dm-accepts_nested_attributes 0.12.0

data/.gitignore

@@ -0,0 +1,26 @@
+test_log
+pkg
+pkg/*
+*/pkg/*
+bundle
+bundle/*
+doc
+*.log
+log
+!log*.rb
+*/log
+log/*
+*/log/*
+coverage
+*/coverage
+lib/dm-more.rb
+*.db
+nbproject
+.DS_Store
+rspec_report.html
+*.swp
+_Yardoc
+*/ri
+measurements
+tmp
+

data/CHANGELOG

@@ -0,0 +1,970 @@
+[47b2757 | Thu Dec 03 13:42:02 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Use jeweler, add standard dm-more tasks and bump to 0.12.0
+
+[2a08a89 | Mon Nov 23 16:46:40 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Simplified require statements
+
+[e24cefe | Mon Nov 23 16:53:50 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Use :required => true instead of :nullable => false
+
+    This means that dm-accepts_nested_attributes now
+    depends on dm-core >= 0.10.2
+
+[4c30792 | Mon Sep 21 17:06:53 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Fixed error message in case no appropriate relationship was found
+
+[41f43e8 | Thu Sep 17 12:19:11 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Simplified code doesn't depend on class_inheritable_reader
+
+    This means that dana is now usable inside modules,
+    which is necessary for proper dm-is-remixable suppport
+    and my dm-is-localizable plugin in particular.
+
+[95041df | Sat Sep 12 20:02:37 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Removed current limitations section from README
+
+    This no longer applies.
+
+[780f8a0 | Sat Sep 12 14:15:05 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Refactored to support ruby 1.9
+
+    This should fix the following error that was
+    previously thrown on ruby 1.9.1
+    
+      implicit argument passing of super from method
+      defined by define_method is not supported
+    
+    Also, it really should have been this way from
+    the beginning. Using define_method was completely
+    unnecessary.
+
+[95ab6bf | Thu Sep 03 15:42:08 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Finally, all specs pass!
+
+    * I'm not really pleased with the way it works now
+      though. If anyone has better suggestions, I'd be
+      glad to hear them.
+
+[503e6d9 | Wed Sep 02 13:00:56 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Removed now unnecessary workaround
+
+[f47516a | Fri Aug 28 15:22:56 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Destroying m:n now works, 1:1 and 1:n don't :)
+
+    * Getting weird can't modify frozen object errors
+      in has_1_spec and has_n_spec when running specs
+      for destroy.
+
+[07122e4 | Fri Aug 28 15:20:24 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Allow dm-constraints to migrate the constraints
+
+[e0d017a | Wed Aug 05 17:21:45 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Strip whitespace
+
+[19d1871 | Wed Aug 05 17:13:16 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Support STI scenarios
+
+    * Previously, calling accepts_nested_attributes_for
+      on an STI subclass resulted in errors because
+      options_for_nested_attributes wasn't inherited.
+    
+    * Hopefully this commit fixes that. No specs to prove
+      that though, so far.
+
+[d36e7ee | Wed Aug 05 17:11:15 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Enable 'rake spec' without errors saying dm-core can't be found
+
+    * However, rake spec still yields weird transaction errors like
+    
+      Lock wait timeout exceeded; try restarting transaction
+
+[1d51f5b | Tue Aug 04 13:08:41 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Bumped version
+
+[2fa2de4 | Mon Aug 03 07:53:51 UTC 2009] Angel N. Sciortino <contact@angeliccomputing.com>
+
+ * update operations now perform validations too
+
+    * This is achieved by calling resource.attributes = ...
+      followed by resource.save under the hood.
+    
+    * Once Resource#update calls Resource#save inside dm-core
+      this shouldn't be necessary anymore (but wouldn't hurt
+      either)
+
+[286ac11 | Mon Aug 03 18:19:02 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Commented currently unspecced relationship declarations
+
+    * These introduce lots of spec failures because dm-core
+      seems to establish not nullable FK properties on the
+      target end of the relationship.
+    
+    * For reference, the specs started failing with
+      http://is.gd/20SH8
+
+[571ba2e | Fri Jul 31 16:20:25 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Removed obsolete section from README
+
+[9a9bbe1 | Fri Jul 03 14:26:21 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * spec fixtures for has_1_through and has_n_through with n intermediaries
+
+[ae1acfa | Fri Jul 03 13:55:32 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * added additional spec expectations for nested attribute writers
+
+    * Added explicit expectations to prove that calling
+      nested attribute writers never alters the state
+      of the storage engine directly. Persistence will
+      only happen when #save is called.
+
+[7b0194f | Wed Jul 01 19:41:38 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * support composite keys when creating new nested resources
+
+    * refactored the previous constant into it's own method
+      to support composite keys and also allow overwriting
+
+[a9efa26 | Wed Jul 01 18:45:49 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * no assumptions on the length of the m:m target collection
+
+    * Maybe this will never be necessary, but I'm still thinking
+      about how this whole #assign_or_mark_for_destruction method
+      could be refactored. I don't like the fact the name itself
+      already implies that this method has 2 different purposes.
+
+[d2e0a9d | Wed Jul 01 18:20:36 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * fixed renaming regression in #normalize_attributes_collection
+
+[34efa0f | Wed Jul 01 18:19:57 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * don't sort attributes in #normalize_attributes_collection
+
+    * I can see no real advantage in doing so, let's keep it simple
+
+[5a17003 | Wed Jul 01 18:18:11 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * idiomatic datamapper usage
+
+    * this changes the spec error to a simple unmet expectation
+      compared to an fk violation error before
+    
+    * this is kind of weird, since existing_record is the same
+      in both cases
+
+[fa4edde | Wed Jul 01 16:46:47 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * no need to do anything if the m:m target collection is empty
+
+[599da7f | Wed Jul 01 16:45:20 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * don't expect the target_collection to always be of length 1
+
+    * actually this is the case in the current implementation,
+      but that may change, and the code is simpler that way
+
+[2ce2b3f | Wed Jul 01 16:40:22 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * respect composite keys at the target end of m:m relationships
+
+[0331a9f | Wed Jul 01 16:05:04 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * mark the correct resources for destruction in m:m scenario
+
+    * However, it doesn't work, because the target resource always
+      gets destroyed before the intermediaries
+
+[a1e8da4 | Mon Jun 29 15:39:19 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * removed meaningless documentation
+
+[a07dcef | Thu Jun 25 03:26:08 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * added spec that proves that m2m deletion wipes out all intermediaries
+
+    * this is bad bad behavior and needs to be fixed
+
+[c8d21a0 | Thu Jun 25 03:04:20 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * all specs pass with or without dm-constraints
+
+    * deleted an invalid spec that expected dm-constraints
+      to be present
+
+[6913c23 | Thu Jun 25 02:42:13 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * allow github to build a datamapper 0.10 compatible gem
+
+[d6abeb1 | Thu Jun 25 02:39:03 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * removed rspec tmbundle logger support
+
+[bc119f9 | Thu Jun 25 02:34:45 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * support running specs with or without dm-constraints
+
+[84609fb | Thu Jun 25 01:57:03 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * updated datamapper 0.10.0 compatibility information
+
+[bda6fbc | Thu Jun 25 01:53:38 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * cleaner code for many_to_many deletion
+
+    * also added TODO note about handling this with
+      mark_for_destruction like all the other
+      relationship kinds.
+
+[bce5322 | Thu Jun 25 00:58:11 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * support deleting of many_to_many associations (all specs pass)
+
+    * Deletion is currently triggered when calling the nested
+      attribute writer. This is most probably not desired and
+      is likely to change.
+    
+    * Example:
+    
+      Person.has(n, :projects, :through => :project_memberships)
+      @person.projects_attributes = {
+        :id => project.id, :_delete => true
+      }
+    
+      This will delete the associated project_membership
+      and then the project
+
+[a8f2a9a | Wed Jun 24 15:52:48 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * mark_for_destruction? now really returns true or false (not nil)
+
+[8127032 | Wed Jun 24 15:44:27 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * corrected specs for atomic commits
+
+    * those specs aren't really shared, but only valid
+      for situations with no :reject_if guard that returns
+      true.
+    
+    * only 1 spec failing now and that's deleting
+      has(n, :through)
+
+[226a5a6 | Wed Jun 24 04:00:41 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * added hopefully last missing documentation
+
+[51b5fdb | Wed Jun 24 03:54:41 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * finished documenting all modules and methods
+
+    * also fixed a bug where attributes weren't sent
+      to the :reject_if guard in case the guard was
+      specified as Symbol or String.
+
+[faeaf5b | Wed Jun 24 02:05:08 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * #save_self no overwritten on every resource
+
+    * this makes more deletion specs pass and is actually
+      really obvious. wonder why it took me so long.
+    
+    * still a few atomic commits related specs failing
+      but slowly getting there.
+    
+    * also, this needs my current dm-validations work
+      available at
+      http://github.com/snusnu/dm-more/commits/validation_context_work
+
+[2582c12 | Wed Jun 24 02:02:36 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * docfixes and a minor internal implementation change
+
+[097bf9d | Wed Jun 24 02:01:13 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * minor internal method renaming.
+
+    * actually, i'm considering refactoring this method
+    * i don't like the name, it already indicates that
+      this methods serves 2 purposes which isn't good.
+
+[5cd44f9 | Wed Jun 24 01:58:52 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * no need to name useless rest arguments
+
+    * in fact, it may very well be safe to remove them
+      altogether. think more about this.
+
+[666ee63 | Wed Jun 24 01:57:56 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * use mysql in specs since dm-constraints are broken for sqlite3
+
+[92d25ff | Wed Jun 24 01:57:37 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * use dm-constraints in specs for more safety
+
+[9c5f2af | Wed Jun 17 16:16:00 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * updated Manifext.txt and gemspec to reflect the current structure
+
+[717d870 | Tue Jun 16 13:17:34 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * adapted to latest dm-core
+
+    * still the same specs failing
+
+[0ae2f1a | Mon Jun 15 15:02:02 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * less invasive dm-core overwrite in Resource#save_parents
+
+    * i don't know if that's the best way, but it works for now
+
+[dc188ee | Mon Jun 15 15:01:06 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * minor spec refactoring to run the shared examples all the time
+
+[178c99b | Sat Jun 13 20:20:44 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * added comments about what goes wrong
+
+[9f2279e | Sat Jun 13 18:52:47 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * styles matter, i like this better
+
+[123c70f | Sat Jun 13 18:26:21 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * refactored shared specs for has(n, :through) associations
+
+    * factored out real shared behavior and run it from
+      all examples
+    
+    * causes 1 failure in "should perform atomic commits"
+
+[2393d0d | Sat Jun 13 18:25:57 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * refactored shared specs for has(n) associations
+
+    * factored out real shared behavior and run it from
+      all examples
+    
+    * causes 1 failure in "should perform atomic commits"
+
+[f786acd | Sat Jun 13 18:25:11 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * refactored shared specs for has(1) associations
+
+    * factored out real shared behavior and run it from
+      all examples
+    
+    * causes 1 failure in "should perform atomic commits"
+
+[ad7d968 | Sat Jun 13 17:45:58 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * refactored shared specs for belongs_to associations
+
+    * factored out real shared behavior and run it from
+      all examples
+
+[a14e6fd | Sat Jun 13 16:07:36 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * made all but 3 specs pass
+
+    * this means that all specs except those that test
+      destroying of an associated resource via the
+      nested attribute accessors, works.
+    
+    * TODO this spec change needs more thinking. I'm
+      not yet sure if not creating a Person in this case
+      is the right behavior.
+
+[8de57d1 | Sat Jun 13 16:05:44 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * adapt overwritten save behavior to latest dm-core changes
+
+[e3837fe | Sat Jun 13 16:05:18 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * don't use chainable in client code
+
+    * it should be sufficient that dm-core declared
+      the method as chainable
+
+[4121c57 | Sat Jun 13 16:04:26 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * removed debug statements
+
+[80b9774 | Fri Jun 12 20:59:12 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * added debug statements
+
+[3d316f8 | Fri Jun 12 20:58:56 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * specs now explicitly assign related resources to the association
+
+[04a87ad | Thu Jun 04 15:44:20 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * removed unneeded assertions
+
+    * deleting a related resource still not working
+    * transactions still not working
+    
+    * other than that, it seems quite fine
+
+[6a2e07f | Tue Jun 02 20:17:56 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * include save behavior only if a resource accepts nested attributes
+
+[d2c27c6 | Tue Jun 02 19:59:41 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * spec formatting and useless assertion removal
+
+[5f8d0e1 | Tue Jun 02 19:39:28 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * added a few TODO and FIXME comments
+
+[a317852 | Tue Jun 02 19:05:20 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * moved shared specs to spec/shared directory.
+
+    * this reduces clutter from the main specs
+    * also it will hopefully help with refactoring
+      the specs to provide better shared specs that
+      can be used for all integrations specs
+
+[3ab51dc | Tue Jun 02 18:24:23 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * moved rspec_tmbundle_support to spec/lib
+
+[3579f97 | Tue Jun 02 18:19:03 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * formatting and some assertion changes for belongs_to specs
+
+[7752b87 | Tue Jun 02 18:06:08 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * nice formatting for spec fixtures.
+
+[4c18a35 | Tue Jun 02 17:55:25 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * use parent saving behavior for *all* resources.
+
+    * breaks lots of specs that were already passing
+    * i don't know why this is, but nevertheless i think
+      that this is probably the way to go.
+
+[653f615 | Tue Jun 02 17:33:15 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * disabled mark_for_destruction for now
+
+[15a5174 | Thu May 28 18:06:23 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * docs for the resource extensions
+
+[22dea73 | Thu May 28 17:40:55 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * more docs for the model extensions
+
+[a109f10 | Thu May 28 16:19:29 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * allow deletion of many_to_one target via generated writer
+
+    * this makes all belongs_to_specs pass, however it doesn't
+      work for has_1_specs.
+    
+    * before this changeset, both belongs_to and has_1 failed
+      at the same spec (deleting the related resource). i wonder
+      why that is ...
+
+[81eda71 | Thu May 28 16:16:50 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * docs for one method to have yard style in place
+
+[b50b192 | Thu May 28 16:01:00 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * added a few TODO notes
+
+[8652a89 | Thu May 28 16:00:49 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * readability ftw
+
+[89d02cf | Thu May 28 14:52:14 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * better names for internal methods
+
+    * accepts_nested_attributes_for_relationship_to_#{type}
+      is now
+      accepts_nested_attributes_for_related_#{type}
+
+[04e39ed | Thu May 28 14:41:03 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * fixed reject_new_record guards (gone bad while refactoring)
+
+    * belongs_to and has_1 specs only have 1 failure
+      related to deleting the record
+    
+    * TODO: find out why mark_as_destruction doesn't
+      do the right thing when called from #save
+
+[7052c72 | Thu May 28 14:39:23 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * removed comment in shared spec that doesn't apply to every usage
+
+[8479ee6 | Thu May 28 13:32:39 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * fixed nested_attributes reader (gone bad in the refactoring)
+
+[b72f510 | Thu May 28 13:16:24 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * better formatting for save methods
+
+[d8535cc | Thu May 28 13:05:05 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * changed internal method name convention
+
+    * accepts_nested_attributes_for_#{type}_relationship
+      is now
+      accepts_nested_attributes_for_relationship_to_#{type}
+
+[1fdf6b5 | Thu May 28 12:37:35 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * rspec_tmbundle_support now uses Extlib::Logger
+
+    * DataMapper::Logger is now Extlib::Logger
+
+[a2d4f4e | Thu May 28 04:29:58 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * removed resource_spec from Manifest and gemspec
+
+[2e89d38 | Thu May 28 04:29:29 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * no need to expect explicitly ordered spec files
+
+[76c06a3 | Thu May 28 04:28:53 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * i guess whitespace isn't necessary there
+
+[0e309e1 | Thu May 28 04:16:56 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * bypass needless calls to association_name all the time
+
+    * key options_for_nested_attributes by relationship
+      instead of association_name. this eliminates the
+      need to repeatedly fetch the relationship by its
+      name and provides more concise codes in general.
+    
+    * removed the relationship method since it's not
+      needed anymore
+    
+    * removed the reject_new_nested_attributes_guard_for
+      method since it's not needed anymore
+    
+    * number of passing specs is still the same
+
+[dd859c3 | Thu May 28 04:12:44 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * include save_or_destroy in save, no need to separate it
+
+[fdd29b0 | Thu May 28 02:36:01 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * removed dead meaningless parameters
+
+    * also, let the code speak for itself if it can.
+      in this case, the comments didn't offer any
+      additional value
+
+[da8747c | Thu May 28 02:23:00 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * typo
+
+[8c6a061 | Thu May 28 02:22:16 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * methods should have a single task
+
+    * don't validate the association_name inside
+      assert_valid_options_for_nested_attributes
+
+[da2d474 | Thu May 28 02:21:15 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * explicitly include plugin functionality
+
+    * don't alter existing datamapper api by simply
+      requiring files. instead, do it explicitly as
+      this also serves a documentation puropose since
+      it's more obvious what's going on.
+
+[ae50006 | Wed May 27 23:31:47 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * every case statement wants an else branch
+
+    * although the plugin will guarantee not to enter
+      this code path, who knows what people try to do.
+      it's a (private) method, so someone could use
+      it in a possibly wrong way so why not inform
+      him/her if that happens.
+    
+    * also, i think that the use of 'next' instead of
+      deeply nested if/unless/case reads nicer
+
+[38ece56 | Wed May 27 23:17:36 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * avoid mutating parameters when it's not necessary
+
+[38225d6 | Wed May 27 23:15:32 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * made normalize_attributes_collection self contained
+
+[b42cf92 | Wed May 27 21:50:16 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * removed specs for ':reject_if => :foo' for now
+
+    * added a TODO note to remind me to add them back again,
+      once i have decided where to best put them.
+
+[57ccf94 | Wed May 27 21:41:59 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * check proper exception class in specs
+
+[6fa76ee | Wed May 27 21:11:54 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * replaced meaningless specs with one that makes sense
+
+[361a5d1 | Wed May 27 21:01:04 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * don't use bang! methods when not appropriate.
+
+    * see http://is.gd/HqLo for a detailed explanation
+
+[4e8014b | Wed May 27 20:52:07 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * extracted method mainly for documentation purposes
+
+[5e22497 | Wed May 27 20:51:04 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * replaced explicitly defined helper with dm-core api access
+
+[e981ce6 | Wed May 27 19:10:46 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * cleaned up weird condition syntax
+
+[9a6318c | Wed May 27 19:10:20 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * better usage of RelationshipAccess
+
+    * it's not necessary to call relationship! from
+      inside accepts_nested_attributes_for.
+      The association_name is *definitely* valid at
+      this time.
+
+[6e8c49f | Wed May 27 18:20:47 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * model extension now less invasive and easier to extend
+
+    * Added relationship/relationship! methods to
+      provide a cleaner interface to the relationships
+      defined on a model. This hides the implementation
+      detail that relationships are stored in a hash
+      which is bound to change anyway. Also, it allows
+      to *fail fast* when trying to access a relationship
+      that is not defined in a model (via relationship!)
+    
+    * restructured the files to better encapsulate and
+      communicate what's in them.
+    
+    * better use of dm-core methods allowed previously
+      explicitly declared methods to be removed.
+
+[5084804 | Tue May 26 18:55:12 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * imported dkubb's version of Resource#save for dana
+
+    * unit specs all pass
+    * belongs_to specs pass except for deletion (not yet integrated)
+    * all has_x specs still have failures
+
+[44fd13d | Fri May 22 20:52:53 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * having serious troubles trying to get 0.10.0 to work :-(
+
+[aed8d1e | Wed May 20 02:53:32 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * NOT compatible with datamapper 0.10.0
+
+    * a list of errors when running unmodified specs:
+    
+      uninitialized constant
+      DataMapper::Associations::RelationshipChain
+    
+      undefined method `transaction'
+      for #<Person @id=nil @name="snusnu"
+    
+      undefined local variable or method `create'
+      for #<Person @id=nil @name="Martin">
+      (probably extlib related?)
+    
+      The property 'name' is not accessible in Profile
+      (probably transaction related)
+
+[eecfa47 | Wed May 20 02:45:48 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * updated version to 0.10.0
+
+[55d6029 | Wed May 20 02:40:51 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * removed usage of rubygems
+
+[9ae06e1 | Wed May 20 02:40:39 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * upgraded datamapper dependencies to 0.10.0
+
+[b704849 | Wed May 13 04:30:18 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * updated CHANGELOG
+
+[b621475 | Wed May 13 04:29:30 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * bumped version to 0.0.6.
+
+[91923dd | Wed May 13 04:18:49 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * respect dm-is-remixable (thx Asher Van Brunt for pointing this out)
+
+    Since class_inheritable_accessor is not available on Module, dana
+    bailed out for dm-is-remixable. This should hopefully fix this issue.
+
+[ef8e1ca | Mon May 11 00:17:43 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * added CHANGELOG rake task and file (taken from manveru/ramaze)
+
+[52c55bc | Wed Apr 29 04:50:32 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * bump tiny version for extlib/hook compatible gem.
+
+[c34edae | Wed Apr 29 04:46:43 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * now compatible with extlib/hook. (last commit was missing files)
+
+[8e42229 | Wed Apr 29 04:34:20 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * workaround extlib/hook not respecting hooked but redefined methods.
+
+[578e658 | Wed Apr 29 02:46:56 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * only alter common DataMapper namespace if really necessary.
+
+[7da8378 | Wed Apr 29 01:41:07 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * less brittle specs with no more auto_migrations before(:each)
+
+[f2a2fcb | Wed Apr 29 01:38:00 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * updated Manifest and gemspec to reflect spec cleanings
+
+[dd0a4a8 | Fri Apr 24 17:29:41 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * removed fixtures and specs for customized has(n, :through) associations.
+
+    It makes no sense to spec all different options for the association
+    macros in integration specs. Rather, it should be specced once and
+    for all (in unit specs) that dm-accepts_nested_attributes infers the
+    right kind of associations. Think about it, then do it!
+
+[ac7c008 | Fri Apr 24 17:27:12 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * be explicit about before(:each)
+
+[7a56d17 | Fri Apr 24 17:26:35 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * removed commented code
+
+[03bc5c2 | Fri Apr 24 17:25:25 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * don't auto_migrate! before(:each) spec run (takes half the time)
+
+[b6f5bf5 | Tue Apr 21 14:52:22 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * removed debugging outputs
+
+[001c791 | Sat Apr 18 14:31:35 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * added red specs for transactions. make them green.
+
+[9852f03 | Sat Apr 18 03:07:04 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * formatting
+
+[3fceda3 | Sat Apr 18 02:34:54 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * no need to require gem dependencies in spec_helper.rb
+
+[2118726 | Wed Apr 15 18:57:22 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * initial support for wrapping every save inside a transaction.
+
+    * this probably needs more thinking, also funny is that specs if
+      called from textmate rspec bundle all pass, whereas when called
+      from the shell, some of them (only has(1)) break.
+      investigate this
+    * dbussink also suggested it might be a good idea to add an option
+      that allows to turn transaction usage off.
+    * also, look for some api to infer wether a backend supports
+      (nested) transactions or not. we could default to true if it does,
+      but still leave the option to turn it off nevertheless.
+
+[abdc211 | Wed Apr 15 14:18:07 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * added sanitize_nested_attributes to provide a hook for doing this.
+
+    * see http://is.gd/sz2d for the motivation behind this new method
+
+[158b6ea | Wed Apr 15 02:50:38 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * fixed broken generated gemspec. find out why this happens.
+
+[5603c57 | Wed Apr 15 02:49:54 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * updated manifest to include new specs from integration branch
+
+[4c5e656 | Tue Apr 14 18:25:02 UTC 2009] Rupert Voelcker <rupert@rupespad.com>
+
+ * added #{association_name}_attributes reader
+
+    Signed-off-by: Martin Gamsjaeger <gamsnjaga@gmail.com>
+
+[6d26213 | Tue Apr 14 03:09:34 UTC 2009] Rupert Voelcker <rupert@rupespad.com>
+
+ * added associated attributes reader (doesn't do anything yet)
+
+    Signed-off-by: Martin Gamsjaeger <gamsnjaga@gmail.com>
+
+[32dcfe8 | Wed Apr 15 02:20:00 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * initial support for dm-validations
+
+    * this desperately needs specs! currently only basic not null
+      constraint specs pass (well, there are no other).
+    * saving a resource along with its associations should always
+      be done inside a transaction! figure out how to do that!
+    * validation errors from associated resources should be copied
+      to the parent resource. this currently happens, but it's not
+      specced at all so consider the behavior undefined.
+
+[fef1578 | Wed Apr 15 01:37:14 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * initialize dm logger before setting up dm adapter
+
+[8d7ede8 | Wed Apr 15 01:35:50 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * allow to call puth helper like puts without parameter
+
+[4e204a1 | Tue Apr 14 01:19:30 UTC 2009] Rupert Voelcker <rupert@rupespad.com>
+
+ * now works with 'renamed' has many relationships
+
+    Signed-off-by: Martin Gamsjaeger <gamsnjaga@gmail.com>
+
+[8f20aad | Mon Apr 13 00:33:40 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * build that gem\!
+
+[cc02415 | Sun Apr 12 22:57:26 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * updated manifest
+
+[d0f0dce | Sun Apr 12 22:57:08 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * use constants from Rakefile in tasks/install
+
+[b45ea52 | Fri Apr 10 16:12:11 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * added get_#{association_name} helper for (many/one)_to_one associations.
+
+    This is mainly to support common usecases in fields_for. This method
+    will return a new record if the associated object is nil at the time
+    it is accessed, and the associated object otherwise. This should help
+    fields_for to render the right id and name attributes for the input
+    fields. This method is not generated for (one/many)_to_many associations
+    since these return an Collection anyway and this allow to call build
+    or get on it. A possible use case would look something like this:
+    
+    <% fields_for @location.get_geo_point do %>
+     ...
+    <% end =%>
+    
+    The whole purpose of this commit is to provide something to fields_for
+    that is guaranteed to not be nil.
+
+[8acb396 | Fri Apr 10 02:28:48 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * no .rb extension in require statements
+
+[bd6dfe6 | Fri Apr 10 01:52:45 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * added passing specs for reject_if procs
+
+[2a1d309 | Thu Apr 09 19:55:52 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * updated README
+
+[b0fcf44 | Thu Apr 09 19:52:14 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * update now works for has(n, :through) -> all works now
+
+[a8ba5f6 | Thu Apr 09 19:44:48 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * destroy now works for has(n, :through)
+
+[2b5d583 | Thu Apr 09 19:11:35 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * explain what's going on when creating a new has(n, :through) child entry
+
+[b8a4be4 | Thu Apr 09 18:59:44 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * updated README to reflect latest changes
+
+[1efbb4c | Thu Apr 09 18:56:27 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * create now works for has(n, :through)
+
+[71d3d73 | Thu Apr 09 16:07:41 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * added a few spec helpers
+
+[2783be5 | Thu Apr 09 03:06:59 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * more robust (backwards compatible) monkeypatch.
+
+[c49516e | Thu Apr 09 01:42:23 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * README cosmetics
+
+[974d7b5 | Thu Apr 09 01:38:01 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Updated README to reflect the latest status.
+
+[69eb9ff | Thu Apr 09 01:28:10 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * reworked association saving. all except has(n,:through) work now
+
+[1200617 | Wed Apr 08 18:57:58 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * really minor README cosmetics
+
+[bda2607 | Wed Apr 08 17:44:34 UTC 2009] snusnu <gamsnjaga@gmail.com>
+
+ * Push it
+