papermc/Spigot-Server-Patches/0149-Use-TerminalConsoleAppender-for-console-improvements.patch

622 lines
28 KiB
Diff
Raw Normal View History

Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
From bd039fe4f27ea8f45a60714fb121a7e639eb0b94 Mon Sep 17 00:00:00 2001
From: Minecrell <minecrell@minecrell.net>
Date: Fri, 9 Jun 2017 19:03:43 +0200
Subject: [PATCH] Use TerminalConsoleAppender for console improvements
Rewrite console improvements (console colors, tab completion,
persistent input line, ...) using JLine 3.x and TerminalConsoleAppender.
New features:
- Support console colors for Vanilla commands
- Add console colors for warnings and errors
- Server can now be turned off safely using CTRL + C. JLine catches
the signal and the implementation shuts down the server cleanly.
- Support console colors and persistent input line when running in
IntelliJ IDEA
Other changes:
- Server starts 1-2 seconds faster thanks to optimizations in Log4j
configuration
diff --git a/pom.xml b/pom.xml
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
index 3dc6c2a3f5..b1f008738b 100644
--- a/pom.xml
+++ b/pom.xml
@@ -41,10 +41,27 @@
<scope>compile</scope>
</dependency>
<dependency>
- <groupId>jline</groupId>
- <artifactId>jline</artifactId>
- <version>2.12.1</version>
- <scope>compile</scope>
+ <groupId>net.minecrell</groupId>
+ <artifactId>terminalconsoleappender</artifactId>
+ <version>1.2.0</version>
+ </dependency>
+ <dependency>
+ <groupId>org.jline</groupId>
+ <artifactId>jline-terminal-jansi</artifactId>
+ <version>3.12.1</version>
+ <scope>runtime</scope>
+ </dependency>
+ <!--
+ Required to add the missing Log4j2Plugins.dat file from log4j-core
+ which has been removed by Mojang. Without it, log4j has to classload
+ all its classes to check if they are plugins.
+ Scanning takes about 1-2 seconds so adding this speeds up the server start.
+ -->
+ <dependency>
+ <groupId>org.apache.logging.log4j</groupId>
+ <artifactId>log4j-core</artifactId>
+ <version>2.8.1</version>
+ <scope>runtime</scope>
</dependency>
<dependency>
<groupId>org.ow2.asm</groupId>
2019-04-27 06:26:04 +00:00
@@ -230,10 +247,18 @@
<transformer implementation="org.apache.maven.plugins.shade.resource.AppendingTransformer">
<resource>META-INF/services/java.sql.Driver</resource>
</transformer>
+ <transformer implementation="com.github.edwgiz.mavenShadePlugin.log4j2CacheTransformer.PluginsCacheFileTransformer" />
</transformers>
</configuration>
</execution>
</executions>
+ <dependencies>
+ <dependency>
+ <groupId>com.github.edwgiz</groupId>
+ <artifactId>maven-shade-plugin.log4j2-cachefile-transformer</artifactId>
+ <version>2.8.1</version>
+ </dependency>
+ </dependencies>
</plugin>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
diff --git a/src/main/java/com/destroystokyo/paper/console/PaperConsole.java b/src/main/java/com/destroystokyo/paper/console/PaperConsole.java
new file mode 100644
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
index 0000000000..cd6e259239
--- /dev/null
+++ b/src/main/java/com/destroystokyo/paper/console/PaperConsole.java
@@ -0,0 +1,40 @@
+package com.destroystokyo.paper.console;
+
+import net.minecraft.server.DedicatedServer;
+import net.minecrell.terminalconsole.SimpleTerminalConsole;
+import org.bukkit.craftbukkit.command.ConsoleCommandCompleter;
+import org.jline.reader.LineReader;
+import org.jline.reader.LineReaderBuilder;
+
+public final class PaperConsole extends SimpleTerminalConsole {
+
+ private final DedicatedServer server;
+
+ public PaperConsole(DedicatedServer server) {
+ this.server = server;
+ }
+
+ @Override
+ protected LineReader buildReader(LineReaderBuilder builder) {
+ return super.buildReader(builder
+ .appName("Paper")
+ .completer(new ConsoleCommandCompleter(this.server))
+ );
+ }
+
+ @Override
+ protected boolean isRunning() {
+ return !this.server.isStopped() && this.server.isRunning();
+ }
+
+ @Override
+ protected void runCommand(String command) {
2018-07-19 19:23:18 +00:00
+ this.server.issueCommand(command, this.server.getServerCommandListener());
+ }
+
+ @Override
+ protected void shutdown() {
2019-05-05 23:24:37 +00:00
+ this.server.safeShutdown(false);
+ }
+
+}
diff --git a/src/main/java/com/destroystokyo/paper/console/TerminalConsoleCommandSender.java b/src/main/java/com/destroystokyo/paper/console/TerminalConsoleCommandSender.java
new file mode 100644
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
index 0000000000..685deaa0e5
--- /dev/null
+++ b/src/main/java/com/destroystokyo/paper/console/TerminalConsoleCommandSender.java
@@ -0,0 +1,17 @@
+package com.destroystokyo.paper.console;
+
+import org.apache.logging.log4j.LogManager;
+import org.apache.logging.log4j.Logger;
+import org.bukkit.craftbukkit.command.CraftConsoleCommandSender;
+
+public class TerminalConsoleCommandSender extends CraftConsoleCommandSender {
+
+ private static final Logger LOGGER = LogManager.getRootLogger();
+
+ @Override
+ public void sendRawMessage(String message) {
+ // TerminalConsoleAppender supports color codes directly in log messages
+ LOGGER.info(message);
+ }
+
+}
diff --git a/src/main/java/net/minecraft/server/DedicatedServer.java b/src/main/java/net/minecraft/server/DedicatedServer.java
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
index 4b1f8c5373..d34f772fae 100644
--- a/src/main/java/net/minecraft/server/DedicatedServer.java
+++ b/src/main/java/net/minecraft/server/DedicatedServer.java
@@ -86,6 +86,9 @@ public class DedicatedServer extends MinecraftServer implements IMinecraftServer
if (!org.bukkit.craftbukkit.Main.useConsole) {
return;
}
+ // Paper start - Use TerminalConsoleAppender
+ new com.destroystokyo.paper.console.PaperConsole(DedicatedServer.this).start();
+ /*
jline.console.ConsoleReader bufferedreader = reader;
Updated Upstream (Bukkit/CraftBukkit/Spigot) Upstream has released updates that appears to apply and compile correctly. This update has not been tested by PaperMC and as with ANY update, please do your own testing Bukkit Changes: 93e39ce1 Clarify documentation regarding getMaterial with legacyName = true c3aeaea0 Improve dependency tracker 14c9d275 Add support for transitive depends in load access warning c8afe560 SPIGOT-5526: Add EntityEnterBlockEvent 6bb6f07d SPIGOT-5548: Show error that hints towards plugins misusing reflection ed75537d SPIGOT-5546: Fix bad depend access using wrong provider in message 4e4c0ee9 Fix buggy classloader warning triggering for all classes 89586a4c Print warning when loading classes from depends that have not been specified d4fe9680 Fix bug where disablePlugin could remove ConfigurationSerializable classes from other plugins 85e683b7 Add additional checkstyle checks 612fd8e1 Correct max page count in BookMeta docs fa8a9781 Correct max title length in BookMeta docs CraftBukkit Changes: ab13a117 SPIGOT-5550: Cancelled ProjectileLaunchEvent still plays sound for eggs 44016b1d SPIGOT-5538: Using javaw to run GUI prints input error e653ae76 SPIGOT-5526: Call EntityEnterBlockEvent for bees trying to enter hives 6515ea49 SPIGOT-5537: Bee nests generated by growing trees near flower have no bees d82b3149 Remove unused CraftWorld.getId method 10763a88 Change some block == AIR checks to isAir to catch CAVE_AIR Spigot Changes: f2c1cd15 Rebuild patches bcd458ad Reformat patches
2020-01-28 19:43:57 +00:00
// MC-33041, SPIGOT-5538: if System.in is not valid due to javaw, then return
@@ -125,6 +128,8 @@ public class DedicatedServer extends MinecraftServer implements IMinecraftServer
DedicatedServer.LOGGER.error("Exception handling console input", ioexception);
}
+ */
+ // Paper end
}
};
Updated Upstream (Bukkit/CraftBukkit/Spigot) Upstream has released updates that appears to apply and compile correctly. This update has not been tested by PaperMC and as with ANY update, please do your own testing Bukkit Changes: 93e39ce1 Clarify documentation regarding getMaterial with legacyName = true c3aeaea0 Improve dependency tracker 14c9d275 Add support for transitive depends in load access warning c8afe560 SPIGOT-5526: Add EntityEnterBlockEvent 6bb6f07d SPIGOT-5548: Show error that hints towards plugins misusing reflection ed75537d SPIGOT-5546: Fix bad depend access using wrong provider in message 4e4c0ee9 Fix buggy classloader warning triggering for all classes 89586a4c Print warning when loading classes from depends that have not been specified d4fe9680 Fix bug where disablePlugin could remove ConfigurationSerializable classes from other plugins 85e683b7 Add additional checkstyle checks 612fd8e1 Correct max page count in BookMeta docs fa8a9781 Correct max title length in BookMeta docs CraftBukkit Changes: ab13a117 SPIGOT-5550: Cancelled ProjectileLaunchEvent still plays sound for eggs 44016b1d SPIGOT-5538: Using javaw to run GUI prints input error e653ae76 SPIGOT-5526: Call EntityEnterBlockEvent for bees trying to enter hives 6515ea49 SPIGOT-5537: Bee nests generated by growing trees near flower have no bees d82b3149 Remove unused CraftWorld.getId method 10763a88 Change some block == AIR checks to isAir to catch CAVE_AIR Spigot Changes: f2c1cd15 Rebuild patches bcd458ad Reformat patches
2020-01-28 19:43:57 +00:00
@@ -136,6 +141,9 @@ public class DedicatedServer extends MinecraftServer implements IMinecraftServer
}
global.addHandler(new org.bukkit.craftbukkit.util.ForwardLogHandler());
+ // Paper start - Not needed with TerminalConsoleAppender
+ final org.apache.logging.log4j.Logger logger = LogManager.getRootLogger();
+ /*
final org.apache.logging.log4j.core.Logger logger = ((org.apache.logging.log4j.core.Logger) LogManager.getRootLogger());
for (org.apache.logging.log4j.core.Appender appender : logger.getAppenders().values()) {
if (appender instanceof org.apache.logging.log4j.core.appender.ConsoleAppender) {
Updated Upstream (Bukkit/CraftBukkit/Spigot) Upstream has released updates that appears to apply and compile correctly. This update has not been tested by PaperMC and as with ANY update, please do your own testing Bukkit Changes: 93e39ce1 Clarify documentation regarding getMaterial with legacyName = true c3aeaea0 Improve dependency tracker 14c9d275 Add support for transitive depends in load access warning c8afe560 SPIGOT-5526: Add EntityEnterBlockEvent 6bb6f07d SPIGOT-5548: Show error that hints towards plugins misusing reflection ed75537d SPIGOT-5546: Fix bad depend access using wrong provider in message 4e4c0ee9 Fix buggy classloader warning triggering for all classes 89586a4c Print warning when loading classes from depends that have not been specified d4fe9680 Fix bug where disablePlugin could remove ConfigurationSerializable classes from other plugins 85e683b7 Add additional checkstyle checks 612fd8e1 Correct max page count in BookMeta docs fa8a9781 Correct max title length in BookMeta docs CraftBukkit Changes: ab13a117 SPIGOT-5550: Cancelled ProjectileLaunchEvent still plays sound for eggs 44016b1d SPIGOT-5538: Using javaw to run GUI prints input error e653ae76 SPIGOT-5526: Call EntityEnterBlockEvent for bees trying to enter hives 6515ea49 SPIGOT-5537: Bee nests generated by growing trees near flower have no bees d82b3149 Remove unused CraftWorld.getId method 10763a88 Change some block == AIR checks to isAir to catch CAVE_AIR Spigot Changes: f2c1cd15 Rebuild patches bcd458ad Reformat patches
2020-01-28 19:43:57 +00:00
@@ -144,6 +152,8 @@ public class DedicatedServer extends MinecraftServer implements IMinecraftServer
}
2019-04-27 06:26:04 +00:00
new org.bukkit.craftbukkit.util.TerminalConsoleWriterThread(System.out, this.reader).start();
+ */
+ // Paper end
System.setOut(new PrintStream(new LoggerOutputStream(logger, Level.INFO), true));
System.setErr(new PrintStream(new LoggerOutputStream(logger, Level.WARN), true));
diff --git a/src/main/java/net/minecraft/server/MinecraftServer.java b/src/main/java/net/minecraft/server/MinecraftServer.java
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
index 8d6a089007..284793e4bf 100644
--- a/src/main/java/net/minecraft/server/MinecraftServer.java
+++ b/src/main/java/net/minecraft/server/MinecraftServer.java
2019-07-20 04:01:24 +00:00
@@ -57,7 +57,7 @@ import org.apache.commons.lang3.Validate;
import org.apache.logging.log4j.LogManager;
import org.apache.logging.log4j.Logger;
// CraftBukkit start
-import jline.console.ConsoleReader;
2019-04-27 06:26:04 +00:00
+import joptsimple.OptionSet;
import org.bukkit.Bukkit;
import org.bukkit.craftbukkit.CraftServer;
2019-04-27 06:26:04 +00:00
import org.bukkit.craftbukkit.Main;
2019-12-11 23:43:22 +00:00
@@ -161,7 +161,7 @@ public abstract class MinecraftServer extends IAsyncTaskHandlerReentrant<TickTas
public OptionSet options;
public org.bukkit.command.ConsoleCommandSender console;
public org.bukkit.command.RemoteConsoleCommandSender remoteConsole;
- public ConsoleReader reader;
+ //public ConsoleReader reader; // Paper
public static int currentTick = 0; // Paper - Further improve tick loop
2019-04-27 06:26:04 +00:00
public java.util.Queue<Runnable> processQueue = new java.util.concurrent.ConcurrentLinkedQueue<Runnable>();
public int autosavePeriod;
2019-12-11 23:43:22 +00:00
@@ -212,7 +212,9 @@ public abstract class MinecraftServer extends IAsyncTaskHandlerReentrant<TickTas
2019-07-20 04:01:24 +00:00
this.K = s;
// CraftBukkit start
this.options = options;
+ // Paper start - Handled by TerminalConsoleAppender
// Try to see if we're actually running in a terminal, disable jline if not
+ /*
if (System.console() == null && System.getProperty("jline.terminal") == null) {
System.setProperty("jline.terminal", "jline.UnsupportedTerminal");
Main.useJline = false;
2019-12-11 23:43:22 +00:00
@@ -233,6 +235,8 @@ public abstract class MinecraftServer extends IAsyncTaskHandlerReentrant<TickTas
LOGGER.warn((String) null, ex);
}
}
+ */
+ // Paper end
Runtime.getRuntime().addShutdownHook(new org.bukkit.craftbukkit.util.ServerShutdownThread(this));
2019-04-27 06:26:04 +00:00
}
// CraftBukkit end
@@ -951,7 +955,7 @@ public abstract class MinecraftServer extends IAsyncTaskHandlerReentrant<TickTas
org.spigotmc.WatchdogThread.doStop(); // Spigot
// CraftBukkit start - Restore terminal to original settings
try {
- reader.getTerminal().restore();
+ net.minecrell.terminalconsole.TerminalConsoleAppender.close(); // Paper - Use TerminalConsoleAppender
} catch (Exception ignored) {
}
// CraftBukkit end
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
@@ -1475,7 +1479,7 @@ public abstract class MinecraftServer extends IAsyncTaskHandlerReentrant<TickTas
2019-04-27 06:26:04 +00:00
@Override
public void sendMessage(IChatBaseComponent ichatbasecomponent) {
- MinecraftServer.LOGGER.info(ichatbasecomponent.getString());
+ MinecraftServer.LOGGER.info(org.bukkit.craftbukkit.util.CraftChatMessage.fromComponent(ichatbasecomponent, net.minecraft.server.EnumChatFormat.WHITE));// Paper - Log message with colors
}
2019-04-27 06:26:04 +00:00
public KeyPair getKeyPair() {
diff --git a/src/main/java/net/minecraft/server/PlayerList.java b/src/main/java/net/minecraft/server/PlayerList.java
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
index 56266a7729..77f1ddc8a8 100644
--- a/src/main/java/net/minecraft/server/PlayerList.java
+++ b/src/main/java/net/minecraft/server/PlayerList.java
2019-04-27 06:26:04 +00:00
@@ -76,8 +76,7 @@ public abstract class PlayerList {
2019-04-27 06:26:04 +00:00
public PlayerList(MinecraftServer minecraftserver, int i) {
this.cserver = minecraftserver.server = new CraftServer((DedicatedServer) minecraftserver, this);
- minecraftserver.console = org.bukkit.craftbukkit.command.ColouredConsoleSender.getInstance();
- minecraftserver.reader.addCompleter(new org.bukkit.craftbukkit.command.ConsoleCommandCompleter(minecraftserver.server));
+ minecraftserver.console = new com.destroystokyo.paper.console.TerminalConsoleCommandSender(); // Paper
// CraftBukkit end
2019-12-11 23:43:22 +00:00
this.k = new GameProfileBanList(PlayerList.b);
diff --git a/src/main/java/org/bukkit/craftbukkit/CraftServer.java b/src/main/java/org/bukkit/craftbukkit/CraftServer.java
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
index 2e0f443d9f..742c21480a 100644
--- a/src/main/java/org/bukkit/craftbukkit/CraftServer.java
+++ b/src/main/java/org/bukkit/craftbukkit/CraftServer.java
@@ -42,7 +42,6 @@ import java.util.logging.Level;
2019-04-27 06:26:04 +00:00
import java.util.logging.Logger;
import javax.imageio.ImageIO;
import it.unimi.dsi.fastutil.objects.Object2ObjectLinkedOpenHashMap;
-import jline.console.ConsoleReader;
2019-04-27 06:26:04 +00:00
import net.minecraft.server.Advancement;
import net.minecraft.server.ArgumentEntity;
import net.minecraft.server.Block;
@@ -1118,9 +1117,13 @@ public final class CraftServer implements Server {
return logger;
}
+ // Paper start - JLine update
+ /*
public ConsoleReader getReader() {
return console.reader;
}
+ */
+ // Paper end
@Override
public PluginCommand getPluginCommand(String name) {
diff --git a/src/main/java/org/bukkit/craftbukkit/Main.java b/src/main/java/org/bukkit/craftbukkit/Main.java
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
index 93fcfd5f39..0e76e34664 100644
--- a/src/main/java/org/bukkit/craftbukkit/Main.java
+++ b/src/main/java/org/bukkit/craftbukkit/Main.java
2019-04-27 06:26:04 +00:00
@@ -13,7 +13,7 @@ import java.util.logging.Logger;
import joptsimple.OptionParser;
import joptsimple.OptionSet;
import net.minecraft.server.MinecraftServer;
-import org.fusesource.jansi.AnsiConsole;
+import net.minecrell.terminalconsole.TerminalConsoleAppender; // Paper
public class Main {
public static boolean useJline = true;
@@ -179,6 +179,8 @@ public class Main {
}
try {
+ // Paper start - Handled by TerminalConsoleAppender
+ /*
// This trick bypasses Maven Shade's clever rewriting of our getProperty call when using String literals
String jline_UnsupportedTerminal = new String(new char[] {'j','l','i','n','e','.','U','n','s','u','p','p','o','r','t','e','d','T','e','r','m','i','n','a','l'});
String jline_terminal = new String(new char[] {'j','l','i','n','e','.','t','e','r','m','i','n','a','l'});
@@ -196,10 +198,18 @@ public class Main {
// This ensures the terminal literal will always match the jline implementation
System.setProperty(jline.TerminalFactory.JLINE_TERMINAL, jline.UnsupportedTerminal.class.getName());
}
+ */
+ if (options.has("nojline")) {
+ System.setProperty(TerminalConsoleAppender.JLINE_OVERRIDE_PROPERTY, "false");
+ useJline = false;
+ }
+ // Paper end
if (options.has("noconsole")) {
useConsole = false;
+ useJline = false; // Paper
+ System.setProperty(TerminalConsoleAppender.JLINE_OVERRIDE_PROPERTY, "false"); // Paper
}
if (Main.class.getPackage().getImplementationVendor() != null && System.getProperty("IReallyKnowWhatIAmDoingISwear") == null) {
@@ -227,7 +237,7 @@ public class Main {
System.out.println("Unable to read system info");
}
// Paper end
-
+ System.setProperty( "library.jansi.version", "Paper" ); // Paper - set meaningless jansi version to prevent git builds from crashing on Windows
System.out.println("Loading libraries, please wait...");
MinecraftServer.main(options);
} catch (Throwable t) {
diff --git a/src/main/java/org/bukkit/craftbukkit/command/ColouredConsoleSender.java b/src/main/java/org/bukkit/craftbukkit/command/ColouredConsoleSender.java
deleted file mode 100644
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
index fdf2f075e2..0000000000
--- a/src/main/java/org/bukkit/craftbukkit/command/ColouredConsoleSender.java
+++ /dev/null
2019-04-27 06:26:04 +00:00
@@ -1,72 +0,0 @@
-package org.bukkit.craftbukkit.command;
-
-import java.util.EnumMap;
-import java.util.Map;
-import jline.Terminal;
-import org.bukkit.Bukkit;
-import org.bukkit.ChatColor;
-import org.bukkit.command.ConsoleCommandSender;
-import org.bukkit.craftbukkit.CraftServer;
2019-04-27 06:26:04 +00:00
-import org.fusesource.jansi.Ansi;
-import org.fusesource.jansi.Ansi.Attribute;
-
-public class ColouredConsoleSender extends CraftConsoleCommandSender {
- private final Terminal terminal;
- private final Map<ChatColor, String> replacements = new EnumMap<ChatColor, String>(ChatColor.class);
- private final ChatColor[] colors = ChatColor.values();
-
- protected ColouredConsoleSender() {
- super();
- this.terminal = ((CraftServer) getServer()).getReader().getTerminal();
-
- replacements.put(ChatColor.BLACK, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.BLACK).boldOff().toString());
- replacements.put(ChatColor.DARK_BLUE, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.BLUE).boldOff().toString());
- replacements.put(ChatColor.DARK_GREEN, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.GREEN).boldOff().toString());
- replacements.put(ChatColor.DARK_AQUA, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.CYAN).boldOff().toString());
- replacements.put(ChatColor.DARK_RED, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.RED).boldOff().toString());
- replacements.put(ChatColor.DARK_PURPLE, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.MAGENTA).boldOff().toString());
- replacements.put(ChatColor.GOLD, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.YELLOW).boldOff().toString());
- replacements.put(ChatColor.GRAY, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.WHITE).boldOff().toString());
- replacements.put(ChatColor.DARK_GRAY, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.BLACK).bold().toString());
- replacements.put(ChatColor.BLUE, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.BLUE).bold().toString());
- replacements.put(ChatColor.GREEN, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.GREEN).bold().toString());
- replacements.put(ChatColor.AQUA, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.CYAN).bold().toString());
- replacements.put(ChatColor.RED, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.RED).bold().toString());
- replacements.put(ChatColor.LIGHT_PURPLE, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.MAGENTA).bold().toString());
- replacements.put(ChatColor.YELLOW, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.YELLOW).bold().toString());
- replacements.put(ChatColor.WHITE, Ansi.ansi().a(Attribute.RESET).fg(Ansi.Color.WHITE).bold().toString());
- replacements.put(ChatColor.MAGIC, Ansi.ansi().a(Attribute.BLINK_SLOW).toString());
- replacements.put(ChatColor.BOLD, Ansi.ansi().a(Attribute.UNDERLINE_DOUBLE).toString());
- replacements.put(ChatColor.STRIKETHROUGH, Ansi.ansi().a(Attribute.STRIKETHROUGH_ON).toString());
- replacements.put(ChatColor.UNDERLINE, Ansi.ansi().a(Attribute.UNDERLINE).toString());
- replacements.put(ChatColor.ITALIC, Ansi.ansi().a(Attribute.ITALIC).toString());
- replacements.put(ChatColor.RESET, Ansi.ansi().a(Attribute.RESET).toString());
- }
-
- @Override
- public void sendMessage(String message) {
- if (terminal.isAnsiSupported()) {
- if (!conversationTracker.isConversingModaly()) {
- String result = message;
- for (ChatColor color : colors) {
- if (replacements.containsKey(color)) {
- result = result.replaceAll("(?i)" + color.toString(), replacements.get(color));
- } else {
- result = result.replaceAll("(?i)" + color.toString(), "");
- }
- }
- System.out.println(result + Ansi.ansi().reset().toString());
- }
- } else {
- super.sendMessage(message);
- }
- }
-
- public static ConsoleCommandSender getInstance() {
- if (Bukkit.getConsoleSender() != null) {
- return Bukkit.getConsoleSender();
- } else {
- return new ColouredConsoleSender();
- }
- }
-}
diff --git a/src/main/java/org/bukkit/craftbukkit/command/ConsoleCommandCompleter.java b/src/main/java/org/bukkit/craftbukkit/command/ConsoleCommandCompleter.java
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
index befcc19f9b..5510266fb1 100644
--- a/src/main/java/org/bukkit/craftbukkit/command/ConsoleCommandCompleter.java
+++ b/src/main/java/org/bukkit/craftbukkit/command/ConsoleCommandCompleter.java
2019-05-06 02:58:04 +00:00
@@ -4,20 +4,31 @@ import java.util.Collections;
2019-04-27 06:26:04 +00:00
import java.util.List;
import java.util.concurrent.ExecutionException;
import java.util.logging.Level;
-import jline.console.completer.Completer;
import org.bukkit.craftbukkit.CraftServer;
import org.bukkit.craftbukkit.util.Waitable;
2019-04-27 06:26:04 +00:00
+
+// Paper start - JLine update
+import net.minecraft.server.DedicatedServer; // Paper
+import org.jline.reader.Candidate;
+import org.jline.reader.Completer;
+import org.jline.reader.LineReader;
+import org.jline.reader.ParsedLine;
+// Paper end
import org.bukkit.event.server.TabCompleteEvent;
public class ConsoleCommandCompleter implements Completer {
- private final CraftServer server;
+ private final DedicatedServer server; // Paper - CraftServer -> DedicatedServer
- public ConsoleCommandCompleter(CraftServer server) {
+ public ConsoleCommandCompleter(DedicatedServer server) { // Paper - CraftServer -> DedicatedServer
this.server = server;
}
+ // Paper start - Change method signature for JLine update
2019-05-06 02:58:04 +00:00
@Override
- public int complete(final String buffer, final int cursor, final List<CharSequence> candidates) {
+ public void complete(LineReader reader, ParsedLine line, List<Candidate> candidates) {
+ final CraftServer server = this.server.server;
+ final String buffer = line.line();
+ // Paper end
Waitable<List<String>> waitable = new Waitable<List<String>>() {
@Override
protected List<String> evaluate() {
2019-05-06 02:58:04 +00:00
@@ -29,25 +40,37 @@ public class ConsoleCommandCompleter implements Completer {
return tabEvent.isCancelled() ? Collections.EMPTY_LIST : tabEvent.getCompletions();
}
};
- this.server.getServer().processQueue.add(waitable);
+ server.getServer().processQueue.add(waitable); // Paper - Remove "this."
try {
List<String> offers = waitable.get();
if (offers == null) {
- return cursor;
+ return; // Paper - Method returns void
2019-04-27 06:26:04 +00:00
+ }
+
+ // Paper start - JLine update
+ for (String completion : offers) {
+ if (completion.isEmpty()) {
+ continue;
+ }
+
+ candidates.add(new Candidate(completion));
2019-04-27 06:26:04 +00:00
}
- candidates.addAll(offers);
+ // Paper end
2019-04-27 06:26:04 +00:00
+ // Paper start - JLine handles cursor now
+ /*
final int lastSpace = buffer.lastIndexOf(' ');
if (lastSpace == -1) {
return cursor - buffer.length();
} else {
return cursor - (buffer.length() - lastSpace - 1);
}
+ */
+ // Paper end
} catch (ExecutionException e) {
- this.server.getLogger().log(Level.WARNING, "Unhandled exception when tab completing", e);
+ server.getLogger().log(Level.WARNING, "Unhandled exception when tab completing", e); // Paper - Remove "this."
} catch (InterruptedException e) {
Thread.currentThread().interrupt();
}
- return cursor;
}
}
diff --git a/src/main/java/org/bukkit/craftbukkit/util/ServerShutdownThread.java b/src/main/java/org/bukkit/craftbukkit/util/ServerShutdownThread.java
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
index 70f8d42992..449e99d1b6 100644
--- a/src/main/java/org/bukkit/craftbukkit/util/ServerShutdownThread.java
+++ b/src/main/java/org/bukkit/craftbukkit/util/ServerShutdownThread.java
2019-04-27 06:26:04 +00:00
@@ -17,7 +17,7 @@ public class ServerShutdownThread extends Thread {
server.close();
} finally {
try {
- server.reader.getTerminal().restore();
+ net.minecrell.terminalconsole.TerminalConsoleAppender.close(); // Paper - Use TerminalConsoleAppender
} catch (Exception e) {
}
}
diff --git a/src/main/java/org/bukkit/craftbukkit/util/TerminalConsoleWriterThread.java b/src/main/java/org/bukkit/craftbukkit/util/TerminalConsoleWriterThread.java
deleted file mode 100644
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
index 99564fed7c..0000000000
--- a/src/main/java/org/bukkit/craftbukkit/util/TerminalConsoleWriterThread.java
+++ /dev/null
@@ -1,58 +0,0 @@
-package org.bukkit.craftbukkit.util;
-
2019-04-27 06:26:04 +00:00
-import com.mojang.util.QueueLogAppender;
-import java.io.IOException;
-import java.io.OutputStream;
-import java.util.logging.Level;
-import java.util.logging.Logger;
-import jline.console.ConsoleReader;
-import org.bukkit.craftbukkit.Main;
-import org.fusesource.jansi.Ansi;
-import org.fusesource.jansi.Ansi.Erase;
-
2019-04-27 06:26:04 +00:00
-public class TerminalConsoleWriterThread extends Thread {
- private final ConsoleReader reader;
- private final OutputStream output;
-
- public TerminalConsoleWriterThread(OutputStream output, ConsoleReader reader) {
- super("TerminalConsoleWriter");
- this.output = output;
- this.reader = reader;
2019-04-27 06:26:04 +00:00
-
- this.setDaemon(true);
- }
-
2019-05-06 02:58:04 +00:00
- @Override
- public void run() {
- String message;
-
- // Using name from log4j config in vanilla jar
- while (true) {
- message = QueueLogAppender.getNextLogEvent("TerminalConsole");
- if (message == null) {
- continue;
- }
-
- try {
- if (Main.useJline) {
- reader.print(Ansi.ansi().eraseLine(Erase.ALL).toString() + ConsoleReader.RESET_LINE);
- reader.flush();
- output.write(message.getBytes());
- output.flush();
-
- try {
- reader.drawLine();
- } catch (Throwable ex) {
- reader.getCursorBuffer().clear();
- }
- reader.flush();
- } else {
- output.write(message.getBytes());
- output.flush();
- }
- } catch (IOException ex) {
- Logger.getLogger(TerminalConsoleWriterThread.class.getName()).log(Level.SEVERE, null, ex);
- }
- }
- }
-}
diff --git a/src/main/resources/log4j2.component.properties b/src/main/resources/log4j2.component.properties
new file mode 100644
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
index 0000000000..0694b21465
--- /dev/null
+++ b/src/main/resources/log4j2.component.properties
@@ -0,0 +1 @@
+log4j.skipJansi=true
diff --git a/src/main/resources/log4j2.xml b/src/main/resources/log4j2.xml
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
index 722ca84968..620b9490e5 100644
--- a/src/main/resources/log4j2.xml
+++ b/src/main/resources/log4j2.xml
@@ -1,17 +1,14 @@
<?xml version="1.0" encoding="UTF-8"?>
<Configuration status="WARN" packages="com.mojang.util">
<Appenders>
- <Console name="SysOut" target="SYSTEM_OUT">
2019-04-27 06:26:04 +00:00
- <PatternLayout pattern="[%d{HH:mm:ss}] [%t/%level]: %msg%n" />
- </Console>
<Queue name="ServerGuiConsole">
<PatternLayout pattern="[%d{HH:mm:ss} %level]: %msg%n" />
</Queue>
- <Queue name="TerminalConsole">
- <PatternLayout pattern="[%d{HH:mm:ss}] [%t/%level]: %msg%n" />
- </Queue>
+ <TerminalConsole name="TerminalConsole">
+ <PatternLayout pattern="%highlightError{[%d{HH:mm:ss} %level]: %minecraftFormatting{%msg}%n%xEx}" />
+ </TerminalConsole>
<RollingRandomAccessFile name="File" fileName="logs/latest.log" filePattern="logs/%d{yyyy-MM-dd}-%i.log.gz">
- <PatternLayout pattern="[%d{HH:mm:ss}] [%t/%level]: %msg%n" />
+ <PatternLayout pattern="[%d{HH:mm:ss}] [%t/%level]: %minecraftFormatting{%msg}{strip}%n" />
<Policies>
<TimeBasedTriggeringPolicy />
<OnStartupTriggeringPolicy />
@@ -24,10 +21,9 @@
<filters>
<MarkerFilter marker="NETWORK_PACKETS" onMatch="DENY" onMismatch="NEUTRAL" />
</filters>
- <AppenderRef ref="SysOut" level="info"/>
<AppenderRef ref="File"/>
- <AppenderRef ref="ServerGuiConsole" level="info"/>
<AppenderRef ref="TerminalConsole" level="info"/>
+ <AppenderRef ref="ServerGuiConsole" level="info"/>
</Root>
</Loggers>
</Configuration>
--
Improve mid tick chunk loading, Fix Oversleep, other improvements Process loads outside of any canSleep check. Original intent was to only apply those restrictions to generations but realized I had some checks higher up the call chain. Reworked the back off strategy to just run every 1 millisecond per world, and to apply the per tick limit to generations only. This guarantees that your chunk will load with at most around 1ms delay. Additionally, fire midTick processing in a few more places, notably the oversleep section so we can keep processing loads here too which has a large up to 50ms window... Speaking of oversleep, we had a bug in our implementation changes for Timings that caused oversleep to not sleep the correct amount. Because we now moved it into the NEXT tick instead of THIS tick, the value of nextTick had already been increased to +50ms, resulting in the risk of sleeping more than it should, but, more importantly, this caused every task that was trying to NOT run during oversleep to actually run during oversleep. This is now fixed. Another small tweak is to the /tps command, to no longer show the star when TPS is right at 20. Due to ineffeciencies in the sleep precision, TPS is commonly 20.02. This causes the star to show up almost constantly, so now only show it if we actually hit a real "catchup". This commit also improves the changes to the CallbackExecutor, in that it now is also recursion safe. It was possible that the executor could run tasks out of desired order if the executor task scheduled more executor tasks. We solve this by ensuring new additions do not enter the currently iterated queue. Each depth level will have its own queue. Fixes #3220
2020-04-26 03:47:29 +00:00
2.26.2