*** holycow has quit IRC *** jcater has quit IRC *** reinhard has joined #gnuenterprise *** johannesV has joined #gnuenterprise *** jcater has joined #gnuenterprise *** siesel has joined #gnuenterprise hey siesel long time no see hi reinhard i had a quite close look at the rcp stuff and i was wondering and what do you think for the dynamic object stuff currently, for every new instance of an object new methods are registered under [].method is that correct? yes why don't we simply pass a kind of "self" parameter to the methods and register them only once? both methods work what was your reason to choose this method? hmmm. I try hard to remember, but ...:) .. I seem to forgot it. Probably there is some info in the TODO document. didn't find something however the registering of new methods for every dynamic object causes some problems for example some (most) rpc libraries don't support unregistering of methods do you have rpc clients outside gnue? btw i changed both py_xmlrpc and pw_xmlrpc to transport all our datatypes transparently which rpc libraries do you mean? No, no clients outside of gnue I will have a look at the rpc librarys now siesel: can you do me a favour please? you have python 2.3, right? reinhard: I've created a new parameter called "objref_passing" to set how obj_ref can be passed yes. (2 times) can you please try this: >>> raise Exception, 'blöd' Traceback (most recent call last): File "", line 1, in ? Exception: blÃ¶d ok now this: >>> raise Exception, u'blöd' oops >>> raise Exception, u'blÃ¶d' Traceback (most recent call last): File "", line 1, in ? Exception>>> ok so still the same in python 2.3 i think i'll define a gnue.common.i18n.Exception class that does unicode -> local encoding cool :) or maybe we have to override "handle_exception" or how this is called but this must be fixed yes. IMHO this is a bug in python re self parameter everytime you pass a unicode error message this error is raised? i didn't want to make *you* do it i just wanted to ask if you were ok with that change no problem. siesel: everytime a unicode error message is passed, it is simply "swallowed" and very often we have the problem that we do things like I have some time now, and there are some things still missing in RPC raise Exception, _("message %s") % param when param is unicode, the % fails because the _(..) will return an already encoded string now i also see the need for the u_() function --- reinhard: unicode: does it allways happen when a message is raised, or does it just happens when printing the stack trace? just when printing the stack trace i think not sure siesel: i will care for that if you really have some time i think it would be much more important to fix forms under windows (just my 2 cents) currently, we don't use dynamic objects at all in appserver yes, but I haven't setup a full development env under windows yes so this is not a #1 priorit ah ok nevertheless i would think it's important :) :) bbl stop one question :) ok, lr /lr/later/s *** SachaS has joined #gnuenterprise hi everyone hi johannesV hi siesel hi SachaS how are you? have to restart X fine :) siesel I took some time to look at appserver but need reinhard ;) as I come across problems *** dcmwai has quit IRC *** dcmwai_ has joined #gnuenterprise so you have a running appserver now? siesel yes. the sample is working so i thought i start to make my own sample but now i run into issues here and there I will be here for some time, so just ask if you've questions, possibly I can help out ok. hey siesel is there an event u attend somewhen soon? ok. time to restart X *** SachaS has quit IRC probably linuxtag *** btami has joined #gnuenterprise *** SachaS has joined #gnuenterprise hi btami, wb SachaS hi siesel and all hi brami hi SachaS siesel what are you working on? back SachaS: rpc-code ok. reinhard: I've found out, why I used the [SDSSD..SSS].function way to pass the OID siesel: ok, please tell me :) it allows clients to work even when they don't have access to the GRPC file. and the self parameter won't? you think the problem is that the client won't know what object type the dynamic object is of? if an method returns an object, it need both unique id and object type one solution would be passing them both yep or making the "object id" something likey "object type:somevalue" another to read the object type from the grpc file hi reinhard siesel: i like the first one better yes, but object type:somevalue, doesn't have a fixed length siesel: so a function could return different object types, depending on something siesel: the fixed length thing is an ugly hack IMHO we must find something else to indicate an object anyway hi SachaS as ugly as gnue_id ;) no oh gnue_id is a question i have ;) why :) siesel: the probem is what happens when a rpc function returns a string that by chance *has* a length of 40 problem even ok, you mean this part of fixed length. SachaS: shoot :) yes, that's BAAAAD. reinhard: i was wondering if I can setup my sample application without touching the .gsd files by hand. siesel: what does the "allow_none" parameter do? so i add the gnue_id property in the gnue_classes.gfd form SachaS: should work SachaS: be careful to make what about passing a dictionary instead of a string: ( "objref":"sdafsdfasdf", "objtype":"Session.bla") module = gnue but then when I get the .gsd file via the ggcvs I do not get the gnue_id property in my classes. name = id type = id if allow_none is true, None values can be passed via XMLRPC lenght = 0 scale = 0 siesel: why don't we set that to True generally? siesel: i like that idea siesel: but then, we must encode normal dictionaries somehow reinhard. was that for me? module=gnue, name=id, type=id, length=0, scale=0 ? siesel: we need some means of "escaping" or something like that SachaS: yes reinhard: and in my own module? same? yes module=mymodule, name=id, type=id etc? the objectid must always be gnue_id no module=gnue name=id type=id always always module=gnue even within my class? yes is that confusing? :) ok will try. siesel: the problem is that we want to transport types over xmlrpc that are not valid xmlrpc types especially objects and if we use some other structure to transport those objects no matter if we choose dictionaries, tuples or strings we again have the problem that we basically loose the ability to use the structures in a standard way for example we could do dictionaries (which i like most) for objects we would do {type="object" objtype=object_type objid=object_id} then if we want to transport a dictionary, we would have to convert that as The reason behind the objects was to have the ability to transfer objects with gnurpc even if we use xmlrpc {type="dictionary" value=the_real_dictionary} siesel: i understand siesel: and i think that it's ok to do it siesel: just trying to think about a better solution than "if len(result) == 40" siesel: another question siesel: which of the rpc protocols were actually functional *before* my changes? I think we should the the dictionary way and add a "totaly standart conformant" parameter siesel: i would try to fix them just pyro ok is pyro available as deb? yes for woody or only for sid? only sid ok btw. there are problem with appserver and some db drivers: popy and pypgsql doesn't work pypgsql f.e. return string values instead of unicode hmm, got Berkeley DB error again is this a svn bug in 1.0 ? getting DB error too. running svnadmin recover now thx siesel: btw. i'm not sure about one thing in win32 uidriver in meny places i'v changed to textEncode(...) as reinhard did it in wx but not sure about # display unicode chars properly if type(value)==UnicodeType: value=textEncode(value) #.encode('mbcs') in _base.py line 506 the win32 ui now works well with my textencoding=cp1250 can you try your chinese forms? *** reinhard_ has joined #gnuenterprise *** reinhard has quit IRC btami: ok, will do later today siesel: thanks btami: it was siesel who changed those things in wx not me * reinhard does not want to take the praise for things he didn't do ok,sorry :) or s/praise/blame ;-) hmm, designer doesn't fixed yet regarding textEncoding/unicode :( l8r *** btami has quit IRC reinhard: did you see my comments from last night? SachaS: i think yes didn't we already talk about them? siesel: did you already run the database recovery? reinhard: one was that I get an error when I remove a property, using the gnue_classes.gfd form. oh didn't see that SachaS: the nullable button should now work perfectly in gnue_classes.gfd let me try to remove a property reinhard: yes, but it seems to be stopped siesel: what do you mean with "stopped" ? SachaS: i can reproduce the error, which means it will be fixed in 10 min ;-) svnadmin is blocked by some svnserve processes siesel: ah reinhard ;) SachaS: fixed, just waiting for svn to come up again reinhard: i am your current tester :) i will try with the id id stuff. i think it works. SachaS: no joke, this is invaluable help you're doing *** dimas has joined #gnuenterprise reinhard: do you have root access on ash? siesel: no jcater!!!! donut, donut, donut!!! lol * SachaS makes the happy dance - created gnue designer form with appserver as datasource. form is working ! SachaS: cool the id id stuff is good i mean is working, its a bit confusing but working. proabaly you could do that behind the scene later. I have no idea what it is and atm dont want to know. next test: referenced property :) SachaS: yes, it will be done behind the scenes SachaS: for a reference property in a form, you could look at the sample form the country is a reference siesel: FWIW: unicode exceptions are fixed and i've introduced a u_() function will commit asap you mean, when svn is running again :( exactly Recovery completed. The latest repos revision is 5435. siesel: thanks *** sjc has joined #gnuenterprise reinhard: what is the scale property of a type? integer oh you mean the meaning? length of a number is the total number of digits scale is the number of digits after the decimal point for example 1234.567 has a length of 7 and a scale of 3 now this commit takes long... oh i see. thanks. so length includes the digits after the point? yeah :) yes siesel: Transmitting file data ..Read from remote host svn.gnuenterprise.org: Connection timed out did you kill the svnserve process? hehe when svn is down, the developers have time to "communicate" on email ;) yeah :) bbl tonight (probably) *** reinhard has quit IRC what? reinhard is gone? dah. *** sjc has quit IRC see yas *** SachaS has quit IRC *** jbailey has joined #gnuenterprise *** gsoti_away has joined #gnuenterprise *** thierry__ has joined #gnuenterprise *** thierry_ has quit IRC *** johannesV has quit IRC *** siesel has quit IRC *** jbailey has left #gnuenterprise *** reinhard has joined #gnuenterprise *** bigbrother has joined #gnuenterprise *** bigbrother has joined #gnuenterprise *** jcater has quit IRC *** jcater_ has joined #gnuenterprise *** wendall911 has quit IRC *** btami has joined #gnuenterprise reinhard: i'v tested gnue on XP again, and find 2 bugs you use strptime() from mx.DateTime in forms GFDisplayHandler.py and in pw_xmlrpc 's typeconv.py ,but there is no strptime() in mx.DateTime on win32 :( hmmm, who is DavidPesta, sounds like a hungarian name... *** SachaS has joined #gnuenterprise hi everyone hi reinhard hi SachaS hi btami *** jcater_ has quit IRC *** jcater__ has joined #gnuenterprise svn down again? I got a connection refused try again I forgot that I killed the anonymous checkout trying to figure out an earlier problem ok yep working again *** dneighbo_ has joined #gnuenterprise *** reinhard has quit IRC *** reinhard has joined #gnuenterprise wb reinhard * ajmitch mutters hi ajmitch ajmitch is it difficult to make a .deb of PySablot for gnue_designer? *** dneighbo has quit IRC SachaS: what do you want to do with pysablot/designer? btami: dont know :) installing it, maybe try reports later pysablot was used as a general (xslt) report filter, but there is another: libxslt it has .deb ok. so instead of pysablot, libxslt works as well? yes and it has a deb. then I will take libxslt thanks btami. np *** ogger has joined #gnuenterprise *** ogger has quit IRC *** ogger has joined #gnuenterprise * SachaS makes happy dance. referenced property is working *** btami has quit IRC *** ogger has quit IRC *** wendall911 has joined #gnuenterprise *** ogger has joined #gnuenterprise anyone still here? sorta hi reinhard did you get my email? *** ogger has quit IRC no don't think so no? its a document I wrote. no, didn't get it you tried reinhard@gnue.org? yeah I think yep reinhard@gnue.org please try reinhard.mueller@bytewise.at its a 450K document not sure if @gnue.org works at the moment ok. no problem ok. ok. sent what are you doing reinhard? gnue or just logged on? actually not very much watched tv ok did you get my email? not yet what? but it can take a minute or two i think seems it does not get through i am using someone elses mail server ... my emails might get recjected. oh otherwise i can send it here you could use email.aon.at as smtp server to send a mail to me ;) ok, that works, too :) i put up a wireless ap at my neighbours place and a wireless ap at my place ;) saves me lots :) ok. i like the screenshots ;) its realy a quick and dirty documetn may i ask what license it is under? or more specific may we put it on our website and would you provide the "source code"? of course :) i think this is what at least some of the people on the list are looking for gpl and the gpl docu licence? yeah i think so too, with some screenshots etc. i will read through it tomorrow just to make sure that we don't publish some bad misundertsandings *** ogger has joined #gnuenterprise and will put it on website afterwards thanks a lot! ok. better read through. its really a start or even a draft that's ok it's much better than everything we have now taken together haha what did you use to write it? forgive me: open office ;) no problem quick and dirty that's free software and those lines ... i cant figure out how to remove them ;) well. referenced properties work SachaS: you might have to forgive me if i give it a shorter title ;-) that was the main thing for me for tonight oh isnt that what that verbatim / non-verbatim part is of the gnu docu licence? haha but yeah :) no such restrictions on this document well of course not the title, but the filename := :) true ;) so ... next thing is to look at procedures a little bit more if look at the last screenshot