*** Morphous has joined #gnuenterprise
<someon> holycow: How goes the setup of your local repository?
*** Amorphous has quit IRC
*** btami has joined #gnuenterprise
*** reinhard has joined #gnuenterprise
<someon> WRT: Date/Time fields...  the extensible ERP I use at the moment stores Date Time fields as Unix Timestamps UTC...
<someon> Not sure if that's doable or not...
<someon> Date is stored at Unix Timestamp UTC of <Date> 00:00:00
<someon> So if you search the db on a date or date/time >= start date, it just compares the ints.  pretty easy for the SQL to do...
*** johannesV has joined #gnuenterprise
<someon> It does do nice things for searching eg. > DATE or < TIME
<someon> Other thing it does...  if you enter a TIME in a DATE/TIME field, it assumes you mean today
<someon> Blah.  Logs haven't written in like 1/2 hr.
*** someon has quit IRC
*** SachaS has joined #gnuenterprise
*** dimas has joined #gnuenterprise
<SachaS> hi reinhard
<reinhard> hi SachaS
<SachaS> had a good last night. could get reports to do a "serien brief" with data from my invoices.
<SachaS> has nothing to do with appserver but got me very excited.
<SachaS> a couple of days ago, gnue navigator had problems switching between froms. now it has no issues.
<SachaS> reinhard: did you see the screenshots?
<SachaS> actually, it seems my server is down :(
*** dimas has quit IRC
<btami> reinhard: i get errors with intro.gfd like
<btami> Runtime Error occured:
<btami>  Traceback (most recent call last):
<btami>   File "<unknown_executioncontext>", line 4, in _messageBox
<btami>   File "<string>", line 1
<btami> SyntaxError: can't assign to operator (<string>, line 1)
<btami>  
<btami> same comes from _jumpto, _about
<reinhard> btami: when does it happen?
<btami> seems the new trigger code doesn't accepts "-" in trigger names 
<reinhard> ah
<reinhard> thanks
<btami> np
<reinhard> will fix
*** kilo has joined #gnuenterprise
<kilo> morning all
<SachaS> hi kilo
<kilo> SachaS: can i take a look at the screenshots you mentioned last night?
<kilo> @weather LHUD
<bigbrother> kilo: The current temperature in Szeged, Hungary is 8°C. Conditions: Mostly Cloudy.
*** btami has quit IRC
*** dimas has joined #gnuenterprise
<SachaS> kilo. my server is down :(
<kilo> oh, i see
<SachaS> my friends server is down as well...might be, that they have a network problem.
*** PeterD has joined #gnuenterprise
<PeterD> Hi Reinhard
<PeterD> tried snapshot dated 20040330 and found problem with instaling Forms:
<PeterD> Traceback (most recent call last):
<PeterD> File "setup.py", line 229, in ?
<PeterD> setup().run()
<PeterD> File "/usr/local/gnue/lib/python/gnue/common/setup/GSetup.py", line 283, in run
<PeterD> _setup.set_params (setup_params)
<PeterD> File "setup.py", line 90, in set_params
<PeterD> samp19files = self.allfiles ("samples/connections")
<PeterD> File "/usr/local/gnue/lib/python/gnue/common/setup/GSetup.py", line 246, in allfiles
<PeterD> return [directory + "/" + file for file in os.listdir (directory) \
*** lekma has joined #gnuenterprise
<reinhard> PeterD: just a second
<reinhard> i'm on the phone
<reinhard> ok
<reinhard> here now
<reinhard> hi PeterD
<PeterD> hi reinhard
<PeterD> cannot install Forms
<reinhard> found the problem
<reinhard> i can provide a quick fix for you
<reinhard> can you edit setup.py?
<PeterD> got it ready
<reinhard> ok
<reinhard> search for "samp19"
<reinhard> the first line you find there, please simply delete
<PeterD> got it
<reinhard> the second line you find,
<reinhard> please also delete
<reinhard> and change the line above
<reinhard> to end in a ] instead of a ,
<reinhard> that should fix this
<reinhard> thanks for testing and reporting btw
<reinhard> you seem to be the only one playing with our snapshots currently
<reinhard> at least you are the only one reporting bugs
<reinhard> which is something we appreciate *highly*
<PeterD> loks like
<reinhard> did you already try appserver?
<PeterD> not yet
<reinhard> ok
<PeterD> forms are ok now
<reinhard> excellent
<SachaS> uploaded new version of appserver dev. manual (includes howto create a pdf mail bulk report with appserver)
<SachaS> also has a table of contents ;)
<reinhard> :)
<reinhard> SachaS: great
<SachaS> reinhard: i could not believe how easy it is to create a pdf bulk mail with reports!
<reinhard> it's so much more fun coding when people are acutally doing something with it :)
<reinhard> SachaS: that's good
<SachaS> reinhard: i had a test on this afternoon: i want to extend a class :)
<reinhard> SachaS: I have not looked at reports until now
<SachaS> ..that is ... i will do a test of extending a class soon.
<PeterD> I got appserver runing with forms now
<SachaS> reinhard: the canned tuna fish is also on my check-it-out list
<PeterD> thanks
<SachaS> canned-tuna-fish = canned list property
<reinhard> PeterD: great
<reinhard> PeterD: happy that i could help you
<lekma> SachaS: where did u upload 'new version of appserver dev. manual' ?
<SachaS> lekma. hold on
<SachaS> http://www.gnuenterprise.org/tools/appserver/docs/
<SachaS> the how to guide
<lekma> thxs
<PeterD> I have to go, by to all
<reinhard> bye PeterD
*** PeterD has left #gnuenterprise
<lekma> the dcl doesn't seem to work on gnuenterprise.org...
<lekma> how can i report a bug?
*** kilo has quit IRC
<reinhard> lekma: write it here
<lekma> well... ok: import-tags are broken
<SachaS> lekma: reinhard is a living dcl
<lekma> i.e. import tags in form
<lekma> and i'm too new to python to understand why
*** reinhard_ has joined #gnuenterprise
*** reinhard has quit IRC
*** jamest has joined #gnuenterprise
<reinhard> hello jamest
<reinhard> (was that good?) ;-)
<jamest> hello reinhard
<jamest> the people in here are so friendly
<reinhard> :)
<reinhard> jamest: lekma had some problems with import tags in forms definitions
<reinhard> that's something i have absolutely no idea how it works
<jamest> i did some looking yday but don't have a single usage example
<jamest> and will have to dig into the code to figure it out again :(
<lekma> jamest: can i help?
<jamest> sure, here's a bat
<jamest> stand outside my office door and if a user gets too close......
<lekma> :)
<jamest> what is the error you are getting?
<jamest> reinhard: did you see sashas screenshots yday? 
*** dsmith has joined #gnuenterprise
* jamest thinks you'd already left
<lekma> DB000:   File "C:\Python23\lib\site-packages\gnue\common\definitions\GParser.py", line 431, in primaryInit
<lekma> DB000:     form = self._xmlParser.loadFile(handle, self.findParentOfType(None)._app, initialize=0)
<lekma> DB000:   File "C:\Python23\lib\site-packages\gnue\forms\GFParser.py", line 68, in loadFile
<lekma> DB000:     initialize, attributes={"_instance": instance,
<lekma> DB000: AttributeError: 'NoneType' object has no attribute '_parameters'
<lekma> when i try to import a simple datasource definition
<lekma> which works perfectly well in a form
<jamest> is your import library a complete form?
<jamest> i seem to recall needed that at one time
<lekma> nope only a datasource...
<lekma> you mean i must have: <form/> <logic/>...
<jamest> could you try enclosing that in a <form></form>
<jamest> i think so
<jamest> it's been ages, in fact it was pre <logic> gfd format last time I messed w/ this stuff 
<reinhard> jamest: no, didn't see them
<reinhard> and today his server is down
<jamest> :(
<SachaS> shall I send them (in one tar file) to you reinhard?
<jamest> ah, heck, he's here
<jamest> that !@#!@# capital S always gets me
<reinhard> SachaS: you all made me so curious :)
<reinhard> yes please :)
<SachaS> thanks :)
<jamest> SachaS: cc them to me too please
<jamest> if you don't mind
<SachaS> jamest: can i send it to oyu here ? ddccsend?
<jamest> i don't think that'll work
<jamest> let me change irc clients and we can try
<jamest> brb
*** jamest has quit IRC
*** jamest has joined #gnuenterprise
<SachaS> ok. files flying.
<jamest> thanks
<reinhard> SachaS: 
<reinhard> thanks
<lekma> jamest: doesn't work with <form> enclosure
<SachaS> welcome welcome ;)
<lekma> jamest: same error
<jamest> lekma: ok, can you send me the sample forms, i'll try and take a quick look
<jamest> as I don't recall any of this stuff :)
<jamest> SachaS: care if I pop those up on gnuenterprise.org in my personal dir
<jamest> case someone else wants a look
<lekma> jamest: ok, in a min
<SachaS> jamest go for it
<jamest> cool
<jamest> http://www.gnuenterprise.org/~jamest/misc/sashas/
<jamest> perms are now open :)
<SachaS> can you change from sashas to sachas ?
<SachaS> please :)
<jamest> gach!
<jamest> yet
<jamest> yes
<SachaS> hey, if I have an account on ash, i might also have a pub web directory ....
<jamest> you do
<SachaS> so i could put up my stuff there as well...
<jamest> yes
<SachaS> ok not today :)
<SachaS> cool.
<jamest> i can move it there now in fact if you like
<SachaS> no no keep it.
<SachaS> thats ok.
<jamest> ok http://www.gnuenterprise.org/~jamest/misc/sachas/
<jamest> since SachaS didn't want a gender change
<SachaS> haha
<jamest> one of my user is sasha
<jamest> so the fingers are trained
<SachaS> in australia most sacha's are girls, that was sort of funny :)
<jamest> you're a he though right?
<SachaS> right
<jamest> ok
<jamest> reinhard: so where are we wrt date/time handling in common
*** dneighbo has joined #gnuenterprise
<reinhard> jamest: jcater proposed to choose the variant
<reinhard> forms/reports should tell what data they expect, common should try to convert into that format
<reinhard> as opposed to
<reinhard> common should try to figure out what the heck is coming over the wire from the database
<reinhard> and actually i agreed
<reinhard> and it seems like that would break our dtasources API
<reinhard> (if we ever had a clearly defined one :))
<jamest> lol
<jamest> i guess i still dont see how this will get us around the issue of the different dbsig drivers returning bad representations of time
<jamest> (the rounding)
<holycow> james, thx for the screen caps
<holycow> those look m arkedly different from the packages in debian now
<jamest> would our drivers, based upon requested type, try to format the values as strings then convert to requested type
<jamest> holycow: that's all SachaS's work, i just put a copy up since his server is down
<jamest> holycow: i think he said 2 weeks effort on his part
<holycow> *nod* looks lovely none the less
<holycow> yeah? neat :)
<reinhard> jamest: re rounding i agree, but i'm not sure if we *need* time values with fractions of a second for business applications
<reinhard> i believe it's not possible to have them portably over different databases
<jamest> reinhard: the problem comes in the update
<reinhard> jamest: re type conversion
<reinhard> i don't think we would always take the way via strings
<reinhard> more like
<jamest> as we basically do a update foo set blah, blah, blah where field1=field1OrigValue, field2 = f2ov, field3 =f3ov, etc, etc
<reinhard> if requested_type == 'date':
<reinhard>   if isinstance (value, mx.DateTime.DateTime)
<reinhard>     return value
<jamest> rounding errors mean the update will fail every time
*** holycow has left #gnuenterprise
<reinhard>   elif isinstance (value, StringType)
<reinhard>     try
*** holycow has joined #gnuenterprise
<jamest> reinhard: makes sense
<reinhard>       return mx.DAteTime.strparse (value)
<reinhard>     except:
<jamest> what I was trying to do was not loose precision
<reinhard>     raise InvalidTypeError
<reinhard> etc.
<jamest> as without an exact match we can't do updates
<reinhard> yes
<reinhard> well
<reinhard> the question is
<reinhard> how do those fractions of a second *get* into the database first?
<reinhard> other than with a default = now ()
<jamest> that's how 
<jamest> but all my legacy apps do that
<jamest> and expect that
<jamest> and I can see places where you may want to have data w/ high level of precision
<jamest> if we simply say "we don't do fractional seconds" then we become useless for those applications
<jamest> i'm thinking scientific experiments with output from instrumentation
<reinhard> i think then you are bound to psycopg for postgres
<reinhard> maybe other drivers have parameters where you can set the format for the string they generate
<jamest> doesn't seem to be part of the std api 
<reinhard> correct
<reinhard> IMHO DBSIG2 is ill defined
<jamest> agreed
<reinhard> leaves much too much open
<holycow> okay so i've tried setting up gnue from debian packages but i'm told to upgrade to the latest packages
<holycow> should i download latest nightly from the website ... or?
<lekma> jamest: here are my forms
<holycow> not sure what svn is in other words :)
<reinhard> jamest: the other possibility would be to store dates as strings - weee
<jamest> holycow: svn is a cvs replacement
<jamest> reinhard: no fscking way
<jamest> that java app in inherited did that
* jamest shudders
<reinhard> jamest: even another idea
<reinhard> does LIKE work on date values?
<reinhard> we could compare dates with LIKE instead of with =
<reinhard> if we know that we have a rounding driver
<jamest> i was thinking some dbs don't do like
*** derek has quit IRC
<jamest> lekma: thanks
<lekma> jamest: you're more than welcome :)
*** holycow has quit IRC
<lekma> jamest: oh and designer doesn't seem to like import tags as well...
<jamest> :(
<SachaS> lekma, dont get frustrated with the import tag if it is not working. you might solve it temporarly in another way....
<lekma> SachaS: i know but it will be a pain in the a..
<SachaS> i know i know
<SachaS> uploaded new appserver dev. manual. includes: a) procedures with parameters, b) how to call procedures, and c) how to create a master/detail form (heavily screenshots based)
<SachaS> its a 1.3 MB pdf file. dah?
<jamest> cool!
<SachaS> its a ~400K openoffice file
*** lekma- has joined #gnuenterprise
*** lekma has quit IRC
*** lekma- has quit IRC
*** lekma has joined #gnuenterprise
<jamest> lekma: ok, installing
<chillywilly> blah
<chillywilly> anyone know anything about EDI?
<reinhard> sure
<chillywilly> what are some hipper alternatives? SOAP? XMLRPC ;P
<reinhard> whaddaya wanna know?
<reinhard> EDI is asynchronous
<reinhard> well
<reinhard> at least if you're talking about EDIFACT
<chillywilly> there's a mailbox type thing going on right?
<reinhard> well
<chillywilly> at the "VAN"
<reinhard> actually, EDIFACT only defines a file structure
<reinhard> and nothing about the transport of the file
<chillywilly> called 3 places only got through to one place to find out that they aren't a van but they do cdocument translation
<chillywilly> then he tells me no business's documents are the same
<reinhard> but it's quite usual to use some weird proprietary mailbox systems with edifact files
<chillywilly> that sorta defeats the purpose of the standard then eh?
<chillywilly> hmmm
<reinhard> well
<reinhard> for example, for invoices
<reinhard> edifact defines like some hundred fields
<reinhard> of which about 10 are mandatory
<reinhard> and the rest is optional :)
<chillywilly> what is the format of the files?
<reinhard> it's a 3-level record structure
<chillywilly> I dig some googling before but I can't find any real information
<chillywilly> just generic crap about what EDI is
<reinhard> you have segments that start with a segment identifier (3 char) and end with a single quote
<chillywilly> so it is text?
<reinhard> every segment is divided into field groups
<reinhard> yes it is text
<chillywilly> bleh
<reinhard> field groups are divided by +
<chillywilly> are they going to XML now?
<reinhard> every field group is divided into fields by :
<reinhard> no
<reinhard> edifact is 30 years older than xml
<chillywilly> why do I see them mention it on the EDI group's web site?
<chillywilly> well anyway
<reinhard> and IMHO will be there when xml has been gone again
<chillywilly> go on ;)
<chillywilly> I am listening :)
<reinhard> :)
<reinhard> that was actually the basics
<chillywilly> can I find some docs on edifact?
<jamest> lekma: :(((
<chillywilly> ok
<reinhard> every message is composed of segments
<jamest>     self._app = self._instance #TODO: Imports are broken do to
<jamest>                                #TODO: switch from _app to _instance
<jamest> i think this may be my fault
<jamest> er, dereks, i mean dereks fault
<lekma> jamest: :)
<jamest> whew, that was a close one
<lekma> jamest: so i have to find another way to do what i want...
<jamest> right now yes
<jamest> as i can't imagine it's a quick fix
<jamest> else we wouldn't have put in the todo 
<jamest> i'll make a note to dig a bit more tonight and tomorrow 
<lekma> jamest: thanks
<jamest> np, sorry for the trouble
*** kingpin2k has quit IRC
*** wendall911 has joined #gnuenterprise
<lekma> SachaS: would you care to share the navigator code of the screeshots at http://www.gnuenterprise.org/~jamest/misc/sachas/ ??
<SachaS> can receive dccsend?
<lekma> SachaS: i think so
<SachaS> ok lets try
<lekma> can u retry my accept list was wrong
<SachaS> ok. try again.
<SachaS> ok.
<SachaS> lekma at one point soon all source will be available
<lekma> thank you very much
<jamest> i'm about done setting that up
<SachaS> ok
<jamest> done
<jamest> reinhard: i moved gnue-contrib to it's own unix group so that access to one doesn't mean access to both
*** lxf has joined #gnuenterprise
<jamest> it works for me (tm)
<jamest> do we want/need a commit email from gnue-contrib?
*** jcater has joined #gnuenterprise
*** lekma has left #gnuenterprise
<reinhard> jamest: i don't think we want commit mails
<reinhard> oh
<reinhard> april 1 is heading
<jamest> i meant as a seperate mail list
*** jcater has quit IRC
<reinhard> eeeek
<reinhard> weeeeee
* reinhard runs and screams
<reinhard> appserver is running under windows!!!
<reinhard> and it works!
<reinhard> that's not right
<wendall911> lol
<reinhard> we must do something against that!
<reinhard> would anybody agree that it's about time for a release?
<jamest> yes
<jamest> i'd like to try and knock a few more bugs out that have been reported though
<jamest> but nothing major
<jamest> i don't think
<reinhard> would you think we could risk using the new plugin loader for datasources?
<reinhard> it seems to work nicely for rpc
<jamest> is it worth it for the release?
<reinhard> i think so
<jamest> or does it gain anything vs releasing then flipping over in svn 
<reinhard> it would give much better diagnostics in case of driver load failure
<reinhard> so if ppl use release and have problems with a driver dependency or something like that
<reinhard> they would get probably better error messages
<jamest> are the drivers already converted or is this major undertaking?
<reinhard> there is close to nothing to change in the drivers
<reinhard> all you have to do is exchange the code that *loads* the drivers
<reinhard> the only change necessary in the drivers is actually not importing the dbsig2 module delayed
<reinhard> but immediately
<reinhard> so that import of the dbdriver fails if not all dependencies are available
<jamest> didn't we do that for a reason?
<jamest> i seem to recall needed drivers to load to provide info to designer(?)
<reinhard> we did that because the old plugin loader couldn't handle failed imports properly
<jamest> or was it docs
<reinhard> hmmm
<reinhard> maybe it was the docs
<jamest> i thought we had something that loaded them all and did something
<jamest> but my memory sucks
*** lekma has joined #gnuenterprise
<reinhard> i think we'll ask jcater
*** dimas has quit IRC
*** lekma has quit IRC
<jamest> also
<jamest> what's the state of forms on win32?
<jamest> are things still seriously broken
<reinhard> nope
<reinhard> seems to work
<reinhard> just tested today
<jamest> great!
<jamest> my wife doesn't like me on her windows box
<jamest> i break too much stuff :)
<jamest> is this error still there?
<jamest> I can't use non ascii characters at all in triggers code, even in
<jamest> comments. I've remember a debug message saying somthing about locale,
<jamest> and looked for the reason. It's the debug line, at GBaseApp that says:
<jamest> 'Unable to load locale information falling back...'
<jamest> btw, i'm cool with a release in case i forgot to mention it
<reinhard> jamest: IIRC this line was commented out months ago
<reinhard> that debug message
<jamest> hmmm, i thought that bug report was against 0.5.3
*** lxf has quit IRC
<reinhard> jamest: ok, we can keep the delayed loading and still use new plugin loader
<jamest> ok
<jamest> i'm not against losing that delayed loading
<jamest> just against breaking something :)
<reinhard> jamest: i figured jcater can't do his dbdriver info page without delayed loading
<jamest> that sounds right
<reinhard> anyway
<reinhard> i tested non-ascii characters in triggers
<reinhard> and it works for me
<jamest> :)
<jamest> cool
<jamest> hope I'm not annoying
<reinhard> however i have of course to use utf-8 in the xml file
<jamest> i've started tracking mail threads from the ML
<reinhard> because utf-8 is the default for xml files
<jamest> and tagging some as important if they have bugs
<reinhard> regardless of what locale you have
<reinhard> maybe that was his problem
<jamest> i18n and localization are completely forgien concepts to me
<jamest> I'm (l)american
<reinhard> :)
<jamest> so I get really confused about what is doing what
<jamest> locale sets things like curency formats right?
<jamest> and date
<reinhard> locale sets everything that's different per country
<reinhard> date format, currency format, number format, language, character set
<reinhard> are the most important
<jamest> and utf-8?
<jamest> that's a i18n font encoding?
<reinhard> utf-8 is a special charset
<reinhard> not sure what you know about charsets
<reinhard> ascii defines 128 characters
<jamest> isn't unicode a mapping of several charactersets
<reinhard> so other countries than the us took the remaining 128 codes to map to what they need
<reinhard> different countries have different charsets
<reinhard> so there is things like latin-1, latin-2, ibm CP 850 etc.
<jamest> win-1251
<jamest> etc
<jamest> right?
<reinhard> people figured that they should start to use more than a single byte per character
<reinhard> jamest: right
* jamest has to support lots and lots of different character sets hear acutally
<jamest> in TeX and mail and web
<reinhard> so people invented unicode
<reinhard> which is a 4-byte character set
<reinhard> that means every character takes 4 bytes
<reinhard> and the have defined codes for quite every character that exists in any living language and in most dead languages we know about
<reinhard> like egyptic hyroghlyphes (sp?) etc.
<reinhard> now using a 4-byte character set for files, screen output, keyboard input etc is quite inefficient
<reinhard> and notable breaks compatibility with any given existing system
<reinhard> so people invented utf-8, which is a mapping of unicode into a variable lenght character code
<reinhard> where character codes from 0 to 127 use a single byte
<reinhard> codes from 128 to 16384 (i think) use two bytes
<reinhard> and so on
<reinhard> up to 6 bytes
<jamest> ah
<reinhard> the advantage of this utf-8 encoding is (among others) that files only containing standard ascii (< 128) characters are 100% identical with their ascii representation
<reinhard> well
<reinhard> am i boring?
<reinhard> :)
<jamest> no
<jamest> it's interesting
<reinhard> ok then i'll continue a bit
<jamest> i didn't realize utf-8 did that
<reinhard> 'till you sleep
<reinhard> ;-)
<SachaS> Zzzzz
<reinhard> usually applications use unicode (read: 4-byte character) strings internally
<reinhard> and convert to utf-8 on read from file, write to file etc
<reinhard> also most communication between processes is still with 8-bit strings to retain compatibility
<SachaS> dinner time. will read the log.
<reinhard> that's why no database sends unicode over the protocol
<reinhard> all use some encoding so that old, not unicode aware apps can still use the same protocol
<reinhard> that's why all databases have to have an encoding
<reinhard> quite some time, all of datasources used the *encoded* strings instead of unicode (multy-byte) strings internally
<reinhard> which led to problems when for example
<reinhard> the database encoding differs from the encoding of your terminal
<reinhard> that is the encoding you'd like your exception text in :)
<reinhard> now the real problems arise when you mix unicode (multibyte) and 8-bit strings
<jamest> so the encoding tells app how to map from unicode to some local characterset?
<reinhard> jamest: exactly
<reinhard> jamest: first unicode problem that appears quite often is the use of str()
<reinhard> str(x) will fail if x is a unicode (multibyte) string and contains non-ascii characters
<reinhard> because str() simply doesn't know which mapping to use
<jamest> is that the reason for the "%s" % value   replacements in dropdowns?
<reinhard> that's a kind if cheat
<reinhard> "%s" % x does
<reinhard> * leave x for 8-bit strings and return an 8-bit string
<reinhard> * leave x for unicode strings and return a unicode string
<reinhard> * make a str(x) for all other types
<jamest> interesting
<jamest> so what you're saying is don't use str()
<reinhard> so it's basically the same as str(x) with the exception that it doesn't break with unicode
<reinhard> don't use str() when the expression can contain non-ascii characters
<jamest> python strings support unicode right?
<reinhard> yes
<jamest> but their str() doesn't?
<reinhard> you can do about all string operations on unicode
<reinhard> str() is defined to always return StringType
<reinhard> and like i said above
<reinhard> they can't know which encoding you would want
<jamest> because the encoding isn't tied to locale
<reinhard> normally yes
<jamest> as you may need app x in english, app y in japanesse
<reinhard> but as i said
<reinhard> wlel
<reinhard> well
<reinhard> not like that, but rahter
<reinhard> you may have your terminal window using charset a
<reinhard> for your X font you use charset b because the glyphs look better
<jamest> (as a lamerican it almost sounds easier if you guys all just used my language of choice and thru out these extra characters ;)
<reinhard> and your database backend works with charset c
<reinhard> jamest: :)
<jamest> argh!
* jamest head just popped
<reinhard> so *within* a single app you might work with 3 encodings or even more
<reinhard> in fact i'm quite sure that siesel and dcmway have this case
<reinhard> with 3 encodings
<reinhard> where wx should use encoding a, tracebacks on the terminal window should use b
<reinhard> and their database is probably installed with utf-8 encoding
<jamest> so what is stored in the db fields?
<reinhard> that depends on the database
<jamest> unicode values?
<reinhard> IIRC postgres uses unicode internally
<jamest> ok, this helps some
<reinhard> but you have to tell them which encoding to use over the protocol
<reinhard> so you can't ask postgres to *send* you unicode
<reinhard> postgres will always encode somehow
<reinhard> like actually all other databases i know of
<reinhard> another problem that occures quite often is
<jamest> ok
<jamest> wow
<reinhard> the % operator
<jamest> i think my brain is full
<reinhard> with the % operator, unicode is dominant
<reinhard> so if any of both sides are unicode, the result will be unicode, too
<reinhard> which means that python will have to convert the non-unicode side to unicode
<reinhard> which again fails if the non-unicode side contains non-ascii characters
<jamest> based upon the defined encoding for the app?
<reinhard> nope
<reinhard> we have not a single defined encoding for the app :)
<reinhard> we can have 3 different encodings
<reinhard> :)
<jamest> thankfully it's time for lunch
<reinhard> so basically there are 2 paths to go
<reinhard> :)
<jamest> as i can't take much more :)
<reinhard> a) do not use unicode at all, use encoded strings everywhere
<reinhard> b) use unicode everywhere and decode/encode just before read/write from file and/or protocol
<reinhard> where for a), you would have to define for every single string parameter
<reinhard> *which* encoding this parameter is expected to have
<reinhard> like
<reinhard> "this parameter must be encoded in the database encoding
<reinhard> "this parameter must be encoded in the current terminal's encoding"
<reinhard> etc.
<reinhard> thus we tried to move towards b)
<reinhard> where actually one of the biggest moves happened when siesel accidentally committed a change that he didn't mean to commit yet :)
*** _florin_ has joined #gnuenterprise
<_florin_> hello
<reinhard> hi _florin_
<jamest> hello _florin_
<jamest> reinhard: sure it was
<jamest>  <reinhard> Damn jcater and jamest, why they always so wishy washy about unicode support!
<jamest>   <siesel> you need unicode?
<jamest>   <reinhard> yes, but they just don't understand!
<jamest>   <siesel> Ooops.
<jamest>   <siesel> Might want to cvs up
<jamest> :)
<_florin_> hi reinhard, jamest
<jamest> reinhard: how soon you thinking release?
*** ogger has quit IRC
*** sjc has joined #gnuenterprise
<reinhard> jamest: today?
<reinhard> j/k
<SachaS> haha
<SachaS> good entertainment, to read the logs :)
<reinhard> jamest: seriously,
<reinhard> jamest: i would think by end of this week?
<SachaS> james bond release.
<reinhard> SachaS: yes :)
<reinhard> bbl
*** ogger has joined #gnuenterprise
*** kg_kilo has joined #gnuenterprise
*** jcater has joined #gnuenterprise
<jamest> hello ogger kg_kilo jcater
<kg_kilo> hi jamest
<jamest> reinhard: i'm ok with release this week 
<jamest> i'm wondering if it's worth the efforts to try and knock out some reported issues
<jamest> and do a prerelease on the list
<jamest> we seem to be getting better feedback now
<jamest> jcater: reinhard is intested in a release
<jamest> <reinhard> jamest: i would think by end of this week?
<jcater> are we ready?
<jcater> I'm ok w/it
<SachaS> reinhard: have the masterdata_customer class. have the test_testBusinessClass. the test_testBusiness class added a new test_testProperty property to the masterdata_customer business class. From the test_testBusiness business class I can successfully access that new property of the masterdata_customer business class.
*** johannesV has quit IRC
<jamest> jcater: he was merging the new plugin system to the db drivers
<jcater> lots and lots of recent changes
<jamest> i didn't see an issue as it works with the import trick you put in
<jcater> but I would like to get these bugfixes out too, though
<jamest> i seem to recall that import trick needed for doc generation?
<jcater> yes
<jamest> ok
<jamest> i'm also wondering about all the changes, i figure we need to prerelease
<jamest> at least once
<jamest> and there are several bugs I wouldn't mind fixed
<SachaS> reinhard: a bug :)
<SachaS> reinhard: i have a test module.
<SachaS> one sec. checking again.
<chillywilly> anyone ever look at SQLObject?
<chillywilly> seems intersting
<chillywilly> doesn't give you enough control over the schema
<chillywilly> though
<SachaS> reinhard: no bug :)
<chillywilly> it would be nice to use bigserial for primary keys instead of serial
<kg_kilo> SachaS: is your server still down?
<jamest> kg_kilo: looking for the screens?
<jamest> http://www.gnuenterprise.org/~sachas/screens/
<kg_kilo> yes, exactly
<kg_kilo> thanks!
<SachaS> kg_kil no but there is a new link.
<reinhard> kg_kilo: please tell btami that forms and appserver work nicely under windows
<reinhard> jamest: I'm ok with any bug you want to fix :)
<jamest> :)
<jamest> i'll try and go thru them tonight
<reinhard> i'm not 100% certain we need prereleases at all
<SachaS> reinhard: test 2 works also. test 2 is: we have: masterdata_customer.test_testProperty . test2_testBusinessClass can access masterdata_customer.test_testProperty successfully. (access only assign and read. have not tested commit). must test with commit.
<reinhard> now that our nightly snapshots are built as release tarballs
<reinhard> however what would be nice is
<reinhard> when we defined a string freeze
<reinhard> and would give translators 2 or 3 days time to update translations
*** btami has joined #gnuenterprise
<btami> hi
<reinhard> hi btami
<reinhard> <reinhard> kg_kilo: please tell btami that forms and appserver work nicely under windows
<jamest> hi btami
<reinhard> wrote that 5 minutes ago :)
<btami> hi all
<jamest> btami: i have a windows error that was reported to webmaster@gnuenterprise.org
<btami> reinhard: thx
<jamest> you want?
<jamest> =)
<btami> sure
<jamest> where can I forward it?
*** reinhard has quit IRC
<btami> btami at freemail dot hu
<kg_kilo> btami: reinhard told that forms and appserver worked nicely under windows 8-))
<jamest> sent
<btami> k
<jamest> btami: any outstanding issues w/ a windows release?
<jamest> i think we're targeting this week
<btami> i'm ok with that
*** ogger has quit IRC
*** ogger has joined #gnuenterprise
<btami> an addition to benefits os parameters in sql execute:
<btami> http://sourceforge.net/forum/forum.php?thread_id=1041549&forum_id=30917
<btami> in kinterbasdb forum David Rushby's explanations are usually excellent 
*** reinhard has joined #gnuenterprise
<btami> reinhard: you tried my 2. setup exe or svn?
<reinhard> of course i tried your setup.exe
<reinhard> ah err
<reinhard> i still had problems with the connection form
<reinhard> nearly forgot that
<reinhard> but everything else worked
<btami> on Me?
<reinhard> yes on you
<reinhard> ;-)
<reinhard> yes on Win ME
<btami> lol
<btami> ok
<reinhard> also that appserver works means that the new plugin loader works
<btami> yes
<btami> i like it
<btami> the new setup exe has a plus advantage
<btami> it contains all unmodified (installed) gnue tree (*.py)
<btami> so peop can fix bugs
<reinhard> i can confirm that the connection dialog is broken also on gnu
<btami> or udate fixed files from svn
<reinhard> btami: yes, that's good
<reinhard> btw
<reinhard> can you defined "canned" package selections?
<reinhard> could make sense to define a "client" install
<jamest> btami: what's the new setup.exe using?
<reinhard> where you install forms, reports, desinger, but not appserver
<btami> py2exe+inno
<jamest> i didn't think mcmillian could package .py files and change 
<jamest> ah, so mcmillian is gone then?
<btami> reinhard: sure, i can
<btami> jamest: yes
<jamest> ah, cool
<btami> jamest: all files you need in gnue-common/packaging/win32
<btami> the _main_ trick was to exclude "gnue" in py2exe setup.py
<btami> and just append the . dir to syspath in all startup scripts
<btami> btw. re str() and % 
<btami> they not limited to ascii, but
<btami> want to convert with sys.getdefaultencoding()
<btami> whitch is ascii by default
<btami> but on python startup can set with setdefaultencoding()
<btami> as it was discussed before
<btami> why we cant use it
<SachaS> so is there  a bug system or is there none?
<btami> anyhow, using unicode internally in all gnue is the best
<SachaS> i have an issue with Class Repository.
<wendall911> reinhard: have you looked at SQLite at all?
<SachaS> no issue with Class Repository.
<SachaS> :)
<SachaS> i have an issue with Class Repository. :) feature request to set default values in Class Repository.
<reinhard> wendall911: not specially
<reinhard> why?
<wendall911> reinhard: looks kind of useful for quick evaluations. Doesn't require setup or configuration... Not ideal for production, but may be useful for small scale things.
<reinhard> jcater: you think you have some time to gnue today?
<reinhard> wendall911: ah
<reinhard> wendall911: was not sure what you mean
<reinhard> wendall911: we have an SQLite driver in common
<jcater> reinhard: not sure
<jcater> why?
<wendall911> reinhard: ok, cool, didn't see it.
<reinhard> wendall911: if you have it installed and have some experience with it
<reinhard> we would highly appreciate if you tested gnue with sqlite and gave us feedback
<wendall911> reinhard: might be nice to use it for an evaluation install for app server.
<reinhard> as i'm not aware of anybody using it currently
<wendall911> reinhard: ok, I'll give it a shot, just looks interesting is all
<reinhard> jcater: i've finished transition of dbdrivers to new plugin loader and could commit
<reinhard> jcater: however, this would break your website generation code
<reinhard> jcater: as the DRIVERS= lines would be removed
<reinhard> wendall911: thanks
<reinhard> jcater: but i could let the DRIVERS= in there for the time being
<reinhard> and we can remove them later once you've adapted the website code
<wendall911> reinhard: SQLite doesn't support alter tables, will that be an issue?
<reinhard> depends on what you want to do :)
<reinhard> for quick evaluation i'm pretty sure it will be no issue
<reinhard> for production work it could be hard to update an application then :)
<SachaS> reinhard: i start to see the need to extend business classes
<wendall911> reinhard: yeah, would require some extra coding for updates and such if it was used as a solution
<SachaS> i have started with the masterdata module, where i have the inventory business class
<SachaS> everything was fine, i added the property i was using in the beginning (name, description, id etc).
<SachaS> now i am working at the inventory system.
<jcater> why not leave for now
<jcater> and I'll fix asap
<jcater> (which might be tonight)
<wendall911> reinhard: I think the same issue applies to oracle 8 and partly with oracle 9 iirc
<reinhard> SachaS: isn't that wonderful how you more and more understand the reasons why i'm doing stuff the way i do?
<reinhard> if only women could do that too ;-)
<reinhard> jcater: i'll commmit and leave the DRIVERS = in there
<SachaS> now i see, that the inventory business class needs more properties of the masterdata_inventory business class, such as "reorder point" and "reorder quantity". so now i am thinking, should i add them in the masterdata module, or in the inventory module.
<reinhard> jcater: ok?
<SachaS> :)
<SachaS> haha
<SachaS> can you give your thoughts on my example?
<SachaS> its like there are 100+ different ways to do it ...
<reinhard> SachaS: I'd try to see modules as "features" that can be wanted or not wanted
<reinhard> for example, info like "reorder point" are parts of the feature "automatic ordering"
<reinhard> or however you would call it
<SachaS> yeah
<reinhard> so i would probably create a module "autoorder" and put these fields there
<SachaS> i see the point
<reinhard> as well as the logic behind that
<reinhard> because there might be companies that don't use that feature
<reinhard> for any reason
<SachaS> yeah the logic would be in the autoorder module.
<reinhard> and then these fields should not even be there
<SachaS> the thing is
<SachaS> yeah.
<SachaS> damn
<SachaS> :)
<SachaS> it would been the old inventory form is still ok.
<reinhard> which of course makes it not so easy to design a form for inventory :)
<SachaS> eg the person who adds new items does not have to know aobut reorder quantity and point
<SachaS> yeah thats the point
<reinhard> yeah
<SachaS> so a inventory or autoorder guy must have to provide these infors
<SachaS> so a inventory or autoorder guy must have to provide these infos
<SachaS> so there would be a new form.
<reinhard> that's why we must go for autogenerated forms
<SachaS> a new form in the autoorder module.
<reinhard> that are not statically defined by xml
<SachaS> autogenerated forms? a new subproject? :)
<reinhard> but dynamically created from the list of available fields
<SachaS> reinhard: one question concerning extending a class
<jcater> I don
<SachaS> i was extending a class and then there was the "nullable" option
<reinhard> but that will require much beating for the forms guys ;-)
<SachaS> i have not tested it
<reinhard> should work
<jcater> I don't see that as the only option, fwiw
<SachaS> but ... if an extended property has the not null
<jcater> just security added to forms
<SachaS> could that cause problems, as the original forms do not have those properties? in theory, it should be ok. because they still use the "orignial" class.
<reinhard> jcater: you're refering to the reorder point problem in forms?
<reinhard> SachaS: I don't see where that could cause problems
<jcater> yes
<SachaS> in which table does the new, extended property go?
<reinhard> SachaS: in the same table
<SachaS> so if there is a NOT NULL statement (in the SQL def.)
<btami>   File "/home/tamas/svn/gnue/gnue-common/.cvsdevelbase/gnue/common/datasources/drivers/DBSIG2/Connection.py", line 258, in makecursor
<btami>     (s, p) = self.getattr ('_Connection__param_' + paramstyle) (s, p)
<btami> AttributeError: Connection instance has no attribute 'getattr'
<btami> ------------------------------------------------------------
<btami> t
<reinhard> jcater: the point is that if the reorder module is not installed, the fields do not even exist in the backend
<SachaS> and the user addes a new inventory (orignial class), then he can never add those new NOT NULL properties.
<reinhard> SachaS: the user needs to have a form that also shows the newly added properties
<btami> reinhard: sorry for interrupt
<reinhard> or the initialization procedure must init the property
<SachaS> ahh but the original form ... does not avhe those field. 
<reinhard> btami: no problem
<btami> i get this when presing save
<reinhard> SachaS: that's why we need dynamically created forms
<reinhard> but i repeat myself ;-)
<SachaS> reinhard: so there is a problem ...
<reinhard> btami: latest svn up?
<btami> yes
<reinhard> ah oh
<SachaS> ok. i see how you want to have all these modules independent....
<SachaS> <btami> yes
<SachaS> dah? no comment above
<SachaS> all those modules should be independent
<reinhard> btami: just a second
<reinhard> i always thought every pthon object has a getattr () method
<SachaS> but if you have a combination of these modules ... you get into a bit more compilcated waters
<reinhard> doh
<reinhard> it's getattr (object, name)
<SachaS> as i am currently building one fat (in fact slim ;) app
<reinhard> not object.getattr (name)
<SachaS> where i am aware of all modules installed.
<SachaS> so thats the easy one.
*** kg_kilo has quit IRC
<SachaS> if i am aware of all those modules, and their classes etc, then I can simply access them etc. or change forms, so everything works together.
<reinhard> SachaS: i think you now get the point :)
<SachaS> ok. i have not started to think how to rather design modules, than an app. and all the installed modules result in the app.
<reinhard> btami: can you try again please?
<SachaS> all the modules make the app
<SachaS> not the other way around....
<SachaS> ok ok.
<SachaS> ah ;) and i remember you talking about having "module" packages (as in debian packages)
<btami> reinhard: it's OK now, thx
<reinhard> btami: great!!
<reinhard> btami: could you please test with useParameters = 1 ?
<reinhard> (in gnue.conf in [common] section)
<SachaS> reinhard: quickly. can you say if the concept of modules will implement the package (as in HR, Finance, AccountsReceivable) are
<btami> i'v used it with =1
<reinhard> btami: cool
<reinhard> SachaS: sorry, parse error with that sentence ;-)
<btami> have to sleep now, good nigt all
<SachaS> remove the trailing are
*** btami has quit IRC
<reinhard> ok
<reinhard> in that case
<reinhard> yes.
<SachaS> ok. so module = concept to implement packages
<SachaS> somehow ~
<reinhard> yes
<SachaS> ok.parse error! haha
*** nickr has quit IRC
<SachaS> ok. i think i got another thing tonight. i will do a little bit more on my fat application compared to the desired independent modules making together the app
*** nickr has joined #gnuenterprise
<SachaS> well reinhard: i bet you have your thoughts on how to realise those independent modules.
<reinhard> yes :)
<SachaS> ok. cool approach. its interensting. i think you tried to tell me that last time. and i just did not get it :)
<SachaS> i feel like throwing away my app before I even got started and to scratch my head on those "independent" modules. those lego bricks ;)
<SachaS> yeah i start to see my app as a granite block. there are modules in it, but they are pretty much interlinked with each other...eg. if a class (from one module) isnt there, it simply does not work.
<SachaS> ok. if you want to have such a "module builds the app" approach, then you need to know what you are doing. ok enough.
<SachaS> damn i am adding a property called "inventoryChecked" to the sales module. dah? :)
<SachaS> to a sales order. dah? :)
<SachaS> ok enough.
<SachaS> i just realise, i have no clue what i am doing.
<SachaS> instead of having this information "the inventory system has checked this sales order" in the sales order, it should go into a "inventory system sales order processing" table. and that table references the sales order.
<jamest> hello nickr
<nickr> hi
<chillywilly> nick nick nick na nick nick niiiick, nickelodeon
<nickr> no
<chillywilly> yes
* SachaS removed inventoryChecked from sales_salesOrder :)
<nickr> wocher
* SachaS added salesOrderTransaction business class in the inventory module.
<SachaS> hmmm inventory relies on sales order. hmmm looks also wrong.
<SachaS> but better than before
*** dsmith has quit IRC
* SachaS removes the masterdata_reorderPoint and masterdata_reorderQuantity from the masterdata_inventory biz class.
* SachaS makes the masterdata_inventory(should be item).inventory_reorderPoint
<reinhard> SachaS: there is a very basic programming rule
<reinhard> don't write a program while you are still confused
<reinhard> ;-)
*** ogger has quit IRC
<reinhard> however
<reinhard> think i broke enough for today
<reinhard> night all
*** reinhard has quit IRC
<SachaS> :)
<SachaS> how about learning by doing?
<jamest> shit
<jamest> looks like I'm going to have to resign
<SachaS> what jamest?
<jamest> <reinhard> don't write a program while you are still confused
<jamest> no more gnue for me
<SachaS> :)
<jamest> bbl
*** jamest has quit IRC
*** nenhum has joined #gnuenterprise
*** sjc has left #gnuenterprise
*** jcater has quit IRC
*** jcater has joined #gnuenterprise
*** nenhum has left #gnuenterprise
*** _florin_ has quit IRC
* SachaS likes how stable appserver is (when adding, removing, changing business classes, modules, procedures etc)
*** jamest has joined #gnuenterprise
*** wendall911 has quit IRC
<SachaS> jamest: in a form, there is commit(), why is there no rollback() ?
<jamest> because no one asked for it :)
<jamest> that's basicallly a clear()
<jamest> something is probably named badly :)
<SachaS> ok. but there is no clear() on the level of commit() ... i can call commit() but have to call block.clear() ?
<SachaS> is that right?
<jamest>  [01:44:30 am] <jamest>because no one asked for it :)
<jamest> i'm sure we can add it :)
<jamest> let me finish up what I'm doing and I'll take a look
<SachaS> :) do you think thats of use?
<jamest> not really or else someone would ask for it
<jamest> oh wait
<jamest> you *are* asking for it
<SachaS> <reinhard> SachaS: there is a very basic programming rule
<SachaS> <reinhard> don't write a program while you are still confused
<SachaS> but there is also
<SachaS> "but they dont know what they are doing"
<jamest> please read our web site
<jamest> no where do we claim we know what we are doing
<SachaS> i think a rollback() might be useful, as useful as a commit()
<jamest> :)
<SachaS> :)
<jcater> but then there'd be no gnue
<SachaS> i think i would like business class instance procedures. Then again i am confused what bound procedures are.
<jamest> gnue-import?
<jamest> does that work?
<SachaS> i am pretty sure a bound procedure is not a business class instance procedure.
<jcater> what is that?
<SachaS> a bound procedure?
<jamest> NAME
<jamest>        gnue-import - GNUe Schema Importer
<jamest> SYNOPSIS
<jamest>        gnue-import [ options ] file
<jamest> DESCRIPTION
<jamest>        GNUe  Schema  Importer  reads data from a GNUe Schema Definition and imports it
<jamest>        into a connection.
<SachaS> ah you are talking about gnue-import
<jamest> SachaS: i no nothing about appserver yet
<jamest> er know
<jamest> jcater: gimcvs
<jcater> um
<jcater> if it does work
<jcater> it's poorly named imho
<jcater> but that was gonna be a part of gnue-schema
<jcater> I dunno if it works or not though
* SachaS makes a feature request: a rollback() on the same level as commit() :)
* jamest thwaps SachaS
<jamest>  [01:45:58 am] <jamest>let me finish up what I'm doing and I'll take a look
<jamest> so give me a minute! ;)
<jamest> i'm docing
<jamest> and I just *love* docing
<jcater> jamest: can you finish fixing designer?
<SachaS> docing? writing documentation?
<jamest> bring docs up to date
<jamest> making a script to make it a tiny bit easier 
<SachaS> ok. thanks anyway.
<jamest> SachaS: it's minor change I bet
<jamest> i just want to finish this up
<jamest> give me 10 minutes
<SachaS> great.
<SachaS> its 4am here
<jamest> gack!
<SachaS> i think i have 10 minutes :)
<jcater> geeze
<jcater> late night?
<SachaS> a couple of late nights in a row ....
<jamest> gnue is just so good
*** jbailey has joined #gnuenterprise
<jamest> he can't put it own
<jamest> s/own/down
*** jbailey has quit IRC
<jcater> it's like crack!
<jcater> only more dangerous
<jamest> why do we have
<jamest> gdes and gfdes 
* jcater wants to phase out gfdes
<jamest> sigh
<jcater> (but that was 2 yrs ago, and don't think ppl are using gdes yet :)
<jamest> blocks have a processClear and a processRollback
<jamest> both do a rollback on the datasource
* SachaS uses only gdes
* jamest thwaps SachaS again
<SachaS> dah?
<jamest> don't encourage him!
<jamest> :)
<jamest> jcater: remove gfdes
<jamest> that'll motivate us
* jcater thwaps jamest
<jamest> I'm going to make it rollback() instead of clear at the form level
<jcater> to me
<jcater> clear implies clearing out current record
<jcater> not reversing all records in memory
<jcater> so I agree
<jamest> yeah
<jamest> not sure about clear in block
<jcater> hmm
<jamest> iirc we say clear or something in the menu
<SachaS> appserver can create business objects in memory. they only get stored to the db, when a commit occurs.
<jcater> block-level commit may be dangerous
<jcater> really, it should be either form-level (for all connections)
<jamest> iirc we added that for the papo team
<jcater> or connection.commit()
<jamest> SachaS: you running svn right?
<SachaS> jamest: yes svn.
<jamest> ok, you want to test this for me
<jamest> or do i need to make a test form
<SachaS> jamest: yes svn. i can do a test.
<SachaS> a rollback test?
<jamest> yes
<jamest> it's global so
<jamest> rollback()
<jamest> will work
<SachaS> ok.
<jamest> DB_RUNRECOVERY: Fatal error, run database recovery
<jamest> rock!
<jamest> ok, commited
<jamest> if anything you may get ui corruption
<jamest> but at 4am i'm not sure you'll notice
<SachaS> gnaraloo:/home/sacha/gnuenterprise# svn up gnue
<SachaS> svn: Berkeley DB error while opening environment for filesystem /var/svn/gnue/db:
<SachaS> DB_RUNRECOVERY: Fatal error, run database recovery
<jamest> so I should be in the clean
<SachaS> ok. got it.
<jamest> would someone shoot the svn server for me
<jamest> i'm starting to wonder if setting up apache2 for svn would be a good thing
<jcater> me too
<jcater> hmm
<jcater> the more I think about it
<jcater> the more I think you should
<jcater> you were volunteering, right?
<jcater> :)
<jamest> um
<jamest> sure
<jamest> why not
<jamest> not like I have anything else to do
<jcater> just think of it as Grand Theft Auto: Fscking Ash Up
<dneighbo> rofl
<jamest> dneighbo: i've taken care of the dtd
<jamest> those with small children or delicate consitution please shut off your screens
<jamest> FUCKING SVN!!!!!!!!!
<jamest> ok, turn them back on now
<jamest> DB_RUNRECOVERY: Fatal error, run database recovery
<SachaS> how does a sql update look like?
<SachaS> update () db where (x = y)
<SachaS> how do i set the field and the value?
<SachaS> it has been ... ages. :)
<jcater> update jamest set mood="calmer";
<SachaS> thanks. yep
<jamest> ahhhhhh
<jamest> jcater: could you commit that please
<jcater> hang on
<jcater> have a few more to do 
<jamest> wiser, smarter, sexier
<jamest> course those aren't moods
<jcater> insert into todo (person, task) values ('jamest','svn update');
<jcater> insert into todo (person, task) values ('jamest','designer crashing');
<jcater> insert into todo (person, task) values ('jamest','web-po');
<jcater> insert into todo (person, task) values ('jamest','gnue-contrib emails');
<jcater> insert into todo (person, task) values ('jamest','gnue-sb');
<jcater> insert into todo (person, task) values ('jamest','adrius');
<jcater> insert into todo (person, task) values ('jamest','ajrs');
<jcater> insert into todo (person, task) values ('jamest','math');
<jcater> oh, yeah
<jcater> commit;
<jamest> argh!!!!!!!!!!!!!!!!!!!!!
<jcater> @weather 38654
<bigbrother> jcater: The current temperature in Olive Branch, Mississippi is 50°F. Conditions: Clear.
<SachaS> jamest: as far as I can tell is rollback working.
<jamest> the ui ok?
<SachaS> ui?
<SachaS> do you want to know if the icons are set accordingly?
<jamest> fields all clearing properly
<jamest> in master/detail forms
<SachaS> i am running commit() / rollback() so appserver stores/clears allocated business objeccts.
<SachaS> have not checked in the form ...
<SachaS> hmmm
<SachaS> ok. will test with another form, where i have master/detail in the form.
<chillywilly> jcater: you forgot begin;
<jcater> ah
<jcater> damn postgres
<jcater> begin is implied in oracle
<jcater> (btw, this is one of my pet-peaves with psql)
<chillywilly> bleh
<jcater> (if I'm in interactive mode, it should damn-well assume I may want to rollback)
<jamest> jamest@jamest$ killall postmaster
<jamest> w00t!  I'm saved!
<SachaS> ok. just as a remark: rollback() is not in the namespace of a buiness object procedure.
<jcater> ?
<SachaS> jamest: as far as i can tell is rollback working in forms. all fields get cleared, and the save icon gets shaded again.
<jamest> wow
<jamest> that cant be right
<jamest> it worked
<SachaS> ok. good night.
<jamest> thanks
<jamest> for testing
<jamest> nite
*** SachaS has left #gnuenterprise
*** mdean has quit IRC
*** holycow has joined #gnuenterprise
<jamest> hi holycow
<holycow> hey :)
<holycow> jamest, dude, boy do i need to get cracking on gnue
<holycow> in need to do lists and project tracking that actually works!
<holycow> heh
<holycow> getting to be hard to track everything
*** jamest has quit IRC
<holycow> okay, so i executed the subversion code above... what next?
<holycow> i realize they are all python scripts... did that place everything in the right places or are there other steps for proper installation?
<jcater> what did you do?
<holycow> well uninstalled the gnome packages and ran the svn command from topic
<holycow> it looked like it downloaded everything... i presume it must be in a subversion folder somewhere now, correct?
<jcater> yes
<jcater> should be a gnue/ directory
<jcater> cd gnue/gnue-common; ./setup-cvs.py --auto
<jcater> will set up symlinks so you can 
<jcater> (doesn't need root)
<jcater> that creates a script in ~/bin/ for all the gnue tools
<jcater> to run them directly from the subversion checkout that you did
<jcater> (i.e., doesn't require you to properly install them all)
<holycow> i can't fine any documentation on the svn tigris website
<holycow> :/ sorry for the stupid q's i'm a bit lost as this is so new to me on a lot of levels
<jcater> http://svnbook.red-bean.com/
<holycow> ah!
<holycow> danke :)
<chillywilly> anyone use optimistic locking much :)?
<chillywilly> seems like a nifty strategy for concurrency
<Vee2d2> red-bean rocks
<holycow> heh yes
<holycow> extremely helpfull
<holycow> ah, gnue now supports 2 tier and n tier
<holycow> neat
<jcater> ?
<holycow> i always thought of it as 3 tier, always requiring gnue-appserver
<jcater> ah, no, not at all
*** mdean has joined #gnuenterprise
*** jcater has quit IRC