*** derek has joined #gnuenterprise
*** johannesV has joined #gnuenterprise
<johannesV> good morning
*** btami has joined #gnuenterprise
<btami> good morning
<johannesV> good morning
*** sjc has joined #gnuenterprise
*** reinhard has joined #gnuenterprise
<reinhard> good morning
<johannesV> good morning
<ajmitch> hi
*** sjc has quit IRC
*** johannesV_ has joined #gnuenterprise
*** johannesV has quit IRC
*** siesel has joined #gnuenterprise
<siesel> hi
<reinhard> hi siesel
<siesel> hi reinhard
<siesel> btw. i found some strange bug in case of using sqlite and appserver
<siesel> sqlite just allows one thread to access one object
<reinhard> you talking about sqlite3?
<reinhard> or sqlite2?
<siesel> sqlite3
<reinhard> ok
<reinhard> we are working on that
<reinhard> we = johannesV_ :)
<johannesV_> ?
<reinhard> erm, not sure though
<reinhard> let me precise
<reinhard> we are working on some strange behaviour with sqlite3
<reinhard> not sure if it is related to your problem
<siesel> currently i try to reproduce it .. wait a second ..
<siesel> hmm, murphy anti law is working.  If you want to reproduce a bug .. it will work fine :(
<reinhard> lol
*** kilo has joined #gnuenterprise
<kilo> good morning
<siesel> hi kilo
<siesel> reinhard, I got the error again. just by opening three form instances and changing the content of the same record
<reinhard> what does it say?
<siesel> http://www.gnuenterprise.org/~jan/Screenshot.png
<siesel> it occured when calling "close"
<reinhard> woot
<reinhard> you sure there isn't a session mismatch?
<reinhard> i.e. one session tries to close a connection that another session opened
<reinhard> ?
<siesel> how could this happpen?
<reinhard> by passing the wrong session id parameter
<reinhard> not sure if it really makes sense what I'm saying
<reinhard> and my understanding of the error message is limited starting with the fact that I thought appserver would fork new processes and not make threads
<reinhard> at least per default
<siesel> sorry i switched to threading
<reinhard> ah oh
<reinhard> what happens when you switch to forking?
<reinhard> oh
<reinhard> I think I got it
<reinhard> you close the HTTP connection in between?
<reinhard> so you get a new thread on the server side for the new HTTP connection
<reinhard> so when switching to forking, you should receive an error message that this session does not exist
<siesel> yes, that would be the reason
<reinhard> btw
<siesel> yep
<reinhard> if you want beautiful error messages, you can cut what you get from the backend at occurances of character 0x91
<reinhard> first part is the exception type
<reinhard> err
<reinhard> first part is the exception kind (system, user, application)
<reinhard> second part is the exception class
<reinhard> third part is the error message
<reinhard> fourth part is the complete traceback
<reinhard> off to lunch
<siesel> I already thought about making some better error page. Then I will remember it
<reinhard> siesel: the idea of exception kind is
<siesel> guten hunger
<johannesV_> please note: appserver really *needs* persistent http connections
<reinhard> that for "user" exceptions you would only show the error message
* johannesV_ off to lunch 
<reinhard> while for other kinds you could show a "detail >>" button to show the traceback
<reinhard> you can look at the gtk2 implementation of the error message box to get an idea
<reinhard> off now for real, too
<siesel> ok. thx.
* siesel is missing real german/austrian food ;)
<siesel> sushi is great, but ....
*** btami has quit IRC
*** lekma has joined #gnuenterprise
<lekma> hi all
<lekma> from an rpc client perspective what would be the cleanest way to logout?
<lekma> _call(sessobj, "close")
<lekma> ?
<lekma> and then closing the connection?
<lekma> or is closing the connection enough?
<lekma> i know reinhard explained it to me yesterday, but i'm still confused...
*** kilo has left #gnuenterprise
<reinhard> calling "close" on the session object is the clean way
<reinhard> although just closing the connection should also clean up everything server side
<reinhard> you can even keep your connection open and close the session and open a new one
<reinhard> without closing the connection in between
<reinhard> or have several sessions within one single HTTP connection
<reinhard> then just close one session and leave another one open
<reinhard> however when the HTTP connection is closed, all things are kinda "emergency cleaned up"
<lekma> reinhard: thx for the clarification
<reinhard> looking at the code, I'm not really sure how perfect this "emergency cleanup" is, so it would be preferable to explicitly close the connection.
<lekma> you mean explicitely close the session, don't you?
<reinhard> err
<reinhard> yes of course
<reinhard> off for a while, kids want to play with the computer ;-)
<reinhard> cu
*** reinhard has quit IRC
*** btami has joined #gnuenterprise
*** jamest_ has joined #gnuenterprise
*** mawimawi has joined #gnuenterprise
<mawimawi> hi, is there a minitutorial for gnue-designer somewhere? found nothing in wiki so far. I want to start with something like a "hello world" - application...
<mawimawi> I managed to install gnue, have a postgresql database running, can start gnue-designer but then I'm stuck :(
<dimas> mawimawi: "hello world" type application does not require designer, so do not be stuck
<mawimawi> not if you want a database with table "continents", table "countries" and would like to browse through the continents and see the countries in a sub form ;-)
<siesel> if I remeber right, there are some tips for gnue-designer in the forms tutorial
<mawimawi> that might be something bigger than "hello world", but it covers the world, right?
<siesel> johannes: still here?
<siesel> can a persistent HTML/1.1 connection pass a proxy?
<mawimawi> ah, found something when googling for gnue forms: "www.gnuenterprise.org/tools/forms/docs/Developers-Guide.pdf" - thanks! hope it gets me started :)
<siesel> yes, thats it
*** jcater has joined #gnuenterprise
*** reinhard has joined #gnuenterprise
<mawimawi> hm. now I created my form in designer (with two fields from my table), but when I select "run form", it gives me:
<mawimawi> DB000: XML markup error in '[unknown]' at line 1:
<mawimawi> DB000: <unknown>:1:1: not well-formed (invalid token)
<mawimawi> in the background shell window. any ideas?
<mawimawi> ha! saving the form and opening it with "$ gnue-forms myform" worked!
<mawimawi> now I can navigate through my continents :)
<siesel> so "Hello world !" :)
<mawimawi> yessss!!!! success!
<mawimawi> but it's strange that I can't do that inside designer.
<mawimawi> I'm using the svn-version of gnue, all modules installed system-wide (with sudo python setup.py install) . might that be the reason?
<siesel> could be a designer bug, ...
<siesel> it shouldn't be connected to your installation I think
<siesel> but concerning designer you probably better ask jcater or jamest
<siesel> lekma, you around?
<siesel> what are your next steps?
<siesel> I got the rpc part of the webfrontend updated and want to include it into xul now.
<mawimawi> which file extension does gnue-designer want?
<mawimawi> I cannot start "$ gnue-designer myform" - because it tells me "Init'ing \n I do not know what to do with a . file!"
<reinhard> usually .gfd
<mawimawi> thanks. that works.
<mawimawi> "File-
<mawimawi> Open" seems to be broken in gnue-designer's cvs version.
*** klasstek has joined #gnuenterprise
<siesel> first XUL version of the webfrontend is working: http://www.gnuenterprise.org/~jan/XUL%20forms%20client.png
<reinhard> looks very nice
<SachaAway> siesel doing xul :)
*** SachaAway is now known as SachaS
<SachaS> nice
<siesel> reinhard: thx
<siesel> hi SachaS
<SachaS> hi siesel
<siesel> reinhard: now I just have to reactivate the parser and create the xul out of a gfd
<siesel> and do some clean up
*** derek has quit IRC
<siesel> Probably I can publish an pre-alpha XPI tomorrow
<siesel> SachaS, you are still doing ebXML?
<SachaS> siesel yes ebXML.
<SachaS> have you come across ebXML in Japan?
<siesel> no,
<SachaS> ok.
<siesel> I just care for organisational structure and process organisation now :(
<siesel> but I want to switch to IT stuff, and I'm not shure what thing I should start to learn to get a better job.
<SachaS> and you were thinking of ebXML?
<siesel> yes, as its an open standart, so I can get information easily
<SachaS> nice :)
<siesel> in which sectors is it already in use?
<siesel> automotive, finance, ...., SME?
<SachaS> high tech (telecommunications) in europe
*** derek has joined #gnuenterprise
<SachaS> automotive in us and now in europe
<SachaS> apparently people seem to  do ebXML in asia (japan south korea, hong kong?)
<siesel> I think you pointed me to some ebXML client in Hongkong some years ago.
<siesel> client like client server I mean.
<siesel> I think I have to investigate the topic a bit deeper.
<siesel> ok, have to go to sleep. good night all
*** siesel has quit IRC
*** btami has quit IRC
*** mawimawi has quit IRC
*** derek has quit IRC
*** derek has joined #gnuenterprise
*** sjc has joined #gnuenterprise
<SachaS> siesel: a good source of open source ebxml implementations is: http://www.freebxml.org
*** johannesV__ has joined #gnuenterprise
*** johannesV_ has quit IRC
<jamest_> after a resultSet.post() it should contain values that were alterd by the backend shouldn't it?
<reinhard> nope
<SachaS> siesel: if you need an ebXML contact in Japan, I am sure I can get you one.
<jamest_> reinhard: i swore there was support for refreshing now
<SachaS> jamest_: no swearing here
<SachaS> ;)
<reinhard> http://www.gnuenterprise.org/tools/common/docs/api/public/gnue.common.datasources.drivers.Base.ResultSet.ResultSet-class.php#post
<reinhard> read the docs ;-)
<reinhard> you must call requery subsequently
*** johannesV__ has quit IRC
<jamest_> or set requery to true on the __init__?
<reinhard> no
<reinhard> the point is that you would call the commit() of the connection between the post() and the requery()
<jamest_> then what is the requery param for on the __init__?
<jamest_> requery - If this is set to True, the ResultSet reqeries its values from the backend after posting, in case a backend trigger has changed something.
<reinhard> it basically decides whether the requery() does something or not
<jamest_> ah, ok
<reinhard> it is usually initialized from the datasource
<jamest_> so the description is a bit off on the docs on the init then yes?
<reinhard> so you always have the same code calling both post() and requery()
<reinhard> but if you do <datasource requery="False"> then requery() simply does nothing
<reinhard> brb
<reinhard> back
<reinhard> I agree that the description is not good
<reinhard> trying to improve it now
<reinhard> improved it (I hope)
* derek notes having docs for a free software project is kind of taboo
<derek> having current and meaningful docs may very well be illegal in some countries and/or states
<jamest_> reinhard: still sounds like it automatically does it sorat
<jamest_> sorta
<reinhard> well it *does* it automatically
<reinhard> not calling requery() is a mistake either way, no matter whether you want the backend to be queried again or not
<jamest_> oh i'm setting here playing with the wording and so far anything I write I dislike more than what's there
<reinhard> (that's what I wanted to express with the changed comments actually)
<jamest_> :)
<reinhard> I'm open to suggestions of course :)
<jamest_> yeah, i was trying, but i think it'll do
<reinhard> the method name "requery" is misleading to start with
<reinhard> but it started with really only doing this
<reinhard> but ended up doing other things as well
<reinhard> post() and requery() should actually be post_phase_1() and post_phase_2()
<jamest_> ah
<jamest_> ok
*** lekma has quit IRC
<reinhard> cu all
*** reinhard has quit IRC
*** jcater has left #gnuenterprise
*** jcater has joined #gnuenterprise
*** yure has joined #gnuenterprise
*** kilo has joined #gnuenterprise
*** jcater has left #gnuenterprise
*** jcater has joined #gnuenterprise
*** jamest_ has quit IRC
*** ajmitch has quit IRC
*** ajmitch has joined #gnuenterprise
*** sjc has quit IRC
*** klasstek has quit IRC
*** derek has quit IRC
*** derek has joined #gnuenterprise
*** kilo has left #gnuenterprise
*** derek has quit IRC
*** jamest has quit IRC
*** derek has joined #gnuenterprise