*** lekma_ has joined #gnuenterprise
*** lekma has quit IRC
*** dimas has joined #gnuenterprise
*** reinhard has joined #gnuenterprise
*** holycow has joined #gnuenterprise
*** sjc has joined #gnuenterprise
*** SachaS has quit IRC
*** btami has joined #gnuenterprise
<btami> good morning
*** johannesV has joined #gnuenterprise
<lekma> morning everyone
<johannesV> good morning
<reinhard> good morning Europe!
<dimas> good morning
<johannesV> reinhard, i've just tried that unicode-thingy you've mailed me .... :)
<johannesV> RUN: <class __main__.TestPsycopg at 0x404312fc>
<johannesV> returns Unicode(fetchmany): NO
<johannesV> handles Unicode Statements: NO
<johannesV> /home/johannes/bin/gcvs: line 9:  7620 Speicherzugriffsfehler  /usr/bin/python2.3 "$@"
<johannesV> that's a segfault :)
<reinhard> wow
<reinhard> for unicode parameters?
<johannesV> yeah
*** kilo has joined #gnuenterprise
<btami> johannesV/reinhard: in form's field attrib of field tag can contain only lowercase chars, unless it's not updated
<btami> is it normal?
<reinhard> no
<reinhard> that is a bug probably introduced by my ongoing changes
<reinhard> thanks for reporting
<btami> np
*** holycow has quit IRC
<btami> Traceback (most recent call last):
<btami>    File "<unknown_executioncontext>", line 12, in blkItem_fldQuantity_POST_CHANGE_onchange
<btami>    File "/home/tamas/svn/gnue/.cvsdevelbase/gnue/common/logic/NamespaceCore.py", line 317, in __call__
<btami>     return self._objectFunction(*args, **kwargs)
<btami>    File "/home/tamas/svn/gnue/.cvsdevelbase/gnue/forms/GFObjects/GFBlock.py", line 708, in callFunction
<btami>     parameters)
<btami>    File "/home/tamas/svn/gnue/.cvsdevelbase/gnue/common/datasources/GDataSource.py", line 208, in callFuncOfCurrentRecordsetEntry
<btami>     self._currentResultSet.current.call (name, params)
<btami>    File "/home/tamas/svn/gnue/.cvsdevelbase/gnue/common/datasources/drivers/Base/RecordSet.py", line 502, in call
<btami>     raise errors.ApplicationError, u_("Function call on empty record")
<btami>  ApplicationError: Function call on empty record
<btami> got it with gnue-invoice (in gnue-contrib) grid.gfd
<btami> when trying to create a new invoice
<reinhard> btami: and? is it true?
<reinhard> is the record you call a function of empty?
<btami> reinhard: i know nothing about it's internals
<btami> just testing things
<btami> it worked before
<reinhard> hmm
<reinhard> let me test something
*** dcmwai has joined #gnuenterprise
*** dimas has quit IRC
<reinhard> btami: ok, I can reproduce, will look into it
<reinhard> well actually
<reinhard> as gnue-invoice uses *exclusively* mixed-case field names
<reinhard> it might well be related to the other bug you reported
<reinhard> btw the goat is nice ;-)
<kilo> it is one of Jason's
*** sjc has quit IRC
*** btami has quit IRC
*** StarDeep has joined #gnuenterprise
*** mnemoc_ has joined #gnuenterprise
<reinhard> bbl
*** reinhard has quit IRC
*** mnemoc has quit IRC
*** jamest has joined #gnuenterprise
*** jamest has quit IRC
*** tiredbones has joined #gnuenterprise
*** dimas has joined #gnuenterprise
*** kilo has quit IRC
*** jamest has joined #gnuenterprise
*** johannesV_ has joined #gnuenterprise
*** johannesV has quit IRC
*** titopbs has joined #gnuenterprise
*** johannesV_ has quit IRC
*** johannesV has joined #gnuenterprise
<dimas> Added a fish-hook element to address_person
<dimas> johannesV: what does a fish-hook mean in here?
<johannesV> that's a column of a table referencing itself
<johannesV> address_person.address_senior is a reference to address_person
<dimas> how and when is it usefull?
<dimas> is it used in contrib projects?
<dimas> oh, in that sample
<johannesV> hmm, i don't know ... but that doesn't matter ...
<johannesV> i think lekma had such classes ...
<dimas> ok
<johannesV> the problem was the order in which appserver did the processing of such related records
<lekma> to explain what was our case
<lekma> some cleaver consultant (namely yours truly)
<lekma> decided to put offers and orders in the same class
<dimas> :)
<lekma> and transform offers to orders by copying data and referencing the offer
* lekma thinks he must be the dumbest designer in the worlds
<johannesV> :)
<dimas> world_S_ :)
<lekma> :)
*** titopbs has quit IRC
<lekma> one thing that i really learned the hard way with this project is:
<lekma> do one thing at a time
<lekma> and don't rely on expected behaviour
<lekma> TEST
<lekma> :)
<johannesV> that's true ...
<dimas> you guys are just great!
<johannesV> but we've startet to integrate unittests quite a lot now
<johannesV> this should make things a little more stable ...
* jamest looks out the window at all the kids gnue'n in the yard
<jamest> johannesV: svn settling down any?
<johannesV> jamest, what do you mean ?
<jamest> i run svn in production here
<jamest> and try to follow the curve just a bit
<jamest> w/ all the changes in datasources recently i wasn't sure how svn head is, vs a week or so ago
<johannesV> oh, wow ... :)
<johannesV> hmm, i don't know wether reinhard's changes are *all* bug-free atm, but at least appserver's unittest runs fine
<jamest> cool, most my apps are custom against gnue-common
<johannesV> (but i've found a little bug in requery () though ... it fails if the requeried record does not exist ...)
<jamest> and so far the only thing that bit me was loss of implied fields if none specified
<jamest> and requery is what I'm wanting to gaim
<jamest> gain
<johannesV> maybe we should consider a unittest for that db-driver tree
<jamest> i planned to write some
<johannesV> that'd be really great
<jamest> but work on the house is taking more time that expected
<jamest> so i imagine it'll be may
<johannesV> hm, yeah, better in may then never .... :)
<jamest> lol
<jamest> hmmm
<jamest> i thought we allowed for sql passthru on datasources
<jamest> so I could manually do an insert / update / or whatever
* jamest needs to manipulate a postgresql sequence from a python script
<johannesV> i think that was eliminated by reinhard yesterday ... but I'm not sure about that
<jamest> lol, timing is everything I guess :0
<jamest> er, :)
<tiredbones> On page 13 in the Common Developer's Guide there is a reference to a registration.gsd file, does is matter where I put this file?
*** kilo_ has joined #gnuenterprise
<jamest> tiredbones: no
<jamest> you specify that file as a command line argument to gnue-schema
<jamest> so it should work like a form
<jamest> in that it can be anywhere
<tiredbones> Thanks jamest.
*** StarDeep has quit IRC
<tiredbones> jamest: would it be good practice to put all gsd file relating to giving package in a separate directory?
<jamest> i don't know
<jamest> if using appserver i believe it supports a directory where you put these
<jamest> wait, i'm thinking gcd's
<kilo_> gcd files yes
<jamest> what i do here is still based on .sql
<jamest> but i put all table definition scripts in a stucture like
<jamest> dbname/tables/tablename.sql
<tiredbones> Thanks jamest and kilo.
<tiredbones> I keyed in the registration.gsd and then ran the command "gnue-schema --vendor=postrgesql registration.gsd"Having gotten an error, I looked at gnue-schema --help. Should --vendor parameter be replace with --connection?
<jamest> looks like it
<johannesV> tiredbones, gnue-schema takes a connection name, which you will have defined in your connections.conf
*** SachaS has joined #gnuenterprise
<lekma> jamest: are you fluent in gnue-reports?
<lekma> :)
<SachaS> johannesV what is a fish hook?
<dimas> :)
<johannesV> :)
<lekma> :)
<johannesV> a table is called to have a fish-hook if it has a column referencing it's own table
<SachaS> a reference to the same class ?
<SachaS> ok.
<johannesV> if you draw an erd it looks like a fish-hook ...
<johannesV> right, a ref to the same class
<johannesV> (if speaking in appserver terms)
<SachaS> ok. thanks.
<johannesV> np
<jamest> lekma: not really, every time i create a new report i have to relearn it
<lekma> :(
<johannesV> bbl
*** reinhard has joined #gnuenterprise
<reinhard> jamest: re passthrough SQL
<reinhard> DBSIG2.Connection still supports an "sql (query, parameters = None)" method
<reinhard> returning a 2-dimensional array for the result
<reinhard> sql1 for something that is expected to only return a single value
<reinhard> and sql0 for statements that are expected to return nothing
<reinhard> but if you really just want to do arbitary inserts, deletes, and updated
<reinhard> updates
<reinhard> you might want to look at the (new) Base.Connection.insert/update/delete methods
<reinhard> that don't require you to build SQL but accept fieldname/value dictionaries
<reinhard> plus they are portable to non-SQL backends in case that ever matters
<jamest> reinhard: i'd have to call postgresql specific functions to manipulate sequences
<jamest> as they aren't really fields
<reinhard> so you would stick to sql()
<reinhard> gotta run again
<reinhard> btw jamest: current SVN seems to have a problem with uppercase letters in field names
<reinhard> will fix tomorrow probably
<reinhard> l8r all
*** reinhard has quit IRC
<tiredbones> jonannesV; Well, if I look at what was setup in the connectios.conf vs. that of the command that was entered, there is no clue as to what that connection name should be.Here is the connection.conf that was entered:
<tiredbones> [pros]
<tiredbones> s/[pros]/[prod]
<tiredbones> comment = Production DB
<tiredbones> provider = psycopg
<tiredbones> host = rdbms.example.com
<tiredbones> dbname = prod
<tiredbones>  [web]
<tiredbones> comment = Web DB
<tiredbones> provider = psycopg
<tiredbones> host = www.example.com
<tiredbones> dbname = web
<tiredbones> the command entered;
<tiredbones> gnue-schema --vendor=postgresql registration.gsd
<tiredbones> --vendor probalbe should be --connection
<tiredbones> So, if it is --connection what part of the connections.conf should --connections point to?
<dimas> to [prod] or [web]
<dimas> just through away []
<tiredbones> dimas: Is this what I what, --connection=prod?
<tiredbones> s/I what/ I want/
<dimas> should be that
<tiredbones> Thanks
<tiredbones> When I run the gnue-schema command I get an error informing me that psycopg module not installed. However, I just verify that I did installed it using synaptic and it does show that it is installed.
<dimas> can you give an exact message of the error?
<tiredbones>  File "/usr/lib/gnue/python/gnue/common/apps/plugin.py", line 168, in find
<tiredbones> DB000:     raise LoadError, (name, result)
<tiredbones> DB000:  LoadError: Cannot load plugin psycopg
<tiredbones> DB000: The following plugins failed:
<tiredbones> DB000: * gnue.common.datasources.drivers.postgresql.psycopg: DependencyError: Module 'psycopg' is
<tiredbones> not installed
<tiredbones> I just ran a find command on psycopg.py, it comes up empty. hmm
<dimas> try to run  gnue-common/src/apps/plugin.py
<dimas> i do
<dimas> gcvs  plugin.py
<dimas> it gives more info as why plugin was not loaded
<tiredbones> what is gcvs command
*** jcater has joined #gnuenterprise
<dimas> it just adds gnue modules into pythonpath
<dimas> it is for cvs installation
<tiredbones> I use deb's. gnue-common a command?
<tiredbones> I use deb's. is gnue-common a command?
<dimas> no
<tiredbones> when I try man gnue-common, it comes up empty
<dimas> try
<tiredbones>  gnue-common
<dimas> python plugin.py
*** sjc has joined #gnuenterprise
*** Amorphous has quit IRC
*** tiredbones has quit IRC
<dimas> is there a "serial" fieldtype in postgresql?
<nickr> like something that gets incremented with each now record?
<dimas> yes
<nickr> There is something akin to this
<nickr> hmm, trying to remember the details
*** johannesV has quit IRC
<nickr> you cerate a sequence and set the default value of a column to the next value of the sequence, I think
<dimas> yes, i remember this
<nickr> create sequence serial_seq; create table serialized ( serial_number integer default nextval(serial_seq) );
<nickr> I think.
*** Amorphous has joined #gnuenterprise
<dimas> nickr: thanks, someone is trying to setup a billing apps for asterisk, will look into how it create the db myself
*** tiredbones has joined #gnuenterprise
<jamest> i think there is a serial type too
<dimas> nickr: there is serial in postgresql from 7.3 and up
<dimas> CREATE TABLE tablename (colname SERIAL);
<dimas> yep
<jamest> which automates the creation
<jamest> of the sequence
<jamest> but you lose control of naming the sequence
<jamest> and iirc the sequence isn't autodropped if you drop the table
<jamest> so manual gives you better control
<dimas> jamest: thanks
*** titopbs has joined #gnuenterprise
<lekma> night all
*** lekma has quit IRC
*** Amorphous has quit IRC
*** Amorphous has joined #gnuenterprise
*** jamest has left #gnuenterprise
<nickr> interesting
<nickr> I think I'll stick to manual sequences :)
*** reinhard has joined #gnuenterprise
*** jamest has joined #gnuenterprise
*** jamest has left #gnuenterprise
*** jcater has quit IRC
<tiredbones> When the connections.conf file  is read do all the db have to exist?
<tiredbones> What I mean is can I put a entry in the connections.conf without using it?
<tiredbones> I'm running this command and I getting and message telling me that db gnue does not exist.
<tiredbones> This is what I'm running
<tiredbones>  gnue-schema --connection=prod registration.gsd
<tiredbones> It appears that the prod db is ok.
<tiredbones> Can someone tell how to solve this?
<tiredbones>   File "/usr/lib/gnue/python/gnue/common/datasources/drivers/DBSIG2/Connection.py", line 299, in makecursor
<tiredbones> DB000:     cursor.execute (s)
<tiredbones> DB000:  ProgrammingError: ERROR:  syntax error at or near "-" at character 30
<tiredbones> DB000:
<tiredbones> DB000: CREATE SEQUENCE registration_-1216706252_seq;
<reinhard> erm
<reinhard> that doesn't look like a message that the db gnue doesn't exist
<tiredbones> I went ahead and created the gnue db so that I could get by the error message.
<tiredbones> Now I'm getting this error.
<reinhard> where do you have this gsd file from?
<tiredbones> Is this error telling me that my gsd file is wrong?
<reinhard> at first sight yes
<tiredbones> This gsd file is the one in "Common Developer's Introduction Guide" on page 13.
<reinhard> you planning to write common based non-gnue apps?
<tiredbones> I'm trying to do all the guides that has something to tech me at the moment.
<reinhard> ouch
<reinhard> seems like you found a bug in gnue-schema
<reinhard> ************************
<reinhard> johannesV
<reinhard> gnue.common.datasources.drivers.Base.Schema.Creation.Creation._getSequenceName
<reinhard> uses the id() of an object for generating a "unique" name
<reinhard> but the id can contain a - singn (it can be negative)
<reinhard> ********************
<reinhard> tiredbones: you can work around that bug by using shorter table and field names
<reinhard> so table name + field name < 27 characters
<tiredbones> ok, thanks
<reinhard> FWIW I am not sure how current or outdated this documentation is
<tiredbones> It looks like its outdated, but some of the stuff works and I'm getting some idea what's happening.
<tiredbones> This info is good to know in the overall scheme of things.
<reinhard> kilo_: oh my, I just saw your .dia files in gnue-common/doc
<reinhard> they cover *exactly* what I am messing around with for some weeks now
<kilo_> lol
<reinhard> so they will be heavily outdated for the next release
<reinhard> :(
<kilo_> np
<kilo_> i will redo them
<kilo_> i am happy you fight this monster
<reinhard> kilo_: I see some overlap between your .dia files and our epydoc documentation
<reinhard> hmm
<reinhard> common-techref.lyx
<kilo_> dia was earlier i think
<reinhard> This is version 0.0.1 of the Technical Reference it covers GNUe Common 0.3.0
<kilo_> hehee
<reinhard> building-static-distro.txt refers to wxgtk 2.3
<reinhard> and some other quite old versions :)
<kilo_> ok, good night now
<reinhard> night
*** kilo_ has quit IRC
<tiredbones> reinhard: I hope I'm not causing use extra work, doing the things that I do.
<reinhard> no problem
<tiredbones> reinhard: I shorten a values to be < 27 and it worked.
<reinhard> good night all
*** reinhard has quit IRC
*** sjc has quit IRC
*** tiredbones has quit IRC
*** dcmwai has quit IRC
*** jcater has joined #gnuenterprise
*** holycow has joined #gnuenterprise