*** SachaS has quit IRC *** havoc has quit IRC *** sjc has joined #gnuenterprise *** johannesV has joined #gnuenterprise good morning johannesV: hello *** havoc has joined #gnuenterprise hi dimas *** btami has joined #gnuenterprise hello btami btami: sorry to hear about the father :( hi dimas thx btami, i've read it in the log today, and i feel very sorry too *** havoc has quit IRC johannesV: thx, what do you think about introducing an optional feature for dbdrivers not supporting rowid thingy something like tablename_id as keyfield with an autogenerated id is this make sense ? * btami is thinking loud i haven't follwed this discussion in very much detail, so you'd better ask reinhard ok btami, you mean instead of gnue_id ? yep in 2 tier yes, i see well, i could imagine something like a _rowIdField setting per driver (if it's not there already) defining the 'id-field' but, as mentioned before, i don't have a very deep insight regarding this topic atm *** reinhard has joined #gnuenterprise *** dimas has quit IRC *** dimas has joined #gnuenterprise *** btami has quit IRC *** nickr has quit IRC *** johannesV has quit IRC *** sjc has quit IRC *** Vee2d2 has quit IRC *** dcmwai has quit IRC *** mnemoc_ has quit IRC *** chillywilly has quit IRC *** ajmitch has quit IRC *** cilkay has quit IRC *** reinhard has quit IRC *** derek has quit IRC *** tiredbones has quit IRC *** Amorphous has quit IRC *** elwis has joined #gnuenterprise *** elwis has left #gnuenterprise *** Vee2d2 has joined #gnuenterprise *** sjc has joined #gnuenterprise *** kilo has joined #gnuenterprise *** johannesV has joined #gnuenterprise *** reinhard has joined #gnuenterprise *** dcmwai has joined #gnuenterprise *** derek has joined #gnuenterprise *** mnemoc_ has joined #gnuenterprise *** tiredbones has joined #gnuenterprise *** chillywilly has joined #gnuenterprise *** ajmitch has joined #gnuenterprise *** Amorphous has joined #gnuenterprise *** cilkay has joined #gnuenterprise *** nickr has joined #gnuenterprise *** btami has joined #gnuenterprise reinhard: what should I provide for new designer release? NEWS? kilo: people were noticing problems with designer yesterday I think they are not sorted out when calling "run form" from menu there was a traceback seemed to come from initializing the debug system twice or something like that when doing 'run form' a new gfinstance is created, so that seems logical... it's setDebug which actually should be called in GFClient, not in GFInstance but designer has a menu entry a user can use to set debug level how to change that then? oh hmmm will I really have to fire up designer now? :) will look at it asap (~ 1 hour) as to release see gnue-common/doc/technotes/00011.txt :-) gnue-designer/src/forms/Instance.py line 371 *** aries_mindworks has joined #gnuenterprise hi * aries_mindworks testing gaim's irc functions hi aries_mindworks kilo: could you please check what the value of sys.__stderr__ is before that call to setDebug ? wee dicovered a bug just now in forms/Debugger.py sys.__stderr__ is redirected to a class this class must implement a flush () method that's all (that flush() would have to flush() the underlying file) i see although I consider it quite bad to redirect __stderr__ (as that is usually the *original* stderr) it is redirected to a simple windows that shows the debug messages setDebug would take a second parameter with a file handle or compatible object so it *should* work to pass that _stderrcatcher as second parameter to setDebug I think Jason thought that this new windows is more comfortable than seeing debug messages in somewhere in the console I agree, but the above would be a cleaner way to achieve the same however this would probably need testing also, what I'm missing is resetting the debug to the original state hi reinhard: going to provide a new designer release or shall I apply the patch? which would mean that debug messages of designer wouldn't work any more after a form was run ajmitch: I think we want to provide a new release ok * ajmitch has only 2 weeks before the hoary release reinhard: it is reset in Debugger.py line 80 kilo: that's not enough blame Jason if you look at setDebug in GDebug.py it is all his code and in fact i cannot say that i feel comfortable hacking this... kilo: I understand that :) kilo: if you look at setDebug in GDebug.py, you can see that the current sys.__stderr__ is saved as the filehandle to write debug and error messages to resetting sys.__stderr__ in Debugger.py won't help, because the _fh in GDebug will still point to the _stderrcatcher of designer which then points to a non-existing window which actually *might* even be the reason for the problems with exiting from a debugged form * btami hopes finally reinhard will fix this bastard :) Debugger.py line 144 does this, doesnt it? well, actually there's parts of gnue that i don't want to mess with :) kilo: ahh now I see what broke it GDebug._fh is only half of the truth __stderrcatcherr.filehandle is the other half * kilo locks reinhard and jason in a room for a day and throws away the key. very very far... so instead of setting sys.__stderr__ and GDebug._fh _stderrcatcher should do GDebug.__chatchStderr (self) and in _restore, GDebug.__catchStderr (self.origfh) but anyway the flush() method needs implemented this doesn't change and that will fix the traceback * kilo feels that reinhard wont touch this... I can't even test properly actually I'm not even sure if I have wx installed here btami: hi kilo/btami: from the code i read that adodbapi and interbase have problems with cursor.rowcount it must have now I have read in DBSIG2 docs that for some db's rowcount is only available after the first fetch is there any chance you can test that? I'd like to implement the handling of broken rowcount in the general DBSIG2 driver so adodbapi and interbase wouldn't need special handling sure thanks where is rowcount used? we've started a form connected to a firebird db and it works *** elwis has joined #gnuenterprise kilo: please compare interbase's DataObject._createResultSet with the standard DBSIG2 one * kilo lost deep in designer code DBSIG2 just does i just reported as btami was eating pizza... recordCount = cursor.rowcount if self._strictQueryCount: rs._recordCount = recordCount while interbase does a query like SELECT count(...) from... to determine the _recordCount of the resultset and I wondered if cursor.rowcount is broken for interbase or if there is another reason why interbase does it differently than the other DBSIG2 drivers from kinterbasdb docs: rowcount attribute Although KInterbasDB's Cursors implement this attribute, the database engine's own support for the determination of "rows affected"/"rows selected" is quirky. The database engine only supports the determination of rowcount for INSERT, UPDATE, DELETE, and SELECT statements. When stored procedures become involved, row count figures are usually not available to the client. Determining rowcount for SELECT statements is problematic: the rowcount is reported as zero until at least one row has been fetched from the result set, and the rowcount is misreported if the result set is larger than 1302 rows. The server apparently marshals result sets internally in batches of 1302, and will misreport the rowcount for result sets larger than 1302 rows until the 1303rd row is fetched, result sets larger than 2604 rows unti l the 2605th row is fetched, and so on, in increments of 1302. As required by the Python DB API Spec, the rowcount attribute "is -1 in case no executeXX() has been performed on the cursor or the rowcount of the last operation is not determinable by the interface". wee thanks btami *** aries_mindworks has left #gnuenterprise elwis: sorry, odbc driver has not been maintained for years; I've finally removed it from last week as it was not functional for quite some time :( elwis: as to master detail, I think you misunderstood even if it sounds wrong, the master is always the "main" table of the form, in your case the order and the "detail" is then the customer bbl reinhard: no odbc, not in the future either then? elwis: maybe you have to try the adodbapi dbdriver the problem is that I don't have windows and can't help redoing the driver for odbc should not be a very big problem reinhard: Good for you ;) (at least if you don't need introspection support) but I can't do it :) bbl now for real btami: How would I try out the adodbapi driver? btami: Maybe I'll start with walking over to my colleague who wrote the ODBC drivber :) elwis: here is my two examles from connections.conf [xls] provider = adodbapi oledb_provider = Microsoft.Jet.OLEDB.4.0 data_source = C:\home\proba.xls extended_properties="Excel 8.0;HDR=Yes" [booksale] provider = adodbapi oledb_provider = Microsoft.Jet.OLEDB.4.0 data_source = C:\home\booksale.mdb database = valami and another one: [tabl_dbf] comment = tabl_dbf.dbf adodbapi-val provider = adodbapi oledb_provider = vfpoledb data_source = c:\home\tabl_dbf.dbf Thanks a lot, I'll check those out we have some docs generated from *.py sources for all our dbdrivers like this: http://www.gnuenterprise.org/tools/common/databases/adodbapi.php Thanx that will be useful when showing my win-friends at work Won't work with our ERP system though ;) A quick one, how well does the menu work? I saw Saschas "BabyERP" screenies and want one :) *** sjc has quit IRC it's far from perfect, but works Thanks for the info, seem to lack som documantation, both the Navigator and Reports yes any help are welcome Actually, I was planning to write my "GNUe for the helpless" .. I'm the man for it :) cool ! it will be a success story I'' give you notice when it's ready, then someone could check it, and maybe put it on the wiki For other helpless ones.. ;) it will sell out in the first week... maybe I'll try to do business from that instead from my GNUe cash-registry.. elwis: knowledge about gnue is a terse good on this world ;-) bbl *** btami has quit IRC * reinhard is happy about the increasing number of translations we get for this project now we seem to have de, hu, nl, and it with people that constantly maintain them or at least intend to do so also ru (sorry dimas) :) *** dimas has quit IRC *** Vee2d2 has quit IRC *** kilo has quit IRC *** Vee2d2 has joined #gnuenterprise bbl *** kilo has joined #gnuenterprise *** havoc has joined #gnuenterprise *** lekma has joined #gnuenterprise hi everyone reinhard: I'm back, and i have to spell out my intention to get the knowledge to translate to swedish elwis: great! reinhard, it looks like gnue-schema is broken now ... hmm, if i add a 'primarykey' definition to the DataSourceWrapper's attribute in gnue-schema it seems to work again ... i think gnue-schema needs an 'optimization run' ... :) *** kilo has quit IRC *** jamest has joined #gnuenterprise *** jamest has joined #gnuenterprise Traceback (most recent call last): File "/usr/lib/gnue/appserver/geasSessionManager.py", line 399, in call result = s.call (classname, obj_id_list, procedurename, parameters) File "/usr/lib/gnue/appserver/geasSession.py", line 659, in call result.append (instance.call (procdef, parameters)) File "/usr/lib/gnue/appserver/geasInstance.py", line 362, in call method = cx.buildFunction (proceduredef.gnue_name, proceduredef.gnue_code, File "/usr/lib/gnue/appserver/repository.py", line 373, in __getattr__ result = getattr (self.__object, attr) File "/usr/lib/gnue/appserver/language/Object.py", line 99, in __getattr__ res = sm.load (self.__sid, self.__class, [self.objectId], [name]) File "/usr/lib/gnue/appserver/geasSessionManager.py", line 358, in load result = s.load (classname, obj_id_list, propertylist) File "/usr/lib/gnue/appserver/geasSession.py", line 537, in load result.append (instance.get (propertylist)) File "/usr/lib/gnue/appserver/geasInstance.py", line 261, in get return [self.__getValue (aProperty) for aProperty in propertylist] File "/usr/lib/gnue/appserver/geasInstance.py", line 254, in __getValue return self.__convert (value, propertydef, DbValueError) File "/usr/lib/gnue/appserver/geasInstance.py", line 124, in __convert return unicode (value) UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 220: ordinal not in range(128) this one is new yay, traceback ;) nothing like a good traceback in the morning lekma, i get the same error if i use german umlaut ... seems as if there's something broken .... (i'm using sample.gfd [if reinhard is reading this]) i can't seem to find out what has changed between yesterday and today that triggers this error yes, i've thought the same ... was it working for you yesterday ? yep that was the same proc that triggered the commit pb you fixed yesterday so i assume smthing changed in between you fixed it and now cause neither the proc nor the way it's called changed hmm, i've to run now but i'll look into it in the evening ... maybe reinhard could have a look at it ... thx *** elwis has quit IRC *** titopbs has joined #gnuenterprise anybody can help with reports? *** jcater has joined #gnuenterprise *** johannesV_ has joined #gnuenterprise *** johannesV has quit IRC jcater: can you help with reports? i have a strange pb i finally made a master/detail report (still first stepS) but i have one more detail line than available in data :) i have a master (a bill) which has 2 details (bill items) but when i fire the report i get 3 lines in the detail table the first one being repeated at the end the backend is appserver and i double checked in db: there really is only 2 records *** kilo has joined #gnuenterprise *** SachaS has joined #gnuenterprise well in fact if you make master/detail datasources in reports the result is completly screwed in the previous description the detail that appears first and third isn't even related to the master hi gnuers hi SachaS hi lekma *** dimas has joined #gnuenterprise lekma: plz check gnue-invoice's report in ggnue-contrib, iirc it has m/d kilo: thx but blame btami for it 8-) can't seem to find a gnue-contrib is an other svn trunk ? s/is/is it it's in it's own repository co gnue-contrib instead of gnue or download a tarball from http://www.gnuenterprise.org/cgi-bin/viewcvs.cgi/gnue-contrib/ *** jamest has left #gnuenterprise *** wendall911 has joined #gnuenterprise *** elwis has joined #gnuenterprise *** jamest has joined #gnuenterprise *** monolithic has joined #gnuenterprise *Ouch* I'm on page 3 in my "GNUe for the brainless" guide... and I just reached how to install the packages. is the correct packagename "gnue-forms-wxgtk"? Or is that Hoary only? packages are different in different distro that's what I thought, didn't recognize the name. I'll stick to Hoary only then hm packages that required for gnue that is not sure about gnue-* packages themselfs Oh. I thought wxGTK is "always" needed on *nix platform? Not true? i have no wxGTK right now at all here is altlinux.org distro on one machine So, what do you use for your ui? GTK? yep Wait a second? How does gnue-forms make that desicion if you don't specify a parameter? What's the "default"? Cause my forms looked somehow more polished on the Novell-GNOME Desk then on my Hobby Hoary? there is DefaultUI in gnue.conf elwis: it's a bit confusing, sorry about the package names the *plan* was to have gnue-forms-wxgtk gnue-forms-gtk2 gnue-forms-curses gnue-forms-kde3 and the first one to run smoothly was wxgtk then, when the others were finished, debian had stopped accepting of new packages (in anticipation of sarge release) so to have everything there, we packed everything in gnue-forms-wxgtk (which was already accepted) so currently, gnue-forms-wxgkt contains *all* ui drivers but that will change as soon as we seem to be able to get the other packages into debian clever, clever, so if I mention to install gnue-forms-wxgtk, they'll have it all :) ... today... yep, sorry for that confusion however, the package names are identical for ubuntu and debian (and, fwiw, the package content) but then I shouldn't mention any need of having wxGTK installed? GTK2 would also do? yes but you *should* probably mention that the defaultUI should be set in gnue.conf not until I checked that setting myself.. I didn't know that.. :) the difference you see with wx (the default) might come from different wx versions brb changed to gtk2, it's still not as nice as the Novell GNOME, supose the themes comes in to play though. Is it possible to mess with fonts for example? yes, I think it's the themes bbl *** reinhard has quit IRC Hello all hi monolithic Hate to sound odd, but I would like to help w/ the project Not a programmer, but am a small business owner there are few the same here Don't wanna sound pushy, but I do have ideas about what I would like to see implemented monolithic: you could help greatly What can I do w/o getting in the way? i think that depends more on what you'd like to do than what we need done share with your ideas of what and how you see it need to be done asking you to do something you don't enjoy will chase you away :) jamest, I never really have gotten involved w/ something like this before. Should I just start rambling? First, is gnue useable at all right now? the tools most definately reinhard: for the logs: result of connection.requery() is not unicode !!! *** titopbs has quit IRC monolithic: you're free to rambling in here but i can't promise timely responses i have to go beat a co-worker *** reinhard has joined #gnuenterprise doh! reinhard: for the logs: result of connection.requery() is not unicode !!! I think that was a good hint :) Gnue obviously is about general accounting (accounts payable, receivable, etc.), but here is what I would like to see. invoicing estimating job costing scheduling & dispatch integrate w/ common office apps Are any of these in the works? there is gnue-invoice in contrib gnue is mainly about the tools to create business apps quickly and easily we are to have some apps built upon it but they never seem to get much past concept for instance in house my invoicing system is all gnue based I'm an hvac contractor, but I hobby w/ gnu/linux at home. I would love to see gnu/linux for contractors like me. monolithic: hvac? However, biggest hurdles are standard apps, special apps, and even custom apps (like load calcs, etc.) heat & air conditioning ok is svn dead again? seems to be try now great, that unicode-problem is fixed ! thanks reinhard lekma, please svn up and try again ... i'm shutting off apache-ssl and postgresql on ash unless someone has a reason we need to run them johannesV_: yep it works now reinhard: thx bb bbl * lekma is away: feeding the cat... *** johannesV_ has quit IRC *** titopbs has joined #gnuenterprise gack!!!! python expert question def foo (bar = {}): print bar bar ['x'] = 100 foo () foo () ---------- what does the above code print? interesting so the default dict is only initialized once eh i wouldn't have thought that http://www.python.org/doc/2.3.5/ref/function.html nice little trap for example we do things like def __init__ (self, parent, initialData = {}, defaultData = {}) and later we modify the initialData in the code and voila i wonder if that is the reason for some oddities that i get in datasources once in a great while the next record inserted gets the data of the previous one jamest: there are quite a few potential reasons for oddities hidden in datasources ;-) lol they've only been rewritten what now 3 times? but this has made me believe like I was insane you are er, i mean, really? I called the __init__ with initialData not set at all and had data! in function header I give default value of {} and right as the first line of the function I did print initialData and got {'oid': 81203} and was like "wtf??"" reinhard: while youre in there (datasources) one thing i've noticed is that gnue datasources leave connections in transaction so if I want to update the structure of a table in postgresql I have to tell all my forms users to exit the application as they block the table updates isn't that a DBSIG2 module problem? i don't recall our sql code issueing a FOR UPDATE on the selects which i think would cause that no, it doesn't but psycopg & friends do a BEGIN right on connect (IIRC) sigh and immediately after commit and rollback that's special *Ouuch* My neck, my back.. my brain.. page 11 and finally the structure of a .gfd file.. * reinhard remembers monty python's "doctor! doctor! my brain hurts ... me" *********************** johannesV: you can reverse your change in gnue-schema it was a bug in datasources, fixed it thanks for leading me to this awful bug *********************** *** elwis has left #gnuenterprise *** nickr has quit IRC *** dcmwai has quit IRC any1 knows an msn client for linux ? *** nickr has joined #gnuenterprise I think that's an oxymoron :) my wife wants to chat with friend who use that thing... kilo: kopete offers it though i've not used it hmm, i think i will recommend her to use skype... *** dcmwai has joined #gnuenterprise *** jamest has left #gnuenterprise *** jcater has quit IRC *** reinhard has quit IRC *** wendall911 has quit IRC hi SachaS *** mnemoc has joined #gnuenterprise *** kilo has quit IRC *** mnemoc_ has quit IRC *** titopbs_ has joined #gnuenterprise *** titopbs_ has quit IRC *** titopbs has quit IRC *** monolithic has left #gnuenterprise *** jamest has joined #gnuenterprise *** titopbs has joined #gnuenterprise *** jcater has joined #gnuenterprise *** titopbs has quit IRC *** holycow has joined #gnuenterprise *** titopbs has joined #gnuenterprise *** titopbs has quit IRC *** jcater has quit IRC *** holycow has quit IRC *** cilkay has left #gnuenterprise