formatter module which is imported in common/utils/TextUtils.py does it specific for the latest Python's version? I have no idea, don't even know if that is used anywhere can you chech in common/utils/TextUtils.py, do you have it there? check just a second seems to work I can import, and also it is used in web page generation off to lunch python from binary build is complaining checking again off to lunch *** btami has quit IRC *** johannesV_ has joined #gnuenterprise *** johannesV has quit IRC Connection instance has no attribute '_defaultBehavior_' is it mostly a config issue? no, most probably an issue of mixed versions i.e. old pyc files hanging around after svn up or something like that try clean new checkout of common reinhard: do we have svn working? hah good question :) svn works for me after updating the host key strange that i have cvs install working and can't get another one in /usr/local/gnue on the same machine you get that _defaultBehavior_ error message? yep you definately have some old Connection.py[c|o] files hanging around reinhard: how did you update the host key? just delete the old one from known_hosts ok the new one will then be added automatically which ip of the new server? ~$ host svn.gnuenterprise.org svn.gnuenterprise.org has address 70.158.36.94 *** wt has joined #gnuenterprise reinhard: still can't svn, could you send me new Connection.py by email? i have gnue-common 7910 that's ok you just should make sure that you don't have older Connection.pyc somewhere else e.g. find . -name "*.pyc" | xargs rm already did that, and cvs install works fine *** jamest has joined #gnuenterprise trying to install in /usr/local strange and there is no "old" insteall in /usr/local that you installed over? *** kilo has left #gnuenterprise double checking now, thought i moved the whole gnue dir first than installed again reinhard, what about removing internal Session from appserver's auth-agent ?! ah to get authentication data? I'd vote for replacing that with a temporary session right just like for filter values yep *** barsik has joined #gnuenterprise *** barsik is now known as dimas_ so svn is up but the site is down? reinhard: reinstalled into /usr/local - still the same problem *** andre has joined #gnuenterprise jamest, yes looks like that ... dimas_: strange. can you look at the traceback to see which files are actually used? which dbdriver do you use? can anybody check - if there is navigator/build/scripts-2.3/gnue-navigator done for windows it is not renewed during install under linux and just moved to /usr/local with broken paths to interpreter psycopg i'm using psycopg too, and, well, it's ok here ... dimas_: please try to change connections.conf from "psycopg" to "psycopgdrv" this is a new name for the driver that will make 100% sure gnue won't fetch an old driver version from anywhere it is only needed for appserver side, right? if you use appserver, then yes cvs - working, usr/local do not with psycopgdrv? this is creazy crazy hi ! Good Morning arrrgh dimas_: found it hi andre! Hi! Mr. reinhard I'm trying to use a soundex query in postgresql (fuzzystring) how can I use this on a gfd file in gnue-designer in pgAdmin sql I use this query select * from people where soundex_string(name) like '%' || soundex('Schneider') || '%' the soundex_string get all soundex for any word The result of the phonetic search above: Neitzke / Naitzke / Scheibler / Schaibler / Schneider / Schnaider "Edgar Blásio Schneider" "Nicolau Schneider" "Walter Schneider" "Leonardo Schneider" "Hélio Luiz Schneider" "José Ari Schneider" "Sandra Feix" "Roque Schneider" "Claudemir Sander da Silva" "Clarice Schneider" "Clari Schneider" "Ervino Schneider" "Sandro Lowe" "Lotário Schneider" "João Tadeu Schneider Paraíba" "Daniel Schneider" "Albino Ângelo Santarosa" "Aloísio Schneider" "Marcos Aurélio Schneider" "Pedro Canísio Schneider" "Sandra Sofia Schroeder Wissmann" "Carlos Schneider" "Sedemir Schneider" "Perci Delfino Schneider" "Nelson Schneider" "Cláudia Sander da Silva" "José Vicente Santurio Ramos" "Marlene Schneider" "João Leonardo Schneider" That's my soundex rules: Neitzke / Naitzke / Scheibler / Schaibler / Schneider / Schnaider Eloi / Eloe / Noeli / Noili Algeo / Algeu Moraes / Morais Stuermer / Stiermer / Stüermer (ue=ie=üe) Júlio Basso, Julho Baso, Julio Baço Dresch / Drech / Dreck / Drec / Drek (Sch=Ch/Ck=C ou K) Schenckel / Chenckel / Xenckel / Schenkel (Sch=Ch=X=K) Christian / Cristian Therezinha / Terezinha / Teresinha / Terezina / Teresina / Theresina / Therezina Lehmen / Lehnen / Lenen / Lemen / Lemem / Lehmem Massuchini - Massochini - Massochine - Mazzochini - Mazochini - Masochini - Masoquini Ignacio / Inacio / Inácio / Ignácio Vociechoski = Vociecoski = Vociecoski = Vociecosqui = Vociequosqui Luís = Luiz what's the idea? sorry by the overdose of information ps. My full name is "Andre de Souza Pinto" make a phonetic search in the gnue-forms using a simple query my email address is andre@aspinformatica.com.br andre: I don't think you can do this other than with a trigger dimas_: can you svn up meanwhile? btami: the win32-driver still doesn't treat multilines right ... if not, the fix is easy: drivers/Base/Connection.py contains a line "_defaultBehaviour_ = None" -> should be _defaultBehavior_ (without u) damn american vs. british spelling differences jumping into a multiline-edit, entering some text "foobarbarbaz" shows, that after pressing the cursor does *not* move, and after the next character (b) the cursor stays in front of the "b" reinhard: could we add to the undocumented colon: operators to do phonetic search dimas_: however I wonder why this breaks for installed but not for cvs jamest: is the : operator already there? yes i thought we only *talked* about it oh not sure how portable this would be no, i'm pretty sure I commited it i'm not sure about that either *** kilo has joined #gnuenterprise [ds@exodus gnue]$ svn update Password: Password: Password: Permission denied (publickey,keyboard-interactive). svn: Connection closed unexpectedly yeah, it's commited you can do queries with :>:value again for the logs: btami, sorry for blaming you, but if i'm using just instead of to add a line-break it's working fine ! If we can mount a dynamic query in runtime, in some event, the portability stay in the programers hand jamest: what could be done to fix ssh login? dimas_: i was out all weekend but last I knew jason was moving subversion to apache2 server and trying to eliminate|reduce ssh access to the box so I have no clue where he's at in that process dimas_: looks like the ips all point to the new server yes, connect goes there Is there any form to write a dynamic query in the "F9->Perform query" event ? In Run-Time ? Using a gfd file ? I don't think so andre: no sniff ! you could possibly write a pre-query trigger then raise an exception at the end of it but you'd get an error dialog ideally query and commit could have override triggers as I have forms that could use that functionality where instead i've resorted to putting a button on the screen Thanks dimas_: try ssh now trying.. *** jcater has quit IRC works now, thanks jamest, jcater reinhard, the dropdown-pb is handled by win32 driver correctly ... i think this is because of the different kind of dropdowns excellent probably because of the different signals e.g. windows generates a focus-in signal and gtk doesn't yeah hmm if i run a win32 forms client against appserver ... quitting the form raises a socket.error 104 at the server-side looks like the win-client does not close the socket properly ... but this would be a bug in the python xmlrpc library on windows, wouldn't it? hm, i'm still not sure where to nail down that bug maybe it's something within our xmlrpc since i've created a small test-server and client ... and that one is working fine well, even common.rpc.server and common.rpc.client are working fine so maybe forms is simply not closing the connection? yep rpc.client and rpc.server (in gnue.common) are working without an error *** dneighbors has joined #gnuenterprise yes, or it does not release that ServerProxy (or it's PersistentTransport) *** dneighbors is now known as derek but if i'd add a debug-print into appserver-driver's close () method ... *** kilo has quit IRC *** jcater has joined #gnuenterprise [bars@exodus forms]$ gnue-forms --debug-level=1-4 town.gfd : bad interpreter: No such file or directory\PYTHON.EXE reinhard: the same what i said about navigator under linux dimas_: can you please explain again what you do here? I don't understand how you get PYTHON.EXE under linux yeah, looks like i've found it the transport's socket doesn't get closed automatically will look for a clean solution tomorrow ... (as this would mean to access the ServerProxies' ClientAdapter instance) reinhard: may be installation was once done for windows via windows share, so build/ scripts was there and they just moved to /usr/local during linux installation, i have no other explanation as of yet maybe you have to remove the whole /build directory when changing the OS i was expected /build to be rebuilded completely each new install it might well be that distutils don't expect completely different operating systems to be installed from the same source honestly I've never tried dimas_: i don't think that's the case i'm pretty sure it only rebuilds what it thinks needs it a setup.py clean should wipe the build dir dimas: appearantly I had comparable problems - in my release script I explicitly rm -rf the build directory before I start setup.py I can't remember what *exactly* the problem was bbl jamest: it would be ok if not met combined with few other errors like today jamest: not a problem, thanks yeah, i hate when they stack up *** andre has quit IRC *** derek has quit IRC dimas_: however I wonder why this breaks for installed but not for cvs reinhard: dirty /build dirs the cause for this ah good and bad good because it shows I was actually guessing right (mixture of different versions) bad because I didn't figure out how it happened how things go well when you steps no out of linux... johannesV_: you removed the commitAll() functoin from GConnections as obsolete but it wasn't tagged obsolete, what replaces it as it broke my timeclock app here commit on the correct connection? *** andre has joined #gnuenterprise i'm doing that now via a getConnection('name') then commit it was at one point added as a quick hack as we needed it for appserver now it's not needed any more so we thought safe to remove it i grew to use it in my apps as it was less code than grabbing the connections and commiting them and covered any connections i opened not a huge deal if I'd actually done extensive testing prior to rolling out common this AM if you need it feel free to re-add it i'm on the fence as specifing them individually would be more descriptive code ... and safer as you might have situations where you have other, independent connections open at the same time where I use it on generally only have 1 connections that you don't want to commit right then i'll adjust my code hmmmm would it be bad form to do GConnections.commit('connectionName') as I don't need the connection for anything else does anonymous svn work now? jamest: I'm not sure. The connection manager should *manage* connections, but maybe not *operate* on them true when it comes to object oriented design, I sometimes imagine objects being little people shouting at me "you ask me to do *what*??" ;-) but of course it's your deal i'm fine with that i think short term I need to put commitAll back in but with an obsolete warning as i find I'm using that more than i expected of course, i really don't need to re-commit that though do I :) *** psu has joined #gnuenterprise *** andre has quit IRC *** dneighbors has joined #gnuenterprise hi dneighbors hi I heard we got haxored ;-( *** btami has joined #gnuenterprise cu *** reinhard has quit IRC *** btami has quit IRC *** btami has joined #gnuenterprise *** btami has quit IRC *** btami has joined #gnuenterprise psu looks that way *** dneighbors is now known as derek btw: i don't like sunderland's new away kit or the fact they are sitting at the bottom of the league table but am happy they are back in premiere league *** sjc has quit IRC I prefer not to think of it as haxored as much as a forced migration plan is it completely dead? if so i would like to harvest the IP address so that i can forward mail to the new server (mx record is now changed) for any requests coming to the old ip address so i dont lose more mail the onl;y saving grace for sunderland is that newcastle were doing almost as badly yeah but they got own er owen but had spent many £millions more i watched the manchester city vs bolton game poor man city they must have hit the frame at least 10 times the keeper for Bolton was incredible *** sjc has joined #gnuenterprise derek, I doubt you've lost mail... it's just queued up changing the mx record should be all you need postfix isn't running on the old machine website isn't showing anything to me so mail queues will just hold it until the mx record changes psu, squint harder but then my isp is notorious for caching "Alert: the document contains no data." jcater my fear is that the machine has been down several days can't believe that the front page would be firefox-incompatible ;-) and it will take several das for MX to filter out most servers only queue mail for 3 days *** wt has quit IRC *** johannesV_ has quit IRC *** btami has quit IRC *** btami has joined #gnuenterprise *** johannesV_ has joined #gnuenterprise *** btami has quit IRC *** btami has joined #gnuenterprise *** johannesV_ has quit IRC *** psu has left #gnuenterprise *** menomc has joined #gnuenterprise *** mnemoc has quit IRC *** menomc is now known as mnemoc *** btami has quit IRC *** btami has joined #gnuenterprise *** kilo has joined #gnuenterprise *** yangoon has joined #gnuenterprise *** btami has quit IRC *** wt has joined #gnuenterprise *** klasstek has joined #gnuenterprise *** yangoon has quit IRC *** jcater has quit IRC *** jcater has joined #gnuenterprise *** jamest has left #gnuenterprise *** wt has quit IRC *** jcater has quit IRC *** klasstek has quit IRC *** derek has quit IRC *** jcater has joined #gnuenterprise *** kilo has quit IRC *** sjc has quit IRC *** jcater has quit IRC *** jcater has joined #gnuenterprise *** dneighbors has joined #gnuenterprise *** kilo has joined #gnuenterprise *** sjc has joined #gnuenterprise *** johannesV has joined #gnuenterprise good morning good morning hi dimas hello johannesV 'd morning hi kilo dobroje utro! *** btami has joined #gnuenterprise morning kilo: :) (with the help of btami...) hi btami hi dimas hm, it's really tricky to catch that mouse-click to a dropdowns -button ... johannesV: are you working on gtk2 ui now? yeah cool :) hmm :) i have 2 bugs for you then :) ok *** reinhard has joined #gnuenterprise the scrollbar doesn't moves when i click in the bar not on the arrows nor on the slide, but inside the bar another one is the multiline edit when it scrolls down the cursor is out and not visible or something... so the new chars that are input are not visible either yeah yeah, i've already saw that one what about that scrollbar ok try to scroll with mouse click inside the bar which bar ? of a form ? the scrollbar's bar of what ? i mean, how can i reproduce that. how to create such a scrollbar have a small form with a big row-count ? (grid-block) ? yes, grid-block good morning all http://www.gnuenterprise.org/~btami/scrollbar.gfd hi reinhard hi reinhard btami, scrollbar moves properly fro me s/fro/for * ajmitch should see gnue-* in breezy in the next day or so hmm hm, looks like i've found a usable event for gtk2-combos to catch that click-pb johannesV: you can scroll one PAGE up and down with a mouse click? the record is moved to the next/previous if i click on the arrow-buttons, and the record moves as i click the grab-button moving it around btami: does that work in win32 driver? you assume the record to jump page-wise if you click into the scrollbar (not on a button of the scrollbar) ? yep dimas: yes hm, how many records form a page ? the row-count of the block ? yep shift+pgup/pagedown do the same btami: can you check, it does not work if focus is not in a grid? and size of the scroll button does not reflect size of the datasource (pages to scroll) btami, ok, i'll look at that scrollbar-thingy too dimas: in win32? yes will look into dimas: i'v tried it on gtk2, and it doesn't work too i mean the scrollbar, if the focus is not in the grid may be bbl *** btami has quit IRC