*** holycow has quit IRC *** Morphous has joined #gnuenterprise *** Amorphous has quit IRC *** johnf has quit IRC *** SachaS has joined #gnuenterprise *** reinhard has joined #gnuenterprise hallo reinhard *** thierry_ has joined #gnuenterprise *** thierry has quit IRC uploaded latest appserver application dev. manual can i create a table of contents with OO ? nevermind. SachaS: is it possible that calling a procedure from a form trigger *always* returns a list? reinhard: when i tested with string, it was a string. i will check that again. eg, when i was returning a string, it returned a string, i thought. i try again. reinhard: i was wrong the string example returns a list as well. ok i understand that then i think i found the error yeah returning a boolean, returns a list with a boolean inside SachaS: fixed thanks. now i have to change all forms again ;) thanks very much for that ;) haha no great. :) jk reinhard: do i understand the idea of a list property right? example: i want a customer business class with a list property this list property has all customerOrders of this customer. could that be an example of a list property? reinhard: when will you release 0.0.7 ? ;) SachaS: you understand list properties absolutely correct currently, to achieve that, you have to do find ('customerOrder', conditions = {'customer': self.gnue_id}) if the list property would be named "orders" then self.orders would be a shortcut for exactly the above so to speak list properties are "canned" find statements ok. great. so not sooo difficult for you to add it ;) eventhough its intended for 0.2.0 yeah it's not so high on priority because it's just a shortcut ok. you can do anything without list properties as well as to 0.0.7 i want to finish this before: * gsd's should also support table extension, not only table creation * some dbdrivers have issues with time fields (that is time without date) after that is fixed, i'd like to do 0.5.4 of common/forms/designer and the james bond release of appserver :) immediately after this we will most probably do bound procedures and calculated fields in appserver james bond release haha cool. table extensions... and then i think we have essentially what we need for 0.1.0 of appserver what is that? that is a gsd that you apply on an existing database like oh ok. you have a production database with several tables you want to add a module you didn't use before yes :) that module adds columns to existing tables ok true yes and that not only for modules but also classes etc basically, that you dont have to drop the database first...and start from scratch? i think i did interrupt you ... sorry. james bond 007 :) must be an insider joke of release managers haha no you didn't interrupt me the idea is that with gsd files you should not only be able to add tables to your database but also to add columns to existing tables weather is so good today must go work in the garden *** rm-garden has quit IRC *** lxf has joined #gnuenterprise *** btami has joined #gnuenterprise hi all i'v just uploaded an unofficial setup .exe from svn snapshot http://www.gnuenterprise.org/~btami/Gnue-Tools-0.5.3.99-svn5545.exe please try/test it! it was created with py2exe and enables adding additional drivers/pligins under the bin directory s/pligin/plugin *** reinhard has joined #gnuenterprise *** _florin_ has joined #gnuenterprise hi reinhard hi btami read about the windows exe's that's great i think i will test thx although i do not like windows very much ;-) me too :) <_florin_> hi everybody i'v changet 5 box from win to UHU-Linux in last months changed even and i'm a Linux fan now :) but i know that my customers use win boxes, and if i want to change our foxpro(dos) app to gnu i have a good win32 client sorry for my poor english and typing, sometimes i lose words/letters :) btami: no problem :) <_florin_> does anybody know what this svn-error means: svn: subversion/libsvn_ra_svn/marshal.c:434: vwrite_tuple: Assertion `opt || cstr' failed. <_florin_> Aborted btami: it's the same for me _florin_: please remove the directory gnue-common/src/logic and try again seems like subversion is getting along too well with directory renaming s/is/isn't/ btami: it's the same for me, i also need frontends for windows yes, this is the only reason i deal with win32 ui+packaging <_florin_> thanks reinhard, seems to work now... :-) * btami not a mazochist btami: :) btami: the installation program looks highly professional it's Inno, not me :) a (really) minor nitpick: my understanding is that for *installing* and *using* gnue you don't have to agree to the GPL for modification and redistribution you have but i'm not sure anyway it was inherited from old installer stuff, made by jcater or jamest ok so it's probably correct this way i just added the connection.gfd (from jan) and more (db)drivers btw. jason didn't liked that gfd to much... i plan to enhance it with a "test connection" button reinhard: sorry for asking, but i'v forgot, have you fixed the "where unicode bug" ? i mean the "update ... where ..." *** jamest has joined #gnuenterprise you know, Petõfi :) btami: i was not able to reproduce it hmm, ok btami: the connection.gfd doesn't work for me on win32 ? win me is that win32? yes :) ok :) was not sure if i need nt/2000/xp win32 = all win from 95 ok thanks what error you get? global name "GDataObjects" is not defined hmm, other forms(intro.gfd) are OK? let me try intro works but segfaults on closing sigh i don't like win me which provider does (should) work for postgres? *** lekma has joined #gnuenterprise psycopg and pyPgSql ok with zipcode sample the login dialog is not visible on the screen and i get a traceback with "no error message available" but when i type in username and password for the database blindly i get the form and it works on the other hand appserver doesn't run maybe there are dependencies missing wrt xmlrpc it's the win32 ui or wx? erm just a second only happens for win32 ui i'v put defaultui = win32 in gnue.conf wx ui everything works fine login dialog works in wx ok, thx and no traceback in wx and no segfault on closing the form in wx btami: i hear the 0.5.3 release of forms is broken for windows? dropdowns? l8r l8r *** reinhard has quit IRC *** jamest has left #gnuenterprise *** jamest has joined #gnuenterprise *** btami_ has joined #gnuenterprise hey btami_ you get my earlier question? *** _florin_ has quit IRC *** btami has quit IRC *** kingpin2k has left #gnuenterprise *** kingpin2k has joined #gnuenterprise *** lxf has quit IRC *** jcater has joined #gnuenterprise *** _florin_ has joined #gnuenterprise *** reinhard has joined #gnuenterprise jcater: how is that internet connection coming :) horrible reinhard: are you here? and taking appserver bug reports? yes no j/k sure :) from svn a few days ago Should all the modules match - forms,report etc? probably would be best given the issues we've hit :) OK I'll try it out - be back in a few BTW every time I use "./setup.py install --prefix" I have be root - is this right yes OK I get "no previously-included directories found mathcing 'etc/.svn' and 'doc/.svn' I all OK? hmm did it continue? look like it installed? I think those can be ignored ok I get "importerror: cannot import name Connection" what the crack talk about messed up buffers DB000: File "/home/dneighbo/cvs/gnue/gnue-common/.cvsdevelbase/gnue/common/datasources/drivers/DBSIG2/Connection.py", line 98, in makecursor DB000: cursor.execute (s) DB000: psycopg.ProgrammingError: ERROR: Relation "gnue_parameter" does not exist DB000: DB000: SELECT gnue_name,gnue_procedure,gnue_length,gnue_id,gnue_scale,gnue_type FROM gnue_parameter WHERE ((gnue_procedure = '0000000000000000000000000000101E')) ORDER BY gnue_id dneighbo@latitude:~/cvs/gnue$ there we go ok i get that when running the sample first "svn from a few days ago" is a generation when talking of appserver :) and gnue_parameter doesnt exist i reloaded the .sql file even i recommend updating to current svn anyway i am reupping svn as we speak and trying again you should in samples dir make postgresql to regenerate the sql script for postgres hmmm svn: Obstructed update svn: failed to delete dir 'gnue-common/src/logic': local mods found within. svn: Left locally modified or unversioned files svn: dneighbo@latitude:~/cvs/gnue$ any ideas on this one? you have to delete that directory shoudl i delete that dir ok that seems to be a bug in svn seems svn can't handle directory renames properly at least not in that case problem is that directory has non-svn files in it so it is hesitant to delete it which sucks, imho ah i.e., the *.pyc files bb in 15 minutes watching news on tv *** lekma has left #gnuenterprise dneighbo@latitude:~/cvs/gnue$ ~/bin/gacvs Exporting our services via xmlrpc (port 8765) ... ... GNUe Application Server up and running ... dneighbo@latitude:~/cvs/gnue$ wondering if appserver died holy error from hell it popped the error in a dialog box so i cant cut and paste i THINK i gave it bad user name and pasword and that is a HELL of an error is it in the background? iirc, you have to pass it -Z (or --no-detach) to keep it running in the foreground http://www.gnuenterprise.org/~dneighbo/screenshots/appservererror.png that is svn from 5 minutes ago with a new make postgresql and a new setup-pgsql.sh running appserver then trying svn of gfcvs against sample.gfd jcater: are there any good tutorials out there on gnue-reports? basically i use kmusicdb to catalog all my music so i have a large database of music in postgres its report tool sucks i wanted to generate some reports with gnue-reports figured it would be a good way to help remember how to use it (as well as test it) if there is good docs started i can add to them if there are not good docs started i would like to help :) not really or at least none I know of is the report file (grd) still all hand created? and what samples work i.e. if i can get a working sample i think i could probably get a working report and some documentation foobulations works most of my use of gnue reports is RTF merges any non rtf merges? and I do have a lot of those now for me personally, no but jamest does and the examples work afaik as in this case i think i want an html output i.e. i want to generate a list of my collection for a webpagae and generate my "wantlist" for a webpage * derek goes to dig you around for questions? dood we have foobulations in chinese that is freaking insane :) back dneighbo@latitude:~/cvs/gnue/gnue-reports/samples/foobulations$ ~/bin/grcvs monthly.grd --filter html -d mytest.html subtitle="Foo Bar" startdate="04/10/2000" enddate="05/20/2004" The use of the "common.rpc.GComm" module is deprecated. Please use the "common.rpc.server" or "common.rpc.client" module instead. GNUe Reports Client Version 0.1.4.99 GNUe Common Version 0.5.3.99 Error: Parameter "--filter" specified, but no value supplied. For help, type: gnue-reports --help dneighbo@latitude:~/cvs/gnue/gnue-reports/samples/foobulations$ derek: you said wrong password error does that mean it worked when you provided correct password? derek: all flags must come before all parameters this isn't dos =) jcater: i fixed the error i.e. i committed a file wee derek commits code!! syntax is: gnue-reports [options] [param1=foo] [param2=foo2] ? ah, the warning of GRPC? ;) hmm from gnue.common.rpc import GComm had to be changed to isn't that one of the signs of the Armaggedon? from gnue.common.rpc import client in GRRun.py cool our parameters are broken i can do -d filename.html ? it is forcing me to do -d=filename.html no it should either be -d filename or --longname=filename not broken you are doing wrong syntax imho report name must follow all options you are misinterpreting the error look back at your command line ok then look at my syntax I posted dneighbo@latitude:~/cvs/gnue/gnue-reports/samples/foobulations$ ~/bin/grcvs monthly.grd --filter=html -d=mytest.html subtitle="Foo Bar" startdate="04/10/2000" enddate="05/20/2004" ************************************************************ Attempting to log into "GNUe Test Database" (gnue_reports): User Name: works i flipped things around and it doesnt error my bad funny it actually works the other way if you put the = in there just not well :) ************************************************************ Unable to load PySablot we still rely on sablot? if so do we have packages somewhere for python 2.3 you aren't listening ~/bin/grcvs monthly.grd that's all the options you passed everything after the report name is considered a parameter, not an option i.e., same as ~/bin/grcvs monthly.grd month=01-dec-03 syntax is: gnue-reports [options] [param1=foo] [param2=foo2] ~/bin/grcvs --filter html -d mytest.htmlmonthly.grd subtitle="Foo Bar" startdate="04/10/2000" enddate="05/20/2004" is what you want err well, with a space before monthly.grd of course yeah i misread the amle i have it working other than it complaining about pysablot derek commited code? I'd change your report-filters.conf /msg jcater i thought we revoked his access! to use tabulator instead of sablotron derek: wrt error message can you please try again as a second exception that happened in the except block obscured the real error message tabulator? derek: and, btw, you might have noticed that tracebacks are now *correctly* transferred from the server to the client :) hmmmm DB000: File "/home/dneighbo/cvs/gnue/gnue-common/.cvsdevelbase/gnue/reports/adapters/filters/SimpleTabulation/tabulator/Adapter.py", line 68, in close DB000: raise "Filter configuration file is missing the 'formatter=' parameter" DB000: Filter configuration file is missing the 'formatter=' parameter dneighbo@latitude:~/cvs/gnue/gnue-reports/samples/foobulations$ look at the sample i must have an older style report filter reinhard: yes i like they are transfered to the client, just no way to easily paste here :) reinhard: when you say "try again" do mean update svn? yes derek: fwiw, you might want to run with --debug-level 0 as I think all exceptions are relayed to the console too then we're still fine-tuning how errors are given to users yes personally, I want a message to pop up that sayus "Stop doing that, dummy" but jamest protested jcater: using the sample report filter and get following jcater: i'm even seriously thinking about a "short - detail" dialog box Oh yeah.....I can feel it Error processing css file: /home/dneighbo/gnue/share/gnue/filters/SimpleTabulation/tabulator.css dneighbo@latitude:~/cvs/gnue/gnue-reports/samples/foobulations$ reinhard: reupping cvs er svn jcater: like "Error opening forms: DriverError: No such database [Detail] and when you click "Detail" button, it gives you the traceback reinhard: yes exactly what I want to see (just have had bigger things to work on :( ) jcater: i also had another crazy idea while showering today I'd like some sort of bug reporting system worked intot hat too jcater: excellent idea and a "Save to file" or "Copy to clipboard" option now my idea: we could write a dbdriver that encapsulates the "plugin.list()" function (read-only of course) so you could create a form that lists all your available plugins of a given type reinhard: after update asources/drivers/postgresql/Base/Connection.py", line 100, in connect DB000: self.sql ("SET CLIENT_ENCODING TO '%s'" % self._pg_encoding) DB000: File "/home/dneighbo/cvs/gnue/gnue-common/.cvsdevelbase/gnue/common/datasources/drivers/DBSIG2/Connection.py", line 117, in sql DB000: cursor.close () DB000: psycopg.InterfaceError: already closed dneighbo@latitude:~/cvs/gnue$ when trying to run ~/bin/gacvs oops jcater: any ideas on the reports error message? (i.e. the css issue) give me a minute please :) did anyone get my mail on the ML about the hotfix? i've sent it from home and from work and not seen it yet jcater: it sesems to have created the mytest.html file jamest: i have NOT gotten anything BUT i dont generally get the mails until a few hours after everyone else i sent the first 1 yesterday the second about 5 hours ago its on gnu time you have seen gnu enterprise and gnu hurd you still expect mail to be delivered in a timely fashion ;) derek: ok, fixed my fix :) reinhard: how is that fix going? as it seems to have broken reports :) yes never commit something witout prior testing :( should be fixed again jcater: ROCK got my first ARTIST report off gnue-reports about 10 second hack of monthly.grd i think i can do a tutorial on this reinhard: testing appserver then i have to take family to lunch *** sjc has joined #gnuenterprise derek: this fix will only fix the exception handling that is you should get a traceback again but that should point to the real problem source this time appserver WORKS it works it works * derek jumps up and down who are you? let me try one more think er thing derek: no need to jump up and down now you're about the last one in this channel to get it to work ;-) reinhard: that's not derek j/k okay same error as before if you run gacvs w/o the -Z option it will NOT work doing withthe -Z option it does work i have to leave cant file report but you should get other traceback but you should be able to recreate derek: you can also try to add timeout = 10 in your connections.conf for [appserver] just a second i dont think tha tis the issue (btw) File "/home/dneighbo/cvs/gnue/gnue-common/.cvsdevelbase/gnue/appserver/data.py", line 299, in _createResultSet return datasource.createResultSet (condition_tree) File "/home/dneighbo/cvs/gnue/gnue-common/.cvsdevelbase/gnue/common/datasources/GDataSource.py", line 279, in createResultSet resultSet= self._dataObject.createResultSet(conditions,readOnly,sql=sql) File "/home/dneighbo/cvs/gnue/gnue-common/.cvsdevelbase/gnue/common/datasources/drivers/Base/DataObject.py", line 89, in createResultSet readOnly=readOnly, masterRecordSet=masterRecordSet, sql=sql) File "/home/dneighbo/cvs/gnue/gnue-common/.cvsdevelbase/gnue/common/datasources/drivers/DBSIG2/DataObject.py", line 185, in _createResultSet cursor = self._connection.makecursor (query) File "/home/dneighbo/cvs/gnue/gnue-common/.cvsdevelbase/gnue/common/datasources/drivers/DBSIG2/Connection.py", line 100, in makecursor (exc_type, exc_value, exc_traceback) = sys.exc_info () NameError: global name 'sys' is not defined ------------------------------------------------------------ errgh doesnt seem like a "timeout" issue to me * derek runs to lunch * reinhard goes to fix his fix that fixed his previous fix reinhard: you should test before you commit * jamest runs jamest: :P jamest: that's something that is really a problem w/ python if you have a bug in an exception handling code you don't know it until the exception really happens for example both (!) oracle drivers have a "raise" statement to indicate unavailable oracle python library that both have an invalid exception class :) and as i can't reproduce derek's error i can't even test the exception handling code sigh derek: if you read this derek: the exception again occurs in the exception handling code derek: should be fixed now derek: but the result will be (as it should have been before) that you see the real exception happening -- did anybody already test new trigger handling code? and if yes, did anybody encounter any problems? not yet i'm still putting my system back together here what did johannes do? most important thing is he encapsulated every trigger/procedure code into a function which means you can now use "return foo" in a trigger to return a value not sure if that is of any use for triggers but we need it for appserver procedures hmmm i don't think that'll hurt a thing other than that reinhard: i created a method in the procedure and tried to call the method in the procedure. i get an error message (not sure if creating a method is clever, was just a test): Error occured during method invokation: Traceback (most recent call last): SyntaxError: unqualified exec is not allowed in function 'createSalesOrder' it contains a nested function with free variables (, line 4) long ago i think forms expected return value to indicate error, but i think thats now a raise jamest: sure anyone running apache2 jamest not me. no apache2 jamest: at least, you can now do a "return" in the code to indicate a successful exit before the end of the code jamest: other than that, he tried to improve error handling and such things cool like if there is a traceback in a trigger i think it would be handy you will get the traceback handed over if you could call named triggers like functions in the trigger as the "message" of the exception Sacha: I don't understand what you mean with "created a method in a procedure" yeah i think that was no good idea :) a nested method a method inside a method. thats silly ;) Sacha: you can nest a python function in a procedure i tried this and it worked: def add (a, b): return a + b print add (12, 13) this is possible in a procedure but you can not nest a *method* within another method ok. i wonder why i got the error message then . reinhard: how about when you access business object properties within the method? *** _florin_ has quit IRC reinhard: actually ok. my procedure might not be OK. if they are using it with python 2.3 or greater they can, I think just fyi (not that they would necessarily want to :) jcater: no, they can't or did I just pull a derek :) and not pay attention to what we're talking about doh! I think I did sorry because what they define is not really a method of a python object but rather a normal "global" python function that gets a "self" parameter and therefore looks like it would be a method I was thinking in terms of forms methods not appserver just ignore me nevermind :) Sacha: of course you can access properties in methods that's the whole point of methods in 90% of the cases :) reinhard: i mean in that nested method. ah then you have to pass a "self" parameter but really as parameter that is you define def submethod (self, foo, bar): ... and then you have to call in your main code submethod (self, x, y) not self.submethod (x, y) ok. yes i see. have to fix my code first. jcater, once your here. i have this multiline text window (in the classrepository.gfd), where I have to edit the code. now the code is longer than the window, so there is the scrollbar on the right. but there is a terrible flickering and when i click inside, the cursor is somewhere x lines below. (using wx) might try with gtk2. is there a way to check the syntax of a python program (except checking the color highlightning of python mode in emacs :)? just check syntax but to not run the code. yes try to import it ok. thanks. syntax is ok then. Sacha: yes that's on our known annoyance list (the flickering/odd behavior of the multiline edit) can I get a good screenshot of the classrepository form being used? for my screenshots section of the web site one sec !config list supybot.plugins.ChannelLogger
V: flushImmediately, timestamp, noLogPrefix, rotateLogs, and filenameTimestamp nice I need to upgrade he's contimplating releasing a 0.77.1 @version Vee2d2: The current (running) version of this Supybot is 0.77.0. The newest version available online is 0.77.0. jcater, can you receive ddcsend? can you email it to me? or, wait maybe it worked ok, got it thanks ur welcome *** johnf has joined #gnuenterprise I've started using gvim after years of using joe and you'd think I was a kid in a candy store.. Hi all Anyone figure out what do I have to do for SUSE 9.0 to get the database connection to work? reinhard: when i call the procedure from the forms, then the procedure works. Sacha: sorry, i think i lost track on your problem you are still talking about the nested function? reinhard: no its fine. ah ok i have not tried it again, as its too troublesome to edit the procedure in the text widget. are you working on the procedure calling a procedure? the most annoying thing is that you can't scroll through the lines with crsr up/down no that should work now i think there was not commit. johnf: i think jamest has put a howto in our wiki yeah the scrolling is bad. ah yeah johnf: i wanted to make a joke and forgot: upgrade to debian ;) or yeah there is no scrolling Was that since last night - because yesterday they couldn't get to work Sacha: i think i committed it trunk/gnue-appserver/src/language r5551 reinhard: Fixed typo. that was it oh the fixed type ok. sorry. oh johnf: what's your problem? * reinhard goes to read the log I'm getting an error everytime I attempt to run the states.gfd. hey johnf ahh the quickfix! i've tried posting a mail to the list all day none are going thru jcater: if I do a gfcvs --help I see two -u options: one for --username and one for --interface you'll need to install the hotfix1 from the downloads page seems 0.5.3 went out w/o pygresql driver files The error is either no "Driver" or no connection either in at line 41, or 42 depending on which driver I use. jamest: it's the typo in the pygresql driver, right? reinhard: it's completely missing only an __init__.py file Jamest - are you talking to me johnf jamest: what? reinhard: download common 0.5.3 tarball johnf: did you get that hotfix1 from the web site? NO I have not - I guess I'll do that now datasources/drivers/postgresql/pygresql/ is missing the Driver.py file i tried posting this to the list last night from work and it didn't go thru from work this morning and it didn't go thru and this afternoon from home i know my home systems is flaky now but my work system is fine so i don't know wtf is going on Sacha: how odd will have to look at that Sorry I can't find the hotfix - can you provide the URL johnf: sure, let me get it You can get a hotfix for GNUe-Common 0.5.3 at http://www.gnuenterprise.org/downloads/current.php Recent nightly snapshots or svn are also ok. the direct url is http://www.gnuenterprise.org/downloads/current/GNUe-Common-0.5.3-Hotfix1.tar.gz I just downloaded the hotfix - was is the correct install? it's in a readme inside it, but i'll cut n paste here since the damn ML hates me To install change to the directory ABOVE where you extracted GNUe-Common-0.5.3 then tar zxvf GNUe-Common-0.5.3-Hotfix1.tar.gz Sample Install: ----------------------------------------------- jamest@linux:~> cd Documents/Software/ jamest@linux:~/Documents/Software> ls GNUe-Common-0.5.3 GNUe-Designer-0.5.3 GNUe-Forms-0.5.3 jamest@linux:~/Documents/Software> tar zxvf ~/GNUe-Common-0.5.3-Hotfix1.tar.gz src/ src/datasources/ src/datasources/drivers/ src/datasources/drivers/postgresql/ src/datasources/drivers/postgresql/pygresql/ src/datasources/drivers/postgresql/pygresql/Info.py src/datasources/drivers/postgresql/pygresql/Driver.py src/datasources/drivers/postgresql/pygresql/__init__.py jamest@linux:~/Documents/Software> jamest: i even know why this happened :( jamest: this file was just added some minutes before i did the release i wondered about that jamest: there must have been a MANIFEST there from some previous test as everything seemed ok in svn head jamest: and it didn't rebuild the list jamest: and it didn't rebuild the MANIFEST we really need to release sigh yes common trigger stuff makes me a bit nervous jcater: how usable is your create_release script yet? and it would be nice to slip in the new trigger functions requested on the list jamest: i think we will relese in a week or so there are also some minor fixes i'd like to do before the release you think it's worth freezing common for a release i hate to stop the trigger work though :( i was also wondering if the changes going on will be enough to jump common to 0.6.0 * jamest doesn't recall what's on the roadmap or where it's at I installed and it appears to be in the correct area - but I still get the error. However, I tar zxvf from myhome and it looks like it placed the files into the correct area The install suggest that I tar from GNUe-Common.0.5.3 started gfcvs with -u qt option after login: Fatal Python error: PyEval_RestoreThread: NULL tstate /home/sacha/bin/gfcvs: line 9: 14844 Aborted /usr/bin/python2.3 ${SCRIPT} "$@" /home/sacha/bin/gfcvs: line 9: 14844 Aborted /usr/bin/python2.3 ${SCRIPT} "$@" reinhard: i think calling a form.commit() does not work anymore. what happens? reinhard: did start appserver with debug-level=5 but nothing goes to the server. in form: have a print statement before "form.commit()" and a print statement after "form.commit()". both print statements print. weird reinhard: I've been working on create-website so create-release is currently nothing but a few ideas not useful at all yet jcater: ok reinhard: since the appserver doesnt work w/o -Z should it maybe have the -Z be default? derek: it works for most people except for you :) concerning the qt3 problem: i am using debian unstable: python2.3-qt3c102 (3.8-3) what is the syntax to sort a database by a column? oh just saw, there is a python2.3-qt3 package. * derek thinks there used to be something basically i want to sort some columns in my report have very short time to play so if someone knows great, if not i will look up iirc, orderby="field1,field2" sort options are on my list thanks concerning qt problem: python-qt3 uses python2.3-qt3c102 same Fatal Python Error. well if anyone else uses kmusicdb hopefully by tomorrow there will be a LOT of new reports for you hmmm DB000: File "/usr/lib/python2.3/site-packages/_xmlplus/sax/expatreader.py", line 353, in start_element_ns DB000: AttributesNSImpl(newattrs, qnames)) DB000: File "/home/dneighbo/cvs/gnue/gnue-common/.cvsdevelbase/gnue/common/definitions/GParser.py", line 279, in startElementNS DB000: raise MarkupError, tmsg DB000: gnue.common.definitions.GParser.MarkupError: Error processing tag [I do not recognize the "orderby" attribute dneighbo@latitude:~/cvs/gnue/gnue-reports/samples/foobulations$ may be order_by I always forget ok jcater: on that css error is there an explanation is it that a css is optional anad i just need to provide one? css? it's order_by Oh yeah.....I can feel it Error processing css file: /home/dneighbo/gnue/share/gnue/filters/S on/tabulator.css dneighbo@latitude:~/cvs/gnue/gnue-reports/samples/foobulations$ oh and order_by does work :) does that file exist? no dneighbo@latitude:~/cvs/gnue/gnue-reports/samples/foobulations$ less /home/dneighbo/gnue/share/gnue/filters/SimpleTabulation/ .svn csv.xsl fo.xsl html.xsl kspread.xsl text.xsl btw: reports still rocks i think i am going to be able to do some of personal website just using reports :) I need to document some of the stuff I use it for at work we have a lot of stuff where I have users edit RTF files (as templates) and GNUe Reports uses those as templates to print out (or email, etc) fairly impressive reports i use it w/ hylafax some of our more critical stuff is done that way now all our contracts sent out to publishers use this i can turn all those notes over too as well as contracts we email or fax to potential schools my users love it but my usage is much simpler than jcaters :) the downside is you can't automatically do detail-like rows but for form letters, etc, it rocks bbl I really need to give reports some more luvin but don't see that happening in the near future as I want/need curses and input mask support in forms first well, I need input mask support and others have made it clear they need curses :) sigh i need case insensitive queries @weather 38654 jcater: The current temperature in Olive Branch, Mississippi is 79°F. Conditions: Mostly Cloudy. sigh @weather 65506 jamest: Error: No such location could be found. @weather 66506 jamest: The current temperature in Manhattan, Kansas is 63°F. Conditions: Light Rain. Vee2d2: you know, another feature they might consider is letting us specify whether bigbrother responses should get logged or what plugins do/don't get logged no need for @weather, @ping, @seen, etc to get logged but that's not really a big deal Just a general Announcement - Johnf is working on SUSE 9.0 (with the help of Jamest)! Many thanks to all that helped and a special thanks to Jamest! jcater: where'd the roadmaps wind up johnf: thank you for sticking with it hopefully it proves worth the effort http://www.gnuenterprise.org/tools/common on left bar, iirc reminds me, though I left those off the developers/ pages @weather 85233 derek: The current temperature in Gilbert, Arizona is 79°F. Conditions: Partly Cloudy. jamest: I slightly changed the look of those, though used pail colors so it wasn't so harsh fine w/ me the first copies were ripped off from KDE yip sigh we're no where near the 0.6 mark are we reinhard: the form.commit() works in another form. but in the form, where I call first a procedure does not work. jamest: I'm not above admitting we were overly optimistic of the feature plans for 0.5.x and move some to 0.6.0 then again, I only use svn at work so version # doesn't matter to me as much reinhard: i know why it does not work: i set the form to read only. dah. no wonder. then i have to think how i can make appserver "commit" the created business object. reinhard: procedure called another procedure creating 100 business objects and then committing those 100 new business object. appserver passed :) i'm blown away by the recent progress in appserver you guys are making the rest of the project look slow :) jamest: if i would not have met reinhard in person, i would think he's a machine ;) and johannesV. and the rest :) creating 1000 business objects appserver passed another 1000 business objects Sacha: what kind a biz objects you creating? i am implementing a sample application so i can enter a customer order so a customer order is a business object then i create a new sales order of that customer order also a business object so i create a new sales order business object by: newSalesOrder = new('sales_SalesOrder') and then I can assign the properties of the sales order business object by: Sacha: i am a machine Sacha: what you saw when we met is only a hologram ;-) newSalesOrder.sales_shippingAddress = self.customer.masterdata_shippingAddress :) haha jamest: so you see creating a new business object is easy. and i created about 2000 just before and stored them in the database jcater: if you are going to work on input masks apart from the fans now blowing faster, nothing changed ;) i'd like to cooperate i also have some ideas in that direction reinhard: have you seen the techdoc on niput masks ok btw jamest: yes the direction I'm taking now is to use yacc/lex type processors i.e., not reimplement the wheel on token processing *** _florin_ has joined #gnuenterprise also, this wasn't for public consumption yet but http://www.gnuenterprise.org/cgi-bin/viewcvs.cgi/jcater/test-masks/ (all my http requests time out, so that might not be quite right) was a test implementation I was working on outside of the main tree reinhard: I'm open to suggestions brb again * Sacha is going to create 10000 business objects just for fun. forms is of course in a state, where it does not refresh, redraw the window, as the procedure call is snychronous *** btami_ has quit IRC while appserver is creating 10000 business objects, i could add a new customer order with another client and created one sales order of that customer order appserver passed again .... why are there only 2208 rows in the sales order database? jcater: from the top of my brain an imput mask could be implemented as a python object that implements a defined interface in any case, it should be able to react on a keypress and tell the form how to behave yes, definitely like you have a input mask for dates unfortunately, that's what makes it so hard :) that is mm/dd/yyyy for example then imho the form should already display / / for an empty date then if you press "1" the form should look 1 / / if you press "2" then it should be 12/ / yes, if you look at our old formatting/BaseMask.py, that's pretty much the approach we tool tool=took and the cursor should already be *after* the / that is, there should be no need to enter the / and on dates like the american 1/2/04 ? do you display that as 1/2/04 normally? or would you say 01/02/04? an idea was that *if* you pressed / users will bitch if they have to enter 010204 then the leading zero will be filled so users can enter 010204 or 1/2/4 they'll want to 1/4/04 and just after they typed 1/ so it'd have to take the / as move to next section the entry will look like 01/ does that make sense? yes jamest: exactly, or whatever the section separator is for us in austria/germany, it's the dot today is 27.03.2004 also especially for dates maybe users would like parts that default to the current date like they only have to ente enter 1/2 and the form makes it 01/02/2004 but they don't have to enter the year if it's the current year * jamest pictures jcater sitting as his monitor....crying however *** holycow has joined #gnuenterprise i'm not sure if i'm going nuts now but i'd think the implementation of format masks should be a plugin structure where we have a date plugin a telephone number plugin (maybe) a vat code plugin for our european vat id's or whatever that was in the technote but i think that's already proposed in the technote yeah :) and though jcater forgot that existed he did same it was a similar system sigh, my nfs server is acting up reinhard: triggers seem to behave normally reinhard: 2nd test: i created 10000 business objects and committed them. OK. reinhard: 1st test: i created 10000 business objects and connected a second client also creating 4-5 business objects and comitting them. the 10000 business objects of the 1st client were committed. when i checked the number of entries in the db table for test 2 the number was 10000 when i checked the number of entries in the db table for test 1 the number was 0 jamest: good to hear Sacha: can you try to reproduce in a simpler environment? that is well not sure can you please give the actual order of things for the 1st (failed) test? ok. before the test i created (and commited) a customer order bo (business object) then in a form i called procedure one() of the customer order bo procedure one() called procedure two() of the customer order bo 10000 times procedure two() creates a new sales order bo and assigned properties of the new sales order bo when the form returns from procedure one() i call form.committ() in the form. while client 1 (fomr) was busy with procedure one() i started client 2 with client two i created a customer order bo then called procedure two() of the customer order (creating a new sales order bo) when procedure two() returned I called form.committ() in the form of client 2 . ok *should* work then i checked with psql the table of the sales order and there were the 2008 i created before this test. do you have a log of the SQL statements that were passed to the database? i did run the appserver without debug level, so no sql statements do you use postgres? on debian? yes postgres on debian on debian sid shall i start the test again with appserver in debug mode? to get the sql statement? there might be a debug log on /var/log dah :) for test 2 i dropped the db and started fresh maybe in /var/log/syslog, you see it hehe that's a good sign hmmm there is something in syslog ... not sure if its from test 1 Mar 27 23:34:48 gnaraloo postgres[15731]: [4-1] ERROR: could not serialize access due to concurrent update hmmm do you get all SQL statements in the syslog? i would have to run test 1 again. ok i dont know ... i dont get sql statements in syslog ok. new test1 should also create new postgresql log file. *** dneighbo has joined #gnuenterprise reinhard: test1 successful. all records in db excellent did add 6 entries in client 2 while client 1 was creating business objects *** _florin_ has quit IRC *** johnf has quit IRC *** derek has quit IRC hi all :) hi holycow hey, has any discussion gone on in terms of comparing something like zope.org and gnue? k. probly not :) nuthin on google. holycow: i think zope has a completely different target can you expand on that? not flaming, i am researching different frameworks :) trolling i mean i'm not technical enough to get a good enough grip on either just yet my understanding is that zope is about web services is that correct? yes gnue is not (mainly) about web services the web is merely one of many possible frontends for gnue no? how so? to me the difference between web services and traditional applications is the technology used to describe the frontend to interact with some datasource int he backend? (even not sure if the web frontend is functional yet) ah okay, so gnue framework is not constrained to web frontend many frontends are plausible exactly :D maybe it could be said that the target of gnue is a much more general one than the target of zope or maybe it could be said that zope is a specialized version of gnue *nod* okay that starts to put me squarely into the gnue ballpark but i know too little of zope to be sure holycow: i'm currently working on a web based po system using gnue one person that saw what i've done called it a mini zope i've been thinking about the work i do and the tools i need, and i look at all the disparate unrelated toolsets and its a bit disconcerting jamest, really? :) which is giving it way too much credit as it can't done anywhere near everything zope can do is there a url anywhere or private project? it's in our gnue-contrib svn repository jamest, well from what i've seen of gnue, the term 'not yet' comes to mind ah! *nod* okay time to start exploring then the question came up as i was looking at zope and thinking that the web based part is extremely limiting but it would seem possible to implement a zope type system using pieces of gnue i think you can now (in zope ) do an import zope from python i saw a reference to that the other day then i look around and looked all the little tools on linux for prject management, etc. and thought all of that could be rewritten in gnue... just wasnt sure *nod* i think you guys are really onto something here, waaaaaay ahead of the curve one of the BIG problems is that everything that we do as individuals and various organizations is that everything we do is project management we have all these disparate pieces of information from email, to text files to graphics, to various data formats... and its extremely difficult to start making links between the information it seems this is what bill gates meant when he said 'why cant i search no only for file names but the information in the files them selves' and making their file system have db properties in the next version you guys are building a wonderfull framework to compete with some of their ambitions, at least thats how it seems to me on the periphery one last q, in terms of widget sets, if someone wanted to implement a gant chart type application as you find in mr. project and such... is the current widget set deep enough to allow for such applications or are we mainly focusing on form based dev right now? forms is targeted at making it quick to build data entry forms however the core of gnue ( gnue-common ) doesn't require forms so nothing would stop you from writing a custom app with any ui widget sets you wanted and still utilize common, appserver, reports and forgo using forms itself aha! k. that answers a whole bunch. basically, all our tools can stand alone at the heart of it all is gnue-common *nod* i need to run to the store, bbiab thanks for the info jamest np *** jcater has quit IRC *** jcater_ has joined #gnuenterprise night all *** reinhard has quit IRC *** sjc has left #gnuenterprise is the ML dead? *** jamest has left #gnuenterprise jamest: can i call committ on a block?= or only form? form.committ() i think ML is very slow same error message. will send error message to reinhard. *** jamest has joined #gnuenterprise maybe I should jast use commit() instead of committ() ;) night all nite *** Sacha has quit IRC *** jbailey has joined #gnuenterprise there was the recent request for controlling from trigger the shift-pageup, shift-pagedown behaviour in forms anyone have any thoughts on instead giving them jumpRecords(number) hi hi *** jamest has quit IRC *** jamest has joined #gnuenterprise *** jbailey has quit IRC *** jbailey has joined #gnuenterprise *** jamest has left #gnuenterprise *** jbailey has quit IRC * jcater_ is away: I'm busy * jcater_ is away: I'm busy hrm I thought an empty /away signaled you were back