blah anyone awake? * derek is hi, you know anything about fixed point math? * ajmitch is awake but busy been running a new report on this lame system and there seems to be some significant rounding errors ;P ajmitch: fine then ;P good :) although when yo only keep 2 decimals around that can't be too accurate... you* I think I'd like to beat some people :) *** reinhard has joined #gnuenterprise hi good morning or evening ;) Mon Feb 7 01:02:44 CST 2005 I guess *early* morning ;) heh work station says 1:08am 1:04 would be correct I guess you know anything about using FixedPoint reinhard ? you should use ntp I do in Python? no what about in practice ;) or by some other means not really, no never worried about rounding errors? ;P no I always dealed with monetary amounts where 2 decimal digits are enough and floating point math was always exact enough for my purposes k I thought jcater was messing with it FixedPoint that is yes I remember he told about some problems some problems with the module or problems he has that led him to use the module? had* I don't remember I didn't follow it closely enough :( ok I will ask him tomorrow if he's around I should sleep...at least now I know I am not insane ;) it's just some stupid rounding issues ahhh just saw something in a ChangeLog 2005-01-06 17:38:28 Rev 6873 jcater * src/apps/__init__.py: removed decimal class until I get it working in Python 2.1 wth is the decimal class? something he wrote? I think that's something he picked up somewhere and put it in gnue-common/external g'night night *** btami has joined #gnuenterprise good morning good morning *** sjc has joined #gnuenterprise *** dimas has joined #gnuenterprise *** aries has joined #gnuenterprise hi *** kilo has joined #gnuenterprise good morning hi aries reinhard: reinhard: if you read the logs, Forms' po contains (C) FSF 2000-2004. shouldn't it be 2005 instead? hi btami: hi kilo: hi hi aries kilo: should be changed whenever something is changed in the po file I think it is *now* that something has changed which file do you mean? hu.po? de.po? in each po, as it comes from... wait I updated de.po, you can update hu.po if you want uidrivers/wx/UIWXSplashScreen.py:81 it comes from what you wrote :) huh? it is in a string that is to be translated oh no, that's awful should i change tha source and commit? I would propose to remove the year from the splash screen output I bet we will miss that forever yeah and I doubt the splash screen is legally relevant anyway where's a lawyer? is it safe to remove it? changing stuff in splash screens is often ignored until after the release :) kilo: I think it's safe to remove it if i commit it now, do i f*ck up your release? no it's ok done thanks well i was kinda astonished when i saw that i needed to translate that message kilo: I also fixed it in hu.po *** johannesV has joined #gnuenterprise ah, thx reinhard: i get this raceback from gnue-invoice (hu) if i try to modify customer, then come back and invoice item autommits itself Traceback (most recent call last): File "/home/tamas/svn/gnue/.cvsdevelbase/gnue/appserver/geasSessionManager.py", line 295, in load return s.load (classname, obj_id_list, propertylist) File "/home/tamas/svn/gnue/.cvsdevelbase/gnue/appserver/geasSession.py", line 513, in load instance = self.__findInstance (classdef, object_id, propertylist) File "/home/tamas/svn/gnue/.cvsdevelbase/gnue/appserver/geasSession.py", line 458, in __findInstance raise InstanceNotFoundError, (classdef.fullName, object_id) InstanceNotFoundError: A 'BILLING_Head' osztály '83149234333458361087285718605396' példánya nem található it was reported to johannesV before and i thought it was fixed :( ? :) johannesV: i confirm it is a live bug kilo, so we have to trace it ... what do i have to do ? gnue-invoice, start head.gfd, edit Buyer, Query/ExecQuery, quit subform, then select Cust1 as Buyer, Cash as Payment, add P1 as product and 1 as Quantity, press Enter and there it comes * aries is now online where is module BILLING defined ? packages/supply-chain/billing ah, so i need to update packages too .. ? hmm, no gnue-invoice/scripts/setup-db.sh does everythinh ah, no, i had a type in my commandline ? typo ? kilo, we don't need setup-*.sh anymore since appserver can integrate that gcd's itself ok, now starting head.gfd it hurts me very very much when you kick my wee little baby... 8-)) ok, have it open. now what shall i change ? kilo gnue-invoice, start head.gfd, edit Buyer, Query/ExecQuery, quit subform, then select Cust1 as Buyer, Cash as Payment, add P1 as product and 1 as Quantity, press Enter and there it comes there's the following error: Traceback (most recent call last): File "", line 12, in calcDateDue File "/usr/lib/python2.3/site-packages/mx/DateTime/DateTime.py", line 493, in __init__ self.days = days + weeks*7 TypeError: unsupported operand type(s) for +: 'NoneType' and 'int' :) don't fill dates bah, can't load invoice-hu.gsd ERROR: insert or update on table "item_item" violates foreign key constraint "fk_item_item_item_unit" we usually leave them dates empty or copy/paste the one that is displayed johannesV: svn up gnue-contrib but i've updated it just 5 mins ago ... you can reproduce without hu too go for the english one In Revision 211 kilo, is this that sample with the bound/unbound block thingy ? johannesV: no ok, now let's do it step by step please ... i've integrated invoice-en.gsd now and have some sample data available i start Head.gfd the foreign-key error was resolved by btami, but you hand-edited the gsd files later. edit Buyer you mean press edit-button, or just select one ? press edit button ok on the subform press Query, then Exec Query leave thru the door ok back in Head, select Cust1 from Buyer ah, ok, i see i think this is quite obviouse then erm, obvious * kilo shut up? or continue tour? standby :) red lamp, /me in standby yeah the subform runs a rollback ! when you press 'new query' isnt it a different session??? no we cannot start different session atm, since all that sessions are based upon the same connection ok, I go back to Visual Basic... but this is a shortcoming we do no very long now i think we've discussed such a feature about one or two month ago ... but there must be common-support for this in order to realize it johannesV: is it a bug (the traceback), or we do something wrong? * johannesV phone *** sjc has quit IRC btami, i would call it a missing feature the problem is starting a new form (subform/dialog) does not start a new appserver-session knowing this you'd better avoid such situations (for the moment) I'm not sure if a separate session would solve that problem reinhard, it will, because the rollback will happen in another session and theirfor will not thorw away the billing-head instance created in the first session (which is the actual problem i think) but the newly inserted record wont'be available in the other session? or did I understand something wrong reinhard, no it won't but as i see things it does not have to be visible there are there any plan to improve the developer's guide? ;) yes (to aries) the problem is quite tricky though; just imagine to 'modify' a customer in the sub-form .. you won't see these changes in the first session (which would be the downside of havine separate sessions) does resetFroeignKey requery the table? kilo, yes, but this does not bring in data modified by the other session (due to serialization) ??? the resetForeignKey called after subform closed hmm, but maybe i'm wrong ... depends on when the transaction started in the first session btami, migth work ... btami, but it's quite theoretical since we do not have a different session in the subform :) but we will or not :) problem is the 'rollback' throwing away all 'fresh' data from the master-form if you do not rollback in the subform it should work (on a commit for example) the fix i did a few weeks ago was regarding commit (by the way :) ok, i'm off for lunch now ... *** sjc has joined #gnuenterprise * ajmitch sees yet more gnue packages to put together :) ajmitch: yeah, common and forms is out reinhard: ok, thanks this'll be the last common release for awhile? aries: however, it's a matter of time ajmitch: we hope so, yes ok actually we planned to let you do the splitting with this release so this one will be done & uploaded in the standard way and then split for 0.5.13-2 so that people at least have 0.5.13 in sid excellent I can do the same with forms, btw since I tested splitting them back at 0.5.5 or 0.5.7... great oooh, a whole big family of gnue debs kilo: nah, not really the question would be what to put as default yeah I would be tempted to say choose what has the least dependencies that would be sqlite and curses ui curses * kilo was late but otoh that might not be what the majority wants * ajmitch is worried that he has a ~/debian/gnu/old/new dir ;) lol still downloading.. wrt the gnue-common-split I would propose to name the packages gnue-dbdriver-psycopg, gnue-dbdriver-mysql, gnue-dbdriver-sqlite etc not gnue-common-psycopg etc that's what I was intending, actually as we *might* end up at some point with gnue-rpcdriver-xmlrpc, gnue-rpcdriver-soap etc right if someone works on the various RPC drivers yes :) I've seen some commits from siesel in the rpc directory maybe he's started again playing with soap or pyro reinhard/ajmitch: do you have any vision how gnue-packages setup will look like, and how they will be packaged ? btami: total world domination! oh sorry.. well debian is easier to do because of the usefulness of apt-get *** aries has quit IRC btami: you mean our "source tarballs" or the debian/rpm... packages? tarballs will probably just include what we have now under gnue-packages sigh, after midnight again debs would install the gld and gcd files to /usr/share/gnue/appserver/some/sub/directory and then SIGHUP the running appserver at least that's my vision of things :) if appserver is running yes exactly reinhard: write this down somewhere, I need to learn appserver properly ;) ok, i see gacvs --help-config ============================================================ Name:modulepathDefault Value:/home/reinhard/gnue/share/gnue/appserver Semicolon-separated list of paths to load modules from ============================================================ but well I see we also need an appserver adminguide besides an appserver devguide yeah and I don't know enough to write it yet off to lunch ok * ajmitch ought to go to bed, after midnight now ajmitch: go to bad! :) ajmitch: go off to lunch instead lol @ping pong *** johannesV has quit IRC *** johannesV has joined #gnuenterprise bbl *** btami has quit IRC *** mnemoc has joined #gnuenterprise *** tiredbones has joined #gnuenterprise *** mnemoc_ has quit IRC *** jamest_ has left #gnuenterprise < sorry :| reinhard: hi, i need to inject legacy data into some appserver modules. how can i generate valid gnu_id-s? reinhard: i had the hope of that been a sequence, but it is a varchar(30) .oO( why? )o mnemoc, it's a varchar(32) because we're going to use UUID's (or sometimes called GUID's) for this to inject legacy data you can use any string you'd like so "foobar_my_id" could be a valid gnue_id mnemoc: i generate gnue_ids like this: TABLENAME000..0001 and 0002 and so on, thru 000F then 0010 etc etc i'll commit a new module "UUID.py" to gnue-common today in the afternoon this will provide functions to create gnue_id's as well great :) kilo: i'll do that, thanks mnemoc: this way i can be *almost* sure they are unique mnemoc, i think this is the best way, as it makes keeping referential integrity quite easy yes I agree with what johannesV said importing several related tables can be a major PITA if you have sequencials as primary keys what form your UUIDs have? mnemoc: you can fill anything you like into gnue_id for example I import data from another app with customers where I have customer numbers mnemoc, i've implemented all 4 types (timebase, random, namebased md5/sha1) reinhard: yes, i'll use $TABLE$LOOP$COUNTER, but i wanted to know what i should not write there and I generate gnue_id's like VCS_IMPORT_004711 (e.g. for customer number 004711) following the draft of Leach/Salz (or rpc-rfc) ic mnemoc: the generated gnue_id'x will be 32 char hex numbers (128 bit numbers written as hex) so everything that's not a valid hex number will be sure to not overlap with a generated id thanks, i'll play with that so dont name a table ABBA 8-)) johannesV: implemented where? kilo: *G* mnemoc, it will be commited this afternoon ... kilo: or deadbeef lol bbl *** kilo has quit IRC In directory "/gnue-common//packaging/debian" there are two files; rules and README.cvs, are these still valid? tiredbones: you want to build debian packages yourself? we will have our new releases in sid within a day or two ajmitch now being official debian developer No! I'm looking for the package Gettext. In trying to install gnue-common I get an error that Gettext needs to be install. I can find the dam thing any place. *** btami has joined #gnuenterprise Correction "can" should have been "can't" in my last message. *** jamest has joined #gnuenterprise Is Gettext only used with PHP? brb phone *** apauley has quit IRC tiredbones: you try to install our debian package, from our release, or from svn? I used svn. so you would run setup-cvs.py instead of setup.py any specific reason why you don't stick to released version? *** johannesV has quit IRC Sure, Would you perfer that I use the tarballs? I just happen to come across some material that suggest one to use svn. depends on what you want to do if you just want to start with playing a little around *** ajmitch has quit IRC I would really recommend using the debian packages (provided you run debian) Should I scrach everthing and download the tarballs? *** ajmitch has joined #gnuenterprise Can you give me a source where the deb packages are? *** ncjp_ has joined #gnuenterprise in sid and sarge (sid is newer) *** kilo has joined #gnuenterprise *** jamest has quit IRC *** johannesV has joined #gnuenterprise *** SachaS has joined #gnuenterprise *** jcater has joined #gnuenterprise *** ncjp has quit IRC mnemoc, if you svn update you should be able to use "gnue.common.utils.uuid" to generate new gnue_id's johannesV: do you know a friendly python module to read dbf files? mnemoc, gnue-common has a dbdriver for reading dbf-files johannesV: currently i'm using clip (www.itk.ru) for my migration, but i would like to do it on python and learn johannesV: oh! so you should be able to do something like "provider = dbf" in your connections conf * johannesV has to look into drivers ... can i mix two connections on a gnue app? sure .oO( gnue seems to be much powerfull than what i found on the docs )o johannesV: can you add _that_ missing dbf.py to external ? mnemoc: another solution can be, if you have a win box, you can use adodbapi dbdriver to read .dbf files and use psycopg or pypgsql driver for postgres missing dbf.py? you can test pg 8.0 on win32 now :) http://www.fiby.at/dbfpy.html <--- is this what i need to install then? btami, i think dbf.py is not only this single file, but also binnum.py and strutil.py mnemoc: the gnu-common's dbf driver starts with "import dbf" mnemoc, yes this should work (at least it was this version i've used for changing the common's dbf driver) can't we put them into common's external dir ? *** titopbs has joined #gnuenterprise *** titopbs has joined #gnuenterprise johannesV: i think this doesn't needs binnum.py nor strutil.py http://www.gnuenterprise.org/~btami/dbf.py it was in the gnue tree at r3516 don't know why it was deleted later... btami, i think this was another module which wasn't working svn co -r 3516 svn://svn.gnuenterprise.org/var/svn/gnue/trunk/gnue-common/src/dbdrivers/dbf dbf don't know at least not with common's dbdriver never tried it i think using the fiby-module is best choice atm i'm debating w/ myself on porting that to dbfpy as have a (very) small in house app that needs to write dbf the code works fine but i'd be nice to have it consistant w/ my other apps which all use common hmmm, dbf.py is readonly :( right dbfpy has experimental support to write which has worked ok for me for a long time cool but i don't know about updating or if it'd force a rewrite of the whole dbf file (that's what I do now as it's only generating 2000 record files ) so, you just have to rewrite the dbf dbdriver using dbfpy :) yep btami, you can try if it's working with that dbf.py module (as i do not use it anyway :) maybe mnemoc might volunteer for testing ... mnemoc: anyhow, we tried to convert .dbf files into firebird with adodbapi/kinterbasdb, and it worked well after we discovered that something is wrong... yes, kilo knows more :) he made the conversion but tha wrong thing was the foxpro driver .oO btami: is it too hard to use gnue-common to do this convertion? mnemoc: no, not hard at all * mnemoc needs to learn Python * kilo loves downloading Win2K SP... 132 MB... pie thong *** holycow has joined #gnuenterprise *** jcater has quit IRC *** jcater has joined #gnuenterprise does any body knows how i can make the cartesian product of lists values in python?? how do you maka cartesian product? ie i have ['fr', 'en', 'es'], ['titi', 'lolo'] and i want ['fr', 'tit'], ['fr', 'lolo'], ['en', 'titi'],... i found a module called probstat but it has a bad memory leak under 2.3.4 and it doesn't seem to be maintained anymore...:( zip() nickr: it doesn't seem to be what i need would you share an example... oh I see your problem I misread your example. :) http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/302478 Title: Generating combinations of objects from multiple sequences; Submitter: David Klaffenbach; Updated: 2004/08/29; Version: 1.0; Category: Algorithms; Description: The function combine takes multiple sequences and creates a list in which each item is constructed from items from each input sequence, and all possible combinations are created. If that description is confusing, look at the example in the docstring. It's a pretty simple transformation. The function xcombine is similar, but returns a generator rather than creating the output all at once. * jcater bitchslaps bigbrother jcater: thx for the pointer @more that was pretty cool of bigbrother you should be nicer to him jcater jcater: seems the qt uidriver's time will come in this year http://www.trolltech.com/newsroom/announcements/00000192.html yes :) * jcater saw that this morning I got quite excited maybe you pla to rewrite gue-designer in qt ? I'll be switching to 100% QT in-house s/pla/plan btami: I wouldn't rule that out honestly I would prefer a ui independent designer :) but yes, as a first step, it would make sense to finish forms qt uidriver that's a much harder task than the ui independent client tools designer is rather complex stuff hmmm qt driver is rather broken now I was using it at one point jcater: really? at one time, the only issue I had was getting the x,y coordinates right it was always off at the top only major issue there's always issues :) we didn't change much reinhard: I'm sure I broke it knowing me :) and IIRC there weren't too many incompatible changes in the framework that would break existing drivers if any at all but in fact I would be happy about a working qt uidriver interesting announcement well, as of this weekend, I'm living alone again last time I lived alone is when I wrote designer and reports so maybe I'll have another coding spree soon :) as the one I'm apporaching as my first potential paying gnue customer is a kde user *** wendall911 has joined #gnuenterprise *** allen has joined #gnuenterprise ping *** allen has quit IRC i've looked at qt once and it looked very ugly ... (since i'm using gtk2 ... :) qt devs say the same thing about gtk2 funnily enough :) hehe *lol* *** johannesV has quit IRC *** jamest has joined #gnuenterprise *** btami has quit IRC *** titopbs has quit IRC On the Community WiKi page there are instruction for set up PostgreSQL are still valid? no is anybody here with write access to that wiki? *please* remove this document tiredbones: please ignore that document, it applies to appserver 0.0.6 where now we are at 0.4 (or at least close to it) tiredbones: did you install the debian packages now? Well, I installed using deb packages from sid and now I'm looking for for doc. ok you want to use postgres? please make sure you have the debian package python-psycopg installed also make sure that postgres accepts tcp connections from localhost When I run the "find" command on anything gnue, I don't see anything that point to some docs. you can test that with "psql -h localhost template1" there should be docs in /usr/share/doc/gnue-* setting up the database requires some manual steps for now, but we are actively working on removing the need for those steps (for this reason we didn't document them all too well - they will be obsolete soon) also, we generally suck at writing docs the only thing I find in /usr/share/doc are dtds. hi, you have dbf provider for migration purposes. how should i define schemas to not get .dbf files modified? i would like to migrate from dbf to appserver using GNUe stuff tiredbones: you should find samples there, too what debian packages did you install? mnemoc: you have an existing application accessing the dbf, and you can modify that app? in that case I would recommend writing an export to .gsd format in the old app My connection to postgres is ok.Its been awhile since I installed that. I installed all of them. tiredbones: please be aware that there are 2 ways of accessing postgres - sockets and tcp. default is only sockets are allowed tiredbones: then you should have a /usr/share/doc/gnue-appserver with an examples subdirectory and much more reinhard: and after i have a 1:1 .gsd describing those .dbf? mnemoc: wait, not sure if you can generate a gsd... reinhard: i can write it by hand the gsd? reinhard: gsd=xml file describing each dbf structure?, yes reinhard: or you mean data? both oh can't be done using the provider? *** titopbs has joined #gnuenterprise mnemoc: I think this would very much be the target of gnue-integrator ... of which not much more than a README file exists... reinhard: :\ what I would do in your place is try to somehow export the data from the dbf tables then with some sed, awk and other magic create a gsd from that should i match appserver structure on that gsd file or can that be done procedurally? s/structure/schema/ mnemoc: not sure I understand that question .dbf has one 'schema', appserver has another tiredbones: you have installed python-psycopg and tested postgres with tcp? tiredbones: then you should be able to run can i generate a .gsd using dbf schema, process it and inject it to appserver or i do need to generate a .gsd with modules' schema? gnue-setupdb reinhard: what's the easier/safer way to go? mnemoc: what I would do (again YMMV) is create the appserver schema by writing gcd files testing that all with demo data err test data well you know what i mean... SachaS: could you send me your (obsoleted) PDF's OOo source? yes :) and if it works import your old data using the new schema i would create a gsd with *only* data reinhard: I'll go check those out, thanks as appserver will care about db schema anyway tiredbones: if you have questions *please* ask our documentation sucks so hard it *really* hurts that's why SachaS once wrote that postgres install document I have python-psycopg-1.1.81-1. but then again he had to learn it's impossible for any documentation writer to keep up with johannes' development speed tiredbones: the version shouldn't matter actually kilo yes. what is your email? can you run gnue-setupdb ? SachaS: sorry, I hope it's ok for you to officially mark your doc as obsoleted do you set it obsolete? but currently it seems like it does more harm than good :( thats cool we can even remove it thats ok. with me * reinhard hasn't even access to the wiki SachaS: I would *love* to see it updated ok so i will go and search for the user password yeah me too :) or otherwise any contribution of you to our documentation as what you did there really rocked but it's amazing what happened to appserver in the last 6 months kilo sent whoever updates a parts should also update docs to suit looking back I can only shake my head SachaS: thx how did johannes ever find the time to implement all this :) reinhard: how can i make readgcd drop obsolete relations? I think SachS doc was great to bad it was out of date. Sorry for spelling your name wrong SachaS reinhard: will take a look at wiki now *** SachaS_ has joined #gnuenterprise archivist: that's easier said than done. for example it would mean that people not firm enough in English to write proper docs are not allowed to code... mnemoc: readgcd does not drop anything on purpose otherwise it could happen that accidentally reading e.g. a corrupted or incomplete gcd would make you loose data ic so readgcd only *adds* but does not *delete* from the db reinhard: how do you handle schema changes then? usually I leave the old column there and when I've checked *at least* 3 times that everything has moved correctly I remove the old column manually (with psql) of course while I'm still in development and only using test data anyway I normally export the most important data to gsd (you can do that from appserver with gnue-gsdgen) drop the db completely and let appserver rebuild it from scratch * reinhard never would have imagined that it's so much work to write a perfect system that fits all needs of the world ;-) reinhard do u have a production gnue env? :) SachaS_: yes, the good old hotline application reinhard: does Appserver make good whisky? kilo: yes, but I didn't GPL *that* module on purpose reinhard when did upgrade prod. last? ;-) * archivist needs beer http://www.gnuenterprise.org/cgi-bin/twiki/view/Main/UserMANUALs SachaS_: I run hotline with appserver svn version of today reinhard: non-GPL is ok, i only ask for a little share 8-))) ta kilo kilo: thank you very much for updating the wiki sniff sniff :) so now you have no manual new users may as well give up edit and update dont through! throw even *sigh* I think I could write documentation for a week or so and still it wouldn't be somewhere near complete archivist: SachaS has just sent me the source for that pdf, i promise to produce something useable from that SachaS: where did you sned it? archivist: you are still entitled to ask here whatever you want or need to know who does the releases for windows? and we *do* accept *every* piece of documentation with *kisses* kilo freemail (well, or without kisses, if you prefer;-)) kg_kilo freemail mnemoc: btami asking here is not much use when my box is at home 7 miles away yep kilo: thanks :) SachaS: ah, ok, thx cd / ls bin boot etc home porn sorry for that noise. hmm is that home and porn or home\ porn ? *G* jcater: well, I always have to make that a separate partition ;-) *** SachaS has quit IRC when will hat fairy come and tell me I have three wishes? 1. a complete set of current documentation for gnue (plus translations of course) 2. a second instance of johannesV she's gone forever. haven't you seen Shrek 2 ??? 3. a few customers paying me something for all this hacking :) and according to Nietzsche, God is also dead... your chances are not looking good... kilo: god is not dead. he's just concentrating his efforts on a less ambitious project ;-) to put it clear: I am *not* attempting to collect all knowhow about appserver within bytewise and then sell it for a lot of money although ... I am starting to like that idea ... j/k and then it will be called ByteWi$e... rofl reinhard for world domination... I guess I would have to share with jcater -- I am completely helpless when it comes to reports don't you want to register in CIA? they probably keep an eye on me anyway I mean I use encrypted mail, and I have been in Amsterdam... if you dominate, you dont have to make reports... kilo: do you only like whisky, or do you also drink southern comfort? I had a user at a customer who used to chat via icq with a guy named osama... lol they even talked on the phone I still have to lol when I think that in the hurd, the first process to load the kernel and all other processes is named /bin/laden laden == to load in german and marcus brinkmann (main hurd developer) is german and, as a side note, you can meet lots of hurd people at fosdem this year reinhard: As aonais uisge-beatha tha mi nam neoni. Without whisky I am but nothing. why do you ask? * reinhard loves celtic/gaelic -- it reminds of the LOTR * reinhard is drinking southern comfort right now aye, Tolkien was a linugisiticisiitctic (southern comfort is a liqueur(sp?) made of whisky in case somebody doesn't know) ... and it reminds of enya kilo: do you know enya? did you know that about 86% of Ireland's milk production is spent on making whisky liqueur like Baileys * reinhard feels he is getting *way* offtopic kilo: lol I dont know her but would love to 8-))) anyway I love her music kilo: I love baileys I always drink baileys in remembrance of our former debian maintainer ;-) come to Szeged, we have one opened and nobody drinks it... kilo: don't tease me, I *will* come to szeged at some point in time you know I have a sister-in-law from debrecen i know i know and I've been in budapest twice already what about a meeting in Vienna about the same distance for you and us virtually at any time you want well... not really... but you know what I mean ;-) I will be in vienna several times over the next few months I'll let you know as soon as I know the dates ok for me * reinhard still wonders why there isn't a photo neither of btami nor of kilo in the web and there is a good whisky shop there... When I run "gnue-appserver" , the last traceback error is- LookupError: unknown encoding: eg_gb. What am i missing?\ sorry- that should have been en_gb tiredbones: with that case? erm try to run gnue-appserver with reinhard: yes LC_ALL=C gnue-appserver it seems there are some bugs in python wrt encoding with some locale, python returns completely bogus encodings ***************** **** johannes ****** ***************** if you read the logs: might be something worth to track down how can python come to an *encoding* of en_gb ----- No matter how I enter it, I get the same error. would it help if I copy the traceback here? no what does "locale" issue for you? reinhard: LC_ALL=, all other values are en_US LANG=C gnue-appserver what does that give? same error still with en_gb? can you please check if you installed the "locales" package? (please notice we're in high debugging mode now...) (means I'm not telling that you did *anything* wrong but trying to find out where our mistake lies) oooohhhhhhh --- commit from siesel... I know. I just hope this info helps you folks. Yes the error is still en_gb. tiredbones: it will help, I promise in fact I won't give up till I found it :) did you check the "locales" package? should I test these debs before uploading them? :) reinhard: If I type locale and get a reply doesn't that imply that locale is installes? ajmitch: no need to test, what are users good for? j/k ajmitch: I can test (but not today) if you upload to ~ajmitch/debs I feel wanted now. hmm however there was no big change, so I don't think there is an urgent *need* to test tiredbones: no, the "locale" command is part of the libc6 package forms complaining about a gfd containing tab chars which you are *very* likely to have installed *** lekma has quit IRC ajmitch: tab characters are invalid in forms for ages but there might be sample forms that are outdated for ages... * ajmitch should test with a form that isn't one he wrote ;) ah ok :) well there's only one thing we suck more than in writing documentation in providing up to date samples :(( heh sup dudes? * ajmitch goes to upload common & forms chillywilly: you were chosen to write a complete up to date documentation for gnue including install from tarball, install from svn, and configuration chillywilly: thanks for volunteering HAH fat chance! chillywilly: sorry * reinhard blames the bottle of southern comfort standing beside him well, it's carneval here, isn't it? * reinhard really loves "don't drink and su(1)" reinhard: chillywilly said he was going to do it for sure the other day, better check the logs and not let him off the hook on this one chillywilly: how'd ice fishing go? *lol* * archivist has independent logs for adjudication ummm, it was ok but we did not catch anything chillywilly: you *really* did ice fishing? but scotty had fun so I guess that is all that matters and you got a chilly willy? reinhard: yes no, my willy stayed toasty if you must know tiredbones: did you check locales package? I'm still trying to find what deb package its in. locales is a deb package sorry to not put that clear I would assume so. How else would it get installed on a sarge system? I am as debian addict as can be, so always expect me to talk in debian package names "locales" *is * the name of the debian package apt-get install locales well, if that the case dpkg -l locale - no find. I guess I'll install it now yes and then please tell me if that changes anything *** wendall911 has quit IRC I should have added the "s" to locale. The package is installed. you mean it already was installed before? "don't drink and su(1)" :) Yes, it was already install before. fsck I argee! you run sarge or sid? I run sarge, but I pulled the gnue stuff from sid. that's about exactly the same configuration as I have except that it works perfectly for me even if I set LANG=en_US all your output of "locale" is en_US? not en_US.something? Should the version numbers on gnue deb pkg match that of the tarballs on the web site? no, they are a bit behind ajmitch just is about to upload the new versions (we had a release today) reinhard: already done tiredbones: but there has been *no* change that could influence this behaviour wrt locales All the output from locale is en_US ok we get into even deeper debug mode now (you will earn a lifetime mention in our THANKS file for this) please enter Don't let me waste your time. * ajmitch doesn't think he's mentioned anywhere in the thanks ;) ajmitch: you aren't :( tiredbones: just a sec reinhard: that's ok, I don't need to be tiredbones: please call python THANKS ajmitch and then enter >>> import locale >>> locale.getdefaultlocale() what does that give you? a dram o' whisky? hmmm... my wife just came home ... will be in and out you or she? 8-))) chillywilly: :P the output is "en_US, "en_gb" grrr ... for me itĂ„s ['en_US', 'ISO8859-1'] I don't get it :(( From this output it looks like I doing something wrong. I enter gnue-appserver and get about a page of errors, with the last error being what I reported on. http://rafb.net/paste :) chillywilly: what does http://rafb.net/paste imply? *** jcater has quit IRC paste the trace back in there and give the link ;) *** kilo has quit IRC *** kilo has joined #gnuenterprise chillywilly: did it work? *** SachaS_ has quit IRC gacvs==gnue-appserver? mnemoc: yes gfcvs too? that equals gnue-forms ok :) gncvs is gnue-navigator, gdes is designer grcvs is gnue-reports etc etc i prefer current names :) gacvs is the svn version of gnue-appserver you can use the released version for production and the svn version for testing/development on the same machine ic what --connection means to appserver? --database behaves like --connection and --connection fails to read connections.conf :\ --database points to an entry in connections.conf. and --connection? --connections (NB with s in the end) points to the connection.conf file location oh lalalala everybody sunshine *** wendall911 has joined #gnuenterprise *** jcater has joined #gnuenterprise No section: 'appserver' <--- what does this mean when you actualy have an appserver section on your connection.conf? (the one that comes on sample.connections.conf) tiredbones: I think we have to debug more tomorrow tiredbones: is that acceptable to you? it was the lack of gnue.conf :) mnemoc: good you found out ... too much southern comfort to be productive any more :) and sabine has come home, so need to spend some tiime with her off for today night all see you all tomorrow appserver-devguide said that only with: gnue-readgcd address.gcd i can gnue-forms appserver://appserver/form/address_person, is that true with last appserver? tiredbones: I will keep on debugging tomorrow if you still have enough patience mnemoc: yes night all *** reinhard has quit IRC reinhard: take all the time you. mnemoc: yes, true *** jamest has left #gnuenterprise .oO( we doesn't work then )o why* gnue-readgcd run without any errors? yep then you started gnue-appserver, right? start it with the -Z option so that it wont demonize itslef yes --loglevel ? no need for logleve and then in another terminal you start gnue-forms appserver://appserver/form/address_person if you connection name IS appserver s/you/your gnue-forms appserver://appserver/form/address_person works but gnue-forms appserver://appserver/form/region_region don't :( re-readgcd is safe? yes but stop appserver before doing that ok worked :\ thanks :) look at the file sample.gcd in gnue-appserver/samples there is no region module or region class in there region_region is mine ah, ok, sorry don't owrry i should said it before :) you may want to look at gnue-packages then yes, i'll do and gnue-contrib is also worth it brb :) good night *** kilo has quit IRC *** sjc has quit IRC *** titopbs has quit IRC *** jcater_ has joined #gnuenterprise *** holycow has quit IRC *** holycow has joined #gnuenterprise *** jamest has joined #gnuenterprise *** holycow has quit IRC *** ccfiel has joined #gnuenterprise *** ccfiel has quit IRC *** jcater__ has joined #gnuenterprise *** jcater_ has quit IRC *** wendall911 has quit IRC *** titopbs has joined #gnuenterprise *** jamest has quit IRC *** jcater_ has quit IRC *** titopbs has quit IRC