*** Morphous has joined #gnuenterprise
*** Amorphous has quit IRC
*** johannesV has joined #gnuenterprise
*** sjc has joined #gnuenterprise
*** reinhard has joined #gnuenterprise
<johannesV> reinhard, great job you did on dropdowns/fk-stuff
<johannesV> seems to be working quite well ...
<johannesV> but i had a situation where it produced a dirty form (from the beginning) which i cannot get clean
<johannesV> so ctrl-c was my friend
<reinhard> johannesV: I know of some remaining problems
<johannesV> btw. i'll change form-generation to reflect recent changes then
<johannesV> if a prop is a 'info' property: do we set it as 'usual' fk-field (like a dropdown) but with 'label'-style, ...
<johannesV> ... or should we add a <block> for the dropdown-source including the info-field ?
<johannesV> (the latter method produces that weird situation)
<johannesV> the first one seems to work
<johannesV> (i would tend to the first solution anyway, but i'm curious why the second isn't working as well)
<reinhard> I would go the latter way actually
<reinhard> have never even considered the first way
<reinhard> that weird behaviour will be fixed ASAP (expect a fix today)
<johannesV> ok
<reinhard> I think the latter is less overhead
<johannesV> but there should nothing wrong with the first one, since it uses the same fk-source, but just another fk-description
<johannesV> right
<reinhard> because it's a "normal" block
<johannesV> that's the difference
<johannesV> using fk- will produce th _allowedValues-dictionaries twice
<johannesV> (at least)
*** kilo has joined #gnuenterprise
<kilo> good morning
<reinhard> good morning kilo
<kilo> hi reinhard
<reinhard> kilo: we are *not* trying to tease you, btw
<reinhard> ;-)
<kilo> i don't believe that 8-))
*** dimas has quit IRC
*** dimas has joined #gnuenterprise
<johannesV> F****
<johannesV> i cannot find a signal emitted by a gtk.Combo when the drop-down is closed, neither when it's opened
<reinhard> can you somehow monitor what signals are emitted at all?
<kilo> johannesV: you will not get that signal. i've tried that also...
<kilo> reinhard: re teasing me
<kilo> DENNIS:
<kilo>     Oh, King, eh, very nice. And how d'you get that, eh? By exploiting the workers! By 'anging on to outdated imperialist dogma which perpetuates the economic and social differences in our society. If there's ever going to be any progress with the...
<reinhard> that's from the holy grail, isn't it?
<kilo> ayer
<johannesV> kilo, i've now installed a 'catch and print all signals' ...
<johannesV> and it seems i get something ...
<johannesV> will continue hacking ...
<kilo> iirc it did not work like the docs said it should
<johannesV> which docs ... *lol* they tell nothing about such a signal
*** btami has joined #gnuenterprise
<btami> hi
*** holycow has joined #gnuenterprise
<kilo> no, i remember reading something about combo signals
<johannesV> damn combos
<johannesV> there seems to be really no signal emitted on popup/close of the dropdown
<johannesV> ahh...
<johannesV> i get a map-signal when it opens ...
<kilo> hmm
<holycow> mornin guys
<johannesV> ok, map and unmap aren't the right thing ...
<johannesV> cause they're only called once ..
<kilo> reinhard: your latest changes are great
<reinhard> unfortunately they do not yet work
<kilo> in win32 dropdowns are refreshed every time
<kilo> it is gtk that does not work quite prfect
<kilo> ie
<reinhard> kilo: there are other problems remaining as well
<reinhard> 1. it does not work if there is a block assigned to the fk datasource
<reinhard> 2. updating other info on the screen doesn't work when you select the empty line in the dropdown
<kilo> well btami has extensively tested your latest changes in gnue-invoice, in windows
<btami> yep
<reinhard> and?
<btami> reinhard: if you have a test gfd, please let share, and i can test it on win32 too
<kilo> and win32 ui works great. if you edit a dropdown's choice list in a subform, when closing the subform the choises are refreshed automatically in the dropdown
<kilo> in gtk2 it is only true if you have selected a value from the dropdown and called the editor subform that way
<kilo> maybe there are errors (you probably know that better), but we were happy to see these changes
<reinhard> what i'm actually targeting at is things like in appserver's sample.gfd
<reinhard> two dropdowns for the same key (country code and country name)
<reinhard> when you select in one dropdown, the other dropdown updates magically
<kilo> morning holycow
<btami> have to go for kids, bbl
*** btami has quit IRC
*** cilkay has joined #gnuenterprise
<reinhard> johannesV: I think we have to go the non-block way for info fields
<johannesV> ok
<johannesV> anyway, i cannot find a way to solve the combo-problem
<johannesV> i don't understand why they have left such a signal
<johannesV> even simulating such a signal is quite hard to do
<reinhard> any other signal that could be used?
<reinhard> something that is fired more often but still somehow sensible?
* reinhard seriously considers adding a field.setValue to GFDisplayHandler.addText
<kilo> will it not fire field-changed triggers with every character you type?
<reinhard> it will
<reinhard> erm
<reinhard> it should
<reinhard> but appearantly it doesn't...
<kilo> and is that what you want?
<kilo> doesnt a field change only when you move away from the entry?
<reinhard> no
<reinhard> but I think it's closest to what I want
<reinhard> currently yes
<reinhard> sigh
<kilo> let me guess: you s*ck with dropdowns...
<reinhard> must get something to eat
<reinhard> bbl
*** jcater has joined #gnuenterprise
<reinhard> jcater: any chance you can have a look at my latest changes regarding dropdowns?
<jcater> I'll try
<jcater> but today is looking bad for me
<jcater> (I am in at 5:00am :)
<jcater> I'm looking at commit logs now
<jcater> so, if I understand this right
<jcater> you've basically made any fk datasources that are bound to a dropdown
<jcater> to force the datasource's current record to match the dropdown's current selection
<jcater> ?
<jcater> it looks non-intrusive to me (i.e., I don't see how this can break anything -- it only adds new capabilities)
<reinhard> jcater: exactly
<reinhard> that will enable forms to have several dropdowns or entries
<reinhard> pointing to the same fk datasource
<reinhard> and always be in sync with each other
<jcater> cool
<reinhard> for example you have a country code dropdown and a country name dropdown
<reinhard> both have the same fk datasource
<chillywilly> bah...
<reinhard> you're all up early today?
<reinhard> :)
<chillywilly> I was up an hour ago
<chillywilly> finsihed scotty's boat for sub scouts
<chillywilly> cub*
<reinhard> hmmm
<reinhard> hmmmm
<reinhard> what would others think when every single keystroke in an entry would generated pre-change and post-change events?
<reinhard> a) that's ok
<reinhard> b) acceptable for dropdowns, but not for other entries
<reinhard> c) no way, are you crazy?
<jcater> actually
<jcater> expanding on our naming convention
<jcater> that would be a 3rd trigger
<jcater> on-change
<jcater> I think
<kilo> i agree with jcater
<holycow> reinhard, what are the consequences of the option?
<reinhard> the consequences are
<reinhard> currently pre-change and post-change are more or less fired at the same time as pre-focusout and post-focusout
<reinhard> it would be fired much more often then
<reinhard> jcater: when would the on-change be fired?
<reinhard> I mean
<holycow> ah, i would personally go with b, or perhaps be able to define it on a per field basis as an attribute?
<reinhard> I can either call setValue on the field, or I can not call it
<holycow> if you had a large text box for inputing notes, that wouldn't be appropriate i guess
<holycow> but
<holycow> if you had a contact field, you could use the input for lookahead features say
<reinhard> holycow: that's about the idea
<reinhard> a large text box would probably not define an on-change trigger
<holycow> if its an attribute one can turn on, on a per field basis, i say wonnerfull
<holycow> ah okay
<reinhard> holycow: i think you are actaully right about the attribute
<reinhard> hmmm
<reinhard> jcater: you still here?
<reinhard> (sorry to bug you)
<jcater> kinda
*** jamest has joined #gnuenterprise
*** jamest has quit IRC
*** johannesV_ has joined #gnuenterprise
*** holycow has quit IRC
*** johannesV has quit IRC
*** johannesV_ has quit IRC
*** johannesV has joined #gnuenterprise
*** jamest has joined #gnuenterprise
<reinhard> johannesV: dropdowns are ready for further beating^H^H^H^H^H^H^H^Htesting
<johannesV> great
<johannesV> but the 'empty'-record problem in the block isn't solved, is it ?
<reinhard> no
<reinhard> you can not use a block on the same datasource than a dropdown
<reinhard> just about to document that in BUGS
<johannesV> ok
*** kilo has quit IRC
<johannesV> reinhard: i think i've got another one ...
<johannesV> use sample.gfd, start and exec query (f8,f9)
<johannesV> change the country-code by typing AT (for example)
<johannesV> then change the record (without leaving the entry, using the cursor-keys)
<johannesV> you're still in country-code
<johannesV> now type x
<johannesV> and it changes to RUx (if the next records country code is RU)
<reinhard> that's an old one
<reinhard> i remember that from months ago
*** btami has joined #gnuenterprise
<reinhard> that's a GF-cursorpos vs. real-cursorpos issue
<johannesV> hmm, but it seems as if there is an 'old' value flying around
<johannesV> after pressing a new key the old value comes up again
<johannesV> (old == from last record)
<reinhard> ?
<reinhard> you said it changes to RUx not ATx
<johannesV> right, but RU is the code from the other record
<johannesV> after moving to the second record and the moving back to the first one
<johannesV> ah
<johannesV> now i see, i haven't explained properly
<johannesV> after moving back to the first i see AT, but after pressing x i see RUx
*** btami has quit IRC
<johannesV> reinhard: dropdown-problem should be solved now
<reinhard> johannesV: cool, thanks
<reinhard> I still can't believe we have *really* working dropdowns finally
<johannesV> seem so :)
<johannesV> but i think we're on the right way ...
<johannesV> so the next big point would be that 'search'-stuff
<reinhard> did you already fix that "committing empty master" problem?
<johannesV> yes, an empty master get's commited automatically
<johannesV> if a detail is pending
<johannesV> i thought to remove that pending bug from common regarding gnue-schema
<johannesV> (but i've to rework gnue-schema a bit)
<johannesV> doing this i've planned to integrate dependency-checking too (so the order of definitions no longer matters)
<jamest> working dropdowns?
<johannesV> like i did on gcd's
<johannesV> jamest: sounds unbelievable, doesn't it ? :))
<jamest> i don't believe it
<jamest> i can't believe it
<jamest> i WON'T believe it
<jamest> something has to still be broken
<johannesV> jamest: please try to find something, and then - if you're lucky to do so - tell us :))
<jamest> let my svn up
<jamest> as i had one i hit this week
<jamest> lets see if it's still there :)
<johannesV> ok, thanks
<jamest> definately works differently
<jamest> as this form used post-change to apply a query filter
<jamest> now the query is immediate instead of upon exit
<jamest> which i think is good
<reinhard> jamest: it is much more "responsive" against the user
<jamest> the only bad thing I could think of would be down arrowing thru the filters
<reinhard> as long as you have the dropown in dropped condition
<reinhard> it shouldn't issue a post-change for cursor up/down
<jcater> I wish you guys had an oracle install :)
<jcater> as my oracle is still fscked
<reinhard> wrt dropdowns?
<reinhard> or wrt something completely different?
<jcater> completely different
<reinhard> ah
<reinhard> still the same old problem then?
<jcater> yeah
<jcater> ui corruption
<reinhard> sorry to hear that
<reinhard> sounds really weird
<jamest> i was talking down arrow from a closed dropdown
<jamest> fwiw i'm not saying that's an issue
<reinhard> jamest: ah yes that does browse through the allowed values for wx
<jamest> just generates a lot trigger firing if the use navigates like that
<jamest> but i like the behaviour better now
<jamest> and users will learn quickly not to do that in the form :)
<reinhard> i think so, too
<reinhard> fwiw
<reinhard> the on-change is fired on every keypress *only* in a dropdown
<jamest> so yes, the issue I had was fixed
<reinhard> not in a normal entry
<reinhard> to be precise
<jamest> i still don't believe the dropdowns work
<reinhard> it is fired on every keypress if the field has a fk source
<reinhard> so the fk source can sync with the selection
<jamest> that would go against the laws of the natural order
<reinhard> lol
<reinhard> to keep you satisified
<reinhard> 2 issues still remain
<reinhard> 1. you can't use a fk datasource as a datasource for a block at the same time
<reinhard> (but you can use the same fk datasource for several fields)
<reinhard> 2. when the fk_description for one of the allowed values is empty, the text "None" is displayed in the dropdown :)
<johannesV> now you won't use a forms-trigger to keep other fields of the fk-source in sync
<johannesV> and this change is really a *good* change since it makes forms having less code
<jamest> i'm all for less code
<johannesV> bbl
*** johannesV has quit IRC
*** johannesV has joined #gnuenterprise
*** holycow has joined #gnuenterprise
<reinhard> off for today
<reinhard> l8r all
*** reinhard has quit IRC
*** wendall911 has joined #gnuenterprise
*** wendall911 has quit IRC
*** johannesV has quit IRC
*** havoc has quit IRC
*** johannesV has joined #gnuenterprise
*** havoc has joined #gnuenterprise
*** wendall911 has joined #gnuenterprise
*** kilo has joined #gnuenterprise
*** sjc has quit IRC
<kilo> @server
<bigbrother> kilo: tolkien.freenode.net
<kilo> jcater: i'm playing with Supybot at home. do you need to set the irc server name? i can't make it to log in
*** jamest has quit IRC
*** kilo has quit IRC
*** wendall911 has quit IRC
*** jcater has quit IRC
*** jcater has joined #gnuenterprise
*** ajmitch_ has joined #gnuenterprise
<jbailey> Andrew!
<jbailey> You need to save me from Reinhard!
<ajmitch_> me?
<ajmitch_> why me?
<jbailey> Because you are the keymaster.
<jbailey> The one holding the source.
* ajmitch_ only just got networking going here :)
<jbailey> The new Debian packages!
<ajmitch_> oh
<jbailey> And perhaps if I upload them, he'll shut up about them for a day or two. ;)
<ajmitch_> ah
<ajmitch_> good luck with that ;)
<jbailey> Well, I was thinking the day or two would be the weekend when I'm not usually online. =)
<ajmitch_> by the way, NX seems quite fast, even to home :)
<jbailey> Are you used to a dialup line that people on the street regularily use for a latrine?
<jbailey> IP by Avian carrier might feel fast to you. =)
<ajmitch_> heh
<ajmitch_> using cable here at the moment
*** jamest has joined #gnuenterprise
<ajmitch_> jamest!
<jamest> sorta
<ajmitch_> ah
<jamest> more like jamest!  lurking while cleaning house
<ajmitch_> you poor chap
<jbailey> He's always doing chores.
<ajmitch_> it's an excuse not to fix bugs :)
*** holycow has quit IRC
*** jamest has quit IRC
*** jcater has quit IRC
*** ajmitch_ has quit IRC
*** ajmitch_ has joined #gnuenterprise
*** dcmwai has joined #gnuenterprise