*** reinhard has joined #gnuenterprise
<apauley> Hi reinhard
<reinhard> good morning
<apauley> I read that copyrights have to be transferred to the FSF. Do you differentiate between core GNUe software and 3rd party software?
<reinhard> yes of course
<reinhard> you can write your applications with gnue and distribute them on your own as you like
<reinhard> you can even contribute them to our "gnue-contrib" repository where we collect "inofficial" gnue apps
<apauley> OK, nice
<reinhard> and if you aim at developing something that will become "core gnue" we would request you to transfer copyright
<apauley> And does these 3rd party apps need to be GPL compatible?
*** SachaS_ has quit IRC
<reinhard> apauley: depends: if they use parts from other GPL apps (like the base modules kilo and btami are working on), then yes
*** dimas has quit IRC
<reinhard> if you write your app from scratch and just use gnue-appserver, gnue-forms etc as infrastructure, then you can regard it like
<reinhard> you write a C program and use gcc to compile it
<reinhard> your app can be any license you like
<reinhard> however of *course* we recommend using the GPL for everything :-)
*** johannesV has joined #gnuenterprise
<johannesV> good morning
*** kilo has joined #gnuenterprise
<reinhard> hello johannesV
<reinhard> and kilo
<kilo> good morning
<kilo> hi reinhard
<johannesV> reinhard, i'm still reworking readgcd ... but it looks like i can commit changes today ... :)
<reinhard> excellent
<reinhard> it's *really* about time for a 0.3.90 release
<johannesV> as i went further, doing quite awful things to my gcd files the old version showed up a lot of bugs
<johannesV> so i started a comple rewrite
<reinhard> bugs == unexpectable behaviour when gcd is corrupt?
<johannesV> yeah
<reinhard> so it's "missing fault tolerance" rather than "bug" ;-)
<johannesV> i still dislike the way a general markup-error get's reported
<johannesV> yeah, could say so
*** sjc has joined #gnuenterprise
<johannesV> well, not always
<johannesV> for example if one says <module name="foo_bar_baz">
<johannesV> it just went in
<johannesV> so we had corrupt gnue_* table entries
<johannesV> as well as db-tables 'foo_bar_baz_something'
<johannesV> or you say <class name="foo" module="invalidOrNotExisting"> ...
<johannesV> things like this, there are quite a lot of them ...
*** btami has joined #gnuenterprise
<kilo> johannesV: have you seen btami's last commit, in gnue-invoice?
<kilo> could you try something: drop the invoice db on postgres, and remove those lines added in the last commit from invoice-hu.gsd, then try setup-db-hu.gsd. does it work for you taht way?
<johannesV> hm
<johannesV> kilo, can you describe the problem ?
<kilo> the problem is, that the row added in btami's last commit is referenced. and it was missing for some weeks now from the gsd. hjowever, i have set up the db several times, and no error messages appeared.
<kilo> yesterdat btami wanted to set up the db on windows, and got the error message from the firebird db engine. and we find that strange
<johannesV> kilo, as soon as i will commit my changes of appserver such scripts get obsolete ...
<johannesV> :)
<kilo> ah
<kilo> but the gsd will not or will it?
<johannesV> no, not the gsd
<johannesV> but gcd/glds get imported automatically
<johannesV> one can define a "modulepath" containing a list of directories, and appserver searches this dirs recursivley for gcd's and gld's and import them as needed
<johannesV> and if appserver get's a sig-hup the same procedure runs again
<johannesV> modulepath can be specified on the commandline as well as in gnue.conf
<johannesV> so usually there's no need for a setup-script which integrates gcd's or gld's anymore (as appserver can handle this himself)
<kilo> okay, okay, but this problem i mentioned above is a problem regarding the gsd itself
<johannesV> ok, i'll drop my test-db and recreate it using setup-db.sh, right ?
<johannesV> kilo, which gsd shall i import now ?
<kilo> setup-db-hu.sh
<kilo> <kilo> could you try something: drop the invoice db on postgres, and remove those lines added in the last commit from invoice-hu.gsd, then try setup-db-hu.gsd. does it work for you taht way?
<johannesV> INSERT INTO ITEM_Item (sales_price, gnue_createdate, gnue_id, item_type, item_longname, item_unit, item_shortname, sales_vat) VALUES ('111.0', '2004-10-14 13:28:02', '04158385481881526447158761849009', '17527929513744924781765541414167', 'Szolgáltatás 1', '87347199152063042221294657166838', 'S1', '61602943621389663558226281311093')
<johannesV> (if i remove the lines btami added yesterday)
<johannesV> ERROR:  insert or update on table "item_item" violates foreign key constraint "fk_item_item_item_unit"
<johannesV> DETAIL:  Key (item_unit)=(87347199152063042221294657166838) is not present in table "item_unittype".
<johannesV> which is correct ...
<kilo> ok, that is correct
<johannesV> if i do revert the change, everything's fine ...
<kilo> but it does not give an error message when using firewirb...
<kilo> firebird
<johannesV> it does not ?
<johannesV> let me try here ...
<johannesV> wow
<kilo> does that mean you also got no error messages?
<johannesV> no, i don't get an error using firebird
* johannesV still checking
<kilo> but on the windows version btami got error messages
*** dcmwai has quit IRC
<kilo> btami says maybe his windows' firebird version is 1.5.1 or 1.5.2
<johannesV> ah, got it
<johannesV> the constraint is missing
<johannesV> well, what an obvious error this is ....
* johannesV claps his face ...
<johannesV> kilo, let me finish the new readgcd, and i can tell ya if the missing constraint get's added on firebird
<kilo> ???
<kilo> hmmm, ok. 5 minutes enough? 8-)))
<johannesV> if you do "show table item_item" in isql you won't find the constraint to item_unit table, do you ?
<johannesV> kilo, no will take until afternoon
<johannesV> but i'll take this bug into my notes, so i can have a look at the reason later ...
<kilo> hmm, i cant wait until afternoon. we have received order for gnue-invoice from 1425 companies 8-)))
<reinhard> kilo: rofl
<reinhard> kilo: I see you have to start burning the CD really now :)
<reinhard> the CDs
<reinhard> :q
<reinhard> hmmm... this is not vi??
<kilo> yeah, i've rented all the cd-burning capacity in town
<kilo> johannesV: no, no constraint to item_unit
<johannesV> kilo: i get the following error (on firebird) now:gnue-schema:
<johannesV> ERROR POSTING RECORD # 1
<johannesV> (-530, 'execute.isc_dsql_execute: violation of FOREIGN KEY constraint "FK_ITEM_ITEM_ITEM_UNIT" on table "ITEM_ITEM". ')
<btami> yep
<johannesV> and if i look at the table item_item i do have 4 constraints now
<kilo> 4?
<btami> what firebird version you use?
<johannesV> CONSTRAINT FK_ITEM_ITEM_ITEM_TYPE:
<johannesV>   Foreign key (ITEM_TYPE)    References ITEM_ITEMTYPE (GNUE_ID)
<johannesV> CONSTRAINT FK_ITEM_ITEM_ITEM_UNIT:
<johannesV>   Foreign key (ITEM_UNIT)    References ITEM_UNITTYPE (GNUE_ID)
<johannesV> CONSTRAINT FK_ITEM_ITEM_SALES_VAT:
<johannesV>   Foreign key (SALES_VAT)    References TAX_VAT (GNUE_ID)
<johannesV> CONSTRAINT PK_ITEM_ITEM:
<johannesV>   Primary key (GNUE_ID)
<johannesV> -
<johannesV> SQL> show version ;
<johannesV> ISQL Version: LI-V1.5.1.4500 Firebird 1.5
<johannesV> Firebird/linux Intel (access method), version "LI-V1.5.1.4500 Firebird 1.5"
<johannesV> Firebird/linux Intel (remote server), version "LI-V1.5.1.4500 Firebird 1.5/tcp (Boston)/P10"
<johannesV> Firebird/linux Intel (remote interface), version "LI-V1.5.1.4500 Firebird 1.5/tcp (Boston)/P10"
<johannesV> on disk structure version 10.1
<johannesV> the main question remains: why was it created correctly for postgres, but wrong for firebird ...
<kilo> ah, ok, i wasnt counting in the PK
<kilo> my computer says
<kilo> ISQL Version: LI-V1.5.0.4231 Firebird 1.5 Release Candidate 8
<kilo> maybe it is a firebird bug, that is corrected in the 1.5.1 version
<johannesV> no, so it should still be missing the constraint
<johannesV> but if i use the 'new readgcd' the constraints are ok
<kilo> ah, i see
<johannesV> ah
<kilo> well then it is not only optimization...
<johannesV> get the same error on postgres now ...
<johannesV> so the bug is really in readgcd
<johannesV> (which is fixed right now)
*** psf has joined #gnuenterprise
*** sjc has quit IRC
<kilo> johannesV: great, thx
<ajmitch> hi
<btami> hi
<kilo> hi
<reinhard> hmmm
* reinhard just noticed that we have a *lot* of documentation to update before we can release 0.4.0 final
<kilo> and what about the long-forgotten forms release?
<reinhard> gotta check if we must release common at the same time
<reinhard> which would mean we have to wait
<reinhard> as common must be released with appserver
<btami> bbl
*** btami has quit IRC
<reinhard> hmmm... no
<reinhard> seems like it's upward compatible
<reinhard> so we could release common, then forms, then appserver
<reinhard> bbl
*** psf has quit IRC
*** tiredbones has quit IRC
*** kilo has quit IRC
*** tiredbones has joined #gnuenterprise
*** btami has joined #gnuenterprise
*** btami_ has joined #gnuenterprise
*** btami has quit IRC
*** titopbs has joined #gnuenterprise
*** jamest has joined #gnuenterprise
*** dimas has joined #gnuenterprise
<reinhard> jamest: do you have any pending changes in forms or common?
*** johannesV_ has joined #gnuenterprise
<mnemoc> reinhard: when do you plan to release?
<reinhard> common and forms: today a prerelease
<mnemoc> :D
<reinhard> final release by end of week or so
<reinhard> appserver: today or tomorrow a -beta1
<reinhard> further steps depending on the feedback
<mnemoc> feature frozen -beta1?
<reinhard> yes
<mnemoc> great
<mnemoc> i'll postpone my installations then
<reinhard> johannesV has announced to commit the last missing feature for 0.4 appserver today
<johannesV_> yeah, i'm nearly finished with exceptions ...
<reinhard> johannesV_: you also have uncommitted changes in common, right?
<johannesV_> no, i don't
<johannesV_> should i ?
<reinhard> GParser?
<johannesV_> i haven't changed GParser right now
<johannesV_> will postpone these changes after committing the appserver
<johannesV_> as it's a bit cosmetic though ... :)
<reinhard> ah
<reinhard> well then I'll wait for those changes with the prerelease of common
<reinhard> (I suspect we will also have string changes with that GParser excpetion stuff)
<johannesV_> ok, i think i can do that GParser thingy right this afternoon ...
<reinhard> no hurry
<reinhard> tomorrow is early enough, too :)
*** johannesV has quit IRC
<reinhard> please go on and finish appserver first
<jamest> reinhard: no pending changes
<reinhard> jamest: thanks
<reinhard> jamest: well, just to ask for your opinion, too
<reinhard> currently, when GParser sees an error in the XML markup
<reinhard> you get an exception with a traceback through GNUe internal code
<reinhard> but more or less no info about the file and the line number where the parse error occured
<jamest> i thought the traceback included line #
<archivist> any updates to INSTALL and setup.py since .3.5 as I just tried them for the first time last night
<jamest> wait, i'm thinking of triggers
<reinhard> well now a few months ago johannesV_ has improved the exception handling to be able to display *anything* you want instead of the traceback
<reinhard> for example for exceptions that occured at server side, we now display the server side traceback on the client instead of the client side traceback
<reinhard> and something similar for the tarcebacks that hapen in trigger code
*** jcater has joined #gnuenterprise
<reinhard> and we had the idea to display better info about the xml parse error instead of a traceback that the user can't read anything from anyway
<johannesV_> reinhard, appserver is committed ...
<reinhard> somehow like a traceback through the XML file instead of the python traceback
<reinhard> jamest: what would you think about that?
<reinhard> the basic idea is to not display a traceback all through gnue-common and gnue-forms/appserver for a xml parse error
<reinhard> as the traceback might confuse the person that wrote the xml file
<reinhard> johannesV_: great, thanks
<johannesV_> reinhard, but i have to do the datetime.* fix for geasGsdgen.py
<reinhard> johannesV_: ok
<jamest> reinhard: i think that's a good idea
<reinhard> jamest: thanks for your verbatim feedback
<reinhard> ;-)
<reinhard> j/k
<jamest> lol, sorry
<jamest> i miss being a state worker
*** wendall911 has joined #gnuenterprise
<jamest> as they actually expect me to do something here :)
<jcater> reinhard: I've always hoped we could do what you're describing
<jcater> but never knew where to begin
<jcater> (or had time myself, for that matter)
<jcater> just my $0.02
<reinhard> jcater: excellen, thank you too for your feedback
<jcater> I hate getting an XML traceback
<jcater> as it's quite useless
<reinhard> yes
<reinhard> that was very much the starting point of this whole plan :)
<reinhard> btw I've finished another (last) step for automatic release process
*** titopbs has quit IRC
<reinhard> whenever a release is uploaded to ash, the next web page rebuild automatically deletes all obsoleted prereleases
<jcater> cool
<johannesV_> hmm, since GParser.loadXMLObject () does not have the meaning of a file, but only get's a stream to process it's quite difficult to do something like "Hey, you have an xml markup error in your file foo at line 100 ..."
<reinhard> new optional parameter with file name?
*** titopbs has joined #gnuenterprise
<jcater> reinhard/ johannesV_ : I think it's easier than that
<jcater> file objects have a .name attribute
<johannesV_> hmm, jcater that's a good point ...
<johannesV_> but i'd need the filename in the xmlSchemaHandler instance too
<johannesV_> so i could do something like: dh = handler (); dh._filename = stream.url
<johannesV_> or we could change the api of xmlSchemaHandler instance by adding an optional constructor-parameter 'filename'
<johannesV_> so it would look like: dh = handler (filename = stream.url)
<johannesV_> jcater, what would you prefer ?
<jcater> I'm ok with the latter
<johannesV_> (the latter would mean changing all xmlSchemaHandler classes)
<johannesV_> but it's not too much
<johannesV_> as i would see it as the 'cleaner solution'
<johannesV_> but i'm still not sure if we have a 'file-like object' all the time (like loadFromBuffer () in forms)
<johannesV_> yeha
<johannesV_> err
<johannesV_> yeah
<reinhard> I think if we have no file, it would make sense to show something like "stdin" or such
<johannesV_> AttributeError: 'cStringIO.StringI' object has no attribute 'url'
<reinhard> or [unknown]
<johannesV_> reinhard, good idea
<johannesV_> i'm ok with that
<jcater> you know, the python libraries we use to do our triggers
<jcater> have you pass in the stream
<jcater> and a "name"
<jcater> which lets us say "On-New-Commit" as the "name"
<jcater> maybe that's a better solution
<jcater> where normally it can be the file name
<jcater> or in special cases, it can be like a subform
* jcater is thinking out loud
<jcater> I'm ok with what reinhard said
<reinhard> jcater: actually I like your idea better
<reinhard> or a combination
<reinhard> priority 1 - explicit parameter
<reinhard> priority 2 - stream.url
<reinhard> priority 3 - "[unknown]"
<jcater> reinhard: I like that idea best
<jcater> it would be a little more work initially
<jcater> but we don't have to change the various tools' code
<jcater> to have them keep working
<reinhard> well, that's not our problem, jcater
<reinhard> ;-)
<jcater> lol
<jcater> btw, just so you know
<jcater> I'm thinking of moving ash to my rack
<jcater> jamest and I were discussing it last night
<jcater> I'm getting 5Mbps/1.5Mbps service here
<jcater> and have all the large UPS's, etc
<jcater> but in the process, ash will be upgraded to a better machine
<chillywilly> hi
<btami> hi chillywilly
<chillywilly> anyone know of any FOSS OLAP solutions?
<reinhard> no
<johannesV_> bbl
*** dcmwai has joined #gnuenterprise
*** dcmwai has quit IRC
*** SachaS has joined #gnuenterprise
<reinhard> off for today
<reinhard> bye all
*** reinhard has quit IRC
*** dcmwai has joined #gnuenterprise
*** dcmwai has quit IRC
*** cilkay has quit IRC
*** btami has quit IRC
*** wendall911 has quit IRC
*** sjc has joined #gnuenterprise
*** holycow has joined #gnuenterprise
*** titopbs has quit IRC
*** wendall911 has joined #gnuenterprise
<johannesV_> jcater, i've commited a first shot of an improved exception handling in GParser (and friends). maybe you can find a few minutes to double-check or try it out ...
<johannesV_> in order to keep compatibility i haven't changed the inteface of the GParser.xmlHandler class by adding an additional parameter
<johannesV_> so after creating the xmlHandler the url (which might get overriden by the loadXMLObject () function) and the parser instance both get passed into the handler
<johannesV_> the latter is usefull for retrieving line numbers from the parser
<johannesV_> additionally all objects created while parsing a file get the properties '_url' and '_lineNumber' set
<johannesV_> this allows a phased init function to refer to this information
<johannesV_> (like GDataSource does with it's sorting- and order_by-tags ...)
<johannesV_> well, just have a look at it and let me know if you'd like to have something changed ...
<johannesV_> off for today
<johannesV_> bye all
*** johannesV_ has quit IRC
*** tiredbones has quit IRC
*** titopbs has joined #gnuenterprise
*** tiredbones has joined #gnuenterprise
*** SachaS_ has joined #gnuenterprise
*** SachaS has quit IRC
<archivist> testing 2 things at once gnue and new subdomain can anyone see http://www.pastebin.archivist.info/pastebin.php?show=3
<archivist> show error i got last night installing gnue
<holycow> i cant see it
<archivist> only set dns 10 minutes ago
<holycow> yeah
<holycow> domain could not be found
<holycow> it probably didn't propagate to our dns servers yet
<holycow> got an ip based subdomain for that?
<holycow> ip based path i mean?
<archivist> try http://www.archivist.info/pastebin.php?show=3
<holycow> Object not found!
<archivist> hmm working internally ip is  217.35.103.47
<holycow> that works
<holycow> let me add the / part
<holycow> nope
<archivist> hmm http://www.pastebin.archivist.info works externally (remote to a friends box)
<holycow> not here
<archivist> dns might by querying the slave which may not be updated yet
<archivist> http://www.archivist.info/pastebin
<holycow> i can see the page
<holycow> pasting generates error
<archivist> i get errors as well  when pasting
<archivist> im just testing the dns via dnsreport, the slave has not been updated yet
<archivist> ah bug in pastbin setup as well
<archivist> any way setup.py gives me a ImportError: No module named gnue
<archivist> but no help as to the cause is it db not setup /permissions/config
*** tiredbones has quit IRC
*** jamest has quit IRC
*** tiredbones has joined #gnuenterprise
*** titopbs has quit IRC
*** cilkay has joined #gnuenterprise
<archivist> bug on the website
<archivist> http://www.gnuenterprise.org/cgi-bin/viewcvs.cgi/gnue/trunk/
<archivist> An Exception Has Occurred
<archivist> Python Traceback
<archivist> Traceback (most recent call last):
<archivist>   File "/usr/lib/python2.3/site-packages/viewcvs/viewcvs.py", line 3235, in main
<archivist>     request.run_viewcvs()
<archivist>   File "/usr/lib/python2.3/site-packages/viewcvs/viewcvs.py", line 268, in run_viewcvs
<archivist>     import vclib.svn
<archivist>   File "/usr/lib/python2.3/site-packages/viewcvs/vclib/svn/__init__.py", line 27, in ?
<archivist>     from svn import fs, repos, core, delta
<archivist> ImportError: No module named svn
<wendall911> gotta love python cgi
<wendall911> archivist: I don't know if anyone is here now that can fix that, may want to file a bug report
<wendall911> http://www.gnuenterprise.org/community/bugs.php
<archivist> im there at the moment just reading page i see yr not using bugzill like everybody else
<archivist> and i have to swap pcs to email
*** tiredbones has quit IRC
<wendall911> archivist: like everybody else?
<wendall911> archivist: there are alot of different systems, bugzilla is common, but not disproportionately
<archivist> i dont particularly like it except its a web interface which is useful when this pc hanst got direct email
<archivist> and i have to vnc to a box that does
<archivist> bug posted
<wendall911> archivist: understood
* archivist wonders how to file a bug when it bounces recipient.
<archivist> Remote host said: 550 <www-support@gnuenterprise.org>: Recipient address rejected: User unknown in local recipient table
<archivist> not friendly
<archivist> hence need for bugzilla and working software
*** mnemoc has left #gnuenterprise
*** sjc has quit IRC
*** wendall911 has quit IRC
*** holycow has quit IRC
*** cilkay has quit IRC
*** jcater_ has joined #gnuenterprise
*** lekma has joined #gnuenterprise
<lekma> hi all
*** jcater_ has quit IRC
*** jcater has quit IRC
*** dcmwai has joined #gnuenterprise
<chillywilly> anyone know how I can change the install prefix of a deb?
<ajmitch> it shouldn't be done
<ajmitch> because it has all the paths relative to /
<ajmitch> at best you can do dpkg -i --root=/target <file.deb>
<ajmitch> or --instdir=/target
<chillywilly> ok, well I think I will just rebuild my machine
<ajmitch> why?
<chillywilly> my / partition is too small
<chillywilly> was trying to isntall some OOo 2.0 beta debs which install into /opt
<ajmitch> ah..
<ajmitch> why would you want to install them?
<chillywilly> cause I like the new OOo
<ajmitch> yes, but why crap debs into /opt?
<chillywilly> what is available?
<ajmitch> you could just move /opt to another partition & bind mount
<ajmitch> ubuntu hoary has OOo 2 beta debs, iirc
<chillywilly> I don't have a free partition
<ajmitch> you have free space on any partitions?
<chillywilly> I can make a symlink
<chillywilly> yea I do
<ajmitch> symlink, bind mount, do whatever suits
<ajmitch> I'm moving my system to LVM at the moment because of these issues :)
<chillywilly> I don't blame you
<chillywilly>  /var on my server fills up all the time too
<chillywilly> are you talking about doing mount --bind?
<ajmitch> yeah
<ajmitch> Filesystem            Size  Used Avail Use% Mounted on
<ajmitch> /dev/mapper/vg1-var   9.9G  4.7G  5.2G  48% /var
<chillywilly> why doesn't mount --bind /opt /home/danielb/opt work?
<chillywilly>  /opt is on / though...so maybe that is why it doesn't work