reinhard (~reinhard@M1250P019.adsl.highway.telekom.at) joined #gnuenterprise. aaron_ (coyote@CPE-24-163-193-174.mn.rr.com) got netsplit. jrs (~jrs@dsl093-129-017.sfo2.dsl.speakeasy.net) got netsplit. steveb (~steveb@202.0.44.163) got netsplit. InsanitY (insanity@co58147-a.olden1.ov.home.nl) got netsplit. chillywilly (~danielb@CPE-24-167-199-51.wi.rr.com) got netsplit. ra3vat (~ds@ics.elcom.ru) got netsplit. derek (~dneighbo@cpe-24-221-112-50.az.sprintbbd.net) got netsplit. InsanitY (insanity@co58147-a.olden1.ov.home.nl) returned to #gnuenterprise. chillywilly (~danielb@CPE-24-167-199-51.wi.rr.com) returned to #gnuenterprise. steveb (~steveb@202.0.44.163) returned to #gnuenterprise. jrs (~jrs@dsl093-129-017.sfo2.dsl.speakeasy.net) returned to #gnuenterprise. aaron_ (coyote@CPE-24-163-193-174.mn.rr.com) returned to #gnuenterprise. ra3vat (~ds@ics.elcom.ru) returned to #gnuenterprise. derek (~dneighbo@cpe-24-221-112-50.az.sprintbbd.net) returned to #gnuenterprise. jcater (~jason@cpe-066-061-083-220.midsouth.rr.com) left irc: "Client Exiting" wtg (~merlinx@ppp1021.qld.padsl.internode.on.net) joined #gnuenterprise. btami (~tamas@ip102-205.ktv.tiszanet.hu) joined #gnuenterprise. fixe (~fixe@dsl093-081-253.chi2.dsl.speakeasy.net) left irc: Read error: 104 (Connection reset by peer) jcater (~jason@cpe-066-061-083-220.midsouth.rr.com) joined #gnuenterprise. howdy huh? ah oh i forget AI's need no sleep lol jcater (~jason@cpe-066-061-083-220.midsouth.rr.com) left irc: "Client Exiting" johannes_ (~johannes@M1259P002.adsl.highway.telekom.at) joined #gnuenterprise. reinhard_ (~reinhard@M1254P024.adsl.highway.telekom.at) joined #gnuenterprise. reinhard (~reinhard@M1250P019.adsl.highway.telekom.at) left irc: Read error: 60 (Operation timed out) johannes (~johannes@M1250P007.adsl.highway.telekom.at) left irc: Read error: 110 (Connection timed out) Nick change: reinhard_ -> reinhard ra3vat (~ds@ics.elcom.ru) left irc: "Client Exiting" ra3vat (~ds@ics.elcom.ru) joined #gnuenterprise. Stoke_ (~stoker@ppp209.tc-22.roc-pt.ny.localnet.com) joined #gnuenterprise. SachaS (~Sacha@dialup-196-58.wasp.net.au) joined #gnuenterprise. Stoke (~stoker@ppp329.tc-2.roc-pt.ny.localnet.com) left irc: Read error: 113 (No route to host) lupo (lupo@pD9E6804D.dip.t-dialin.net) joined #gnuenterprise. dimas (~dimas@195.218.176.222) joined #gnuenterprise. CleeK (~CleeK@csahut.net1.nerim.net) joined #gnuenterprise. hi allù ù hi CleeK i discovered gnue today :) nice you can be the first trying the hot 0.5.0 :) whooohoo! 0.5.0 is out :) cheer and dance! huh :) really ? really really :) look at the Topic for #gnuenterprise yep ToyMan (~stuq@smtp.dstoys.com) joined #gnuenterprise. woaw, there r many packages in this project r they all working ? :) more or less :) morning all morning good day Mon Apr 7 14:33:24 CEST 2003 lun avr 7 14:33:07 CEST 2003 :) tu es francais? oui :) you too ? no i'm german ok but i had french for 7 years in school cool :) I don't know 1 word of german you don't need to I studied spanish a little bit but that's all i had russian for 12 years in scool :) but forgot all :) as long as we all can speak and write english, everything's fine that's cool, i would like to know how to speak russian ask ra3vat or arturas they got some cool security hackers :) r u all developpers of gnue ? (me) trying to use/patch/develop i'm just someone who disturbs jcater and jamest (the most active developers wrt forms, designer, reports, navigator, common) to include my bug fixes :) johannes, reinhard and jan (=siesel) develop the application server the rest of the ppl here do diverse stuff ok is gnue ready to be used in an enterprise context ? yes at least in two-tier mode or should i wait for version 1.0 ? :) it's a good access replacement wait a moment, i remember some list of active users in a KC edition access ? database you mean ? I don't knwow SAP very well, but gnue should be in the future a free replacement of this soft, no ? http://kt.zork.net/GNUe/gnue20030308_71.html#3 microsoft access Cleek: yes GNUe is currently targeted for smaller businesses, because we can't claim the stability and scalability of SAP, even with 1.0. but yes, the technical spirit is inspired by SAP and quite a few other packages (Oracle Forms). I suppose that sap at this time is really bigger than gnue actually, currently it's "just" a set of database-aware applications ok so you can write most database and ERP related stuff with it i just wouldn't bet a Fortune 500 company on GNUe yet :) SAP is a monster you can do anything with it, and that's IMHO its biggest fault dsmith (dsmith@oh-strongsvillecadent1-1d-224.clvhoh.adelphia.net) left irc: "later.." when you design with "one size fits all", you'll have at some time the result that "one size fits noone". Nick change: dimas -> mercy do u know compiere ? www.compiere.org the best thing to know what gnue is, install it and write a simple form for an already existing database. compiere is well known around here :) ok that is one of the standard question :) actually, compiere is not really a free software project heh :) you don't have a faq on gnue.org :) we have! oh !! sorry :) i didn't see it :) btw, has anyone seen neil tiffin (neilt) in recent days or weeks? no jamest (~jamest@gw.math.ksu.edu) joined #gnuenterprise. jamest! the one and only :) :) what are the next steps for 0.6? :) deprogram (~hmason@66.94.69.196) joined #gnuenterprise. step 1) finish the features promised in the 0.5.x feature plan :) www.gnuenterprise.org/feature-plans but I forgot to upload new ones last night :( :) i'll try the qt forms driver now it's not done the only fully working driver is wx the others will flesh out over the course of the 0.5.x series having said that I think qt is usable on simpler forms it doesn't work with foreign key stuff in dropdown lists hi jamest he doesn't even show the entries :( well i'v just committed to fix for dbdrivers (typos) but some direcoty renaming needed: sapdb/sapbd -> sapdb/sapdb hm navigator still not usable? that's the one thing that bugs me most :( Nick change: CleeK -> CleeK-PulA what? lupo: try the zipcode example i tried to make a gpd with a form in it and the form is not displaed i mean for the dropdown lists that sample works hrm gpd was working for me yesterday btami: how's win32 going? i have just 1 packaging error now gnue-head/samples/tutorials/forms/zipcode/zipcode.gfd is forms 0.4 format it's wery strange lupo forms/samples/zipcode not samples our samples are a mess :( btami: you mean a typo in the driver right not a misnamed directory in the repository the dialogs in forms with datasources are working but in intro/gauntlet not nope doesn't work with qt driver lupo: AH! qt dropdowns I think are not done yet qt is very rough at the moment sigh with wx it works fine, don't worry Action: jamest was happy about 0.5.0 now I'm reminded how far we have to go lupo: all but the last dialog work for me on debian woody is that what you are seeign? you mean the zipcode stuff? jamest: there will be no samples-0.5.0? what dialogs? s/lupo/btami btami: i didn't create one our samples are a complete mess we need to pull them from forms and put in samples i _think_ and make them more educational mkay meaning less duplication of features more simple examples gauntlet is not a sample of that i made gauntlet as a way to break forms :) jamest: can you rename the sapbd dir to sapdb? where at? i looked and didn't see it if not, i have to change the __init__.py where is the dir drivers/sapdb/sapbd fsck yeah, just a minute thanks Stoke (~stoker@ppp570.tc-1.roc-pt.ny.localnet.com) joined #gnuenterprise. Action: lupo is away: buying a lab smock have to pick-up kids and go home bye btami (~tamas@ip102-205.ktv.tiszanet.hu) left irc: reinhard: you around? yep when releasing last night the new common I bundled up a appserver too and put up as a release but I didn't know the stat (notice no official announcements of the release :) IIRC you guys were ready for a release to match the new common should I leave it up or pull it? let me put it this way we haven't reached a specific milestone yet with appserver but it's been a long time since the last release and a lot of incompatible changes so the release is good for 0.0.3 and it makes sense to keep it Nick change: mercy -> dimas btw thanks for the effort of doing it dsmith (~dsmith@borg.altus.cc) joined #gnuenterprise. Stoke_ (~stoker@ppp209.tc-22.roc-pt.ny.localnet.com) left irc: Read error: 113 (No route to host) jbailey (~jbailey@atlas.fundserv.com) joined #gnuenterprise. released? Like "released" type of released? Or a pseudo-fuzzy "almost released" type of released? as in http://www.gnuenterprise.org/downloads/current.php released just not announced except in IRC bundle her up and slap her in sid oooo, i see if images work well enough for you Action: jamest recalls you needing them Is it worth packaging appserver yet? dimas (~dimas@195.218.176.222) left irc: "÷ÙÛÅÌ ÉÚ XChat" deprogram (~hmason@66.94.69.196) left irc: Read error: 60 (Operation timed out) dunno from earlier today i asked reinhard if he minded my doing that release we haven't reached a specific milestone yet with appserver but it's been a long time since the last release and a lot of incompatible changes so the release is good for 0.0.3 and it makes sense to keep it So, I'm guessing no. (since he told me not to bother before) ok did you decide to stick it out w/ setup.py? I'm asking because both jcater and myself dislike the current setup system (not enough to switch to autoconf though :) and I think you'll see a nicer setup system soon i was wondering how much of the .deb creation we could automate jbailey: i think we if we package the next release for appserver we will still be early enough for sarge :) I would like to understand what you hate about autoconf. The concerns you listed seem to largely have been solved in the most recent releases. reinhard: Parse error my personal issues for disliking it It's a pain in the ass wrt to missing macros It's syntax is a pain in having to remember is this where we've went from AM_FU_BAR to AC_BAR_FOO jbailey: actually i would like to have appserver in sarge when sarge freezes :) jamest: deb creation is quite simple jamest: which part do you want to have automated? I used to use it religously for all my c programs but really hated using it jbailey: but 0.0.3 is more a snapshot than a release Setup.py is pure python, it works on windows, it's easy to understand and extend in any way I desire jamest: Aside from the Windows issue, which I ackowledge still exists - a configure.in for python should be all of 6 lines. with dependency checking? with --skip-tests? I mostly wish that gnue would use it so that it complied with the GCS and worked as the principal of least surprise. jamest: Dependencty checking is implemented by automake. with the custom install we currently do wrt to samples/docs/etc To run the tests you have to intentionally run make check. yeah, i've used automake/autoconf what I wanted to see setup.py --make-deb What is the maintainer name? What is the foo? reinhard: Cool. Just let me know. What is the bar? ETc? Are you sure? Creating GNUe-Forms-0.9.0.deb I really dislike when upstream packages do things like that. GCS? The One True Way to build a deb packages is to run 'dpkg-buildpackage' with the right options. jamest: GNU Coding Standards. ah so to switch we'd have to pick up autoconf and automake and break our ability to install on windows Action: jamest is thinking....... jamest: The auto* tools are only for building the dist file. They shouldn't be necessary once the configure script is written. jamest: Or so it's supposed to work. they are necessary for the developer Yes. for cvs I don't care about the user :) j/k so what are we taking about...... It would take a hardier soul than mine to develop on windows. my auto* skills are rusty makefile.in and configure.in right? jamest: It's not ideal for windows. That's correct. not ideal == not work at all right? jamest: Which is why I haven't just sent a patch to you guys to take care of it. :) jamest: not work without extra tools. I've never gained auto* skills. I'm a Makefile by hand kind of guy. jamest: What I had proposed originally was that if you all use Unix to do development, you can add a hook to the Makefile to generate the windows executables and the installer. jamest: You need a Makefile.am, it generates Makefile.in from that. jamest: So the idea would be that from your GNU/Linux system, you'd run; ./configure, make windows-install if we add a hook to make the exe doesn't that mean we'd still need a setup.py or has autotools replaced that functionality of dist-utils? jamest: And from there it would generate everything you needed to just install on windows since setup.py makes the exe Phone afk. back Nick change: CleeK-PulA -> CleeK I'm not so familiar with what setup.py does. I had looked at a package a while back that was a windows installer and the packages could be created from Unix. jcater (~jason@w202.z065105010.mem-tn.dsl.cnc.net) joined #gnuenterprise. well, actually I'm no thinking clearly setup.py doesn't do the .exe mcmillian installer does that it's been ages since I've messed with it though (thanks btami! :) setup.py does automatic building of .tgz, .zip, and .rpm files it allows us to write any python we want as part of the install process so we use it to autogenerate sh or .bat scripts at install time it handles proper placement of all files including docs, config files, images That placement, certainly in a posix environment, is handled by the autotools. I'm just reading the mcmillan-inc.com web page right now. so what are we talking about bottom line? which files need maintained in autoconf? We're talking that unless there's a useful solution for the Windows platforms, autotools won't do it for you. The only files in autoconf are: configure.in and at least one Makefile.am, but up to as one per directory. sorry, I talk about autoconf as autoconf and everything needed to make it work (automake, etc) So do I. if we need to test to see if a python module is importable? does that require a C code fragment? No, I think you should be able to do that in autoconf. I'd have to look - When I did my couple of Python bits, I wasn't looking for python modules. (wihtout C code. No point in using C for a python module!) my only experience with auto* was using c/tcl and I just remember it being more painfull a process than setup.py Most of mine is with C and C++. Lemme look to see if there's a macro for detecting python module dependencies. for the record: i think auto* sucks. I think setup.py sucks too. jamest: You've made that clean. I just don't yet see why you think it sucks. =) because it is difficult to use it has a strange grammar AC_FOO_BAR AM_BAR_FOO AM_WHO_KNOWNS that you have to know to use AM_THIS_NOW What is strange about that? You have macros and they take parameters. which is replaced a few months later by AC_THAT_WAS_THEN which breaks my stuff They don't generally replace macros much anymore. and there doesn't seem to be much logic to the macro names so I can't derive what to use next I recall my auto* days as constaint digging thru the available macros to see if something was already setup it's a very powerfull system I admin youssef (~youssef@adsl-128-36.wanadoo.be) joined #gnuenterprise. Action: jcater wants a custom installation program personally but it's written to provide a language like C all the stuff it needs to be portable across inconsistant systems Ah. there's AC_PY_CHECK_LIB(foo) that looks to see if foo.py is in the python library path. s/admin/admit hello python doesn't have the portability issues that C does (IMNSHO) hello youssef so I'm not 100% sure we need that complexity to be fair the one place python does need to worry about it (file mappings) setup.py sucks Action: youssef is interested in memory translation and translation it handles windows/unix installs differently which I hate Most apps don't need the complexity of the autotools. But it's very nice to know that ./configure; make; make install will work on 95% of the apps out there. I wanted to see a much nicer custom install too jcater i wonder if we can autocreate the files auto* needs or if that is a maintenance nightmare with the custom installer? or in general? What's your vision for a custom installer? i'd rather see a dict based dependency (with the installer) system along the lines of what GParser, GConfig do so that a gnue package defines a few dicts in it's setup and doesn't worry about how things get done do GNUe need translators for French ? right now each setup.py is custom youssef: we're alway interested in people willing to do translation I'm not sure we have anyone doing french translations at the moment I envision the installer being like what jamest said but also it provides a nice front end for users used to oracle-like installers it could also spit out packages, possibly like .deb setup files how can a translator begin to translate ? with dialog support It seems wrong to put that in the installer. And not in the build system. sigh Action: jcater thinks jbailey is thinking too "C" :^) jcater: In what way? I have an installer on Debian. It's called 'apt-get'. Similar for RedHat, 'rpm'. yeah It would be wrong for those to themselves produce packages for other systems. and people are having to hand-maintain those installers sigh no the installation generator could spit out packages jcater: They'll always need to be hand maintained - The requirements for each distribution is different. Action: jcater needs to get back to work but for the record, I am against auto* then we steal the plug-in idea from our drivers to allow a redhat-plugin, debian-plugin, mandrake-plugin that live in the installer and only need done once in the setup system and not per package it's a dream I know :) ./setup.py --just-work-dammit i'd be very interested to see if we can make setup.py spit out the auto* stuff then we'd both be happy and people that prefer one system over the other could use it as it really isn't one or the other Action: youssef is busy Nick change: youssef -> youssef_busy see you good luck bye youssef_busy (~youssef@adsl-128-36.wanadoo.be) left #gnuenterprise ("Fermeture du client"). youssef_busy: I'm not sure who (if anyone) coordinates translation. Maybe derek knows. darn Navigator? What's navigator? right now? not much it's a menu system that launchs apps,reports, other programs ah, okay. So nothing i'll need for Debian then. it can add it's process def file to the gnome or kde menu and it does connection pooling so if you log into a form via navigator it holds that connection for other gnue tools you launch read :single logon oh hmm. it's not getting much TCL but I depend upon it here at work I don't really want to require that someone have to go into a tool to get into the various GNUe pieces. Will it eventually be transparent if you go right into one of the apps directly? so the users have to learn 1 command to get to a menu of the other stuff it's not quite like that in the wx mode you have a tree on the left Sale enter order hang on Sales order entry HR personnel maintenance benefits crap General zipcode maint coutry maint it uses gpd file (gnue process definition) where you can define steps in a process run this form run this app generate this report as the user navigates the tree a html frame on the right can contain any help screens you deem needed so the user can browse thru tasks and select the task they need to accomplish it (should) then guide them thru the process i'm not sure how we'd deal w/ connection pooling in the outside apps I don't know if python support shared memory segments where I find navigator handy i used to use forms symlink trick to make things like school_maint launch the school_maint form as the number of forms users used increased the number of command they had to learn increase now they remember 1 in soviet russia, the commands remember you doh! sorry, recovering from slashdot reading this mornign BLECH. This is the part I hate most about updating this package. Is getting the changes all right for the setup.py CleeK (~CleeK@csahut.net1.nerim.net) left irc: "BitchX: now with wings" huh? I have to modify the setup.cfg.in BUt it changes with every release. This time some lines aren't needed anymore, so I have to renumber the patch offsets. what do you have to do to it? can you send me a sample? jcater provided the original, but it's basicalyl setting 'prefix' to be 'stage' in setup.cfg.in and then create a site.cfg (site.cfg.debian) hmason (~hmason@66.94.69.196) joined #gnuenterprise. jamest: I'm guessing that it's so that during build, it doesn't try to write to readonly directories. gnue-common-0.5.0-1_all.deb ready for testing. I need to hack something together to create a connections.conf file. Hmm ello jbailey: ? chillywilly: morning you mean a whiptail like thing? or just a default? jamest: Yeah. I'd do it in debconf, though, so it could be in whiptail, dialog, etc. A default one would fix the bug for now, but I should probably offer an automated setup. hi Jon I'm a bit mixed. I'm not raelly sure how I should approach it. I thin having a systemwide connections.conf is probably not the right thing for many people. Hmm Maybe I should just put an empty default one in /etc/gnue (in case system-wide is wanted), and then hack it to look in ~/.gnue for another one. most debian packages give you commented out configs and/or a config that you need to change to suit your needs but at keats you have an example to work with least better than writing on from scratch in most cases true dat the default could just contain a boilerplate for the [gnue] connection Sure. I don't like default config files with examples in them - It's best provided in a man page. then the install dialog says before this is useful modify it to suit your needs (that way it's still there after the config file is changed) or something like that chillywilly: There should be no reason to have an install dialog for gnue on debian. well it helps to be reminded and told of the config file ansd what you have to do imho chillywilly: It's a violation of debconf policy Can't do it. =( Action: chillywilly thinks of dhcpd which he installed the other day why do they do it for dhcpd Hmm. Is gnue known to work with wxgtk2.4? yes it's prefered chillywilly: Because dhcpd is an essential package, and not configuring it could leave you unable to access the machine. oh chillywilly: So in that case, it's a bug not to tell you. hrrrm jamest: Thanks! in fact hurry up with getting sid into sarge into stable sorry I am not familiar with policy as we want to dump support for wx2.2 :) the dialog thing isn't that important anyway chillywilly: It's all good. That's why you guys keep me around. =) grumble, grumble, libc maintainer, grumble, grumble or are people still blaming things on that dude? jamest: libc? glibc's fine.. for now. =) gotom wants to break it again. =) er, damn s/libc/glibc jbailey: tell him no break it again? how? ;) Should I still be using python2.2, or should I use python2.3? whichever you prefer i think i run 2.3 at home chillywilly: I did. So far he's listening. But there's alot of bugs that the newer package fixes. chillywilly: The down side is that it breaks 2 arch's badly, so would halt testing again. sigh you know there's no 2.3 right? that's alpha/cvs ooooooo jcater: Ah, really? crap, that's right i forgot that jcater: I just noticed that there' python2.3 packages and alot of modules. =) it's in alpha status Okay. I'll stick to 2.2 then. =) Someone will file a bug when I need to switch. Action: chillywilly tries to stick with 2.2 as far as the dependency anything >= 2.0 should work (like happened when they noticed that I wasn't using wx 2.4 yet) Action: jamest doesn't always like the fact sid debs depend on the latest and greatest jamest: only problem is depending on python's modules so do we avhe a release? i.e., what if you have python2.2, python2.3 and python2.3-xml Action: derek doesnt have time to read backlog :) and it's running against python2.2 then read the topic ;P derek: I'll have sid packages for you to test soon. it was release last night so you could test it the deb would pass the dependency check, but wouldn't be runnable as you had the time so I think that's why they have to do specific python point releases Action: jamest thwaps derek back ah, makes sense club wife works for went out business (she got call late last night) ouch jamest: Is there a sample empty connections.conf somewhere that I could use? so my plans to do work last night were thwarted :( touch connections.conf jbailey: how far off are sid's? empty enough? derek: About 30 minutes. jamest: Is that enough to start the program? Action: derek has some work to clear off plate, but can test after lunch can i ask someone about navigator? i don't think you even need it at all if you do then it's a bug i think jamest: Lovely. Then beat Derek for filing it against the debian package instead of upstream. with pleasure Action: jamest thwaps derek Action: jamest thwaps derek again Action: jamest thwaps derek a 3rd time hey jason, you want in on this? Action: chillywilly thwaps derek with a smelly trout jamest: Can I tag the bug upstream, wontfix then and just hope that it's fixed in 0.5.1? =) i think so ought to let jcater comment as maybe he did make that a requirement last I knew it wasn't I think it's always been a requirement but not saying it should be Action: jamest scratches his head you mean I have to take back the thwapings? No. It's still an upstream bug. __init_() got multiple values for keyword argument 'connections' TypeError, this is gack! what program? mom Nick change: jbailey -> mom YeS? you installed GNUe on your mom? Ewww. Action: derek doesnt remember what bug i filed Nick change: mom -> jbailey that's just so wrong on so many levels but i think the deb package should be placing a connections.conf go hom :) navigator regardless of whether the program demands it DB001: [GNObjects:165] Running form thisworks.gfd DB010: [UIwxpython:248] connections is DB000: Traceback (most recent call last): DB000: File "/home/lupo/cvs/gnue-head/.cvsdevelbase/gnue/navigator/UIwxpython.py", line 188, in OnTreeItemActivated DB000: object.run() as gnue w/o out a database is pretty useless DB000: File "/home/lupo/cvs/gnue-head/.cvsdevelbase/gnue/navigator/GNObjects.py", line 166, in run DB000: self._controller._clientHandlers[self.type](self) DB000: File "/home/lupo/cvs/gnue-head/.cvsdevelbase/gnue/navigator/UIwxpython.py", line 239, in runForm DB000: self._runForm(formfile) DB000: File "/home/lupo/cvs/gnue-head/.cvsdevelbase/gnue/navigator/UIwxpython.py", line 251, in _runForm DB000: ui=ui, disableSplash=1) DB000: TypeError: __init__() got multiple values for keyword argument 'connections' ah, crap Action: derek thinks the programs may not need it that's longhand for, expect a new navigator release real soon now but the packaging should create one roflmao probably let derek work his magic today and break every damn package he tries derek: I'm not certain it should. then do a 0.x.1 release derek: It's quite possible for there to be no system-wide connections. jbailey: my point is lets not do this the "painful" way lets do it the "easy" way that is if we dont provide a connections.conf support work goes up 400% as that will be the most FAQ it will already be asked over and over The right way is to provide tools to create them. I HAVE FOUND A BOOOOG because we will have to show them how to change values properly yes yes yes The user should be able to run (or have run automatically on the first time) a wizard that will handle the config file for them. The sysadmin can edit the one in /etc by hand if it's needed. um (the user one that I think should be in ~/.gnue) I think we need to add support for user-level connections.conf but I don't think the debian packages should support that jcater: Details. =) until the main release does Action: jcater thinks that's going to be a big headache for the debian maintainer navigator/src/UIwxpython.py, line 249: remove "self" as a parameter! lupo: does that remove your bug? yep it creates another, but much later :) sigh the new one is GFInstance instance has no attribute 'setForm', line 264, same file nonetheless, line 249 is impossible according to everything i know about python :) jbailey: can you make common create the connections.conf derek: I can. i.e. prompt people for driver, server, database etc derek: Yes. then allow them to 'skip' it derek: But like I said, I'm not sure that's the right thing. and latter they can come back and do dpkg-reconfigure gnue-common derek: If it prompted the user the first time they pulled up one of the apps, it would be better. derek: It's goign to take more thinking about and then some elbow grease either way. I don't see very many personal connections.conf files except for developers Action: derek is pro anything that prevents i apt-get install gnue-* and it doesnt do anything that is it designer forms etc get put on menu system somewhere derek: The problem with the prompting is that then I have to be able to parse the connections.conf files and deal with any custom changes the users might make without breaking the file. jcater/jamest at one time we discussed Yeah, I keep meaning to add the menu entries. =) if you ran forms jcater: Hard to say. They'd be used here, because everyone needs different connections. and didnt pass it a form it would pop a dialog to select a valid forms file did we ever do that? or did we decide against it? okay, fixed that as well. jcater: And the users should be able to connect to new datasources without calling IT. (so it can't require root access) what is "unknown step type form" in navigator but again you are talking developers I didn't say there was no use for user-level but the majority of them won't be lupo, let me try to run my dept .gpd file thru navigaot jcater: Perhaps we have different definitions of developers. =) Action: jamest swore he did this How about adding a directory, /etc/gnue/connections.d or something, where there can be lot's of conf files, so the main .conf doesn't need to be edited? developer: over caffinated, undersexed, social outcast that talks to machines. And have a default .conf file that just talks to a text flatfile db or something? jamest: what version do you use? #gnuenterprise: mode change '+o jbailey' by ChanServ!ChanServ@services. Topic changed on #gnuenterprise by jbailey!~jbailey@atlas.fundserv.com: We've released new versions of appserver,common,designer,forms,navigator, and reports. http://www.gnuenterprise.org/downloads/current.php Share and Enjoy! | Running Debian unstable? Add this to your sources.list: deb http://people.debian.org/~jbailey gnue sid completely untested of course. =) Now back to my regularily scheduled.. err.. Action: jbailey looks at my watch lunch it looks like. Action: jbailey is away: food lupo: 0.4.x in production, was going to 0.5.x this week mkay CVS head is currently 0.5 more or less? cvs head is 0.5 i think unless someone commited in last 15 hours okay then i use 0.5 == CVS head do you think we can go through my observations? sure I'm looking at it now here too okay Action: lupo has a lot of coffee, so be prepared :) i've only got 30 min then we'll make it quick i see that in navigator/src/UIwxpython.py, line 264, an outdated method "setForm" (which does not work) is used. I suppose this should be "buildForm" by now. indeed, when using buildForm instead of setForm and debuglevel=10, then i get all the typical forms startup debug output, so i conclude this must be right method to be called. ok, hang on jcater: on timeline where is designer ability to support navigator er that is where i can make navigator screens with designer i think other things should take priority over that derek: i think that navigator itself should have priority over its design tool that's why i'm extensively testing it right now and jamest is my core team victim (CTV) [tm] sigh this is all screwed up i don't think i'll have her running before lunch but this is major roadblock to my deployment here so it will be running soon mkay i hope i have given you some pointers. actually, it's about two simple fixes as far as i can see want a patch? sure dsmith (~dsmith@borg.altus.cc) left irc: "later.." crap crap? yeah crap i'm not sure how you'd get this to run with only a two line patch hehe maybe I need more caffinee this is screwed up formfile = os.environ['INSTALL_PREFIX']+'/'+formdir+"/"+step.location so on my system I get /usr/local/gnue//usr/local/math/forms/formname.gfd Action: chillywilly definitely needs more caffiene not real useful :) what should the default behaviour be? if the step.location starts with a / then that's absolute lupo: i agree Action: derek asked for where it fit in roadmap not pushing for designer support as part of this release jamest: yep. else it tries formdir+"/"+step.location if that fails then step.location ? jamest: yes, that would be the best behavior. there's no navigator-support@gnuenterprise.org, shall i send it to you, anyone important to CC? btw, i have used a relative path for that, so no probs for me. just to me jamest@math.ksu.edu would be best or should it flip the last two jamest@gnuenterprise.org? jamest@math.ksu.edu would be best reinhard (~reinhard@M1254P024.adsl.highway.telekom.at) left irc: Read error: 104 (Connection reset by peer) lunch! reinhard (~reinhard@M1254P024.adsl.highway.telekom.at) joined #gnuenterprise. jamest: you have deserved that. :) hi reinhard oh, rejoin reinhard (~reinhard@M1254P024.adsl.highway.telekom.at) left irc: Remote closed the connection oh how yummyu got smb to mount win2k machines over vpn if ONLY i could get adaytum to run under wine i think i woudl be golden is there someway to enable file compression over smb? reinhard (~reinhard@M1254P024.adsl.highway.telekom.at) joined #gnuenterprise. jamest: okay, i'll send you a new fix. it let's navigator work, the only downside is that forms doesn't work _within_ navigator. i hope that helps. derek: i'm not sure whether smb supports file compression lupo: i imagine it doesnt Action: derek is jazzed i got it to work at all but over modem line would be nice to have faster way to get dir listings and such Action: jbailey is back (gone 00:46:26) derek: according to google, it doesn't. derek: Tried my debs yet? =0 =) nope working :( lunch is soon then will grab hehe theo de raadt is way cool :) Cool, thanks! http://www.globetechnology.com/servlet/story/RTGAM.20030406.whack46/GTStory derek: you could likely do a ssh tunnel for smb traffic which you could then enable compression over SachaS (~Sacha@dialup-196-58.wasp.net.au) left irc: Excess Flood SachaS (~Sacha@dialup-196-58.wasp.net.au) joined #gnuenterprise. lupo: your second patch is getting closer :) give me a bit to see what I can do what did you mean by it opens forms in an own window in the navigator window, left side, i click on the process step where the form should be loaded. the login dialog pops up, i enter my data and a forms client is started in a separate window that's how it aways worked the navigator window stays open in it's window, the form launches seperate so you can launch multiple forms from the same menu what it's doing for me that I don't like so it's not a bug, it's a feature :) is open a form and select the jump to record icon and then? jcater: hmmm interesting approach Action: derek is thinking it better to first try to get vpn client working so can use broadband doesn't it make the navigator window pop to the top? oh over the launched form jamest: no, not for me jamest: but it produces a KeyError: _jumpto :) ah lol nevermind prior to lunch I think i got the form mostly working but not the path crap you need to do # # Build the form tree # instance.addFormFromFile(formfile) # # Start the instance # instance.addDialogs() #instance.buildForm(form) instance.activate() why the addDialogs()? i think the dude that created that system is fscked in the head lol dialogs are sexy as I have no clue why he made it seperate step roflmoa mao jamest: can you commit that? yeah, just a sec i've butchered the formfile stuff above it commited thanks works as expected please test it out good :) i'm going to pull the current release jbailey: what got uploaded? pull? gnue-common - The shared library for many items of the GNU Enterprise Frameworkgnue-designer - A Rapid Application Development tool for GNU Enterprise. pull gnue-forms-wxgtk - An XML-based forms painter gnue-reports - An XML-based output-independent reporting system is what i see out there 0.0.3 is unusable oh 0.0.4 will be usable thaaaaat "pull" :) cool all four seem to be new Action: jamest counts the packages derek lists 1...2....3 hmmm gnue-designer didnt paste but it was updated ah ok jbailey: errors galore on install SyntaxError: invalid syntax File "/usr/lib/gnue/lib/python/gnue/reports/adapters/filters/Standard/Base/psutils/PSOutputStream.py", line 45 PS_HEADER = """\ %!PS-Adobe-3.0 """" ^ SyntaxError: invalid token is last error before it finally died but there were several before it Standard/* can be stripped out for the .debs i will message them to you ah hell i will post them here as most look like python errors so jcater/jamest may want to comment Setting up gnue-forms-wxgtk (0.5.0-1) ... /usr/lib/gnue/lib/python/gnue/forms/uidrivers/curses/UIdriver.py:0: SyntaxWarning: name '_EVENTPROCESSOR' is assigned to before global declaration [13:00] Last message repeated 1 time(s). Setting up gnue-designer (0.5.0-1) ... Setting up gnue-reports (0.1.1-1) ... File "/usr/lib/gnue/lib/python/gnue/reports/adapters/filters/Standard/Base/Parser.py", line 161 'sortoptions': { ^ SyntaxError: invalid syntax File "/usr/lib/gnue/lib/python/gnue/reports/adapters/filters/Standard/Base/psutils/FontMetrics.py", line 47 for line in handle.readlines() ^ SyntaxError: invalid syntax File "/usr/lib/gnue/lib/python/gnue/reports/adapters/filters/Standard/Base/psutils/PSOutputStream.py", line 45 PS_HEADER = """\ %!PS-Adobe-3.0 """" ^ SyntaxError: invalid token File "/usr/lib/gnue/lib/python/gnue/reports/adapters/filters/Standard/Base/Parser.py", line 161 'sortoptions': { ^ SyntaxError: invalid syntax File "/usr/lib/gnue/lib/python/gnue/reports/adapters/filters/Standard/Base/psutils/FontMetrics.py", line 47 for line in handle.readlines() ^ SyntaxError: invalid syntax File "/usr/lib/gnue/lib/python/gnue/reports/adapters/filters/Standard/Base/psutils/PSOutputStream.py", line 45 PS_HEADER = """\ %!PS-Adobe-3.0 """" ^ SyntaxError: invalid token Setting up gnue-common (0.5.0-1) ... File "/usr/lib/gnue/lib/python/gnue/common/datasources/drivers/odbc/mxodbc/Driver.py", line 128 def _buildQuery(self, conditions={}forDetail=None,sql=""): ^ SyntaxError: invalid syntax File "/usr/lib/gnue/lib/python/gnue/common/datasources/drivers/odbc/mxodbc/Driver.py", line 128 def _buildQuery(self, conditions={}forDetail=None,sql=""): ^ Anyone actually know what these errors mean? SyntaxError: invalid syntax Standard/* should be stripped out for the .debs... it can be removed from the setup.py if need be oodles and oodles of unable to delete foo because not empty for common and reports it will play a big role in 0.2.0 of reports, but wasn't used at all in 0.1.0 Action: derek suspects its due to directory structure changes in the two jcater: In both designer and reports? let me know why its time to apt-get update; apt-get install gnue-reports gnue-designer gnue-common gnue-forms-wxgtk s/why/when is there one in designer? ./gnue-designer-0.5.0/build-tree/GNUe-Designer-0.5.0/src/reports/Standard ./gnue-designer-0.5.0/build-tree/GNUe-Designer-0.5.0/build/lib/gnue/designer/reports/Standard ./gnue-designer-0.5.0/build-tree/GNUe-Designer-0.5.0/stage/lib/python/gnue/designer/reports/Standard that's definitely not a usable part of designer yet derek: Does forms work? I don't think it'd hurt to remove it, but am not 100% positive jbailey: i didnt attempt to try based on it erroring out er i guess it didnt error out only warninigs Oh, I didn't see the error up there. let me give it a whirl Tx. Can I otherwise consider derek's gnue-forms warning reported here, or do I need to drop it into dcl? jbailey: we should really do all bugs via dcl, but jamest jcater prefer irc :) derek: I go with whatever method the people who fix them prefer. and since they are squashing the bugs .... I don't think the forms issue is an error just a warning jbailey: forms passed my most 'basic' test latitude:/home/dneighbo/cvs/gnue# gnue-forms --version GNUe Forms Version 0.5.0 GNUe Common Version 0.5.0 latitude:/home/dneighbo/cvs/gnue# jcater: Looks like. It would be nice if I could silence it somehow, though. well, seeing as how curses wasn't functional/intended to be .deb-ianized this release you could squash it real easy :) um forms appears to work jcater: True. Should I just rm -rf the curses directory? probably jcater: curses isnt debianized i.e. i dont see a gnue-forms-curses has anyone of you thought about role based access control in navigator? (jamest: i've already asked you) derek: right that's what I just said there's no reason to keep the curses/* stuff in gnue-forms-wxgtk Yeah. I'll sort out that properly when there's another frontend that's ready to package. lupo: yes derek: documented or in your mind? QT was *very* close to being ready in fact much of navigator was started off of stanley kleins document but we really wanted to get a release out it's been too long about process definitions his document of course talked much higher level than anything implementation based error: package directory 'src/uidrivers/curses' does not exist Blech where can i find this proposal? Where else do I need to prune it. Just src/GFClient.py? Or is thre a Makefile type thing that also mentions it? right now we are far from an implementation jcater: you aware bug i reported at pre8 still exists wrt to lookup boxes? which one yes ok any chance a 0.5.1 will fix that? hopefully though if I knew what was causing i5t or this one of those, we did really bad thing, time to rethink? it'd be fixed in 0.5.0 :) as we use in a few places in gnue-sb im hoping very soon to get something useful derek: so where's the document :) What's the right solution for the gnue-common bug? lupo: let me find it thx http://www.gnuenterprise.org/project/docs/security.html talks about security and has sections on RBAC but its not what i was thinking of derek: i will read thru it so i have a base for work and/or discussion bye then Action: lupo is away File "/usr/lib/gnue/lib/python/gnue/common/datasources/drivers/odbc/mxodbc/Driver.py", line 128 def _buildQuery(self, conditions={}forDetail=None,sql=""): ^ SyntaxError: invalid syntax brb reinhard (~reinhard@M1254P024.adsl.highway.telekom.at) left irc: "This sentence no verb" lupo (lupo@pD9E6804D.dip.t-dialin.net) left #gnuenterprise. siesel (jan@xdsl-213-196-208-10.netcologne.de) joined #gnuenterprise. hi all hi :> Nick change: hmason -> deprogram *sniff* there are no debian packages for powerpc :( What? They're arch: any err. arch: all rather hmm, I'll try again. you running unstable? on i386 yes. on powerpc woody + parts of unstable It's not in testing right now, so if your sources aren't pointed at unstable, you might not get it. I pointed to people.debian.org/~jbailey oh hmm. Sorry, that's my bad then. but forgot, that there is just a binary-i386 subdir ;) They're arch: all, but I have them in a binary-i386 subdir because that's what my script wants. I just need to figure out how apt-ftparchive works a bit better and it would be fine. wget + dpkg work too :) its just a heap of dependencies missing in woody :) SachaS (~Sacha@dialup-196-58.wasp.net.au) got netsplit. SachaS (~Sacha@dialup-196-58.wasp.net.au) returned to #gnuenterprise. chillywilly (~danielb@CPE-24-167-199-51.wi.rr.com) left irc: "leaving" chillywilly (~danielb@CPE-24-167-199-51.wi.rr.com) joined #gnuenterprise. reinhard (~reinhard@M1254P024.adsl.highway.telekom.at) joined #gnuenterprise. dsmith (~dsmith@borg.altus.cc) joined #gnuenterprise. chillywilly (~danielb@CPE-24-167-199-51.wi.rr.com) left irc: "leaving" chillywilly (~danielb@CPE-24-167-199-51.wi.rr.com) joined #gnuenterprise. ToyMan (~stuq@smtp.dstoys.com) left irc: "Client Exiting" reinhard (~reinhard@M1254P024.adsl.highway.telekom.at) left irc: "No problem is so interesting that you can't just walk away from it" Stoke (~stoker@ppp570.tc-1.roc-pt.ny.localnet.com) left irc: Read error: 110 (Connection timed out) jbailey: were you going to brew new debs with fixes ? i have a few moments to test again derek: I don't have fixes for the problems yet. Vee2d2: you around today? on hpone - slow responses what are the problems? jbailey: i thought jamest/jcater gave direction on deleting many of the things to remove the errors? jamest: should all be in back log that's really not his problem is it? that'd be ours jamest: not sure if he is using autotools probably his problem or are they package issues on debian if using setup.py prolly yours that is the directories have to be in cvs as you all are 'working' on those features just they aren't done so no need to be in the packages at least thats what i gathered from the ui driver errors/warnings (curses) and the reports/common errors (some of them) wrt: standard adapters forms error was curses reports errors were with standard filters (not adapters) designer appeared to have no errors common had errors with mxodbc driver Action: derek didnt see you or jcater respond to that one the other one i think jcater said just pull the directories out of the packaging as they are not 'ready' for 0.5.x here is the mxodbc error: SyntaxError: invalid token Setting up gnue-common (0.5.0-1) ... File "/usr/lib/gnue/lib/python/gnue/common/datasources/drivers/odbc/mxodbc/Driver.py", line 128 def _buildQuery(self, conditions={}forDetail=None,sql=""): ^ SyntaxError: invalid syntax File "/usr/lib/gnue/lib/python/gnue/common/datasources/drivers/odbc/mxodbc/Driver.py", line 128 def _buildQuery(self, conditions={}forDetail=None,sql=""): ^ SyntaxError: invalid syntax jbailey: i assume soon we shouuld be switching to gnue-forms-wxgtk doesnt try to build curses and instead we see gnue-forms-curses, gnue-forms-qt and gnue-forms-gtk2 ;) but also on gnue-common i think it might make sense to have gnue-common-psycopg gnue-common gnue-common-mxodbc etc where gnue-common is truly the base package then you add db drivers, rpc transports etc with additional packages but for now im just down with something that works :) derek: I should be, some time later on. also for sanity sakes we need to see what it will take to back port packages to woody derek: Yes - I'll work with you guys to split it out. derek: Do each of those have different dependencies? The only reason I would want to do that is to minimise dependency hell. yes, all those have different dependencies yip diff dependencies (and reason why we want them seperate) as 'size' wise they arent big enough to justify breaking them apart but to do proper dep checks it will make sense (imho) dont know if jcater/jamest agree, but suspect they do is the deb stuff in place in cvs? Action: derek is chomping to do more playing (so i can beat on gnue-sb) Action: derek thinks everything is ok (as far as working) without having new packaging as i dont use mxodbc and am not playing with standard filters ;) or curses jcater: what do you need from me to help fix the foreign key lookup bug better traces? screenshots? sample form and db? a patch would work well ToyMan (~stuq@65-73-51-150.bras01.glv.ny.frontiernet.net) joined #gnuenterprise. hmmm interesting narrowed down a little further if you do the commit all drop downs go to set value but then as soon as you jump to a new entry they all refresh to proper values DB005: [GDataObjects:526] Master record changed; Notifying Detail Objects i think is what is 'changing' things Is what deb stuff in cvs? None of my stuff is, no. it appears it doesnt 'notify detail objects' on the commit but rather waits until you leave the record you committed? that's a completely different part of the logic notifying detail objects means to load a different child resultset bingo thats definitely the behavior this is similar to the 'commit' button not being activated if i do a commit the lookup boxes dont get 'refreshed' until i leave the record i was on during the save so say a record has 10 fields if do an insert and enter data and commit the lookup boxes freak out if i move from entry to entry in the current record they stay freaked out as soon as i move off that record they refresh properly the good news... this bug isnt nearly as it first appeared i would mark it annoying/cosmetic instead of critical ;) Action: derek would patch, but its getting sufficiently complex in the model that it takes a good day to reunderstand it if you dont look at it daily deprogram (~hmason@66.94.69.196) left irc: "BitchX: more nutritious than a six-pack." resetForeignKey and allowedValues appear to both be functioning properly jamest (~jamest@gw.math.ksu.edu) left irc: Read error: 60 (Operation timed out) dsmith (~dsmith@borg.altus.cc) left irc: "later.." icltlfatppl (~icltlfatp@adsl-66.218.46.51.dslextreme.com) joined #gnuenterprise. fixe (~fixe@dsl093-081-253.chi2.dsl.speakeasy.net) joined #gnuenterprise. forms needs interactive debugger :) gnue-forms --interactive-debugger so has their been an "official" release or you still squishing bugs? ya official was last night ouch dneighbo@latitude:~/cvs/gnue-sb/item/forms$ gnue-forms --interactive-debugger item_category_maint.gfd > (0)?() (Pdb) what is it wanting me to do? enter doesnt seem to step through I forget how to use it it's kick ass but without a cheatsheet it's useless jamest showed me how to use it before rofl we need to get his notes in a techref ok just making sure i didnt miss the cluetrain somewhere it will do everything you would expect Action: derek assumed there were magic keybindings but without a list of commands it's hard to get started :) does help do anything? cause gdb will spit out some stuff for you that way d seems to stop it :) jcater: i know where the bug happens i just dont enough about forms to know where in 'code' that is the one event that could help narrow it down is in 5 or 6 different objects :) jamest (~jamest@gw.math.ksu.edu) joined #gnuenterprise. sup jamest lost a ups took out me firewall I think my users really like it when something like that happens HEH bbl jamest (~jamest@gw.math.ksu.edu) left irc: "[x]chat" icltlfatppl (~icltlfatp@adsl-66.218.46.51.dslextreme.com) left irc: "Client Exiting" jbailey (~jbailey@atlas.fundserv.com) left irc: "Client Exiting" jcater (~jason@w202.z065105010.mem-tn.dsl.cnc.net) left irc: "Client Exiting" ToyMan (~stuq@65-73-51-150.bras01.glv.ny.frontiernet.net) left irc: Read error: 104 (Connection reset by peer) jamest (~jamest@adsl-64-216-111-196.dsl.tpkaks.swbell.net) joined #gnuenterprise. jcater (~jason@cpe-066-061-083-220.midsouth.rr.com) joined #gnuenterprise. jamest: you have cheat sheet (or url) for interactive debugger? hey chillywilly's suggestion of help worked :) egads there needs to be a way to set the break point Action: derek will have to look when i have more time g'd night siesel (jan@xdsl-213-196-208-10.netcologne.de) left irc: "Client Exiting" jamest (~jamest@adsl-64-216-111-196.dsl.tpkaks.swbell.net) left irc: Read error: 104 (Connection reset by peer) jamest (~jamest@ppp-66-142-210-175.dsl.tpkaks.swbell.net) joined #gnuenterprise. ToyMan (~stuq@65-73-51-150.bras01.glv.ny.frontiernet.net) joined #gnuenterprise. jamest (~jamest@ppp-66-142-210-175.dsl.tpkaks.swbell.net) left irc: Read error: 104 (Connection reset by peer) wtg (~merlinx@ppp1021.qld.padsl.internode.on.net) left irc: "You ain't got debian you ain't got shit!" jamest (~jamest@ppp-66-142-210-5.dsl.tpkaks.swbell.net) joined #gnuenterprise. wtg (~merlinx@ppp1021.qld.padsl.internode.on.net) joined #gnuenterprise. jamest_ (~jamest@ppp-66-142-211-238.dsl.tpkaks.swbell.net) joined #gnuenterprise. jamest (~jamest@ppp-66-142-210-5.dsl.tpkaks.swbell.net) left irc: Read error: 104 (Connection reset by peer) Nick change: jamest_ -> jamest ToyMan (~stuq@65-73-51-150.bras01.glv.ny.frontiernet.net) left irc: "Client Exiting" jamest (~jamest@ppp-66-142-211-238.dsl.tpkaks.swbell.net) left irc: Read error: 110 (Connection timed out) jamest (~jamest@adsl-64-216-111-217.dsl.tpkaks.swbell.net) joined #gnuenterprise. am I here? you are i think unless you meant mentally, then the answer is 'no' er 'not quite' Action: jamest thwaps derek with his shinny new netgear MR814 it's about the only thing it's good for rofl hows the testing going? had soccer practice was waiting for new debs (earlier) i wouldn't wait on the debs wife has pampered chef party so im vacating to go watch college basketball finals then will be back to test ok Action: derek spent most of afternoon working and trying to fix the bug of refreshForeignKey being broked have it narrowed down to 'when' it happens just have to find that when in the code tried using interactive debugger but stepping through the code to get to right point is too painful what? have to learn how to use interactive debugger better what bug? or at least put a 'break' in the code in the section closest to failure so i dont have to step through really basic stuff like reading the xml file what bug? jamest: read... this bug bothers me enough im gonna touch the code... muhahahahahahaha Action: derek can imagine jamest calling jcater and going... wtf... get on line bi0tch he is gonna rape our baby jamest: the bug um WHAT BUG!?!?!?!?! if you have a foreign key lookup that gets 'refreshed' from the current table i.e. you dont have to shut down and restart form to get the item to show in the dropdown when you 'commit' every entry on the form displays the same value (i.e. old records) until you exit the current row then it refreshes them properly so if i ahve field1 field2 field3(lu) -------- -------- ---------- [20:47] Last message repeated 3 time(s). and then at i insert a new record what calls the refresh? is it automatic and enter field1 and filed2 and push commit or is it triggered via a functoin it saves but all lines on field3 go to 'foo' until i move to a different record jamest: let me check do you have a simple sample form I can test with? i will probably have time to look at this in the next day or so block trigger fk_key="id" fk_source="dtsItemCategory" max_length="34"/> inpParent.resetForeignKey() ]]> the resetForeignKey() works no problem the problem is the 'fields' get refreshed like crap internally you can checkout gnue-sb ok, will you be here for a bit? and gnue-sb/item/forms/item_category_maint.gfd will do this everytime i may need help installing it about 5 to 10 minutes then will be back about 2 hours later if Vee2d2 is around he is kosher with things too ok, no promises on tonight i'm JavaBoy(tm) at the moment javaBoy meet .netGirl have fun ;) Action: derek is away: taxman JavaBoy(tm)? why? i have a java app I need to run under windows did you write it? no :) i'd go into more details but this channel is logged i think you know my feelings about java :) NDA? no k politics, wouldn't want some people to stumble across it HEH Action: chillywilly thinks it is a bit too late jamest (~jamest@adsl-64-216-111-217.dsl.tpkaks.swbell.net) left irc: Remote closed the connection --- Tue Apr 8 2003