fix(flatbar): stop the swap optimization trading identical copies

TryImprove kept a swap when the trial run looked fuller than before, but
the trial charges each type's self-nesting gain again even when the run
already holds that type. Trading a placed copy for an identical unplaced
one therefore looked fuller by the gain and repeated forever: FlatBar ran
into the benchmark's 5-minute timeout on tabbed_cog_layout once the
all-edges MBR fix gave one part new rotations (it finished in 7 s before).

The swap is now judged on the merged run, so every kept swap really
leaves the run fuller and the loop terminates. Swaps the old test kept
that do fill the run are still kept.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

OpenNest.Engine/NestingEngines/FlatBar/BarPacker.cs

@@ -174,7 +174,13 @@ internal static class BarPacker
                     trial.Items.Add(next);
                 }
 
-                if (trial.Remaining < before)
+                // Judge the swap on the merged run: the trial charges gains again for types the
+                // run already holds, so trading a copy for an identical one looks fuller and the
+                // swap would repeat forever.
+                var merged = new PackedSegment(segment.Capacity);
+                merged.Items.AddRange(segment.Items);
+                merged.Items.AddRange(trial.Items);
+                if (merged.Remaining < before)
                 {
                     remaining.Add(removed);
                     segment.Items.AddRange(trial.Items);