05 Mar, 2010
1 commit
-
* The problem ocurred because the body of PublishedArticles is empty and it must call the body of the article to which it references to. This is done through the to_html method. * Also included fast_create defaults for blog. (ActionItem1395)
20 Jan, 2010
2 commits
-
(ActionItem1367)
-
Conflicts: po/de/noosfero.po po/pt/noosfero.po
18 Jan, 2010
1 commit
-
* Not allowing people to signup without a name * Adding a migrate to set identifier as name for people that are already without a name * Changing the field "Full Name" to "Name" and to become red when not filled * Adding new db:schema (ActionItem1352)
13 Jan, 2010
1 commit
-
(ActionItem1357)
18 Dec, 2009
1 commit
-
This way we can reuse that logic (fast creation with a category and/or adding a category) in other tests for search functionality.
11 Dec, 2009
1 commit
-
This time: * test/functional/search_controller_test.rb * test/functional/profile_editor_controller_test.rb The time gain was small, so perhaps we need to formulate another strategy for functional tests. The way we are doing them, they always require hitting the database.
25 Nov, 2009
1 commit
-
(ActionItem1255)
31 Oct, 2009
1 commit
-
* Moved all object creation helper methods into Noosfero::Factory * Added a new and faster version of create_user. * Adjusted some test cases tests to use the new factory methods This comprised removing a lot of bad smells from the tests, involving eliminating uneeded database hits. * reviewed ProfileTest for speed as a case study. The test suite used to take ~ 40 seconds to run, now it takes ~ 14 seconds. If we do this to all test suites, we can dramatically reduce the time we need to run all the tests * Added a rake task test:smells to list potential test smells. Started with probably unecessary and/or too slow database hits * In the end, running all the tests now takes ~ 13 minutes on Colivre's development machine, instead of the ~ 30 minutes it used to. (ActionItem1279)
27 Oct, 2009
1 commit
-
* transformed script/feed-updater into a controller script. It starts and stops together with the production system (script/production) * moved the update (daemon) logic into FeedUpdater class. It knows which feeds must be updated and when, and when it should stop running. * concentrated the fetch (download) logic into FeedHandler class. It knows how to update a specific feed. * implemented the concept of "enabled" and "expired" in both ExternalFeed and FeedReaderBlock. The feed updater looks for feeds that are both enabled *and* expired to update. * Disabled feed reader blocks get re-enabled when their address is changed. * fixed a bug that made some feeds crash when using PostgreSQL (CGI::unescapeHTML transforms Numeric Character References into iso-8859-1 data and PostgreSQL won't accept that into a UTF-8 database) * Removed sleep(1) calls from script/production, they don't seem to be useful * Added an index for the type column in `blocks` table. (ActionItem1243)