Fix eight review findings on the render-bounds diagnostics

Corrects a false comment example, fixes two tests that couldn't detect their
own regressions, adds two more read-back checkpoints around Main_OnCommand so
a drift report self-locates, guards a spurious zero-vs-zero coincidence match,
and softens two sentences that overclaimed cause or defect.

src/core/capture/render_window.cpp

@@ -70,7 +70,11 @@ std::string describeBoundsDrift(double reqStart, double reqEnd,
     // straight back, so anything but the same bits is a value REAPER changed.
     if (storedStart == reqStart && storedEnd == reqEnd) return {};
 
-    std::string s = "REAPER did not keep the render bounds it was handed -- asked for [" +
+    // Says only that the two differ, not why -- a legitimate clamp (negative start,
+    // end past project end) reads back differently for the same reason a precision
+    // defect would, and this sentence cannot tell those apart.
+    std::string s = "REAPER read back different render bounds than it was handed -- "
+                    "asked for [" +
                     exactly(reqStart) + "s, " + exactly(reqEnd) + "s), read back [" +
                     exactly(storedStart) + "s, " + exactly(storedEnd) + "s).";
     if (sampleRate > 0) {

src/core/capture/render_window.h

@@ -2,7 +2,9 @@
 // render_window — pure frame arithmetic for a capture's requested window: the
 // frame count a project-time range occupies, whether a render whose bounds come
 // from the selected items' own extent already prints that window, and the two
-// diagnostics that say where a short render lost its frames.
+// diagnostics that bound a short render without locating it: whether the stored
+// RENDER_* bounds round-tripped, and whether the shortfall matches a millisecond-
+// floor coincidence.
 // NO REAPER types; unit-tested by tests/test_render_window.cpp.
 
 #include <string>
@@ -54,7 +56,7 @@ bool itemExtentPrintsWindow(double reqStart, double reqEnd,
 // A COINCIDENCE OF COUNTS, not a claim about how anything resolved the end: nothing
 // renders from this number and no capture path asks for it. Whole-millisecond values
 // are recognized within a nanosecond, because a decimal millisecond is not always one
-// in binary (0.029 * 1000 lands just below 29) and a bare floor would drop a
+// in binary (1.007 * 1000 lands just below 1007) and a bare floor would drop a
 // millisecond from a window already on the grid. A nanosecond is far under one frame
 // at any rate we render, so a real sub-millisecond remainder still floors.
 long long msFlooredEndFrameCount(double startSeconds, double endSeconds,