*** reinhard has joined #gnuenterprise
*** ferg has joined #gnuenterprise
*** johannesV has joined #gnuenterprise
<johannesV> good morning
<dimas> good morning
*** btami has joined #gnuenterprise
<btami> good morning
*** kilo has joined #gnuenterprise
<reinhard> good morning all
<kilo> good morning
<btami> reinhard: our yesterday problem solved
<btami> the datasource attrib was missing from block :)
<btami> (cut and pasted from another gfd)
<reinhard> oh
<reinhard> that can't really work
<btami> yes :)
*** ferg has left #gnuenterprise
<docelic> Morning all
<kilo> reinhard: is there any technique to do includes in a gcd file?
<kilo> our gcd files grow rapidly
<kilo> and they will become hard to maintain soon
<kilo> lots of procedures, long code parts
<kilo> and if they could be included from an external file, it would help maintenance very much
* btami dreams on a preprocessing step in readgcd
<reinhard> #include "foo.gcd"
<btami> s/dreams/dreaming
<reinhard> would be cool....
<btami> yep
<kilo> good old C...
<reinhard> there is some xml magic to do includes, isn't it?
<reinhard> or do I mix up things here?
<kilo> there is something
<reinhard> maybe you can even just have different gcd files for the same module
<kilo> but not sure how
<reinhard> i.e. have the same <module> tag in several files
<reinhard> but then different <class> tags
<kilo> can we???
<reinhard> I didn't try
<reinhard> but it could work
<kilo> hmmm
<johannesV> maybe you mean <class name="foo" library="foo.gcd"> ... :)
<reinhard> library???
<johannesV> i've tried something like that quite long ago and it didn't work very well, but i gave it up without putting more efforts on it
<reinhard> ah ok
<johannesV> the same as for gfd-files with dialogs and forms
<reinhard> I misunderstood that it is already there
<johannesV> ah, or was it <import-class name="foo" library="foo.gcd"/>
<johannesV> that's that importable-tags thingy
<btami> reinhard: arn't filter classes filtered per definition? i mean if i have year and company filters, they are not filtered with selected login data ?
<btami> f.e. if i want a procedure of year class returning workdays of a year, it have to have a year parameter ?
<reinhard> IIRC the filter class itself is not filtered
<btami> ok, thx
<reinhard> i.e. if you filter by year and you open a form with the years, you see all years
<btami> yes
* btami likes filters
*** johannesV has quit IRC
*** johannesV has joined #gnuenterprise
<johannesV> reinhard, i've discoverd the concept of postgres' atttypmod field for numeric values
<johannesV> length = (attypmod - 4) >> 16
<johannesV> scale = (attypmod - 4) & 0xFFFF
<reinhard> aha
<reinhard> quite close to what i suspected actually :)
<johannesV> thos nasty 4 bytes to subract were a hard thing to find out :)
<johannesV> i've forgotton about that so i've got quite different results
*** holycow has quit IRC
<btami> bbl
*** btami has quit IRC
*** yure has joined #gnuenterprise
*** kilo has quit IRC
*** jamest has joined #gnuenterprise
*** btami has joined #gnuenterprise
*** psu has joined #gnuenterprise
*** yure has quit IRC
*** jcater has joined #gnuenterprise
<reinhard> good morning jcater
<reinhard> you have a minute?
<jcater> yeah
<jcater> (but there may be lag time :)
<reinhard> IIRC you were involved in the creation of gnue-schema, right?
<reinhard> and certainly in introspection functions of datasources
<reinhard> I talked with johannesV the other day about the representation of a db schema
<reinhard> gnue-schema uses an object tree
<reinhard> introspection uses a Schema object with childSchema objects
<reinhard> and johannesV's schema creation functions use a dictionary structure to describe the db schema
<jcater> okay
<reinhard> and we (johannesV and me) somewhat agreed that this is suboptimal
<reinhard> and it might make sense to unify this
<reinhard> and the gnue-schema object tree seemed the most powerful approach of representing a db schema in python
<jcater> yes
<jcater> Schema objects came first
<jcater> but weren't meant (at the time) to be storable GParser objects
<jcater> gnue-schema came after that, and needed to be storable
<reinhard> so johannesV is working on changing both schema creation and schema introspection to work with gnue-schema GParser objects
<jcater> I'd have no problem getting rid of Schema objects
<reinhard> ok
<reinhard> that's what I wanted to hear
<reinhard> :)
<jcater> and moving to the GParser object
<reinhard> I think it could make it easier to do a schema designer module at some point in gnue-designer
<reinhard> instead of merely a schema viewer
<jcater> yeah
<reinhard> ok
<reinhard> actually johannesV already has started work on that
<reinhard> just wanted to make sure you don't see any problem in that path
<jcater> well
<jcater> the only potential roadblock I see
<jcater> is the introspection stuff would give more implementation-specific details
<jcater> but that's probably not too big a deal
<reinhard> can you explain what you mean?
<reinhard> not sure if I understand
<jcater> gnue-schema is lowest-common denominator
<jcater> but introspection also returned the equivalent information
<jcater> but also included, for example, postgresql-specific data types too
<jcater> in case your tool knew about those
<jcater> but that may not be needed
<jcater> I just think it does it now
<reinhard> yes it returns "nativetype"
<reinhard> we might add that to the schema xml definition
<johannesV> that nativetype won't be needed, does it ?
<reinhard> as a field that is (probably) ignored on read but is output on write
<johannesV> since the type="" length="" precision="" combination of a field-def get's translated to a proper nativetype automatically again
<reinhard> johannesV: somebody *might* want to know what native type it was in the db, just as info
<johannesV> but, yes, the GSField object still has a 'nativetype' attribute
<reinhard> e.g. was it a CHAR or a VARCHAR (for db's that do have different types for that)
<johannesV> (which won't get saved on a dumpXML () though)
<reinhard> johannesV: that seems to make sense
<reinhard> so the info is there as pure information
<johannesV> right
* johannesV adding constraint introspection right now ...
<reinhard> jcater: thanks for your comments
<psu> hi all
<reinhard> hey psu
<jcater> no problem
<reinhard> long time no see
<psu> I see we've got a namecheck on slashdot today
<psu> hi reinhard
<psu> http://it.slashdot.org/article.pl?sid=05/05/18/1817227
<reinhard> nice
*** mnemoc has joined #gnuenterprise
<PhUrl> so what is Steuerpflichtige innergemeinschaftliche Erwerbe
<reinhard> das ist, wenn du eine UID-Nummer hast und im EU-Ausland was kaufst
<reinhard> dann kriegst du eine Rechnung ohne Mehrwertsteuer
<reinhard> erm sorry for switching to German
<PhUrl> reinhard: danke
<PhUrl> so it is irrellevant for most things i do?
<PhUrl> if i bill you for software, it is aussergemeinschaftlich
<docelic> nur wenn do im EU-Ausland was kaufst ;-)
<PhUrl> ok
<docelic> du*
<PhUrl> Danke
<PhUrl> does anyone work with elster?
<reinhard> PhUrl: luckily I'm Austrian ;-)
<docelic> I learned German in school, 9 years
<PhUrl> docelic: not bad
<PhUrl> reinhard: i forgor
<PhUrl> we have met before, in frankfurt
<reinhard> oh
<reinhard> mike
<reinhard> long time no see :-)
<docelic> PhUrl: oscommerce ;-)
<PhUrl> cool
<mnemoc> how operative is gnue-report?
<PhUrl> yeah, i am working on sqlledger/lxerp at the moment
<docelic> PhUrl: did you investigate Interchange at any point ? (www.icdevgroup.org)
<PhUrl> docelic: yes
<PhUrl> i met with this very pleasant person, danb who declared me crazy
<PhUrl> i looked into using it
<PhUrl> currently i am working on a project for someone that uses oscommerce and sqlledger
<docelic> aha
<docelic> (btw, /me <-- Interchange developer)
<PhUrl> cool
<PhUrl> well, i looked into interchange
<PhUrl> and found that it was very complex from the coding
<PhUrl> i tried to hack some of the code
<PhUrl> sqlledger is also similar
<PhUrl> it does not use strict or warnings
<PhUrl> full of hacky perl
<PhUrl> and has many dangerous segments of code
<docelic> Well, depends how you look at Interchange. Currently I think the main problem is documentation (either lack of it, or badly written), which I am therefore redesigning since last year
<PhUrl> like evaling strings out of a web page as function pointers
<docelic> (btw, we switched to use warnings in all code just last week)
<PhUrl> does interchange use strict and warnings?
<PhUrl> oh
<PhUrl> wow
<PhUrl> good
<docelic> and use strict was always there, of course
<PhUrl> well, not in lxerp/sqlledger
<PhUrl> it is really badly coded
<PhUrl> full of copy and paste
<PhUrl> and hardly any structure
<PhUrl> even the structure it has is wrong
<PhUrl> the html is mixed up with the database
<docelic> (btw do you folks mind the conversation going slightly off-topic?)
<PhUrl> anyway
<PhUrl> that is what i am working on
<PhUrl> i hope to be able to determine of gnue is able to be the business server for the sqlledger
<PhUrl> will it be possible to port the perl code to gnue?
<psu> docelic : I think "offtopic" is not really a concept we worry about much around here ;-)
<PhUrl> it needs to have a seperation of presentation from database access
<psu> so keep talking
<docelic> PhUrl: yes, that's crucial
<PhUrl> has anyone used perl to talk to gnue?
<PhUrl> i would be willing to strip out the business logic from lxerp and attach it to gnue
<PhUrl> if it was possible to keep some of the code in perl
<psu> PhUrl: I guess in GNUe terms you are talking about "triggers"
<psu> i.e. things that you program in GNUe Forms, or Reports or Appserver
<PhUrl> ok
<psu> that actually contain the "business logic"?
<PhUrl> i am thinking about the appserver
<PhUrl> like, how to calculate the tax
<PhUrl> or what the workflow of an order is
<PhUrl> or how to transform an order to a bill
<PhUrl> all the stuff that is business logic
<PhUrl> if i can port that to use gnue as a basis
<PhUrl> and then transform the data model over
<PhUrl> it would be great
<PhUrl> there are alot of people wanting to use lxerp
<psu> I know we talked, before Appserver really got going
<PhUrl> now it is going good?
<psu> of providing "hooks" so you could write triggers for Forms or Reports in
<PhUrl> ok
<psu> something other than python (C was the main suggestion)
<psu> Don't know if this also applies to Appserver
<PhUrl> i guess it would be possible to rewrite the perl in python
<psu> but presumably the same logic is there
<PhUrl> i need to see
<PhUrl> at least :
<PhUrl> if i use the gnue for handling a bunch of stuff
<PhUrl> like the data dictionary
<PhUrl> and forms
<psu> if your perl is typically kludgy, a re-write in python might be better
<PhUrl> yeah
<psu> as python tends to force cleaner code
<PhUrl> ok, i need to take another look
<psu> but as I say, you might be able to plug in the perl anyway
<PhUrl> it might be easier to just look what gnue has and add the misisng things from lxerp
<PhUrl> and gnue supports a web front end and a form
*** johnix has joined #gnuenterprise
<johnix> Hello
*** johannesV has quit IRC
<johnix> ping?
*** johannesV has joined #gnuenterprise
<reinhard> johnix: pong
<johnix> hello Reinhard
<reinhard> hi joh
<reinhard> erm
<reinhard> johnix
<johnix> did I meet you somewhere before?
<reinhard> I don't think so
<reinhard> but it might be
<reinhard> at least I can't remember :)
<johnix> Maybe you can help me. I'd like to use Gnue to write a small timetracking app for my company
<johnix> I've used Gnue-designer on Debian, but it segfaults often
<reinhard> yes
<johnix> is it supposed to be useable, or should I seek another tool?
<reinhard> gnue-designer has problems with some versions of gtk and wxpython
<reinhard> gnue *is* usable
<reinhard> most people use designer just to create a basic template of their forms and then manually tweak the gfd file with an xml or text editor
<johnix> ok, so Emacs it will be.
<johnix> I've had trouble setting things up and getting the big picture
<reinhard> yes
<reinhard> we suck hard at documentation :(
<reinhard> but we try to be friendly to people who come to IRC and ask things :)
<johnix> There are files, databases, app servers, clients...
<reinhard> the problem might be
<johnix> ok... then that's how it will be...
<reinhard> well one of the problems at least
<reinhard> that you can do so many different things with gnue
<reinhard> you can use it 2-tier or 3-tier
<reinhard> you can write a full blown app or just create a form or report on top of an existing database
<johnix> ok, I want to use it 3-tier
<reinhard> etc etc
<reinhard> cool
<johnix> so I should use the app server
<reinhard> you have seen http://www.gnuenterprise.org/tools/appserver/docs/manual/devguide/index.html ?
<johnix> so I guess I have to describe both the database and the app server in  /etc/gnue/connections.conf
<johnix> no
<reinhard> johnix: about connections.conf
<reinhard> you have to have an entry for the database on the machine where gnue-appserver runs
<reinhard> and an entry for the appserver on the machine where your gnue-forms, gnue-reports etc run
<johnix> but if I develop the application on the same machine?
<reinhard> which is not necessarily the same machine
<reinhard> but it *can* be of course :)
<johnix> should I forget about the appserver during development?
<reinhard> no
<reinhard> au contraire
<johnix> does it make a difference?
<reinhard> appserver will be the core of your development
<johnix> ok
<johnix> and will I still use files to store interface definitions and suchM
<johnix> ?
<reinhard> think of appserver (better: gcd files) defining the *logic* in your application
<reinhard> and the forms and reports (gfd, grd) merely the "look"
<reinhard> and yes, you will use files
<reinhard> look at the doc link I gave you above, it should give a basic idea on how you develop with appserver
<johnix> ok. I'm doing it right now
<johnix> Now, I have another problem
<johnix> I can't start the app server
*** jamest has left #gnuenterprise
<johnix> DB000:  ProgrammingError: (1146, "Table 'GNUE.gnue_module' doesn't exist")
<johnix> this is as far as I've gotten
<reinhard> ok
<reinhard> you installed debian packages?
<johnix> yes
<johnix> (should I say, "of course"?)
<reinhard>  /usr/share/doc/gnue-appserver/configuration.txt.gz
<reinhard> should be your friend
<reinhard> bbl in 30 min
<johnix> ok, thanks
<johnix> I'm not staying either
<johnix> LUG meeting in 35 minutes, and I still have to eat
<johnix> but thanks
<docelic> LUGs are so 90's ;-))
<reinhard> back
*** sjc has joined #gnuenterprise
*** md-afk has joined #gnuenterprise
*** PhUrl has quit IRC
*** johannesV has quit IRC
*** btami has quit IRC
*** psu has quit IRC
*** kilo_ has joined #gnuenterprise
*** jamest has joined #gnuenterprise
*** yure has joined #gnuenterprise
*** jamest has left #gnuenterprise
<reinhard> time to go to bed
<reinhard> night all
*** reinhard has quit IRC
*** dimas has quit IRC
*** yure has quit IRC
*** kilo_ has quit IRC
*** sjc has quit IRC
*** jcater has quit IRC
*** wendall911 has joined #gnuenterprise
*** jamest has joined #gnuenterprise
*** SachaS__ has quit IRC
*** ramiroluz has joined #gnuenterprise
<ramiroluz> hi all, its my first time on the chanel. Someone can help me?
<jamest> hi, whats up?
<ramiroluz> I need to call a database trigger on Postgres in a form. I didn't found it in the docs.
<jamest> i'm rusty but last I knew we didn't have a sql passthru for that
<jamest> at least one exposed to the trigger namespace
<jamest> let me check
<jamest> ok
<jamest> completely untested/unverified/etc,etc
<ramiroluz> :)))
<jamest> but it looks like a datasource object has a call method
<jamest>   def call (self, methodname, parameters):
<jamest>     """
<jamest>     Call a function of the backend.
<jamest>     It is highly recommended to post this record, it's chain of master records
<jamest>     and all of it's details before calling this function, so the backend is up
<jamest>     to date when executing the called backend method.
<jamest>     @param methodname: Name of the function to call.
<jamest>     @param parameters: Dictionary with parametername/value pairs.
<jamest>     @return: Return value of the function that was called.
<jamest> so I think you can do in a trigger
*** dcmwai has joined #gnuenterprise
<jamest> dtsName.call()
<jamest> but  i don't use DB level triggers so I can't be certain if/how this works
<ramiroluz> Ok, thank you. I will try it. Its easier as I was thinking. I hope this work. :))
<jamest> that makes 2 of us :)
<ramiroluz> good bye.
<jamest> bye
<jamest> if that doesn't work
*** ramiroluz has left #gnuenterprise
<jamest> nevermind :)
*** wendall911 has quit IRC
*** holycow has joined #gnuenterprise
*** jcater has joined #gnuenterprise
*** holycow has quit IRC
*** jamest has quit IRC
*** dimas has joined #gnuenterprise
*** jcater has quit IRC
*** Galleon has joined #gnuenterprise
*** Galleon has left #gnuenterprise
*** jcater has joined #gnuenterprise