Patch Tablefield for header row orientation per widget

Here's an updated patch that also changes the CSS to make it use the selected "Header orientation"

Actually, that one wasn't quite right. On some themes it would lead to "vertical" not highlighting the first column in the first row. This one is better.

Here's a Travis build with both patches:

https://travis-ci.org/github/panopoly/panopoly/builds/680706682

Hm. The patches fail testing on Travis :-(

Hm. I'm not sure why this update hook is required when the tablefield patch does a very similar one:

+/**
+ * Move 'Header orientation' from formatter settings to default field value.
+ */
+function tablefield_update_7008() {
+  $instances = field_info_instances();
+  foreach ($instances as $entity_type => $entities) {
+    foreach ($entities as $bundle => $fields) {
+      foreach ($fields as $field_name => $instance) {
+        $field_info = field_info_field($field_name);
+        if ($instance['display']['default']['type'] == 'tablefield_default') {
+          if (isset($instance['display']['default']['settings']['header_orientation'])) {
+            // Copy 'Header orientation' from field formatter into the
+            // fields default value.
+            $instance['default_value'][0]['tablefield']['rebuild']['header_orientation'] = $instance['display']['default']['settings']['header_orientation'];
+            unset($instance['display']['default']['settings']['header_orientation']);
+            field_update_instance($instance);
+          }
+        }
+      }
+    }
+  }
+}

See https://www.drupal.org/files/issues/2020-04-22/tablefield-header-orienta...

Here's a new Travis build of this patch, though: https://travis-ci.org/github/panopoly/panopoly/builds/681217048

Here's a new patch that is just a drush fu panopoly_widgets after running the tablefield update. It removes the update hook from #8 because I'm hoping that's not actually necessary. Let's see how this does on Travis.

Gah, that was a profile patch but only of panopoly_widgets which is the worst of both worlds. :-) Here's just a patch for panopoly_widgets so the two patches will get merged when the tests run.

EDIT: Here's the travis build: https://travis-ci.org/github/panopoly/panopoly/builds/686265769

That patch also failed, but I think it's because the tests aren't running the individual .make files, but the overall profile one. So, here's a profile patch that includes both the panopoly_widgets and panopoly_test changes - let's see how this does!

EDIT: Here's a new Travis build: https://travis-ci.org/github/panopoly/panopoly/builds/686280497

That patch failed too. :-( It looks like there's nothing in Panopoly that's setting scope="row" on those row headers. That's probably a tablefield bug? Anyway, I've adjusted the tests here just so we can get what we have committed.

EDIT: Here's the Travis build: https://travis-ci.org/github/panopoly/panopoly/builds/686300049

This finally passed! So, I just merged it. :-)