*** SachaAway has quit IRC *** johannesV has joined #gnuenterprise good morning yikes morning for you already that is a sure sign it is my bed time :) *** reinhard has joined #gnuenterprise bbl good morning *** kilo has joined #gnuenterprise good morning so is it a good day or a bad one? *** btami has joined #gnuenterprise im just getting ready for bed and would like to know if tomorrow is worth waking up for ;) hmmm hard to answer that btami \o/ hi all *** johannesV_ has joined #gnuenterprise *** johannesV has quit IRC * johannesV_ back *** yure has joined #gnuenterprise *** erkki__ has joined #gnuenterprise *** erkki_ has quit IRC *** erkki__ has quit IRC *** erkki_ has joined #gnuenterprise *** erkki_ has quit IRC *** Jave27 has joined #gnuenterprise *** erkki has joined #gnuenterprise *** yure has quit IRC hello, a question about appserver procedures.. how are concurrent procedures handled? is some sort of locking/queueing included in appserver or is it up to the application writer to do that? can you be a bit more precise about the word "concurrent" ? do you have two sessions running the same procedure at the same time ? appserver doesn't introduce it's own locking ... this is done by the backend (by using transactions) lol, reinhard the mail you sent to my gnue address arrived *now* :) appserver usually forks a new process per session (on connect) *** yure has joined #gnuenterprise yeah, I mean two sessions running the same procedure, at the same time.. can the two procedures run in paraller or do they run in some sort of queue for example, incrementing/decrementing a counter in a procedure, by getting the value first, adding something to it, and then saveing the value.. does appserver some how force the procedures run after eachother, or must some sort of locking be included in the procedure itself bbl :w *** kilo has quit IRC as i've mentioned before appserver starts a new process per session, so there's no queue and both procedures are running parallel (depending on the os and it's concept of multitasking :) regarding locks: this depends on the backend you're using if you can access the same information from within multiple transactions ... but i think you'd like to have some sort of 'manual' locking in such a situation ok, thanks, I think things got clearer.. does appserver automatically use transactions for its stuff? .. presuming that the backend supports transactions.. erkki: yes, it does ok, thanks I'm still thinking about the locking issue I asked about earlier.. is it possible to "hint" appserver to use certain sql features? for example, to tell postgres to use certain locks within a query/transaction? there is no support for locking in appserver at all so far and we avoid implementing backend specific features locking is on our todo list, although way down ok, understandable.. was just thinking about the feature to fall-back on sql level for certain things but I suppose that would invalidate the whole point of the application-server.. gotcha right now you can implement a complete application on sqlite then at one point you go into production you can switch to postgres with changing a single line in connections.conf (more or less) then when app grows, you can easily migrate to maxdb if you like yeah and there might be a time when you can, for whatever purpose, use csv files as your backend .. when someone writes the backend driver?:p there *is* a csv backend driver, but it only supports read-only access so far oh it's good for testing reports without real db hmmm..... actually I just think of a project where it could really be cool to use csv as a backend it's a simple address maintenance with forms (without appserver) yep storing things in csv would enable the user to open his data "tables" directly in excel or OOo without any exporting for creating a mailing, for example *** sjc has joined #gnuenterprise why should one change from postgres to maxdb ? bbl maybe because he's consultant and wants to sell more working hours ;-P .. gotta add that to my bag-of-tricks?:p gnue-gsdgen -o test -c appserver Loading class repository ... Reloading class repository ... DB000: Traceback (most recent call last): DB000: File "/usr/lib/gnue/common/rpc/drivers/xmlrpc/ServerAdapter.py", line 278, in do_POST DB000: params) DB000: File "/usr/lib/gnue/common/rpc/drivers/xmlrpc/ServerAdapter.py", line 124, in call DB000: result = self._tcpServer._dispatch (method, params) DB000: File "/usr/lib/python2.4/SimpleXMLRPCServer.py", line 406, in _dispatch DB000: return func(*params) DB000: File "/usr/lib/gnue/appserver/geasSessionManager.py", line 294, in request DB000: result = s.request (classname, conditions, sortorder, propertylist) DB000: File "/usr/lib/gnue/appserver/geasSession.py", line 339, in request DB000: classdef = self.__getClassdef (classname) DB000: File "/usr/lib/gnue/appserver/geasSession.py", line 128, in __getClassdef DB000: return self.sm.repository.classes [classname] DB000: File "/usr/lib/gnue/common/utils/CaselessDict.py", line 140, in __getitem__ DB000: raise self.__notFoundError, tuple ([dkey] + list (self.__exceptionArgs)) DB000: ClassNotFoundError: Class 'gnue_module ' not found in class repository any ideas anyone?:) the appserver connection is working, and has a working schema inside.. at least forms works with a small inventory app latest release version of tools.. hmmm *** SachaS has joined #gnuenterprise strange errm no not strange you must run gnue-gsdgen on the db connection, not on the appserver connection ok, will try that.. getting another error when trying with the db gnue-gsdgen -o test -c gnue Loading class repository ... Reloading class repository ... Class repository loaded Building list of classes and calculating dependencies ... Generating schema definition ... Exporting data of class 'INVENT_itemSize' ... DB000: Traceback (most recent call last): DB000: File "/usr/bin/gnue-gsdgen", line 34, in ? DB000: gsdGen.run () DB000: File "/usr/lib/gnue/appserver/geasGsdGen.py", line 162, in run DB000: self.__exportClass (classname) DB000: File "/usr/lib/gnue/appserver/geasGsdGen.py", line 205, in __exportClass DB000: self.__normalDataDump (cDef, fieldlist, rows) DB000: File "/usr/lib/gnue/appserver/geasGsdGen.py", line 223, in __normalDataDump DB000: (dts, fields) = self.__openSource (classDef, fieldlist) DB000: File "/usr/lib/gnue/appserver/geasGsdGen.py", line 787, in __openSource DB000: attributes = {'name' : "dts_%s" % table, DB000: AttributeError: geasSessionManager instance has no attribute '_internalSession' ok do you run svn or release? release the svn server was down a few days ago:p svn co http://svn.gnuenterprise.org/gnue/trunk gnue svn: REPORT request failed on '/gnue/!svn/vcc/default' svn: REPORT of '/gnue/!svn/vcc/default': 400 Bad Request (http://svn.gnuenterprise.org) seems like a svn internal db error or something.. ouch I can't reproduce the error at the moment will be back later, must run ok *** derek has quit IRC *** derek has joined #gnuenterprise erkki: is it possible that you have some version mismatch? e.g. some parts of a svn install mixed with a release install? bigbrother_: weather 85233 derek: The current temperature in Weather Underground is 45°F. reinhard: possibly, I had a old install of gnue maybe 6 months ago.. now, when I installed the newer versions, I had to manually remove the pyc files to get things running ah oh and the svn thingy, possibly a error on my side also.. as a checkout in another machine worked for me too ok the error above seems to be a mixture of versions will try cvs and then see if the gsd error is there also or not ok *** derek has quit IRC *** derek has joined #gnuenterprise *** Jave27 has quit IRC should file="asd.py" tags in gcd files work with relative paths? appserver seems to require full paths, when reading classes from a "module directory" s/tags/attributes/ problem is that putting a program in the background makes it forget its current working directory so relative paths don't work if you run appserver in background *** jamest has joined #gnuenterprise ok gpl 3 draft is out http://gplv3.fsf.org/draft *** johannesV__ has joined #gnuenterprise *** johannesV_ has quit IRC reinhard: update on the gsdgen problem.. got the same error when I installed the release version on a clean machine, so no version mismatch things.. tried cvs then and gsdgen worked .. why is it always, that I get gnue bugs that are fixed on the cvs and not fixed on the current "release":p.. should try to learn to test cvs more often:p hmmm ok so it must have been fixed by chance we didn't even know this bug existed... *** btami has quit IRC reinhard, that's not true i've fixed that bug while i migrated to the object oriented rpc interface that was where i've found it ... *** yure has quit IRC *** johannesV__ has quit IRC oh ok ./gnue-navigator/doc/gnue-navigator.dtd oops wrong paste buffer:p Traceback (most recent call last): File "/home/erkki/gnue/.cvsdevelbase/gnue/navigator/UIgtk2.py", line 723, in __about text % (VERSION,'-','-','-','-')) TypeError: flag values must be strings, ints, longs, or tuples that's the right one.. navigator using gtk2 ui frontend.. "about" menu "help>about"* huiii I think it's been a time that I've last looked at navigator... *** kilo has joined #gnuenterprise good night *** reinhard has quit IRC @seen btami kilo: btami was last seen in # 6 hours, 34 minutes, and 8 seconds ago saying: yep kilo: btami was last seen here 6 hours, 34 minutes, and 11 seconds ago saying: yep damnz *** SachaS_ has joined #gnuenterprise *** SachaS has quit IRC *** jamest has left #gnuenterprise *** sjc has quit IRC *** kilo has quit IRC *** derek has quit IRC *** erkki has quit IRC *** jamest has joined #gnuenterprise *** jamest has quit IRC