Fix extruder-map bounds, popover lifetime, and arrange thread safety

Review findings on the IDEX/IQEX parallel printing code, all in paths the
feature owns.

- physical_extruder_map lookups used ConfigOptionVector::get_at(), which
  clamps an out-of-range index to values.front() rather than reporting a
  miss. The map holds one entry per nozzle while filament ids index slots,
  and nothing caps the slot count at the nozzle count, so a project authored
  with more filaments than the printer has extruders silently addressed the
  primary's head: pressure advance pinned to the wrong carriage, and
  skip-primary loops suppressing whichever head sat at pem[0]. Bounds-check
  at all four sites and treat the miss as "no mapping" (-1). Covered by a new
  imex_pem_tool_for test; the header note now warns against get_at here.

- IMEXFilamentPickerPopover leaked a top-level window per ghost click:
  wxPopupTransientWindow::Dismiss() only hides, and never reaches OnDismiss().
  Destroy from an OnDismiss() override and dismiss the picker through
  DismissAndNotify(), which is the path a successful pick takes.

- ArrangeJob read PartPlate's IMEX zone cache from the worker thread, where
  a cache miss rebuilds GLModel members with no GL context current while the
  GUI thread may be painting them. Snapshot the zones in prepare(), on the
  main thread, already converted to plate-local coordinates.

- The mode grid anchored its row window to the Primary's gantry row. A window
  as tall as the grid can only start at row 0, so this drew tiles for tools
  that do not exist and hid real ones. Render the whole grid instead; a
  Primary outside it is a data problem the zone layout already reports.

- Build the mode tooltip from one format string rather than two catalog
  fragments concatenated around a runtime value, so translators can move the
  mode name within the sentence, and register IMEXModesCtrl.cpp for string
  extraction.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

src/libslic3r/GCode.cpp

@@ -3418,9 +3418,14 @@ void GCode::_do_export(Print& print, GCodeOutputStream &file, ThumbnailsGenerato
         const auto plate_head_map = parse_imex_head_filament_map(
             print.objects().front()->config().imex_head_filament_map.value);
         const ConfigOptionInts& pem = print.config().physical_extruder_map;
-        const int primary_physical = pem.values.empty()
-            ? -1
-            : pem.get_at((int)initial_extruder_id);
+        // Bounds-check rather than get_at(), which clamps out-of-range to values.front(). The
+        // clamped value is used as a skip-primary sentinel below, so a filament id past the end
+        // of the map would suppress whichever head sits at pem[0]. -1 matches no head.
+        const int primary_physical =
+            ((int) initial_extruder_id >= 0 &&
+             (int) initial_extruder_id < (int) pem.values.size())
+                ? pem.values[(int) initial_extruder_id]
+                : -1;
         for (int logical : imex_secondary_logical_slots(
                 get_imex_active_tools(print), primary_physical, plate_head_map, pem))
             if (logical < (int)is_extruder_used.size())
@@ -3910,7 +3915,15 @@ void GCode::_do_export(Print& print, GCodeOutputStream &file, ThumbnailsGenerato
             // loop can skip the primary head (which emitted PA via the normal path).
             // Then pem-invert each active physical head back to its first routed filament
             // for the PA setting lookup. Guarded on non-empty pem above.
-            const int initial_physical = m_config.physical_extruder_map.get_at((int)initial_extruder_id);
+            // Bounds-check rather than get_at(), which clamps out-of-range to values.front():
+            // a clamped initial_physical would make this loop skip whichever active head equals
+            // pem[0], leaving that head with no PA at all. Same reasoning as the second-layer
+            // temperature loop below. -1 matches no head, so every active head is emitted.
+            const int initial_physical =
+                ((int) initial_extruder_id >= 0 &&
+                 (int) initial_extruder_id < (int) m_config.physical_extruder_map.values.size())
+                    ? m_config.physical_extruder_map.values[(int) initial_extruder_id]
+                    : -1;
             for (int tool_idx : get_imex_active_tools(print)) {
                 // Unlike the second-layer temperature loop, the primary is skipped here:
                 // set_extruder() above already emitted its PA with the pem tool qualifier.
@@ -5937,9 +5950,17 @@ LayerResult GCode::process_layer(
             // transition at all. `tool_idx` is physical; the printing head uses this layer's
             // own filament, the parallel carriages resolve through the head map.
             const int num_filament_columns = (int)print.config().nozzle_temperature.values.size();
-            const int initial_physical = m_config.physical_extruder_map.values.empty()
-                ? -1
-                : m_config.physical_extruder_map.get_at((int)first_extruder_id);
+            // Bounds-check rather than get_at(): get_at() clamps to values.front(), which would
+            // make initial_physical the primary's head for any first_extruder_id past the end of
+            // the map. A secondary that happens to sit on that head would then take the
+            // "initial" branch below and be given the wrong filament's transition temperature,
+            // while never receiving its own. -1 matches no tool, so every head takes the
+            // resolved path instead.
+            const int initial_physical =
+                ((int) first_extruder_id >= 0 &&
+                 (int) first_extruder_id < (int) m_config.physical_extruder_map.values.size())
+                    ? m_config.physical_extruder_map.values[(int) first_extruder_id]
+                    : -1;
             for (int tool_idx : get_imex_active_tools(print)) {
                 const int logical = (tool_idx == initial_physical)
                     ? (int)first_extruder_id