*** someon has joined #gnuenterprise *** holycow has joined #gnuenterprise Heya Holycow. or is Moo more appropriate? *** jcater has quit IRC hey dude :) either or i've been called worse :) howya doin? DNS change? Not bad. Wish I could figure out Python... but alas, I'm as dense as a bag of hammers i dun think so, mebbe shaw is playing funny again I guess. You're just showing an IP. and i'm a newb so hey two of a kind :) at least i can read python, all other languages give me a headache A pair beats a ton of nothing a pair of newbies beats an ace of python? hmm.. logic flaw there somehow. lol How're things in Ample? bored, reading comics, waiting for gnome to be release u toying with python huh? Erm.. Python -> GNUe... hehe :) i havent had a lot of time to play with gnue :/ or is it GNUe -> Python *** |dcmwai| has joined #gnuenterprise sigh I noticed in the logs though that security is going to get a look or two... That would be cool i sent this to the list but SuSE and I have agreed that I don't really need to send email Though I'm not sure that security can be achieved much on 2-tier http://www.gnuenterprise.org/cgi-bin/twiki/view/Main/SuSEQuickStart i think it was a matter of security hasn't been addressed in appserver but wouldn't take much to add it is there a similar doc for Debian? Nods, that's kinda what the logs were showing. the source build is the same on sid http://www.gnuenterprise.org/cgi-bin/twiki/view/Main/DebianUnstableQuickStart lists the package dependencies i guess i didn't confirm so between the 2 one should be ok When I asked before about security, I was told that I could push the XML-RPC over https:// so, that would give "over-the-wire" security, so all that would really be needed would be ACLs and Password security. The latter appears to be possible through good programming erm.. that is for AppServer stuff... not 2-tier *** chillywilly has quit IRC *** chillywilly has joined #gnuenterprise *** |dcmwai| has quit IRC *** |dcmwai| has joined #gnuenterprise someon, i really don't know much about appserver i do know I need sleep badly :) nite jamest: who doesn't? 'night my wife she told me to come to bed soon about 3 hours ago :) <|dcmwai|> who is your wife.. jamest <|dcmwai|> haha : that lady that lives with me :) <|dcmwai|> I know I meant what is her name and did she here before? lindsay and never in here she's not a geek like me <|dcmwai|> haha <|dcmwai|> Faster goto bed... |dcmwai| i may need to pick your brain about aria soon as I really want to see about using it <|dcmwai|> else you will e lock out of the bed room :) :) nite *** jamest has quit IRC *** |dcmwai| has quit IRC Question for log readers (or developers if they are awake)... if multiple entries in connections.conf, then how does forms/designer/etc. know where to do user authenication? Looks like the loggers will have to get that one holycow: You work evenings or something? You tend to come on pretty late... no regular hours, but long hours Eww. Been there done that. after work i try to pick up as much linux related knowledge as i can so ... growth hurts i guess :) Nods. I'm working on developing converts. *nod* but its for a good cause, gotta exterminate the ms plague, i refuse to bow to the monopoly I was first introduced to the cause about 10 years ago. Got hooked about 5 years ago my self about 2 years ago, 1 year ago i took the plunge its not so much getting rid of a plague even, but restoring balance to the marketplace I'll still have to read the docs when I think about recompiling kernel at least in the few bits i can affect Not even that for me. It's about HAVING choice *** |dcmwai| has joined #gnuenterprise heh :) i'm getting there slowly, although debian makes it dead easy right exactly I'm liking Knoppix on a HD install, though I'm told that KDE 3.2 is a HUGE improvement. i like kde3.2 a lot try kanotix it installs a pure debian to hd, although live cd still uses lots of knoppix bits Knoppix seems to too, but I took about 3 hours cleaning up packages I didn't want thats not what i heard, but not sure have you tried d-i? so far i thought it excellent for autodecting various oddball pcs configs *** dimas has joined #gnuenterprise No.. Knoppix worked wonderfully... even registers 4 CPUs on a 2xXeon(HT) server :) lol really? sweet I took time taking games packages etc off the Knoppix Debian install Pulled about 100 MB out. i want d-i to get ldap support d-i? that way i can use ldap based config files to setup machines as per my spec :) debian installer the new one thats going to be in sarge Ahh. This is for a machine that cannot be connected to the net, so I'd have to download the CDs... Knoppix was great for that too. Though kanotix looks interesting. The author... Jorg Schirottke ... Looks familiar... http://debian.tu-bs.de/kanotix/KANOTIX-BUG-HUNTER-04/ erm.. name looks familiar. *nod* he seems quite talented great with hardware stuff *** johannesV has joined #gnuenterprise look... an AppServer person :) oi. 128 MB minimum? We coerced Knoppix to do KDE on 96. *** |dcmwai| has quit IRC nite *** holycow has quit IRC *** reinhard has joined #gnuenterprise *** someon has quit IRC *** kilo has joined #gnuenterprise *** |dcmwai| has joined #gnuenterprise good morning. reinhard a print in an application server can be very helpful for debugging. SachaS: sure, but you would not use _() for that ok. does _() encode a string into unicode/utf-8? it encodes into the currently selected local encoding unfortunately, the appserver (on the server side) doesn't know about *your* (client side) local encoding so it can't work *** |dcmwai| has quit IRC ok. 'ning reinhard: gacvs --selftest with locale en_US passes OK but with encoding hu_HU it fails 1. lépés: Session Manager indítása... 2. lépés: Session megnyitása (felhasználó: 'hacker', jelszó: 'secret')... 3. lépés: Objektum lista létrehozása... 4. lépés: Első példány beolvasása... DB000: Traceback (most recent call last): DB000: File "/home/gabor/gnue-svn/gnue-appserver/scripts/gnue-appserver", line 31, in ? DB000: server.run (); DB000: File "/home/gabor/gnue-svn/gnue-common/.cvsdevelbase/gnue/appserver/geasRpcServer.py", line 169, in run DB000: self.selftest () DB000: File "/home/gabor/gnue-svn/gnue-common/.cvsdevelbase/gnue/appserver/geasRpcServer.py", line 238, in selftest DB000: print _(""" DB000: UnicodeDecodeError: 'ascii' codec can't decode byte 0xe9 in position 35: ordinal not in range(128) *** btami has joined #gnuenterprise *** M7U90nt has quit IRC hi kilo hi SachaS *** M7U90nt has joined #gnuenterprise kilo: thanks for reporting this kilo: can you svn up and try again please? of course beautiful!!! thx reinhard for fastest_committer_in_gnue_universe! :) *** dimas has quit IRC *** thierry__ has joined #gnuenterprise *** thierry_ has quit IRC *** nickr_ has joined #gnuenterprise *** nickr has quit IRC reinhard: i do have an encode traceback. reinhard: maybe my fault... will check again. SachaS: can i have a look at the doc you produced? kilo I am updating it and then i am uploading it to the wiki ok then, cause i looked at the twiki but couldnt read it, onlyu "test" kilo can you get files via irc? if so, i send it to you now lets give it a try 500K! must be the screenshots ;) actually the screenshots are what I have to update. hmm.. now i wonder where xchat put it... check in the xchat properties ;) fund it found off for lunch johannesV: are you around? kilo/btami: i have removed all mx.DateTime.strptime can you please test now if it works again under windows? yes, but not now :) maybe later afternoon at home btami: no problem just lost in appserver sample... kilo: can i help you? *** btami has quit IRC classrepository.gfd, procedures page i select module "gnue", then run query, and then focus out the dropdown and it looses its value *** btami has joined #gnuenterprise not sure what you are trying to do procedures is acutally a detail to class kilo: in order to modify a procedure you have to select a class first procedures are detail to gnue_class, so without a masterrecord you can't modify or select a detail yeah so first, on the class tab, select the class you want i haven't found a way to "disable" procedures as long as no record is selected in classes and then change to the procedure tab and you can browse through the procedures of the current class btw. should we remove the old gnue_module.gfd and gnue_classes.gfd from svn ? yes (IMHO) ok very unclear it is ok, found it removal: yes btw when starting the form it says there is an error in one of the document's triggers kilo: i agree that it is not really intuitive if you have a better idea how to manage that please tell if one comments out line 45 then there is no error message kilo: the problem is that we have a 3-part master-detail procedure is a detail of class parameter is a detail of procedure reinhard: ok, will think about it, havent done so yet problem is clear thanks funny i don't get any error message when starting the form Runtime Error occured: There is an error in one of this document's triggers. Traceback: Trigger "unknown_executioncontext" (unknown), line 3 'NoneType' object has no attribute 'current' exceptions.AttributeError: ("'NoneType' object has no attribute 'current'",) i think we will need forms experts to get some light on this error message but thanks anyway kilo: does this error occur everytime you start the form ? have you updated your db with the new base.gsd ? (gnue_parameter) yes and yes hmm, strange ... updated from svn this morning *** M7U90nt has quit IRC hi johannesV we havent met ;) SachaS: hi do you have a new form to create modules and classes ? or why do you want to removed the gnue_classes.gfd and gnue_modules.gfd SachaS: yes, classrepository.gfd ok. good. reinhard i have a TypeError: encode() argument 1 must be string, not None\n'> ---) traceback created a simple form with designer and when I run it, after i log on, i get the traceback. * SachaS understands how quickly documetnation can get outdated ;) SachaS: can you paste the last few lines? ok return ([classdef.properties[p].column for p in propertylist])\n\n File "/home/sacha/gnuenterprise/gnue/gnue-common/.cvsdevelbase/gnue/appserver/classrep/Base.py", line 253, in __getitem__\n raise self._itemNotFoundError (key)\n\n File "/home/sacha/gnuenterprise/gnue/gnue-common/.cvsdevelbase/gnue/appserver/classrep/Property.py", line 100, in _itemNotFoundError\n return PropertyNotFoundError (self.__class.fullName, key)\n\n File "/home/sach a/gnuenterprise/gnue/gnue-common/.cvsdevelbase/gnue/appserver/classrep/Property.py", line 36, in __init__\n ClassRepositoryError.__init__ (self, msg)\n\n File "/home/sacha/gnuenterprise/gnue/gnue-common/.cvsdevelbase/gnue/appserver/classrep/Base.py", line 36, in __init__\n gException.__init__ (self, text)\n\n File "/home/sacha/gnuenterprise/gnue/gnue-common/.cvsdevelbase/gnue/common/apps/i18n.py", line 161, in __init__\n exceptions.Excepti on.__init__ (self, o(message))\n\n File "/home/sacha/gnuenterprise/gnue/gnue-common/.cvsdevelbase/gnue/common/apps/i18n.py", line 144, in outconv\n return message.encode (encoding, enc_policy)\n\nTypeError: encode() argument 1 must be string, not None\n'> ---)------------------------------------------------------------ thats on the client side just a sec DB005: [ResultSet:65] ResultSet created DB001: [Base:295] RPC detected Exception of type "exceptions.TypeError", value: "encode() argument 1 must be string, not None". DB000: debian-amd - - [23/Mar/2004 13:17:50] "POST / HTTP/1.0" 200 - this is on the server side can confirm kilos problem, when starting classrepository.gfd I get a traceback too nice form johannesV *** havoc has quit IRC *** derek has quit IRC *** Vee has quit IRC *** btami has quit IRC *** thierry__ has quit IRC *** dcmwai|AWAY has quit IRC *** chillywilly has quit IRC *** btami has joined #gnuenterprise *** thierry__ has joined #gnuenterprise *** chillywilly has joined #gnuenterprise *** dcmwai|AWAY has joined #gnuenterprise *** derek has joined #gnuenterprise *** Vee has joined #gnuenterprise SachaS: we are having issues with translations that occur in trigger and procedure code SachaS: please try running in the C locale *** havoc has joined #gnuenterprise reinhard: traced down the problem to GFField.py line 238 self._block._resultSet.current[self.field] = value seems like the resultSet is None on startup and if the trigger runs (and it does run) and tries to set the value of the field it fails kilo: have you found out which trigger has the problem ? yes, line 45 in the form, name=newClass it fires on startup and causes trouble as the resultset it tries to act on is still None SachaS: can you please dcc me the gsd and gfd's i need to reproduce this behaviour? i can't reproduce it (or mail if your mail works) *** lxf has joined #gnuenterprise reinhard: which one? the encode one? kilo: would you svn up the classrepository.gfd and try it again ? SachaS: yes johannesV is looking at the other one (kilo's) testers chasing developers, developers chasing testers ;) hunger chasing kilo SachaS: must run bbl in 1 hour johannesV: svn upped but still get the error the block itself is not None hmm, ok, so there's no recordSet in the block yes, exactly this is strange, cause this is the on-newrecord trigger :) question is why it is fired at startup hmmm, it's an unbound block which means no datasource in the background *** psu has joined #gnuenterprise kilo: please remove the trigger name="newClass" the whols trigger or just nthe name? s/whole/the it's useless anyway because the "setClassName" will be fired when the procedure-tab is selected you can remove the complete trigger (newClass) ok i'm curious if the same exception is raised by the setClassName-trigger it works now like magic ok, but now the classname-label on the procedure-tab renders wrong if you first query some classes and then insert a new one could you add this else-part to the "setClassName"-trigger: if len (classname) and len (moduleId): module = blkClass.fldModule.allowedValues () [0] [moduleId] blkUnbound.fullClassName.set ("%s_%s" % (module, classname)) else: blkUnbound.fullClassName.set ("not set") ups, this is the complete trigger-code ... *** lxf has quit IRC added the else part, trying now johannesV: the error reoccurs this way ok so it's ok ... everything else would have been mysterious ... kilo: you can remove the else-part ok, did so btw i played with it a little and didn't see that it rendered wrong when inserting a new record kilo: oh, so it's my forms/wx/whatever going wrong ?? :) kilo: so i'll remove the newClass-trigger-code from classrepository.gfd if i inserted a new record, it correctly displayed the name even before commiting the insert have you inserted the record after there was another classname displayed (like address_person) ? yes johannesV question: do you write your forms by hand? ok, good, so i'll change the gfd SachaS: yes ok. i love vim ! dont you love designer ? ;) i've tried the designer a few times, but it segfaults way too much :) btw with vim writing xml is quite handy kilo: new version of classrepository.gfd is comitted johannesV: ok, thx i am just not into all the forms tags etc johannesV: maybe a wx version difference between our systems is the clue? kilo: might be, as reinhard has thought of something similar btami thought of that too *** dsmith has joined #gnuenterprise *** jamest has joined #gnuenterprise rm-away: wouldn't you know just when I tell you freeswan is dead they revive it :) rm-away: appserver might not like camelCase for properties :| SachaS: what do you mean: anotherPropert != AnotherProperty ? yeah i think its the camelcase i had a class called "inventory" and for that class I had a property "item_id" so reinhard thought, thats not a good idea so i changed the property to "itemId" now i get tracebacks, when I run the form :| so must be the camelCase of properites hmm, let me try ... ok. ok, so where's your problem ? *** btami has quit IRC ok. i have a class called "inventory" and i have a property called "nameId" with that inventory. *** kilo has quit IRC ok when i create a form with designer to access the appserver and to get the nameId field. thats ok. i can work with designer and save the form. then when i run the form i get to the login and then i get a traceback and bybye the form has 1 field only the nameId field so i create another form where i get a property which has not camelCase, such as the "name" property of the "inventory" class. then, that form works nicely. so i guess its somewhere a problem with the camelcase. ok, would you check the gfd file ... ok. look for a this tag should has at least one -tag is the field-attribute of the field-tag in camelcase or not ? it does yes not camelcase so this might be the problem change it to match the propper case ok. yep that was it. but ... when i select the datasource of the appserver in designer, the list has the non camelcase listed. question: does appserver wrongly return non camelcase or does designer change to lowercase? yeah i saw that; i've to check appservers' introspection module. maybe the properties are string.lowered () there SachaS: don't know now, but i'll check it cool. SachaS: fixed please svn up back johannesV what was it? appserver's schema.discovery.introspection did a string.lower () on the property names think that was me who made that mistake :) ;) - 1 bug oh i have to update the forms manually, then ;) no problem, thanks for fixing it johannesV! no problem *** lxf has joined #gnuenterprise *** dneighbo has joined #gnuenterprise johannesV yep now they are upper case. i mean camelcase ok johannesV I think there is another .lower or .upper somewhere. when i commit the form. i get another traceback along the line it complains, that Property was not found. sigh anyone know if our windows installs are fried? gnue-common install nope nope, dont know *** derek has quit IRC johannesV did you see my comment here? SachaS: yes, i saw it :) which property isn't found ? i assume the camelcase one... i cant tell from the traceback... *** dcmwai|AWAY has quit IRC *** thierry__ has quit IRC post you in priv window SachaS: this seems to be the same encoding problem as before yes. *** thierry__ has joined #gnuenterprise *** dcmwai|AWAY has joined #gnuenterprise i am not sure, i think thats another problem ... that i cannot display any errors... not sure at all. SachaS: you still have that encoding problem? i'm desperatly trying to reproduce it and can't reinhard: we found the cause it was the idea with camelCase :) first the appserver did str.lower() when presenting the items to the designer then the forms i created had a problem, that the appserver could not find the fields all lowercase. so must be the reverse now. but i still want to know why the error message isn't issued correctly brb phone i think thats the second part of the problem. well dont look at me ;) *** GNUe587 has joined #gnuenterprise jamest sucks!!! cool, i never tried our web irc before *** GNUe587 has quit IRC * johannesV is away: airport heh a friend of mine in US was once told by his boss "I want you to miss at least 3 flights a year" (for work) - "as otherwise you're getting to the airport too early & wasting valuable company time" * psu wonders if reinhard applies the same rule to johannesV nope especially because he's not going to fly but he's picking up his wife there ;-) jamest: we are in the process of realizing that it could be a problem that appserver's field names are case sensitive do you think that could become a big problem? it seems that designer does some string.lower(fieldname) stuff before it saves a form what?!?! sigh, sql isn't case sensitive is it it isn' it isn't but do we require all dbdrivers to not be case sensitive or don't we reinhard: actually the designer does write the database field names correctly with camelCase. oh so there are some string.lower() in forms?? will look maybe in forms, or in rpc or in server-receving side i thought we were case sensitive reinhard: johannesV did fix the appservers introspection before, allowing the appserver to provide field names in camelcase. basically he removed the str.lower() yeah saw that