*** Morphous has joined #gnuenterprise *** Amorphous has quit IRC *** jamest has joined #gnuenterprise *** holycow has quit IRC *** holycow has joined #gnuenterprise *hmmm* i installed gnue but the forms designer doesnt want to start up do i need gtk componenets as well? nm got it :) okay, where do i find a basic how to make my first form tut? i see nothing on the site that helps a newb :/ *** thierry_ has quit IRC *** thierry_ has joined #gnuenterprise *** reinhard has joined #gnuenterprise anyone around? i have a couple of questions regarding setting the app server from the howtowriteaneweapp....pdf the first instruction says to 'gsscvs-vendor=postgresql-output=postgresql.slq get-appserver-started.gsd i have no clue what that means, is it a command? where? by what? holycow: hi you have installed gnue-appserver? latest svn snapshot? hi reinhard, not cvs, debian package, it makes a difference i take it? oh yes what i mean is, where do i use that command, bash doesn't like it, su to postgres doesnt do the trick... command is not available in standard debian package then? i.e. it is a gnue command? holycow: our last release (0.0.6) was a month ago and is already completely outdated we are already considering to release 0.0.7 within a week or so now in debian sid, we have 0.0.5 of appserver i think that says it all :) ah! k :) danke i did look at some of the samples very interesting approach reminds me a lot of what we do on lotus notes/domino *** jcater_ has joined #gnuenterprise *** jcater has quit IRC *** holycow has quit IRC *** sjc has joined #gnuenterprise *** reinhard has quit IRC *** gsoti_away has joined #gnuenterprise *** reinhard has joined #gnuenterprise *** Morphous has quit IRC *** ajmitch has quit IRC *** Stoke_ has quit IRC *** Morphous has joined #gnuenterprise *** ajmitch has joined #gnuenterprise *** Stoke_ has joined #gnuenterprise * derek notes in our samples we need to probably be putting notes so we say run gnue-forms myform.gfd but in the NOTES a the top we say if you are running CVS version you s/gfcvs/gnue-forms or the likes as when we move to better production packages im not sure we want to be redoing all our docs :) *** SachaS has joined #gnuenterprise *** lxf has joined #gnuenterprise *** lxf has quit IRC *** jcater_ has quit IRC *** jcater has joined #gnuenterprise *** lekma has joined #gnuenterprise anybody home? *** lekma has left #gnuenterprise *** jamest has joined #gnuenterprise *** gsoti_away has left #gnuenterprise *** jamest has left #gnuenterprise *** jamest has joined #gnuenterprise *** jbailey has joined #gnuenterprise *** siesel has joined #gnuenterprise hi *** siesel has quit IRC *** jbailey has quit IRC *** jbailey has joined #gnuenterprise *** siesel has joined #gnuenterprise hi siesel hi jamest the webpage is cool jcater's been working like mad on it yeah, I saw it a week ago, but its improving day by day *** dneighbo has joined #gnuenterprise night *** siesel has quit IRC *** derek has quit IRC jamest: you here? yes i've been looking at some issues with our dbdrivers my impression is that we would gain quite a bit by using parameters for our sql statements has that ever been considered seriously? parameters? huh? reinhard: i don't follow now we use DBSIG's "execute" with only one parameter we do cursor.execute (sql_string) all DBSIG2 drivers have an option to use a second argument cursor.execute (sql_string, parameters) then sql_string can be something like "SELECT foo, bar FROM table WHERE baz=%(baz)s AND frob=%(frob)s" and parameters can be a dictionary with "baz" and "frob" being keys you want to pass the dict into the execute yess but that's not universal is the reason we didn't do it the format of parameters is db-specific (could be list, could be dict) format of the parameter notation is also db-specific could be :name could be ? could be %s DBSIG2 seems to define 4 or 5 ways it can be done could be %(field)s and there is a way to test which way it is done imho we lose more by doing it the param wayu not gain more what we would gain would be * the driver would care for proper escaping of strings * the driver would care for proper date/time formatting I think you will be unpleasantly surprised w/that second one not sure what we would lose as dates are so nonstandard i.e., with oracle you have to pass in an OracleDate some you have to pass in an mxDateTime jcater: AFAICT which is why we were standardizing the way we were I'm not necessarily opposed to changing it every DBSIG2 driver must define a function to accept dates that is all drivers accept a .timestamp (y, m, d, hh, mm, ss) the reason why i was looking at this is that the current dbdrivers don't do escaping properly (tested for postgres, but i figure it's true for most of them) and i thought why reimplement something that is already implemented I just see it as a loss of flexibility not a gain I'm not against changing I just remember looking at parameters when we first did them and they were more trouble to keep straight than the benefit provided the escaping would be nice this seems like a fairly trival thing to test out we just then have to have 5 different ways of building SQL statements jcater: did you test which driver uses which param format? i do recall us looking at them and rejecting them , but my memory isn't as good as jcaters i could only test for postgres, and all 4 postgres drivers use python format the %(name)s format anybody got a non-postgres dbsig2 module installed? i don't wouldn't this just impact the DBSIG2 base driver? jamest: correct jamest: well, actually * jamest is wondering if it'd be a small enough change to add as a gnue.conf option enableExperimental DB or something then ask people to test in the next release if it breaks a lot of drivers then we have our answer :) all descendants would have to define a variable that holds the actual dbsig2 driver module what drivers are acutally in use? because I'm not against the change, i just recall it being rejected do people use other databases than oracle, postgres, mysql and firebird? reinhard: oracle uses :foo notation jcater: can you please in python import your oracle dbsig2 module and tell me what .paramstyle gives you ? >>> import DCOracle2 >>> print DCOracle2.paramstyle numeric ok thanks hmm that's acutally :1 notation I've never used numeric style with it I use :foo * jcater learned something new... it supports both :) with the dbsig driver? yes ok you've never tested if it would support python format also, did you? :) oracurs.execute ("""update kb_transmissions set recv_sent = :today where recv_sent is null and recv_ready = 'Y'""", today=today) is one of my statements reinhard: no I'm so used to :foo notation as that's what oracle uses in everything not just python wow but anyway like I said I'm not opposed to that change if you want to do it I just am not as sure as you are about the benefit jcater: i'm concerned about the loss of flexibility you mentioned i've already partly done it to test and it fixed at least two issues for postgres i think i could really enable it as a config paramter (for the time being) or do you have a better idea for how to fix quoting? i fear that quoting rules are quite different for different databases i think they are *** chillywilly has quit IRC *** sjc has quit IRC night all *** reinhard has quit IRC *** Vee has quit IRC *** mdean has quit IRC *** Vee has joined #gnuenterprise hey vee you here? i'm reviving gnue-sb commits i'll disable output of those to gnue-commits unless enough people think it should go there or unless someone sets up a seperate commits channel for that, i'm not going to mess w/ it however *** wayneg has quit IRC *** jbailey has quit IRC gnue-sb people? hello? i need help w/ the schema layout ok * derek is here cool we have jamest@calvin:~/shared/svn/gnue-sb/gnue-sb$ ls AUTHORS COPYING INSTALL README ROADMAP.in TODO ap contacts docs gl item pointofsale schema system utils * derek is trying fix a fscked up mozilla install (that is not upgradeable) for my daughter to do a research paper :) did jcater add the pointofsale? the schema * derek cant quite remember and if so is he maintaining that there or in contrib now? not the code gnue-pointofsale is in the mail gnue tree atm er main? the schema is in -sb though ah has been as it predates -contrib anyway how do I build a complete schema i need aria working w/ postgresql there is schema in gnue-sb/aria/schema but it sounds like there is merger of that into gnue-sb/gnue-sb im not sure if we made a gnue-schema style tehre yet i also wish to add purchase order stuff which I would assume should go in gnue-sb/gnue-sb/purchasing/schema? i dont see an arias there? how did you check out? hell if i know its been more than 6 months :) we were trying to get the arias guys to convert and fix things so we were on the most current version svn co svn+ssh://www.gnuenterprise.org/var/svn/gnue-sb/trunk gnue-sb i lost track fo them jamest: oh we moved it to ash? * derek jumps for joy if we did it's there dcmwai is in here and he's aria yeah but i mean lost track of their work dcmwai wasnt doing the work anyway, i think my user/role layout is more flexible at first glance grabbing svn now im sure it is the way they did things wasnt all that incredible as i don't hard code access rules as fields, it appears at quick glance they did in their genuser table let me know when you have it sorry got side tracked kids tearing up house grabbing now... (forgot to give password so it timed out) jamest: im trying like mad to remember what we did here :( i know we were trying to move arias over to our schema AND at same time to try to break it into smaller parts to tackle converting it to GNU Enterprise problem is i dont recall where we left off we probably need to start from scratch again as they didnt use our tree (and are still actively developing) in fact, now its coming back to me they were supposed to look at our layout of nola and start making arias similar then we were going to use their tree sigh they are if you go to their site they talk about using gnue-reports in the recent posts hmmm but i dont think they switched over yet we really need to talk to josh their technical lead (i know you have like zero time) and we need to try to convince them to switch arias to our svn imho so we can fix things up and be running out of a single tree or at least get them to switch to our schema format so that we can be more productive lol gnue productive oh wait you're serious um, yeah. i agree =) judging from what dmcai (sp?) said other day to jamest josh isn't developing it any more not sure if that was just short-term stuff, or long-term i forgot about that josh is just last person i have on record as being our technical contact we either need to work with them or fork let me see if I can find that reference but maintaining two trees will get unweidly at a minimum we need to be using same schema format so we can at least share a schema (which is probalby most important for what jamest wants) well look at gnue-sb it is kind of messy i was thinking we originally were going to put all the .gsds out in the directory where their modules were i'll be keeping my user/role system which already would break aria but then it kind of looks like we did it in the root dir and so now that it has been so long im not recalling what we did or why :) so jamest you are looking at a complete fork? * derek is interested in your user/role system i have no idea what I'm looking at as if it is pluggable maybe i can use at the county it's very basic and is part of web-po the intire system is plugable well in gnue-sb there should be item and contacts with roles defined by the plugins those are the only two pieces i recall playing much with and both have their schemas located within their folder so i have im thinking the other folders were added because we were starting to pull over arias im not recalling what the gnue-sb/schema/ directory was for user_id | organization_id | role_name ----------------------+-----------------+---------------- jamest@math.ksu.edu | 3 | po:user jamest@math.ksu.edu | 4 | po:user jamest@math.ksu.edu | 3 | receiving:user lindsay@phys.ksu.edu | 3 | po:user im kind of thinking it was a way to generate a master schema po:user is a role why is role name something:user it can be anything just thinking if you are separating at any time the : but I have a po section and a receiving section why not hav it be two columns like these sections define their role names role_function and role_type role_function = po that's not what that is role_type = user my po section makes all it's roles start w/ po: simply so I know which sectoin defined the rolw tole role role can also be site_manager dereks_boss jasons_cat whatever, but in my po plugin I did this to keep it strait for me hmm we've started to do this internally to gnue events too at county im debating about groups vs roles or whether to a large degree they are just different ways of looking at same thing you need to check out gnue-contrib and look in web-po the schema is there but no admin interface also web-po isn't just for po you can make any plug-in you like put it in the right dir and it'll become part of your system wtf did i use bad syntax or there rerally gnue-forms in gnue-contrib? ? fsck i think wrong syntax i'm sure of it svn co svn+ssh://svn/gnuenterprise.org/var/svn/gnue-contrib i sitll have head in ass about svn i used svn co svn+ssh://www.gnuenterprise.org/var/svn/gnue/trunk gnue-contrib *** loco1 has joined #gnuenterprise does gnue works with python 2.3? loco1: yes derek: that checks out gnue/trunk and puts in gnue-contrib dir but your syntax was wrong too so i feel better :) mmmmm, ok, something else is wrong sigh a typo :) loco1: what's up i got it now good i expect patches daily bwahahaahah I'd follow the installation procedure in http://www.gnu.org/software/gnue/community/docs/GNUe-Win32-CVS-HOWTO.html and no luck ouch DB000: UnicodeDecodeError: 'ascii' codec can't decode byte 0xf0 in position 9: o rdinal not in range(128) requires apache2 first I fix an error importing datetime any reason for apache2 ? loco1: what are you doing that gives you that error? what did you download (a regular release? snapshot? svn?) open designer how did you install it? what operating system and version derek: it requires mod_python 3.1 which requires apache2, older mod_python's don't support session management win2000 following http://www.gnu.org/software/gnue/community/docs/GNUe-Win32-CVS-HOWTO.html jamest: ouch python 2.3 wxpython 2.4 apache2 just isnt my friend thats all okay i just installed on win2000 on friday :) did you grab a snapshot? svn or a release (0.5.3) gnue-svn-2004-03-17 hmm i would either go with todays snapshot OR the current release i would start with current release ok, lets try unless you are wanting to help develop the tools more than write applications be sure to delete the folders you created for that 2004-03-17 before installing the new stuff install common first first I will try the current release then forms then designer then appserver, reports, navigator etc in any order you like jamest: looking briefly now at web-po no way i can install and play as apache 2 is satan on my machines nope it's the schema in there as it's not user friendly install yet but the start of the roles is there and look in src/sections silly question but do we still have mdeans documentor for gnue-schema? at po.py the one that makes really pretty html files for .gsd files? at the top you see me define the roles, and the actions in the section the whole po:foo thing is just what I called them, it has no real meaning yeah im not sure at this point i need this level at countya i need more traditional groups plus data security as there are really only two types of users (today) analysts and liasions morrons and pains in the ass er, wait? was that a questoin? and of course it depends on who you ask.. which one is the moron and which is the pain in the ass or on the day sometime a persons got to pull double duty rofl :) *** loco1 has left #gnuenterprise *** SachaDinner has quit IRC *** fil_c has joined #gnuenterprise hi *** fil_c has quit IRC *** genbie has joined #gnuenterprise hi genbie