docs: say that HrtRateFit::toSeconds returns producer-epoch, not wall, seconds

The fit keeps the slope and discards the intercept, so the result counts from
the producer's boot. Tasks 4 and 7 compose it with ClockOffset::map, which is
correct, but the bare name invites passing it straight to a plot axis.

Also unwrapped the stalled-clock assertion from behind `if (fit.ready())` —
that branch never runs, so the test confirmed nothing.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Client/udpscope/tests/TimeBaseTest.cpp

@@ -102,10 +102,12 @@ TEST(HrtRateFit, SurvivesAStalledClock) {
     for (int i = 0; i < 64; i++) {
         fit.add(12345u, 1000.0 + i * 0.01);   /* hrt never advances */
     }
-    /* A degenerate fit must not produce a rate that would divide by zero. */
-    if (fit.ready()) {
-        EXPECT_GT(fit.ticksPerSecond(), 0.0);
-    }
+    /* A degenerate fit must not produce a rate that would divide by zero, so it
+     * must decline to be ready at all. Guarding this behind `if (fit.ready())`
+     * would make the test vacuous: the branch never runs and a fit that
+     * declared itself ready with a rate of 0 or NaN would pass unnoticed. */
+    EXPECT_FALSE(fit.ready());
+    EXPECT_DOUBLE_EQ(fit.ticksPerSecond(), 0.0);
 }
 
 TEST(TimeSignalScale, UsesNanosecondsForUint64AndMicrosecondsOtherwise) {