After applying this patch, phpstan throws a bunch of additional output here, like Class PhabricatorAuthPasswordQuery extends generic class PhabricatorCursorPagedPolicyAwareQuery but does not specify its types: R and such. Probably that is totally fine. Just saying, as a phpstan newbie.

Put minus icon back in

Pasting full unit test output here before garbage collection kicks in:

Assertion failed, expected values to be equal (at PhutilSearchQueryCompilerTestCase.php:303): Function compilation of query: cat title:="Meow Meow"
Expected vs Actual Output Diff
--- Old Value
+++ New Value
@@ -1,18 +1 @@
-Array
-(
-    [0] => Array
-        (
-            [0] => 
-            [1] => and
-            [2] => cat
-        )
-
-    [1] => Array
-        (
-            [0] => title
-            [1] => exact
-            [2] => Meow Meow
-        )
-
-)
-
+false

1. Go to http://phorge.localhost/search/query/advanced/
2. In the Query field enter z a a "a phrase" a a a a ~a -b =foo title:u which includes duplicate values
3. Click Search button

Don't output "1 Terms" but "x Term(s)"

Before:

Hi and welcome, thanks for bringing this up! I guess you could Edit Question and change the View Policy of this Ponder Question to either "Trusted Contributors", or something more restrictive via the "Custom Policy" option, like administrators and subscribers, or such.

Looks like https://we.phorge.it/source/phorge/browse/master/src/infrastructure/query/policy/PhabricatorCursorPagedPolicyAwareQuery.php$2118-2125 and https://we.phorge.it/source/phorge/browse/master/src/infrastructure/query/policy/PhabricatorCursorPagedPolicyAwareQuery.php$2499-2506 already do this silently but only for JOIN clauses.

Hi, please explain specifically what "support for large monitors" means to you.
Large monitors are "supported" as all the content on a Phorge page is correctly displayed, layouted, readable on large monitors, as your screenshot proves. :)

Hi, could you please describe actual problems due to the current layout? Free space is not an issue per-se, for example limiting the content width makes sense.
I'm also a bit wondering why you would put/use the browser full screen when having such a ratio?

diff --git a/src/view/phui/PHUIObjectBoxView.php b/src/view/phui/PHUIObjectBoxView.php
index dce9d57568..13a0a3d061 100644
--- a/src/view/phui/PHUIObjectBoxView.php
+++ b/src/view/phui/PHUIObjectBoxView.php
@@ -318,6 +318,10 @@ final class PHUIObjectBoxView extends AphrontTagView {
       $lists,
       $this->table,
       $pager,
+      // When the anchor link scrolled us below the $showhide search parameters
+      // box, repeat the error message hidden and unseen at the top of the page
+      ($this->showHideOpen == true ? $this->formErrors : null),
       $this->renderChildren(),
     );

^ This would simply be ugly repeating of the error message below the search params box on http://phorge.localhost/maniphest/query/advanced/.
For testing the patch behavior with a results list, replace the line
($this->showHideOpen == true ? $this->formErrors : null),
with
($this->showHideOpen == true ? $header : null),

Before and after Chromium Lighthouse a11y result comparison of some Phorge pages for patches I'm going to attach.
(Goodhart's Law applies, as with any misguiding incentives.)

Also nothing seems to use these functions?:

[acko@fedora phorge]$ grep -r compileLiteralQuery .
./phorge/src/applications/search/compiler/__tests__/PhutilSearchQueryCompilerTestCase.php:          $literal_query = $compiler->compileLiteralQuery($tokens);
./phorge/src/applications/search/compiler/PhutilSearchQueryCompiler.php:  public function compileLiteralQuery(array $tokens) {
[acko@fedora phorge]$ grep -r compileStemmedQuery .
./phorge/src/applications/search/compiler/__tests__/PhutilSearchQueryCompilerTestCase.php:          $stemmed_query = $compiler->compileStemmedQuery($tokens);
./phorge/src/applications/search/compiler/PhutilSearchQueryCompiler.php:  public function compileStemmedQuery(array $tokens) {

Nice find. Applying this patch removes about 250 errors in phpstan output here.

Running phpstan analyse, before this patch I get Call to an undefined method ArcanistWorkflow::runWorkflow(), after this patch I get Method ArcanistWorkflow::runWorkflow() invoked with 1 parameter, 0 required, both in src/workflow/ArcanistWorkflow.php:227. That looks like an improvement indeed.

Yes to the idea however a related elephant in the room (IMHO) are those non-standard types like list or map or wild being replaced, no clue if that has any effects on...something™ (Diviner docs? Code?) as I haven't managed to understand the consequences of code in https://we.phorge.it/source/arcanist/browse/master/src/parser/PhutilTypeSpec.php yet. (Technically speaking map and list seem to be just arrays, indeed.)