This avoids us having to support requireless environments inside
cc.strings.
I do kinda wonder if os.loadAPI-loaded files should also have their own
shared "require", just so we're not loading 10 copies of cc.expect.
Build system:
- Switch to our new maven server. This has a cleaner separation between
published packages and mirrored packages, to avoid leaking those into
other people's builds.
- Update Gradle and Loom versions.
Code:
- Link to definitions instead in the breaking changes page.
- Fix several unused variable warnings.
Other:
- Remove unsupported Minecraft versions from the issue template.
Due to the earlier commits, the only functionality this block entity
adds is to register the command API. This commit:
- Add the command API when constructing the ServerComputer instead.
This is not a good long-term solution (I think we need to make API
factories more powerful), but is sufficient for now.
- Replace usages of CommandComputerBlockEntity with a normal
ComputerBlockEntity.
- Move the command permisssion checks to a new
ComputerFamily.checkUsable method (from
CommandComputerBlockEntity and ViewComputerMenu). I don't feel great
about putting new functionality in ComputerFamily (trying to move
away from it), but I think this is fine for now.
- Use this method from within the computer menu and computer block, to
check whether computers can be interacted with.
- Remove ViewComputerMenu, as it now no longer needs any special
is-usable logic.
Historically we used to have separate menu types for computers and
pocket computers, as the screen had to be initialised with the correct
terminal size.
However, as of c49547b962 (which was
admittedly two years ago now), we have the terminal available when
constructing the screen, and so the code for the two is identical.
This change actually merges the two screens, replacing usages of the
pocket computer UI with the computer one.
Historically (and according to the docs) getAudioTitle returned "false"
when the drive was empty (or had invalid media), and "null" when the
disk had no item. This was accidentally changed in a later refactor --
this change fixes that behaviour.
We keep getting bug reports on 1.20.1 about an Optifine bug that causes
Forge's capabilities to not work (#1458). The cause of this bug is not
immediately visible to users, and can be very confusing when hit.
Optifine have not released a fix for this bug (despite it being reported
a year ago), and we continue to receive bug reports about it.
Nobody likes it when mods complain about other mods. So much Minecraft
drama can be traced back to this, and it's a slippery slope to go down.
I've tried to keep this as unobtrusive as possible — it's just a chat
message at world join, and it'll turn off if the bug is fixed.
Turtles currently read their textures from a single 128x128 sprite
sheet. Most of this texture is unused which means we end up wasting a
lot of the block texture atlas[^1].
This change splits up the turtle textures into individual 32x32
textures[^2], one for each side, and then an additional backpack
texture.
I'm very sorry to any resource pack artists out there. The
tools/update-resources.py script will update existing packs, but does
not (currently) handle non-standard resolutions.
[^1]: It used to be worse: https://github.com/dan200/ComputerCraft/issues/145
[^2]: Turtle textures are a bit weird, in that they mostly *look* 16x16,
but have some detail in places.
In 1.20.1, Forge and Fabric have different "common" tag conventions (for
instance, Forge uses forge:dusts/redstone, while Fabric uses
c:redstone_dusts). This means the generated recipes (and advancements)
will be different for the two loader projects. As such, we run data
generators for each loader, and store the results separately.
However, aside from some recipes and advancements, most resources /are/
the same between the two. This means we end up with a lot of duplicate
files, which make the diff even harder to read. This gets worse in
1.20.5, when NeoForge and Fabric have (largely) unified their tag names.
This commit now merges the generated resources of the two loaders,
moving shared files to the common project.
- Add a new MergeTrees command, to handle the de-duplication of files.
- Change the existing runData tasks to write to
build/generatedResources.
- Add a new :common:runData task, that reads from the
build/generatedResources folder and writes to the per-project
src/generated/resources.
- Use enums for key and mouse actions, rather than integer ids.
- Change TerminalState to always contain a terminal. We now make
TerminalState nullable when we want to skip sending anything.
- Update Gradle to 8.7
- Configure IntelliJ to build internally, rather than delgating to
Gradle. We've seen some weird issues with using delegated builds, so
best avoided.
- Remove gitpod config. This has been broken for a while (used Java 16
rather than 17) and nobody noticed, so I suspect nobody uses this.
- Add the core TeaVM jar to the runtime the classpath, to ensure
various runtime classes are present.
- Fix computer initialisation errors not being displayed on the screen.
The terminal was set to the default 0x0 size when logging the error,
and so never displayed anything!
Rather than handling right clicks within the block entity code, we now
handle it within the block. Turtles now handle the nametagging
behaviour themselves, rather than overriding canNameWithTag.
Minecraft.hitResult may /technically/ be null when rendering a turtle.
In vanilla, this doesn't appear to happen, but other mods (e.g.
Immersive Portals) may still take advantage of this.
This hitResult is then propagated to BlockEntityRenderDispatcher, where
the field was /not/ marked as nullable. This meant we didn't even notice
the potential of an NPE!
Closes#1775
This fixes several issues with @Nullable fields not being checked. This
is great in principle, but a little annoying in practice as MC's
@Nullable annotations are sometimes a little overly strict -- we now
need to wrap a couple of things in assertNonNull checks.
This theoretically allows you to use the emulator to run the test suite
(via --mount-ro projects/core/src/test/resources/test-rom/:test-rom),
but not sure how useful this is in practice.
This tells Create that modems will pop-off if their neighbour is moved,
and so changes the order that the block is moved in.
We possibly should use BlockMovementChecks.AttachedCheck instead, to
properly handle the direction modems are facing in. However, this
doesn't appear to be part of the public API, so probably best avoided.
Fixes#948
When the terminal data is not present, width/height are set to 0, rather
than the terminal's width/height. This meant we'd create an empty
terminal, which then crashes when we try to render it.
We now make the terminal nullable and initialise it the first time we
receive the terminal data. To prevent future mistakes, we hide
width/height, and use TerminalState.create everywhere.
Fixes#1765
- Mention the timer event in os.startTimer. Really we should have a
similar example here too, but let's at least link the two for now.
- Fix strftime link
I have mixed feelings about speaker.playSound. On one hand, it's pretty
useful to be able to play any sound. On the other, it sometimes feels
... maybe a little too magic?
One particular thing I don't like is that it allows you to play
arbitrary records, which sidesteps both a vanilla mechanic (finding
record discs) and existing CC functionality (disk.playAudio). We now
prevent playing record tracks from the speaker.
In 5d8c46c7e6, we switched to using UUIDs
for looking up computers (rather than an integer ID). However, for
compatibility in some of the command code, we need to maintain the old
integer lookup map.
Most of the code was updated to handle this, *except* the code to remove
a computer from the registry. This meant that we'd fail to remove a
computer from the UUID lookup map, so computers ended up in a phantom
state where they were destroyed, but still accessible.
This is not an issue on 1.20.4, because the legacy int lookup map was
removed.
Fixes#1760
The two mod loaders expose different methods for this (Forge's method
takes a ItemPropertyFunction, Fabric's a ClampedItemPropertyFunction).
This is fine in a Gradle build, as the methods are compatible. However,
when running from IntelliJ, we get crashes as the common code tries to
reference the wrong method.
We now pass in the method reference instead, ensuring we use the right
method on each loader.
BYTECODE WAS NOT SUPPOSED TO BE REWRITTEN
YEARS OF DEBUGGING REMAPPING FAILURES yet NO ACTUAL SOLUTION FOUND.
Wanted to use Mixins for anyway for a laugh? We had a tool for that: it
was called "FABRIC LOOM".
"Yes, please produce completely broken jars for no discernable reason"
Statements dreamed up by the utterly Deranged.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
This removes our two mixins used on Forge:
- Breaking progress for cabled/wired modems.
- Running client commands from chat click events. We now suggest the
command on Forge instead.
Occasionally we get issues where the mixin annotation processor doesn't
write its tsrg file in time for the reobfJar/reobfJarJar task. I thought
we'd fixed that cb8e06af2a, but sometimes
we still produce missing jars - I have a feeling this might be to do
with incremental compilation.
We can maybe re-evaluate this on 1.20.4, where we don't need to worry
about remapping any more.
We were seeing some strange issues in the Fabric test code where we
tried to load the implementation from a different classloader. This
ensures that the classloaders are consistent.
Due to the asynchronous nature of main-thread tasks, it's possible for
them to be executed on peripherals which have been detached. This has
been known for a long time (#893 was opened back in 2021), but finding a
good solution here is tricky.
Most of the time the method will silently succeed, but if we try to
interact with an IComputerAccess (such as in inventory methods, as seen
in #1750), we throw a NotAttachedException exception and spam the logs!
This is an initial step towards fixing this - when calling a peripheral
method via peripheral.call/modem.callRemote, we now wrap any enqueued
main-thread tasks and silently skip them if the peripheral has been
detached since.
This means that peripheral methods may start to return nil when they
didn't before. I think this is *fine* (though not ideal for sure!) - we
return nil if the peripheral has been detached, so it's largely
equivalent to that.
Double chests peripherals were getting reattached every time there was a
block update, as the inventories were not comparing equal (despite being
so!). We now check for a couple of common cases, which should be enough
for vanilla/vanilla-like inventories.
I actively Do Not Like This Code, but do not see a good alternative.
This should never happen, but apparently it does!? We now log an error
(rather than crashing), and include the original BE (and associated
block), as the BE type isn't very useful.
See #1750. Technically this fixes it, but want to do some more poking
there first.
Here's a fun bug you can try at home:
- Create a new world
- Spawn in a pocket computer, turn it on, and place it in a chest.
- Reload the world - the pocket computer in the chest should now be
off.
- Spawn in a new pocket computer, and turn it on. The computer in chest
will also appear to be on!
This bug has been present since pocket computers were added (27th March,
2024).
When a pocket computer is added to a player's inventory, it is assigned
a unique *per-session* "instance id" , which is used to find the
associated computer. Note the "per-session" there - these ids will be
reused if you reload the world (or restart the server).
In the above bug, we see the following:
- The first pocket computer is assigned an instance id of 0.
- After reloading, the second pocket computer is assigned an instance
id of 0.
- If the first pocket computer was in our inventory, it'd be ticked and
assigned a new instance id. However, because it's in an inventory, it
keeps its old one.
- Both computers look up their client-side computer state and get the
same value, meaning the first pocket computer mirrors the second!
To fix this, we now ensure instance ids are entirely unique (not just
per-session). Rather than sequentially assigning an int, we now use a
random UUID (we probably could get away with a random long, but this
feels more idiomatic).
This has a couple of user-visible changes:
- /computercraft no longer lists instance ids outside of dumping an
individual computer.
- The @c[instance=...] selector uses UUIDs. We still use int instance
ids for the legacy selector, but that'll be removed in a later MC
version.
- Pocket computers now store a UUID rather than an int.
Related to this change (I made this change first, but then they got
kinda mixed up together), we now only create PocketComputerData when
receiving server data. This makes the code a little uglier in some
places (the data may now be null), but means we don't populate the
client-side pocket computer map with computers the server doesn't know
about.
- Remove "initial connections" flag, and just refresh connections +
peripherals on the first tick.
- Remove "peripheral attached" from NBT, and just read/write it from
the block state. This might cause issues with #1010, but that's
sufficiently old I hope it won't!
Our GatedPredicate hack was clever, but also fundamentally didn't work.
The predicate is called before extraction, so if extraction fails (for
instance, canTakeItemThroughFace returns false), then we still think an
item has been removed.
To fix that, we inline StorageUtil.move, specialising it for what we
need.
This feels a little overkill, but nice to standardise how this code
looks.
There's a bit of me which wonders if we should remove
IPeripheral.equals, and just use Object.equals, but I do also kinda like
the explicitness of the current interface? IDK.
This ensures the client decoder is in sync with the server. Well, mostly
- we don't handle the anti-jerk, but that should correct itself within a
few samples.
Fixes#1748
The original runtime error reporting PR[^1] added a "cc.exception"
module, which allowed coroutine managers (such as parallel) to throw
rich errors, detailing the original context where the error was thrown.
Unfortunately, the change to parallel broke some programs (>_>, don't do
string pattern matching on your errors!), and so had to be reverted,
along with the cc.exception module.
As a minimal replacement for this, we add support for user-thrown
exceptions within our internal code. If an error object "looks" like an
exception ("exception" __name, and a message and thread field), then we
use that as our error information instead.
This is currently undocumented (at least in user-facing documentation),
mostly because I couldn't figure out where to put it - the interface
should remain stable.
[^1]: https://github.com/cc-tweaked/CC-Tweaked/pull/1320
This adds support for computer selectors, in the style of entity
selectors. The long-term goal here is to replace our existing ad-hoc
selectors. However, to aid migration, we currently support both - the
previous one will most likely be removed in MC 1.21.
Computer selectors take the form @c[<key>=<value>,...]. Currently we
support filtering by id, instance id, label, family (as before) and
distance from the player (new!). The code also supports computers within
a bounding box, but there's no parsing support for that yet.
This commit also (finally) documents the /computercraft command. Well,
sort of - it's definitely not my best word, but I couldn't find better
words.
When rendering non-origin monitors, we would fetch the origin monitor,
read its client state, and then cache that on the current monitor to
avoid repeated lookups.
However, if the origin monitor is unloaded/removed on the client, and
then loaded agin, this cache will be not be invalidated, causing us to
render both the old and new monitor!
I think the correct thing to do here is cache the origin monitor. This
allows us to check when the origin monitor has been removed, and
invalidate the cache if needed.
However, I'm wary of any other edge cases here, so for now we do
something much simpler, and remove the cache entirely. This does mean
that monitors now need to perform extra block entity lookups, but the
performance cost doesn't appear to be too bad.
Fixes#1741
Historically, computers tracked whether any world-visible state
(on/off/blinking, label and redstone outputs) had changed with a single
"has changed" flag. While this is simple to use, this has the curious
side effect of that term.setCursorBlink() or os.setComputerLabel() would
cause a block update!
This isn't really a problem in practice - it just means slightly more
block updates. However, the redstone propagation sometimes causes the
computer to invalidate/recheck peripherals, which masks several other
(yet unfixed) bugs.
- colors.toBlit now performs bounds checks on the passed value,
preventing weird behaviour like color.toBlit(2 ^ 16) returning "10".
- The window API now uses colors.toBlit (or rather a copy of it) for
parsing colours, allowing doing silly things like
term.setTextColour(colours.blue + 5).
- Add some top-level documentation to the term API to explain some of
the basics.
Closes#1736
