*** kilo has joined #gnuenterprise
*** btami has joined #gnuenterprise
<btami> good morning
<kilo> good morning
*** johannesV has joined #gnuenterprise
*** johannesV has quit IRC
*** johannesV has joined #gnuenterprise
<johannesV> good morning
*** kilo_ has joined #gnuenterprise
*** kilo has quit IRC
*** kilo_ is now known as kilo
*** reinhard has joined #gnuenterprise
<reinhard> hi all
<reinhard> (not saying good moring on purpose as that would sound like I just got up from bed ;-))
<reinhard> johannesV: did you see jamest removed your test exception?
<reinhard> would've been a good example how a comment would help others understanding what one means ;-)
<johannesV> ?
<johannesV> yeah, i've seen it beeing removed ...
<reinhard> he probably didn't understand it was a feature of forms to test exception handling on different ui's
<johannesV> ah, ok ... yeah that's right
*** erkki__ has joined #gnuenterprise
*** erkki has quit IRC
*** btami_ has joined #gnuenterprise
*** kilo_ has joined #gnuenterprise
*** johannesV_ has joined #gnuenterprise
*** johannesV has quit IRC
*** yure has joined #gnuenterprise
*** erkki has joined #gnuenterprise
*** kilo has quit IRC
*** btami has quit IRC
*** kilo_ is now known as kilo
*** erkki__ has quit IRC
*** erkki__ has joined #gnuenterprise
*** erkki has quit IRC
*** erkki__ is now known as erkkie
*** erkkie has quit IRC
*** erkkie has joined #gnuenterprise
<erkkie> I think I got irssi configured for now:p
*** erkkie_ has joined #gnuenterprise
*** erkkie has quit IRC
*** erkkie_ is now known as erkkie
*** reinhard has quit IRC
*** reinhard has joined #gnuenterprise
<erkkie> does the cfield tag in reports datasource conditions support something like name="TEST_classAId.classBId.field" where TEST is the module, classA and class B are the classes of appserver, class a has a reference of classB and field is a property of classB
<erkkie> right now I'm getting a error saying, that TEST_classB has no property field, but the class has that property:p
<erkkie> previously I just used custom calculated properties, like TEST_getFieldOfB, but I was trying to clean up some reports and class definitions
<reinhard> erkkie: wow, you are a power user of appserver ;-)
<reinhard> it would probably be something like
<reinhard> TEST_classAId.TEST_fieldnameofclassA.TEST_fieldnameofclassB
<reinhard> where fieldnameofclassA is the field name that references class b
<erkkie> reinhard: not quite power user yet:p.. just trying to understand & clean something I did in gnue a few months ago:p
<erkkie> oh, so just add the module reference to every class..
<erkkie> will try:p
<reinhard> note in the middle there's not the class name but the field name
<erkkie> ok
<reinhard> hmmm
<reinhard> actually you wouldn't need the classA at all
<reinhard> because that would already be in the datasource definition
<erkkie> yeah, the datasource was, let's say, classX
<erkkie> :p
<erkkie> anw, it worked now, thanks:)
<reinhard> ah ok
<reinhard> two hooks
<reinhard> gotcha
*** Jave27 has joined #gnuenterprise
<reinhard> seriously, I think your project might be one of the biggest ones done with appserver so far....
<reinhard> all we know of so far is what is in gnue-contrig
<reinhard> contrib
<erkkie> hah
<reinhard> would be nice to hear some feedback
<erkkie> well, it's not so big actually.. might send it to you when I clean it up a bit
<reinhard> not to say how happy we would be to see it in gnue-contrib as an example :-)
*** jamest has joined #gnuenterprise
<reinhard> [not wanting to pressure in this direction of course, but would appreicate it a lot!]
<erkkie> haha
<jamest> erkkie: was it you asking for the key strokes to trigger
<erkkie> jamest: yeah
<jamest> i got it sorta working last night but wanted to bounce stuff on reinhard before I commit it
<jamest> reinhard: did you see what erkkie was requesting?
<reinhard> jamest: I noticed you removed a feature from forms ;-)
<jamest> lol, yeah, i loved that feature 2 as it appealed to me for 2 reasons
<erkkie> reinhard: it's a small inventory management system so far... has been working for like 3 months without any major problems.. so now I'm being brave and thinking about adding ltsp thin clients using forms/appserver as POS terminals
<jamest> 1.  it was a deliberately coded exception upon certain into
<jamest> 2. the debug message looked like the meaningless stuff I use for debug messages :)
<jamest> that check has been in our event loop since the ~6200 rev I can't believe it was there that long
<jamest> on erkkie's feature I made it so forms first tries to see if there is a command defined for the keystroke
<jamest> if not and if alt and ctrl is depressed it then builds an event named IIRC requestUSERCOMMAND with a trigger name of
<jamest> [CTRL-][ALT-]<KEY>
<jamest> the GFInstance then tries to run a trigger named "KEY-%s" % triggerName
<jamest> so alt-g would fire a trigger named KEY-ALT-G in the form
<reinhard> jamest: sounds sane
<reinhard> fwiw, the exception *really* was a feature
<reinhard> because it made it possible to test exception handling code for different ui's
<reinhard> :)
<jamest> lol, really?
<reinhard> yes, johannes introduced it to test the new error message windows
<jamest> sorry it seemed to be an unneeded check in the keypress event loop for a keycode that I couldn't generate
<reinhard> ctrl-e (e as in exception) :-)
<reinhard> but I guess we can leave it out
<jamest> didn't work
<reinhard> was meant temporary anyway
<jamest> as ord('e') here is 101
*** kilo has quit IRC
<jamest> but ctrl-e generated a keycode of 5 when the ctrl key is pressed
<reinhard> nevermind, better to leave it out anyway
<jamest> :)
<reinhard> users would look strange if this error would be generated...
<reinhard> :)
<jamest> in my tinkering I put the trigger code outside of GFKeyMapper but I really think that's the best place for it
<jamest> to build a command string out of any unknown command if the alt or ctrl are depressed otherwise return none
<jamest> but I'm so rusty I wanted to verify with you guys that would be ok
<jamest> or that I wasn't overlooking the feature added somewhere else
<erkkie> jamest: sound cool.. will surely try it out once you commit it
<erkkie> reinhard: I suppose I could make a "fork" of my app and add it to gnue-contrib
<reinhard> jamest: we haven't done that much in forms lately, it's not like we had completely gone through the code (like we did in common)
<reinhard> so just do as you think it would be best, if it breaks something we'll look into fixing together I suppose
<reinhard> erkkie: why a fork?
<jamest> reinhard: ok, cool
<erkkie> I might change some specific things that might not make sense in a example app
<jamest> reinhard: did you guys ever decide about the wx 2.6(?) ui driver?
<reinhard> erkkie: if you want, you can do that, but you wouldn't have to
<reinhard> erkkie: we have very specific things in gnue-contrib, it's not that it were official samples
<reinhard> just more like "look what i've done with gnue"
<reinhard> jamest: yes, we did. there will be a  -u wx26
<reinhard> and we will see if we can build it in a way that can replace -u win32 and -u gtk2
<johannesV_> wx26-driver is about to be finished this week
<reinhard> erm
<reinhard> s/we/johannes/ :-)
<jamest> ah, it's just not in svn
<jamest> ok
<reinhard> johannes likes big commits :)
<erkkie> reinhard: oh, ok, will think about it.. I'm sure I'll add something into contrib
<jamest> does the checkbox in 2.6 not suck?
* jamest hopes
<reinhard> hehe
<reinhard> IIRC the checkbox in 2.4 was more an issue of being gtk1 instead of gtk2
<johannesV_> jamest, can you be more precise about the "sucking" thing with 2.4-checkboxes ?
<reinhard> you had to be outstandingly smart to recognize whether it is checked or not
<jamest> in 2.4 a checkbox is a raised or lowered box
<reinhard> it looked like a button pressed in or out
<reinhard> yes
<jamest> i never had an issue with it (must be what reinhard said :)
<jamest> but my users HATE it
<johannesV_> ah, ok ... so the answer is: yes it won't suck anymore
<reinhard> jamest: roflmao
<jamest> one of the biggest complaints I get in fact
<reinhard> jamest: we are planning to do bad bad things to forms RSN
<reinhard> like vbox/hbox-oriented positioning as an alternative to character based
<reinhard> implementing a real grid that looks like a grid and behaves like a grid
<reinhard> forms looking still good if they are resized (up to having more lines in a grid depending on available space)
<jamest> how?
<jamest> you adding a layout mgr that translates to positions?
<reinhard> and all of that most probably resulting in more or less compatible changes in gfd
<jamest> or something else?
<reinhard> we plan to use the layout manager where it already exists
<reinhard> like in wx26
<reinhard> and do something as-smart-as-possible in non-layout-mgr ui's like curses
<reinhard> but wx would give us layout management even for win32 (which is per se pixel based)
<jamest> cool
<reinhard> but that will mean changes in designer :(
<reinhard> and that's our main headache
<johannesV_> yeah ... that's really odd ...
<reinhard> and of course it's always s/we/johannes/
<reinhard> especially when it comes to the actual work ;-)
<jamest> sorry guys
<reinhard> uh, that last remark was targeted at *me*, not at anybody else
<reinhard> I always show of what "we" are doing, and in fact johannes does the work
<reinhard> not meaning to push you or anything
<jamest> :)
<jamest> the only reason I worked on forms last night was because I refused to put in another 12 hour day for my job
* jamest was being defiant
<jamest> :)
<jamest> ooooo, maybe I shouldn't have said that in here
<jamest> it's ok johannesV_ 12 hour days rock!
<johannesV_> :)
<reinhard> lol
*** btami_ has quit IRC
*** jcater has left #gnuenterprise
*** ajmitch_ has joined #gnuenterprise
*** ajmitch has quit IRC
<chillywilly> oooh, there's activity here ;)
<jamest> :)
<jamest> reinhard: do you know of any way to force the postgres drivers to not hold the transactions open ?
<jamest> my users keep forms open all the time on their desktops
<reinhard> I am not sure if it is possible
<jamest> and I can't even create new tables using REFERENCES table(column) as those transactions block the table creation
<reinhard> I think you have to close the connection
<reinhard> so forms would have to log out from the connection and log in again
<reinhard> all the time
<jamest> how difficult would it be to have gnue connections time out
<jamest> as this sucks here
<reinhard> which doesn't sound like a good idea in general
<reinhard> hmmm
<jamest> as I have about 25 people with various forms open
<reinhard> I don't know how this timeout would be done
<jamest> and it's easier to wait until after hours than to track down who has what open and then just know "oh form A" needs closed
<jamest> we store the username/passwd right?
<reinhard> yes
<jamest> and we keep the connection open so we're not building and tearing down every query
<reinhard> right
<reinhard> currently
<jamest> let me think out loud
<jamest> what about a timer thread set to wake up every user specified seconds
<jamest> and if the connection says I haven't been used since the last time you checked
<jamest> then the connection to the db is closed but the gnue connection stays open
<reinhard> yes
<jamest> the next time it's called it says I need to reopen the db connection before I do anything
<reinhard> maybe just reset the timer at each activity
*** SachaS has quit IRC
<jamest> i still don't grok why an open form is blocking on the tables it references though
<jamest> as connections are pooled aren't they?
<reinhard> erm
<jamest> or is that only the connection mgr and each form gets it's own connection
<reinhard> yes, per instance of gnue-forms running
<reinhard> different independent processes can't pool connections :)
<reinhard> [except you would use appserver] - scnr :)
<jamest> what about when ran thru navigator?
<jamest> i seem to recall navigator providing the connection at that point
<reinhard> one pool per navigator instance
<reinhard> I would say
<reinhard> but if 2 users have forms open, you certainly have 2 connections
<jamest> all my users launch from navigator
<jamest> and keep a handful of forms open they use throughout the day
<jamest> but easily idle them for 10 to 15 minutes at a time then do 2 to 20 minutes of work in them
<jamest> so if I set a 2 or 3  minute timeout on holding the db connection open i think i'd be golden
<jamest> actually this might help me with another issue i've started seeing
*** ajmitch has joined #gnuenterprise
<jamest> where I deadlock
*** jcater has joined #gnuenterprise
<jamest> i've not found the app that is causing it yet as typically when it happens the sales force needs to db up *NOW*
*** klasstek has joined #gnuenterprise
<jamest> reinhard: would something like this work in the new datasources?
<reinhard> brb phone
*** btami has joined #gnuenterprise
*** CirrusImage has joined #gnuenterprise
*** derek has quit IRC
*** ajmitch_ has quit IRC
*** derek has joined #gnuenterprise
*** derek has quit IRC
*** derek has joined #gnuenterprise
*** docelic has joined #gnuenterprise
<docelic> Howdy folks
<derek> sup
<reinhard> jamest: yeah, I think that would be ok
<reinhard> based on a parameter like "autologout=(minutes)"
<btami> hi docelic,long time no see
<reinhard> would be deep in Connection.py
<reinhard> and thus pretty generic (usable for all backends)
<reinhard> must run
<reinhard> cu all
*** reinhard has quit IRC
<docelic> btami, hey yes :) I was evaluating Bond (www.treshna.com) to use instead of GNUe, but I think I gave up on it. There are no usable mailing lists, code repository or documentation and at least the main author can't spell a sentence without at least two mistakes..
<docelic> Even though it has some nice ideas. For example designing GUI directly in glade...
<docelic> Any of you seen Bond and have comments about it ?
<btami> not me
<btami> arn't thresne the firm where kiwis was here befor ?
<btami> i mean treshna
<jamest> not me
<btami> http://www.gnuenterprise.org/community/sponsors.php
<docelic> Hm.. interesting..
<docelic> You also credit Andru Hill , who is treshna lead guy..
<derek> he was one of original developers of appserver
<derek> before it was rewritten in python
*** johannesV_ has quit IRC
*** johannesV_ has joined #gnuenterprise
*** sjc has joined #gnuenterprise
*** btami has quit IRC
*** johannesV_ has quit IRC
*** dimas has quit IRC
*** sjc has quit IRC
*** kilo has joined #gnuenterprise
*** yure has quit IRC
*** klasstek has quit IRC
*** Jave27 has quit IRC
*** jamest has quit IRC
*** yure has joined #gnuenterprise
*** derek has quit IRC
*** derek has joined #gnuenterprise
*** yure has quit IRC
*** jamest has joined #gnuenterprise
*** derek has quit IRC
*** kilo has quit IRC
*** jcater has quit IRC
*** derek has joined #gnuenterprise
*** dimas has joined #gnuenterprise
*** jamest has left #gnuenterprise
*** jcater has joined #gnuenterprise
*** jcater has left #gnuenterprise
*** derek has quit IRC
*** derek has joined #gnuenterprise