*** reinhard has joined #gnuenterprise *** btami has joined #gnuenterprise *** Morphous has quit IRC *** johannesV has joined #gnuenterprise *** Amorphous has joined #gnuenterprise *** holycow has quit IRC *** dimas has quit IRC *** Morphous has joined #gnuenterprise *** Amorphous has quit IRC johannesV: i'v just tried the type="boolean" in parameters with monthly.grd and when i change it's value, click in a checkbox, i get tamas@TAMAS-PC:~$ grcvs -u gtk2 monthly.grd DB000: Can you feel that GNUe-Common love?

DB000: Traceback (most recent call last): DB000: File "/home/tamas/svn/gnue/gnue-reports/scripts/gnue-reports", line 29, in ? DB000: GRRun().run() DB000: File "/home/tamas/svn/gnue/.cvsdevelbase/gnue/reports/client/GRRun.py", line 245, in run DB000: omitGNUeXML=self.OPTIONS["exclude_xml"]) DB000: File "/home/tamas/svn/gnue/.cvsdevelbase/gnue/reports/base/GREngine.py", line 112, in processReport DB000: omitGNUeXML=omitGNUeXML) DB000: File "/home/tamas/svn/gnue/.cvsdevelbase/gnue/reports/base/GRReport.py", line 139, in process DB000: dest.write ( parameters.getRunOptionsAsXML() ) DB000: File "/home/tamas/svn/gnue/.cvsdevelbase/gnue/reports/base/GRParameters.py", line 89, in getRunOptionsAsXML DB000: xml += ' %s\n' % \ DB000: File "/usr/lib/python2.3/xml/sax/saxutils.py", line 30, in escape DB000: data = data.replace("&", "&") DB000: AttributeError: 'bool' object has no attribute 'replace' btw. good morning :) btami, thanks for reporting, i will check it asap ok, so let's have a look at this ... btami, this seems to be a bug in reports btami, please svn up and try again it's ok now, thx anyway, i think that handling of parameters need a bit more TLC *** kilo has joined #gnuenterprise good morning hi kilo hi johannesV, what a quiet day it was yesterday... well, i had a holiday yesterday ... :) i know had you got good weather there at least? here was 25 C yesterday no, it was rainy yesterday :( but this way i had some time to fix that param-form and navigator-thingy in the evening johannesV: you should finish with reports today, nav tomorrow, then friday is for forms. than monday you could help debian so that sarge could be released tuesday... 8-)) as we talk about sarge has anybody heard from ajmitch_? our debian packages are *way* outdated kilo: *lol* kilo: no luck kilo: monday is holiday ;-) ROTFL oh, then sarge could be released only wednesday??? no, johannesV likes to hack on holidays too :) l8r all off to customer *** reinhard has quit IRC johannesV: you've played with forms recently a little bit what do you think, could a new trigger be implemented for a dropdown that fires at the selection of an entry, while the focus is still on the dropdown? i'd have to check this, but there's no such an event internally anyway selecting an entry from a dropdown fires a 'setValue' so it's not different from entering the value one character by another i think you agree it would be nice if upon entry selection some info on that entry could be displayed for the user for example ? hotline, selecting a contact... name, phone, email could be displayed *** stan has joined #gnuenterprise kilo, this should be done automatically by selecting the apropriate 'record' from the lookup-source so no trigger would be needed anyway ... it's a known bug in forms (see BUGS file) yes, that is one way, and i know that is buggy but there can be scenarios where it is not appropriate eg you choose a payment method and the invoice's dateDue field could be automatically computed based on dateIssued and paymentMethod this can be achiev by a trigger, but it should be fired when selecting an entry now it is hooked on on-change and that means it fires only when the user changes focus to another entry *** dimas has joined #gnuenterprise *** btami has quit IRC *** stan has left #gnuenterprise johannesV: OTOH, if i declare a field as nullable=false in the gcd, and would like to fill it in OnValidate, the form cannot be commited so can i not fill non-null fields in onvalidate? must i fill them before? just a sec kilo, you're right ... *lol* testing for NOT NULL constraints is done *before* execution of OnValidate-triggers ah, that WAS an answer 8-))) defined by modules of the class i see. is it by design? yes for now :) the idea behind was: all OnValidate triggers can rely on the basic integrity so a trigger can be sure, all fields reqeusted to be not null are not null the question arises if it makes sens to change fields in OnValidate yes it does 8-)) well, maybe for invoice numbers ? yes yes yes hmm i would have to discuss it with reinhard ... from a technical point of view it would be easy to change yes, it is clearly a design question imho onvalidate is there to check fields for validity. if it cannot be null, onvalidate must make sure it is not null. it can throw back the commit with an abort and make the user fill in fields with sane values, or it can fill them up byu itself. but in the second case the possibility to throw the commit is still there, so the second approach imho is more flexible. am i clear? * kilo is unsure... kilo, it's not very complicated the only thing we've to make clear is wether constraint-checking comes before or after execution of all OnValidate triggers if it is done afterwards a trigger has still the chance to throw an exception (although it's not the same) yes, i see, i only try to conclude my thoughts online... an persuade myself which is the better one problem if it is done afterwards still remains other OnValidate triggers cannot relay on valid reference-properties any more it has not been guaranteed that a ref to another class is filled in so you could not do something like "address_person.address_country.address_name == 'foobar'" in an OnValidate of address_person class because you cannot know if address_person.address_country is ok but - still - it is a design question trigger code would have to do extra checks and it's ok ... as reinhard would say, hmmm :) ok, other problem then invoice. you know the drill... M/D, detail is autocommit, autonextrecord it works ok, but if i put in a trigger to an invoice field or entry (does not count) that uses blkSomething.call(), it simply wont commit the detail records you can see it in gnue-invoice recent commit which form? inv_head.gfd enable trigger in line 149 for example ah, i see /msg everyone he sees. that is good. solution is due in 5 minutes... hmm i try to understand you gfd-construction ... :) hmmmmmm this is really strange the call () does an implicit store that's ok but why are the recs missing on commit kilo: got the problem but i have no solution atm it's in the combination of data.py and dbdriver the record (item) is listed as 'changed' but has an internal state of 'inserted' i need to dig much deeper .. *** kilo has quit IRC *** kilo has joined #gnuenterprise kilo, you there ? the problem is much more complicated, and i'm not sure if appserver is doing wrong :) johannesV: sorry i've been cut off seems as if the inv_head column is not set properly ... but i do not know where it will be lost ... hmm, that should be set automagically as it is an M/D... right, but it seems as if this is *not* done something happens during that implicit store this would be a forms-thingy the call does a post () on the recordset, so the procedure sees all changes made in the form and it works ok if there is no call()... and one more iteresting problem try uncomment line 79 in the same file. it causes infinite loop on commiting an invoice line... got to go now, bbl *** johannesV_ has joined #gnuenterprise *** johannesV has quit IRC *** jcater has joined #gnuenterprise *** reinhard has joined #gnuenterprise @weather LHUD kilo: The current temperature in Szeged, Hungary is 70°F. Conditions: Partly Cloudy. Humidity: 53%. Dew Point: 56°F. Wind: SSE at 9 mph (14 km/h). Pressure: 29.84 in (1010 hPa). Visibility: 12.0 miles (20.0 kilometers). UV: 0 out of 12 kilo: as to setting not nullable fields in OnValidate if you have an OnValidate procedure where you set the field there's no use in defining the field non-nullable bbl *** jcater_ has joined #gnuenterprise *** jcater has quit IRC hmmm, i disagree OnValidate is an implementation while the fact a field mustn't be null can be a constraint from the outside world db design and db use can differ both wrt time, place and humans I do not see the big difference between defining a field non-nullable and defining an OnValidate both define rules that have to be fulfilled that an instance can be saved both are defined in the same place actually I see nullable="False" as a shortcut for "if foo is None: raise bar" in OnValidate *** sjc has joined #gnuenterprise can we change the process order then? kilo: i still believe it perfectly makes sense to check non-nullable before running OnValidate to make sure that OnValidate doesn't have to handle all cases of properties not set however if others agree with you i could live with changing it now it's 1:1 i would like to hear the voice of others like lema, SachaS, or other appserver users (sorry, but i won't count btami as he might be biased ;-)) brb reinhard, could you please read the mail regarding db-driver issues i've found today kilo: i've tracked down your master-/detail-troubles and we're working on a solution but it might take till tomorrow bbl johannesV_: i've seen the mail and i will need some time to think about it ok anyway - it seems that master-detail with appserver is a kind of lotery atm after trying to add that tabular-form support to labels.py today i came to the conclusion a 're-structuring' of labels.py would be best ... :) if i think of the next steps (like search-dialogs, permissions (where fields get faded out depending on other criteria) and much more) but that restructuring of labels.py has nothing to do with m/d does it? no it has not ok i meant: there's enough work left (for tomorrow) ah ok so that m/d stuff has a bit time well meanwhile i read through your mail ok *** kilo has quit IRC *** kilo_ has joined #gnuenterprise *** johannesV_ has quit IRC *** kilo_ has quit IRC *** kilo has joined #gnuenterprise *** cilkay has quit IRC reinhard: yet i still have some bad feelings: if OnValidate assumes all non-null fields are to be filled prior to the point when execution gets to OnValidate, I can only achieve that with form triggers that means i put business logic code to ui part checking for non-nulls should be part of the validation process, and where else would it be best to do than in OnValidate? btw lekma is lost somewhere, Sacha is lost in the States... any other active AppServer users out there??? btw it can be very schizofrenic to think about AppServer, sometimes i argue with myself... eh, time for a whisky now, before dwelling into VB kilo: i think that setting a property in OnValidate *and* defining it as non-nullable is redundant but the more i think about it the more you make me unsure i think i have to sleep over it there's another aspect well actually not hmmm (I've learned it from you...) *lol* think think think think think think it's so hard to thing when you're a little bear with a little brain s/thing/think/ *** GNUe110 has joined #gnuenterprise *** GNUe110 has quit IRC oh, that IS lovely... *** cilkay has joined #gnuenterprise really, are there any AppServer users out there??? I am not sure in fact I'm a bit worried about lekma he was last seen 19 days ago i have to finish this week the xul client nearly there i still got some trouble with master/detail was something he wrote then so a) he was successful and doesn't need our help any more or b) he gave up :( 8784 gnue-appserver 19 3 12 4 0 (Andrew Mitchell) from popcon.debian.org 19 people have appserver installed from debian 3 of them voted to use it regularly yes, i am very very curious about lekma too (of those having the populary contest activated) good night all *** reinhard has quit IRC *** cilkay has quit IRC *** cilkay has joined #gnuenterprise *** kilo has quit IRC c) mozilla trolls ate him derek: lemka grrr *** sjc has quit IRC *** holycow has joined #gnuenterprise *** cilkay has quit IRC *** holycow has quit IRC *** jcater has joined #gnuenterprise *** wt has joined #gnuenterprise hi jcater, if you are here, is there a way to add tooltips to the toolbar in the designer? do we not? not that I see *** holycow has joined #gnuenterprise holycow, that is a funny name wt, hola hi yah is it just me or is the documentation that is built in the svn way out of date? coding is way more fun than documenting way out of date? I don't think so which are you looking at? gnue-docbook stuff GNUEnterprise is the Object server implemented in Corba? ah I need to remove /docs/ I am talking about the stuff from the svn tool-specific ones are good not just the stuff on the web http://www.gnuenterprise.org/tools/appserver/docs/ etc *** cilkay has joined #gnuenterprise gnue-docbook is still there? yes reinhard: when you check the logs do we still need gnue-docbook? some of it is relevant * jcater thinks we need to move that to our archive repository like naming standards maybe they are relevant what is a .gcd? appserver file defines "classes" okay basically like database views only with business logic is there a list of the file types? (reinhard may not like my oversimplification :) I like it :-P gcd -- GNue Class Definition --AppServer files gpd -- GNUe Process Definition -- Navigator files (basically defines a menu structure) gfd -- GNUe Forms Definition -- Forms Files grd -- GNUe Reports Definition gsd -- GNUe Schema Definition -- Defines database tables in non-vendor XML format where is the documentation for this stuff? it would be under the individual tool's docs there is usually a .dtd file there, and reference guide some more than others gld -- gnue label definition -- Part of AppServer (i18n support, iirc) gid -- GNUe INtegrator Definition -- Defines transforms from one database to another (not developed yet) um what am I missing? where should one start?