*** SachaS has quit IRC
*** johannesV has joined #gnuenterprise
<johannesV> good morning
*** SachaS has joined #gnuenterprise
*** Muile has joined #gnuenterprise
*** reinhard has joined #gnuenterprise
<reinhard> good morning
<reinhard> alexbodn: the sample database is in zipcode.gsd
<reinhard> you can read that into your db with gnue-schema
*** btami has joined #gnuenterprise
*** Muile is now known as kilo
*** SachaS has quit IRC
*** SachaS has joined #gnuenterprise
<reinhard> ouch
<reinhard> gnue-forms 0.5.x loses all changes without any confirmation if you switch to query mode
<reinhard> customer just learned that the hard way...
<johannesV> bahh
<reinhard> funny enough I just changed it in svn head a few days ago
<alexbodn> hi reinhart. thanks, let me see
<alexbodn> sorry reinhard: no data for the zipcode table.
<alexbodn> i have found compatible data and posted to gnue-dev this morning :-)
<reinhard> yes, we have no data for the zipcode table
<reinhard> only the states table is filled out
*** btami has quit IRC
*** johannesV has quit IRC
*** jamest has joined #gnuenterprise
<alexbodn> hi all.
<alexbodn> i would ask your help to activate the inventory demo: the data is ok in the gnue database, which is in the [gnue] config section.
<alexbodn> there is a query that complains there is no oid. i wish to change this query, but i don't know where is it saved.
*** johannesV has joined #gnuenterprise
<reinhard> alexbodn: ouch
<alexbodn> so?
<reinhard> alexbodn: I am not sure the inventory demo is functional any more at all
<reinhard> I haven't tried it in literally years
<alexbodn> what demo may i try?
<reinhard> are you running release or svn?
<jamest> alexbodn: postgresql dropped the oid column by default in new tables
<jamest> so i imagine you're using a newer postgresql as your storage system right?
<alexbodn> i know about this. i can change this behaviousr, but i'd prefer to change the query (and learn about the system a littel :-))
<alexbodn> s/ousr/our/
<jamest> if a table doesn't define a primary key then our system falls back to using oid
<alexbodn> s/littel/little/
<jamest> depending on what you're using in gnue (like say a .gfd) you can tell the datasource which columns to use as the primary key
<alexbodn> ok, but where is the form data source defined?
<jamest> in a gfd file
<alexbodn> i see. but what option is the primary key?
<jamest> i'm checking now
<alexbodn> i'm here, thanks in advance :-)
<jamest> in the <datasource> tags in a gfd file you can add a primarykey="field,field2,field3"
<jamest> which is a work around when missing a primary key def on a table
<alexbodn> you mean this would be a 'kind of de-facto unique combination'?
<jamest> yes
*** kilo has quit IRC
<jamest> unless things have changed greatly the sql gnue generates used to compare every field along with the oid
<jamest> but it still expected it
<jamest> primarykey IIRC short circuits that and uses the defined fields to determine the proper record
<jamest> reinhard: is that still the case?
<jamest> reinhard: i svn'd up :)
<alexbodn> so no way around oid?
<jamest> it only happens when a table is defined without a PK
<jamest> which is really a bad practice that i'm very guilty of
<jamest> so I imagine the demo was setup missing a PK
<alexbodn> all tables here have pk
<jamest> wha?
<jamest> then it should be fine
<alexbodn> but not in the datasource
<jamest> shouldn't be needed there IIRC as it tries the table first
<jamest> and falls back to OID
<jamest> i think
<jamest> alexbodn: which demo are you running?
<alexbodn> inventory
<alexbodn> may there be more than a form in a gfd (at least here)?
<jamest> no
<jamest> gfd = gnue form definition
<alexbodn> ok
<jamest> you can have a form and multiple dialog defs in a single gfd
<jamest> a dialog is just a form of type dialog used to generate custom input popups
<jamest> which form are you trying to run in there
<jamest> as there are 3
<alexbodn> why should i need different dialogs for one form?
<jamest> an example that I have here
<jamest> in shipping as they scan the UPCs codes of items and put in a box
<jamest> the system checks to see if it was really part of that order
<jamest> if not it pops up a custom dialog that says something like
<jamest> this isn't part of the order, type "YES" to continue
<jamest> as my users tend to just click through anything without reading it so they can get back to whatever it is they really want to do
<alexbodn> i understand: dialogs don't have to represent the form data, but may be any msgbox
<jamest> yes
<jamest> alexbodn: with the latest svn checkout i can't run anything in that demo
<jamest> nor any of my own forms it would seem :)
<alexbodn> i begin to love it, even though my demo shows no data :-( yet!
<jamest> one of the traditional demos that was used a lot in debugging by a few of us was the zipcode stuff
<jamest> i'm not sure what reinhard and johannesV and co use now
<johannesV> still using zipcode
<johannesV> :)
<jamest> but zipcode forms are simple and show several features
<jamest> johannesV: :)
<alexbodn> i have the version packaged in debian testing. the moment i might do serious gnue based development, i'll allways have the latest and greatest :-)!
<alexbodn> so i'll try zipcode. btw, i've posted on gnue-dev a method to fill the zipcode table
<jamest> johannesV: wx drive in svn is toast it would seem
<jamest> wx26 is working....sorta
<jamest> seems like not svn'n up for months is going to cost me a bit ;)
<reinhard> jamest: do you have tracebacks?
<reinhard> can you describe the symptoms?
<jamest> yeah, tracebacks :)
<johannesV> :)
<reinhard> jamest: give them to me!
<jamest> reinhard: give me a bit, i'm figuring out if it's my devel install, my forms, or an issue with the code
<reinhard> or you fixing yourself?
<reinhard> I'm 95% positive it's an issue with the code
<reinhard> I did some changes lately that affect all uidrivers and I did the changes in wx ui without being able to test at all
<jamest> i think i'll take a look as if I have to plone/zope any more today i think i'll lose it
<reinhard> alexbodn: when the db has a primary key it doesn't help if in the <datasource> definition in the GFD it isn't mentioned
<reinhard> jamest: heh, you're gnueing for recreation? ;-P
<alexbodn> right, it stopped complaining only after i have added the primarykey option
<reinhard> jamest: anyway, I'll help in any way I can
<jamest> well, at home I am still working on the house for "recreation"
<jamest> and i've been ploning and javaing for the last few months non-stop at work
<jamest> so it's 8 hours of suck followed by an evening of suck
<jamest> i think I can gnue just a little for mental health reasons :)
<jamest> or is that if I work on gnue i must be mental?
<alexbodn> the zipcode table has the primarykey field zipcode too, while in the zipcode.gfd demo it seems to have been referenced as "zip"
<alexbodn> now i changed zip to zipcode probably in the wrong options :-(
<alexbodn> i mention, i'm using the sql to create the schema, not the gsd file (yet)
<reinhard> is the sql still there?
<reinhard> i thought we completely replaced it by the gsd
<jamest> hmmm
<jamest> we have a bigger issue than i thought
<alexbodn> i took it from another demo :'(
<jamest> as I just created the zipcode table with a PK
<jamest> and the datasource still attempts to use the oid column in the query
<alexbodn> right
<reinhard> yes
<reinhard> because the primary key field can change on inserts if there is a trigger on db side
<reinhard> so the only really reliable record identifier is the oid
<jamest> but the oid isn't in newest postgresql
<jamest> i thought we used the previous values of the fields during updates
<jamest> so every field had to match
<alexbodn> you may create the table with oids option
<alexbodn> brb
<jamest> alexbodn: yes, but then we're not friendly w/ the default postgresql config moving forward
<reinhard> jamest: are oid's depreciated to use or are they just turned off by default but still a valid option?
<alexbodn> i accept it, but as we are programmers, we should just use the documented recipes ;-)
<reinhard> turning off oid usage for postgres is a one-line change
<reinhard> but it takes away the possibility to use on-insert triggers on the db
<jamest> i'm pretty sure it's depreciated
<jamest> but let me grab the latest tarball and read the release notes
<jamest> hmmmmm
<jamest> would we need to drop it completely
<jamest> or could we check the table schema to see if it exists and use it if available
<alexbodn> brb
<jamest> that way we could qualify on tables that use on-insert triggers you have to have an oid
<reinhard> I'm not sure the oid is listed as a column when you query the schema in postgres
<jamest> reading the release notes it looks like OIDs are going to be around for a while
<jamest> just not created by default
<jamest> reinhard: the form isn't getting added as a key in _gfObjToWidget for the wx24 driver
<jamest> how long ago did you change that so I can look at the diffs to see what you were doing?
<jamest> as it's not giving a keyerror in line 298 of the uidrivers/_base/UIDriver
<jamest> actually , scratch that
<jamest> hmmmmm
<jamest> in the base ui driver we have a huge try: except KeyError: block that wraps the entire method
<reinhard> jamest: sorry, on the phone
<reinhard> *long* phone call
<reinhard> :(
<jamest> what's happening is in the 2.4 driver
<jamest> that that in the _buildUI function in the base the line that read
<jamest> supported = self.supportedWidgets [object._type]
<jamest> is generating a key error that is caught by the function wide wrapper
<jamest> and ignored (as it should be for triggers, connections and such)
<jamest> now I'm trying to figure out why 2.4's form widget isn't using the configuration {} setup in it's form/__init__.py
* jamest is verbally going through this in case reinhard has a "Whoops I forgot to change foo" moment
*** klasstek has joined #gnuenterprise
<alexbodn> hi, i'm back :-)
<reinhard> so far I hand't even had a "uuhh.. I messed with this" moment
<alexbodn> hoorai, i did it ! :-P
<alexbodn> i can see the zipcode demos
<jamest> reinhard: :)
<jamest> we ignore a lot of excepts in the buildForm process
<reinhard> yes
<reinhard> that's a part that I didn't go through yet
<jamest> in this case a bad import in form/widget.py
<jamest> was causing it to not generate an error but instead be dropped from the supported widget list
<jamest> lol
<jamest> next!
<reinhard> hmmm... actually giving a primarykey field should shop forms from using oid
<reinhard> <datasource primarykey="foo">
<jamest> reinhard: it's not working
<jamest> i don't know why yet
<alexbodn> it is (for me)
<jamest> as that is my work around in my production system
<jamest> hmmmmmmm
<jamest> maybe I screwed up my test setup
<jamest> as i do use that in production with 8.1
<jamest> sigh, it helps to save the .gfd with the primarykey="" added before testing to see if the change worked
<jamest> reinhard: 24 works again, however it seems that buttons on a form break it during a rollback
<reinhard> jamest: removing the RESPONSE is not good
<alexbodn> pardon my ignorance: in the master/slave zipcode form, how is the link between the blocks deduced?
<alexbodn> only by implied foreign key?
<jamest> reinhard: i sync'd it up with the way wx26 does it
<jamest> let me dbl check
<reinhard> _RESPONSE should be a global variable rather than an instance variable now
<reinhard> which probably was my mistake
<reinhard> to not remove the "self."
<alexbodn> sorry, i found the detaillink option :-)
<jamest> hmmmm
<jamest> 26 doesn't return response either
<reinhard> err that is the exception handler?
<reinhard> not the _ui_show_message_?
<reinhard> erm yes it is
<reinhard> scratch what I said, your changes were perfectly right
<reinhard> I wonder how buttons can break a rollback, as rollback doesn't do anything ui special IIRC
<jamest> doh' i said rollback
<jamest> i meant query
<jamest> if I open zipcode.gfd
<jamest> and immediately press f8/f9
<jamest> i get
<jamest> GFButton instance has no attribute '_field'
<jamest> in GFBLock line 913
<jamest> however that button is associated with a block="zip" in the gfd
<jamest> and if I remove it then she's fine
<alexbodn> pardon me please, but how can i use query datasets, and use say, picture fields?
<jamest> use query datasets?
<jamest> you mean perform a query from a form?  or something else?
<alexbodn> using a select statement
<jamest> that would be read only I believe
<jamest>   <datasource name="dtsBranches" connection="epm" prequery="Y" type="sql">
<jamest>     <sql>
<jamest>       sql statement
<jamest>      </sql>
<jamest>   </datasource>
<jamest> if you need read/write
<jamest>   <datasource name="dtsInvoice" connection="epm" table="invoice">
<alexbodn> is it straight forward too? :-)
<jamest>    <condition>
<jamest>       <and>
<jamest>         <or>
<jamest>           <eq>
<jamest>             <cparam name="queryType"/>
<jamest>             <cconst type="char" value="A"/>
<jamest> then close the tags
<jamest> i could paste the whole example there but it's fugly
<jamest> ah
<jamest> here's  a simple one
<alexbodn> it's wonderfull!
<jamest> from my incoming fax assignment form
<jamest>   <datasource name="dtsXref_image" connection="epm"
<jamest>               order_by="claim_no,entry_timestamp,url" prequery="Y" primarykey="url" table="xref_image">
<jamest>     <condition>
<jamest>       <null>
<jamest>         <cfield name="claim_no"/>
<jamest>       </null>
<jamest>     </condition>
<jamest>   </datasource>
<jamest> also happens to have an image field
<jamest>       <image name="imgURL" Char:height="25" Char:width="94" Char:x="1"
<jamest>              Char:y="3" block="blkXref_image" field="fldUrl"/>
<jamest> in this case the image is stored outside the table with a url field in the table
<alexbodn> i quite love this text based metrics
<alexbodn> i was meaning to calculate the image, like a map or a graph
<jamest> i believe reinhard and friends have added a newer layout style into forms as well
<jamest> i've not looked at it yet though
<jamest> i don't believe you could currently generate an image on the fly
<jamest> though you could display a filesystem image and manipulate it inside triggers
<alexbodn> i'm looking forward to convert access applications to gnue, and later, to finish an autoconvertor
<jamest> which is how i do my fax assignments
<jamest> the user has a set of rotate buttons and a scale dropdown
<alexbodn> i'mm looking forward to involvement with gnue
<jamest> which uses triggers to manipulate the image
<jamest> once the user gets what they want they assign it to the appropriate order (a claim)
<alexbodn> is there a larger quantity of documentation and working examples?
<jamest> fwiw, i have to do something like
<jamest>           # Poor man's refresh.
<jamest>           blkXref_image.nextRecord()
<jamest>           blkXref_image.prevRecord()
<jamest> in the trigger to get the image to refresh, but i imagine that <image> could be enhanced to provide more functionality
<alexbodn> well known everywhere! ;-)
<reinhard> jamest: button query problem should be solved, can you try please?
<jamest> samples are all in gnue-forms/samples and there is a gnue-samples package
<jamest> docs are a weak point
<reinhard> documentation is in the files ending in .py
<jamest> exactly
<jamest> just like plone and zope
<jamest> :)
<reinhard> but you are highly encouraged to aks for help here at any time
<dimas_> where is the program then?
<alexbodn> is there any plan of a web ui?
<reinhard> dimas_: lol
<reinhard> alexbodn: several, all of them stalled
<reinhard> there were attempts on a php based ui
<reinhard> then a pure html ui
<jamest> a javascript ui
<reinhard> then another php attempt
<alexbodn> no problem, gtk2 is wonderful too ;-)
<reinhard> ah, yes, javascript also
<jamest> reinhard: yeah, that did it
<jamest> reinhard: fwiw, that sloppyQuery thing
<jamest> that was put in by request here eons ago when I was consulting
<jamest> and we've never once used it
<reinhard> I've neither used it
<jamest> if it looks as ugly to you as it does to me I wouldn't be against depreciating it
<alexbodn> thanks a lot folks. i'm in. i'll have a dayly intake of gnue, and hope to be also able to contribute
<jamest> what it does if you've not looked
<jamest> is replace your field query values of
<jamest> 'reinhardRocks'
<jamest> with
<reinhard> I know
<reinhard> %foo$
<reinhard> err
<reinhard> %foo%
<jamest> %f%o%o% actually
<reinhard> eek
<jamest> yeah
<alexbodn> :-)
<reinhard> that's indeed pretty ugly
<jamest> they insisted
<jamest> as they wanted to instead of typing
<alexbodn> thanks a lot. bye for now
<jamest> bye
<reinhard> bye alexbodn
<jamest> 11-222-333-4444
<reinhard> hope to see you again here :)
<jamest> to type
<jamest> 22444
<jamest> and hit all matches
<alexbodn> bye all. you are doing great work!
<reinhard> so "foo" would also match "fuck our boss"? ;-P
<jamest> my for johannesV
<jamest> maybe for
<jamest> er, i mean yeah
<jamest> it would
<alexbodn> sure, it's worth!
<reinhard> jamest: if you think nobody uses it you can as well just remove it
<alexbodn> bye all
<reinhard> I won't object at all
<jamest> reinhard: it's somewhat scary how quick you expanded foo into that
<reinhard> lo
<reinhard> lol
<reinhard> jamest: I am going to leave for 2 days in 3 hours
<reinhard> so if you have any questions remaining please ask now
<reinhard> or ask johannes tomorrow :-)
<jamest> i think i'm good for now
*** jamest has left #gnuenterprise
<reinhard> wow
<reinhard> this was a lot less troubles than I expected
<reinhard> :)
*** jamest has joined #gnuenterprise
<reinhard> cu all
*** reinhard has quit IRC
*** johannesV has quit IRC
*** jcater has joined #gnuenterprise
*** jamest has left #gnuenterprise
*** kilo has joined #gnuenterprise
*** klasstek has quit IRC
*** kilo has quit IRC
*** derek has quit IRC
*** curt__ has joined #gnuenterprise
*** curt_ has quit IRC