implement fetchWithAbort and use it for partial page load, mappings & alphabetical index

osma commented 6 months ago

Reasons for creating this PR

Trying to address the AJAX race conditions related to partial page load (#1561).

This PR defines a new JS function fetchWithAbort that wraps the fetch method but also keeps track of pending fetches (per category - there can be different categories) and aborts them early when it's called again.

Link to relevant issue(s), if any

Closes #1561

Description of the changes in this PR

define function fetchWithAbort
use fetchWithAbort when performing partial page load and also when loading mappings for the concept page
use fetchWithAbort when loading entries for the alphabetical index

Known problems or uncertainties in this PR

The PR doesn't interact with the search field autocomplete widget, which is still WIP in PR #1589. I think the fetchWithAbort function would be useful there as well.
Should there be Cypress tests? How would they work?

Checklist

[x] phpUnit tests pass locally with my changes
[ ] I have added tests that show that the new code works, or tests are not relevant for this PR (e.g. only HTML/CSS changes)
[x] The PR doesn't reduce accessibility of the front-end code (e.g. tab focus, scaling to different resolutions, use of .sr-only class, color contrast)
[x] The PR doesn't introduce unintended code changes (e.g. empty lines or useless reindentation)

codecov[bot] commented 6 months ago

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 70.58%. Comparing base (f22c504) to head (fcf5be7). Report is 5 commits behind head on main.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #1609 +/- ## ========================================= Coverage 70.58% 70.58% Complexity 1647 1647 ========================================= Files 32 32 Lines 4321 4321 ========================================= Hits 3050 3050 Misses 1271 1271 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

osma commented 6 months ago

You are right @UnniKohonen . I will increase the timeouts. I think it makes sense to do this on this PR branch rather than makkng a separate PR because the timeout problem is closely related even though it's probably not caused by new changes in this PR.