runMedia3OrFallBack was eleven lines at 0% and isCancellation had never been called by any
JVM test -- ci=0, not merely a missed branch. The seam to reach it has existed the whole
time: ConversionDependencies.hardware, which no unit test had ever set. What kept the path
cold is that every worker test uses EnginePreference.FORCE_SOFTWARE, which never enters
the function, and the probe defaults to UNPARSEABLE, which PERMISSIVE.canDecode refuses --
so even AUTO would have routed straight to FFmpeg for a reason no assertion mentioned.
Both are now stated in setUp rather than inherited.
Four behaviours, each with the mutation that proves it:
hardware failure falls back to software delete the fallback call
... on a *clean* staging file delete staged.delete() before it
cancellation is rethrown, not fallen back delete `if (isCancellation(e)) throw e`
engine.close() runs either way empty the finally block
the display-name fallback (#169) change "input" to anything else
All five confirmed red, then restored.
The cancellation one is the reason this ticket was first in the group. runMedia3OrFallBack
catches Throwable, so without that re-throw a user cancelling a hardware transcode has the
app quietly start a *second* conversion in software -- the one thing cancelling is for.
ForcedFailureTest covers the failure half on a device and does not cover this half at all.
The clean-staging assertion is made where it is observable rather than by reading the
file: the software fake records whether the output existed when it was entered, so a
missing delete shows up as FFmpeg finding a half-written hardware output at the path it is
about to write.
556 -> 560 JVM tests, 0 failures. No production code changed.
ConversionWorker: 22 -> 9 missed lines, 14 -> 7 missed branches.
Line 2090/2348 -> 2103/2348; branch 1022/1340 -> 1029/1340.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The ticket asked for two things. One of them is not reachable from the JVM, and saying
so is most of the value here.
**probeForConcat's catch arm cannot be provoked on this runtime.** Robolectric's
MediaExtractor never throws from setDataSource -- measured across four input shapes: an
unregistered content:// authority, a missing file://, a file of garbage bytes, and an
http:// URL. All four returned normally with trackCount = 0. So a failed read arrives as
an empty track list rather than as an exception and reaches the same
ConcatInput(null, null, 0, 0, 0) by the other road. The catch stays covered only by
ConcatEngineTest on a device. The test file says this rather than implying the arm is
handled.
**What is reachable, and was genuinely missing, is the span.** Both halves were already
covered and neither reached the other: MediaProbeTrackWalkTest pins what concatInputFrom
makes of a track list, ConcatPlannerTest's `an unknown codec is not treated as a match`
pins what the planner does with a hand-built ConcatInput(video = null). The planner's
safety rests on the probe really producing that shape, and the hand-built fixture would
go on passing if it stopped.
Measured rather than claimed: mutating concatInputFrom's initial `video` to a non-null
placeholder leaves ConcatPlannerTest green and turns this red. Dropping the planner's
video null guard turns both red -- so that half was already held, and this file does not
claim credit for it.
The coupling itself is worth writing down: ConcatPlanner guards video against a null codec
and audio not at all, and that asymmetry is correct rather than an oversight --
MediaProbe.shortName returns a non-null String, so a null audioCodec means the track is
absent and two clips with no audio really do match, while a null videoCodec means absent
*or* unreadable. The audio check is safe because the video guard fires first on a clip
nothing could read. Nothing held that.
555 -> 556 JVM tests, 0 failures. No production code changed.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
JMR-devandClaude Opus 5