*** bigbrother has joined #gnuenterprise *** bigbrother has joined #gnuenterprise bad bigbrother * jcater spanks bigbrother corrupting your own config file careful big brother likes spankings *** wendall911 has joined #gnuenterprise can someone give svn a kick to the nads? svnserve 29990 0.0 0.1 2716 408 ? S Jun13 0:00 telnet 62.49.129.34 12333 wtf is that? no clue populous.netsplit.com? this is scaring me ash:~# ps aux|grep svn svnserve 29720 0.0 0.2 2468 564 ? S Jun13 0:00 /bin//sh svnserve 29990 0.0 0.1 2716 408 ? S Jun13 0:00 telnet 62.49.129.34 12333 svnserve 30127 0.0 0.2 2468 560 ? S Jun13 0:00 /bin//sh hmmmm looks suspicious *** bigbrother has joined #gnuenterprise *** bigbrother has joined #gnuenterprise jcater: The write perms for anon access bother me too. But the svn people say they are coming up with a new backend that doesn't have that limitation. *** sjc has quit IRC I don't see any security alerts on the debian page for svn but I see 2 for cvs tongue of frog has the oil :-/ *** bigbrother has joined #gnuenterprise *** bigbrother has joined #gnuenterprise jcater: http://subversion.tigris.org/security/CAN-2004-0413-advisory.txt Workarounds: ============ no known exploits, my ass :) Disable svnserve and use DAV (http://) instead. someone should tell #svn there are exploits well they say that a sophisticated attacker can so some fun stuff... looks like someone found something fun to do... fwiw, 62.49.129.34 hosts svn services so looks like they were relaying from one svn host to another interesting... I'm busy with ash, but could someone find an email address on that site and email their sysadmin? I think that would be common decency to do to tell them to upgrade operatpr@netsplit.com operator or svnmaster@netsplit.com i will email them if someone else has not go ahead *** dsmith has quit IRC *** havoc has quit IRC *** ajmitch has quit IRC *** dimas has quit IRC *** Vee2d2 has quit IRC *** Morphous has quit IRC *** kilo has quit IRC *** kfogel has quit IRC looks like the guy was executing a shell then telnetting to some weird port their admins have been notified *** Vee2d2 has joined #gnuenterprise *** Morphous has joined #gnuenterprise *** kilo has joined #gnuenterprise *** dsmith has joined #gnuenterprise *** dimas has joined #gnuenterprise *** havoc has joined #gnuenterprise *** ajmitch has joined #gnuenterprise *** kfogel has joined #gnuenterprise jcater: are you running checkrootkit or whatever *** btami_ has joined #gnuenterprise *** reinhard has quit IRC jcater or jamest are you here? i believe the refresh="change" for foreign keys is not working (as i expect it to) or i am doing somethign fundamentally wrong i have a simple form to replicate what i am looking to do and how it appears to be failing also i think i have other bigger issues with F/K even if this is fixable basically we have some foreign key lookup lists that have tens of thousands of records that need to be filted down issue #1 1. i can not seem to query in real time the list down that is if i have the list and i try and initquery, execute query on the block that is the fk datasource it does not update the list even if i have refresh="change" set issue #2 2. it appears that these lists get prequeried even if they are not set to prequery="Y" this is major problem for speed reasons having load 11,000 records in a dropdown to only later be queried down to 4 choices seems like a waste of resources i think i need to see if htere is refreshkey method i could manually call via trigger chillywilly: yeah k crap must run to lunch jcater: i see refreshForeignKey() has been obsoleted explains why it does nothing :) derek: i reactivated 'refreshForeignKey ()' today; so you would need an svn update to get it right back :) * johannesV_ back in 5 minutes *** johannesV_ has quit IRC *** ChtiTux has joined #gnuenterprise *** ChtiTux has quit IRC *** johannesV has joined #gnuenterprise * johannesV back :) johannesV cool * derek will give it a whirl glad you said something as i was strarting to resurrect it before i ran to lunch :) grr i deleted GFField.py to update andnow svn appears to be freakign out jcater: i dont want to hose anything neighborsd@soleri:~/svn/gnue$ svn update Password: svn: Berkeley DB error while opening environment for filesystem /var/svn/gnue/db: Invalid argument neighborsd@soleri:~/svn/gnue$ yeah hang tight okay just let me know when its kosher I hate computers * jcater needs a career change desparately *** johannesV has quit IRC *** reinhard has joined #gnuenterprise jcater: What did the #svn people say was the problem? jcater: is it safe to svn up yet? dsmith: we should have upgraded over the weekend derek: not quite compile of svn not done dokie * derek notes mv instead of rm is safer to get rid of svn files before an update :) derek: svn is back up thanks slurped for the record refreshForeignKey() still no worky pooh er reset *** kilo_ has joined #gnuenterprise fsck resetForeignKey() REQUERIES it sigh so i see my "filter" works *** kilo_ has quit IRC but the reset reDraws it which is part of my second problem in that it "auto" loads dropdowns the autoload is happening as part of the loadValues fsck fsck fsck self.__loadAllowedValues (self._fk_datasource.createResultSet ()) i really want this to be something like hmmm i might have an idea grrr its a bitch to debug triggers jcater: if i have dtsFoo how do i access its "result set"? i mean the actual resultset object not traverse through result set * derek has a suspcion that its not available to triggers? * derek is going to try blkName.resultSet but expect not much luck try print dir(blkName) or print dir(dtsFoo) I don't recall if/how to do that okay hmm blkName appears to have a .firstRecord which makes me think its the result set object but didnt get desired results * derek kicks things it appears to query properly and it appears to not load the world but now its coming back blank fsck resultset is none but it exists you little bi0tch sigh i dont think there is anything there is a createResultSet but i think that actually requeries and blk has init and execute but ive already done that * chillywilly kicks derek *** jcater has quit IRC *** dsmith has quit IRC *** btami_ has quit IRC *** havoc has quit IRC *** ajmitch has quit IRC *** dimas has quit IRC *** Vee2d2 has quit IRC *** Morphous has quit IRC *** kfogel has quit IRC *** kilo has quit IRC *** wendall911 has quit IRC *** Stoke has quit IRC *** reinhard has quit IRC *** btami has quit IRC *** nickr has quit IRC *** reinhard has joined #gnuenterprise *** kfogel has joined #gnuenterprise *** ajmitch has joined #gnuenterprise *** havoc has joined #gnuenterprise *** dimas has joined #gnuenterprise *** dsmith has joined #gnuenterprise *** kilo has joined #gnuenterprise *** Morphous has joined #gnuenterprise *** Vee2d2 has joined #gnuenterprise *** wendall911 has joined #gnuenterprise *** btami has joined #gnuenterprise *** Stoke has joined #gnuenterprise *** nickr has joined #gnuenterprise *** jcater has joined #gnuenterprise night all *** reinhard has quit IRC okay now im gonna really kill this bastard i created a getResultFunction and it works er getResultSet and it appears that hte created dropdown now works BUT it STILL dont repaint the dropdown right oh wait it does freaky but youhave to exit the field and come back yeah... minor freaking vitory victory jcater: is getResultSet a good name? or would be getResultSetObject or something be better? i.e. i dont want some one to confuse getResultSet and createResultSet blah blah blah blah blah blah well its committed now * derek does the happy dance blkName.fldName.resetForeignKey(blkFKName.getResultSet()) is how you can reset a block with custom result set *** dsmith has quit IRC night *** btami has quit IRC jamest / jcater fundamental issue with fk datasources auto querying on startup im assuming there was a reason? hell i might have been the one that complained for that feature :) but i think we need institute some way to stop that behavior as it will kill performance for accounting applicatiosn or anything that does dynamic list box values i was going to look at this thinking maybe a gnue.conf file about "auto query fk" y/n would work if "no" is chosen then it woudl look at the query on startup property other wise it by default auto load * derek wants people to comment before i go at it but i dont want to "blah/blah/blah" as chilly woudl say HEH jsut messing with you just* shouldn't be a gnue.conf value should be connection-related i think so you could tell which connections to auto-load and which dont honestly when i think about it.. it should just be a datasoruce thing if i say prequery prequery if i dont dont BUT i think the issue was people didnt get the concept of prequery and so they wouldnt flag it and then their combo boxes would be blank im seeing a queryDefault property * derek now thinks perhaps i could have used that as it gets called by load values BUT ultimately i think it better to have custom queries that you can pass the result to the refresh anyhow... anyone off top of head know where prequery gets read :) somewhere there in... a little bit to the left... err... read this! *** bluesbaron_ has joined #gnuenterprise *** bluesbaron_ has left #gnuenterprise okay i have it reading from prequery value again so question do i make my own "prequery" value (keep from breaking backwards) do i make change and force people to update forms to "prequery" if they really want prequery or last and imho least desirable option make a hack in gnue.conf or connection object to determine whether to read prequery value of the datasource for fk lookups afaik, we've always said you have to use prequery nope so I'd say we just enforce that okay works for me :) seeing how i was probably the stupid bastard that requested we be nice and auto prequery for them BUT i noticed i thinmk if you use wizard for fk in designer derek: if we told you differently, I'm sure we thought you'd never actually use it anyway * jcater ducks it sets to true for you hey... i figure if i commit enough you guys will get unbusy really quick or i will get refamiliar with the code and be dangerous *** dyfet has joined #gnuenterprise *** Vee2d2 has quit IRC *** kilo has quit IRC *** dhill1 has quit IRC hey dyfet *** Vee has joined #gnuenterprise *** wendall911 has left #gnuenterprise *** dcmwai has joined #gnuenterprise *** dyfet has quit IRC *** bluesbaron_ has joined #gnuenterprise *** bluesbaron_ has left #gnuenterprise *** reinhard has joined #gnuenterprise *** johannesV has joined #gnuenterprise *** kilo has joined #gnuenterprise good morning did irc-logs moved, or is there something wrong with the logging-facility ? good morning i think btami or kilo already noted yesterday that the logs are gone good morning aye, jcater kicked bigbro' alive again sometime in the afternoon ah, ok thanks * reinhard goes to finally release appserver, forms, and common hey what's that??? svn works !?! sthing must be wrong... 8-)) *** btami has joined #gnuenterprise *** johannesV has quit IRC *** johannesV has joined #gnuenterprise *** Vee has quit IRC *** Vee has joined #gnuenterprise *** dcmwai has quit IRC so release is out ahhh wow a release still takes 2-3 hours i mean just the release process not the coding involved ;-) reinhard: you installed Mandrake 10, right? is it comes with wx-gtk2 ? yes and it seems to work although i didn't try designer only forms off to lunch hmm, so it's time to support it in gnue... *** kilo has quit IRC *** dsmith has joined #gnuenterprise can anyone reach google? there it goes...weird nevermind then *** kilo has joined #gnuenterprise *** Vee2d2 has joined #gnuenterprise bbl *** btami has quit IRC *** Vee2d2 has quit IRC *** Vee2d2 has joined #gnuenterprise reinhard : at one time i had documented the steps i think for releases we need detailed step by step guide as soon as we have that we can automate some of the steps but worse case any old slug can do a release that way core developers are not spending time on packaging *** Morphous_ has joined #gnuenterprise reinhard: I guess moving to a source install of subversion was worth it :-/ derek: http://www.gnuenterprise.org/tools/common/docs/technotes/00011.txt *** Morphous has quit IRC *** kfogel has left #gnuenterprise *** holycow has joined #gnuenterprise any of you want a gmail account? jcater: sweet *** jamest has joined #gnuenterprise I've been crazy busy with everything else (building house, swamped at work with new projects) so I've slipped into park in my keeping up with gnue.. is there a small business solution out there now? I saw Sachas did something called Baby-ERP the screenshots looked pretty interesting wtf.... we have a small print shop here and are shopping for a solution. gnue is misbehaving here in the strangest way we actually bought printer's plan, but dont like that it doesnt have an accounting package (they expect you to integrate with quickbooks) derek: yeah, reinhard, et al, are breaking our tradition and actually documenting stuff :-/ Vee2d2, from what i can tell getting a small business solution wouldn't be all that hard if we were to take sachas work and move forward on that jcater: heads will roll!!! ;) i think more than anything gnue needs a stable and a development branch tho jcater: okay i have one for you holycow: we did that once yesterday i left work and forms was fine and decided we weren't far enough along i come in this morning and i launch a form does dcmwai still come around? Are they (arias) still interested/attempting merged efforts? and it prompts for login i enter login and it just hangs no splash screen no errors nothing so i wiped my tree redid setup-cvs.py and tried again same things still happening i cant seem to find any instances of forms running elsewhere (it is freaking wild) strange im at a complete loss if i were on windows i would jsut reboot have you run with --debug-level high? does it consistently hang at same time? or perhaps does it hide a traceback? yes no traceback * derek thinks debug high might have clues just a second what is the output when it hangs? neighborsd@soleri:~/svn/omb/budget/forms$ ~/bin/gfcvs resetFK.gfd ['BINARY', 'BOOLEAN', 'Binary', 'DATE', 'DATETIME', 'DataError', 'DatabaseError', 'Date', 'DateFromMx', 'DateFromTicks', 'Error', 'FLOAT', 'INTEGER', 'INTERVAL', 'IntegrityError', 'InterfaceError', 'InternalError', 'LONGINTEGER', 'NUMBER', 'NotSupportedError', 'OperationalError', 'ProgrammingError', 'QuotedString', 'ROWID', 'STRING', 'TIME', 'Time', 'TimeFromMx', 'TimeFromTicks', 'Times and thats it I meant from the debug-level stuff do you get tracebacks with debug-level and does the last debug-level output look the same each time DB007: [DataObject:293] Implicit Fields: {u'agency_class_name': '', u'agency_class': ''} DB005: [DataObject:334] SELECT agency_class_name,agency_class FROM agcl DB001: [Connection:277] DBSIG2 Statement: SELECT agency_class_name,agency_class FROM agcl DB001: [Connection:278] DBSIG2 Parameters: None is where it hangs right now wild i can get to that postgresql server via psql but a select * from agcl; seems to hang * derek thinks i have issues on the postgres server I suspected as much jcater, thank you for the info, i didn't know :) *nod* understood whew all is back to normal i think just have REALLY old machines i think a process to index file system (updatedb) or something was running just put all postgres queries to a halt all is good in oz now http://www.gnuenterprise.org/developers/copyleft.php <-- there a typo in the finally section explains instead of explain section: wtf i could have sworn i committed something to honor prequery="N" in foreignkey datasources but an update of most current svn doesnt have it * derek scratches head wtf my resetForeignKey() chagnes are gone as well * derek triple scratches head what the hell is wrong with svn? * derek KNOWS i committed the resetForeignKey() properly as another developer her svn updated and got the change it uses a database to store files *** johannesV has quit IRC *** johannesV has joined #gnuenterprise *** dneighbo has joined #gnuenterprise jcater svn is fscked in the head BAD i come to david's machine he has BOTH changes that i committed yesterday that you and i DO NOT have after an svn update so yeseterday svn had these changes BUT here is the freaking kicker if i do an svn diff GFField.py it shows that his file only has ONE of the TWO changes NOT committed to his machine it thinks SVN has the resetForeignKey() change that we are not seeing something is broked um fwiw that would be a client-side issue i am at his machine now if it helps trouble shoot not server side but my point is any chance you edited on his maching how did his client get the changes machine and just didn't realize it? the ONLY way was to svn update no i did NOT edit anything on his machine that i am positive of i made chagnes on my machine committed came over and had him udpate and things worked (yesterday) and it holds true as his system still has those changes but you said it shows them as being edited on his machine no the resetForeignKey and the other one i think is a fluke because svn is fscked up i.e. i did NOT make that change on his machine he updated to get it but now svn appears to think that doesn't exist hilld001@minardi:~/svn/gnue/gnue-forms/src/GFObjects$ svn commit Password: Sending GFObjects/GFField.py Transmitting file data .svn: Commit failed (details follow): svn: Out of date: '/trunk/gnue-forms/src/GFObjects/GFField.py' in transaction '54w' svn: Your commit message was left in a temporary file: svn: '/home/hilld001/svn/gnue/gnue-forms/src/svn-commit.tmp' hilld001@minardi:~/svn/gnue/gnue-forms/src/GFObjects$ fsck i lost a change now i had some change in GFBlock.py no idea what i didnt expect it to update cna you make a .py file otu of .pyc file? jcater can you update and see if resetForeignKey() in gffield.py has two params or one? *** wendall911 has joined #gnuenterprise okay GFBlock.py changes made i think i restored my changes can someone PLEASE verify they got committed i.e. do an svn update and check GFBlock.py to see if getResultSet() exists and check GFField.py to see if resetForeignKey() has 2 parameters instead of one it would be GREATLY appreciated *** dneighbo has quit IRC back for a second to clear up some confusion okay i think im back to where i was the other day derek: i had to undo your commits from yesterday because they broke existing forms and the "print 'hi mom'" you added wasn't helpful, either that was a mistake that i meant to clean this morning the "hi mom" you can get back to the state where you were yesterday evening by checking out the revision well i rechecked my changes back in so forms is again broken? i tested for backwards compatiablity on the resetForeignKey() that shoudl work i didnt find out until i walked out the door that the prequery change appears to break SOME instances of forms i checked that back in but commented out the section that was breaking things