Say three where a third test joined, and stop the name claiming to be exact

#80 added SafPickerRoundTripTest's rotation case to @FailsOnEmulatorApi37, because a
real rotation aborts the framework on android-37.0. Three tests carry the marker now --
two in Media3EngineTest, one in SafPickerRoundTripTest -- and five statements still
described two.

Four were counts, and wrong:

  status_check.yml  "notAnnotation removes the two tests that do not pass"
  status_check.yml  "The two API 37 tests the gating row above excludes"
  CLAUDE.md         "the gating leg runs the other 55"          (59 - 3 = 56)
  CLAUDE.md         "do not read a green run as evidence those two tests pass"

The fifth was worse, because it was not a count. The advisory job's header justified its
name with an invariant:

  "It is named for WHAT IT RUNS, deliberately. Both tests drive a full H.264 -> H.265
   hardware transcode through Media3Engine"

The rotation case drives no transcode. So the comment did not merely miscount -- it
asserted a property of the job's contents that had stopped being true, and that property
was the entire argument for the name.

The name is unchanged, deliberately, and the header now says so instead of implying the
question never arose. This is not a required context, it is red on every PR by design, and
it is one people have learned to look for; renaming a check costs more than the imprecision
does. What replaced the invariant is the honest rule: THE MARKER IS THE DEFINITION, NOT THE
NAME -- this job holds the tests that cannot pass on the API 37 emulator image, whatever
their subject.

Two things stay as they were because they are still true. "the two Media3EngineTest cases
that pass here" is correct: that class has four tests and two carry the marker. And the
decoder theory is still a claim about the Media3 pair alone, so it now says so rather than
being read as covering a rotation failure it has nothing to do with.

Nothing about the job's behaviour changes: same name, same continue-on-error, same marker,
same selection on both rows. Verified: yaml parses, five jobs, matrix still 33/34/35/36/37.

The check to re-run when a test next joins or leaves the marker, which is the event that
broke this twice:

  grep -rn "@FailsOnEmulatorApi37" app/src/androidTest --include='*.kt' | grep -v import | grep -c FailsOn

It must equal the number every corrected comment states. It is 3.

Closes #81.

.github/workflows/status_check.yml

@@ -266,7 +266,7 @@ jobs:
           # api-level must be "37.0". A bare 37 is not an SDK package and fails
           # during setup, which cost a run to discover.
           #
-          # notAnnotation removes the two tests that do not pass on this image; they
+          # notAnnotation removes the three tests that do not pass on this image; they
           # run in the advisory job below, off the same marker so they cannot end up
           # in both or neither. docs/api-37-emulator-crash.md has the measurements.
           - label: "37"
@@ -367,7 +367,7 @@ jobs:
           if-no-files-found: ignore
 
   # ---------------------------------------------------------------------------
-  # The two API 37 tests the gating row above excludes, run on their own so they
+  # The three API 37 tests the gating row above excludes, run on their own so they
   # stay visible instead of disappearing behind a notAnnotation.
   #
   # continue-on-error: it reports, it never blocks. That is the whole reason it is
@@ -375,17 +375,29 @@ jobs:
   # job's `E2E API <label>` name, and a check cannot be both required and advisory
   # under one name.
   #
-  # It is named for WHAT IT RUNS, deliberately. Both tests drive a full H.264 ->
-  # H.265 hardware transcode through Media3Engine -- which is exactly what
-  # separates them from the two Media3EngineTest cases that pass here, since those
-  # two never decode video. The current theory about why they fail is in the next
-  # paragraph, where it can be corrected without renaming a check that people have
-  # already learned to look for.
+  # It was named for WHAT IT RUNS, and that is now APPROXIMATE rather than exact.
+  # When this job was created it held two tests, both driving a full H.264 -> H.265
+  # hardware transcode through Media3Engine -- which is exactly what separated them
+  # from the two Media3EngineTest cases that pass here, since those two never decode
+  # video. Since 2026-08-25 it also holds SafPickerRoundTripTest's rotation case,
+  # which drives no transcode at all: a real rotation rebuilds every surface at once
+  # and takes the framework down on this image (INSTRUMENTATION_ABORTED), which is a
+  # different failure from the decoder one below.
+  #
+  # The name is kept anyway, and that is a decision rather than an oversight. This is
+  # not a required context, it is red on every PR by design, and it is one people
+  # have learned to look for -- renaming a check costs more than the imprecision
+  # does. **The marker is the definition, not the name:** what this job holds is the
+  # tests that cannot pass on the API 37 emulator image, whatever their subject. The
+  # theory about the Media3 pair is in the next paragraph, where it can be corrected
+  # without touching the name.
   #
   # THEORY, NOT SETTLED: the exception surfaces at `dequeueOutputBuffer` on
   # `c2.goldfish.h264.decoder`, the emulator's own codec, which gets its frames out
   # of a host-side colour buffer -- the same readback machinery that aborts
-  # surfaceflinger on this image. What is MEASURED is narrower: these two fail on
+  # surfaceflinger on this image. It is about the MEDIA3 PAIR only; the rotation
+  # case above fails for its own reason. What is MEASURED is narrower: those two
+  # fail on
   # the API 37 emulator image; pass at API 36 on this runner under the same renderer
   # AND the same SystemUI-disable path; pass at API 33-36 without that path at all,
   # since nothing below 37 needs it; and pass on a physical Pixel 10 Pro XL at 37. That the decoder is the culprit rather than something else