Compare commits

...
Author SHA1 Message Date
JMR-devandClaude Opus 5 016030f3e4 A1 (#167): pin all three foreground-service regimes, and the boundary between two of them
`ConversionForegroundType.current()` has three arms and the JVM suite executed one.
`robolectric.properties` pins everything to `sdk=36`, and `@Config` appears nowhere in
`app/src/test`, so 3 lines and 3 of 4 branches were cold.

The instrumented test is not a substitute, and the reason is specific rather than
general. `ConversionWorkerTest.foregroundTypeMatchesTheRunningApiLevel` asserts against
whichever API the leg is, so it covers one arm per leg and never the other two -- and the
legs that would cover 33 and 34 are the ones #122 wedges. From
docs/coverage-read-findings.md, an API 33 run reported `received: 60` with
`failed: unknown`: the regime was exercised and that leg could not have said so if it had
broken. This runs all three deterministically in the same ./gradlew invocation.

Four classes, not three. 35 shares its answer with 36 and looks redundant; it is the
whole point. Relaxing `>= VANILLA_ICE_CREAM` to `>` is invisible at every level except
exactly 35 -- measured, not assumed: that mutation failed ForegroundTypeApi35Test alone,
while swapping DATA_SYNC and MEDIA_PROCESSING failed 34, 35 and 36. Without the 35 class
the first mutation survives the suite.

546 -> 550 JVM tests, 0 failures. No production code changed.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-01 21:19:06 -05:00
Jason Ross d354f6470c Merge pull request #166 from JMR-dev/docs/coverage-wave2
Re-measure coverage after wave 2, and write down how a stacked PR merges
2026-08-29 11:33:20 -05:00
JMR-devandClaude Opus 5 dbedfb4708 Re-measure coverage after wave 2, and write down how a stacked PR merges
COVERAGE. 87.1% line / 69.1% branch, 502 tests -> 88.9% line (2087/2348), 75.4%
branch (1011/1340), 546 tests in 76 classes, as #153's five children land.

The branch figure moved for two reasons and the entry now says so, because only
one of them is new tests. The numerator rose 974 -> 1011; the denominator *fell*
1410 -> 1340. Both are the seam work: pulling a `when` out of a lambda inside a
`collect` deletes the coroutine state machine's synthesized branches around it,
and leaves a plain function whose branches a test can choose.
`ConversionViewModel$observe$1$1` went from carrying the whole mapping to six
branches, while the extracted `ConversionViewModelKt` covers 41 of 42 and
`JoinViewModelKt` 38 of 39.

That is worth stating rather than quoting the percentage alone. A number that
rises because the denominator shrank is a different claim from one that rises
because more branches are tested, and this entry has a documented history of
explaining its own movements wrongly.

STACKED PRS. A new Conventions entry, from two traps measured on 2026-08-27
while landing #144-#151.

`gh pr merge` refuses a stacked PR outright -- "must be merged using the
asynchronous merge REST API" -- and so does the plain `/merge` endpoint. The one
that works is `PUT .../pulls/N/merge-async`, which returns a uuid to poll.

The second is worse because nothing looks wrong. GitHub retargets a stacked PR's
base to main when the one below it merges, but asynchronously. Merging five about
thirty seconds apart outran it, so each merged into its own already-merged base
branch. Every call returned `status: merged`, every PR read MERGED, `gh pr list
--state open` was empty, and none of the content was on main.

What caught it was a coverage re-measure two points below what the same tree had
produced an hour earlier -- a fresh `git pull` changed nothing, which is what
made it a question rather than a stale checkout. `git merge-base --is-ancestor`
answers it in one line. #160 is what the recovery cost.

Also recorded: the auto-retarget belongs to the stacking feature. A PR opened
with a plain `--base some-branch` does not retarget when that branch merges, and
has to be moved by hand -- which is what #163 needed.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-29 11:16:20 -05:00
Jason Ross c6e9a480e1 Merge pull request #165 from JMR-dev/test/screen-wiring
W3: the screen wiring, and a narrower hazard than the ticket claimed
2026-08-29 11:13:34 -05:00
Jason Ross d8110d7a62 Merge pull request #164 from JMR-dev/test/viewmodel-setters
W4: the seven settings edits, and three tests that did not bite until they did
2026-08-29 10:57:14 -05:00
JMR-devandClaude Opus 5 6f3966cc69 W3 (#156): the screen wiring, and a narrower hazard than the ticket claimed
The stateful outer composables hand `ConverterScreenContent` and
`JoinScreenContent` a list of `viewModel::` references. No test in the suite had
ever seen that list: the content tests build their own `ConverterActions`, so
they drive the stateless inner and never touch the wiring.

THE TICKET'S PREMISE WAS HALF WRONG, AND CHECKING BEAT ASSUMING. #156 was filed
claiming a transposition of any two of seventeen bindings would survive the
suite. Measured instead of trusted:

  onVideoCodec <-> onAudioCodec   -> REJECTED: "Inapplicable candidate(s):
                                     fun setAudioCodec(codec: AudioCodec)"
  onCancel     <-> onReset        -> COMPILES

Every typed binding -- container, both codecs, preset, suggestion, quality,
engine preference -- takes a distinct parameter type, so the compiler is already
the test. Writing assertions against those transpositions would have been
theatre, and this file says so rather than quietly including them.

WHAT IS ACTUALLY AT RISK is the `() -> Unit` bindings, which are interchangeable
to the compiler: two on the converter screen (onCancel, onReset) and *three* on
the join screen (onJoin, onCancel, onReset). A Cancel button that discards the
finished file, a Start-over that leaves it on screen, or a Join button that
cancels -- each is one wrong word and each ships.

I got that wrong in the first check too: an early run reported the onCancel/
onReset swap as rejected, from a grep-and-exit-code test that misread a stale
build. Re-running it properly printed BUILD SUCCESSFUL with the swap in place.

THE SEAM. `converterActions(viewModel, onPickInput, onConvert, onSave)` and
`joinActions(viewModel, onPickInputs, onSave)`. The launcher-backed actions stay
parameters -- they need an ActivityResultLauncher, which is the part that
genuinely needs a composition, and keeping them out means the rest needs none.

Told apart by effect rather than by a recording double: `reset()` sets the state
to Idle, `cancel()` with no active job leaves it alone (`activeWorkId?.let`,
which SettingsEditsTest pins).

Mutations -- every transposition caught, each by two tests:

  converter onCancel <-> onReset  | 2 tests
  join      onJoin   <-> onCancel | 2 tests
  join      onReset  <-> onCancel | 2 tests
  a typed binding dropped to {}   | 1 test
  a launcher action rerouted      | 1 test

The two-test symmetry is deliberate: one direction alone passes against a wiring
with BOTH actions bound to the same method, which is what a copy-pasted line
produces.

Three guard assertions earned their place during writing -- the picks land
through an injected dispatcher, and without `ParkedPickDispatcher.runAll()` all
three state-based tests sat on Idle and would have asserted nothing. They failed
loudly instead of passing quietly.

`@UnstableApi` on both builders, per CLAUDE.md; lint caught their absence, as it
did in W1.

525 -> 537 tests, 88.0% -> 88.9% line, 70.5% -> 75.4% branch. Gate green.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-29 10:56:44 -05:00
Jason Ross 7595177e81 Merge pull request #163 from JMR-dev/test/join-state-mapping
W2: the join state mapping, and a crash the seam exposed
2026-08-29 10:49:03 -05:00
JMR-dev 1ff5c4463c Merge remote-tracking branch 'origin/main' into test/join-state-mapping 2026-08-29 10:39:55 -05:00
JMR-devandClaude Opus 5 fea480b000 W2 (#155): the join state mapping, and a crash the seam exposed
The join-side twin of W1, deliberately the same shape -- one refactor done twice,
and letting the two diverge would cost more than the duplication. Five arms had
never been chosen by any test, for the same reason: a real ConcatWorker only ever
reaches a terminal state with well-formed output.

WHAT THE SEAM TURNED UP. This line was in the SUCCEEDED arm:

    info.outputData.getString(ConcatWorker.KEY_STRATEGY)
        ?.let(ConcatStrategy::valueOf) ?: ConcatStrategy.REENCODE

`valueOf` throws IllegalArgumentException on a name this build does not define,
and this runs inside a `viewModelScope` collect with no handler -- so it is not a
Failed state, it takes the process down.

Not theoretical. WorkManager keeps finished work about a week, so a downgrade or
rollback hands this build a job enqueued by another one -- the premise
`WorkerEnumFallbackTest` and `JobTags` are both written on. `ConcatWorker` writes
`result.strategy.name`, so a build that added a third strategy would leave this
one crashing on its own completed joins.

The codebase had already made this exact fix one file over, and said why:

    // Looked up rather than `valueOf` -- see the same three reads in
    // ConversionWorker. This one is above the try as well, so a format name this
    // build does not define used to throw past the catch

The matching read on the ViewModel side had not been changed with it. It is now
`ConcatStrategy.entries.firstOrNull { it.name == name } ?: REENCODE`.

PROVEN RATHER THAN ASSERTED. Restoring `valueOf` and running the new test:

    RED: an unknown strategy name is read as a re-encode rather than thrown
      java.lang.IllegalArgumentException: No enum constant
      org.libremediaconverter.model.ConcatStrategy.SMART_CONCAT_V2

REENCODE is the conservative default rather than an arbitrary one: it is the
answer for inputs that do not match, so a job whose strategy cannot be read is
described as the more cautious of the two rather than claimed as a lossless
stream copy. The mutation to STREAM_COPY reddens two tests.

Eight mutations, all red:

  unknown strategy -> STREAM_COPY            | 2 tests
  runAttemptCount ignored, both directions   | 2 tests
  success with no path -> empty Joined       | 1 test
  BLOCKED unfolded from RUNNING              | 1 test
  blank error no longer falls back           | 1 test
  cancellation ignores the caller's state    | 1 test

516 -> 530 tests, 87.7% -> 88.0% line, 70.4% -> 71.3% branch. Gate green:
assembleDebug, testDebugUnitTest, compileDebugAndroidTestKotlin, ktlintCheck,
detekt, lintDebug.

Stacked on W1 (#162), which this mirrors and should not land before.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-29 10:37:13 -05:00
7 changed files with 761 additions and 67 deletions
+44 -3
View File
@@ -130,9 +130,9 @@ install for code that can never run — and on API 37 the full APK does not fit
- The `model` package is excluded from `ReturnCount` and `CyclomaticComplexMethod` only. It is the
decision layer, where one branch is one documented user-visible outcome and the metric counts
answers rather than complexity. Every other rule still applies there.
- **Coverage is reported, not gated** — **87.1% of lines (2025/2324), 69.1% of branches
(974/1410)**, measured 2026-08-27 with `./gradlew :app:jacocoTestReport`, against 502 JVM tests
in 71 classes.
- **Coverage is reported, not gated** — **88.9% of lines (2087/2348), 75.4% of branches
(1011/1340)**, measured 2026-08-29 with `./gradlew :app:jacocoTestReport`, against 546 JVM tests
in 76 classes.
**Every figure this file carried before 2026-08-24 was an artifact, roughly half the real one.**
Robolectric loads classes through its own sandbox classloader with no source location, JaCoCo
@@ -159,6 +159,21 @@ install for code that can never run — and on API 37 the full APK does not fit
chooses a branch the suite had never taken. Line coverage barely notices; branch coverage is the
whole point.
Then #153's five children on 2026-08-29 — 87.1% -> 88.9% line, 69.1% -> **75.4%** branch, 502 ->
546 tests.
**That last branch figure moved for two reasons, and only one of them is new tests.** The
numerator rose 974 -> 1011; the denominator *fell* 1410 -> 1340. Both are the seam work. Pulling
a `when` out of a lambda inside a `collect` deletes the coroutine state machine's synthesized
branches around it, and what is left is a plain function whose branches a test can choose:
`ConversionViewModel$observe$1$1` went from carrying the whole mapping to 6 branches, while the
extracted `ConversionViewModelKt` covers 41 of 42 and `JoinViewModelKt` 38 of 39. So a seam is
worth more than the tests it enables — it also stops the measurement counting scaffolding.
Be careful quoting a branch move on its own for that reason. A percentage that rises because the
denominator shrank is not the same claim as one that rises because more branches are tested, and
this entry has a history of explaining its own numbers wrongly.
And **re-measure before quoting**: this entry was once written quoting 81.4%, measured four hours
earlier, and was already three points stale by the time it was ready to merge.
- **Testable code is not done until it is tested.** If a piece is unit testable, it gets unit
@@ -182,6 +197,32 @@ install for code that can never run — and on API 37 the full APK does not fit
- `kotlin.code.style=official`. Gradle stays Kotlin DSL.
- **A stacked PR does not merge with `gh pr merge`, and `MERGED` is not proof it reached `main`.**
Two separate traps, both measured on 2026-08-27 while landing #144-#151.
`gh pr merge` uses the GraphQL mutation, which refuses a stacked PR outright: *"This pull request
is part of a stack and must be merged using the asynchronous merge REST API."* So does
`PUT .../pulls/{n}/merge`. The one that works is
`gh api -X PUT repos/OWNER/REPO/pulls/N/merge-async -f merge_method=merge`, which returns a uuid
to poll at `.../merge-async/{uuid}` until `status` is `merged` or `failed`.
**The second trap is worse, because nothing looks wrong.** GitHub retargets a stacked PR's base
to `main` when the PR below it merges, but *asynchronously*. Merge a stack faster than that
settles — five PRs about thirty seconds apart, in the case that found this — and each one merges
into its own base branch, which has itself already been merged and left behind. Every call
returns `status: merged` and every one is true. `gh pr list --state open` comes back empty, every
PR shows `MERGED`, and **none of the content is on `main`**.
What caught it was a coverage re-measure reading two points lower than the same tree had measured
an hour earlier; a fresh `git pull` changed nothing, which is what turned it into a question.
`git merge-base --is-ancestor <merge-sha> origin/main` answers it in one line. Do that after
merging a stack, or merge one at a time and re-read `baseRefName` between. #160 is what the
recovery cost.
The auto-retarget belongs to the stacking feature specifically. A PR opened with a plain
`gh pr create --base some-branch` does **not** retarget when that branch merges — it is left
pointing at a dead base and has to be moved by hand.
- **File one-off issues with `tools/github/file-issue.sh`, not `gh issue create`.** `gh issue
create` does not touch the project board, so the issue exists, carries its labels, and is
invisible in the Kanban — indistinguishable from never having been filed. Measured 2026-08-24:
@@ -94,24 +94,61 @@ fun ConverterScreen(modifier: Modifier = Modifier, viewModel: ConversionViewMode
state = state,
settings = settings,
validation = validation,
actions = ConverterActions(
actions = converterActions(
viewModel = viewModel,
onPickInput = { pickInput.launch(arrayOf("*/*")) },
onPreset = viewModel::setPreset,
onContainer = viewModel::setContainer,
onVideoCodec = viewModel::setVideoCodec,
onAudioCodec = viewModel::setAudioCodec,
onSuggestion = viewModel::applySuggestion,
onQuality = viewModel::setQuality,
onEnginePreference = viewModel::setEnginePreference,
onConvert = { requestNotifications.launch(Manifest.permission.POST_NOTIFICATIONS) },
onCancel = viewModel::cancel,
onSave = { suggestedName -> chooseDestination.launch(suggestedName) },
onReset = viewModel::reset,
),
modifier = modifier,
)
}
/**
* Which of the ViewModel's methods each affordance on the screen calls.
*
* ## Why this is a function rather than an argument list
*
* It was an argument list, inside [ConverterScreen], which no test reached: `ConverterScreenContent`
* builds its own [ConverterActions], so every test in the suite drove the stateless inner and none
* of them ever saw the wiring.
*
* Most of the list is safe without a test, and saying so is more useful than pretending otherwise:
* `onContainer`, `onVideoCodec`, `onAudioCodec`, `onPreset`, `onSuggestion`, `onQuality` and
* `onEnginePreference` each take a distinct type, so binding one to another's setter does not
* compile. Verified rather than assumed — swapping `onVideoCodec` and `onAudioCodec` fails with
* *"Inapplicable candidate(s): fun setAudioCodec(codec: AudioCodec)"*.
*
* **[ConverterActions.onCancel] and [ConverterActions.onReset] are the exception.** Both are
* `() -> Unit`, so swapping them compiles silently — also verified — and ships a Cancel button that
* throws the conversion away and a Start-over button that leaves it on screen. That pair is what
* `ConverterWiringTest` exists for.
*
* The three launcher-backed actions stay parameters: they need an `ActivityResultLauncher`, which
* is the part that genuinely needs the composition, and keeping them out means the rest can be
* checked without one.
*/
@UnstableApi
internal fun converterActions(
viewModel: ConversionViewModel,
onPickInput: () -> Unit,
onConvert: () -> Unit,
onSave: (suggestedName: String) -> Unit,
): ConverterActions = ConverterActions(
onPickInput = onPickInput,
onPreset = viewModel::setPreset,
onContainer = viewModel::setContainer,
onVideoCodec = viewModel::setVideoCodec,
onAudioCodec = viewModel::setAudioCodec,
onSuggestion = viewModel::applySuggestion,
onQuality = viewModel::setQuality,
onEnginePreference = viewModel::setEnginePreference,
onConvert = onConvert,
onCancel = viewModel::cancel,
onSave = onSave,
onReset = viewModel::reset,
)
/**
* Everything [ConverterScreenContent] can ask for, in one value.
*
@@ -56,17 +56,38 @@ fun JoinScreen(modifier: Modifier = Modifier, viewModel: JoinViewModel = viewMod
JoinScreenContent(
state = state,
actions = JoinActions(
actions = joinActions(
viewModel = viewModel,
onPickInputs = { pickInputs.launch(arrayOf("video/*")) },
onJoin = viewModel::join,
onCancel = viewModel::cancel,
onSave = { suggestedName -> chooseDestination.launch(suggestedName) },
onReset = viewModel::reset,
),
modifier = modifier,
)
}
/**
* Which of the ViewModel's methods each affordance on the join screen calls.
*
* The join-side twin of `converterActions`, and the transposition risk here is worse: **three**
* `() -> Unit` bindings rather than two. `onJoin`, `onCancel` and `onReset` are mutually
* interchangeable as far as the compiler is concerned, so a Join button that cancels, or a Cancel
* button that starts the job, is a swap nothing but a test would catch.
*
* See `converterActions` for why the launcher-backed actions stay parameters.
*/
@UnstableApi
internal fun joinActions(
viewModel: JoinViewModel,
onPickInputs: () -> Unit,
onSave: (suggestedName: String) -> Unit,
): JoinActions = JoinActions(
onPickInputs = onPickInputs,
onJoin = viewModel::join,
onCancel = viewModel::cancel,
onSave = onSave,
onReset = viewModel::reset,
)
/**
* Everything [JoinScreenContent] can ask for, in one value.
*
@@ -5,6 +5,7 @@ import android.net.Uri
import androidx.lifecycle.AndroidViewModel
import androidx.lifecycle.viewModelScope
import androidx.media3.common.util.UnstableApi
import androidx.work.Data
import androidx.work.WorkInfo
import androidx.work.WorkManager
import kotlinx.coroutines.CoroutineDispatcher
@@ -30,6 +31,108 @@ import org.libremediaconverter.work.jobSnapshots
import java.io.File
import java.util.UUID
/**
* One update about a running join, as WorkManager last reported it.
*
* The join-side twin of `ConversionUpdate`, and deliberately the same shape: this pair of seams is
* one refactor done twice, and letting them diverge would make the two flows harder to compare than
* the duplication costs. There is no `progressPercent` here because `ConcatWorker` publishes none —
* a join is indeterminate.
*/
internal data class JoinUpdate(val state: WorkInfo.State, val runAttemptCount: Int, val outputData: Data)
/**
* What the join screen should show, given what WorkManager last said about the job.
*
* The join-side twin of `conversionStateFrom`, extracted for the same reason and with the same two
* exclusions: the ownership check stays at the call site, and this takes no responsibility for the
* staged file. See that function's KDoc for the argument in full.
*
* Five arms had never been chosen by any test before this was cut out, because a real `ConcatWorker`
* only ever produces a terminal state with well-formed output.
*
* @param cancelled where a cancellation lands, which differs for a reattached job — see [observe].
*/
@UnstableApi
internal fun joinStateFrom(update: JoinUpdate, inputs: List<InputFile>, cancelled: JoinState): JoinState =
when (update.state) {
// BLOCKED is a job waiting on a prerequisite, which the user has nothing to do about and
// nothing useful to be told about. It reads as "starting", like a fresh ENQUEUED.
WorkInfo.State.RUNNING, WorkInfo.State.BLOCKED -> JoinState.Joining(inputs)
// ENQUEUED after a run means a retry is pending -- the same rule, and the same reasoning, as
// the convert side. See `conversionStateFrom`.
WorkInfo.State.ENQUEUED ->
if (update.runAttemptCount > 0) {
JoinState.Waiting(inputs)
} else {
JoinState.Joining(inputs)
}
WorkInfo.State.SUCCEEDED -> joinedFrom(update.outputData)
// A worker that dies before it can report anything leaves no output data at all, and an
// exception's message can be an empty string. Both would read as a failure with nothing said,
// so blank falls back like missing does.
WorkInfo.State.FAILED -> JoinState.Failed(
update.outputData.getString(ConcatWorker.KEY_ERROR)
?.takeIf { it.isNotBlank() }
?: ConcatWorker.GENERIC_FAILURE_MESSAGE,
)
WorkInfo.State.CANCELLED -> cancelled
}
/**
* The `SUCCEEDED` arm. A join that reported success and named no file has nothing to offer.
*/
@UnstableApi
private fun joinedFrom(outputData: Data): JoinState {
val path = outputData.getString(ConcatWorker.KEY_OUTPUT_PATH)
?: return JoinState.Failed(JOINED_WITHOUT_A_FILE_MESSAGE)
return JoinState.Joined(
staged = File(path),
strategy = strategyFrom(outputData.getString(ConcatWorker.KEY_STRATEGY)),
// A join enqueued before the worker reported these carries neither, and the fallback is
// the format such a job really used -- ConcatWorker.request has always defaulted to it,
// and the join screen has never offered a choice.
suggestedName = outputData.getString(ConcatWorker.KEY_SUGGESTED_NAME)
?.takeIf { it.isNotBlank() }
?: ConcatWorker.outputNameFor(ConcatWorker.DEFAULT_FORMAT),
mimeType = outputData.getString(ConcatWorker.KEY_MIME_TYPE)
?.takeIf { it.isNotBlank() }
?: ConcatWorker.DEFAULT_FORMAT.mimeType,
)
}
/**
* The strategy a finished join reported, or [ConcatStrategy.REENCODE] when it named none.
*
* **Looked up rather than `valueOf`, and that is a fix rather than a style choice.** `valueOf`
* throws `IllegalArgumentException` on a name this build does not define, and this runs inside a
* `viewModelScope` collect with no handler -- so the throw does not become a `Failed` state, it
* takes the process down.
*
* Reachable for the reason `WorkerEnumFallbackTest` and `JobTags` are both written on: WorkManager
* keeps finished work for about a week, so a downgrade or a rollback hands this build a job
* enqueued by another one. `ConcatWorker` writes `result.strategy.name` into the output `Data`, so
* a build that added a third strategy would leave this one crashing on its own completed joins.
*
* `ConcatWorker.kt` already made exactly this change for `KEY_FORMAT`, and says why in as many
* words: *"Looked up rather than `valueOf` … a format name this build does not define used to throw
* past the catch."* The same read on this side had not been changed with it.
*
* REENCODE is the safe default rather than an arbitrary one: it is the answer for inputs that do
* not match, so a job whose strategy cannot be read is described as the more conservative of the
* two rather than being claimed as a lossless stream copy.
*/
private fun strategyFrom(name: String?): ConcatStrategy =
ConcatStrategy.entries.firstOrNull { it.name == name } ?: ConcatStrategy.REENCODE
/** A join that reported success and named no file. There is nothing to offer the user to save. */
internal const val JOINED_WITHOUT_A_FILE_MESSAGE: String =
"Joining reported success but produced no file."
sealed interface JoinState {
data object Idle : JoinState
data class Ready(val inputs: List<InputFile>) : JoinState
@@ -251,56 +354,19 @@ class JoinViewModel @JvmOverloads constructor(
// takes ownership of the staged file, and a superseded observation must not do
// that either.
if (!ownership.stillHeldBy(token)) return@collect
_state.value = when (info.state) {
WorkInfo.State.RUNNING, WorkInfo.State.BLOCKED -> JoinState.Joining(inputs)
WorkInfo.State.ENQUEUED ->
if (info.runAttemptCount > 0) {
JoinState.Waiting(inputs)
} else {
JoinState.Joining(inputs)
}
WorkInfo.State.SUCCEEDED -> {
val path = info.outputData.getString(ConcatWorker.KEY_OUTPUT_PATH)
val strategy = info.outputData.getString(ConcatWorker.KEY_STRATEGY)
?.let(ConcatStrategy::valueOf) ?: ConcatStrategy.REENCODE
if (path == null) {
JoinState.Failed("Joining reported success but produced no file.")
} else {
val staged = File(path)
// Take responsibility for the file at the same moment the state
// starts referring to it, so the two cannot disagree.
pendingStaged = staged
JoinState.Joined(
staged = staged,
strategy = strategy,
// A join enqueued before the worker reported these carries
// neither, and the fallback is the format such a job really
// used -- ConcatWorker.request has always defaulted to it, and
// the join screen has never offered a choice.
suggestedName = info.outputData
.getString(ConcatWorker.KEY_SUGGESTED_NAME)
?.takeIf { it.isNotBlank() }
?: ConcatWorker.outputNameFor(ConcatWorker.DEFAULT_FORMAT),
mimeType = info.outputData
.getString(ConcatWorker.KEY_MIME_TYPE)
?.takeIf { it.isNotBlank() }
?: ConcatWorker.DEFAULT_FORMAT.mimeType,
)
}
}
// A worker that dies before it can report anything leaves no output data at
// all, and an exception's message can be an empty string. Both would read as
// a failure with nothing said, so blank falls back like missing does.
WorkInfo.State.FAILED -> JoinState.Failed(
info.outputData.getString(ConcatWorker.KEY_ERROR)
?.takeIf { it.isNotBlank() }
?: ConcatWorker.GENERIC_FAILURE_MESSAGE,
)
WorkInfo.State.CANCELLED -> cancelled
}
val next = joinStateFrom(
JoinUpdate(
state = info.state,
runAttemptCount = info.runAttemptCount,
outputData = info.outputData,
),
inputs = inputs,
cancelled = cancelled,
)
// Read off the result rather than assigned inside the mapping -- see the same
// three lines in ConversionViewModel for why that is the better half of the swap.
if (next is JoinState.Joined) pendingStaged = next.staged
_state.value = next
}
}
}
@@ -0,0 +1,241 @@
package org.libremediaconverter.convert
import android.app.Application
import android.net.Uri
import androidx.media3.common.util.UnstableApi
import androidx.work.Data
import androidx.work.workDataOf
import org.junit.After
import org.junit.Assert.assertEquals
import org.junit.Assert.assertNotEquals
import org.junit.Assert.assertTrue
import org.junit.Before
import org.junit.Test
import org.junit.runner.RunWith
import org.libremediaconverter.join.JoinState
import org.libremediaconverter.join.JoinViewModel
import org.libremediaconverter.join.joinActions
import org.libremediaconverter.model.AudioCodec
import org.libremediaconverter.model.Container
import org.libremediaconverter.model.EnginePreference
import org.libremediaconverter.model.OutputFormat
import org.libremediaconverter.model.QualityTier
import org.libremediaconverter.model.VideoCodec
import org.libremediaconverter.work.ConcatWorker
import org.robolectric.RobolectricTestRunner
import org.robolectric.RuntimeEnvironment
/**
* That each affordance is wired to the ViewModel method it is named after.
*
* ## What this covers that no other test can
*
* `ConverterScreenContentTest`, `ConverterStateAffordancesTest` and `JoinScreenContentTest` all
* drive the **stateless** content composables, which build their own `ConverterActions`. So the
* wiring — the list of `viewModel::` references the stateful outer hands down — was seen by nothing
* in the suite.
*
* ## The hazard is narrower than "seventeen bindings", and this says so
*
* #156 was filed claiming a transposition of any two bindings would survive the suite. That is not
* true, and it was worth checking rather than testing on the assumption:
*
* | swap | result |
* |---|---|
* | `onVideoCodec` ↔ `onAudioCodec` | **rejected by the compiler** |
* | `onCancel` ↔ `onReset` | **compiles** |
*
* Every typed binding — container, both codecs, preset, suggestion, quality, engine preference —
* takes a distinct parameter type, so the compiler is already the test. Writing assertions for
* those would be theatre.
*
* **The `() -> Unit` bindings are the real gap**, because they are interchangeable to the compiler:
* two on the converter screen (`onCancel`, `onReset`) and three on the join screen (`onJoin`,
* `onCancel`, `onReset`). A Cancel that discards the finished file, or a Join that cancels, is a
* one-character mistake that ships.
*
* ## How they are told apart
*
* By effect, not by a recording double. `reset()` sets the state to `Idle`; `cancel()` with no
* active job leaves it alone (`ConversionViewModel.cancel` is `activeWorkId?.let(...)`, and
* `SettingsEditsTest` pins that). Driving each from a non-`Idle` state is therefore enough to say
* which one ran.
*/
@UnstableApi
@RunWith(RobolectricTestRunner::class)
class ScreenWiringTest {
private lateinit var app: Application
@Before
fun setUp() {
app = RuntimeEnvironment.getApplication()
ConversionDependencies.publisher = { RecordingPublisher(app) }
ConversionDependencies.probe = { _, _ -> org.libremediaconverter.model.InputProbe() }
}
@After
fun tearDown() {
ConversionDependencies.reset()
}
// --- the converter screen ----------------------------------------------
@Test
fun `Start over resets, and Cancel does not`() {
// The transposition that compiles. If onReset were bound to cancel, this stays on Ready.
installTestWorkManager(app, Data.EMPTY)
val pick = ParkedPickDispatcher()
val viewModel = ConversionViewModel(app, pickDispatcher = pick)
val actions = converterActions(viewModel, onPickInput = {}, onConvert = {}, onSave = {})
viewModel.onInputPicked(INPUT_URI)
pick.runAll()
assertNotEquals(
"the fixture needs a non-Idle state or neither action is observable",
ConversionState.Idle,
viewModel.state.value,
)
actions.onReset()
assertEquals(ConversionState.Idle, viewModel.state.value)
}
@Test
fun `Cancel leaves the picked file on screen`() {
// The other half. Without it, a wiring with BOTH actions bound to reset passes the test
// above -- and that is exactly what a copy-paste of the wrong line produces.
installTestWorkManager(app, Data.EMPTY)
val pick = ParkedPickDispatcher()
val viewModel = ConversionViewModel(app, pickDispatcher = pick)
val actions = converterActions(viewModel, onPickInput = {}, onConvert = {}, onSave = {})
viewModel.onInputPicked(INPUT_URI)
pick.runAll()
val before = viewModel.state.value
actions.onCancel()
assertEquals(
"Cancel must not throw away the pick the way Start over does",
before,
viewModel.state.value,
)
}
@Test
fun `each settings affordance reaches the setting it is named after`() {
// The typed bindings. The compiler already rejects a transposition among these, so this is
// not that assertion -- it is the cheaper one that each is bound to *something*, and that a
// binding dropped to `{}` during an edit would be caught.
installTestWorkManager(app, Data.EMPTY)
val pick = ParkedPickDispatcher()
val viewModel = ConversionViewModel(app, pickDispatcher = pick)
val actions = converterActions(viewModel, onPickInput = {}, onConvert = {}, onSave = {})
actions.onPreset(OutputFormat.WEBM_VP9)
assertEquals(OutputFormat.WEBM_VP9.spec, viewModel.settings.value.spec)
actions.onContainer(Container.MKV)
assertEquals(Container.MKV, viewModel.settings.value.spec.container)
actions.onVideoCodec(VideoCodec.H264)
assertEquals(VideoCodec.H264, viewModel.settings.value.spec.videoCodec)
actions.onAudioCodec(AudioCodec.FLAC)
assertEquals(AudioCodec.FLAC, viewModel.settings.value.spec.audioCodec)
actions.onQuality(QualityTier.BEST)
assertEquals(QualityTier.BEST, viewModel.settings.value.quality)
actions.onEnginePreference(EnginePreference.FORCE_SOFTWARE)
assertEquals(EnginePreference.FORCE_SOFTWARE, viewModel.settings.value.enginePreference)
actions.onSuggestion(OutputFormat.MP4_H264.spec)
assertEquals(OutputFormat.MP4_H264.spec, viewModel.settings.value.spec)
}
@Test
fun `the launcher-backed actions are the ones the screen supplies`() {
// Not wired to the ViewModel at all, deliberately -- they need an ActivityResultLauncher.
// Asserted so that a later edit routing one of them at the ViewModel is noticed.
installTestWorkManager(app, Data.EMPTY)
val pick = ParkedPickDispatcher()
val viewModel = ConversionViewModel(app, pickDispatcher = pick)
val called = mutableListOf<String>()
val actions = converterActions(
viewModel,
onPickInput = { called += "pick" },
onConvert = { called += "convert" },
onSave = { called += "save:$it" },
)
actions.onPickInput()
actions.onConvert()
actions.onSave("holiday.mp4")
assertEquals(listOf("pick", "convert", "save:holiday.mp4"), called)
}
// --- the join screen, where three are interchangeable -------------------
@Test
fun `Start over resets the join, and Cancel does not`() {
installTestWorkManager(app, workDataOf(ConcatWorker.KEY_OUTPUT_PATH to "/dev/null"))
val pick = ParkedPickDispatcher()
val viewModel = JoinViewModel(app, pickDispatcher = pick)
val actions = joinActions(viewModel, onPickInputs = {}, onSave = {})
viewModel.onInputsPicked(TWO_INPUTS)
pick.runAll()
assertTrue(
"the fixture needs a non-Idle state: ${viewModel.state.value}",
viewModel.state.value !is JoinState.Idle,
)
actions.onReset()
assertEquals(JoinState.Idle, viewModel.state.value)
}
@Test
fun `Cancel leaves the picked files on screen`() {
installTestWorkManager(app, workDataOf(ConcatWorker.KEY_OUTPUT_PATH to "/dev/null"))
val pick = ParkedPickDispatcher()
val viewModel = JoinViewModel(app, pickDispatcher = pick)
val actions = joinActions(viewModel, onPickInputs = {}, onSave = {})
viewModel.onInputsPicked(TWO_INPUTS)
pick.runAll()
val before = viewModel.state.value
actions.onCancel()
assertEquals(before, viewModel.state.value)
}
@Test
fun `Join starts the job rather than cancelling or resetting it`() {
// The third of the join screen's interchangeable trio, and the one whose transposition is
// worst: a Join button bound to cancel does nothing at all, which reads as a dead button.
installTestWorkManager(app, workDataOf(ConcatWorker.KEY_OUTPUT_PATH to "/dev/null"))
val pick = ParkedPickDispatcher()
val viewModel = JoinViewModel(app, pickDispatcher = pick)
val actions = joinActions(viewModel, onPickInputs = {}, onSave = {})
viewModel.onInputsPicked(TWO_INPUTS)
pick.runAll()
actions.onJoin()
assertTrue(
"Join must leave Ready for a running state, not sit still and not go Idle: " +
"${viewModel.state.value}",
viewModel.state.value is JoinState.Joining || viewModel.state.value is JoinState.Joined,
)
}
private companion object {
val INPUT_URI: Uri = Uri.parse("content://test/holiday.mov")
val TWO_INPUTS = listOf(
Uri.parse("content://test/a.mp4"),
Uri.parse("content://test/b.mp4"),
)
}
}
@@ -0,0 +1,200 @@
package org.libremediaconverter.join
import android.net.Uri
import androidx.media3.common.util.UnstableApi
import androidx.work.Data
import androidx.work.WorkInfo
import androidx.work.workDataOf
import org.junit.Assert.assertEquals
import org.junit.Test
import org.junit.runner.RunWith
import org.libremediaconverter.convert.InputFile
import org.libremediaconverter.model.ConcatStrategy
import org.libremediaconverter.work.ConcatWorker
import org.robolectric.RobolectricTestRunner
/**
* Every answer [joinStateFrom] can give, chosen rather than stumbled into.
*
* The join-side twin of `ConversionStateMappingTest`, and the argument is the same one: the mapping
* ran on every test that drove a real `ConcatWorker`, but a real worker only ever reaches a terminal
* state with well-formed output, so five arms had never been *chosen* by anything.
*
* ## The one that is not just coverage
*
* `an unknown strategy name is read as a re-encode rather than thrown` covers a real defect this
* seam exposed. The line it replaces was:
*
* ```kotlin
* .getString(ConcatWorker.KEY_STRATEGY)?.let(ConcatStrategy::valueOf) ?: ConcatStrategy.REENCODE
* ```
*
* `valueOf` throws on a name this build does not define, and this runs inside a `viewModelScope`
* collect with no handler — so it does not become a `Failed` state, it takes the process down.
* `ConcatWorker.kt` had already made this exact change for `KEY_FORMAT` and written down why; the
* matching read on this side had not been changed with it.
*/
@UnstableApi
@RunWith(RobolectricTestRunner::class)
class JoinStateMappingTest {
@Test
fun `a running join is joining`() {
assertEquals(JoinState.Joining(INPUTS), map(WorkInfo.State.RUNNING))
}
@Test
fun `a blocked join looks like one that is starting`() {
// Folded into the RUNNING arm deliberately: a job waiting on a prerequisite is nothing the
// user can act on, and a separate word for it would be noise.
assertEquals(JoinState.Joining(INPUTS), map(WorkInfo.State.BLOCKED))
}
@Test
fun `an enqueued join that has already run is waiting to retry`() {
assertEquals(JoinState.Waiting(INPUTS), map(WorkInfo.State.ENQUEUED, runAttemptCount = 1))
}
@Test
fun `an enqueued join that has never run is simply starting`() {
// The other side. Without it, a mapping that ignored runAttemptCount passes the test above.
assertEquals(JoinState.Joining(INPUTS), map(WorkInfo.State.ENQUEUED, runAttemptCount = 0))
}
@Test
fun `a success that named no file is a failure, not an empty success`() {
assertEquals(
JoinState.Failed(JOINED_WITHOUT_A_FILE_MESSAGE),
map(WorkInfo.State.SUCCEEDED, data = Data.EMPTY),
)
}
@Test
fun `a success carries the strategy the worker actually used`() {
// Not cosmetic: the join screen tells the user whether their files were stream-copied or
// re-encoded, which is the difference between lossless and lossy.
val joined = map(
WorkInfo.State.SUCCEEDED,
data = workDataOf(
ConcatWorker.KEY_OUTPUT_PATH to "/cache/conversions/joined.mp4",
ConcatWorker.KEY_STRATEGY to ConcatStrategy.STREAM_COPY.name,
),
) as JoinState.Joined
assertEquals(ConcatStrategy.STREAM_COPY, joined.strategy)
}
@Test
fun `an unknown strategy name is read as a re-encode rather than thrown`() {
// The defect. A build that added a third strategy leaves finished joins in the queue naming
// it, and WorkManager keeps those about a week -- the premise WorkerEnumFallbackTest and
// JobTags are both written on. With `valueOf` this throws IllegalArgumentException inside a
// viewModelScope collect that has no handler, so it is not a Failed state, it is a crash.
//
// REENCODE rather than STREAM_COPY because it is the conservative answer: describing an
// unknown join as lossless would be a claim the app cannot support.
val joined = map(
WorkInfo.State.SUCCEEDED,
data = workDataOf(
ConcatWorker.KEY_OUTPUT_PATH to "/cache/conversions/joined.mp4",
ConcatWorker.KEY_STRATEGY to "SMART_CONCAT_V2",
),
) as JoinState.Joined
assertEquals(ConcatStrategy.REENCODE, joined.strategy)
}
@Test
fun `a success with no strategy at all falls back the same way`() {
val joined = map(
WorkInfo.State.SUCCEEDED,
data = workDataOf(ConcatWorker.KEY_OUTPUT_PATH to "/cache/conversions/joined.mp4"),
) as JoinState.Joined
assertEquals(ConcatStrategy.REENCODE, joined.strategy)
}
@Test
fun `a success from older work falls back to the format such a job really used`() {
val joined = map(
WorkInfo.State.SUCCEEDED,
data = workDataOf(ConcatWorker.KEY_OUTPUT_PATH to "/cache/conversions/joined.mp4"),
) as JoinState.Joined
assertEquals(ConcatWorker.outputNameFor(ConcatWorker.DEFAULT_FORMAT), joined.suggestedName)
assertEquals(ConcatWorker.DEFAULT_FORMAT.mimeType, joined.mimeType)
}
@Test
fun `a blank name or type falls back the same way a missing one does`() {
val joined = map(
WorkInfo.State.SUCCEEDED,
data = workDataOf(
ConcatWorker.KEY_OUTPUT_PATH to "/cache/conversions/joined.mp4",
ConcatWorker.KEY_SUGGESTED_NAME to "",
ConcatWorker.KEY_MIME_TYPE to " ",
),
) as JoinState.Joined
assertEquals(ConcatWorker.outputNameFor(ConcatWorker.DEFAULT_FORMAT), joined.suggestedName)
assertEquals(ConcatWorker.DEFAULT_FORMAT.mimeType, joined.mimeType)
}
@Test
fun `a failure carries the reason the worker gave`() {
assertEquals(
JoinState.Failed("Not enough free space to join these files."),
map(
WorkInfo.State.FAILED,
data = workDataOf(
ConcatWorker.KEY_ERROR to "Not enough free space to join these files.",
),
),
)
}
@Test
fun `a failure with nothing said still says something`() {
assertEquals(
JoinState.Failed(ConcatWorker.GENERIC_FAILURE_MESSAGE),
map(WorkInfo.State.FAILED, data = Data.EMPTY),
)
}
@Test
fun `a failure whose message is blank falls back like a missing one`() {
assertEquals(
JoinState.Failed(ConcatWorker.GENERIC_FAILURE_MESSAGE),
map(WorkInfo.State.FAILED, data = workDataOf(ConcatWorker.KEY_ERROR to " ")),
)
}
@Test
fun `a cancellation lands wherever the caller said it should`() {
// A join started here goes back to Ready with the picked files; one picked up by reattach
// goes to Idle, because those URIs belong to a process that no longer exists.
assertEquals(
JoinState.Ready(INPUTS),
map(WorkInfo.State.CANCELLED, cancelled = JoinState.Ready(INPUTS)),
)
assertEquals(JoinState.Idle, map(WorkInfo.State.CANCELLED, cancelled = JoinState.Idle))
}
private fun map(
state: WorkInfo.State,
runAttemptCount: Int = 0,
data: Data = Data.EMPTY,
cancelled: JoinState = JoinState.Ready(INPUTS),
): JoinState = joinStateFrom(
JoinUpdate(state = state, runAttemptCount = runAttemptCount, outputData = data),
inputs = INPUTS,
cancelled = cancelled,
)
private companion object {
val INPUTS = listOf(
InputFile(Uri.parse("content://test/a.mp4"), "a.mp4", 1024L),
InputFile(Uri.parse("content://test/b.mp4"), "b.mp4", 2048L),
)
}
}
@@ -0,0 +1,88 @@
package org.libremediaconverter.work
import android.content.pm.ServiceInfo
import org.junit.Assert.assertEquals
import org.junit.Test
import org.junit.runner.RunWith
import org.robolectric.RobolectricTestRunner
import org.robolectric.annotation.Config
/**
* [ConversionForegroundType.current] answers differently on each of the three API regimes, and
* until this file only one of them was ever executed.
*
* `app/src/test/resources/robolectric.properties` pins the whole JVM suite to `sdk=36`, so every
* Robolectric test that reaches a `ForegroundInfo` takes the `mediaProcessing` arm and no other.
* The 33 and 34 arms were cold: 3 lines and 3 of 4 branches, measured on `main` at `d354f64`.
*
* **The instrumented test is not a substitute, and the reason is specific.**
* `ConversionWorkerTest.foregroundTypeMatchesTheRunningApiLevel` asserts against whichever API the
* leg happens to be — one arm per leg, never the other two — and the legs that would cover 33 and
* 34 are the ones issue #122 wedges. `docs/coverage-read-findings.md` records an API 33 run that
* reported `received: 60` and `failed: unknown`: the regime *was* exercised, and that leg could
* not have said so if it had broken. Four `@Config` classes here pin all three arms
* deterministically, in the same `./gradlew` invocation as everything else.
*
* `minSdk` is 33, so none of these is dead code — each is a device someone is running the app on.
*
* **SDK 35 is in the list for the boundary, not for the answer.** It shares its answer with 36,
* which would make it look redundant. It is not: relaxing `>= VANILLA_ICE_CREAM` to `>` is invisible
* at every level except exactly 35, so without this class that mutation survives the suite.
*/
@RunWith(RobolectricTestRunner::class)
@Config(sdk = [33])
class ForegroundTypeApi33Test {
/**
* Zero rather than a named constant because there is no constant to name: API 33 does not
* require a type, and `mediaProcessing` does not exist here to pass. `ForegroundInfo` reads 0
* as "no type at all", which is what this regime wants.
*/
@Test
fun `api 33 asks for no foreground service type`() {
assertEquals(0, ConversionForegroundType.current())
}
}
/**
* API 34 makes a type mandatory and still has no `mediaProcessing`, so `dataSync` is the only
* sensible fit. See [ForegroundTypeApi33Test] for why this file exists.
*/
@RunWith(RobolectricTestRunner::class)
@Config(sdk = [34])
class ForegroundTypeApi34Test {
@Test
fun `api 34 falls back to dataSync, the only type that fits`() {
assertEquals(ServiceInfo.FOREGROUND_SERVICE_TYPE_DATA_SYNC, ConversionForegroundType.current())
}
}
/**
* The first level with `mediaProcessing`, and therefore the one that tells `>=` from `>`.
* See [ForegroundTypeApi33Test].
*/
@RunWith(RobolectricTestRunner::class)
@Config(sdk = [35])
class ForegroundTypeApi35Test {
@Test
fun `api 35 is the first level that takes mediaProcessing`() {
assertEquals(ServiceInfo.FOREGROUND_SERVICE_TYPE_MEDIA_PROCESSING, ConversionForegroundType.current())
}
}
/**
* The level the rest of the suite runs at, asserted here rather than assumed — it is the one arm
* that was already covered, and leaving it out would make this file look like it is about the old
* levels rather than about all three regimes. See [ForegroundTypeApi33Test].
*/
@RunWith(RobolectricTestRunner::class)
@Config(sdk = [36])
class ForegroundTypeApi36Test {
@Test
fun `api 36 keeps mediaProcessing`() {
assertEquals(ServiceInfo.FOREGROUND_SERVICE_TYPE_MEDIA_PROCESSING, ConversionForegroundType.current())
}
}