Views field rendering on multiple instances of one field is broken

Did maybe #2534780: Fatal error rendering fields using an optional relationship caused this break? I'm curious about that.

I think we should expand the test coverage in core/modules/views/src/Tests/Handler/FieldFieldTest.php in this case
What exactly did you asked for in IRC, actually a test or just a pointer?

+++ b/core/modules/views/src/Tests/Handler/FieldFieldTest.php
@@ -376,26 +377,44 @@ public function testComplexRender() {
+    $this->assertEqual(format_date($this->testUsers[4]->getCreatedTime(), 'custom', 'Y'), $executable->getStyle()->getField(4, 'created'));

Even format_date() is not marked as deprecated it would be nicer if we could use \Drupal::service('date.formatter')->format() directly

Tested this, this fixes the issue. Just some nitpicks

1. +++ b/core/modules/views/src/Entity/Render/EntityFieldRenderer.php
   @@ -231,7 +233,7 @@ protected function buildFields(array $values) {
   +          foreach ($display_fields['field_ids'] as $field_id => $field) {
   

   we are not using $field_id here so let's cut it.

2. +++ b/core/modules/views/src/Tests/Handler/FieldFieldTest.php
   @@ -371,31 +372,50 @@ public function testComplexExecute() {
   +
   ...
   +
   ...
   +
   

   Let's remove these double enters

Fixed those nitpicks

Oops forgot to rebase

Manually tested this and 'multiple field instances of the same field' in this case means 'more then 2' (2 instances don't show this bug, 3 do).

Using two different fields (UI wise) with the same field name and entity type (so combining content: 'authored on' and content revision:'authored on') also cause this bug to show itself if you have a total of more then 2 instances combined.

The patch fixes this.

All raised issues in previous comments are fixed. Looks good to me.

Added beta eval.

Random test failure, back to RTBC

+++ b/core/modules/views/tests/modules/views_test_config/test_views/views.view.test_field_field_complex_test.yml
@@ -58,6 +58,45 @@ display:
+        created_2:
+          id: created_1

The id looks incorrect - I guess it is being ignored but rather than fix on the commit I think we need a fix and test.

Thanks, totally missed that.

EDIT: Just saw I accidentally set a .patch extension on that interdiff, oops.

Issue in #19 addressed, and if I understand @alexpott correctly, the issue with the faulty/duplicate id being ignored is out of scope for this issue, so back to RTBC.

Random fail. Back to RTBC.

Committed 7a8cb7a and pushed to 8.0.x. Thanks!