29 Jan, 2016
2 commits
-
Replace url helpers by the path ones
-
This is expected to improve compatibility with reverse-proxies. Still, for full compatibility, regarding redirections, we still may need to set the default url options with the proxy host. Signed-off-by: Rafael Reggiani Manzo <rr.manzo@gmail.com>
22 Jan, 2016
6 commits
-
Hotspot metric configuration creation
-
Calling methods with self improve the readability as it makes clear that it is not a variable but a method! Signed off by: Daniel Miranda <danielkza2@gmail.com>
-
Signed-off-by: Heitor Reis <marcheing@gmail.com>
-
It will never return nil. At most, it will raise a ActiveRecord::RecordNotFound Signed off by: Diego Araújo <diegoamc90@gmail.com>
-
Signed off by: Diego Araújo <diegoamc90@gmail.com>
-
Signed off by: Diego Araújo <diegoamc90@gmail.com>
20 Jan, 2016
23 commits
-
Signed-off-by: Rafael Reggiani Manzo <rr.manzo@gmail.com>
-
Test the path where the provided metric's type is different from the controller's one. Signed-off-by: Rafael Reggiani Manzo <rr.manzo@gmail.com>
-
We were not testing the case when the returned metric_configuration was from a different KalibroConfiguration. This prevents different ownership MetricConfiguration editions and false positives for ownership_authentication helper. Signed off by: Heitor Reis <marcheing@gmail.com>
-
The code is pretty straightforward, just raising an Error in order to force subclasses to implement this. Signed-off-by: Rafael Reggiani Manzo <rr.manzo@gmail.com>
-
There was no test case for failed_action when the action was neither create or update. The coverage was achieved through show action with invalid reading group id. Signed-off-by: Rafael Reggiani Manzo <rr.manzo@gmail.com>
-
It was missing proper mocks for the set_metric! before_action from BaseMetricConfigurationsController. Signed off by: Heitor Reis <marcheing@gmail.com>
-
Also improve error reporting functions
-
Signed off by: Daniel Miranda <danielkza2@gmail.com>
-
Move as much code as possible to BaseMetricConfigurationsController instead of repeating it ad-hoc in the concrete classes. Most of the concrete code has gone away except for some small overrides: - Hotspot's does not have a 'new' action, and therefore can't redirect to it. It also does not accept any MetricConfiguration parameters. - Metric's only accepts 'weight' and 'aggregation_form' MetricConfiguration parameters. It also needs to fetch ReadingGroups to display in the 'choose_metric' action. - Compound's needs to force the Metric type when creating the metric, and must accept parameters for Metric ('name', 'description', 'script', 'scope', 'code') in addition to MetricConfiguration's 'weight'. Additionally: - Update cache clearing to clear caches for Hotspot and Tree configurations - Verify the existence of the KalibroConfiguration and use the object that was fetched whenever necessary, instead of only using the id. - Ensure the KalibroConfiguration and ReadingGroup ids match between the values inside and outside the 'metric_configuration' parameter. Previously it was possible to send one in the URL and update another, which might even have been a security issue. - Allow sending either 'metric_code' or 'metric_name' as a parameter. Both are used in different places in the code, but making them uniform is a job for a later time. Signed off by: Daniel Miranda <danielkza2@gmail.com> -
There will no longer be a @kalibro_configuration_id instance variable, and all the views will use @metric_configuration, with no special cases like Compound previously had. Signed off by: Daniel Miranda <danielkza2@gmail.com>
-
Signed off by: Daniel Miranda <danielkza2@gmail.com>
-
All of it's code is covered by the concrete class tests, and the abstract class tests were nothing like real usage, and will be made totally obsolete with the refactor. Signed off by: Daniel Miranda <danielkza2@gmail.com>
-
- Fix usage of MetricConfiguration models converted directly to hashes as parameters. It didn't reflect how they are actually sent and omitted some necessary parameters like metric_collector_name and metric_name/metric_code, which are now passed. - Fix KalibroConfigurations not being searched by the controller methods. We must be sure they exist, and fail with a 404 if they don't instead of failing only in one of the ownership testing methods. - Update the expectations for cache cleaning to include both Hotspot and Tree MetricConfigurations Signed off by: Daniel Miranda <danielkza2@gmail.com>
-
Signed off by: Daniel Miranda <danielkza2@gmail.com>
-
There are no attributes to be edited on a hotspot metric configuration Signed off by: Diego Araújo <diegoamc90@gmail.com>
-
Now it doesn't redirect to the tree metric configuration creation page after choosing a hotspot metric configuration Adds a bunch of code from removed commit ec4c3ef6f29aa34f3c58819df4f84725aff8bd60 that fixed the view after the modification introduced by 39bc97f568d6989ae98b1f5fd5452515ee54045f. It fixes the view to properly work with the path helper. This is a squash of two commits. Signed off by: Diego Araújo <diegoamc90@gmail.com> Signed off by: Heitor Reis <marcheing@gmail.com>
-
To decide betwen the different routes for tree metrics or hotspot metrics. Signed off by: Rafael Reggiani Manzo <rr.manzo@gmail.com>
-
It has just the prototype for the method so its route works. This route will be useful when creating the upcoming helper method for metric configuration creation form. Signed off by: Heitor Reis <marcheing@gmail.com>
-
This will enable different actions for TreeMetricConfigurations and Hotspot ones. Signed off by: Diego Araújo <diegoamc90@gmail.com>
-
This test is necessary in order to provide proper coverage that allows a refactor. Signed off by: Diego Araújo <diegoamc90@gmail.com>
12 Jan, 2016
3 commits
-
Signed off by: Diego Araújo <diegoamc90@gmail.com>
-
Replace url route helpers by path ones
-
Using URL helpers is not friendly with proxies like Colab. The path helpers have the same effect unless we want sometime to render pages outside Prezento context like embeds or sending emails. Signed off by: Diego Araújo <diegoamc90@gmail.com>
30 Nov, 2015
1 commit
27 Nov, 2015
1 commit
-
Git lab hooks
25 Nov, 2015
2 commits
-
Signed off by: Eduardo Araújo <duduktamg@hotmail.com>
-
GitLab's webhooks depend on status codes from kalibro_processor Signed off by: Rafael Reggiani Manzo <rr.manzo@gmail.com>
23 Nov, 2015
2 commits
-
Signed off by: Daniel Miranda <danielkza2@gmail.com>
-
Gitlab sends the 'X-Gitlab-Event' header for all the webhook requests. Verify it so we reduce the chance that robots call the hooks accidentally. Signed off by: Daniel Miranda <danielkza2@gmail.com>