173 Commits (a2fd3078be8ae9b0f11cf1299f195d27f31d7250)

Author	SHA1	Message	Date
ulferts	05de0a6b3c	employ view_budgets permission check for budgets resource	4 years ago
ulferts	b179f89534	use view time_entries permission on spentTime schema	4 years ago
ulferts	8969afe617	close time entry modal on enter (#8653) ... * close time entry modal on enter As the time entry is saved anyway once the user saves the form, e.g. by pressing enter in the hours field, we can close the whole modal along with it. The "Save" and "Cancel" buttons do not make sense then anymore * attempt to stabilize flickering spec * fix ruby 2.7.1 hash parameter deprecations	4 years ago
Oliver Günther	e1a88a9a40	[34001] Ensure filters always receive the query context ... In the mentioned project, the query copy service fails to execute because a filter has lost its context right after `query.save`. There is a hook for `after_commit` which is not relevant here since we're in a transaction, but the context is lost regardless. Instead, we could ensure whenever accessing the filters, the context is set correctly. https://community.openproject.com/wp/34001	4 years ago
ulferts	38d2707946	Fix/wp sums in db (#8580) ... * remove apparently unused methods * have specs for wp query sums * grouped sums in sql * sql for total sums * alter interface of all_grouped_sums Since we now fetch all sums in one sql statement it no longer makes sense to fetch the group sums individually * remove now unused method * extract method * Add material_costs to summing * add labor_costs to group sums * add overall costs to sum * fix sum grouping descision * fix summable? check * remove work_package_list_summable_columns setting Now all summable columns are always summed. The user no longer needs to select summing up in the settings. Selecting the column to be displayed and activiating sums will suffice * fix flickering spec	4 years ago
ulferts	6826f90ee2	Feature/remove timelog (#8557) ... * rename costs, introduce budgets * move files from costs to budgets * rename cost_object to budget * remove unused code * move hook - should be turned into standard code in the long run * move type attributes change over to budgets * move patch to work_package proper * move budget menu item up * combine reporting, time and cost module * remove rails based time_entries & reports code * rename cost object filter * adapt menu spec expectations * use cost project module name in administration * include timeline labels in migration * properly place budget linking method * fix permitted params * remove outdated routing spec * adapt budget request specs * ensure order of descendent updates * remove outdated specs * fix checking for reporting to be enabled * fix displaying spent units * fix time entries activity event url * reenable current rate tab * fix path on budget page * allow bulk editing of budgets only in one project scenario * fix sanitizing reference in controller * include module required for format_date * fix reference to correct units from work package spent units * linting * remove outdated spec * remove outdated views and permission references * remove acts_as_event from time_entries There is no atom link for time entries * remove acts_as_event from projects There are no atom links for projects * introduce budget filter for cost reports * remove actions added to removed controller * move time entries to the costs module * factor in view_own permission when calculating time entry visibility * linting * move mounting of time entries * include budgets into api v3 documentation	4 years ago
ulferts	655f021b92	add derived dates to the work package resource ... The derived dates are the minimum/maximum of the descendant's dates. As long as the work package is scheduled automatically, the derived and non derived dates should be equal althoug the derived dates will be null in case non descendants exists. But if the user decides to activate manual scheduling the dates can deviate. The derived dates can then help to display the deviation to the user. This commit however does not alter the front end. The dates are eager loaded on the work package index commit using the same mechanims already introduced for spent time. The patched cost attributes, which are also eager loaded, are now eager loaded in the core as well so the patch is removed. As those attributes are now part of the core (although in a different module) it no longer makes sense to rely on patching for this.	4 years ago
ulferts	a2a559e553	remove almost duplicate, but faulty, readonly check ... The writability of the status is already checked at the attribute level	4 years ago
Oliver Günther	2feaec1101	[33678] Priority needs to be a require field ... Priority was marked non-required for it to turn up in the attribute form configuration. But priority _is actually required_. This results in the frontend adding an optional placeholder `-` value, which makes the resource unsavable. This commit adds priority to the form, but keeps it required as it should be. https://community.openproject.com/wp/33678	4 years ago
Markus Kahl	9c4d4ebc46	added :schedule_manually attribute to work package	5 years ago
Markus Kahl	ebdd8e0d8a	minor refactoring	5 years ago
ulferts	9e2287bfcc	rename and inflect according to zeitwerk	5 years ago
Henriette Dinger	bd7f4e4814	Update copyright notice	5 years ago
ulferts	6c4e572f4e	consolidate allowed_* logic in the wp contract	5 years ago
ulferts	7cca2631b0	add projects create form end point to api	5 years ago
ulferts	e37974dbed	actually provide all relevant custom fields for performance hack ... Before, the last iteration on a project determined the custom fields that where hacked to be loaded in a project. That way, when the last project had e.g. no custom fields on a specific type, the custom fields would be missing for all evaluations later on, even though it might actually have custom fields defined for other types. As the results of the first call would be cached, this would then also affect later requests	5 years ago
ulferts	ad847a3892	add type attribute group attributes to cache key ... This is an attempt to fix cases where some custom fields are not displayed. I suspect this will not fix the bug in itself. But it might help to narrow down the problem by preventing the bug to also affect users who are not affected by the bug originally but only because the buggy result was cached for them.	5 years ago
ulferts	d5ee1a7d20	ensure order of cache key	5 years ago
ulferts	a9213a39cc	remove duplicate permissions	5 years ago
ulferts	a880c85a1d	partial caching for wp schema attribute groups	5 years ago
ulferts	5ee801b549	cache roles for project	5 years ago
ulferts	0dcb1777d6	reduce db hits for wp schemas	5 years ago
ulferts	cc41884048	add schema to project api and complete project resource	5 years ago
ulferts	7971a29a8b	don't cache the attribute_groups	6 years ago
ulferts	4ffa71e68d	include a relevant setting in the wp schema cache key	6 years ago
ulferts	a199c94a6a	have permission specific schema caching	6 years ago
ulferts	e8f2ad5516	reenable caching using standard methods	6 years ago
ulferts	41bb309094	disable wp schema caching until it is turned save for it again ... [ci skip]	6 years ago
Oliver Günther	f492113c94	[30275] Allow global type select, restrict available projects ... This allows to select all available types globally, and restrict the projects to the ones that have the type available. The same functionality applies to embedded tables, where a filter on a type will ensure that only the projects with the type selected will be returned. https://community.openproject.com/wp/30275	6 years ago
ulferts	6e5b81dc3f	add schema specific wp contract	6 years ago
ulferts	a5defb73c5	add dedicated permission to regulate assigning a version	6 years ago
Oliver Günther	1b5fc32366	Avoid before blocks and raw params access for IDs ... Grape's validation runs after a `before` block so we should avoid using raw params there and instead using `declared(params)` which returns only the validated whitelisted params, much like a permitted params hash.	6 years ago
ulferts	1850622f20	Feature/versions api update (#7299) ... Feature/versions api update [ci skip]	6 years ago
Wieland Lindenthal	fd88776ceb	Add relations tables to Types via form configuration (#6773) ... * WIP: PrecedesFilter and FollowsFilter for WorkPackages * Create Type Form Groups per Query Type * generalize filters on relations avoiding alias * fix flickering spec * Rename dependency represeners for Follower and Predecessor Filters * Allow configuring templated queries within query configuring modal * Use `{id}` as templated value, instead of a now deprecated `templated` string * Rewrite ParentFilter to follow same structure as other relation filters. * Set Follower and Predecessor filters as hidden in Frontend * Default query shall show closed WPs, too * Remove relation filter type selection from form configuration view as we extend query configuration modal instead. * Allow ExternalQueryConfiguration to be more specific and prepend relation filters in its modal. * Quick unrelated fix on the way: the instruction text should start below the columns input and not to the right of it * - Add all missing relation filters. - Use verbs for naming instead of nouns to stay consistent with UI. * fix typo * Add relation filters to hidden filters and fix a rebase bug * Typo in docs * Fix broken rebase conflict resolution * Add part of and includes filter * adding specs for relation filters * rename RelatesFilter to RelatesToFilter * Rename shared example as direction does not matter anymore. * Fix spec for IncludesFilter * Fix name conflict for `includes` by calling typed_dag's `direct` later. * Fix shared filter examples as the parent filter does not get added per default anymore. * Fix spec after renaming PredecessorFilter to PrecedesFilter * Fix integration spec for type configuration form * Add filter dependency specs * Differentiate between WorkPackageFormChildrenQueryGroup and WorkPackageRelationQueryGroup types * Create inline-create for relation tables * Rewrite filters to use actual relation_type as key * Reload relations on change, change symbols * Fix relation filter i18n strings * Test presence of all relation filters frontend. * Extend integration test for form configuration to not only cover hierarchy, but also undirectional and directional relation types. * Consistent naming of filter dependency representer files * Satisfy codeclimate * Add missing copyright notes * Fixing PartofFilter and its specs * Reverse relation direction when created inline * WIP integration test for Relation Query Component and Children Query Component. * Correctly reload children tables when refreshed * Fix FilterOnUndirectedMixin * Fix removing of relation from embedded table after reload * Adding intergration tests for removing and adding relation or children, initiated form either embedded table or relations tab. * Fix specs * Fix implementation of duplicates and duplicated They are reversed in typed_dag and causes confusion [ci skip]	6 years ago
Oliver Günther	3ba6913960	Implement read-only work package attributes based on status	6 years ago
Jens Ulferts	d04b86da9c	introduce custom eager loading for work package list ... Reduces the amount of AR models instantiated when handing wp#index requests by: * Calculating part of the checksum for caching in the db. This is done by using the id and the updated_at/on field of associated models whose representation is cached in the wp representer. * Reusing instantiated models for multiple associations. e.g instances of Project are used in various places (the work package itself, it's ancestors and children). Instead of using Rails' eager loading which would instantiate a project potentially once for every association path, the code now assigns the same instance to all those associations. Additionally, the commit streamlines the creation of custom field injected representers. Now only one class is created for all work packages when before it would be one class for every uniq set of custom fields. On top of that, the class can be reused if the set of custom fields does not change.	7 years ago
Jens Ulferts	ce21ee20f7	cached representers	7 years ago
Jens Ulferts	5a5e3494f9	turn attribute group into dedicated object	7 years ago
Jens Ulferts	77116b5cdb	ensure not to have user specific query cached in schema	7 years ago
Jens Ulferts	9e4f5053bc	hack having the children table provisioned from the backend	7 years ago
Jens Ulferts	3e86dafad2	extract children rendering into dedicated component	7 years ago
Jens Ulferts	a33b0bbbbd	WIP - return Query as an attribute group ... * only default query for now * only returned when no custom configuration is applied * start to turn attribute groups into ressources	7 years ago
Jens Ulferts	6d359cb02d	turn attribute groups into array of objects	7 years ago
Oliver Günther	6ef4211e77	Bump copyright to 2018 (#6171) ... [ci skip]	7 years ago
ulferts	61281b7f6a	explicitly set author to be nonwritable in wp schema (#5805) ... While the author is writable, one can set it to the current user. It should not be marked as writable as the backend makes that assignment already. But because of this behaviour we cannot rely on the default behaviour for determining whether a property is writable [ci skip]	7 years ago
Oliver Günther	ab9a9baee7	[26002] Add date as writable when type is milestone (#5804) ... https://community.openproject.com/wp/26002 [ci skip]	7 years ago
ulferts	7154caec85	Fix/consolidate representers (#5732) ... * consolidate representers Minimizes the need for extensive payload representers by: * introducing the payload_representer module which removes all non writable propertyies from a representer * introducing the resource/resources/associated_resource/associated_ressources helper methods which provide a framework for having a representer suitable for parsing and rendering Additionally increases the importance of the contracts by using them to calculate the writeable attributes. * remove unused endpoint [ci skip]	7 years ago
Oliver Günther	8956700408	[25374] Remove attribute visibility	7 years ago
ulferts	65c7db3120	avoid side effects on calls to attribute_groups (#5677) ... Without the deep_dup, the call to attribute_groups used to alter the type's attribute_groups deeply nested attribute_groups property. As a result, the second call to attribute_groups would be different from the first one. This became most apparent when rending the schema for a couple of work packages all sharing the same type. As the type was reused to render the schema, but the project might have changed, only the first schema for the type was accurate, while subsequent schemas for that type lacked attributes [ci skip]	8 years ago
ulferts	0c0b16508b	Fix/bump representable (#5465) ... * bump reform and roar -> bumps representer * adapt to changed validation interface * disable initializer patch for now * adapt to changed representable attr interface * can no longer have private methods inside a representer * private no longer possible for representer * bump reform * wip - restyle validation * remove commented out patch * apply injection as prescribed * reactivate reform error symbols patch * remove patch to Hash superfluous wit ruby 2.3 * remove outdated human_attribute_name patch * whitespace fixes * adapt filter name after removal of human_attribute_name patch * adapt filter specs to no longer rely on human_attribute_name patch * fix version filter name * remove reliance on no longer existing human_attribute_name patch * use correct key in journal formatter * remove private from representer * adapt to altered setter interface * reenable i18n for error messages in contracts * no private methods in representer * defined model for contracts * fix validaton * instantiate correct Object * define model for contract * circumvent now existing render method on reform * replace deprecated constant * patch correct reform class - not the module - via prepend * refactor too complex method * replace deprations * remove remnants of parentId * prevent error symbols from existing twice * adapt user representer to altered setter interface * adapt watcher representer to altered setter interface * remove now unnessary patch * adapt setter to altered interface * adapt spec * fix custom field setters * remove parentId from wp representer As the parent is a wp resource, clients should use the parent link instead * adapt spec to changed valid? interface * remove parentId from wp schema * replace references of parentId in frontend * remove TODO [ci skip]	8 years ago