| Age | Commit message (Collapse) | Author |
|
|
|
Use policies for calendar sharing. Refs #5325
|
|
|
|
This will allow us to compile assets with a null_db adapter.
|
|
|
|
|
|
|
|
|
|
Refs #5291 @5m
|
|
Remove the "!archived?" condition which makes no sense here.
|
|
Fix integration_spec_helper after rebase
|
|
- Refactor specs helper to be more generic
- Write missing specs for the Workbenches#show view
We now have failing tests highlighting the bug
|
|
Use policies to determine if a user is allowed to share a calendar,
instead of a hardcoded string
|
|
This method must return the keys of all `control_attributes` in
subclasses of `ComplianceControl`.
Now that we're not using `hstore_accessor`, this breaks. Use
`stored_attributes` to get the keys since we create virtual attributes
using Rails' `store_accessor`.
Refs #5316
|
|
af83/5291-disable-useless-checkboxes-on-referentials-index
5291 disable useless checkboxes on referentials index
|
|
Copy Cédric's JSON to Hstore migration for the `down` here to give us
the ability to roll back.
I'm guessing the solution probably comes from here:
https://stackoverflow.com/questions/28315948/casting-json-to-hstore-in-postgres-9-3/28319295#28319295
Refs #5316
|
|
Fix integration_spec_helper after rebase
|
|
- Refactor specs helper to be more generic
- Write missing specs for the Workbenches#show view
We now have failing tests highlighting the bug
|
|
Add missing styles for disabled checkboxes
|
|
Update the view to match the actual controller behaviour.
|
|
Update TableBuilderHelper to allow the `selectable` param to be a
lambda, thus allowing us to have row-based granularity.
|
|
- Refactor specs helper to be more generic
- Write missing specs for the Workbenches#show view
We now have failing tests highlighting the bug
|
|
We're no longer using this gem. See:
* b54daa7b2d7925131572e5e5d5181cf3515a6057
* 993650cd1739a3c4c2dd6a4dc01110b57ef88f36
Refs #5316
|
|
The places we were using `hstore_accessor` have now been converted to
Postgres JSON fields. Thus we no longer need `hstore_accessor` in order
to add accessors for the fields.
Instead we can use Rails' built-in support for Postgres JSON fields and
create accessors for our JSON hash keys with `store_accessor`. We get
rid of the data types because `store_accessor` doesn't work like that
and JSON already types our values with primitive data types.
Finally, add some extra validation tests now that the vaildation for our
JSON fields should work.
Refs #5316
|
|
Make these columns JSON fields again. Do this because of bugs trying to
validate types with `hstore_accessor`
(https://github.com/devmynd/hstore_accessor/issues/78).
With a JSON field, types are preserved and we don't have to deal with
casting back and forth.
Previously we had converted to Hstore because the Java IEV application
used an older ORM version (maybe Hibernate) that didn't support the
Postgres JSON type. Because of the validation problems with Hstore, the
Java application has been updated to support JSON fields.
Thanks to this Stack Overflow post from 'a_horse_with_no_name' that
describes how to convert from Hstore to JSON types in Postgres:
https://stackoverflow.com/questions/33732529/postgres-how-to-convert-hstore-to-json-datatypes#comment55234342_33732529
Converting in this direction seems to me to be much safer because we're
not losing type information. All the values should be hashes of strings,
which should convert without issue into JSON.
Refs #5316
|
|
#5307 Remove update & create links
|
|
belongs_to referential
|
|
5287 clean links in stop areas index
|
|
|
|
Refs #5307
|
|
For calendars panel, instead of hardcoded string
|
|
|
|
- Get rid of ActiveSupport::Concern in the spec helpers
- Rename misnamed spec files
|
|
|
|
- Fix specs on connections_links/show
- Update pundit view specs helper to use the current referential when it
has already been defined
|
|
Add specs for the view.
Refactor to come
|
|
Refs #5287
|
|
|
|
- add missing policies
- update permissions translator
- update views to check for the permission
- update views helper to check for the permission
- uipdate controllers to check for the permission
|
|
Fix integration_spec_helper after rebase
|
|
Add missing styles for disabled checkboxes
|
|
Update the view to match the actual controller behaviour.
|
|
Update TableBuilderHelper to allow the `selectable` param to be a
lambda, thus allowing us to have row-based granularity.
|
|
- Refactor specs helper to be more generic
- Write missing specs for the Workbenches#show view
We now have failing tests highlighting the bug
|
|
|
|
Refs #5305
|
|
When a "duplicate referential" action is initiated, show a flash message
to let users know that it started and is in progress (and will take a
while since it's an asynchronous task).
Refs #5317
|
|
... instead of user's username in the top bar
|
|
Call `#t` directly since we can and because the rest of the controller
does this without the `I18n.` prefix.
Refs #5317
|
|
* Luc specified a new flash message string to use here. Update the old
one with the new version.
* Move the translation key to `notice.referentials.validate` because we
have multiple "operation_in_progress" messages (this one for
validation, and another for referential cloning/duplication).
* Add a test to confirm that the flash message is set.
Refs #5317
|