Commit Graph

7 Commits

Author SHA1 Message Date
Sérgio Saquetim
64c9cded17
DEV: Added visible/hidden methods on the SelectKit page component (#23018) 2023-08-08 18:45:04 -03:00
Keegan George
c74c90bae5
DEV: Show form templates in the composer (#21190) 2023-05-29 14:47:18 -07:00
Joffrey JAFFEUX
9321299caa
DEV: uses select-kit component for interface spec (#21790) 2023-05-29 10:48:59 +02:00
Alan Guo Xiang Tan
7fdede9f0f
DEV: Speed up slow system tests (#21803)
What is the problem?

Prior to this change, we had a `has_css?(context + ":not(.is-expanded)"`
check when using the select-kit component page object. The problem here
is that this check will end up waiting the full capybara default wait
time if the select-kit has already been expanded. It turns out that we
were calling this check alot of times when the select-kit has already
been expanded resulting in many tests waiting the full default wait
time.

What is the fix?

The fix here is to specify the `wait: 0` option such that we do not wait
and fundamentally, there is no need for us to wait at all here.
2023-05-29 13:31:02 +08:00
Joffrey JAFFEUX
9392bd0f79
DEV: fix tag synonyms flakey specs (#21787)
The fix use the SelectKit component in the spec and improves reliability of SelectKit component to ensure expanded/collapsed state effectively set/present.

Multiple lines have also been removed as they are not necessary, eg: 

- check button is present
- click button

The check is un-necesssary as we won't find the button on click if not present. This kind of checks are only necessary when another element has to be present before the button is show, eg:

- check modal is displayed
- click button

FInally this commit changes the way the SelectKit component initializes component and now uses a css selector instead of a finder, it ensures we are always getting fresh object and allows to build complete selectors: ".context[data-id=1].foo:enabled"
2023-05-28 15:35:55 +02:00
Joffrey JAFFEUX
2535381f44
FIX: ensures tag notification level is changed (#21106)
Following a change in e9f7262813 which prevents the notification level to be returned from the update endpoint, the model couldn't update itself. This commit makes the update manually and adds a test to prevent future regressions.

Note we could also change the backend endpoint, but this should work correctly with minimum risk.
2023-04-17 10:48:41 +02:00
David Taylor
b5721b7b4f
FIX: default_list_filter = none navigation and preloading (#20641)
When a category has default_list_filter=none, there were a number of issues which this commit resolves:

1. When using the breadcrumbs to navigate a `default_list_filter=none` category, adding a tag filter would not apply the no-subcategories filter, but the subcategories dropdown would still say 'none'. This commit adjusts `getCategoryAndTagUrl` so that `/none` is added to the URL

2. When landing on `/tags/c/{slug}/{id}/{tag}`, for a default_list_filter=none category, it would include subcategories. This commit introduces a client-side redirect to match the behavior of `/c/{slug}/{id}`

3. When directly navigating to `/c/{slug}/{id}`, it was correctly redirecting to `/c/{slug}/{id}/none`, BUT it was still using the preloaded data for the old route. This has been happening since e7a84948. Prior to that, the preloaded data was discarded and a new JSON request was made to the server. This commit restores that discarding behavior. In future we may want to look into making this more efficient.

System specs are introduced to provide end-end testing of this functionality
2023-03-14 10:46:05 +00:00