18 Feb, 2016
3 commits
-
Currently translated at 100.0% (24 of 24 strings)
-
Currently translated at 100.0% (53 of 53 strings)
-
Currently translated at 100.0% (2124 of 2124 strings)
17 Feb, 2016
1 commit
16 Feb, 2016
2 commits
11 Feb, 2016
11 commits
-
It is redundant, as the feature that it is supposed to test can be completed tested with less expensive tests.
-
Improve comment paragraph performance Also fix a broken unit test See merge request !788
-
Fixes crash when processing suggest an article task - Removes an old line that was left behind during merge of support for different types of article. It was overriding the correct decision made by #article_type. - Defines SuggestArticle#author_name the same as the sender so noosfero task page doesn't crash in case the processing of the task throws an error, since it will try to show the author name in the task error page. Fixes #166 See merge request !786
-
fix signup with custom_fields Fixes bugs with signup with required custom fields, as well as required custom values not being saved for new communities. See merge request !783
-
- Removes an old line that was left behind during merge of support for different types of article. It was overriding the correct decision made by #article_type. - Defines SuggestArticle#author_name the same as the sender so noosfero task page doesn't crash in case the processing of the task throws an error, since it will try to show the author name in the task error page. Fixes #166
05 Feb, 2016
4 commits
-
Fix feed that has never been fetched succesfully This avoid a crash when container.fetched_at is `nil`. See merge request !782
-
Signed-off-by: Joenio Costa <joenio@colivre.coop.br> Signed-off-by: Marcos Ramos <ms.ramos@outlook.com> Signed-off-by: Marcos Ronaldo <marcos.rpj2@gmail.com>
-
Fix selenium onClick link web step The previous fix at 8064b0da19cedbb3d8089e9fd62e208857aea0ef did not expect links with onClick actions which was specially breaking the category_block feature. See merge request !785
-
The previous fix at 8064b0da19cedbb3d8089e9fd62e208857aea0ef did not expect links with onClick actions which was specially breaking the category_block feature.
04 Feb, 2016
1 commit
-
Depending on timing trying to fill a tinyMCE control will fail with `tinymce.editors.article_body is undefined` (Selenium::WebDriver::Error::JavascriptError), probably because at the point we try to interact with the editor it is still being loaded by the browser. Because of that in case of failure we sleep for 1 second and try again at most 5 times.
02 Feb, 2016
5 commits
-
This was previously done to avoid crashes on development environment, but some selenium tests, for unknown reasons, produces requests where Environment.default returns nil even though all steps passes.
-
Tinymce fields are not filled normally as other fields. In order to fill them, had to use execute_script to set the content of the editor object. This fix might be necessary in other tests as well.
-
Let's isolate those bits of logic under meaningful names to make the code easier to understand.
01 Feb, 2016
3 commits
-
Signed-off-by: Tallys Martins <tallysmartins@yahoo.com.br>
30 Jan, 2016
1 commit
-
Selenium tests were broken in clicking inside the content selection box. Some HTML structures, such as `<li>` tags nested to `<a>` tags, may not get along with Selenium and the browser interaction. But are perfectly fine for the final user. As we do not want to modify the current behaviour in order to pass tests, here we handle Selenium's error properly so tests can still pass without modifying current code. NOTICE: This only works with selenium-webdriver 2.50.0 and up.
29 Jan, 2016
2 commits
-
Don't know exactly why but adding this make all tests pass.
28 Jan, 2016
7 commits
-
Count distinct - Multiple fields in paginate Fix sql query error with `distinct` into a **COUNT()** database function with multiple fields. The problem was caused by the gem `will_paginate` on version **3.0.5**. In their new version **3.0.7** this bug was fixed!! See merge request !777
-
Set correct header for csrf token The header setted from application.js has to be equal to the key used in verified_request? See merge request !778
-
This will avoid, or at least reduce the chances of, having the CI builds falling over on temporary network failures.