Files

T

SoftwareApp 47e48f054e Give a video row a frame of the video, and fix what R8 was hiding

A video row showed a film-strip icon, and a phone holds more video than anything else, so a column of identical icons said nothing about which clip was which. Images and videos now take the same path - EntryIcon hands Coil the file, and for a video the request asks for a frame - which cost exactly one dependency: io.coil-kt.coil3:coil-video, at the same version as the Coil already in use, whose own POM adds nothing that Coil did not already bring (coil-core, kotlin-stdlib, core-ktx). No ExoPlayer, no media3, and no new permission: the decoder is the platform's own MediaMetadataRetriever.

Two details are decisions rather than defaults. The frame is taken one second in, because the first frame of a recording is very often black or the middle of a fade-in, and a thumbnail of a black rectangle is no more useful than the icon it replaced. And a decode that fails - an unsupported codec, or a file only pretending to be a video - settles on the icon rather than leaving an empty square, which the row remembers so it is not retried on every recomposition.

The dependency also nearly shipped a feature that did not work. coil-video registers its decoder through a META-INF/services entry, and in a debug APK that is enough: the file and the class it names are both there, and frames decoded. In the release APK, R8 removes both, because a class mentioned only from a resource looks unused to it. So the feature worked in every build that had been looked at and would have failed silently in the one users install - the worst shape a bug can have, since it needs no code change to "fix" and would have been reported as "video thumbnails don't work" against a build nobody had run. The factory is now registered explicitly against an ImageLoader in FileManagerApplication, which makes it a code reference R8 cannot drop, and the release APK was installed and driven to the Videos library to confirm the frames decode there too. That also closes the gap noted two changes ago, that no ImageLoader was registered at all; its cache sizes are left to Coil deliberately, since they are derived from the device's heap and a number written here would only be a guess.

The same run found a smaller thing that no test could have: opening the search field from inside a library left the scope on "this folder" - the hint read "Search this folder" and a folder toggle sat beside it - while the search in fact ran everywhere, because the view model has always known that a library is not a folder and answered from the index instead. The behaviour was right and the screen was wrong, which is the same shape as the volume-list bug fixed earlier and the same root cause: the switch-on-open asked whether this was the root list rather than whether a folder was open. It now asks for a path, and the scope toggle appears only when there is a folder to choose about, because its two settings did the same thing otherwise. Every other use of isRootView was then checked for the same conflation and none was found. One test was added for the search behaviour; the scope mismatch itself is not testable, being a disagreement between two correct pieces.

324 JVM tests, lint clean. Verified on the phone in both builds: two clips made with ffmpeg (a testsrc2 pattern and SMPTE bars) drew their own patterns as thumbnails, beside a real 4.3 MB video from the device's own storage, and the demo library's three zero-filled stand-in .mp4s fell back to the icon - which exercised the failure path with material that happened to be lying around. The release APK is 2 294 021 bytes and still declares no INTERNET. Not done: the store screenshots still show the home screen from before the library views, and the release build running on a device is in nobody's checklist - which is exactly why the R8 problem survived until now, and is the first thing worth adding.

2026-10-17 01:40:35 +02:00

10 KiB

Raw Blame History

File Manager

An ad-free, one-time-paid Android file manager. It browses and manages the files on your device — locally, with no ads, no account and no analytics.

Nothing ever leaves the device: the app declares no network permission and contains no advertising or tracking SDKs.

Working title. The display name and the store listing are decided before the first Play upload (see docs/ROADMAP.md, M10). The applicationId is already fixed and can never change after the first release.

Status

M13 — video thumbnails: done. A video row shows a frame of the video itself rather than a film icon — the platform's own retriever, through Coil, with a frame taken a second in and the icon kept as the fallback when a file cannot be decoded. Before it: M12 — duplicate finder (size, then the head of every candidate, then the survivors in full), M11 — library views (a Libraries section on the root screen — Images, Videos, Audio, Documents, Archives and Recent — whose rows open ordinary listings of those files, each carrying a count from the search index), and M10 — store assets: the app is named Plain Files, the icon and feature graphic come from the app's own vectors, and docs/STORE_LISTING.md holds length-checked listing copy. Screenshots exist for 5 phone (1080×2160), 4 on a 7-inch tablet (1200×1920) and 4 on a 10-inch tablet (2560×1600) — but they predate M11, so the home screen in them is the old volume list and they will need re-shooting, which the library views make a privacy question as well as a cosmetic one (docs/STORE_LISTING.md).

Tech stack

Area	Choice
Language	Kotlin 2.4.10 (using AGP 9 built-in Kotlin)
UI	Jetpack Compose, BOM 2026.06.01, Material 3
Build	Gradle 9.3.1, Android Gradle Plugin 9.1.0
Storage access	`MANAGE_EXTERNAL_STORAGE` ("All files access") with a Storage Access Framework fallback
Archives	Apache Commons Compress + XZ for Java (both pure Java — no NDK)
SDK levels	`compileSdk`/`targetSdk` 36, `minSdk` 26

targetSdk is 36 because Google Play requires new apps and updates to target Android 16 (API 36) from 31 August 2026.

Project layout

app/src/main/java/de/softwareapp_hb/filemanager/
├── core/      Pure-Kotlin engine (no Android imports!) — paths, sorting, formatting,
│              search matching, copy/move planning, conflict policy
├── data/      Storage backends (all-files access + SAF), volumes, shares coming in
│              from other apps, the trash index (SQLite), file operations, archives,
│              deletion
├── domain/    Browser + operation controllers, models, permission state
├── service/   Foreground service for long-running file operations
├── ui/        Compose theme, browser screen, components, ViewModels
└── text/      Domain enum → string-resource mapping
app/src/test/java/…            JVM unit tests for the core engine

core/ deliberately has no Android dependencies, so the trickiest logic is covered by fast JVM tests.

Building

Prerequisites:

JDK 17+ (JDK 25 verified)
Android SDK with platforms;android-36 and build-tools;36.0.0

# Tell Gradle where the SDK lives (this file is git-ignored)
echo "sdk.dir=$HOME/Android/Sdk" > local.properties

./gradlew :app:assembleDebug        # debug APK
./gradlew :app:testDebugUnitTest    # JVM unit tests for the engine
./gradlew :app:lintDebug            # Android Lint
./gradlew :app:bundleRelease        # Play upload bundle (.aab)

Release signing is optional and driven by a git-ignored keystore.properties (see keystore.properties.example); without it the release build still runs R8/minification but produces an unsigned artefact.

Tracked in git

Source, resources and Gradle build files.
docs/ — roadmap and release documentation.
PRIVACY.md, CHANGELOG.md, and (from M10) store-assets/.

Not tracked: local.properties, keystore.properties, *.jks, build outputs.

Releasing

Release signing is driven by the git-ignored keystore.properties (see keystore.properties.example). Without that file the release build still runs R8 and resource shrinking, but produces an unsigned artefact that cannot be uploaded.

./gradlew :app:bundleRelease                      # Play upload bundle (.aab)
./gradlew :app:assembleRelease                    # APK, for local checks

# Audit what is actually shipped — in particular that there is no INTERNET permission:
$ANDROID_HOME/build-tools/36.0.0/aapt2 dump permissions \
  app/build/outputs/apk/release/app-release.apk

Read docs/PLAY_RELEASE_CHECKLIST.md before uploading. Step 4 is the one that matters most for this app: a file manager must declare "All files access" through Play's Permissions Declaration Form, and have the core purpose documented in its store description. PRIVACY.md is written to match the shipped permissions.

Localisation

All user-visible text lives in app/src/main/res/values/strings.xml, with real <plurals> wherever a count can be one, and the screen and the progress notification resolve their labels through text/TextRes.kt so the two cannot drift apart. English is currently the only locale.

Adding a language is then a file, not a refactor:

app/src/main/res/values-de/strings.xml   # same keys, translated values

Deliberately not localised, and why:

Byte units ("KB", "MB") — symbols, not words, and every file manager and the system Settings app show them the same way.
The timestamp pattern (yyyy-MM-dd HH:mm) — a fixed, sortable format; the zone is passed in explicitly rather than assumed.

Checking that nothing was missed

Two things are easy to get wrong: a literal typed into a composable, and showing an enum constant directly (FileKind.name puts untranslatable English on screen — the details sheet used to do exactly that). The mechanical check is the pseudo-locale:

./gradlew :app:assembleDebug
# Settings > System > Languages > English (XA), then launch the debug build.

The debug build type sets isPseudoLocalesEnabled = true, so every string that came from strings.xml renders accented and bracketed ([Îmàgé]). Anything that looks normal is a string that never went through resources. The release build is unaffected.

One caveat: Google Play's automatic translation translates your store listing, not the app UI. Translated UI text has to be supplied in values-<locale>/.

Known limitations

The trash is a time window, not a bin you manage. A delete stages the item in a hidden .plainfiles-trash folder on the same volume and keeps it for 24 hours; the Trash entry in the overflow menu lists what is still there and puts a whole batch back. There is no "empty it now" action — expired batches are swept the next time something is deleted — and a name taken again in the meantime is never overwritten, so a restored item can arrive beside the new one rather than replacing it.
Where a trashed item came from is remembered in the app's own database, not on the volume. The staged files are the truth and the database is only an index, so the two are reconciled in both directions: an entry whose file has gone is not offered back, and a batch folder no entry mentions is swept.
Sharing a file out only works in the all-files mode. A SAF path is relative to the granted tree rather than a filesystem path, so it cannot be handed to FileProvider. Receiving a share and being picked from another app's chooser work either way: the first needs no filesystem path, and the second falls back to RESULT_CANCELED rather than returning a URI the caller could not read.
RAR, CAB, ISO and WIM are refused, not supported — they need native code, which this project deliberately does not ship. (The Commons Compress jar contains no RAR classes at all.) 7z is different: Commons Compress reads it in pure Java, so it is supported — but the format needs a seekable channel rather than a stream, because its header lives at the end of the file. All-files access can supply one; the SAF fallback cannot, so 7z extraction reports that plainly in folder mode instead of half-working.
7z files can be opened but not created — writing stays ZIP-only.
Archive extraction does not restore original timestamps or permissions, and there is no archive encryption or password support.
Cancellation happens between entries, so a single very large file finishes before a stop is honoured.
A library is not restored across process death. The persisted view state holds a path, and a library has none, so a restart returns to the root screen rather than to Images or Recent.
Long-press multi-select is awkward to script from the host. The gesture that opens the selection bar (input swipe x y x y 900) works on the MIUI test device but has also registered as a plain tap, so compress/extract from the UI wants a manual pass (see docs/EDGE_CASE_TESTING.md).
The display name is still the working title — decided before the first upload (M10).

Monetisation

One-time paid: €4.49 base, with a €2.29 launch price for the first 14 days. No ads, no subscription, no in-app purchases, no free tier.

The price lives on the Play listing, not in the build, which is why there is no Billing client in the dependencies and no com.android.vending.BILLING permission in the APK — that is what keeps "no network permission, nothing leaves your device" true of the shipped artefact rather than just the source. Rationale, market anchors, the rejected alternatives and the Console steps: docs/MONETISATION.md.

⚠️ The price must be set before the first upload — Play allows paid → free but never free → paid.

Competitive context

The category leader is File Manager Plus (AlphaInventor; 100 M+ installs, 4.7–4.8★, free with ads and in-app purchases, and a Data Safety form that shares device IDs with third parties). What it offers, what its users praise, what they complain about, and which of those complaints this app can answer without asking for a single new permission: docs/COMPETITIVE_ANALYSIS.md.

10 KiB Raw Blame History Unescape Escape