Cannot sort a large group of filters items when using the Exposed Grouped information

Wondering if we can make it dynamic, say based on the number fields.

Good idea @swentel ... at that point we should have all fields available ...

Updating the issue summary to make it clear. Also providing the patch with the dynamic version.

Many thanks for the patch, it would be great to get this committed.

We will need automated test coverage for a bugfix for this.

Also, the existing patch fails, so should not be marked RTBC. Once it is passing and has tests, it should also have test runs queued against SQLite and Postgres.

Thanks!

New patch now with tests.

Manually tested this by adding a 4th filter option, and the delta grows with every added group.

Before:

After:

One small thing:

+++ b/core/modules/views/tests/src/FunctionalJavascript/Plugin/views/Handler/GroupedExposedFilterTest.php
@@ -81,6 +81,25 @@ public function testGroupedFilterValuesUI() {
+    $web_assert->assertWaitOnAjaxRequest();

should use $web_assert->waitForField('options[group_info][group_items][4][title]');. Less random failure prone and clearer about what we are expecting to happen.

Thanks @Lendude

We got a fix and some nice new coverage for a spot that can really use some.

Committed 2e02af7 and pushed to 8.4.x. Thanks!

I see this as a non-disruptive bug fix and therefore this could be considered to be backported to 8.3.x.

The one question that crossed my mind when I read the final patch was that I had thought we had a render element that handled automatically sizing this, but @lauriii said he checked for that as well and we don't. He also said that this matches other usecases.

The other thought that crossed my mind was scalability if the count was very large (giant select elements crushing the page), but this should be taken care of automatically by the fix from #1346760: Add a scalable weight select element. (#1518182: Improve #weight select element behavior based on the number of items it turns out is still open, but out of scope here.) So we are good on that front as well.

So this patch makes sense and I think is safe/reasonable for backport to the RC. Thanks @lauriii!

Thanks @xjm for confirming. Committed 94d1a3b and pushed to 8.3.x.