Merge branch 'ml-try-find-element-before-waiting-for-requests' into 'master'

Try to find elements before waiting for requests

See merge request gitlab-org/gitlab!50919

qa/qa/page/base.rb

@@ -188,19 +188,31 @@ module QA
       end
 
       def has_element?(name, **kwargs)
-        wait_for_requests(skip_finished_loading_check: !!kwargs.delete(:skip_finished_loading_check))
-
         disabled = kwargs.delete(:disabled)
+        original_kwargs = kwargs.dup
+        wait = kwargs.delete(:wait) || Capybara.default_max_wait_time
+        text = kwargs.delete(:text)
+        klass = kwargs.delete(:class)
 
-        if disabled.nil?
-          wait = kwargs.delete(:wait) || Capybara.default_max_wait_time
-          text = kwargs.delete(:text)
-          klass = kwargs.delete(:class)
-
-          has_css?(element_selector_css(name, kwargs), text: text, wait: wait, class: klass)
-        else
-          find_element(name, kwargs).disabled? == disabled
+        try_find_element = ->(wait) do
+          if disabled.nil?
+            has_css?(element_selector_css(name, kwargs), text: text, wait: wait, class: klass)
+          else
+            find_element(name, original_kwargs).disabled? == disabled
+          end
         end
+
+        # Check for the element before waiting for requests, just in case unrelated requests are in progress.
+        # This is to avoid waiting unnecessarily after the element we're interested in has already appeared.
+        return true if try_find_element.call(wait)
+
+        # If the element didn't appear, wait for requests and then check again
+        wait_for_requests(skip_finished_loading_check: !!kwargs.delete(:skip_finished_loading_check))
+
+        # We only wait one second now because we previously waited the full expected duration,
+        # plus however long it took for requests to complete. One second should be enough
+        # for the UI to update after requests complete.
+        try_find_element.call(1)
       end
 
       def has_no_element?(name, **kwargs)