01:04 <knowtheory> hmm, Zamous was having fail

01:04 <knowtheory> and blamed dm :(

01:09 <dkubb> yeah, that's too bad

01:10 <dkubb> I would've helped him if he didn't split

01:10 <knowtheory> hey dkubb

01:10 <knowtheory> yeah

01:10 <dkubb> knowtheory: hey

01:16 <dkubb> I'm not entirely sure what his problem was, but I would've said that uniqueness has nothing to do with the intermediaries, but rather the target objects

01:16 <dkubb> the result of most collection operations are a unique set of resources

01:17 <dkubb> by default anyway, or it should be

01:17 <knowtheory> i'm guessing it had to do w/ chaining stuff probably

01:17 <knowtheory> instead of doing User.all.map{|u|u.items}

01:18 <knowtheory> he was doing users = User.all; users.items

01:18 <knowtheory> or something

01:18 <dkubb> User.all.items works fine

01:18 <dkubb> it would return a unique list of Item resources associated with at least one User

01:19 <knowtheory> right

01:19 <knowtheory> he was unhappy about having a uniqed list

01:19 <dkubb> possibly

01:19 <knowtheory> yeah :)

01:19 <dkubb> User.all.items(:unique => false)

01:19 <knowtheory> hard to say w/o more feedback

01:19 <dkubb> ya

01:22 <knowtheory> yay my nexus one has shipped

01:34 <dkubb> knowtheory: you thinking of doing dev for it?

01:35 <knowtheory> thought about it, not seriously though. i need to get into more java dev in general i guess

01:35 <knowtheory> or maybe using V8 to do JS dev for it

01:37 <knowtheory> i'm joining the smartphone age :P

01:37 <knowtheory> i haven't had one up to now

01:38 <knowtheory> and wanted to get an unlocked phone

01:38 <dkubb> I have been cell-phone free for a long time

01:38 <knowtheory> otherwise i probably would have gotten an iphone

01:38 <knowtheory> oh yeah? :)

01:38 <knowtheory> i wanted a mobile computing devices that wasn't my laptop

01:38 <dkubb> I'm plugged in all day long, when I am away from the computer I don't want to be plugged in then

01:38 <knowtheory> because as much as i love my MBP hauling it out all the time isn't exactly the height of convenience, nor does it have mobile connectivity

01:38 <knowtheory> yeah, i know how that feels

03:26 <myabc> Anyone using Rails 3 + DM + multiple repositories?

07:51 <knowtheory> myabc: nope, Merb + DM + multirepo here

07:54 <myabc> knowtheory: hey, how's it going?

07:54 <myabc> knowtheory: do you have Merb::Plugins.config[:merb_datamapper][:use_repository_block] set to true?

07:54 <knowtheory> hey hey

07:54 <myabc> knowtheory: (default is true, AFAIK)

07:55 <knowtheory> yeah, for merb i think the default is true, i'm just verifying to be sure

07:55 <knowtheory> we do use multiple repository blocks. we've actually got it wrapped up in an object for convenience

07:55 <knowtheory> yes, it's set to true

07:55 <knowtheory> why out of curiosity?

07:59 <myabc> knowtheory: well I partly resolved - or at least understand my problem now.. I was wondering why Model#default_repository_name was being observed in console, but not in controllers.

07:59 <myabc> knowtheory: (this is the first time I've used this feature outside of a small Sinatra /CLI apps)

07:59 <knowtheory> interesting

07:59 <myabc> knowtheory: But I had missed the fact that Rails 3 (as Merb), Controller actions are wrapped in a repository block by default.

08:00 <knowtheory> ah yeah.

08:00 <knowtheory> the model we decided upon

08:00 <knowtheory> was to always make the block one is operating in explicit

08:01 <knowtheory> so our Application class requires knowing which repository to use, or it'll throw an exception

08:02 <myabc> knowtheory: Oh, interesting.. that would make sense.

08:02 <isheridan> Is anyone on that is familiar with over-riding accessors in a model class?

08:02 <knowtheory> isheridan: like... overriding properties?

08:02 <isheridan> yes

08:02 <knowtheory> myabc: yeah we have a default repository just to make sure the app boots up, but otherwise, we don't use it.

08:02 <knowtheory> isheridan: what are you overriding properties for?

08:03 <knowtheory> my first question would be, should you rename your property, and create a method/accessor to point to the actual property you want to store stuff in?

08:03 <isheridan> I have two properties that are Strings but I need have the output of the property transformed to an Array.

08:03 <knowtheory> and my second question is, are you just trying to add behavior that could be accomplished using a before, or after block?

08:03 <knowtheory> ah

08:04 <knowtheory> yeh in the event that i've had to do that i tend to have gone for the first option there

08:04 <isheridan> knowtheory: no I am defining methods... here is a gist ... https://gist.github.com/0b74bcceedbb7ce47bec

08:07 <isheridan> knowtheory: When I initiate an object and populate it... all seems good. but when I try to save the new record "nothing" the save returns false but I can soo not error messages or anything that would enlighten me as to what I am doing wrong.

08:08 <isheridan> knowtheory: see no errors that is

08:08 <knowtheory> calling errors on the object doesn't return anything?

08:08 <knowtheory> you can check the attributes hash

08:08 <isheridan> one sec

08:08 <knowtheory> and the dirty_attributes hash to see whether DM is aware that you've changed things

08:09 <namelessjon> attribute_set takes care of dirtyness.

08:09 <knowtheory> mmhmm

08:10 <isheridan> hrmmm this is the error that I get ... :statuses=>["Statuses must be of type String"], :categories=>["Categories must be of type String"]

08:10 <isheridan> yet

08:11 <isheridan> when I look at the object I see this ... #<Filter @id=nil @name="Newly Imported" @statuses="1" @categories="" @start_date=nil @end_date=nil @is_default=true @is_temp=false @created_at=nil @updated_at=nil @user_id=1>

08:11 <knowtheory> what happens when you call the "cateogries" method on the object?

08:12 <knowtheory> it returns [""]?

08:12 <knowtheory> I don't know enough about DM's save chain

08:12 <knowtheory> do you namelessjon?

08:12 <isheridan> yes but that is what I want

08:12 <knowtheory> actually... call "attributes" on the filter

08:12 <knowtheory> if you could

08:12 <isheridan> well it returns []

08:12 <knowtheory> and paste what the hash looks like

08:13 <isheridan> {:categories=>[], :statuses=>["1"], :start_date=>nil, :user_id=>1, :created_at=>nil, :end_date=>nil, :is_default=>true, :is_temp=>false, :name=>"Newly Imported", :updated_at=>nil}

08:13 <isheridan> ugh

08:13 <knowtheory> hehe :) interesting

08:14 <namelessjon> Hmmm. the assignments should definitely be set when the method is called, not (i.e. before save even becomes involved)

08:14 <isheridan> this is why I made sure to use "attribute_set"

08:16 <* namelessjon> nods.

08:18 <isheridan> While it is not the big a deal to work around this... I do not think that this was the intended behavior of "attribute_set" or the over-ride feature.

08:22 <namelessjon> isheridan: It's definitely being called (the method)

08:24 <isheridan> namelessjon: ah you tested it? yeah I thought so too.

08:26 <isheridan> namelessjon: maybe the returned value should be the resulting string ... but what would "attribute_set" usefulness be? handling the dirty stuff?

08:26 <namelessjon> Oh. Of course.

08:26 <isheridan> ?

08:26 <namelessjon> I figured it out, I think.

08:27 <namelessjon> The validator calls your reader method.

08:27 <isheridan> <sigh>

08:27 <namelessjon> When its checking its a string.

08:29 <isheridan> namelessjon: so this means over-riding is useless for me and I should do the data transformation outside of the model object. hrmmm

08:31 <namelessjon> isheridan: You can turn off auto_validation for that field.

08:31 <isheridan> duh /smack

08:31 <namelessjon> isheridan: That removes the primitive validator, and dm handles the saving as a string properly

08:35 <isheridan> so basically in this circumstance one should set ":auto_validation => false" and make sure that you do your own validation in the over-riding method.

08:36 <knowtheory> you can also add your on validator

08:36 <knowtheory> which does what you want it to do

08:36 <knowtheory> rather than what it thinks is the right thing to do of course

08:36 <isheridan> yes of course

08:36 <isheridan> good point

08:37 <isheridan> Thank you both for your help

08:38 <isheridan> I was taring my hair out trying to figure this out.

08:47 <andreich> hello, please could someone point me to :dependent => :destroy in associations? I've searched the documentation but I haven't found where is this mentioned. The ticket on this is marked as resolved

08:48 <andreich> I tried also with dm-constraints, but I get "condition :dependent does not map to a property or relationship" no matter what I include

08:49 <knowtheory> sorry i do not know :(

08:52 <namelessjon> andreich: :constraint => :destroy

08:52 <andreich> just found it

08:52 <andreich> thank you

11:34 <github> datamapper.github.com: master Alex Coles * da9b532 (1 files in 1 dirs): Work-around a Maruku Markdown-parsing warning ... - http://bit.ly/bK259z

11:37 <github> datamapper.github.com: master Alex Coles * b3b55fb (1 files in 1 dirs): Fix typo in Properties documentation ... - http://bit.ly/9e3qsi

13:59 <xshay> hey I'm trying to get a complex aggregate function through DM

13:59 <xshay> SELECT SUM(LEAST(20, value * 0.2)) FROM X WHERE Y

13:59 <xshay> It doesn't seem this is possible?

14:24 <pjwal> Is it possible to create a many to many relationship with a composite key for the many to many table?

15:34 <Tass`> xshay, :value.least => 100 # probably

16:13 <g0bl1n> hi, in OSX (Snow Leopard) I installed PostgreSQL EnterpiseDB package then: gem install do_postgres

16:14 <g0bl1n> It installs fine

16:14 <g0bl1n> but

16:14 <g0bl1n> when trying to connect to a local or remote Postgres server, I get:

16:14 <g0bl1n> dyld: lazy symbol binding failed: Symbol not found: _PQsetdbLogin

16:15 <g0bl1n> Referenced from: /Library/Ruby/Gems/1.8/gems/do_postgres-0.10.1/lib/do_postgres/do_postgres.bundle

16:15 <g0bl1n> Expected in: flat namespace

16:15 <g0bl1n> dyld: Symbol not found: _PQsetdbLogin

16:15 <g0bl1n> any idea ?

17:43 <xshay> Tass`: LEAST isn't an aggregation function, but my point was more "I have this random SQL function and want to push it through DM"

17:43 <xshay> Tass`: would be nice to get all the query building and conditions and scope and stuff, and only drop down to raw SQL for the function

20:21 <chuck> I can never get DataMapper.setup and DataMapper.auto_migrate! to work :(

20:21 <chuck> DataMapper.setup(:default, "sqlite3://#{File.join(File.dirname(__FILE__), 'talker.db')}")

20:21 <chuck> that's my setup line

20:22 <chuck> and I'm calling DataMapper.auto_migrate! after I require my model files, but talker.db is not created

20:22 <chuck> oh.. it's being created at / for some reason... hmm

20:27 <namelessjon> chuck: You probably want an File.expand_path in there.

20:28 <namelessjon> chuck: or alternatively: "sqlite3:#{File.join(File.dirname(__FILE__), 'talker.db')}"

20:45 <chuck> namelessjon, I threw expand_path in there, and it works, thanks!