mbproxy: strip historical phase/wave/plan references from source comments

Comments described the *history* of how the code arrived (phase numbers,
wave IDs, review IDs, dated TODOs) instead of what it does today. That
scaffolding rotted as the codebase evolved. Cleaned 60 source files +
.gitignore; behaviour unchanged (387/387 tests still pass).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

mbproxy/tests/Mbproxy.Tests/Bcd/BcdCodecTests.cs

@@ -10,8 +10,8 @@ namespace Mbproxy.Tests.Bcd;
 /// NOTE on allocation profile:
 ///   BcdCodec is a purely static class operating on value types (ushort, int, tuples).
 ///   It allocates only when constructing exception objects (the error path), never on
-///   the success path. TryGet / hot-path decode callers in Phase 04 will be
-///   allocation-free for valid BCD registers.
+///   the success path. TryGet / hot-path decode callers are allocation-free for valid
+///   BCD registers.
 /// </summary>
 [Trait("Category", "Unit")]
 public sealed class BcdCodecTests
@@ -45,8 +45,8 @@ public sealed class BcdCodecTests
     }
 
     /// <summary>
-    /// Phase 12 (W3 test gap #11) — locks the boundary contract for the `(uint)value > Max16`
-    /// range check. `int.MinValue` cast to `uint` becomes `0x80000000`, which is well above
+    /// Locks the boundary contract for the `(uint)value > Max16` range check.
+    /// `int.MinValue` cast to `uint` becomes `0x80000000`, which is well above
     /// `Max16` (= 9999), so the throw fires cleanly without arithmetic surprise. Prevents
     /// regressions if the bounds check is ever rewritten with a two-sided int comparison
     /// that would underflow on extreme negatives.

mbproxy/tests/Mbproxy.Tests/Bcd/BcdTagMapBuilderTests.cs

@@ -99,9 +99,9 @@ public sealed class BcdTagMapBuilderTests
     }
 
     /// <summary>
-    /// Phase 12 (W2.11) — duplicates within Global itself are now detected
-    /// pre-collapse and produce a DuplicateAddress error. (Before W2.11 the input
-    /// dictionary silently collapsed to last-write-wins, leaving the validator dead.)
+    /// Duplicates within Global itself are detected pre-collapse and produce a
+    /// DuplicateAddress error. (A naive input dictionary would silently collapse
+    /// to last-write-wins, leaving the validator dead.)
     /// </summary>
     [Fact]
     public void Build_DuplicateAddressInGlobal_ReturnsDuplicateAddressError()
@@ -122,9 +122,9 @@ public sealed class BcdTagMapBuilderTests
     }
 
     /// <summary>
-    /// Phase 12 (W2.11) — duplicates within the per-PLC Add list itself are now detected
-    /// pre-collapse. (Cross-list collisions Global vs Add remain the legitimate width-
-    /// override pattern and are NOT errors — see the next test.)
+    /// Duplicates within the per-PLC Add list itself are detected pre-collapse.
+    /// (Cross-list collisions Global vs Add remain the legitimate width-override
+    /// pattern and are NOT errors — see the next test.)
     /// </summary>
     [Fact]
     public void Build_DuplicateAddress_Within_AddList_ReturnsDuplicateAddressError()
@@ -147,9 +147,9 @@ public sealed class BcdTagMapBuilderTests
     }
 
     /// <summary>
-    /// Phase 12 (W2.11) — same-address entries appearing in BOTH Global AND Add are
-    /// the documented width-override pattern (design.md "Hybrid tag resolution"). They
-    /// must NOT be flagged as duplicates; Add wins.
+    /// Same-address entries appearing in BOTH Global AND Add are the documented
+    /// width-override pattern (design.md "Hybrid tag resolution"). They must NOT
+    /// be flagged as duplicates; Add wins.
     /// </summary>
     [Fact]
     public void Build_AddOverridesGlobalAtSameAddress_NoDuplicateError_AddWins()