5 Commits

Author SHA1 Message Date
tkrmagid
229f499465 v0.4.28: wipe video_player_cache on game shutdown
Some checks failed
build / build (push) Has been cancelled
Add ClientLifecycleEvents.CLIENT_STOPPING handler that deletes every
file under <gameDir>/video_player_cache/. Quieter sibling of clearAll()
- no chat notify and no concurrent-download race handling needed since
at shutdown no other code is touching the directory.

Cache repopulation on next launch is already handled: the server's
ServerPlayConnectionEvents.JOIN handler re-broadcasts a PreloadPayload
for every preload_urls and cache_entries entry on every join, so a
freshly wiped cache auto-refills from the server's named entries
without any user action.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-18 18:59:55 +09:00
tkrmagid
6abc7f9475 v0.4.27: don't join decoder worker on close — fix place-then-delete freeze
Some checks failed
build / build (push) Has been cancelled
stopWorker() ran a 2 s bounded join on the client tick thread. When a
user runs /videoPlace and immediately /videoDelete, the decoder worker
is still inside native grabber.start() doing the initial HTTP probe
(probesize=8 MB, analyzeduration=2 s); that call doesn't honor the
running flag and Thread.interrupt() doesn't unblock native I/O, so the
join blocked the tick thread for as long as start() took — exactly the
brief in-game freeze the user reported.

Signal stop (running=false, audio line stop+flush, worker.interrupt())
and return. The worker is a daemon, the audio tail is already silenced,
the Entry has been removed from the map, and the worker's finally still
closes the grabber whenever start()/grab() eventually returns — nothing
observable depends on the grabber being closed synchronously before
close() returns.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-17 03:41:03 +09:00
tkrmagid
e2c63fde7c v0.4.26: keep legacy 5-arg /videoPlace alongside new volume form
Some checks failed
build / build (push) Has been cancelled
Reviewer flagged v0.4.25 broke existing command blocks by inserting
'volume' as a required arg between height and url. Split into two
Brigadier branches: legacy 5-arg defaults volume=50, muted=false; new
6-arg keeps the -1=mute shortcut. Legacy branch uses a single-token url
arg so '50 https://...' parses as the new (int) branch, not as a legacy
url starting with a digit.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-17 03:36:59 +09:00
tkrmagid
5722c299d3 v0.4.25: add volume argument to /videoPlace, -1 mutes
Some checks failed
build / build (push) Has been cancelled
New signature:
  /videoPlace <pos> <facing> <width> <height> <volume> <url>

volume is int -1..100. 0..100 sets percent and clears mute; -1 is a CLI
shortcut that sets muted=true (underlying volume kept at 0.5 so a later
/videoMute false restores audible level without re-typing). Matches the
percent scale shown in the config GUI.
2026-05-17 03:31:13 +09:00
tkrmagid
ecd254cb78 v0.4.24: sync anchor config in vanilla BE update packet
Some checks failed
build / build (push) Has been cancelled
Bug: /videoPlace (and any path that places an anchor) reported success
on the server but the panel was invisible on the client — no quad, no
video. Walking around did not help.

Root cause: VideoAnchorBlockEntity did not override getUpdateTag() or
getUpdatePacket(). The base implementations return an empty CompoundTag
and null respectively, so url/facing/width/height were never carried in
vanilla BE sync. There is also a packet-ordering race between
level.setBlock() (queues a deferred chunk broadcast) and
ServerPlayNetworking.send(SyncAnchorPayload) (writes immediately): if
the payload arrives first, the client drops it because the BE does not
exist yet, then the chunk packet creates the BE with defaults
(url="") and the renderer silently no-ops.

Fix: override getUpdateTag(HolderLookup.Provider) → toNbt(), and
getUpdatePacket() → ClientboundBlockEntityDataPacket.create(this). NBT
key names already line up between toNbt() and loadAdditional(), so
vanilla wraps the CompoundTag in TagValueInput and existing load logic
reads it. Also fixes the 'walk far away, come back' case — that path
has no SyncAnchorPayload, just vanilla chunk re-sync.
2026-05-17 03:21:25 +09:00
6 changed files with 142 additions and 23 deletions

View File

@@ -5,7 +5,7 @@ org.gradle.configuration-cache=false
# Mod
mod_id=video_player
mod_version=0.4.23
mod_version=0.4.28
maven_group=com.ejclaw.videoplayer
archives_base_name=video_player

View File

@@ -3,6 +3,7 @@ package com.ejclaw.videoplayer;
import com.ejclaw.videoplayer.block.VideoAnchorBlockEntity;
import com.ejclaw.videoplayer.client.MusicQuizClient;
import com.ejclaw.videoplayer.client.net.ClientNetworking;
import com.ejclaw.videoplayer.client.playback.VideoCache;
import com.ejclaw.videoplayer.client.playback.VideoPlayback;
import com.ejclaw.videoplayer.client.render.VideoAnchorRenderer;
import com.ejclaw.videoplayer.item.VideoStickItem;
@@ -12,6 +13,7 @@ import net.fabricmc.api.ClientModInitializer;
import net.fabricmc.api.EnvType;
import net.fabricmc.api.Environment;
import net.fabricmc.fabric.api.client.event.lifecycle.v1.ClientBlockEntityEvents;
import net.fabricmc.fabric.api.client.event.lifecycle.v1.ClientLifecycleEvents;
import net.fabricmc.fabric.api.client.event.lifecycle.v1.ClientTickEvents;
import net.fabricmc.fabric.api.client.networking.v1.ClientPlayNetworking;
import net.fabricmc.fabric.api.client.rendering.v1.BlockEntityRendererRegistry;
@@ -82,6 +84,11 @@ public class VideoPlayerClient implements ClientModInitializer {
}
});
// Wipe video_player_cache/ on game exit so preloaded clips don't pile up across
// sessions. Cache entries are re-broadcast by the server on every JOIN, so a freshly
// started game will repopulate the cache automatically.
ClientLifecycleEvents.CLIENT_STOPPING.register(client -> VideoCache.wipeOnShutdown());
VideoPlayerMod.LOG.info("[{}] client initialized", VideoPlayerMod.MOD_ID);
}

View File

@@ -3,7 +3,11 @@ package com.ejclaw.videoplayer.block;
import com.ejclaw.videoplayer.registry.VideoPlayerBlockEntities;
import net.minecraft.core.BlockPos;
import net.minecraft.core.Direction;
import net.minecraft.core.HolderLookup;
import net.minecraft.nbt.CompoundTag;
import net.minecraft.network.protocol.Packet;
import net.minecraft.network.protocol.game.ClientGamePacketListener;
import net.minecraft.network.protocol.game.ClientboundBlockEntityDataPacket;
import net.minecraft.world.level.block.entity.BlockEntity;
import net.minecraft.world.level.block.state.BlockState;
import net.minecraft.world.level.storage.ValueInput;
@@ -123,6 +127,35 @@ public class VideoAnchorBlockEntity extends BlockEntity {
out.putBoolean("autoplay", autoplay);
}
/**
* Vanilla chunk-load BE sync. The base implementation returns an empty tag, which means
* when a client first sees this BE (chunk loads or player walks into range) it gets default
* values — url="" in particular makes the renderer no-op and the panel appears invisible.
*
* <p>Returning {@link #toNbt()} here carries the custom fields in the vanilla packet, so
* we don't depend on the {@code SyncAnchorPayload} arriving before the chunk's block-update
* packet (there's a race: {@code level.setBlock} queues a deferred chunk broadcast while
* {@code ServerPlayNetworking.send} writes immediately; if the payload wins, the client
* drops it because the BE doesn't exist yet, then the chunk packet creates the BE with
* defaults). It also fixes "player walks far away and comes back" — that path has no
* SyncAnchorPayload at all, just vanilla chunk re-sync.
*/
@Override
public CompoundTag getUpdateTag(HolderLookup.Provider provider) {
return toNbt();
}
/**
* Triggers a {@link ClientboundBlockEntityDataPacket} whenever the chunk tracker decides
* this BE needs to push an update. Default implementation returns {@code null} (no packet
* sent on BE change). Combined with {@link #getUpdateTag} above, every BE-state change a
* client sees carries the full config.
*/
@Override
public Packet<ClientGamePacketListener> getUpdatePacket() {
return ClientboundBlockEntityDataPacket.create(this);
}
@Override
protected void loadAdditional(ValueInput in) {
super.loadAdditional(in);

View File

@@ -180,27 +180,24 @@ public class JavaCvBackend implements VideoBackend {
// stale srcAddr — closing the grabber there frees the av_frame plane and the next
// memcpy crashes inside StubRoutines::jbyte_disjoint_arraycopy (exactly the 4K-delete
// crash dump we saw). So the safe rule is: only the decoder thread touches the
// grabber. External stop signals `running=false`, stops the audio line, interrupts the
// worker, and joins briefly; the worker's own `finally` calls grabber.close(). Inside
// the loop, grab() unblocks via the rw_timeout/timeout options (3 s, set in runLoop)
// even on a stuck HTTP read, so the join below normally returns within a frame.
// grabber. External stop signals `running=false`, stops the audio line, interrupts
// the worker, and returns immediately; the worker's own `finally` calls
// grabber.close() whenever grab()/start() eventually returns.
//
// We deliberately do NOT join the worker. close() runs on the client tick thread (via
// VideoPlayback.tick → Entry.close), and the worker can spend several seconds inside
// the native FFmpeg probe at the top of runLoop — probesize=8 MB and
// analyzeduration=2 s do not honor the `running` flag and Thread.interrupt() doesn't
// unblock native I/O. A bounded join() there (the old 2 s) is exactly the "place then
// immediately delete freezes the game for a moment" symptom: the worker hasn't
// entered the grab() loop yet, so flipping running=false has no effect on it until
// start() returns. The worker is a daemon, the audio line is already silenced above,
// and Entry has been removed from the active map by the caller — nothing observable
// depends on the grabber having been closed before this method returns.
Thread t = worker;
worker = null;
if (t != null) {
t.interrupt();
try {
t.join(2000);
} catch (InterruptedException ie) {
Thread.currentThread().interrupt();
}
if (t.isAlive()) {
// Worker still blocked in native grab() — let it finish on its own. Its
// finally still closes the grabber when grab() eventually returns / throws.
// No native pointers leak in the meantime because we don't touch them here.
VideoPlayerMod.LOG.warn(
"[{}] decoder did not exit within 2 s of stop; orphaning until next grab() returns",
VideoPlayerMod.MOD_ID);
}
}
ready = false;
}

View File

@@ -226,6 +226,40 @@ public final class VideoCache {
IN_FLIGHT.clear();
}
/**
* Wipe the cache directory at game shutdown so the user doesn't accumulate disk usage
* across sessions. Distinct from {@link #clearAll()}: at shutdown there are no concurrent
* downloads to race and no player to chat-notify, so we skip the chat ping but still bump
* the epoch (defensive — any straggler write inside the in-flight {@code .part} stream
* will detect the bump and self-clean) and drop indexes.
*/
public static void wipeOnShutdown() {
CACHE_EPOCH.incrementAndGet();
READY.clear();
IN_FLIGHT.clear();
int deleted = 0;
int failed = 0;
try {
Path dir = cacheDir();
if (dir != null && Files.isDirectory(dir)) {
try (var stream = Files.newDirectoryStream(dir)) {
for (Path p : stream) {
try {
if (Files.isRegularFile(p) && Files.deleteIfExists(p)) deleted++;
} catch (Throwable t) {
failed++;
}
}
}
}
} catch (Throwable t) {
VideoPlayerMod.LOG.warn("[{}] wipeOnShutdown failed: {}",
VideoPlayerMod.MOD_ID, t.toString());
}
VideoPlayerMod.LOG.info("[{}] wipeOnShutdown: deleted={} failed={}",
VideoPlayerMod.MOD_ID, deleted, failed);
}
/** Caller-supplied: current set of URLs that are fully cached, for diagnostics. */
public static Set<String> readyUrls() {
return new HashSet<>(READY.keySet());

View File

@@ -22,10 +22,30 @@ import net.minecraft.server.level.ServerPlayer;
import net.minecraft.server.permissions.Permissions;
import net.minecraft.world.level.block.Block;
/** SPEC §4.5.1 — {@code /videoPlace <pos> <facing> <width> <height> <url>} */
/**
* SPEC §4.5.1 — {@code /videoPlace} has two accepted forms:
*
* <ul>
* <li><b>Legacy 5-arg</b>: {@code /videoPlace <pos> <facing> <width> <height> <url>}
* — preserved for existing command blocks. Volume defaults to 50% and muted=false.
* The {@code url} here is a single-token string so it doesn't swallow the {@code volume}
* slot of the new form.</li>
* <li><b>New 6-arg</b>: {@code /videoPlace <pos> <facing> <width> <height> <volume> <url>}
* — {@code volume} is 0..100 (percent) or {@code -1} to start muted. The percent form
* mirrors the GUI slider; {@code -1} is a CLI shortcut so admins don't need a follow-up
* {@code /videoMute} step. {@code url} is greedy here, so URLs containing spaces (rare
* but possible after URL-decoding) still work.</li>
* </ul>
*
* <p>Brigadier disambiguates the two by the type of the 5th argument: integer → new form,
* non-integer token → legacy form.
*/
public final class VideoPlaceCommand {
private VideoPlaceCommand() {}
/** Default volume (percent) applied to the legacy 5-arg form. */
private static final int DEFAULT_VOLUME_PCT = 50;
public static void register(CommandDispatcher<CommandSourceStack> dispatcher) {
dispatcher.register(build("videoPlace"));
}
@@ -38,12 +58,33 @@ public final class VideoPlaceCommand {
.then(Commands.argument("facing", StringArgumentType.word())
.then(Commands.argument("width", IntegerArgumentType.integer(1, 32))
.then(Commands.argument("height", IntegerArgumentType.integer(1, 32))
.then(Commands.argument("url", StringArgumentType.greedyString())
.executes(VideoPlaceCommand::run))))));
// New form: volume (int) + greedy url
.then(Commands.argument("volume", IntegerArgumentType.integer(-1, 100))
.then(Commands.argument("url", StringArgumentType.greedyString())
.executes(ctx -> runNew(ctx))))
// Legacy form: single-token url, no volume slot. Single-token string
// is intentional so "<int> https://..." cannot be parsed as a legacy
// url that happens to start with a number — Brigadier first tries the
// new branch and only falls through here if "volume" isn't an int.
.then(Commands.argument("url", StringArgumentType.string())
.executes(ctx -> runLegacy(ctx)))))));
}
private static int run(com.mojang.brigadier.context.CommandContext<CommandSourceStack> ctx)
private static int runNew(com.mojang.brigadier.context.CommandContext<CommandSourceStack> ctx)
throws CommandSyntaxException {
int volumeArg = IntegerArgumentType.getInteger(ctx, "volume");
String url = StringArgumentType.getString(ctx, "url");
return runWithValues(ctx, volumeArg, url);
}
private static int runLegacy(com.mojang.brigadier.context.CommandContext<CommandSourceStack> ctx)
throws CommandSyntaxException {
String url = StringArgumentType.getString(ctx, "url");
return runWithValues(ctx, DEFAULT_VOLUME_PCT, url);
}
private static int runWithValues(com.mojang.brigadier.context.CommandContext<CommandSourceStack> ctx,
int volumeArg, String rawUrl) throws CommandSyntaxException {
CommandSourceStack src = ctx.getSource();
ServerLevel level = src.getLevel();
BlockPos pos = BlockPosArgument.getLoadedBlockPos(ctx, "pos");
@@ -54,7 +95,12 @@ public final class VideoPlaceCommand {
}
int width = IntegerArgumentType.getInteger(ctx, "width");
int height = IntegerArgumentType.getInteger(ctx, "height");
String raw = StringArgumentType.getString(ctx, "url").trim();
// -1 is the CLI mute shortcut; the BE keeps the underlying volume so an admin can
// /videoMute false later without re-typing a level. Anything 0..100 sets %-volume and
// clears mute.
boolean placeMuted = volumeArg < 0;
float placeVolume = placeMuted ? 0.5F : (volumeArg / 100.0F);
String raw = rawUrl == null ? "" : rawUrl.trim();
// Accept either an http(s) URL or a /videoCache add <name> entry: resolveUrlOrName()
// returns the canonical URL in both cases, or null when a non-URL string didn't match
// any named entry.
@@ -75,6 +121,8 @@ public final class VideoPlaceCommand {
be.setWidth(width);
be.setHeight(height);
be.setUrl(url);
be.setVolume(placeVolume);
be.setMuted(placeMuted);
CompoundTag nbt = be.toNbt();
for (ServerPlayer p : PlayerLookup.tracking(level, pos)) {