paper previously migrated away from using executeIfPossible as this throws a
RejectedExecutionException when the server is shutting down, which is then picked
up by the Connection handler object and causes the player to be kicked without
the intended disconnection message that comes from commands such as /stop, /restart
This was fine, because previously changes made in spigot would just prevent these
packets from being executed anyways. Instead, we'll just use a marker exception
to try to detect this specific state.
Upstream has released updates that appear 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:
58ce1b0f Improve compatibility of new DamageSource API
I cannot test this due to lack of hardware (took 30 minutes getting paper
updated and loaded into IJ on this machine...), however, this is the easiest
patch for now. Ideally, some form of distance check would probably be added to
improve invalidation of this cache.
Upstream has released updates that appear 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:
465c4964 PR-843: Add damage methods to FallingBlock
46ba7c9f SPIGOT-7335: Fix typo in TextDisplay#TextAlignment enum name
CraftBukkit Changes:
818582f40 PR-1169: Add damage methods to FallingBlock
15a3eac66 SPIGOT-7335: Fix typo in TextDisplay#TextAlignment enum name
f01fb4979 SPIGOT-7336: Fix typo in internal method name
Upstream has released updates that appear 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:
de5ba402 PR-848: Mark get/setCatType methods deprecated in addition to the enum
CraftBukkit Changes:
50ee87fbc SPIGOT-7333: Correctly handle null in ColorableArmorMeta#setColor()
790be710b SPIGOT-7332: Persistent Data Container of Chunk not saving on stop or restart.
Upstream has released updates that appear 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:
aeb62aa3 Remove obsolete documentation from ArmorMeta#setTrim
CraftBukkit Changes:
cb6f3c028 PR-1173: Change CraftMinecartTNT to public for consistency
This patch will prevent sending expired keys to players when logging into a server, while the client does have some grace period to prevent issues, if a key has expired beyond the grace period, connecting clients will be disconnected from the server upon join. The client should update keys before this key even expires, and a new key will be broadcasted out if we somehow actually get one.
The paper config has taken precidence over this for a long time
with 0 complaints, and now it's tripping the deprecation warning,
even if timings is disabled in the paper config.
Upstream improperly used the enum name for looking up the vanilla entity
type to get the translation key; temp patch this until upstream fixes this,
also add some quick validation
adding the getHandle method causes some issues with things like ProtocolLib
due to the expectation that all classes which provide a getHandle method
are public.
Upstream has released updates that appear 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:
c7c11188 SPIGOT-2620: Add Player#sendBlockChanges()
f63d2b44 Improve annotation test on parameters
3372e7b4 SPIGOT-1244, SPIGOT-6860, SPIGOT-6874: Various Javadoc and formatting fixes
a1e8a9ab PR-793: Ignore .checkstyle file generated by Eclipse IDE
CraftBukkit Changes:
c2c39089e SPIGOT-2620: Add Player#sendBlockChanges()
8209158db PR-1113: Ignore .checkstyle file generated by Eclipse IDE
Spigot Changes:
4aa5ead2 Rebuild patches
* Make WorldCreator#keepSpawnLoaded return the WorldCreator (Fixes#8321)
Thought that I fixed this before push/merge, apparently not, hence, fix
the return type and add mitigation for this.
* Fix patch/build
* Rebuild patches
Vanilla uses a Random instance which is not thread safe inside of the
constructor for ServerLoginPacketListenerImpl, this is an issue because
instances of that class are created on the network worker threads, which
means that there is a potential (which we have reports confirming) that
multiple users logging on could trip this
in 1.19, mojang made it so that teleporations validate that an awaiting pos was
set in the server when teleporting, thus we need to ensure that this is set when
sending the player pos, otherwise the player will be kicked when the client sends
back the aknowledgement
I do wanna add some leeway to the kick, if message was sent within 30-60
seconds of the last message, assume clock drift and just ignore the kick
(send a message? just silently ignore?), but, I'm not 100% on this one
This is strictly not needed, however, I wanted to write this in part
to test the new forwarding logic, parsing and advertising the latest
featureset is also not exactly a bad thing
Upstream has released updates that appear to apply and compile correctly.
This update has not been tested by PaperMC and as with ANY update, please do your own testing
CraftBukkit Changes:
9d46b56e2 SPIGOT-7009: Stuck on "Loading terrain..." after disconnecting on respawn screen
This really needs a deeper look here, the way updates are handled is
fairly immature, but, this wasn't ever intended to be a large scale thing
Ideally, imho, we'd collect the list of update files into some form of Map,
that way we just have a reference of Name > File refs, and can filter out
cases where there are two versions of a plugin in there and warn expectidely,
but, that creates some complications, you would need to fall back to a dir
scan in the case of a plugin calling loadPlugin, but, it would at least
give us more defined behavior, as well as improve performance here vs
repeatidely trying to deserialise the plugin.yml defs for every file
in there on every load
lets debug mode throw a trace in order to potentially see where
such calls are cascading from easier, but, generally, if you see one setBlock
call, you're gonna see more, and this just potentially causes a flood of logs
which can cause issues for slower terminals, etc.
We can limit the flood by just allowing one for a single gen region,
we'll also only gen a trace for the first one, I see no real pressing need
to generate more, given that that would *massively* negate this patch otherwise
Upstream has released updates that appear 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:
5c21a63f PR-717: Add PlayerHideEntityEvent and PlayerShowEntityEvent
CraftBukkit Changes:
c26a7b54d PR-1000: Add PlayerHideEntityEvent and PlayerShowEntityEvent
fixes a CME due to this collection being shared across threads
This patch was in 1.16.5 for a good while and noticed somebody
getting impacted by the same thing
This commit causes an NPE when getting from the config in some states,
given upstream issue and PR in the works, I have 0 inclination to debug
this
This reverts commit e4358b8217126bbcc3a38b0d17097ad5ab87c50a.
Upstream has released updates that appear 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:
9115281f SPIGOT-6832: Improve Player#getPing docs
CraftBukkit Changes:
fd3478bc7 #967: Store last lava contact location for events
Spigot Changes:
dbf49382 Rebuild patches
58cb9d26 #113: Use simulationDistance for entity activation range base
Upstream has released updates that appear to apply and compile correctly.
This update has not been tested by PaperMC and as with ANY update, please do your own testing
CraftBukkit Changes:
03b725233 SPIGOT-6823: Fix loading custom world in combination with superflat
359d0533a #970: Correct typo in README.md
110492932 Fix per-world worldborder command
Upstream has released updates that appear 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:
9590b610 SPIGOT-6749: Library Loader does not correctly isolate dependencies
CraftBukkit Changes:
11c9299f #940: Fixed a NPE during CraftBlockEntityState#update.
960f3109 #937: Fixes related to unplaced BlockStates and the recent world generation changes.
4faf479e SPIGOT-6754: We ignore any still present TileEntity now when we create a BlockState for a block of type AIR.
a72d5404 SPIGOT-6754: Temporarily restore previous behaviour for tile entities with removed blocks
Spigot Changes:
dc75aca8 Remove redundant 'Log null TileEntity Owner' patch
This re-introduces the issue behind #5872 but fixes#6543
The logic here is generally flawed however somewhat of a nuance,
upstream uses managedBlock which is basically needed to process
the posted entity adds, but, has the side-effect of processing any
chunk loads which has the naunce of stacking up and either causing a
massive performance hit, or can potentially lead the server to crash.
This issue is particularly noticable on paper due to the cumulative efforts
to drastically improve chunk loading speeds which means that there is much more
of a chance that we're about to eat a dirtload of chunk load callbacks, thus
making this issue much more of an issue
Upstream has released updates that appear to apply and compile correctly.
This update has not been tested by PaperMC and as with ANY update, please do your own testing
CraftBukkit Changes:
6719d1f1 SPIGOT-6708: Fix Experience Orb value not being updated in mend event
Mojang made some changes to priorities in 1.17 and it seems that these changes
conflict with the changes made in this patch, which in some cases appears to
cause excessive rescheduling of tasks.
This, however, is not confirmed as such but seems to be the behavior that we're
seeing to cause this issue, if mojang has adopted the changes we suggested,
then a good chunk of this patch may be unneeded, but, this needs a much better
look than I'm currently able to do
Shane Freeder