filament_id v4.1: resolve retired QD_* ids in the Qidi box miss path

QidiPrinterAgent composes QD_<series>_<vendor>_<typeidx> setting ids from
device enums and previously degraded any slot whose id matched no visible
preset to generic-by-type. Plan v4.2 retires every QD_* preset id to the
succession ledger, so insert the ledger walk between the direct match and
the generic fallback: a composed QD_* id now forwards to its family's
minted OF* successor exactly like every other retired id.

Behavior-neutral until the ids are actually retired (the ledger holds no
QD_* keys yet, and live QD_* presets still match directly).

Also documents the code<->ledger lockstep on the hardcoded non-numeric-
series fallback table (QD_1_0_1/_11/_41/_50) and adds a C++ test pinning
that the succession walk is key-format agnostic.

Gates: libslic3r_tests [filament_id] 14 assertions green; libslic3r_gui
compiles.

tests/libslic3r/test_filament_id_succession.cpp

@@ -34,6 +34,14 @@ TEST_CASE("filament_id succession follows forwarding chains", "[Preset][filament
         CHECK(follow_filament_id_succession("OFchainB", forwards) == "OFlive00");
         CHECK(follow_filament_id_succession("OFchainC", forwards) == "OFlive00");
     }
+
+    // QidiPrinterAgent resolves device-composed QD_* protocol ids through this same walk;
+    // the map key format carries no OF assumption.
+    SECTION("device-protocol QD_* keys forward like any other id") {
+        const std::map<std::string, std::string> qd = {{"QD_2_1_11", "OFnew001"}};
+        CHECK(follow_filament_id_succession("QD_2_1_11", qd) == "OFnew001");
+        CHECK(follow_filament_id_succession("QD_2_1_12", qd).empty());
+    }
 }
 
 // Cycles are a ledger-validation error (script check 9) and never ship, but the client