*** mixi^ has joined #gnuenterprise
*** mixi has quit IRC
*** reinhard has joined #gnuenterprise
*** kilo has joined #gnuenterprise
*** johannesV has joined #gnuenterprise
<johannesV> good morning 
<kilo> good morning
<kilo> johannesV: just reading the commit email now.
<johannesV> :)
<johannesV> using firebird-s32-server on my debian is working
<johannesV> it's a bit slow because of some 'additional?' queries ... but it's working
<kilo> svnupping now, will check soon
<kilo> i hated to manually add properties to the database 8-))
<johannesV> :)
<kilo> and it works like charm
<johannesV> well, so i'm happy i could help :)
* kilo is busy filling up the appserver with tons of classes and like
*** derek has joined #gnuenterprise
*** btami has joined #gnuenterprise
*** dneighbo_ has quit IRC
<kilo> johannesV: errr....may i dare to report an error again?
<johannesV> btami, kilo: is there a way to see which statements are executed by a firebird db ?
<johannesV> kilo: yes
*** nickr has quit IRC
<kilo> i cant save a procedure. classes are added, properties are added, but saving a procedure it reports an error
<kilo> GFForm: Váratlan kivétel:
<kilo> ------------------------------------------------------------
<kilo> Traceback (most recent call last):
<kilo>   File "/home/gabor/SVN/gnue/gnue-common/.cvsdevelbase/gnue/forms/GFForm.py", line 544, in commit
<kilo>     block.processCommit()
<kilo>   File "/home/gabor/SVN/gnue/gnue-common/.cvsdevelbase/gnue/forms/GFObjects/GFBlock.py", line 513, in processCommit
<kilo>     self._dataSourceLink._dataObject.commit()
<kilo>   File "/home/gabor/SVN/gnue/gnue-common/.cvsdevelbase/gnue/common/datasources/drivers/Base/DataObject.py", line 151, in commit
<kilo>     self._connection.commit()
<kilo>   File "/home/gabor/SVN/gnue/gnue-common/.cvsdevelbase/gnue/common/datasources/drivers/appserver/appserver/Connection.py", line 100, in commit
<kilo>     self._sm.commit (self._sess_id)
<kilo>   File "/home/gabor/SVN/gnue/gnue-common/.cvsdevelbase/gnue/common/rpc/drivers/Base.py", line 232, in __call__
<kilo>     return self._adapter._runMethod (self._methodname, *args, **params)
<kilo>   File "/home/gabor/SVN/gnue/gnue-common/.cvsdevelbase/gnue/common/rpc/drivers/xmlrpc/pw_xmlrpc/ClientAdapter.py", line 96, in _runMethod
<kilo>     raise client.DistantError, e.faultString
<kilo> DistantError: Invalid value 'None' for property 'gnue_code'
<kilo> though the code editor field was filled with very intelligent code (ie print "blahblah")
*** nickr has joined #gnuenterprise
<johannesV> hmm, have to check ... one moment please
<kilo> ok
<johannesV> kilo: this is working fine for me ... 
<kilo> oh 8-((
<johannesV> i've added a new procedure using parameters and a result .. so
<johannesV> are you sure there was a code in the field ? :))
<kilo> yes, there was
<btami> johannesV: i know it is possible(track sql), koz delphi has a tool for this, but
<johannesV> kilo: can you reproduce the error ?
<btami> i don't know such a tool for linux
<johannesV> btami: ok, but there's no syslog-facility where such stuff is logged into ?
<btami> there is a log file, but i can't find any docs about this issue
<johannesV> ok, thanks
<kilo> yes, error is reproducable
<johannesV> hmm
<johannesV> kilo: can you explain the steps you're doin to me ?
<kilo> sure, in 1 min, i have to advice my friend who wants to buy a good whisky in Scotland
<johannesV> :)
<kilo> whisky advice is done
<kilo> so back to classre...gfd
<kilo> i have a module, in it two classes
<kilo> Classes tab, Query mode, i select my module and execute query
<kilo> there are my two clesses, filled with properties
<johannesV> ok
<kilo> selecting one of them i change to the Procedure tab
<kilo> i see the class name at the top
<johannesV> ok, first i select a class and then change to proc-tab
<kilo> fully qualified name, ie module_classname
<johannesV> ok
<kilo> i write something in each line, but nullable, and also choose the module from the dropdown
<johannesV> do you choose the module by typing or via mouse
<kilo> via mouse
<johannesV> ok
<kilo> no parameters given, i enter code like
<kilo> print "blah"
<johannesV> ok
<kilo> and now i push the save button
<johannesV> ok, so did i
<johannesV> i didn't get an error
<kilo> and nor do i
<johannesV> :)
<kilo> wow, johannesV cured it magically via the internet!!!
<johannesV> yeah !
<kilo> wait, i must have done something
<kilo> maybe only talking about whiskies helped?
* johannesV lookin for some coffee
<kilo> well at least i can produce an error, if i update the procedure code and save
<johannesV> the same procedure you've inserted before ?
<johannesV> same error ?
<kilo> the same proc, yes. i put a second line in the code and tried to save
<johannesV> DistantError: (-902, 'execute.isc_dsql_execute: internal error. ')
<kilo> exactly
<johannesV> ok, so i'll try to catch this one :)
<johannesV> this is strange, cause updates worked well on sample.gfd
<kilo> hehe, ok 8-))
<kilo> btw what does it mean that a procedure is nullable?
<johannesV> this means a procedure is allowed to return "None" as result
<johannesV> otherwise returning 'nothing' or None would raise an error 
<kilo> ok, i see
<johannesV> which should point out where procedures are 'not well defined/programmed'
<kilo> and should parameters be filled right at creation time? or is it required to save the proc first and then add params?
<johannesV> btw. sample.gfd still works well; i think this could be a problem with the blobs
<johannesV> you can add parameters in the same step as defing the procedure
<johannesV> no save-first, save-second needed
<kilo> ok, so it is unlike class-properties
<johannesV> well, actually this should work there too
<johannesV> hmm, at least i think so ..
<johannesV> the internal error is a problem raised by gnue_code's datatype (BLOB)
<johannesV> this was why I initially forced the interbase sql-processor to translate 'unlimited strings' into 'varchar (32767)' instead of blob
<johannesV> blob is always a bit tricky to handle
*** bluesbaron_ has joined #gnuenterprise
<kilo> i used blobs with an interbase db project to store compiled regexp objects in the db, and it worked quite well
*** bluesbaron_ has left #gnuenterprise
<btami> http://kinterbasdb.sourceforge.net/example_code/simple_blob_pickle_example.zip
<btami> shows the way how to use blobs
<kilo> should i send the code to you?
<johannesV> just downloaded that zip-file ...
<johannesV> kilo: yes please send the code 
<kilo> line 199
<kilo> and in the latter line 53
<johannesV> ok, i see
<johannesV> i think the problem is the update-statement
<johannesV> update set gnue_code = 'blahbla' where gnue_code = ? and gnue_id = ?
<johannesV> parameters: ['old text', <buffer-object>, '12313213112']
<johannesV> the first two params won't match in type and value
<johannesV> so if we would change the code so they match in type there's still the db which don't like this (have tried it)
<kilo> ah
<johannesV> this form of an update is diffrent from your samples in that way that the blob-value is used in the condition part
<kilo> yes, i dont use it in comparisons
<johannesV> seems as if there's no quick fix for this problem
<johannesV> there's the more general question how to treat blobs anyway which needs some 'think-over'
<kilo> ok, this problem can be overriden by deleting and adding the procedure again, se it can be lived with
<johannesV> hmm, yeah for the moment; but i don't think this is a 'solution' :)
<kilo> no, but a wuick hack.
<kilo> s/wuick/quick
<kilo> what are the plans about supporting lists and dictionaries?
<johannesV> you mean as datatypes ?
<kilo> yes
<johannesV> there are no real 'plans' to do so, but we have thought about lists (arrays)
<kilo> i see
<kilo> would they cause RPC problems?
<johannesV> not sure about; but using lists as 'parameters' or 'procedure-results' would be easier than making a 'list-type' available as persistent database-type
<johannesV> since rdbms's are quite different in handling such types (if they do so)
<kilo> and what if they were stored as blobs?
<johannesV> *lol*, this will imply we have a 'usable' solution for the 'blob-problem'
<johannesV> anyway, a list of the same type of information is against the codd's normalform
<johannesV> like name_1, name_2, name_3 
<johannesV> this should be a detail-table with one single 'name' field
<johannesV> having multiple entries
<johannesV> instead
<kilo> hmmm, true
<johannesV> anyway, appserver should support list-fields (list-properties) with it's version 0.2
<johannesV> after all we haven't discussed the implementation details
<johannesV> so, this might make the discussion about 'array'-types obsolete
<kilo> ok
*** reinhard_ has joined #gnuenterprise
*** reinhard has quit IRC
<reinhard> bbl
*** reinhard has quit IRC
<kilo> johannesV: we tried using a blob field in where clause
*** Vee has joined #gnuenterprise
<kilo> it works in isql, and in forms when useParameters != 1
<kilo> ie it is commented out in gnue.conf
<kilo> but it doesnt work using appserver, even if useParameters is commented out
<kilo> but it throws an other type of error, this time again boolean matters
<kilo> the sql statement is:
<kilo> UPDATE gnue_procedure SET gnue_code='try:
<kilo>     return self.__cache[kod][1]
<kilo> except:
<kilo>     return None
<kilo> #aaa',gnue_nullable=TRUE WHERE gnue_code='try:
<kilo>     return self.__cache[kod][1]
<kilo> except:
<kilo>     return None' AND gnue_id='60164400889888615052419736243989' AND gnue_nullable='1'
*** havoc has quit IRC
<kilo> gnue_nullable=TRUE, tha is problematic
*** havoc has joined #gnuenterprise
<btami> bbl
*** btami has quit IRC
*** kilo has quit IRC
*** dsmith has joined #gnuenterprise
*** jamest has joined #gnuenterprise
*** johannesV_ has joined #gnuenterprise
*** johannesV has quit IRC
*** mixi^_ has joined #gnuenterprise
*** mixi^ has quit IRC
*** dimas has quit IRC
*** dimas has joined #gnuenterprise
*** jcater has joined #gnuenterprise
<dneighbo> for the record if you ever do a fk lookup
<dneighbo> and you list in the field a max length that is not the same as the value in the lookup
<dneighbo> it will behave MOST oddly on "prepare/execute" queries
<dneighbo> ;)
<dneighbo> hmmm
<dneighbo> jcater: i think that designer keeps resizing my form w/o me telling it too
<dneighbo> i will see if i can verify a simple sample and send to you for verfication
<jcater> if you drag and drop fields onto it
<jcater> from schema navigator
<jcater> yes, it resizes to fit all of the new ones
<dneighbo> hmm i dont think that is the issue (though it could be)
<dneighbo> i think if you make a form smaller than the "toolbar"
<jcater> if it's doing it some other time
<jcater> I've never seen it
<jcater> and it's likely a bug
<dneighbo> when you open the form in designer it resizes it to be big enough for the toolbar
<dneighbo> i will triple check and see if its repeatable and file bug accordingly
<dneighbo> can you have a fk on existing datasource
<dneighbo> er let me rephrase that
<dneighbo> i ahve table foo with field bar and bar desc
<dneighbo> actually never mind
*** reinhard has joined #gnuenterprise
<jcater> cool
<jcater> I like neverminds
<jcater> if you were trying to ask what I think you were
<jcater> then the answer would've been
<jcater> it may let you, but I would be nervous doing so
<jcater> as such a setup may break in the future
<dneighbo> actually i think i changed my question and found the answer
<dneighbo> it works but not really (as it breaks in query mode)
<dneighbo> wtf
<dneighbo> how can a combo contain the value correctly on every record
<dneighbo> but the list is empty?
* dneighbo is not getting this one
<jcater> simple
<jcater> we do that on purpose
<jcater> if you have existing records w/values
<jcater> that are not in the results of the query for that dropdown
<jcater> we don't want the form to bomb
<jcater> as that record does exist
<jcater> if you change the value of that field, however, you wouldn't be able to select the old non-valid value again
<jcater> i.e., we can't enforce the dropdown's foreign key values on existing records
<jcater> if that makes sense
<dneighbo> no
<dneighbo> here is the deal
<dneighbo> i have a field called activity
<dneighbo> i do a lookup to get the description (and valid activities)
<dneighbo> so
<dneighbo> the value in the table is NPRG
<dneighbo> but i do a lookup on the ACTV table to get
<dneighbo> NPRG - NON PROGRAM
<dneighbo> that is what displays
<dneighbo> i go to the next record 9999
<dneighbo> and 9999 - PCARD is displayed
<dneighbo> but if i click on the combo box arrow
<dneighbo> there are no entries
<dneighbo> my point being where the hell is it getting the "description" from to display if nothing is in the box?
<dneighbo> obviously it is binding to the lookup table to get the value
<dneighbo> only thing i can think is i added this datasource in a different manner than the others that do work
<dneighbo> OR
<dneighbo> that this table is already being used as another datasource
<dneighbo> basically i have deptprog as a master table
<dneighbo> it has fields of dept and activity
<dneighbo> then it links to the actv (where actv is the detail table) joined on fy and activity
<dneighbo> and i get a several reporting fields off the actv table for display purposes (all those work fine)
<dneighbo> and the dept lookup works fine
<dneighbo> but i needed an activity lookup
<dneighbo> originally i was going to use the existing datasource of actv (but thought that might be risky)
<dneighbo> so added another datasource for actv to be the lookup
<dneighbo> and now its behaving oddly
<jcater> ah
<jcater> I wish I had a mockup of your database w/dummy data
<jcater> your issues are hard to debug
<jcater> :)
<dneighbo> i can send them
<jcater> bah
<jcater> you were supposed to say "I wish I could... but I could get in trouble"
<jcater> then I'd say, "oh well"
<dneighbo> i did say that
<dneighbo> your irc client must have retranslated on the fly
*** sjc has joined #gnuenterprise
*** bluesbaron_ has joined #gnuenterprise
*** wendall911 has joined #gnuenterprise
*** bluesbaron_ has quit IRC
*** bluesbaron_ has joined #gnuenterprise
*** bluesbaron_ has left #gnuenterprise
*** johannesV_ has quit IRC
<reinhard> cd wahl
<reinhard> sigh
<chillywilly> doh
*** reinhard has quit IRC
*** reinhard_ has joined #gnuenterprise
<dsmith> jamest: 2.6.6 is working great for me, btw
<jamest> dsmith: ?
<jamest> was 2.6.x giving you trouble?
<dsmith> mouse issues
<dsmith> jamest: May it was someone else who told me what modules to load.
<dsmith> s/May/Maybe/
*** jcater has quit IRC
*** jcater has joined #gnuenterprise
*** siesel has joined #gnuenterprise
<siesel> hi guys
<jamest> hi siesel
<siesel> hi jamest, how are your ZOPE servers working at Kansas University?
*** flax07 has joined #gnuenterprise
<flax07> hi there - can anyone tell me the status of the financials package?
<flax07> ...the status bits of the website dont seem to work...
<dneighbo> flax07 we have mostly concentrated on the tools
<dneighbo> and custom applications
*** Vee has quit IRC
<dneighbo> there is no working financial package to download
<dneighbo> there are tools that you could use to write such a package
<flax07> sure - ive been tracking the project for a year or two - hoping that a decent free accounts package would appear!
<flax07> i think along with exchange, and a non-java based office suite - these are crucial to business getting off M$
<dsmith> flax07: the Access Killer
<flax07> yep - as well - openoffice has good, if different to 
<flax07> ...M$, database access.
<flax07> there are some exchange-killer apps appearing - but still nothing that does decent business accounts
<dneighbo> bbiab
*** dneighbo has quit IRC
*** kilo has joined #gnuenterprise
*** dneighbo has joined #gnuenterprise
<dnWork> fwiw: one of the other developers here has hit the "empty" dropdown box problem
*** kilo has quit IRC
*** Vee has joined #gnuenterprise
<dnWork> is there a way to make it so when you enter a field that the text is highlighted already?
<jamest> not currently
<jamest> that shouldn't be hard to do though
<jamest> i think
<jamest> :)
<jcater> funny, I thought it already did that
<jcater> sigh
<jamest> no, it just positions at the end
<jcater> yeah, that's what I said
* jcater coughs
<jamest> holy shit
<dnWork> most windows ide's set the default to highlight the current data
<dnWork> so that is what windows developers and users are expecting :(
<dnWork> at least here
<jamest> neighbors commited 
* jamest runs away in terror
<dnWork> yeah but it was jcater's code
<jamest> ah
<jamest> whew
<jamest> don't do that to me
<jamest> i thought for sure the entire system would collapse
<jamest> svn barely works 24 hrs at a time w/o neighborly love
*** dsmith has quit IRC
*** Vee has quit IRC
*** Vee has joined #gnuenterprise
*** flax07 has quit IRC
*** flax07 has joined #gnuenterprise
*** jcater has quit IRC
*** jamest has left #gnuenterprise
*** sjc has left #gnuenterprise
<dnWork> jcater: you still here?
<dnWork> how do you do sql directly in common apps
<dnWork> actually i probably just need a way to do "distinct"
*** jamest has joined #gnuenterprise
<dnWork> jamest: in gnue common apps (not forms)
<dnWork> how can you set distinct = true
*** reinhard_ has quit IRC
*** siesel_ has joined #gnuenterprise
*** dnWork has quit IRC
*** siesel has quit IRC
*** jcater has joined #gnuenterprise
*** mixi^ has joined #gnuenterprise
*** flax07 has quit IRC
*** R45 has joined #gnuenterprise
*** mixi^_ has quit IRC
*** wendall911 has quit IRC
*** chillywilly has quit IRC
*** dneighbo has joined #gnuenterprise
*** chillywilly has joined #gnuenterprise
*** derek has quit IRC