fix SoftNcRunner GM code normalization.

App/wwwroot/HiAPI-docsite/api/Hi.NcParsers.Syntaxs.FanucSyntaxUtil.html

@@ -228,7 +228,12 @@ Class FanucSyntaxUtil
   </h3>
 
   <div class="markdown level1 summary"><p>G05.1 High-precision contour control (Fanuc AICC / Nano Smoothing).
-Q1 enables, Q0 disables.</p>
+Q1 enables, Q0 disables. Optional R{n} selects the precision /
+smoothness level number (R1..R10) — captured here so the parameter
+does not leak into the standalone Parsing.R tag and so
+<a class="xref" href="Hi.NcParsers.LogicSyntaxs.FanucPathSmoothingSyntax.html">FanucPathSmoothingSyntax</a> can record it
+in the <code>PathSmoothing</code> modal section for bidirectional NC-text
+reconstruction.</p>
 </div>
   <div class="markdown level1 conceptual"></div>
 
@@ -262,7 +267,11 @@ Q1 enables, Q0 disables.</p>
 
   <div class="markdown level1 summary"><p>G43.4: Fanuc TCPM (Tool Center Point Management / RTCP).
 Fanuc-specific — not in <a class="xref" href="Hi.NcParsers.Syntaxs.GenericSyntaxKit.html">GenericSyntaxKit</a>.
-Siemens equivalent: TRAORI. Heidenhain equivalent: M128.</p>
+Siemens equivalent: TRAORI. Heidenhain equivalent: M128.
+TerminateWords intentionally only <code>M</code>: H (offset id) commonly
+appears after move axes or other G modifiers in the same block
+(e.g., <code>G43.4 Z5. H1</code>, <code>G43.4 G54 H1</code>). Using G/X/Y/Z
+would truncate scope before H and lose the offset id.</p>
 </div>
   <div class="markdown level1 conceptual"></div>