*** reinhard has joined #gnuenterprise
*** sjc has joined #gnuenterprise
*** johannesV has joined #gnuenterprise
*** aries has joined #gnuenterprise
<aries> hi
* aries looking for v2.4 python and wxpython
<reinhard> hi
<reinhard> I have no idea where you could get these, sorry :(
*** dcmwai has quit IRC
<aries> reinhard: how can i accelerate the designer's migration to v2.5?
<reinhard> aries: I have never looked at designer at all, so I can't really tell
<reinhard> my understanding is that you just have to fix that stuff that doesn't work any more ;-)
<reinhard> maybe talk to jcater
<reinhard> but he is U.S. timezone
<reinhard> I am quite confident that he would appreciate any help in 2.5 migration
<reinhard> as it's getting more and more of a problem
<johannesV> do we have a concept of distinct in datasources ?
<johannesV> yeah, we do
<johannesV> using distinct="True" in the datasource tag
<johannesV> but it does not work with appserver ...
<johannesV> but have to check it without appserver
*** jproby has joined #gnuenterprise
<aries> reinhard: ok :)
*** sjc_ has joined #gnuenterprise
<johannesV> ok, i have the results
<johannesV> gnue.common supports 'distinct' in general (using two-tier) but appserver does not
<johannesV> that's bad
<johannesV> the problem is: how to pass such information via appserver's rpc-interface
<johannesV> it would be another param to the request method, wouldn't it ?
<reinhard> hmmm
<johannesV> a better idea would be to add something like 'hints' to a request
<johannesV> which could be a dictionary for example
<reinhard> I've actually never considered distinct selects as a feature needed in appserver
<johannesV> allowing us to add a lot of such things, like: 'distinct': True, '...': ...
<johannesV> i doo
<johannesV> erm
<johannesV> i do
<johannesV> shall i give an example ?
<reinhard> yes please
<johannesV> ok
<johannesV> i have artists
<johannesV> i have discs
<johannesV> and i have tracks
<johannesV> where a disc has multiple tracks
<johannesV> a track can have multiple artists
<johannesV> and now i want to know: which albums have tracks from a given artist
<johannesV> but i want to see every album only once
<reinhard> brb phone
<johannesV> there's another problem we've to solve: direction of an 'order by' (asc/desc)
<johannesV> but this has a low priority
*** sjc has quit IRC
<reinhard> johannesV: I really agree with the asc/desc problem
<reinhard> however with distinct I fear we're opening pandoras box
<reinhard> and will end up implementing complete SQL functionality in appserver
<reinhard> I am really unsure what to do here
<reinhard> what other things would come next?
<reinhard> what else is possible in SQL besides distinct?
<reinhard> subselects?
<reinhard> outer joins?
<reinhard> group by?
<reinhard> unions?
<johannesV> right, group by is a good point :)
<johannesV> but, what solution can you suggest in my problem then ?
<reinhard> one possible solution would be to automatically maintain (in OnValidate or such) a class that contains the artists per disc
<reinhard> but spontaneously I don't like that very much
<johannesV> ok, so that way i would have another crossref-table
<johannesV> (which is completely redundant)
<reinhard> yep
<johannesV> is that an english word ?
<reinhard> that's why I don't like it
<reinhard> yes, I think it's proper English :)
<johannesV> i think having a master->detail->detail construct would be very common
<johannesV> and having such situations will rise the need of special handling
<reinhard> hmmmmm
<reinhard> hmmmmmmmmmmmmmmm
<reinhard> I want to see all customers that ordered a given product
<johannesV> right
<johannesV> but you want to see them only once, no matter if they've ordered it ten times
<reinhard> now I think I see the need somehow
<reinhard> although that even might be a matter of reports
<johannesV> well, not if you want to have that in forms
<johannesV> (like me :)
<reinhard> brb again phone
<ajmitch_> hi
<reinhard> hi ajmitch_
<reinhard> johannesV: I still think that the concept of "distinct" can't be taken directly from SQL for appserver
<reinhard> my understanding is that distinct in SQL operates on field values
<reinhard> I'd think we would want to operate on classes
<reinhard> really need to think about it a bit
<reinhard> phone again
<johannesV> so maybe this is where *lists* come in ...
<aries> reinhard: http://www.wxpython.org/migrationguide.php =)
<reinhard> johannesV: yeah, might well be
*** R45^2 has joined #gnuenterprise
*** R45 has quit IRC
*** sjc has quit IRC
*** dimas has quit IRC
*** lekma has joined #gnuenterprise
<lekma> hi all
*** jamest has quit IRC
*** sjc has joined #gnuenterprise
<johannesV> lekma, was it you looking for case-insensitive queries ?
<lekma> yep
<johannesV> have you noticed the new condition elements <upper> and <lower> ?
<lekma> nope
<lekma> in Session.find() ??
<johannesV> in common.GConditions ... and theirfore in session.find () too
* johannesV looking for a sample ...
*** SachaS has joined #gnuenterprise
<johannesV> ok
<johannesV> if you do it in the static condition (xml) it could look like this :
<johannesV> <condition>
<johannesV>   <upper>
<johannesV>     <cfield name="address_name" />
<johannesV>   </upper>
<johannesV>   <cconst name="FOO" />
<johannesV> ...
<johannesV> and, in a find you'd write
<lekma> ok
<johannesV> session.find ('class', ['eq', ['upper', ['cfield', 'address_name']], ['const', ...]]] ...)
<johannesV> of course you can use 'lower' too (if you prefer that one)
<johannesV> using this upper/lower will be transformed into an SQL-function upper/lower
<johannesV> (if supported by the backend)
<reinhard> johannesV: out of curiosity, did you test that with non-ascii characters (umlauts, accents etc)
<reinhard> ?
<johannesV> hmm, i'm not sure ...
<johannesV> if appserver comes to evaluate such a condition it's using string.upper ()
<johannesV> (or lower)
<johannesV> i'd have to check, what the backend is doing if it get's an UPPER function (which is at least sql-99 i think)
<reinhard> >>> import string
<reinhard> >>> x = "müller"
<reinhard> >>> print x.upper()
<reinhard> MüLLER
<reinhard> :(
<johannesV> this is lowercase, isn't it ?
<reinhard> yes
<johannesV> hm, how can this be converted without errors then ?
<reinhard> >>> import locale
<reinhard> >>> locale.setlocale (locale.LC_ALL, '')
<reinhard> 'de_AT@euro'
<reinhard> >>> print string.upper(x)
<reinhard> MÜLLER
<reinhard> so it actually should work :))
<johannesV> well, if x is unicode it's ok
<reinhard> ah yeah
<reinhard> great
<reinhard> we use unicode throughout
<reinhard> so it should be ok at least for the places where appserver does it
<johannesV> if such a value is given by an xml file it depends on it's encoding and the way these characters are specified
<reinhard> you sure?
<reinhard> i thought parsing an xml file will always return a unicode string
<johannesV> does it ?
<reinhard> not 100% sure
<reinhard> but i *think* it does
<johannesV> might be, yes ...
<johannesV> anyway, lekma should be able to do case-insensitive queries now ...
*** jamest has joined #gnuenterprise
<reinhard> johannesV: asking on purpose because I think lekma's users will be french speaking IIRC
<reinhard> (is that correct, lekma?)
<reinhard> and those people usually deal with tons of accents
<johannesV> i can do some tests later in the afternoon ...
<reinhard> johannesV: np
<reinhard> I think it will depend on the DB anyway
<reinhard> and we won't be able to do much about it if the DB breaks UPPER and LOWER for non-ascii characters
*** johannesV_ has joined #gnuenterprise
<reinhard> reinhard: johannesV: np
<reinhard> reinhard: I think it will depend on the DB anyway
<reinhard> reinhard: and we won't be able to do much about it if the DB breaks UPPER and LOWER for non-ascii characters
<johannesV_> that's right ...
*** SachaAway has quit IRC
*** johannesV has quit IRC
*** lxf has quit IRC
*** jcater has joined #gnuenterprise
*** aries has quit IRC
*** dcmwai has joined #gnuenterprise
*** johannesV_ has quit IRC
*** dcmwai has left #gnuenterprise
*** SachaS has joined #gnuenterprise
<lekma> reinhard, johannesV: unicode insensitive search works great!!!! (at least from the few test i ran)
<lekma> reinhard, johannesV: thx
<lekma> :)
<reinhard> lekma: is it correct that you speak french?
<lekma> yep
<lekma> and users here also
<reinhard> just asking as the fr.po are somewhat outdated
<lekma> :)
<reinhard> but I guess you don't have the time to update them...
<lekma> i had one a few month ago, that i never sent
<lekma> but from that time
<lekma> i think i should do it again
<reinhard> now *if* you want to do it
<reinhard> it would be a good chance to fetch the prereleases I'm just preparing now
<reinhard> :)
<lekma> at the end of january, i'll have a bit more time, and if you remind me, i'll send a dutch and french .po for appserver
<reinhard> cool
<reinhard> you can bet I'll remind you ;-)
<lekma> :)
*** chillywilly has quit IRC
*** cilkay has quit IRC
*** chillywilly has joined #gnuenterprise
*** cilkay has joined #gnuenterprise
*** chillywilly has quit IRC
*** chillywilly has joined #gnuenterprise
*** cilkay has quit IRC
*** cilkay has joined #gnuenterprise
<reinhard> jcater: I can't find the gnue-reports-filter script
*** johannesV has joined #gnuenterprise
<reinhard> jcater: if you have time
<reinhard> jcater: at first sight it seems to me like you had forgotten to commit gnue-reports/scripts/gnue-reports-filter
<reinhard> also, ajmitch asked if we could provide a really good default sample.reports-filters.conf
<reinhard> will wait for your feedback before doing prerelease of reports
<reinhard> thank
<reinhard> bbl
<reinhard> thanks
<reinhard> bbl
*** reinhard has left #gnuenterprise
<lekma> bye
*** lekma has quit IRC
*** johannesV has quit IRC
*** wendall911 has joined #gnuenterprise
*** psu has joined #gnuenterprise
*** SachaS has quit IRC
*** wendall911 has quit IRC
*** RumBa138 has joined #gnuenterprise
<psu> hi RumBa138
*** holycow has joined #gnuenterprise
*** RumBa138 has quit IRC
<chillywilly> anyone know of any open source/free knowledge base software?
*** reinhard has joined #gnuenterprise
<psu> chillywilly: as in wiki-like?
<reinhard> jcater: thanks for committing these 2 files
<psu> or more formal?
<reinhard> is gnue-labelmaker a script that should be installed, too?
<chillywilly> psu: no as in a searchable one that is integrated with an FAQ
<holycow> chillywilly, knowledge based software?
<psu> not aware of anything
<chillywilly> freshmeat is being helpful
* holycow pets reiser4
<holycow> reiser3 even
<holycow> i consider my file system the only true knowledge based software
<psu> I *suppose* you could sort-of use DCL - set each knowledge base issue as a ticket & allow anon access
<psu> but that's "round hole-square pegging"
<holycow> everything else is just tools to manipulate the 'knowledge'
<holycow> or some hack to story knowledge in a hiearchical way, i often find this limiting at the end of the day, and getting info out of such a system is usually harder than getting it in
<ajmitch_> morning lads
<holycow> senor
<psu> hi ajmitch_
<reinhard> hi ajmitch_
<ajmitch_> reinhard: any last minute deb requests before I go offline for a couple of weeks? ;)
<reinhard> you do seriously?
<reinhard> or you kidding?
* reinhard just sent out prereleases today
<ajmitch_> hmm
<ajmitch_> I'm serious about being offline
<reinhard> starting with?
<reinhard> and for how long?
<ajmitch_> starting in a few hours
<ajmitch_> until the 11th or so
<reinhard> Jan 11th?
<ajmitch_> yes
<reinhard> well, then don't worry
<ajmitch_> not feb 11th :)
<reinhard> we've waited much longer for debs than that in the past ;-)
<ajmitch_> :P
<reinhard> well, unless you envision debian to release sarge at the 10th
<reinhard> ;-)
<ajmitch_> I'm sure they will release 'soon' ;)
<ajmitch_> current debs are what will be in ubuntu's hoary release
<ajmitch_> since upstream version freeze is this week
<ajmitch_> since ubuntu seems to be the only one releasing at the moment... ;)
<holycow> the gnome drop into sarge was super smooth this time around
<holycow> i wonder if the ubuntu team had anything to do with that
<ajmitch_> probably
<reinhard> holycow: IMHO you can bet on that
<ajmitch_> I know jordi did a fair bit of work there
<ajmitch_> bbl...
<reinhard> well
<reinhard> I don't care *that* much about what's in ubuntu
<reinhard> because that won't be the state of the art for another 3 years :-)
<holycow> me neither, i do hope they find enough customers to keep up the momentum tho
<reinhard> holycow: didn't want to say I don't care what's in ubuntu
<reinhard> just we already know we'll get another chance to get something in in 6 months
<holycow> *nod*
<holycow> :) i hear ya
<reinhard> where in debian we'll get another chance to get something into stable in 3 years or so....
<holycow> one would hope this will change however
<holycow> i'm of the position that all debian needs to do is what gnome did, go to time based releases
<holycow> all it would mean is a stable freeze more often, devs wouldnt need to work any harder
<reinhard> holycow: I support > 10 debian servers for customers
<holycow> your kidding?
<holycow> really?
<reinhard> and I can tell you I am *happy* that there isn't a new stable version every 6 months
<holycow> wow
<holycow> you are a god, i think thats great!
<holycow> :)
<reinhard> so I do security updates now and then and other than that the servers run with uptimes
<reinhard> holycow: you're kidding I think :-)
<holycow> lol no!  i  want to hear more about debian servers out there :)
<reinhard> with uptimes > 200 days
<holycow> infact i want to get into that game my self, theres just no point in running anything else
<holycow> really?
<holycow> nice!
<holycow> yeah having just security updates, makes it super easy
<holycow> although you still haveto pay attention to the updates
<reinhard> well
<reinhard> these servers do a very limited set of things
<reinhard> and even more important
<holycow> as all servers should
<reinhard> they are all more or less identical
<reinhard> (which, btw, is a piece of cake to achieve with debian)
<holycow> heh :) *nod*
<holycow> when i found debian, i couldn't believe it actually
<reinhard> that's what I really love about debian, you can script everything in installation and configuration
<holycow> i haven't got that far into debian config ...
<psu> debian rox our sox
<holycow> the thing i like about debian is it all makes sense
<holycow> i don't understand one fucking thing about suse  or redhat or why they do what they do
<reinhard> :)
<holycow> reinhard, you must be happy about d-i then
<holycow> that seems to have some real potential
<reinhard> meanwhile I'm down to between 25 and 35 minutes for a standard server install
<holycow> i think it now has the capacity for preseeded installs
<psu> holycow: does this just mean your mind works the same way as a debian maintainer? There's a scary thought ;-)
<holycow> altho frankly i just copy hds, and then copy over the bootsector, much faster
<holycow> psu, haha :)
<reinhard> for which i cash EUR 700,- lump sum
<holycow> reinhard, you can get that down just by copying from a haster hd
<holycow> i have removable hd's
<holycow> all debian installs work on all machines
<reinhard> holycow: I tried that once
<holycow> the only thing that doesnt autoconfig is xserver
<psu> reinhard : every business needs its cashcow & that sounds like yours ;-)
<reinhard> psu: the funny thing is that
<psu> holycow: no offence intended
<holycow> non taken :)
<reinhard> a) i'm still cheaper than others
<reinhard> and
<psu> wouldn't diss a cow with you about ;-)
<holycow> hahaha
<reinhard> b) others need about a day for setting up a server with same function
<reinhard> however
<reinhard> c) those others use suse :-)
<holycow> i'm putting together a whole bunch of turnkey function based server products
<holycow> based on debian
<holycow> mail server
<holycow> antispam server
<holycow> firewall
<holycow> gnue app server hopefully :)
<holycow> aolserver is looking good too
<reinhard> holycow: yes debian is excellent for that
<holycow> accounting server
<holycow> etc.
<psu> reinhard : would you see that 25-35 minutes going up once GNUe Appserver becomes part of "the standard server set up"? ;-)
<holycow> i'd like for people to call me up say 'i need this functionality' and i know i can say 'no problem give me 3 hours'
<reinhard> psu: probably depends on the db backend in use
* psu is just trying to show off that he still remembers diddly squat about this project...
<reinhard> I think installing postgres would add 5 minutes, but not much more
<reinhard> psu: :-)
<holycow> postgres is simple
<holycow> terrific piece of software
<psu> The big boys (Oracle etc) might be a bit harder
<reinhard> yeah
<reinhard> johannes played with maxdb
<psu> And Microsoft SQL Server on debian would be approaching infinity
<reinhard> it took him about 4 hours to create a database :-)
<holycow> heh
<psu> unless by some miracle it works under WINE
<reinhard> psu: re cash cow
<reinhard> it's crazy but you can earn most money when you make use of the stupidity of your competitors
<psu> reinhard: yes
<reinhard> I *really* saw people installing completely standard server functions with SUSE
<reinhard> not even web server or such a thing
<reinhard> only ssh, telnet, outgoing mail (sendmail)
<reinhard> and a cron job for a tape backup
<reinhard> and samba
<psu> SUSE isn;'t really a "server" GNU/Linux distro anyway?
<reinhard> and it took them > 8 hours
*** jamest has left #gnuenterprise
<psu> gak
<holycow> reinhard, i ran into that actually
<reinhard> and they still didn't figure out how to make non-ascii filenames work with samba
<holycow> the problem with suse is they do all their own shit
<holycow> for no good reason
<holycow> and theres no real documentation as to what is what and where the bleep it is
<holycow> i used it for a while but it drove me insane
<reinhard> holycow: the main problem with suse IMHO is that they provoke semi-knowledge
<psu> If you need a GNU/Linux box to run as a server, use RedHat if
<reinhard> (not sure if that is a word)
<holycow> reinhard, actually you got it, exactly
<psu> you have a problem with debian
<holycow> redhat?
<holycow> no thank you
<reinhard> they encapsulate everyhting with their silly yast
<reinhard> and you end up with clicky clicky
<reinhard> and not knowing what you do actually
<holycow> right exactly
<reinhard> and if it doesn't work you're as lost as you are with windows
<psu> well at least with Redhat you can get someone who's "certified" ;-)
<reinhard> psu: lol
<reinhard> still debian has the best config tool of all gnu/linux distros
<psu> certified as what we won't say
<reinhard> vi :)
<holycow> i personally am liking userlinux
<holycow> supposedly basically debian
<holycow> with support for a subset of packages
<psu> anyway, time for bed
<psu> back to work tomorrow
<holycow> separating the corporate from debian is good, hopefully there is enough momentum to add things like certification
<holycow> later psu
<reinhard> l8r psu
<reinhard> good night
<psu> respext
<reinhard> holycow: that's very much like ubuntu
<holycow> not at all
*** psu has left #gnuenterprise
<holycow> ubuntu is a fork even though they claim not to be
<reinhard> ah
<holycow> userlinux is uses all official debian packages, nothing more nothing less
<reinhard> so user linux will not do changes in packages?
<holycow> the only caveat is patches
<reinhard> ah
<reinhard> gotcha
<holycow> they will fork a package if a patch is required for security and not immediately accepted into debian
<holycow> yeah thats the goal, not change packages
<holycow> *hmmm*
<holycow> mozilla thunderbird saves its email in mbox format
<holycow> thats nice, easily imported anywhere
<holycow> lets see what evolution saves its email as :)
<holycow> probably some xml format
<holycow> hehe :)
<holycow> mbox as well
<holycow> cool
<holycow> reinhard, i can give you a million reasons why i hate microsoft but ...
<holycow> just one is having to recover corrupted outlook files
<holycow> i had some poor fuck call me, his main hd was toast and his backup usb hd failed at the same time
<holycow> i managed to recover a corrupted outlook file, but it lost like 5% of its content
<holycow> running all sorts of recovery tools was almost pure failure
<holycow> luckily ms hid a recovery tool in win2k about 15 folders deep, but that worked
<holycow> anyway it still trunkated and recovered only shit
<holycow> file was about 1 gig so it was horrendeous but i don't understand why those people do shit like that
<holycow> for being so rich and having smart people, they sure make life hard
<reinhard> night all
*** reinhard has left #gnuenterprise
*** C6 has joined #gnuenterprise
*** C6 has left #gnuenterprise
*** R45^2 has joined #gnuenterprise
*** R45 has quit IRC
*** ajmitch_ has quit IRC
*** sjc has quit IRC
*** dcmwai_ has joined #gnuenterprise