_m7u90nt (~mdupont@p5080F90D.dip.t-dialin.net) left irc: "Leaving"
holycow (~rtaylor@h24-86-227-162.vf.shawcable.net) left irc: "Leaving"
sjc (~sjc@cpc2-seve3-4-0-cust112.popl.cable.ntl.com) joined #gnuenterprise.
bddebian (~bddebian@ip68-4-154-50.oc.oc.cox.net) left irc: "Client exiting"
sjc (~sjc@cpc2-seve3-4-0-cust112.popl.cable.ntl.com) left irc: "working"
snewpy_ (~markl@m034-236.nv.iinet.net.au) joined #gnuenterprise.
snewpy (~markl@2001:db0:1019:1:99e2:930e:fec5:17bd) left irc: Read error: 54 (Connection reset by peer)
reinhard (~reinhard@M1250P006.adsl.highway.telekom.at) joined #gnuenterprise.
<reinhard> ajmitch: you here?
<reinhard> ajmitch: any problems with the .deb's?
dimas (~ds@195.218.177.46) joined #gnuenterprise.
<reinhard> bbl
reinhard (~reinhard@M1250P006.adsl.highway.telekom.at) left irc: "I haven't lost my mind. I know exactly where I left it."
ogger (~ogger@57.Red-213-96-148.pooles.rima-tde.net) joined #gnuenterprise.
ogger (~ogger@57.Red-213-96-148.pooles.rima-tde.net) left irc: Remote closed the connection
ogger (~ogger@57.Red-213-96-148.pooles.rima-tde.net) joined #gnuenterprise.
Nick change: snewpy_ -> snewpy
btami (~tamas@napalm.napnet.hu) joined #gnuenterprise.
SachaS (sacha@217-173-231-57.dialin.powersurf.li) joined #gnuenterprise.
<SachaS> hi everyone
jbailey (~jbailey@CPE00062579efac-CM014260028338.cpe.net.cable.rogers.com) joined #gnuenterprise.
<SachaS> see yas
<SachaS> bye
SachaS (sacha@217-173-231-57.dialin.powersurf.li) left irc: "Leaving"
jcater (~jcater@cpe-066-061-071-147.midsouth.rr.com) joined #gnuenterprise.
thierry_ (~thierry@AStrasbourg-251-1-1-7.w82-127.abo.wanadoo.fr) joined #gnuenterprise.
thierry (~thierry@AStrasbourg-251-1-1-7.w82-127.abo.wanadoo.fr) left irc: Read error: 110 (Connection timed out)
jamest (foobar@adsl-66-142-221-144.dsl.tpkaks.swbell.net) joined #gnuenterprise.
jbailey (~jbailey@CPE00062579efac-CM014260028338.cpe.net.cable.rogers.com) left irc: "Client exiting"
wendall911 (~wendallc@wendallc.icehouse.net) joined #gnuenterprise.
holycow (~rtaylor@h24-86-227-162.vf.shawcable.net) joined #gnuenterprise.
holycow (~rtaylor@h24-86-227-162.vf.shawcable.net) left irc: Remote closed the connection
holycow (~rtaylor@h24-86-227-162.vf.shawcable.net) joined #gnuenterprise.
holycow (~rtaylor@h24-86-227-162.vf.shawcable.net) left irc: Client Quit
reinhard (~reinhard@M1250P006.adsl.highway.telekom.at) joined #gnuenterprise.
<chillywilly> stupid question but can you import a class symbol from a module or do you have to fully qualify it like: Module.Class() ?
<reinhard> from Module import Class
<chillywilly> hmmm
<chillywilly> yea
<chillywilly> wtf does it complain that it can't import certain stuff
<chillywilly> blah
reinhard (~reinhard@M1250P006.adsl.highway.telekom.at) left irc: Read error: 104 (Connection reset by peer)
reinhard (~reinhard@M1250P006.adsl.highway.telekom.at) joined #gnuenterprise.
<reinhard> bah
<reinhard> does anybody know how to enter bios setup on a compaq ML370?
<wendall911> reinhard: many compaq systems require a boot disk to access the bios
<wendall911> reinhard: they make good boat anchors ;)
<reinhard> oh my god
<wendall911> reinhard: I had to make a boot disk to even be able to boot to a cdrom
<reinhard> well, i already installed woody and everything works fine
<reinhard> except
<reinhard> 1. i don't know how to silence down the ventilator
<reinhard> 2. i can't access the scsi card (a ncr 3c895)
<wendall911> reinhard: i've had a problem with them before, there is a configuration disk you can run
<Vee2d2> the compaq I have has it's bios stuff on a special partition on the hd
<reinhard> yeah i have such a partition, too
<reinhard> but lilo refuses to boot that partition
<Vee2d2> /dev/hda3             1         3     12064+  12  Compaq diagnostics
<Vee2d2> hmm.. 
<wendall911> someone at Compaq should be slapped for putting bios information on the hd
<Vee2d2> reinhard: is it possibly the F10 key while it's booting up to access?
<reinhard> it says so, yes
<reinhard> but it doesn't
<reinhard> well
<wendall911> http://h18007.www1.hp.com/support/files/server/us/download/13227.html
<wendall911> reinhard: it is a windows executeable
<reinhard> probably i've overwritten the master boot sector with lilo :(
<wendall911> reinhard: I'm sure they must have the disks available in another format
<reinhard> wendall911: cool, thanks
<Vee2d2>     * LILO should be installed to the /boot partition and not to the MBR (Master Boot Record). Installing LILO to the MBR prevents the F10 key from correctly starting the System Configuration Utility on boot up. 
<reinhard> i can use that
<Vee2d2> hehe
<reinhard> i have a windows machine standing around somewhere :)
<reinhard> Vee2d2: oh ok
<reinhard> figured that much :)
<Vee2d2> http://h18000.www1.hp.com/products/servers/linux/compaq-howto-v8-4.html#_4.0 explains how to get around it in lilo's config
<wendall911> reinhard: what's strange is I swear I downloaded the system configuration utility before and it wasn't a windows .exe
<wendall911> reinhard: I couldn't have make the disks ;)
<reinhard> Vee2d2: isn't it great that i work with un*x for 12 years now
<reinhard> and with gnu/linux for 6 years or so
<Vee2d2> :)
<reinhard> and i can still learn new things on a daily basis?
<reinhard> :))
<Vee2d2> ditto
<reinhard> btw thanks all
<reinhard> this is not only the best support channel for debian
<reinhard> this is also the best support channel for virtually anything :)
<reinhard> brb
reinhard (~reinhard@M1250P006.adsl.highway.telekom.at) left irc: "I haven't lost my mind. I know exactly where I left it."
<chillywilly> hmmmm
<chillywilly> read back scroll some
<chillywilly> I've tries using php4 with apache2 and for a lot of php apps I can get apache2 to segfault
<wendall911> chillywilly: depends on the versions you are running
Action: Vee2d2 is happy with apache 1.3.x
<chillywilly> wendall911: I used mpm-prefork apache with a compiled php4 the latest off the website at that time
<chillywilly> I built the damn thing
<chillywilly> dcl for instance will not work 
<wendall911> chillywilly: debian ancient sources, or current sources?
<chillywilly> just segfaults apache2
<chillywilly> sources of what?
<wendall911> chillywilly: apache
<chillywilly> 4.3.4 or some crap
<chillywilly> I used the debian package
<wendall911> chillywilly: heh
<chillywilly> for apache2
<chillywilly> heh?
<wendall911> chillywilly: it is way too old for use
<chillywilly> so why don't they make a new one?
<chillywilly> fucking A
<wendall911> chillywilly: I don't know
<chillywilly> I don't want to build them both
<chillywilly> hell I don't want to build either one
<wendall911> chillywilly: I had some problems on a RH box, same problem
<chillywilly> I want some damn packages
<wendall911> running 2.0.48 and php 4.3.4 without problems
<chillywilly> 2.0.48 is the version of the package
<wendall911> hmm
<wendall911> strange
<chillywilly> I compiled tons of shit into php
<wendall911> chillywilly: I used to get segfaults before 2.0.45, but not any longer
<chillywilly> basically pulled configure swicthes from the script in the php4 package
<wendall911> chillywilly: I use a lot of things also
<chillywilly> this is on debian?
<wendall911> chillywilly: no
<wendall911> chillywilly: under Gentoo, but I have run both from sources as well recently
<chillywilly> hmmm
<chillywilly> well it was unusable
<wendall911> chillywilly: yeah, we had to go back to 1.3 at the university 
<wendall911> chillywilly: Suse packages for 2.x were serving up blank pages all the time, and the segfault issue
<wendall911> chillywilly: it must be something that the package maintainers are screwing up as far as I can tell
<chillywilly> I suppose I could build them both
<chillywilly> that just kinda sucks though
<wendall911> agreed
<chillywilly> I like how debian lays out the config but I suppose I can steal it
<wendall911> chillywilly: yeah, it is easy to setup whatever you like...I have my own that I use
<wendall911> chillywilly: mine is very similar to debian...it is well thought out
<wendall911> the debian way
<chillywilly> it can be a pain in the ass to get something compiled just right
<chillywilly> will all the options you want, etc.
<wendall911> yep
<Vee2d2> > 5) You can keep up with the discussion regardless of what time zone you
<Vee2d2> >    live in
<Vee2d2> Well the trick here is to just not sleep.  That's what Vee does. ;)
<Vee2d2> heh
Action: Vee2d2 looks at derek
<wendall911> lol
<Vee2d2> sleep is for users
reinhard (~reinhard@M1250P006.adsl.highway.telekom.at) joined #gnuenterprise.
<reinhard> for the record
<reinhard> (in case anybody finds this log when googling :))
<reinhard> i managed to run bios setup from the downloaded disks
<reinhard> but no luck
<reinhard> neither with the noise
<reinhard> nor with the 53c895 scsi controller
<reinhard> but however
<reinhard> i will live with the noise and buy an adaptec controller :)
<reinhard> oh, btami fixed win32 mo files
<reinhard> cool
<jcater> reinhard: make sure the adaptec controller card you get works in 2.6
<jcater> in case you want to upgrade later
Action: jcater always liked adaptec controllers, but 2.6 support seems sparse
<jcater> maybe it's just the models I have
<wendall911> jcater: scsi has a long way to go with 2.6
gsoti_away (~gsoti@adsl-68-23-176-107.dsl.chcgil.ameritech.net) joined #gnuenterprise.
dneighbo_ (~dneighbo@ip68-109-180-32.ph.ph.cox.net) joined #gnuenterprise.
sjc (~sjc@cpc2-seve3-4-0-cust112.popl.cable.ntl.com) joined #gnuenterprise.
dneighbo (~dneighbo@ip68-109-180-32.ph.ph.cox.net) left irc: Read error: 110 (Connection timed out)
btami (~tamas@napalm.napnet.hu) left irc: 
Action: ajmitch wonders where jbailey has wandered off to
<ajmitch> I hope he turns up in the next few days
<jcater> I bet it's not to a steak house
<jcater> (just a guess)
<chillywilly> haha
<chillywilly> hey dudes
<ajmitch> can you get soy steak?
<ajmitch> or tofu steak? :)
<jcater> mmm
<ajmitch> hi chillywilly
<jcater> soy steak is great
<jcater> marinate steak in soy sauce overnight
<chillywilly> I think I had an intersting phenomenom occur in my python code
<ajmitch> heh
<jcater> then seer, or grill
<jcater> mmmmm
<ajmitch> sounds nice
<chillywilly> one module was importing the other adn then at the top fo taht one it wanted to import the other...this did not work too well
<ajmitch> I'm wanting him to upload some gnue packages, but he forgot to reply to my last email & upload them :)
<jcater> chillywilly: yeah, have to watch out for that
<chillywilly> hehe
<jcater> we've hit that in gnue many times and had to work around it
Action: ajmitch had better package appserver
<chillywilly> usually means you need to considering shuffling some code around
<ajmitch> before reinhard asks me about it again ;)
<chillywilly> making new modules or something ;)
<jcater> yeah
<reinhard> ajmitch: :)
<chillywilly> cool thing is python did not get stuck in any kinda infinite loop but the modules had little to no symbols in them
<chillywilly> ;P
<chillywilly> would be better if it threw an exception...I suppose it is more overhead to detect circular references in imports
<chillywilly> oh well I learned my lesson
<ajmitch> reinhard: how big were the changes from 0.0.5 to 0.0.6?
<reinhard> quite some
<ajmitch> ok
<reinhard> erm
<ajmitch> I might have to install some example bits then
<reinhard> probably big enough to have it repackaged
<ajmitch> not sure if they are already ;)
<reinhard> if that is the question :)
<ajmitch> oh dear
<reinhard> oh
<reinhard> you could send me the debs to reinhard@gnue.org
<reinhard> for appserver and common
<ajmitch> the packaging is simple though
<reinhard> and i could test if you want
<ajmitch> alright, I will
<reinhard> (tomorrow that is)
<ajmitch> or you could download them from my computer
<chillywilly> so ajmitch did you get that freakin' package yet or do I have to fly to AU pick it up then fly to NZ and deliver it personally?
<reinhard> chillywilly: you mean AT, not AU probably :)
<chillywilly> probably would be faster ;)
<chillywilly> no
<chillywilly> au
<chillywilly> autrailia
<reinhard> oh ok
<ajmitch> chillywilly: no, and I didn't expect to get it yesterday - perhaps monday.. if not, then you can go & beat them
<chillywilly> ok :)
Action: ajmitch is suspiscious of this appserver...
<chillywilly> it's written by commies
<chillywilly> ;P
<chillywilly> commie gnu hippies
<ajmitch> :1:> ls -la gnue-appserver_0.0.*.orig.tar.gz
<ajmitch> -rw-r--r--    1 ajmitch  ajmitch    314304 Oct 23 13:17 gnue-appserver_0.0.5.orig.tar.gz
<ajmitch> -rw-r--r--    1 ajmitch  ajmitch    314297 Mar 14 12:25 gnue-appserver_0.0.6.orig.tar.gz
wendall911 (~wendallc@wendallc.icehouse.net) left irc: "Download Gaim: http://gaim.sourceforge.net/"
<reinhard> rofl
<ajmitch> it's only a few bytes smaller!
<chillywilly> less code?!?
<reinhard> no
<jcater> less code is not a bad thing
<reinhard> we just chaged the code so that gzip can compress it better ;-)
<chillywilly> jcater: sure
<jcater> for a while there, each gnue-forms release was a bit smaller than it's predecessor
<ajmitch> heh
<chillywilly> refactor
<chillywilly> !!!
<reinhard> jcater: while we are talking about code shrinking
<chillywilly> shrinkage
<reinhard> i saw a TODO somewhere about making a general procedure to load plugins
<chillywilly> heh
<reinhard> that could be used to load a dbdriver
<reinhard> as well as an rpcdriver
<reinhard> and maybe even an uidriver
<reinhard> from name
<jcater> reinhard: yes
<jcater> I don't recall the specific TODO
<jcater> but such a thing makes sense
<jcater> one issue you'll have
<reinhard> i think it was in the RPC code
<jcater> is every loader like that we've written
<jcater> btami has had to wrap our code with hardcoded lists
<jcater> to get it working with mcmillan
<jcater> which I hate, but I guess that's what he has to do
<jcater> well, hate is a strong word
<jcater> I dislike
<jcater> anyway, my point is
<reinhard> this is the DRIVERS = [...]
<reinhard> stuff?
<jcater> yes
<reinhard> in the __init__.py files?
<jcater> though a more prevelant example
<jcater> look in gnue-designer/src/templates/__init__.py
Action: ajmitch mutters
<chillywilly> defeats the whole purpose ;) now you have a stinking list to maintain ;)
<jcater> search for McMillan
<ajmitch> I shoul really install the new gnue common packages before trying to package appserver & forms
Action: ajmitch loves sudo
<jcater> with designer's wizards/templates, I very much wanted it to be where anyone could drop in a .py file, and have a wizard available
<ajmitch> Password:
<ajmitch> Are you on drugs?
<ajmitch> Password:
<jcater> but with this wrapper approach, that will be harder to do
<jcater> anyway, my point is
<jcater> if we do a generic loader like that
<jcater> you have to take into account this necessary workaround for McMillan
<jcater> (or we have to find a better solution for that case)
<reinhard> ok
<reinhard> thanks for the hint
<chillywilly> building a cross compiler takes a while
<chillywilly> la la la
<ajmitch> E: gnue-appserver: binary-without-manpage gnue-gsdgen
<ajmitch> who wants to write a manpage for me? :)
<ajmitch> ooh, there is one!
Action: ajmitch looks to see why it wasn't picked up
<reinhard> ajmitch: it's new
<reinhard> wasn't there in 0.0.5
<ajmitch> yes
<ajmitch> and it's not recorded in setup.py
<reinhard> jcater: you remember why that workaround was necessary?
<reinhard> ajmitch: oh
<jcater> reinhard: btami is better person to ask
<jcater> but
<jcater> iirc, we work by doing a directory scan
<jcater> looking for python modules
<jcater> in mcmillan, there is no directory to scan
<jcater> mcmillan packs all python in its own internal structure
<reinhard> oh
Action: chillywilly borrowed some of that code
<jcater> and intercepts python's module loading code
<reinhard> i understand
<jcater> reinhard: do you have all of gnue checked out?
<jcater> including gnue-pos?
<ajmitch> ok, that problem's fixed..
<reinhard> i think so
<ajmitch> I have it all checked out
<jcater> look at gnue/gnue-pos/src/POSClient.py, around line 74
<chillywilly> I really need to learn to make a debian package
<jcater> I'd prefer if the module loader worked that way
<chillywilly> ajmitch: can you make me apache 2.0.48 and php packages ;)
<jcater> i.e., in the dbdriver part
Action: ajmitch does an svn up
<chillywilly> hehe
<ajmitch> chillywilly: nope, sorry, I don't have that much time free
<jcater> if you have provider=foobar
<reinhard> ajmitch: fixed
<jcater> the loader would first try
<ajmitch> reinhard: thanks
<jcater> import gnue.common.datasources.drivers.foobar
<jcater> then if that didn't work, would just try
<jcater> import foobar
<chillywilly> ajmitch: but they don't have to be perfect or even follow debian policy ;P
<jcater> I do that to get customizations for gnue-pos
<jcater> i.e., I have a python module on my machine called market (my store)
<jcater> so for some of the drivers specific to my setup, I can do
<jcater> provider=market.mydriver
<jcater> and not have to put market in the gnue.common.* source tree
<jcater> just a thought...
<jcater> but would allow others to extend gnue better
<reinhard> not sure
<reinhard> you have to remember
<reinhard> we have a hierarchical driver structure
<reinhard> that is for example
<reinhard> popy driver is in 
<reinhard> gnue.common.datasources.drivers.postgresql.popy
<reinhard> so it basically has to search through all subpackages of "drivers"
<reinhard> to find out where popy is
<chillywilly> dircache :)
<jcater> I still think it could work
<jcater> i.e.,
<jcater> if provider=popy:
<jcater> You first look for 
<jcater>   gnue.common.datasource.drivers.popy
<jcater>   exists --> try: 
<jcater>       gnue.common.datasource.drivers.popy.popy
<jcater>         if doesn't exist --> find first working driver under drivers.popy
<jcater>   not exists --> search thru all drivers.*.* for a popy
<jcater>   if that fails, try to import popy
<jcater>   
[17:47] Last message repeated 1 time(s).
<jcater> the only thing different from what we currently do is that last "if that fails..."
<reinhard> jcater: ok, gotcha
<ajmitch> alright, I'll bbl :)
<reinhard> jcater: also, i think a means of --list-drivers could be cool
<jcater> definitely, yes
<ajmitch> reinhard: I'll give you the URL of where my packages are built tomorrow
<ajmitch> bbl
<reinhard> what is already working now is
<reinhard> thanks ajmitch
<reinhard> what is already working now is
<reinhard>   for m in list ('gnue.common.datasources.drivers', 'Connection'):
<reinhard>     print m.__name__
<reinhard> this issues
<reinhard> gnue.common.datasources.drivers.special.configfile
<reinhard> gnue.common.datasources.drivers.special.static
<reinhard> gnue.common.datasources.drivers.special.unbound
<reinhard> gnue.common.datasources.drivers.sybase.sybase
<reinhard> gnue.common.datasources.drivers.ingres.ingres
<reinhard> gnue.common.datasources.drivers.sapdb.sapdb
<reinhard> gnue.common.datasources.drivers.csv.csv
<reinhard> gnue.common.datasources.drivers.informix.informix
<reinhard> gnue.common.datasources.drivers.dbf.dbf
<reinhard> gnue.common.datasources.drivers.sqlrelay.sqlrelay
<reinhard> gnue.common.datasources.drivers.sqlite.sqlite
<reinhard> gnue.common.datasources.drivers.adodbapi.adodbapi
<reinhard> gnue.common.datasources.drivers.postgresql.pypgsql
<reinhard> gnue.common.datasources.drivers.postgresql.psycopg
<reinhard> gnue.common.datasources.drivers.postgresql.pygresql
<reinhard> etc.
<reinhard> list (a, b) means
<reinhard> list all modules under a that define symbol b
<jcater> is list() a function you did, or some feature of python I didn't know about?
<reinhard> i did it :)
<jcater> nice
<reinhard> and btw
<reinhard> of course it doesn't need the DRIVERS = []
<jcater> I need that in my website generation thing
<reinhard> it works with listdir :)
<reinhard> or (mcmillan speaking)
<reinhard> it works with listdir :(
<jcater> there *has* to be some way to get at mcmillan internals
<jcater> to do an equivalent function there
Action: jcater hasn't looked, but there's bound to be something
<reinhard> i figured there is no other way to find out the modules and subpackages of a package than to read the directory
<jcater> even if it's undocumented
<reinhard> as a last resort
<jcater> I wonder how epydoc does it
<jcater> probably the same way
<reinhard> i'd nearly say we'd generate the __init__.py with the DRIVERS = [] dynamically on .exe generation
<reinhard> if that works
<jcater> that's an idea
<jcater> I'd rather take that route
<jcater> than give up this dynamic loading
<chillywilly> is epydoc better than the pydoc module?
<reinhard> jcater: while we're talking about drivers
<chillywilly> I was considering making my mod_python server be self documenting cause I am lazy like that, but pydoc has little to no documentation
<jcater> I like epydoc
<jcater> jamest picked it from the others
<chillywilly> ok
<jcater> not sure why he picked it vs, say, pydoc
<jcater> but I'm sure he had a good reason
<reinhard> i think the very low code per file ratio in the dbdrivers makes the directory structure quite complex
<reinhard> it might be much easier to overview if we had a single file for a driver
<reinhard> (at least for the DBSIG2 ones, which are generally only a few lines of code)
<reinhard> but that's a personal opinion
<jcater> I'm indifferent
<jcater> personally, I don't mind the individual files
<jcater> but yeah, some of them are rather small
<jcater> but for the bigger ones, I do like them split
<jcater> and I guess I was just being consistent
<reinhard> take the postgresql Base driver for example
<reinhard> the RecordSet and ResultSet files are completely useless (AFAICT)
<reinhard> the 4 actual drivers for postgres are not more than 10 lines of code each
<reinhard> (plus comments and documentation for website)
<jcater> possibly
<reinhard> and when i tried i've made it to break them down to 2 lines each :)
<jcater> we've whittled them down to nearly nothing
<jcater> I'm sure they were bigger at one point :)
<reinhard> sure :)
<jcater> I'm looking at mcmillan source code
<reinhard> don't get me wrong
<reinhard> not trying to offend you or anything
<jcater> and I bet we could patch it if nothing else
<reinhard> just thinking out loud
<jcater> I'm indifferent
<jcater> on one hand, they are really small
<jcater> (almost too small)
<jcater> on the other, they are consistent with the other drivers which may be larger
<jcater> but then again, we've put so much into the DBSIG2 driver now, none of them are really big
<reinhard> yeah
<reinhard> except for appserver, dbf and specialfile
<reinhard> which aren't DBSIG2 based
sjc (~sjc@cpc2-seve3-4-0-cust112.popl.cable.ntl.com) left irc: "sleeping"
jamest (foobar@adsl-66-142-221-144.dsl.tpkaks.swbell.net) left irc: Remote closed the connection
jamest (foobar@adsl-66-142-221-144.dsl.tpkaks.swbell.net) joined #gnuenterprise.
<jcater> hmm
<jcater> there's an ArchiveViewer.py in McMillan's source tree
<jcater> which does exactly what its name implies
<jcater> gets file names, etc, from McMillan archives
<reinhard> hmmm
<reinhard> sonds very interesting
<reinhard> we have another (easy to solve) problem
<reinhard> because we can't distinguish real from base drivers
<reinhard> we will get drivers like Base or DBISG2 also listed
<reinhard> we can suppress this by not exporting a Connection there
<reinhard> but for example a _Connection
<reinhard> or a Connection_
<reinhard> or a BaseConnection
<jcater> my rule of thumb was to only catch lowercase module names
<reinhard> you'll get a problem under windows with that :)
<jcater> but if need be, we can go back to the old way, of prefixing the former with _
<jcater> so it would be _dbsig instead of DBSIG
<reinhard> not necessary
<reinhard> because i check for a specific symbol
<jcater> ok
<reinhard> we can change the identifier "Connection" to be different
<jcater> ok
Action: ajmitch returns
<jamest> lol
<jamest> We upgrade to subversoin 1.0
<ajmitch> hmm?
<jamest> yesterday 1.0.1 is released
<ajmitch> heh
<ajmitch> typical :)
<jcater> what's the easiest way to do a feature request to a debian maintainer?
<jcater> or bug report
<jcater> is there a handy dandy client, or have to go to a web page?
<reinhard> reportbug
<jcater> thanks
<ajmitch> who are you harassing?
<jcater> two things
<jcater> 1. apache2's /etc/init.d/apache2 script gives you the wrong path to /etc/default when it tells you to edit that file
<ajmitch> I  hope you're not filing bugs against gnue packages..
<jcater> 2. It'd be nice if subversion's svn and svnadmin were shell scripts that first sourced in something like /etc/default/svn, then ran the actual svn clients
<jcater> so you can set umask there
<ajmitch> 1. ah yes, I see what you mean there..
<jcater> (as that is what the svn handbook tells you to do, oddly enough)
<jcater> w00t
<jcater> reportbug is nice
<jcater> shows me my bug is already reported
<ajmitch> already reported, already ignored ;)
<jcater> apparently
<reinhard> jcater: how does that look like:
<reinhard> find "postgresql.popy": gnue.common.datasources.drivers.postgresql.popy
<reinhard> find "pygresql": gnue.common.datasources.drivers.postgresql.pygresql
<reinhard> find "mysql": gnue.common.datasources.drivers.mysql.mysql
<reinhard> find "oracle": gnue.common.datasources.drivers.oracle.cxoracle
<reinhard> find "nonexistent": None
<jcater> that last line, how did it arrive at that?
<jcater> well, the oracle one
<reinhard> like before
<reinhard> it chooses the first driver it finds under "oracle"
<jcater> this doesn't do any initializing, does it?
<reinhard> it tests if the import goes through
<reinhard> i'm sure i have no oracle libs installed here :)
<jcater> the way it was intended to work
<jcater> is if you specified "oracle", it would return the first driver that had all the prerequisites
<jcater> so I'm wondering if find should return a tuple if more than one could work
<jcater> and let the init code cycle thru until it finds one that does
<reinhard> or pass a function to find() that does the test
<jcater> (does that make sense?)
<jcater> or, yeah
<jcater> either way
<reinhard> or...
<reinhard> we define every plugin must have a function to test whether all prerequisites are fullfilled?
<reinhard> hmmm
<reinhard> i think the testfunction parameter to find() is the most generic way
<jcater> I think all find has to do is catch the AdapterNotInstalled exception
<jcater> (which might need to be moved into a generic .exceptions module)
<jcater> I'm ok with either approach
<jcater> (either a testfunction to find(), or a driver.loadable() test
<reinhard> my thinking is
<reinhard> the testfunction would get the module as parameter
<reinhard> it can always do
<reinhard> def testfunction (m):
<reinhard>   return m.loadable ()
<reinhard> :)
<jcater> true
<reinhard> would you be ok if we defined the testfunction
<reinhard> to raise an exception if the test fails
<reinhard> and we can GDebug the exception message
<jcater> I imagine
<reinhard> hmmm
<reinhard> currently there's no way in our dbdrivers to find out if prerequisites are installed
<reinhard> other than try to establish a connection
<reinhard> well, to instantiate a connection object
<jcater> try: import Connection as _test; except AdapterNotInstalled: return 0
<jcater> iirc
<jcater> or that was the idea
<reinhard> it doesn't work
<reinhard> that doesn't fail
<reinhard> find "postgresql.popy": gnue.common.datasources.drivers.postgresql.popy
<reinhard> find "pygresql": gnue.common.datasources.drivers.postgresql.pygresql
<reinhard> find "mysql": None
<reinhard> find "oracle": None
<reinhard> find "nonexistent": None
<reinhard> i had to test with
<reinhard>   module.Connection (None, None, {})
<reinhard> not sure if that succeeds on every module
<reinhard> no doubt best would be a module.Test function
<reinhard> ok
<reinhard> i've committed gnue.common.apps.plugin
<reinhard> i'll let you look at it
<reinhard> the code that searches for all subpackages and modules in a package is in __modules ()
<reinhard> probably the fix for McMillan has to go there
<reinhard> it's time for bed for me
<reinhard> night all
reinhard (~reinhard@M1250P006.adsl.highway.telekom.at) left irc: "If you think there is good in everybody, you haven't met everybody."
gsoti_away (~gsoti@adsl-68-23-176-107.dsl.chcgil.ameritech.net) left irc: "Client Exiting"
gsoti_away (~gsoti@adsl-68-23-176-107.dsl.chcgil.ameritech.net) joined #gnuenterprise.
holycow (~rtaylor@h24-86-227-162.vf.shawcable.net) joined #gnuenterprise.
gsoti_away (~gsoti@adsl-68-23-176-107.dsl.chcgil.ameritech.net) left #gnuenterprise.
Nick change: dneighbo_ -> derek
jamest (foobar@adsl-66-142-221-144.dsl.tpkaks.swbell.net) left #gnuenterprise ("Kopete 0.8.0 RC1 : http://kopete.kde.org").
<Vee2d2> I love coffee. 
<chillywilly> ok
<chillywilly> good for you
<Vee2d2> whats you been upto lately, chillywilly?
<Vee2d2> still sculpting greatness out of perl?
<chillywilly> nope
<chillywilly> python
<Vee2d2> cool
<chillywilly> :)
<Vee2d2> hey, wasnt it you that brought up that REST stuff a while back?
<chillywilly> yea
<Vee2d2> a cow-orker ran acrost it the other day and I seemed to recall it getting kicked around in here for a spill (in discussion, I mean)
<chillywilly> well it is an intersting take on the whole web architecture
<ajmitch> hi Vee2d2, chillywilly
<Vee2d2> 'sup ajmitch, how's nz?
<ajmitch> it's ok
<Vee2d2> hah, donald trump is gonna host SNL on 4/3
<holycow> heh
<holycow> the king of combovers
<holycow> he mast have a pretty good cueball by now
--- Sun Mar 14 2004