*** SachaS has joined #gnuenterprise
*** holycow has quit IRC
*** dimas has joined #gnuenterprise
*** SachaS has quit IRC
*** reinhard has joined #gnuenterprise
* reinhard loves to see work starting again on qt driver
*** reinhard has quit IRC
*** sjc has joined #gnuenterprise
*** reinhard has joined #gnuenterprise
*** yure has joined #gnuenterprise
*** yure has quit IRC
*** reinhard has quit IRC
*** reinhard has joined #gnuenterprise
*** sjc has quit IRC
*** yure has joined #gnuenterprise
*** yure has quit IRC
*** yure has joined #gnuenterprise
*** jamest has joined #gnuenterprise
<jamest> reinhard: you here?
<reinhard> yes
<jamest> sorry
<jamest> was away
<jamest> do you know of a way to do the following
<jamest> i have a gnue.common.gui.navigator.[qt|gtk2|win32]
<jamest> and a gnue.common.gui.window.[qt|gtk2|win32]
<jamest> i'd like to somehow set the widget set to one of the options then in my code do things like
<jamest> from gnue.common.gui import window
<reinhard> I'd look how "import os.path" works
<reinhard> it's quite the same system AFAICT
<reinhard> as "os" maps to something dependent on the platform
<jamest> and have it be the widget set i selected before
<jamest> ok
<reinhard> FWIW I'd do gnue.common.gui.[qt|gtk2|win32].window rather than the other one
<chillywilly> I'd go for the abstract factory pattern but maybe that's not pythonic enough ;)
<reinhard> so all of qt is in a single tree
<jamest> i thought about do that as well
<reinhard> "navigator" does not relate to gnue-navigator, does it?
<jamest> and looking into using the plugin system
<reinhard> IMHO having forms|navigator|designer specific code in common defeats the purpose of common :)
<jamest> it does not
<reinhard> ok
<reinhard> :)
<jamest> navigator is working name for the top level widget struture that will manage the windows under it
<jamest> you could think of it as the MDI wrapper and main event loop for a ui
<reinhard> ah ok
<jamest> i'm really just playing around now
<jamest> but I'd like to get the basic wrapper up and going w/o any features
<jamest> fwiw
<jamest> any common.gui widget will be an inherited from a class in the desired widget set
<jamest> and there will only be a handfull of common.gnu widgets
<chillywilly> os just compares names in sys.builtin_module_names and then uses if statements to do stuff like import posixpath as path
<chillywilly> import ntpath as path
<chillywilly> import macpath as path
<jamest> i didn't think that would work
<chillywilly> not really the same issue
<reinhard> def init_ui (name):
<reinhard>   if name == 'gtk2':
<reinhard>     from gtk2 import window, navigator
<reinhard>   elif name == 'qt':
<reinhard>     from qt import window, navigator
<reinhard> ...
<reinhard> ?
<jamest> but if you put that in common.gui.__init__
<chillywilly> not very elegant
<reinhard> working with __import__ would be better still
<jamest> wouldn't that screw up imports in say common.gui.gtk if they tried to import common.gui._base
<reinhard> don't think so
<reinhard> as only window, navigator... are redefined
<jamest> you do agree that the plugin.py isn't a good fit here right?
<jamest> as it's not a single class we're interfacing with
<reinhard> not sure
<jamest> it's a library of widgets
<reinhard> dbdrivers are not a single class, either
<reinhard> plugin.py gives you a complete module
<reinhard> it just *checks* wheater a class is there to see if it's a valid module
<jamest> but a module developer would do something like
<jamest> from gnue.common.gui import window
<jamest> from gnue.common.gui import menu
<jamest> and the like
<chillywilly> I wrote a plugin system that uses python metaclasses to validate the class structure...that certain methods were there, etc.
<chillywilly> was fun
<chillywilly> :)
<reinhard> still think it could work
<reinhard> but probably no need to use plugin.py
<reinhard> hmmm
<chillywilly> I still say abstract factory pattern :)
<chillywilly> the classic example is the tooliy independent UI
<chillywilly> toolkit*
<chillywilly> but I know there's gotta be a more pythonic way to implement it
<jamest> reinhard: you guys using hotilne?
<reinhard> yes
<jamest> i added that detail datasource driven dropdown
<jamest> and the form blows up when adding new records
<jamest> modifying old ones is fine
<reinhard> hmmm
<reinhard> in our case the master is a readonly datasource
<jamest> new ones do the insert then puke in the __do_requery
<reinhard> what kind of puke?
<jamest> i'll trace it back now
<jamest> as it's dying on an AttributeError: keys
<reinhard> ok, that happens (IIRC) if the requery issues an empty result set
<reinhard> actually I should raise a better exception in that case
<reinhard> can you check the SQL statements, whether the requery statement is correct?
<reinhard> do you have up to date svn? there was a bug in requery code about a week ago or so
<reinhard> or two
<chillywilly> http://jamesthornton.com/eckel/TIPython/html/Sect08.htm#Heading61
<jamest> i should be bleeding edge now
<jamest> but i'll check
<chillywilly> ASPN is down but it looks like there's some nice recipe in there too
<jamest> chillywilly: thanks
<jamest> let me move my irc to the devel box
*** jamest has quit IRC
*** jamest has joined #gnuenterprise
<jamest> it's blowing up in the Base/RecordSet.py
<jamest> line 597
<chillywilly> wb
<chillywilly> http://www.pasteur.fr/formation/infobio/python/ch18s06.html
<chillywilly> I think his example is weak
<chillywilly> looks like a nice book on programming python though :)
<jamest> reinhard: how'd you make the datasource readonly?  i don't see that as an <datasource> attribute
<reinhard> it is "de facto" readonly
<jamest> oh, so is mine
<reinhard> the master one I mean
<reinhard> so I can't insert a new master
<jamest> i can't eithr
<jamest> what I have is a item form that lets you set an item to a certain family(class) of parts speaker, amplifier, etc
<reinhard> jamest: i added that detail datasource driven dropdown
<reinhard> jamest: and the form blows up when adding new records
<jamest> then on the 2nd form tab i have a dropdown that lists a set of features based upon the item's class
<reinhard> ah
<reinhard> so your master is a dropdown
<jamest> so user can go to a specific radio
<reinhard> and your detail is another dropdown?
<jamest> go to second tab
<jamest> and the dropdown shows the valid "features" a radio can have
<jamest> the user selects a feature (say channels or watts) and types in a value
<jamest> so the datasource that populates the dropdown is a detail of the item table
<jamest> linked to it's field that sets the class of item
<jamest> eventually it'll work like object properties in designe
<jamest> r
<jamest> but for now the form dropdown fakes that functionality
<chillywilly> http://linux.sys-con.com/read/44251.htm?CFID=109398&CFTOKEN=A8CB3A68-FC4D-C858-A92214F8A0588B60
<chillywilly> the funny stuff you find when googling
<jamest> as a speaker may track size and a radio may track watts, mp3 capability, etc as features for that class of item
<jamest> hmmm
<jamest> DB003: 0:00:13.695 [Connection:284] DBSIG2 Statement: SELECT item_id, family_id FROM items WHERE oid IS NULL
<jamest> in __do_requery my self.__wherefields () is set to {'oid':None}
<reinhard> can you please check again if you are up to date?
<reinhard> looks like the bug we fixed 2 weeks ago
<reinhard> and I added an explicit exception for the case when requery doesn't find the record it looks for
<jamest> lol
<jamest> still bombs
<jamest> just in a new spot :)
<reinhard> what exception?
<reinhard> erm hold it
<reinhard> seems like I can reproduce
*** jcater has joined #gnuenterprise
<jamest> reinhard: right now it's an error in the Expception
<reinhard> already fixed that
<reinhard> investigating on the other one
<jamest> i'll revert mine then :)
<jamest> had the warehouse mgr pull me away from my task
<jamest> chillywilly: thanks for the link, nice to see someone else feel that way :)
*** jamest has quit IRC
*** jamest has joined #gnuenterprise
* jamest forgets just how bad kopete in kde 3.3 was
<reinhard> aaaaaaaa!!!!!
<reinhard> fsck
<reinhard> found it
* chillywilly just found out that AJAX ~= XMLHttpRequest crap
<chillywilly> stupid buzz words
* jcater crosses his fingers that whatever it was isn't in the commit I did last night
<reinhard> jcater: no
<reinhard> :)
<jcater> whew
<jamest> jcater: it's a datasource issue
<jcater> with my luck lately, I wouldn't be surprised :)
<jamest> indirectly though you could take the blame
<jamest> after all they did have to read through the old caterized datasource code while crafting the new stuff
<jamest> that's bound to have long lasting effects
<reinhard> jamest: don't forget about derek when talking about blame
<jamest> oh yeah
<jamest> it's been so long
<jamest> last night after thwapping cater for the first time in years
<jamest> it hit me how long it's been :)
<jamest> reinhard: is it a simple fix or something that'll be a while
<chillywilly> same old GNUe
<reinhard> already committed
<jamest> great
<chillywilly> let the beatings commence!
<jamest> i'd already moved that form into production
<jamest> :)
* jcater troutslaps chillywilly 
<chillywilly> owey
<reinhard> jcater: you're at jamest's house?
<jcater> no, he kicked me out sunday
<reinhard> grrrr
<reinhard> when will you finally come to visit *me* ?
<reinhard> ;-)
<jcater> what's the closest airport?
<jamest> reinhard: it was pretty cool
<jamest> i talked to him about what I want to do with the workbench
<reinhard> jcater: you would probably want to go via zurich
<jamest> showed him a flash demo of how designer should work
<jamest> you could see the geek in him be reborn, sheding his thick skin of upper management cruft built up over the last few years
<reinhard> lol
<jamest> fwiw
<jamest> let me find that flash
<jamest> as it kicks ass
* chillywilly is playing with M$ Project :(
<jamest> http://www.netbeans.org/kb/articles/matisse.html
<jamest> go to tthe section
<jamest> To get a feel for what Matisse looks like, watch or download the flash demo by Roman Strobl.
<jcater> the flight cost isn't too bad
<jcater> but I'm a horrible guest, reinhard
<jcater> just ask jamest
<jamest> reinhard: actually everything was ok, except his "No pants after 3pm" policy
<jcater> so you live in germany, reinhard ?
* jcater ducks
<jamest> reinhard: works perfectly now
<jamest> thanks
<reinhard> jcater: austria
<reinhard> (south of germany)
<chillywilly> so you want designer to work like matisse?
<chillywilly> nifty demo
*** SachaS has joined #gnuenterprise
<jamest> chillywilly: i think it would be nice
*** jamest has quit IRC
*** jamest has joined #gnuenterprise
<chillywilly> jamest: get to coding then ;)
<chillywilly> doesn't seem like it would be very simple :)
<jamest> why do you think i showed it to jason
<jamest> it's not I'd want to code it :)
<reinhard> jamest, jcater: I would *seriously* welcome you here if you ever come to europe
<jcater> reinhard, I couldn't before next spring
<jcater> but I might take you up on that
*** jamest has quit IRC
<chillywilly> what about me reinhard ?
<chillywilly> :)
<mnemoc> only them
<chillywilly> HEH
* chillywilly slaps mnemoc around some
<mnemoc> outch
*** jamest has joined #gnuenterprise
*** sjc has joined #gnuenterprise
*** jamest has quit IRC
*** jamest has joined #gnuenterprise
*** jamest has quit IRC
*** mnemoc_ has joined #gnuenterprise
*** mnemoc has quit IRC
*** jamest has joined #gnuenterprise
*** jamest has quit IRC
*** wendall has joined #gnuenterprise
*** jamest has joined #gnuenterprise
*** mixi has joined #gnuenterprise
<mixi> someone knows if there is a channel for wx widgets?
*** jamest has quit IRC
<mixi> or possibly help me as i have a memory problem in a wx widgets app
<mixi> no leaks, but it consumes loads
*** dimas has quit IRC
*** chillywilly has quit IRC
*** chillywilly has joined #gnuenterprise
<chillywilly> 3
<chillywilly> bah
*** jcater has quit IRC
*** jamest has joined #gnuenterprise
*** jcater has joined #gnuenterprise
<reinhard> night all
*** reinhard has quit IRC
*** wendall911 has left #gnuenterprise
*** jamest has quit IRC
*** jamest has joined #gnuenterprise
*** SachaS has quit IRC
*** jamest has left #gnuenterprise
*** sjc has quit IRC
*** sjc has joined #gnuenterprise
*** yure has quit IRC
*** jcater_ has joined #gnuenterprise
*** sjc has quit IRC
*** jcater has quit IRC
*** jamest has joined #gnuenterprise
*** mixi has quit IRC
*** wendall911 has joined #gnuenterprise
*** jcater_ has joined #gnuenterprise
*** wendall911 has left #gnuenterprise
<ajmitch> good news is that breezy has both wx 2.4 & 2.6 now
<ajmitch> bad news is that I have to make minor source patches to explicitly select 2.4
*** holycow has joined #gnuenterprise
*** jcater has quit IRC
<jamest> breezy?
<jcater> is that the next ubuntu?
<jamest> that's a toy story character
<jamest> at least I think it is
<jamest> the penguin that lost it's squeeker
<jcater> that's squeek, I thought
<jcater> or squeeky?
<jamest> are we really discussing this in a logged public channel?
<derek> since it is being logged
<derek> :-P
<jamest> lol
<jcater> I was thinking the next debian was etch
<jcater> s/was/is
<jamest> you are right
<jcater> the one scheduled for, what 2015?
<jamest> ubuntu is breezy
<jamest> wow, they have a lot of faith in the streamlined development processes if '15 is the target :)
<ajmitch> yes breezy is the next ubuntu release, due out in october
*** jamest has quit IRC
<ajmitch> sid essentially has the same wx* packaging
<ajmitch> 2.6 is in experimental at the moment
*** jamest has joined #gnuenterprise
<ajmitch> wb jamest
<ajmitch> if reinhard gets new releases out, I can probably push them into breezy
<jamest> thanks
<ajmitch> it would be great to have a useful app that people can install, that could be packaged up
<jamest> ?
<jamest> lol
<jamest> time.sleep is in seconds not milliseconds
*** jcater has quit IRC
*** jamest has quit IRC