*** ccfiel has quit IRC
*** vinsci has quit IRC
*** holycow has quit IRC
*** johannesV has joined #gnuenterprise
*** SachaAway has quit IRC
*** reinhard has joined #gnuenterprise
*** btami has joined #gnuenterprise
<btami> johannesV: thx for pythonizing my fix! :)
<johannesV> :)
<johannesV> well, can you check it on win ?
<johannesV> maybe we should add another level to the mapping
<btami> btw. getlanguage() need something similar
<johannesV> like map = {'plattform': {'enc': 'replace'}, 'plattform': {} ...}
<johannesV> so we could define a mapping per plattform, but it also means we would do the 'same' mapps more than once if they need changes on multiple plattforms
<btami> maybe, but it's a bit "implicit" prigrammig
<btami> i prefer fast but readable code
<johannesV> well, we'll see if it's needed anyway
<johannesV> what about that getlanguage () ?
<btami> yes
<btami> it gives some long strings like Hungarian_Hungary or wat...
<btami> instead of hu_HU
<btami> but it's a bit harder to map all localised windowse languages
<johannesV> hmm
<btami> no, not just localised
<btami> LANG=en gives a long string too
<johannesV> btami, can you check the contents of the dict locale.locale_alias ?
<btami> yes
* btami reboots to XP
*** btami has quit IRC
<johannesV> don't know, but maybe we can use a mapping built into the locale module
*** SachaS has joined #gnuenterprise
<SachaS> morning
*** kilo has joined #gnuenterprise
*** btami has joined #gnuenterprise
<btami> johannesV: may i send you the locale.locale_alias in e-mail ?
<btami> it's a bit long
<johannesV> np, just send it.
<johannesV> locales on win* seems a bit wired
<johannesV> anyway, i've to leave for a meeting now; bbl
*** btami has quit IRC
*** btami has joined #gnuenterprise
*** wt has joined #gnuenterprise
<kilo> good morning
<dimas> hello all
<wt> hi
<wt> I can't get to sleep
<wt> so I will talk with gnue peeps
<wt> anyone here ever use southware?
<kilo> hi dimas
<btami> hi dimas
<btami> hi wt, not me
<wt> btami, you a current user of gnue?
<wt> like in production?
<wt> not just as a toy
<btami> not yet
<wt> and who is the web person?
<wt> I would love to use gnue as a pos
<wt> and inventory tracker
<btami> jcater has a pos in gnue-contrib
<btami> and he is using it in hes book store
<dimas> hi kilo,btami
<btami> you mean "web person" like site admin ?
<wt> yes
<wt> like person that fixes the f'ed links
<wt> like the documentation link
<btami> not official, but jcater too
<wt> is contrib in the svn?
<btami> yes
<wt> so...RMS doesn't like the fact that I called my Lug a Linux User Group and not a GNU/Linux User Group
<kilo>  svn co svn://svn.gnuenterprise.org/var/svn/gnue-contrib where-you-want-to-place-it
<btami> hmm, there is a gnue-pointofsale in the main gnue repo and
<btami> a web-po in contrib by jamest
<wt> btami, it is not compatible with anything I see
<wt> nothing modern
<wt> does the oldpackages stuff even work?
<wt> What is hotline?
<wt> also, will DCL be fixed?
<btami> hotline is reinhard's stuff used in hes office for something i don't know :)
<wt> the instance on the gnue site?
<btami> don't know
<btami> the old pg database is at derek, please ask him
<wt> btami, I don't understand what you are telling me
<kilo> hotline is used to register customer incoming things
<wt> do you mean it is broken because there is no database under it?
<btami> yes
<btami> as i know, he dumped the database for upgrade DCL/pg
<wt> and never reloaded it?
<btami> yes
<btami> as i know :(
<wt> that's nice
<kilo> he would like to tidy it up
<wt> tidy up the database?
<kilo> yes
<wt> you can't do that if you never load it
<btami> but the us folks are lost in space for a while...
<wt> I would love to help
<wt> but RMS prolly would let me in since I have a Lug instead of a Glug
<kilo> what does "prolly" mean?
<wt> slang for probably
<kilo> wasnt in webster even...
<wt> would not let me in, I mean
<wt> chat slang
<wt> not real slang
<wt> it's a lower form of slang
<wt> hehe
<btami> i don't think RMS will let or not let in you into gnue :)
<kilo> ok, it is monday morning, cant even speak hungarian, not even english
<wt> btami, I mean the copyright assignment stuff
<btami> me too
<kilo> i think you should fill in the form and we will see
<wt> enough about RMS
<wt> I am sure they will accept any form
<btami> agreed
<wt> I feel that he is an ass, and that is something I just have to deal with
<btami> don't think so, just fill the form and forget RMS :)
<wt> sorry, I just sent him and inquiry and he sent back a not to nice reply because my group is called GOLUM (Group of Linux Users Memphis)
<wt> too late for correct grammar
<btami> ok, i just hope it's independent from being a gnue folk
<wt> what do I have to do to get the paperwork for copyright assignment?
<wt> does it apply to all code submitted to gnu?
<wt> or is it only good on a per project basis?
<btami> http://www.gnuenterprise.org/community/copyleft.php
<btami> iirc it was a per project
<dimas> wt: why you rejecting to name it GOOGLUM ? :)
*** Amorphous has joined #gnuenterprise
<wt> yes, I am rejecting that hideous beast
<wt> I am the president of the group and we are already 501(c)3
<dimas> i'm j/k
<wt> I know
<wt> I was just sayin
<wt> :-P
<btami> wt: http://www.gnuenterprise.org/developers/assignment.php
<wt> cool
<wt> thanks
<wt> btami, is anyone working on the erp packages?
<kilo> 'When trying to ignore a knock at your door,
<kilo> don't yell out, "No!" when someone asks,
<kilo> "Is anybody at home ?'
<wt> kilo, why not?
<wt> :-P
<btami> Winnie the pooh
<wt> btami, that's one talented bear
<kilo> yes, he is
<btami> kilo and reinhard started the work on packages
<kilo> and they do debate...
<wt> kilo, you debate?
<kilo> no, I do this: http://users.bart.nl/~cjmdbr/images/think_ani.gif
<wt> kilo, what is the methodology of the packages?
<btami> the main problem i think is, there is no spec about what the erp will contain
<kilo> hmmm, methodology, hmmmm..... I do something, then others disagree 8-))))
<wt> kilo, any chance of getting a release of gnue-common here shortly?
<wt> kilo, good call
<btami> but we want to build the database schema
<wt> any paid devs?
<kilo> gnue-common was just released last week
<wt> kilo
<wt> oh
<wt> not updated in debian
<wt> my bad
<kilo> debian? ask ajmitch
<wt> OMG
<wt> kword does an unholy job of importing my pdf resume
<wt> it's nearly perfect
<btami> deb http://ajmitch.dhis.org/debuild/gnue/ ./
<wt> btami, you are a vast wasteland of useful knowledge
<btami> maybe it's not updated as deb source, but the .deb files are there
<wt> who runs that?
<kilo> we dont copy, repeat last transmission
<wt> who runs the repo?
<kilo> ajmitch
<wt> I messaged asking him to update the Packages.gz so that it would work
<kilo> he is a kiwi, must be awake
<reinhard> :q
<reinhard> err
<andrewm> hmm?
<andrewm> wha?
<wt> reinhard, you use gaim?
<wt> reinhard, has it gotten better for IRC?
<kilo> reinhard: watch out...
<kilo> After a discussion with btami we have to this point:
<kilo> According to out database bible, there is only one database, no several databases.
<kilo> A database should be planned from top to bottom, and this way it clearly defines the need for any attributes or properties.
<kilo> Ie top-to-bottom planning identifies if one needs several address fields or address1-5 fields.
<kilo> But in the case of GNUe we stand on the other side. We want to provide a skeleton that can be used anywhere, or at least is adequate in many cases.
<kilo> Not knowing the real needs of solutions we want to help is a direct road to disaster.
<reinhard> wt: no, it hasn't gotten better
<kilo> So first imho we should define very VERY V E R Y clearly what GNUe packages are meant for.
<kilo> Like if we want to provide ar,ap,gl only, then address1-5 fields are enough.
<kilo> But if we want to provide a base for hr and payroll, then we need separated address fields, and our task is to find out what these fields are.
<reinhard> wt: but i also need jabber, and xchat can't do that :(
<wt> reinhard, I have both open
<wt> reinhard, what is your jabber id?
<reinhard> wt: only encrypted jabber to private server for fsfeurope
<wt> oh
<ajmitch_> that's a shame :)
<reinhard> kilo: you think that the addresses for customers will reside in the same table as the addresses for employees?
<wt> kilo and reinhard, there is an international standard for addresses
<kilo> they can reside in the same table, yes
<wt> addresses should be addresses, no
<wt> just all addresses in a separate table?
<wt> or is that too easy?
<wt> that way you could just create a new address when a person's address is changed
<wt> all references to the old would stay in tact
<wt> are you two off implementing?
<wt> http://xml.coverpages.org/ni2003-06-17-a.html
<reinhard> kilo: can you please precise:
<reinhard> kilo: they can reside in the same table, yes
<reinhard> do you mean *can*, *must*, or *should*?
<wt> they should, imo
<wt> they are the same type of object
<wt> why would that not make sense?
<reinhard> they are not the same type of object
<kilo> they should reside in the same table and be linked to different types of objects
<reinhard> the employee and the customer are not the same type of object
<wt> bad ass, this standard implements postal address templating also
<kilo> the address is of same type
<reinhard> they just happen to have some common fields
<wt> addresses are addresses
<kilo> the address links to employee or custoer
<reinhard> like you have purchase order and sales order, also you might have offer
<reinhard> all of them have fields like "item", "price", "discount"
<wt> why would an address link to a employee or customer?
<reinhard> so would you put all prices and discounts in the same table?
<reinhard> just because it's the same information?
<wt> reinhard, are you asking me?
<wt> I am confused
*** dimas_ has joined #gnuenterprise
<reinhard> actually both of you
<reinhard> :)
<wt> I think that an address is simply an address
*** dimas has quit IRC
<btami> yes
<wt> other objects can be composed of addresses and other components
<btami> yes
<wt> therefore, all addresses should be in the same table
<wt> I don't see a compelling reason to break it out
<kilo> wt: i agree totally
<wt> I also do not see any reason to link the address to another object
<wt> it seems to me that the link should go the other direction
<reinhard> now
<kilo> yes, other direction
<reinhard> is a price just a price?
<reinhard> and do we want all prices in the same table?
<wt> is a price structured?
<wt> or is it primitive?
<btami> reinhard: it depends
<reinhard> let's say it consists of amount and currency
<wt> well...I guess you can't store structured data in fields in all the supported databases
<wt> but that doesn't lead to the answer
<wt> I think they should be different
<wt> here is my reasoning
<wt> could there be multiple discounts on an item?
<reinhard> maybe
<btami> reinhard: all things depends on what you want to modell
<btami> in a simplest case 1 field is may be enough for strore the address
<wt> there is a base price
<btami> if no need for parts of the address
<wt> discount could just be like another item on the invoice
<wt> discount for some item
<wt> the item would link to the discount items
<wt> that would put all the prices and discounts in the same table
<wt> I think that sounds reasonable
<wt> discount items would have a negative value?
<reinhard> wt: all of them?
<reinhard> wt: those for offers, those for purchase items, those for sales items?
<reinhard> btami: the question is actually not how many fields
<reinhard> but this:
*** Vee2d2 has quit IRC
<reinhard> if several objects have similar fields, should these fields be pulled out of the object and put together in a common table
<wt> reinhard, depends if the things being pulled out represent an object that the other objects are composed of
<reinhard> and, if yes, what do we gain with it
<reinhard> and what do we lose
<reinhard> what we lose is quite clear to me
<wt> with the approach I listed above, you have history of prices whatnot if you never delete an entry in the table
<reinhard> a) performance (there will be more joins)
<reinhard> b) we lose the ability to structure different in different objects
<reinhard> e.g. have a more detailed structure for the address in employee than in customer
<wt> that's why I said you must make sure that the common things you pull represent objects that are used to make the aggregated objects
<kilo> reinhard: go back to the beginning. we need to define what we are aiming for.
<johannesV> reinhard, is your phone ok ?
<wt> why would you have more detailed addressing for employees than customers?
<kilo> modules that can be used separate AND together
<wt> kilo, that's where you get the base package, right?
<kilo> if we define a customer's address this way and an employee address that way, we store the same kind of data two different ways
<wt> kilo, or maybe there should be an addressing package
<kilo> wt, yes, that is we are debating now
<wt> I think that an addressing package would make sense
<wt> make it include (a) addresses & (b) address templates
*** Vee2d2 has joined #gnuenterprise
<wt> by (a) I mean definition of address objects
*** sjc has joined #gnuenterprise
<reinhard> so actually the question whether an address is an object by its own or not
<wt> with regard to the international standarad
<wt> I think an address is an object on it's own
<reinhard> like if an address had its own life
<reinhard> or would only exist as a group of fields in some other tables
<wt> if you decide to link an address to something, either add a column or better (imho) use another table for the link
<kilo> reinhard: does load-testing of appserver exist? ie several hundred thousand records, joins, etc
<wt> http://www.upu.int/post_code/en/address_elements_en.pdf
<kilo> wt: you need link table, two customers or two emplyees can share the same address
<wt> this is info about the address fields in the standard
<wt> not the actual list though
<wt> kilo, sounds fine
<wt> kilo, that's why I said it was better
<kilo> and one person can have several addresses of different kind
<wt> and people can have a list of addresses that represent all known addresses
<wt> so when one changes their address, a record is kept of prior addresses
<kilo> ie address is flagged active or inactive
<wt> http://www.idealliance.org/adis/ADIS01-1.pdf
<wt> kilo, or the most recent could be considered active?
<wt> or maybe there would be an active address field in the employee table?
<kilo> most recent cannot be considered active, that would mean implicit logic hardwires in the database
<wt> especially since the same address could be active for one and inactive for another
<wt> the active address field is prolly better
<reinhard> bbl
<wt> I am gonna try to sleep now
<kilo> wt: you work in postal business?
<wt> no
*** lekma has joined #gnuenterprise
<lekma> hi everyone
<wt> kilo, why?
<wt> lekma, hi
<kilo> just as you know these standards
<kilo> hi lekma
<lekma> hi guys
<wt> kilo, I have researched it before
<wt> the standard even has xml definitions
<kilo> wt: thx for your thoughts
<wt> which should make it super easy to transform with xslt
<wt> I will bbl
<wt> time for sleep
<wt> http://www.idealliance.org/adis/
<wt> we should see if we can get the xml stuff
<wt> seriously, I am out
<wt> http://www.idealliance.org/ <-- standard data interchange stuff
<wt> http://www.upu.int/post_code/en/international_addressing.shtml <-- blam
<btami> bbl
*** btami has quit IRC
<reinhard> kilo: just before i really leave
<reinhard> so you would introduce an n:n relationship between customer and address?
<reinhard> a customer can have several addresses and several customers can share the same address?
<reinhard> or an employee can share the address with a customer?
<kilo> yes, yes
<reinhard> i think we are targeting at different user groups
<reinhard> i am targeting at companies that would not use SAP because it's too big and bloated
<wt> reinhard, I don't think that bloats the user side
<wt> nor the dev side
<kilo> reinhard: i dont see why it would mean bloated
<reinhard> I do think so
<reinhard> however
<reinhard> i'd love to see a customer entry form
<reinhard> and let it test several of my users
<reinhard> i think you might provide such a form
<reinhard> and i will get feedback from some users
<reinhard> ok?
<wt> I think there would be a way to hide the complexity while keeping the schema the same
<wt> I don't know forms, yet
<wt> and I am waiting for the new version to get into debian
<wt> anyhow
<wt> I am really out now
<wt> for real
<wt> no looking back
*** Amorphous has quit IRC
<lekma> when calling classes gnue_label, gnue_parameter, gnue_procedure, gnue_property with generated forms (ie gnue-forms appserver://....)
<lekma> i get this error
<lekma> Traceback (mos recent call last):
<lekma>   File "<gnue_class.gnue_form>", line 14, in form
<lekma>   File "/usr/lib/gnue/appserver/labels.py", line 261, in __init__
<lekma>    self.properties = self.__getPropertyDict (self.fullName)
<lekma>   File "/usr/lib/gnue/appserver/labels.py", line 298, in __getPropertyDict
<lekma>    record = Property (prop)
<lekma>   File "/usr/lib/gnue/appserver/labels.py", line 80, in __init__
<lekma>    self.fieldLength = self.length or 0 + self.scale or 0 + 1
<lekma>  TypeError: unsupported operand type(s) for +: 'int' and 'NoneType'
<lekma> i copied by hand, so i hope there's no mistake in the traceback
<reinhard> lekma: can you try to change that line to
<reinhard> self.fieldLength = (self.length or 0) + (self.scale or 0) + 1
<reinhard> i.e. add the ()
<lekma> ok, wait a sec
<lekma> reinhard: yep it fixes it
<lekma> thx
<reinhard> lekma: can you give me your real name and your email address?
<lekma> yes
<lekma> Malek Hadj-Ali
<lekma> mh_at_orwex.net
<wt> lekma, where are you from?
<lekma> te _ to avoid adress reco in logs
<lekma> spam...
<reinhard> which is your last name? :)
<lekma> i'm french, living in amsterdam, netherlands
<lekma> Hadj-Ali
<reinhard> ok, thanks
<wt> lekma, you use gnue in production?
<lekma> wt: i'm going to put appserver in prod the 01/01/2005 for a customer
<wt> lekma, are you using any of the packages?
<kilo> wt, you still around?
<lekma> for the moment i'm building my own packages
<wt> kilo, I shouldn't be
<wt> I was asleep for a little while
<lekma> but i intend when the actual project
<lekma> is finished to help build standard packages
<lekma> we already have some thing in stock
<wt> lekma, how big is your customer?
<lekma> picking
<wt> lekma, have you ever used southware?
<lekma> it's a small project for the moment
<lekma> but it can lead to more big ones if it's successful
<wt> good luck
<lekma> southware? nope, don't know
<lekma> we are kind of SAP fluents here
<lekma> we that is the consultants that help me
<wt> lekma, ah
<wt> southware is mainly for small business
<kilo> reinhard: given class A, class B extends it, tehn will they be stored in the same physical table?
<wt> anyone, is there a way to sync up database transactions via gnue?
<wt> I am talking about pushing transactions to a central server in a fault tolerant way from satellite servers
<reinhard> kilo: yes
<wt> like take all transactions for the day and put them in the central server?
<kilo> ah, that is why extending properties must not be null?
<reinhard> they must not be "nonull"
<wt> kilo, we could get really cool and use inherited tables in postgresql
<kilo> i mean must be nullable
<reinhard> kilo: yes :)
* johannesV trying to setup gnue on w2k-notebook
<kilo> large wiener schnitzel is against thinking...
<lekma> kilo: lol :)
<wt> johannesV, don't like life?
<johannesV> well, i'd like to check out that locales-stuff on win*
<johannesV> i don't want to work with it ...
<lekma> johannesV: can i pass language parameter on the command line when i invoke gnue-forms
<lekma> ?
<johannesV> lekma, for a appserver-generated form ?
<lekma> yes
<lekma> well in general that would be cool
<kilo> johannesV: what does this error mean:
<kilo> InvalidParameter: NamespaceElement(gnue.forms.GFObjects.GFField.GFField) at 1104397228
<johannesV> appserver://appserver/form/foobar_form?language=xx_YY
<johannesV> kilo, can you explain that error a bit more ? where does it occur ?
<kilo> here on my computer 8-)))
<johannesV> *lol*
<kilo> i try to set a block's fields in a pre-commit trigger
<lekma> johannesV: does this work only with current svn?
<lekma> ie appserver://appserver/form/foobar_form?language=xx_YY
<johannesV> it should work with latest release
<johannesV> kilo, how are you doing this ? block.field.set (value) ?
<kilo> blk.fld = 'foo'
<kilo> no, it is ok.
<kilo> it happens then during the commit
<johannesV> kilo, changin such a field works for me
<johannesV> do you get a traceback ?
<kilo> change is done. it must be the commit then. traceback yes, a long one. paste it here or send it in email?
<johannesV> if i grep fro InvalidParameter i get matches in rpc-stuff only
<johannesV> s/fro/for/
<johannesV> kilo: of which type is that field you'd like to change ?
<lekma> johannesV: when tring appserver://appserver/form/foobar_form?language=fr i get gnue-forms: Parameter "fr" specified, but no value supplied.
<lekma> i have some labels in fr for class foo_bar
<johannesV> lekma, i assume you replace foobar_form with the name of a real class you have ...
<kilo> both a string and a number
<lekma> yep
<johannesV> kilo, what happens if you do "blk.fld.set (value)" instead of "blk.fld = value" ?
<lekma> oh btw i'm on win2k
<johannesV> lekma can you try gnue-forms "appserver://appserver/form/foobar?language=fr" ? (enclose the arg in double-quotes)
<kilo> johannesV: it is the same
<johannesV> kilo, strange
<johannesV> i cannot reproduce that error
<kilo> ok, i will try to set up a simple form to try and locate the error
<lekma> johannesV: it works but i get this
<lekma> DB000: C:\python23\lib\site-packages\gnue\forms\uidrivers\win32\ToolBar.py:69: FutureWarning: hex/oct constants > sys.ma
<lekma> xint will return positive values in Python 2.4 and up
<lekma> DB000:   himl = win32gui.ImageList_LoadImage(0, iconloc, 24, 0, 0xFF000000, win32con.IMAGE_BITMAP,
<johannesV> kilo, ok, i think this is something wired ... seems as if you pass a sequence-type in somewhere, which has GFObjects-elements
<lekma> but basically it works
<johannesV> lekma, so it seems as if win2k doesn't like that questionmark or that equal-sign (splits up arguments)
<johannesV> lekma, can you give me some hints on setting up gnue on win ?
<johannesV> do i need to install the source-packages somewhere ?
<lekma> well setup python, wxpython, pil, and yor db connector(psycopg for pg...)
<lekma> do you want svn or realeased version?
<johannesV> lekma, released version
<johannesV> (i just want to test some parts of gnue to verify it's behaviour on win2k, i don't want to use it there)
<lekma> if there is no win32 package already you need to setup all and install from source
<lekma> ie setup python, wxpython, pil, and yor db connector(psycopg for pg...)
<lekma> then python setu.py install for each gnue tool
<lekma> s/setu/setup
<johannesV> kilo, thanks i can reproduce it
<johannesV> lekma, so i need the source-packages from gnue-tools ...
<kilo> oh, that is good news
<lekma> johannesV: yes
<johannesV> lekma, ok thanks
<lekma> kilo: is that number you're trying to set a long int?
<lekma> cause xml-rpc doesn't like them
<kilo> it is stored as a string(4)
<lekma> oh, ok
<johannesV> kilo, got it working :)
<kilo> oh
<johannesV> hmm
<johannesV> just a sec ...
<johannesV> ok, now there's another problem :)
<johannesV> kilo, why don't you want the phone-field be modified directly ?
<johannesV> why do you need the unbound ?
<kilo> it is just an example. i want an invoicing form to have auto-generated entries.
<johannesV> the problem is the "unbound" field
*** vinsci has joined #gnuenterprise
<johannesV> if you make it really unbound you'd have to keep in sync with the 'real' field (fldLocPhone)
<johannesV> that auto-generated entries are form-specific ?
*** vinsci has quit IRC
<johannesV> kilo, you could use the default-attribute of the field instead
<kilo> cant seem to be able to receive the file, would you send it via email?
<johannesV> np
<kilo> they are generated, like invoice number...
<johannesV> ah, ok
<johannesV> so that form i've sent you solves the unbound-problem but not you're 'real' problem
<johannesV> at least this is what i feel
<kilo> hmmm, my real problem? national lottery picking wrong numbers instead what i've sent them? 8-)))
<kilo> got to leave now, thx everyone, bbl
*** kilo has quit IRC
*** jamest has joined #gnuenterprise
*** johannesV_ has joined #gnuenterprise
*** johannesV has quit IRC
<johannesV_> hrrr, found a bug in appserver
<reinhard> what??
<reinhard> impossible!!
<johannesV_> gacvs --selftest !
<reinhard> oh
<reinhard> haven't used that in ages...
<johannesV_> so did i :)
<SachaS> dont you guys have test cases? ;)
<SachaS> i admit i do not have either, yet
*** jcater has quit IRC
<SachaS> see ya later
*** SachaS has quit IRC
*** SachaS has joined #gnuenterprise
*** havoc has quit IRC
*** havoc has joined #gnuenterprise
*** kilo has joined #gnuenterprise
*** jcater has joined #gnuenterprise
*** dimas has joined #gnuenterprise
<lekma> DB000: ERROR:  column "gnue_compiledcode" does not exist
<lekma> DB000:
<lekma> DB000: SELECT gnue_name,gnue_type,gnue_class,gnue_code,gnue_length,gnue_id,gnue_compiledcode,gnue_module,gnue_scale,gnue_comment,gnue_nullable,gnue_language FROM gnue_procedure
<lekma> from gnue-gsdgen -d gnue -s -o gnue.new.gsd
<reinhard> lekma: you use svn ?
<lekma> yes
<reinhard> strange
<reinhard> that should actually be fixed
<reinhard> may i ask what you use gnue-gsdgen for?
<reinhard> hmmm.... i get the same error
<reinhard> lekma: fixed, thanks again
<dimas_> kilo: you here? i'd like to talk about packages
<dimas_> well, then read in the backlog..
<reinhard> off to lug meeting
*** reinhard has quit IRC
<SachaS> reinhard; what lug is it? Vorarlberg?
<kilo> dimas: will be available in 3 hours, must deal with kids until then
<lekma> reinhard: shouldn't the fix to share/repository.ini be also applied to src/classrep/repository.ini or only one of them is used?
<dimas_> kilo: my starting point is that i'm trying to gather all phone1, phone2 fields spreaded over db into one place
<dimas_> sorry, was interrupted
<dimas_> so in this i'm rather use "complex approach" to packages :)
<dimas_> on the other hand i do not need to much complexity in address schema
<dimas_> and i think we won't find one schema that will be ideal for everyone
<dimas_> what would be usefull for me is this:
<dimas_> gnue-packages from now provide a script like setup-cvs-db.sh that i (other application developer) could use and customize
<dimas_> that script allow me to easy choose what standart classes i would import from the place where gnue is installed
<dimas_> and i could add my own classes for loading also
<dimas_> regarding complex/simple schema, for example with address classes
<dimas_> there could be helper address classes that define simple entities like country, region, towns and the like
<dimas_> i can import/load them and use/not use as i like
<dimas_> and then there could be address-simple.gcd or address-complex-like-hell.gcd and i could choose which one to work with
<dimas_> they use as much as possible all classes from adress-helper.gcd
<dimas_> and if i need something absolutely different i could write my own gcd
<dimas_> one thing is also needed - forms to deal with either schemas
<dimas_> and all things - setup script, gcd, forms should be in sync always
<dimas_> that way i could with one click install standart classes/official packages and play around
<dimas_> does it make sense? anyone?
*** wt has quit IRC
<SachaS> <dimas_> and i think we won't find one schema that will be ideal for everyone
<SachaS> dimas_ makes sense to me
<SachaS> i think we can foresee that there will be no one fits all solution
<SachaS> so we have to get a flexible solution above this assertion.
<SachaS> dimas_ question. what if a package depends on address?
<SachaS> for example billing class
<SachaS> must there be at least two versions of billing class?
*** dimas has quit IRC
<SachaS> one which depends on smiple_addres.gcd and one based on address-complex-like-hell.gcd ?
<SachaS> one like simple-billing-based-on-address-simple.gcd and one simple-billing-based-on-address-complex-like-hell.gcd ?
<SachaS> what about billing-complex-like-hell ?
<SachaS> would that also need 2 versions, one depending on address-simple.gcd and address-complex-like-hell.gcd?
<SachaS> will this get a nightmare of packages depending on other packages?
<SachaS> does this make sense too? anyone?
<SachaS> billing-complex-like-hell-based-on-address-complex-like-hell.gcd
<SachaS> billing-complex-like-hell-based-on-address-simple.gcd
<dimas_> SachaS: i'm for one and simple solution where we find it :)
<SachaS> you started with the complex-like-hell one ;)
<SachaS> i start to think there is no "simple" solution
*** johannesV_ has quit IRC
<dimas_> SachaS: where i started with complex solution in my phone1,phone2 schema or in proposing multiply gcd's for each every things?
<dimas_> it is what really happens and that is why i think that setup script and accompaning fully functional forms is more important
*** btami has joined #gnuenterprise
* kilo is here
<ajmitch> hi
<kilo> hi ajmitch
<kilo> dimas: i see what you are telling, but am puzzled a little bit. I will need to talk to Reinhard on this matter.
<dimas> ok, thanks
<kilo> btw have you seen the forms with the LOC module? if yes, what do you think about them
<dimas> i did not run them yet
<kilo> ok
<dimas> will do that in next couple days
<dimas> kilo: what about setup for gnue-packages?
<dimas> now i use Makefile that looks like
<dimas> init:
<dimas>         gnuenv
<dimas>         #Init appserver's db
<dimas>         gsscvs --connection=taxcom $(APPSERVERINIT)/gnue.gsd
<dimas>         #gldcvs --connection=taxcom $(APPSERVERINIT)/base-*.gld
<dimas>         #Define base classes
<dimas>         #gcdcvs --connection=taxcom $(APPSERVERPACKAGES)/base/util/cdict.gcd
<dimas>         #gsscvs --connection=taxcom $(APPSERVERPACKAGES)/base/util/insert-code-descr.gsd
<dimas> ...
<dimas> and then
<dimas> my own classes
<dimas>         #Define custom classes
<dimas>         gcdcvs --connection=taxcom auth.gcd
<dimas>         gsscvs --connection=taxcom taxcom-auth.gsd
<dimas>         gcdcvs --connection=taxcom address.gcd
<SachaAway> i did the ssame dimas :)
<dimas> let's put one in svn :)
<kilo> hmmm, about the same for me too
*** btami has quit IRC
*** jamest has quit IRC
*** cilkay has joined #gnuenterprise
*** wendall911 has joined #gnuenterprise
*** sjc has quit IRC
*** kilo has quit IRC
*** qman has quit IRC
*** MiXi has joined #gnuenterprise
*** qman has joined #gnuenterprise
*** MiXi^ has quit IRC
*** gsoti_away has joined #gnuenterprise
*** wendall911 has quit IRC
*** wt has joined #gnuenterprise
<wt> anyone home?
<ajmitch> yep
*** gsoti_away has left #gnuenterprise
<wt> ajmitch, you are here?
<ajmitch> yes
<wt> ajmitch, did you happen to update the .gz files on your deb source?
<chillywilly> ajmitchie
<ajmitch> nope
<chillywilly> yes
<chillywilly> you are the ajmitchie
<chillywilly> :P
<wt> ajmitch, will you?
<wt> ajmitch, you'd rock to high heaven
<chillywilly> you got any money?
<ajmitch> wt: do you have problems with the ones currently in debian?
* ajmitch doesn't usually have his debian area open because of bandwidth limits
<chillywilly> yea the NZ telco is particularly evil
<wt> ajmitch, "binaries" that don't (a) have the same names as development and (b) don't read the username/password out of connections.conf
<wt> ajmitch, unless 5.8 is in sid now
<wt> 5.8 of gnue-common, that is
<ajmitch> dont'  have same name as development? what do you mean?
<wt> gnue-gcd2sql is now gnue-readgcd or something
<ajmitch> aha
<ajmitch> is that a big problem? :)
* ajmitch has updated it
<wt> I can't use the scripts from svn with the stable release....boo hoo :((
<wt> Thanks
<wt> I can't believe gaim finally called itself 1.0.0
<wt> what pushed it over the edge?
* ajmitch shrugs
<ajmitch> wonderful, can't mount a cd
<ajmitch> can't eject it either
<wt> beautiful
<wt> what about as root?
<ajmitch> nope, not at all
<ajmitch> getting plenty of kernel errors about it though :)
<chillywilly> bah
<chillywilly> 2.6?
<ajmitch> yup
<chillywilly> I have experienced similar things
<ajmitch> I can mount the dvd drive underneath it just fine
<chillywilly> mainly when I try to burn CDs
<chillywilly> without the proper drivers
<ajmitch> aha
<ajmitch> 2.6.8(,.1) are broken for cd burning
<chillywilly> oh I haven't been keeping up to date
<chillywilly> Linux gandalf 2.4.26-1-386 #1 Thu Jul 22 12:46:23 JST 2004 i686 GNU/Linux
<chillywilly> woops
<chillywilly> wrong box :P
<chillywilly> Linux frodo 2.6.4-1-686 #1 Sat Mar 13 19:25:38 EST 2004 i686 GNU/Linux
<chillywilly> old
<wt> ajmitch, 2.6.8.1 only allows burning as root
<ajmitch> wt: unless you patch it
<chillywilly> why?
<ajmitch> 2.6.8.1 also has a serious issue with burning audio cds
<wt> chillywilly, the devs say that normal users can do unholy things otherwise
<ajmitch> it'll leak a few hundred MB of RAM
<wt> ajmitch, that's not fun
<chillywilly> wt: what?
<ajmitch> Bad Things include wiping drive firmware
<chillywilly> wtf
<chillywilly> so they roally broke it?
<ajmitch> which is why it's restricted to root only now, by capabilities
<chillywilly> lame
<ajmitch> or you can set cdrecord suid-root ;)
<chillywilly> bah
<wt> they will fix it in time
<chillywilly> 2.6.x claims to be a "stable" kernel
<chillywilly> they lie
<chillywilly> bastids
<wt> chillywilly, no they don't
<wt> you don't have to upgrade 2.6.8.1
<wt> stay at 2.6.7
<chillywilly> so wtf is it even numbered then?
<wt> chillywilly, I guess you haven't read the new development mentality?
<chillywilly> crack heads
<chillywilly> no
<chillywilly> I don't hack on the linux kernel
<wt> chillywilly, what would you rather they do? (1) fix a glaring security hole or (2) not
<wt> the memory leak thing is a bug and bugs happen
<chillywilly> wt: why so defensive?
<chillywilly> wt: did you cause the bug?
<chillywilly> wt: maybe you should try hurd development ;)
<wt> I like Linux
<wt> that's why I am defensive, I guess
<wt> I tried to install hurd once
<chillywilly> it's GNU/Linux :)
<wt> maybe they will get it running on l4 like they promised so long ago eventually
<wt> it's Linux
<chillywilly> if you'r talking only about the kernel
<chillywilly> :)
<chillywilly> you're*
<wt> I am probably not talking about ls or grep when I say that
<wt> chillywilly, are you really rms?
<chillywilly> is your mother rms?
<ajmitch> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=241343
<chillywilly> lol
<wt> that's funny as hell
<wt> lol
<wt> is a .gfd a form?
<wt> ouch....traceback
<wt> if the query fails while using gnue-readgcd
<wt> meaning I need to load another file first
<chillywilly> must be that linux kernel bug ;)
<ajmitch> gnue needs a decent bugtracker
<chillywilly> what about dcl?
<wt> ajmitch, like DCL?
<chillywilly> ?!?!?1
<chillywilly> hehehehe
<wt> oh wait, it's b0rken
<wt> circular dependency
<wt> icky
<wt> nvm
<wt> I thought there was one for a sec
<jcater> woohoo
<jcater> my new assistant started today
<jcater> poor guy
<wt> jcater, what does he do?
<wt> great
<wt> now I can't connect to my data sources
<jcater> general programming/tech support
<jcater> stuff I don't want to do
<wt> jcater, ah
<wt> jcater, anything with gnue?
<jcater> over time
<ajmitch> jcater!
<jcater> not initially though
<jcater> ajmitch!
<wt> what do I use for the provider for postgresql?
<jcater> I use pypgsql in sarge & sid
<jcater> psycopg, iirc, if in woody
<ajmitch> I wonder if reinhard will have an appserver 0.2 out soon
<wt> still unable to connect
<wt> damn
<ajmitch> wt: check that you're not trying to connect via TCP/IP when postgres is setup for only UNIX domain sockets
<wt> ajmitch, it is listening
<wt> I just think the auth is f'ed
<ajmitch> gnue or postgres problem?
<wt> postgres config
<jcater> fwiw, gnue can use unix domain sockets twoo
<wt> my bad
<jcater> provider = pypgsql
<jcater> host = /var/run/postgresql
<jcater> dbname = dataentry
<wt> see...I use localhost -> ::1 so that I can use password auth on the ipv6 interface
<wt> I have ident setup on the ipv4 interface
<wt> I was accidentally connecting the wrong one
<wt> damn
<wt> still unable to connect
<wt> how do I output some debug so I know why it is failing
<wt> all I am getting is that it is unable to connect
<jcater> -d10
<wt> option d not recognized
<wt> this is the designer
<jcater> err, --debug-level 10
<wt> DB001: 0:00:15.468 [DataObject:102] DB-SIG database driver backend initializing
<wt> DB001: 0:00:15.469 [GConnections:302] Attaching to DataObject_Object (object)
<wt> Unable to connect to gnue
<wt> 
<jcater> that's all the info?
<wt> yep
<jcater> what's your connections.conf look like?
<ajmitch> looks like someone finally snapped on the debian-x mailing list ;)
<wt> [gnue]
<wt>       7 comment = GNUe Test Database
<wt>       8 aliases = dcl test
<wt>       9 provider = pypgsql
<wt>      10 host = localhost
<wt>      11 dbname = gnue
<wt>      12 username = wt
<wt>      13 password = xxxxxxxxxxxxx
<wt> ajmitch, about damn time
<wt> ajmitch, link?
<ajmitch> http://news.gmane.org/gmane.linux.debian.devel.x
<wt> what message?
<wt> I hate frames
<wt> what is the subject?
<ajmitch> all of the ones at the top :P
<wt> the no subject ones?
<ajmitch> yes
<wt> I must leave my dwelling in search of cake
<wt> bbl
<wt> jcater, btw appserver starts up just fine
<wt> jcater, so it seems to be a designer issue
<wt> I am outta here for a bit, though
*** wt has quit IRC