*** jcater has quit IRC hmmm, a new Gnome release and its in debaiun experimental debian @weather 85233 dneighbo: The current temperature in Gilbert, Arizona is 66°F. Conditions: Overcast. brrrrr *** Morphous has joined #gnuenterprise *** reinhard has joined #gnuenterprise *** johannesV has joined #gnuenterprise *** _florin_asleep has joined #gnuenterprise 66F... That's not that bad is it? Cater has it worse... *** Amorphous has quit IRC *** chillywilly has quit IRC *** sjc has joined #gnuenterprise *** chillywilly has joined #gnuenterprise *** SachaS has joined #gnuenterprise *** someon has quit IRC *** kilo_ has joined #gnuenterprise *** dimas has joined #gnuenterprise @weather LHUD kilo_: The current temperature in Szeged, Hungary is 7°C. Conditions: Fair. * ajmitch tries @weather NZDN ajmitch: Error: The format of the page was odd. * ajmitch mutters i thought Dunedin airport was quite a big one... *** lekma has joined #gnuenterprise *** btami has joined #gnuenterprise kilo_: hardly big *** sjc has quit IRC hello everybody hi lekma ajmitch: LHUD has grass runways. NZDN is ICAO cat 4 that is it can receive B767 or something like that... *** dcmwai has quit IRC hmm yeah, that's true.. it's classified as 'international' in a sense that is an other story. international means it has an immigration office so international passangers can be received yes but when NZ is 3000km from any other country, it generally means that it can take larger planes :) well, who chose that location for NZ? 8-)) any forms specialist around?? <_florin_> good morning everyone :-) i would like to know if one can set form tag attribute from an 'ON-STARTUP' trigger morning _florin_ i.e. from the same form 'ON-STARTUP' trigger lekma: i dont know but I suggest you just try it ;) : :) *** Egor has joined #gnuenterprise hi Egor Hi. I need a help. How can I Install ERP packages? first you have to write them :) btami: lol Egor: actually, btami is right as we stand now, gnue is a set of tools designed to write ERP systems base, finance also? but there is not "turnkey" erp package yes or to put it more nicely "they all are in planning state" :) Can I use something from gnue-packages as templates? i don't think so there was not much work on packages over the last months the reason is that we always intended to build the packages upon gnue-appserver and gnue-appserver was not quite ready for it until very recently it's just a week or two that people have started to do productive applications with appserver lekma: i don't think you can do that but then reinhard got some inspiration... and BANG appserver is alive How can I use .geas files? .geas? from packagename/defaults folders. It looks like sql files. btami: i was looking in forms source code how i could have access to attributes i didn't find any hooks hmmm, i think there should not be hooks Egor: what we have now in gnue-packages is quite obsolete read the wiki, there is an excellent explanation from james on triggers and forms it's basically built upon geas, which was the predessor of gnue-appserver kilo_: i will due to technical problems we had with geas we ditched it and made a rewrite (which is now gnue-appserver) which pushed us quite back in time but forward in quality :) bbl Thanks reinhard and all. Bye *** Egor has left #gnuenterprise *** PeterD has joined #gnuenterprise Hi all hi PeterD hi btami I woul like to talk about appserve, in particular classrepository.gfd reinhard (the man to talk to) has just gone to lunch ok, I will hang around maybe johannesV is here... hi PeterD dou you need advice on usage? no, serch querry failling on classes page and there are some errors in log speak of the devil... ;-) hi reinhard PeterD: can you paste the errors? hi btw :) yes, but big gnue gnue -1 -1 00000000000000000000000000000000 00000000000000000000000000000000 -1 -1 DB000: Traceback (most recent call last): *** PeterD has quit IRC doh *** PeterD has joined #gnuenterprise *** PeterD has quit IRC *** PeterD has joined #gnuenterprise PeterD: i think you better send mail :) reinhard@gnue.org I will do it my client crashed bbl in about 1 h *** btami has quit IRC <_florin_> is there in gnue-forms any possibility to show/hide entries at runtime? *** kilo_ has quit IRC *** kilo_ has joined #gnuenterprise * johannesV is away: outdoor-working _florin_ there is a hide option for fields (or entries), have you tried that? <_florin_> SachaS: yes, i've tried, but it shows very strange behavior <_florin_> 1st: it doesn't work :-) <_florin_> 2nd: it breaks jumpRecord()-method of GFBlock <_florin_> (that means, if it use jumpRecord() the effects are not visible: still have the old string in the entry widget) <_florin_> gotta run now, have to do some shopping! johannesV: once your back from outdoor work, can you tell me what "Added create/extend support" means for a appserver developper, please? *** kilo_ has quit IRC PeterD: are you here? PeterD: trying to reproduce your problem, but no success so far hi reinhard can you tell me exactly what you are doing? at the moment trying to serch classes properties for module gnue or alternatively i can tell you what i am doing (and it works for me) ok ok i open form then i change to "classes" tab then i press the magifyer icon magnifier then i press cursor down key 2 times to select gnue then i press search icon then i get the "module" class when I press search icon I get nothing have to click cursor on name field then get "module" next have to click on Comment field to see description of module I can't get any properties I'm using snapshot dated 20040330 which frontend? linux or windows? linux mandrake10 when you search for "empty" query i.e. load all records on class page does it work then? doesn't did the sample form work for you? serch wildcat "*" and nothing you should be able to search without entering anything just press the query button, and then the search button immediately afterwards nothing , last few lines of log DB000: return self._adapter._runMethod (self._methodname, *args, **params) DB000: File "/usr/local/gnue/lib/python/gnue/common/rpc/drivers/xmlrpc/py_xmlrpc/ClientAdapter.py", line 98, in _runMethod DB000: result = self._client.execute (method, __args, self._timeout) DB000: File "/usr/lib/python2.3/site-packages/xmlrpc.py", line 217, in execute DB000: return self._o.execute(method, params, timeout, name, passw) DB000: SystemError: error return without exception set ok do i understand correctly that you never got any record on a form with appserver? I make some entry an saved and search for known name, it display it did you ever test forms with 2-tier? in the gnue-forms samples directory there's a zipcode subdir *** lekma has quit IRC i didn't that contains a sql script for postgres and two sample forms that run *directy* against the database without appserver i think checking if that works would help greatly to find out ok I will try whether we are facing an appserver problem or a forms problem you use postgres? yes ok thanks for you patience :) <_florin_> hya! reinhard: doesn't looks good, tried zipcode.gfd search only working for known name otherwise not PeterD: thanks for testing PeterD: actually, that *does* look good PeterD: because it means it's a forms problem which looks good for me as i am an appserver developer ;-) j/k PeterD: you're in australia, correct? the forms masters will be here in about an hour i guess jamest even planned to set up a multiboot machine with several distros to test not sure if he already tried mandrake, but might be yes I'm in australia what time is it there now? * reinhard guesses it's rather late as I remember search work some time ago, so somthing maybe is broken in forms is 11:17 pm maybe it broke with the update from mdk 9 to 10? maybe I rebuilt mandrake 10 obout week ago i'm quite sure jamest will be able to find out what this is reinhard: if have a property of type date and i have to create a date in python. do I do a) mx.DateTime.now() or b) mx.DateTime.now().date ? just now() properties of type "date" hold mx.DateTime values ok. thanks. I will try to hang around for while, but probably my wife will kick me out soon reinhard: did johannesV patches break any old installations ? ;) they shouldn't why? you having any problems? *** jcater has joined #gnuenterprise hey jcater good to have you back!! ;-) no problems. just curious are there any changes yet? does the "create/extend support" a) work, b) how is it used? SachaS: in a nutshell consider somebody installed your babyerp package and works with it for, say, a year now i come and write a new module based on it a year are you kidding? j/k for quality assurance module name "quality" ok. this module adds some properties to the inventory class like checkPeriod lastCheck to define periodical quality checks for the product ok now your user finds out about my module and wants to install it so he basically nees an sql script that does alter table masterdata_inventory add quality_checkPeriod datetime alter table masterdata_inventory add quality_lastCheck datetime err forgot the ; but you get what i mean :) just a sec yes. got it. are the properties allowed to be "NOT NULL", does in that case the script set the previous record values to "None" or something? no NOT NULL is not allowed at least not no now so what we actually did s/we/johannes/ is i could now do gnue-gsdgen quality ok. and it generates a gsd file where only the stuff of the quality module are contained and in the gsd file there was a tag up to now err reinhard: it's not as bad as we thouth, the problem is only with classrepository.gfd there is a tag now you can have

ok. if in the gsd you have create gnue-schema will do CREATE TABLE otherwise ALTER TABLE ADD PeterD: search does work for other forms? ok. nice. reinhar:I relised that zipcode didn't contain any data, so made some entries and work OK PeterD: ok PeterD: how about appserver sample.gfd? reinhard: checked with appserver gnue-classes.gfd and other forms and search worked ah oh gnue-classes.gfd worked? and classrepository.gfd doesn't? gnue-classes.gfd worked but sample.gfd and clasrepository.gfd do not work weird could be something broken in this two forms only? there are new forms? these forms have not been changed for more than a week and several people have worked with them meanwhile of course including myself i will try to reproduce *** jamest has joined #gnuenterprise good morning jamest hello who implemented the language adaptors code? originally i think it was siesel and johannesV reworked a bit of it any problems? well, we used to be able to do raise TriggerError("Message") reinhard:log of appserver when hit "magnified" icon Dispatching: Session.request [5, u'address_country', [[u'eq', None], [u'const', 1], [u'const', 2]], [], [u'address_code', u'address_name']] Dispatching: Session.fetch [5, 35, 0, 5] Dispatching: Session.rollback [5] but trigger error was removed from the trigger namespace jamest: I'll check i'd tried adding it back in on the python adaptor but keep hitting RuntimeError exceptions wanted to find out if something was preventing that from working now reinhard: log when hit search icon Dispatching: Session.request [5, u'address_person', [], [], [u'address_city', u'address_weight', u'address_street', u'address_meettime', u'address_children', u'address_name', u'address_lastmeeting', u'address_human', u'address_country', u'address_born', u'address_zip']] Dispatching: Session.fetch [5, 36, 0, 5] PeterD: these logs are 100% ok hi PeterD but it not serching data? hi jamest it really should be but it's very hard for me to debug when i can't reproduce the behaviour I understand <_florin_> bye! *** _florin_ has quit IRC by the way the "module" page of classrepository.gfd search as should, but "classes" page do not work and cursor arrow chang to clepsydre or something like that which means that something is procesed in the background or hung up? PeterD: sorry, didn't understand the last sentence jamest: the aim of TriggerError is to raise a silent exception? mouse arrow changing shape, so looks like something hung-up in the system when search is executed reinhard: no it's never been fully implemented :( well, thats poor wording hang on trigger errors were to allow the trigger to stop an action inside the trigger all the examples in GFForm.py though seem wack instead look at GFInstance.py def executeExit(self, event): I have to go, try to be back tomorow, g'night night PeterD which allows a pre-exit trigger to abort the form exit *** PeterD has quit IRC ah i plan to dig into this again today as time permits after getting the runtime error last night I figured I'd ask if there was a reason triggererror was removed as it was only one line to add the functionality in the old code :) but now it is wack all exceptions that occur in the trigger code raise a RuntimeError all we would need would be did you find logic that causes that? i think before it was all exceptions except TriggerError raise a RuntimeError we missed that exception :( where is that code located? i can fix it if we can define how it should behave sigh i see it now in the python execute yes that's sad as that function is where I was working last night :) ? one of the places I tried to add triggerexception logic was in that execute and I spent about 20 minutes trying to figure out why i kept getting runtime instead of trigger when the code was on the screen in front of me :) oh was it quite late then? :) actual, um, no wife kicked me off 'puter at 10 IIRC right now I think trigger error behaviour is pretty much broken atm wrt trigger code or wrt forms code? as if you look thru the code forms ah ok i'd very much like a more general approach and if it was only for the name it was only used in GFForm.py and there rather badly it would seem because we use the logic/ part in common for much more than triggers i'd like to have that exception named differently actually, i take that back too like "Abort" forms uses it ok, i must have been on some wack shit last night it would maybe be even more intuitive for a form author I really don't care what it's called if he could write Abort works for me abort (message) but what I'm wondering about is all the places processTrigger is called that doesn't catch that error (so i guess forms is still broken) jamest: so you'll fix it? (it sounded like you have it nearly done) yes ok i can change to abort while Im in there just a final hint - the exception should better be derived from gException so it can handle non-ascii messages, too thanks class TriggerError(StandardError): pass class TriggerAbort: pass class TriggerStop: pass class TriggerSuccess: pass i honestly don't thing any but the first was used i would even go so far to say i honestly think any but the first make sense :) 'cause they are not even derived from Exception :) :) reinhard: I have a problem, when running the postgresql script ( psql babyerpP < PostgreSQL.sql ) SachaS: can you give me details? :) eeewww dropped the database :| well yeah it creates tables, call alter and truncates too long names. that seems all ok. then we go: ERROR: insert or update on table "gnue_property" violates foreign key constraint "gnue_property_gnue_module_fk" DETAIL: Key (gnue_module)=(16439304979149249717617966817902) is not present in table "gnue_module". ERROR: insert or update on table "gnue_property" violates foreign key constraint "gnue_property_gnue_module_fk" DETAIL: Key (gnue_module)=(16439304979149249717617966817902) is not present in table "gnue_module". * SachaS thinks the gnue_ids got shaked too much ;) ok. i am looking closer at the .gsd file SachaS: is it a gsd you wroty by hand? wrote nope, the one i got via: ggcvs --database babyerpP --export both -o babyerp-generated.gsd did you generate today? after update of svn? yes can you mail it to me? *** lxf has joined #gnuenterprise probably has to do with johannes' changes (or dcc) yep. most likely ;) dcc let me have a look at the file first, maybe i can find something ... the question is are the data rows of the gnue_module table *before* or *after* the gnue_property table i think there are no gnue_module data .. :) hold on send it to you doh wait ok. yeah there is no gnue_module_dump in the .gsd generated file. luckily I have an older backup of the gnue_module_dump data SachaS: i think you can wait for the fix will be some minutes, but not more (i hope) or i'll go to beat johannes back to his keyboard ;-) j/k :) * SachaS changes script to make a backup copy of previous .gsd files fixed * reinhard is still fascinated how quick he can understand code that johannes wrote and he's never seen before :) * SachaS is still fascinated how fast reinhard can fix code :) oooo

) is sorted alphabetically yeah that's not the problem the table definitions can be in any order they want the problem is that part *** lxf has quit IRC ok. then I have () starting with module_dump, followed by class_dump so order seems ok in .gsd file. ok * SachaS lost also some classes SachaS: sorry about this mmm thats ok. i should be more cautious you dropped the db? la la la laaaaaa reinhard: seems like the order is in reverse in the .sql file SachaS: fixed property, parameter, procedure, class, module --- fwiw there was some confusion about what gnue-import is for this is a program that can import *data* from a gsd file into an existing database with existing records i.e. it does inserts and updates from a gsd file into a datasource reinhard: is truncating ok? truncating? ALTER TABLE NOTICE: identifier "accountsreceivable_invoicetransaction_accountsreceivable_invoice_fk" will be truncated to "accountsreceivable_invoicetransaction_accountsreceivable_invoic" ALTER TABLE i think so erm i think, too but you should try to have names shorter than 32 characters (this is a postgresql limit) i should svn up again, dont I :) SachaS: yes dah :) * SachaS makes happy dance :) *** jcater_ has joined #gnuenterprise SachaS: those names are way too long man ;) hi jcater_ chillywilly, those are the full qualified property names ... so that can get long. SachaS: for example, i'd use "ar" for "accountsReceivable" or "accrev" err accrec put that's more a personal lazyness :) reinhard: i think I lost scale values.... *** jcater_ has quit IRC actually i like to use the full names .... then i remember what it is ;) (remember filename of the manual document? ;) SachaS: i would appreciate very much if you could confirm if gnue-gsdgen and gnue-schema are ok again now or if they still lose stuff reinhard: you might check if ggcvs also gets the scale values? morning all reinhard: hold on i check on scale again... SachaS: scale works for me good morning jcater good to see you back here ;-) thanks changed my mind =) did derek talk you into coming back? :) reinhard: scale looks good for me. SachaS: thanks have to go jogging. bbl l8r SachaS jcater: did you already look at the raping^H^H^H^H^H^Hchanges i did to forms yesterday? ??? err sorry was meaning jamest: did you already look at the raping^H^H^H^H^H^Hchanges i did to forms yesterday? wrt the dropdowns nope yeah the dropdowns i was too busy screwing up designer which means jcater best screw up something in appserver so the cycle will be complete heh lol :) I love this place * reinhard would seriously *love* to see jamest and jcater playing with appserver but i guess it's more lack of time than lack of interest reinhard: you don't want me in appserver just ask jamest i plan to play honest reinhard: you're right... it's a matter of time I like what you're doing with it it's time but I'm already spread too thin damn dude, with all the donuts you eat? * jamest runs rofl well there is one thing I'd like to see eventually not sure how open you guys are to it but I think GParser could easily be extended to support more than just XML file formats (i.e., it could also support a database backend) it would be nice if the class defs could be make GParser compatable so appserver could gain all the work we've put into that (like almost instantaneous Designer support) but I'm not complaining I'm glad to see you guys using common as much as you are lol i was thinking it'd be nice to see database stored forms like appserver via gparser I think that'd be relatively easy to do i think so too i might be crazy but i think that it'd be nice to see dynamically generated forms taken from introspection information and from access permission information I think that's just a layer that would call forms, imho probably yes forms will take a form def from a string buffer how are you guys thinking about layout management in forms? i'm currently not thinking about it i'd like the see the 0.5.0 series wrap up with all the missing current features as 0.6.0 was cleanup for common and forms i agree very much on this then start thinking about where we want to go jamest: i'm game with that :) as there are all kinds of good ideas out there now not sure if this would mean we'd push the cleaned up one as 1.0 then make 2.0 the target for all the new stuff or what 1.0 ??? you're kidding, right? ;-) forms/designer/common are really pretty well along just teasing :) for 0.5 *** wendall911 has joined #gnuenterprise good morning :) morning wendall911 nice April Fools jcarter ;) he was serious. just quickly realized he couldn't live without his dailly allowance of goat luving that's my take on it anyway reinhard: but that means adjusting the roadmap heh and listing must have features jamest: adjusting the roadmap? yeah jamest: i do that on a weekly basis for appserver :) lol jamest: if you are really honest, you just fear that we will reach appserver 0.1 before you reach forms 1.0 ;-) yeah i'm starting to feel the heat i figured i had at least until 2007 reinhard: I'm thinking a lot about it (layout mgmt) I want to get input masks working first but I put off finishing the QT driver err, sorry users the last problem I hit with the QT forms driver was getting the stuff lined up on screen and I was hoping to address some layout mgmt stuff while fixing that jcater: yes, input masks would be a big improvement to forms things I need in forms itself case insensitive query input masks in loaded records searching and something I'm forgetting :) tweaks that would be nice visible indicator of currenly focused master row when in detail ( i fake w/ triggers now) wow that's all I can think of that would make my users happier ah! working cut n paste :) working dropdowns? i thought you fixed that yday? as i wasn't fast enough for you :)