have
<jamest> datasource got their bars in master/detail's foo
<jamest> jcater: but what are the diffs you forsee for client vs server?
<dneighbo> will either one of you be able to look at that form in next 10 minutes?
<dneighbo> other wise i will take off and comeback later
<jcater> forking, threading, daemon background mode, attaching to RPC mech
<jcater> like I said, hadn't gotten that far, but thought there would be minor differences
<jcater> most functionality is in GBaseApp which both descend from
<jcater> dneighbo: looking now
<jamest> ok
<jcater> jamest:  of course, I could be completely off base
<jamest> i was going to structure the base Comm system like the dbdrivers
<jcater> :)
<jcater> Amazing
<jamest> what?
Action: jcater was thinking same thing last night
<jamest> dneighbo: so does this mean the jcater AI is becomming more human or an I becoming less?
<jamest> an = am
<jamest> :)
<dneighbo> neither, we are programming him for empathy
<jamest> jcater: so we setup a commdrivers dir
<dneighbo> so you will never know if he really was thiking that or if he is just trying to empathize with you inferior intellect
<jamest> in there put orbit-python, omniorb-python, etc, etc
<dneighbo> i mean for good ness sake he is using MSN and windows to try to empathize with those less fortunate souls
<jcater> soap, xml-rpc
<jamest> then put the prefered orb setting in to the gnue.conf
<jamest> right on the soap, xml-rpc, rpc, gnue-leetspeed-2000
<jamest> are we still on same page?
<dneighbo> sigh
<dneighbo> unfortunately no
<dneighbo> this sucks
<jcater> I guess... you were thinking xml-rpc, soap, etc later?
<dneighbo> honestly its the RIGHT thing
<jcater> in the commdrivers structure?
<jamest> no, i was just giving examples
<dneighbo> but GEAS needs this SAME functionality
<dneighbo> so if you do this in gnue-common
<dneighbo> that means it will just be rewritten in C
<dneighbo> and that will just be silliness
<jcater> well
<jamest> dneighbo: not necessarily
<dneighbo> so it might be best to go out of your way to please see if you can make this work with geas too
<jamest> I've offered to do c/python interfaces
<jamest> dneighbo: always my intension
<jamest> intention
<jamest> and they're talking complete rewrite of GEAS anyway
<dneighbo> are you serious?
<jamest> as they found scary things in there
<jamest> very scary
<jcater> um
<dneighbo> if they are it really really really should be considered to do it in python
<jcater> rewrite????
<jcater> um
[19:40] Last message repeated 1 time(s).
<jamest> as when doing queries
<dneighbo> they would get a lot more help
<jamest> i guess geas creates a temp table in the backend
<jamest> that it populates with field query values
<dneighbo> would unify the team and we could always convert it to C later
<jamest> that it then uses to construct the sql statement for the backend
<dneighbo> i.e. use python to prototype it
Action: dneighbo shoves fingers in ears and keeps screaming
Action: dneighbo doesnt want to hear this
<jamest> at least that was my understanding
<jamest> read the logs
<jamest> that doesn't require ears
<dneighbo> roflmao
<dneighbo> that was a good one
<jamest> i hope that we can work together 
<dneighbo> jcater : you testing that form out over there?
<jcater> kinda
<jcater> I can't pick myself up off the floor
<jamest> jcater: are you and I still on same page
<dneighbo> hahaha
<jcater> jamest: as always
Action: dneighbo is worried that if someone gets jcater caffiene
<jamest> as derek's almost always on the page with the MastaWhipper2000 ad
<dneighbo> there could be a python geas in the morning when masta wakes up
<jamest> flogs up to 2^8 coders at a time
<jcater> jamest: how can I pass the password to psql?
<dneighbo> -p yourpassword
<jamest> gack!
<dneighbo> psql -U username -p password tablename
<dneighbo> iirc
<jamest> then it's in your bash history file!
<jcater> damn damn damn
Action: jamest smacks masta!
<dneighbo> hey he asked
<jamest> :)
<jcater> postgres quit working on laptop again
<dneighbo> smack him instead
<jcater> dneighbo: can't look now as postgresql is bitch-slapping me
<dneighbo> remember my idea of security on postgres is open trust
<jamest> psql -h shodan -u 
<dneighbo> who needs stinking passwords?
<jamest> wil lprompt for username and password
Action: jcater is getting an itch to take off work and rewrite geas
<jamest> are we going to python 2.? officially?
<jamest> hell, I'll stick w/ 1.5.2 for now
<jcater> well, seeing as how they will release 2.2 soon, might as well
<jamest> which version should we jump too?
<jcater> 2.1
<jcater> :)
<jamest> 2.0
<jamest> ?
<jamest> 2.0+
<dneighbo> i say 2.1 should be our goal at this point
<jamest> why 2.1?
<jamest> woody contains 2.0 IIRC
<dneighbo> 2.0 to me is more like a beta
<jcater> it has quite a bit more functionality in libraries
<jcater> that 2.0
<jcater> like ability to intercept exceptions
<jcater> et al
<dneighbo> if we require something other thna 1.5.2 we can probably dictate
<jamest> IIRC 2.1 had better C interface too
<jcater> but I don't really care
<dneighbo> i would say 2.2 but thats too new
Action: jcater will support whatever
<dneighbo> 2.1 will probably go in the new distros 
<dneighbo> way i look at it we have such small user base
<dneighbo> go more bleeding edge now
<dneighbo> as once the user base comes it will be hard to change
<dneighbo> 2.1 might seem iffy now, but in 9mos when we have fist full of apps and or users
<dneighbo> we will say damn should have done 2.2 :)
<dneighbo> but i think 2.x would be fine
<dneighbo> i would like to get out of 1.5.2 as there are some good things in 2.x
<jamest> woody contains 2.0.1
<dneighbo> ok lets say 2.x then
<dneighbo> for now
<jcater> until tomorrow?
<jamest> which means next stable debian will have 2.0
<dneighbo> jcater shhhhh
<jamest> but I'm happy w/ 2.1
<jamest> just one more dependency to download :)
<dneighbo> jamest how about this
<dneighbo> i will install 2.1
<jcater> hey, still a fraction of gnucash :)
<dneighbo> if i can get it to install we will use it :)
<dneighbo> else 2.x :)
<jamest> lol, jcater I was about to write that myseld
<dneighbo> if masta can get it to install on his crippled ass machine then i dont mind subjecting others
<jcater> roflmao
<jamest> jcater: great, you know masta install will require us to use 1.3
<jcater> :)
<dneighbo> not ture i have 1.5.2 w/ no fixes running now :)
<dneighbo> but that did install by default
<jcater> lets just rewrite everything in emacs' lisp engine
Action: dneighbo looks for chillywilly 's trout
Action: chillywilly gives all GNUers a good trout slappin'
Action: jcater wants fish for dinner
<jcater> gotta run... wifey expects "quality time"
Action: jamest 's mad goat wonder where fish smell coming from
Action: jcater thought he was having quality time w/his laptop & msn connection
<jamest> lol
<dneighbo> fish for dinner
<dneighbo> wifey time
<jamest> did you get masta's form looked at
<dneighbo> i wont go there
<jamest> ?
<jcater> y'all be on later?
<jamest> yip
<jcater> jamest: postgresql is bitch-slapping me again on my laptop :(
<dneighbo> oh wait he is married never mind :)
<dneighbo> bbl
dneighbo (dneighbo@mail.libertydistribution.com) left irc: [BX] That's Mister BitchX to you
<jcater> on my laptop=with my laptop
<jcater> bbl
jcater (jason@1Cust12.tnt7.memphis3.tn.da.uu.net) left irc: later
<SachaS> bye
SachaS (sacha@chime-a-181.conceptual.net.au) left #gnuenterprise (Client Exiting).
ToyMan (stuq@c5300-3-ip155.albany.thebiz.net) left irc: Ping timeout for ToyMan[c5300-3-ip155.albany.thebiz.net]
chillywilly_ (danielb@d138.as28.nwbl0.wi.voyager.net) joined #gnuenterprise.
chillywilly (danielb@d141.as9.nwbl0.wi.voyager.net) left irc: Ping timeout for chillywilly[d141.as9.nwbl0.wi.voyager.net]
Nick change: chillywilly_ -> chillywilly
derek (derek@cpe-24-221-112-50.az.sprintbbd.net) joined #gnuenterprise.
#gnuenterprise: mode change '+o derek' by ChanServ!s@ChanServ
<chillywilly> masta masta, where's the pay chekc that I've been afta....
<derek> who knows
<derek> must be lost in the mail
<derek> jamest: you get that master/detail single table debugged and fixed?
<derek> is website spelled web site or website
<jamest> derek: update cvs and try again, no I haven't touched it
<derek> ok
<derek> will have to wait
<jamest> who's here?
<jamest> chillywilly? derek?
<jamest> jcater?
<jamest> elvis?
Mr_You (rouzer@mankind.boredom.org) left irc: ircII EPIC4-1.0.1 -- Are we there yet?
<jamest> anyone?
<chillywilly> yes?
<jamest> ok, I'm working on geas driver
<jamest> however I'm also working on what I'm currently calling GComm
<jamest> which is a abstraction layer for various communication protocols (ORBs, XML-RPC, SOAP, etc, etc)
<jamest> that I'll build the geas driver on top of
<jamest> it'll use drivers much like dbdrivers do
<chillywilly> ok
<jamest> I'd love to see this available to both python and C apps
<jamest> but I need input, lots of input
<chillywilly> that's th e3rd thing we have in common
<chillywilly> yes!
<chillywilly> you have helped me remember
<jamest> :)
<derek> input, input, input, input, input, input, input, input
[21:36] Last message repeated 8 time(s).
[21:36] Last message repeated 1 time(s).
<chillywilly> comm stuff, scripting, and db abstraction layer
<derek> input, input, input, input, input, input, input, input
[21:36] Last message repeated 1 time(s).
<jamest> still have a long way to go but I'm testing the framework 
<derek> input, input, input, input, input, input, input, input
[21:36] Last message repeated 6 time(s).
<chillywilly> stop that!
<derek> input, input, input, input, input, input, input, input
[21:36] Last message repeated 2 time(s).
<jamest> chillywilly: break out the trout
<derek> my big item is lets not do the work
Action: chillywilly whacks derek with a trout
<derek> if geas is gonna say they cant use it
<derek> like has been done with dbdrivers
<derek> i.e. get thier input first
<jamest> that's why I'm asking chillywilly for input :)
<chillywilly> just write it in C with python bindings
<chillywilly> ;)
<derek> as its getting silly that we arent reusing things :(
<jamest> I really want to prototype in python w/ C interface
<derek> as to the approach it sounds great to me
<chillywilly> I am gonna do the scripting lib thi sway I think
<jamest> and covert the things that require more speed
<derek> but i thought the db driver approach was great
<chillywilly> well
<chillywilly> sig can make a C interface to python code
<chillywilly> swig
<chillywilly> i really don;t care
<jamest> james pukes
<chillywilly> if it is python
<jamest> I've used swig
<chillywilly> well i just heard about it
<chillywilly> it sucks?
<jamest> it's like the 3rd ring of hell
<chillywilly> eeeek
<chillywilly> ok
<jamest> yes, it sucks
<jamest> but it can be done
<chillywilly> how would you make a C API?
<jamest> chillywilly: that is what we need to figure out :)
<jamest> one thing early in design I'm wondering about
<jamest> how do you configure such a system
<derek> um i think jamest said he likes the sunburn swig leaves and wants to attempt that route ;)(
<jamest> iow what are our needs wrt communicatoins
<jamest> do we want multiple comm channels open 
<chillywilly> nah
<chillywilly> a method selectable at donfig time
<chillywilly> config
<chillywilly> imho
<jamest> well chillywilly
<chillywilly> yes?
<jamest> lets say I'm running geas in purchasing and I'm a huge orbit fan
<jamest> and derek being the dog he is is running geas in marketing using soap
<jamest> now I want my geas to talk to your geas
<chillywilly> hmmmm
<jamest> a single comm channel isn't going to work very well :)
<mdean> can I offer a suggestion out of the blue?
<chillywilly> you whack derek over the head real hard and make him use corba :P
<jamest> mdean: as long as you're not derek we're all ears
<chillywilly> mdean: sure dude
<mdean> hehe
<derek> chillywilly: you work for m$ dont you
<chillywilly> uh no
<mdean> wouldn't it be reasonable to use a uri to specify the comm service?
<chillywilly> SOAP is an M$ protocol isn't it?
<derek> i vote we listen to mdean cause his app is saving me a shit load of time
<chillywilly> and derek you were using SOAP
<chillywilly> mdean: that's sound like what DotGNU wants to do
<chillywilly> RLS: reoursce locator string or some shit
<chillywilly> resource
<jamest> mdean: like?  orbit-python:\\www.mastaserver.com
SachaS (sacha@chime-a-181.conceptual.net.au) joined #gnuenterprise.
<mdean> ja - something like that
<jamest> sounds reasonable
<mdean> gets around the config issue nicely 
<jamest> then do we put these uri's in a config file
<chillywilly> hmmmmmm
<jamest> /usr/local/gnue/etc/comm.conf
<jamest> and how does this work on the server end?
<chillywilly> we parse the config file
<chillywilly> ;)
<chillywilly> wow common config files?
<chillywilly> it can't be?!?
<chillywilly> we're getting too out of control
<jamest> chillywilly: I was hoping we could use GConfig for the comm.conf file
<jamest> chillywilly: as it's basically .INI 
<chillywilly> uuuuh, ok where my C API?
<chillywilly> ;)
<jamest> chillywilly: and we've got a config engine built arround it
<jamest> I'll provide it if people like the idea
<chillywilly> I think a common config system is good
<mdean> jamest: would there be a client library that can handle the URI request?
<jamest> could you expand on that?
<mdean> jamest: the client would know how to deconstruct the URI and contact the server properly
<jamest> if it works like our db system in gnue-common then you'd pass the uri in
<jamest> and be handed back a python object (or c pointer) to a connection
<jamest> now how that connection will represent objects isn't clear yet
<jamest> and how a server (w/o a URI to pass in) would init it's own valid connections isn't clear yet
Action: jamest did say he needed lots of input :)
<mdean> I wish I knew more of your internals to be of better help
<jamest> that's not stopping me !
<jamest> :)
<mdean> heh
<jamest> did I scare people off
<jamest> ?
<mdean> no - I just don't know where to go from here
<mdean> I think that URIs provide a good way of specifying protocol and location, so I butted in ;-)
<chillywilly> no
<chillywilly> jsut arguing in another channel
<chillywilly> with a gay hjppy homosexual libertarian
<mdean> chillywilly?  argue?  unheard of!
<chillywilly> (his words not mine)
<chillywilly> mdean: you don't know me well enough to make such assrtions do you?
<chillywilly> maybe you do...
<mdean> I've seen (read) you in action a few times ;-)
<chillywilly> or mabe dtm was complaining
<chillywilly> well I try damnit
<chillywilly> fuck it all and fucking no regrets
<jamest> chillywilly: input on comm system?
<chillywilly> I like the uri deal
<chillywilly> nice one mdean;)
<mdean> :)
<chillywilly> what do youb do then make a corba-to-soap tunnel or something?
<chillywilly> ctatp
<chillywilly> er
<chillywilly> ctstp
<jamest> that i'm not sure about
<chillywilly> corba to soap tunnel protocol
<chillywilly> tunneling
<jamest> I wouldn't think so
<mdean> you could, but don't have to
<chillywilly> ok
<chillywilly> elaborate then
<jamest> I'd think more like a something <-> internal representation of object or interface
<jamest> so that anything that is handed back by GComm acts the same
<jamest> and the other side is native
<chillywilly> we have an internal GEAS  object representation I think.....
<jamest> i imagine we'd need introspection here
<mdean> ja - like interface://service.location
<jamest> well maybe not introspection
<jamest> chillywilly: IIRC you are right about that
<jamest> so we'd have a new internal object spec
<jamest> that in C looks like foo and in python looks like python object
<chillywilly> you need to change from one to the other then let it do its thing
<chillywilly> internal to CORBA to SOAP to client?
<jamest> internal to published external
<jamest> no CORBA middle step
<chillywilly> maybe we should shoot for one thng at a time and allow protocol mixing later
<jamest> if geas is told to produce interfaces via CORBA and SOAP
<chillywilly> ir in th next iteration of desgin
<chillywilly> oh yea
<chillywilly> just use GComm
<chillywilly> duh
<chillywilly> GComm doe sht edirty work right?
<chillywilly> and I can't type
<jamest> that's is whole reason for existance
<chillywilly> of course
<chillywilly> my brain was blobked for a momenr
<chillywilly> moment
<chillywilly> blocked
<chillywilly> man I am really tired
<chillywilly> ok, so what happens is you create an idl or whatever soap uses and you generate stubs that marshall calls etc. right?
<chillywilly> GComm has to wrap this up somehow
<chillywilly> er wait
<chillywilly> we would have to do GComm stuff
<chillywilly> that would be nternal somehow too
<chillywilly> onyl thing they should have to do is have Gcomm installed and working and then just use it instead
<chillywilly> and it magically does all the crap that orbit, soup, etc. does
<chillywilly> using one or a combination thereof
<jamest> more or less yes
<chillywilly> thing is
<chillywilly> you would still have to define the stubs right?
<jamest> no
<jamest> well maybe
<jamest> needs more research :)
<chillywilly> heheh
<jamest> i know CORBA doestn't require the stubs
<jamest> via it's dynamic something or rather
<chillywilly> nah but langaiges like C and C++ need the stubs I think
<jamest> :)
<jamest> it's going to be "fun"
<chillywilly> python doesn't
<chillywilly> they just read the idl in
<chillywilly> hmmm
<jamest> i don't think you'd want a pure idl
<jamest> as we need to look at how RPC, XML-RPC, etc, etc, do it
<chillywilly> I think I am getting hung up on detailos
<jamest> :)
<jamest> and I think I'm getting too tired to think about it :)
<chillywilly> ;)
<chillywilly> well we do have a an object definition language already
<jamest> the gcd
<chillywilly> and I know that geas objects do not map directly to corba objects
<jamest> but geas doesn't currently pass out objects
<chillywilly> a corba refernce to a datab object
<chillywilly> data
<jamest> yip, we were thinking same thing
<chillywilly> dataobject.idl
<chillywilly> it is a bit of work but I am willing to work on it with you whenever you are ready ;)
<jamest> :)
<jamest> "a bit" he says
<jamest> sounds like a royal bitch to me :)
<chillywilly> that's an understatement
<derek> jamest: he must think jcater is writing it ;)
<chillywilly> derek: how so?
<jamest> jcater is a code machine
<chillywilly> yes I know
neilt (neilt@dialup-209.244.81.186.Dial1.Washington2.Level3.net) joined #gnuenterprise.
#gnuenterprise: mode change '+o neilt' by ChanServ!s@ChanServ
<jamest> bwahahahahaha
<jamest> just the man
<chillywilly> heheh
<chillywilly> run neilt!
<chillywilly> run!
<derek> in fact you two argue about it for much longer
<jamest> jcater will have it done
<chillywilly> who?
<derek> your likely to find your answer in cvs if jcater is listening :)
<jamest> that's my plan
<chillywilly> heheh
Action: neilt trips and falls
<derek> jamest: i was just gonna say its not an all bad plan :)
Action: chillywilly helps neilt up
Action: neilt has fallen and can't get up
<jamest> neilt: you have access to the logs ?
<jamest> derek: what's not a bad plan?
Action: derek calls medialert
<derek> arguing until jcater commits it
<neilt> if they work, yes
<chillywilly> we are not arguing though
<chillywilly> we are brain storming
<jamest> yip
<chillywilly> ;)
<derek> im teasing when i say arguing
<jamest> neilt: they work
<chillywilly> we are tryin to get our non-AI minds to comprehend
<chillywilly> wes should just let the superiror AI implement it eh?
<jamest> neilt: I made mistake of starting geas driver work
<jamest> and was planing to abstract out the communication methods
<neilt> cool
<neilt> and
<jamest> also wanted to work w/ geas team so we share stuff
<chillywilly> and it is a thing we all need
<jamest> plan was to do it so that it's a communications object (C pointer in C and python object in python)
<chillywilly> and it does the communications tranparently for us
<chillywilly> and can mix and match too
<chillywilly> or something
<jamest> planned on supporting multiple channels
<jamest> man, if you can read the logs
<jamest> it'll save me a ton of typing :)
Action: chillywilly has a log
<chillywilly> an xchat log
<jamest> the web works are working
Action: neilt goes to read logs
<neilt> yuck, yuck yuck
<derek> like it that much
<neilt> what did you do the gnuenterprise
<neilt> i quit
<chillywilly> heheh
<mdean> lol
<neilt> tell this is a joke
<neilt> jeeze its 145 k
<jamest> huh?
<chillywilly> uh?
<neilt> the log for today
<jamest> oh read the end of it
<jamest> let me get you a line to search for
Action: chillywilly thinks neilt is smoking some good stuff
<neilt> you know if it had times
<neilt> i could go right to the pertaintent part
<derek> how is it 145k?
<derek> man, i dont recall there being much going on today?
<chillywilly> 1:45, ok
<chillywilly> ?
Action: derek must have been in an alternative gnue universe or something
<chillywilly> oh it is a big ol' log
<neilt> i think the size thing must be hosed
<neilt> it loaded too quick for 145k
<derek> ah
<jamest> <jamest> that's why I'm asking chillywilly for input :)
<derek> no its right i think
<jamest> search for that
<derek> too much pasting in the main channel
<chillywilly> yep
<derek> we need to start #gnueflood
<derek> when doing tech support
<chillywilly> I blame you derek
<derek> lots of xml file and such being pasted
<derek> in supporting forms
<neilt> found it
<derek> and lots of libuuid posting
<derek> and me giving jamest lots o input
<chillywilly> it si all your fault derek
<jamest> anyone but derek can give input
<jamest> he gave enough "input" before when I asked for it
<chillywilly> :P
Action: derek is away: Reloading Crack Pipe
<jamest> hmmm neilt is awful quiet, must be hard to type when fists clenched in rage or shaking with laughter
<neilt> ok
<neilt> this part makes sense
<neilt> gay hjppy homosexual libertarian
<neilt> dont know about the rest
<neilt> did i read too far
<chillywilly> heheh
<chillywilly> he's in #gnu
<chillywilly> he's a real weirdo
<chillywilly> neilt: http://www.spectacle.org/897/finkel.html
<chillywilly> ;P
<jamest> neilt: so it makes no sense?
<chillywilly> when has GNUe ever made sense ;)
<neilt> too many tla's
<neilt> and too late
<jamest> gotcha
<jamest> bottom line
<neilt> maybe i'll re-read tomorrow
<jamest> GComm would have drivers for things like SOAP and ORBs
<chillywilly> but CORBA is a FLA :P
<neilt> as i have said repeatably, "we have the need for speed"
<jamest> we have the need for scalability
<chillywilly> that has nothing to do with it
<neilt> can we walk before we run
<chillywilly> I thought you agreed we do nto want to be tied to CORBA
<jamest> my understanding was you wanted to pull orbit out of geas
<jamest> so i started looking at what that involves
<neilt> that wont happen this year
<jamest> ah
<neilt> i want to support other ways to communicate with clients
<neilt> that does not necessarily mean remove corba
<neilt> if that makes sense
<neilt> i thought soap was an thing like DOM except XML
<jamest> i thought soap was like RPC
<chillywilly> soap is simple object access protocol
<jamest> but I have not followed it at all
<mdean> SOAP is like XML-RPC++
<jamest> lol
<chillywilly> well I think reinhard would agree that this si something to be done asap
<chillywilly> he wqanted me to do it with methods stuff
<chillywilly> I was like
<chillywilly> ummmm
<chillywilly> that would take a lot more time
<jamest> i think not that I've started this mess
<jamest> I'm going to bed
<jamest> :)
Action: jamest is too tired to think about this stuff
<jamest> nite
<chillywilly> night
<mdean> nite
jamest (jamest@fh-dialup-202033.flinthills.com) left irc: [x]chat
Action: neilt will have to read soap docs before expressing an opinion
<neilt> later all
<mdean> night
<chillywilly> l8r
neilt (neilt@dialup-209.244.81.186.Dial1.Washington2.Level3.net) left irc: bb tomorrow
<chillywilly> so soap communication is all xml stuff mdean?
<mdean> yup
Action: chillywilly doesn't know soap....have lookd briefly at xml-rpc
<chillywilly> hmmmm
<mdean> it's XML-RPC for objects
<chillywilly> hrrrrm
<chillywilly> why is this better than CORBA?
<chillywilly> some ppl claim this anyway
<mdean> what?  that's like apples and oranges
<chillywilly> how is it so different I thought they both did the same things
<mdean> well, unless you throw in WSDL and UDDI and some other stuff
<mdean> but something tells me CORBA is faster
<chillywilly> ah
<chillywilly> WSDL? UDDI?
<chillywilly> what are those?
<mdean> FLAs ;-)
<mdean> Web Services Description Language
<chillywilly> ah
<mdean> and I can't recall UDDI off the top of my head
<chillywilly> I've heard of that then
<chillywilly> is it an xml DTD?
<mdean> WSDL lets you publish services in an XML format
<chillywilly> so basically soap is like xml extreme or something?
<mdean> it would kinda be like interface discovery
<chillywilly> I see
<mdean> SOAP is ugly
<mdean> ugly to read, anyway
<chillywilly> like all xml filess
<chillywilly> ;)
<mdean> depends on who makes them :)
<chillywilly> I was just looking at OIFML form the ODMG
<chillywilly> object interchage format ML
<mdean> ah, yes - I've heard of it, but know nothing of it
<chillywilly> another xm dtd
<chillywilly> xml
<mdean> so is it like IDL in XML or what?
<chillywilly> ODL actaully
<chillywilly> object definition langauge
<mdean> ah
<chillywilly> but yea
<chillywilly> practically the same thing
<mdean> I haven't touched CORBA as of yet, but I've done plenty o' COM
<chillywilly> well I haven't done diddly with M$ technologies
<mdean> everybody's so psyched about .Net at work too - except for me, of course
<chillywilly> they are?
<chillywilly> oh gawd
<chillywilly> they wanna rent their computer form M$?
<chillywilly> software
<mdean> yeah - most of them (99.9%) do not know about the licensing changes
<mdean> I always hear "oh really?" when I enlighten them
<mdean> of course, since I'm the black sheep in the dev team, it's taken with a grain of salt
<chillywilly> you're the black sheep?
<chillywilly> heh
<chillywilly> you're the open source/free sogftfware nut?
<chillywilly> the guy who codes for nothing in hias spare time
<chillywilly> ;)
<mdean> yep - "there's one in every company", or so I hear
<mdean> the M$ crap puts food on the table, but I'd rather be working in Linux
<mdean> I told a head hunter in town that Linux and equivalent salary would be what it would take for me to entertain a move to another company
<mdean> of course, KC is not exactly hoppin' for Linux dev people - and my Linux dev is pretty light, but I could pick it up
<mdean> welp, almost 11PM - and I turn into a pumpkin
<mdean> catch y'all l8r
mdean (mdean@arc10x26.kcnet.com) left irc: Client Exiting
Mr_You (rouzer@mankind.boredom.org) joined #gnuenterprise.
SachaS (sacha@chime-a-181.conceptual.net.au) left irc: Ping timeout for SachaS[chime-a-181.conceptual.net.au]
--- Tue Sep 11 2001