1695 Commits (1d492f13b38fc714b50caf440c3b661a7de12f2e)

Author	SHA1	Message	Date
Philipp Tessenow	8fa15af8d8	lazy load project scopes	12 years ago
Philipp Tessenow	d9e9536bfd	let migrations run again	12 years ago
Christian Rijke	9be08a767e	Add a method to ensure the correct load order of core model classes with a circular dependency.	12 years ago
Philipp Tessenow	ef9ae5ffa3	use scope instead of named_scope	12 years ago
Philipp Tessenow	1f4f73c1d3	merge upstream changes (1f45cd3 - pull request #5). ... awesome nested set was deleted, so we take only half of the changes	12 years ago
Jens Ulferts	c9780734d8	lets AR handle casting of db condition for assignable_members	12 years ago
Philipp Tessenow	479086af1a	fix class load order	12 years ago
Philipp Tessenow	3affd94490	Revert "wrap scope in lambdas to let the migrations run again" ... This update repairs rake db:migrate but breaks rails s. I'll look into this This reverts commit 7a405355f4.	12 years ago
Philipp Tessenow	7a405355f4	wrap scope in lambdas to let the migrations run again	12 years ago
Jens Ulferts	0d3be2f958	no longer serializes changes as they will be anyway ... Assigning an already serialized attribute to an attribute that will be serialized by active record is disabled due to security issues (rails 2.3.17)	12 years ago
Jens Ulferts	2c22694e98	eager loads objects associated to the issue's relations	12 years ago
Martin Linkhorst	a188dfd713	Project.all_public can also be written as Project.public.all or just Project.public	12 years ago
Jens Ulferts	e28a52e05a	removes explicit casting of changes to yaml ... Journal changes are serialized anyway. This double serialization worked fine until a security risk was identified and a fix for that was created which prevents already serialized data to be serialized again (rails 2.3.17)	12 years ago
Jens Ulferts	74e21f7770	refactors querying for assignable users ... It is still not as fast as it should be. Even after refactoring it takes 1-2 seconds before the answer is calculated (used to be 5-6).	12 years ago
Philipp Tessenow	ddc6cdec54	force message to update its watchers and watcher_users when creating a new watcher in its after_create hook	12 years ago
Martin Linkhorst	a860d6979a	change the way the cache gets invalidated when a setting value is changed. now deletes all cache keys containing the setting's name	12 years ago
Martin Czuchra	daa880645b	Fixes problems that may occur in combination w/ plugins and the new API. ... Adds RESERVED_IDENTIFIERS to Project class, since route re-definitions in certain plugins may need this information in order to not overwrite the level_list collection route on projects.	12 years ago
Martin Linkhorst	2d2851b2d7	self.translated_attributes will only find persisted translations. therefore translations not yet persisted will be added manually.	12 years ago
Martin Linkhorst	857d4eaa8c	manually implement validates_presence_of :some_translated_attribute and :scope	12 years ago
Jens Ulferts	239cdf789f	fixes I18n mail parsing ... The mail handler parses a received mail for attributes. Doing so, it searches for: 1) the humanized key 2) the attribute translated to the user language 3) the attribute translated to the default language Prior to this fix, 2) and 3) where not possible as the responsible method #extract_keyword! always received a string.	12 years ago
Jens Ulferts	bb1a12d910	fixes change's path_from utf8ified assignment	12 years ago
Martin Linkhorst	d8e3513b08	we need a watchable for this validation	12 years ago
Martin Linkhorst	252fb107ea	revert an old commit and assign watcher parameters through mass assignment again. instead, add a validation in the model that ensures that the user is allowed to watch the watchable.	12 years ago
Martin Linkhorst	3ae5b2fac7	fix repository tests	12 years ago
Martin Linkhorst	a3675092c8	fix role tests	12 years ago
Christian Ratz	60c9e3d752	allow to set a default version on issue create via mail handler	12 years ago
Martin Linkhorst	5dc3cb2240	the default sort order in the gem version of awesome_nested_set is called :order_column. also fix a test that didn't expect the children to be sorted by name, now it does...	12 years ago
Martin Linkhorst	37867b4859	forcing attributes to be set is done differently since rails 3.2+: assign_attributes({ … }, :without_protection => true)	12 years ago
Martin Linkhorst	dd9515d745	remove cvs scm from openproject core	12 years ago
Martin Linkhorst	96dacf3150	remove bazaar scm from openproject core	12 years ago
Martin Linkhorst	2ff10f4c54	remove mercurial scm from openproject core	12 years ago
Martin Linkhorst	4b53920965	remove darcs scm from openproject core	12 years ago
Philipp Tessenow	8b661f1e2b	simplify Project#add_member	12 years ago
Jens Ulferts	3adca91704	overrides awesome_nested_set methods for issues to allow for fixing the ... root_id The root_id provides the scope for a nested_set tree. The implicit constraint of the Issue implementation is that there is but one tree per scope. Or in other words, the root_id is part of the nested_set attributes for an issue. This is not the case for Project, which is also using nested_set but does not set a scope. Project's are ordered that way. The root_id should always point to the root of an issue tree. awesome_nested_set only looks for nodes without a parent in the provided scope to determine root nodes. For example, if the root_id of an issue should point to the node itself, this issue is considered to be a root node. This is regardless of the value of the parent_id. For such a node, nested_set methods will provide wrong issues when queried for root, ancestors and similar methods. Fixing a missconfigured root_id attribute is achived by traversing up the parent chain (parent_id) till an issue without a parent is found. The parent_id attribute is considered to be correct. The root_id is not really necessary for awesome_nested_set to work. But employing the field should be beneficial performancewise. Updating an issue tree requires less nodes' left/right attribute to be altered when the forrest is scoped.	12 years ago
Jens Ulferts	705200ec33	matches journaled field against regexp to be able to cope with dynamic field names ... Some plugin, e.g. acts_as_attachable and acts_as_customizable, dynamically add associations to a model following a schema, e.g attachments_[n], where n is increased over time.	12 years ago
Jens Ulferts	98ddfe1b98	moves OpenProject formatters into JournalFormatter::Base subclasses ... This should separate the concerns between the plugin and the application. There are classes for Attachments, CustomFields and Diff.	12 years ago
Gregor Schmidt	c94c0789bb	Async loading of project list in project jump box ... * Adding special URL endpoint to get all visible projects at once (w/o paging - as in /projects/index.json) * Adding openProject JS object to carry useful information and helper methods for JS land * Adding helper method to User to get the number of visible projects w/o loading all of them * Reducing dependency to DOM structure of select2 - using API access instead * Close menu when select2 list is closed	12 years ago
Philipp Tessenow	578590693c	revert performance optimization for Rails2 ... touch_journaled_after_creation was optimized for ActiveRecord 2.3.x ActiveRecord 3.x should already be optimized (and should not trigger all the callbacks on journal.touch)	12 years ago
Martin Linkhorst	02e58ee814	put possible values for available settings into a lambda block because available_settings class variable can change at runtime (was different before)	12 years ago
Philipp Tessenow	edc058cc90	document Project.allowed_to_condition. with doc-string from redmine	12 years ago
Philipp Tessenow	a20c3068e0	fix typo in name of TimeEntry#earliest_date_for_project	12 years ago
Philipp Tessenow	33fd40decf	composed multiple redmine commits (thanks goes to jpl) to let MailHandler tests pass.	12 years ago
Martin Linkhorst	ef6680636e	fix two order clauses which create unambiguous sql queries	12 years ago
Martin Linkhorst	f6746c5327	bug: watcher uses redmine's safe_attributes so we need to include it	12 years ago
Gregor Schmidt	d95a75ec23	Perf opt - touching journaled w/o callbacks ... Initially the journaled was only touched on journal creation to make sure, that journal updates, that do not update the journaled element itself (like adding a note) cause a change in updated_(on|at). This has gone a bit overboard, since simply touching the journaled causes a roundtrip to the db and triggering of all save callbacks on journaled in all cases. These changes try to avoid the db roundtrip by only saving when things actually changed and avoid triggering the callbacks by using update_all instead of save.	12 years ago
Gregor Schmidt	9755b931dd	Extending API of recalculate_attributes_for ... This way we may not need to fetch an entry from the DB, where an existing object is already present. This might be of special intrest, when plugins alias-method chain recalculate_attributes_for and want to save a separate save call.	12 years ago
Gregor Schmidt	ccc97fcbfa	Optimization: Avoiding save if nothing changed ... this should avoid more callbacks being triggered and should therefor improve performance	12 years ago
Gregor Schmidt	ab968a52fc	Micro-optimization - no initialization when admin	12 years ago
Gregor Schmidt	9b3a33ebd0	Using already loaded information if available ... Falling back to light-weight loading if information was not already present	12 years ago
Jens Ulferts	4644f44087	eager loads child wiki_menu_items ... * as the wiki_menu items are queried to be presented in the menu, the children are always needed anyway	12 years ago