Please don't talk to me about this. The first couple of hours of this
update were quite enjoyable, and then the rest was one of the most
miserable times I've had modding.
This has been a real slog, partly due to some large MC changes (item
models are a great change, but a pain to adapt to), and partly due to
mental health reasons — honestly, I've opened up my IDE so many times,
and then just closed it because I've hated the thought of even working
on this.
I will publish this to my maven, so mod authors can depend on it, but I
have no plans to publish a 1.21.4 version. 1.21.5 is right around the
corner (again, with some cool, but no-doubt painful changes), and I
need some time to focus on some breaking changes.
This commit actually includes the 1.21.3 update — the git history got so
messy here, so I just clobbered the whole thing. Sorry.
== Rendering ==
- Remove TBO monitor renderer: There was a big overhaul to how shaders
are defined and loaded in 1.21.2. It might have been possible to
update the monitor shader code to this version, it doesn't see much
use nowadays, so let's just delete it.
This is a real shame — the TBO renderer was one of my favourite
projects I've worked on. Unfortunately, it just doesn't seem worth
the ongoing maintenance burden. It lives on in the standalone
emulator :D.
- Similarly, the VBO rendering code got a bit of an overhaul. We no
longer use a custom VBO subclass, and instead just hack vanilla's to
support changing the number of vertices rendered.
This does mean we need to construct a MeshData, rather than a raw
ByteBuffer. This isn't too hard, but not sure how it'll play with
Iris. Given recent vanilla performance improvements, maybe we can
remove our Unsafe code and use a normal BufferBuilder now.
- Remove our custom emissive model code, now that vanilla supports
it. We should add emissive textures to some other models at some
point.
- Remove mod-loader specific model code, and replace it with vanilla's
ItemModel. This does constrain the design of turtle upgrade modellers
quite a bit — we now only accept an untransformed BakedModel or a
transformed ItemStack model. We may relax this in the future,
unclear.
This change does mean that updsidedown turtles are broken. RIP :(.
- Entity rendering now separates reading state from the entity from
actual rendering. This means we need to pass some extra state around
for item frames. Easy on Forge, but requires a mixin on Fabric.
== Recipes ==
There were several major changes to ingredients this update. The code
here hasn't been very well tested right now — might be nice to add some
game tests for this.
- Ingredients can no longer be constructed directly from a tag key (it
needs to be fetched from the current registries), so the recipe
generation code needs a bit of a reshuffle.
- DiskRecipe now accepts a custom list of ingredients, rather than
being hard-coded (fixes#1755). Recipes can now return custom
`RecipeDisplay`s used to show a recipe in the crafting book. We use
this to replace the impostor recipes.
I'm not entirely sure how well this'll play with other recipe
mods. Here's hoping.
- Similarly, our recipe mod integration has been updated to use
RecipeDisplay. We had to do this as ingredients no longer accept
arbitrary ItemStacks (only a specific item).
== Misc ==
- Blocks/items now need to know their ID ahead of time (so they can
compute their description). This requires some reshuffling to the
registration code, but it's pretty minor.
- updateShape and neighborChanged no longer take a direction (the
Orientation is mostly null) and so invalidates all redstone and
peripherals.
- All the positions were lowered by one in game tests. It's a good
change (they now match the positions in structures), but annoying to
update for!
This bumps them to be 48x48, which allows them to be downscaled to a
mipmap level of 4. We possibly should bump these to be 64x64 (actual
power of two), but I kinda want to avoid that, as it's so much wasted
space. If this does become a problem, we should probably put these on a
separate atlas instead.
Honestly, the whole design around volume and playSound/playAudio is a
little janky — it probably should be a separate setVolume method which
updates directly. But too late to change that now, so let's do what we
can.
See #2108
ComputerItem was just an empty subclass, so no sense keeping around. We
probably could get rid of CommandComputerItem too (just do an `instanceof
GameMasterBlock` in `getPlacementState`), but there's no rush.
We now register these separately, rather than relying on the implicit
IMedia. This allows us to share a bit more logic between
PocketComputerItem and AbstractComputerItem. This doesn't make much
difference on 1.20.1, but does help a bit more on 1.21.1.
Now, hear me out, what if instead of having three @Nullable annotations,
we had *four*?
I've been wanting to switch away from javax.annoations for a while. The
library has been deprecated for ever and, unlike other @Nullable
annotations, the annotation is attached to the parameter/function
itself, rather than the type.
We use JSpecify rather than one of the alternatives (JetBrains,
CheckerFramework) mostly because it's what NullAway recommends. We keep
CheckerFramework around for @DefaultQualifier, and JB's for @Contract.
There are some ugly changes here — for instance, `@Nullable byte[]` is
replace by `byte @Nullable`, and `@Nullable ILuaMachine.Factory` is
`ILuaMachine.@Nullable Factory`. Ughr, I understand why, but it does not
spark joy :).
We now suggest alternative table keys when code errors with "attempt
to index/call 'foo' (a nil value)". For example: "redstone.getinput()",
will now suggest "Did you mean: getInput".
This is a bit tricky to get right! In the above example, our code reads
like:
1 GETTABUP 0 0 0 ; r0 := _ENV["redstone"]
2 GETFIELD 0 0 1 ; r0 := r0["getinput"]
3 CALL 0 1 1 ; r0()
Note, that when we get to the problematic line, we don't have access to
the original table that we attempted to index. In order to do this, we
borrow ideas from Lua's getobjname — we effectively write an evaluator
that walks back over the code and tries to reconstruct the expression
that resulted in nil.
For example, in the above case:
- We know an instruction happened at pc=3, so we try to find the
expression that computed r0.
- We know this was set at pc=2, so we step back one. This is a GETFIELD
instruction, so we check the key (it's a constant, so worth
reporting), and then try to evaluate the table.
- This version of r0 was set at pc=1, so we step back again. It's a
GETTABUP instruction, so we can just evaluate that directly.
We then use this information (indexing _ENV.redstone with "getinput") to
find alternative keys (e.g. getInput, getOutput, etc...) and then pick
some likely suggestions with Damerau-Levenshtein/OSD.
I'm not entirely thrilled by the implementation here. The core
interpretation logic is implemented in Java. Which is *fine*, but a)
feels a little cheaty and b) means we're limited to what Lua bytecode
can provide (for instance, we can't inspect outer functions, or list all
available names in scope). We obviously can expand the bytecode if
needed, but something we'd want to be careful with.
The alternative approach would be to handle all the parsing in
Lua. Unfortunately, this is quite hard to get right — I think we'd need
some lazy parsing strategy to avoid constructing the whole AST, while
still retaining all the scope information we need.
I don't know. We really could make this as complex as we like, and I
don't know what the right balance is. It'd be cool to detect patterns
like the following, but is it *useful*?
local monitor = peripheral.wrap("left")
monitor.write("Hello")
-- ^ monitor is nil. Is there a peripheral to the left of the
-- computer?
For now, the current approach feels the easiest, and should allow us to
prototype things and see what does/doesn't work.
In the original implementation of our prettier runtime errors (#1320), we
wrapped the errors thrown within parallel functions into an exception
object. This means the call-stack is available to the catching-code, and
so is able to report a pretty exception message.
Unfortunately, this was a breaking change, and so we had to roll that
back. Some people were pcalling the parallel function, and matching on
the result of the error.
This is a second attempt at this, using a technique I've affectionately
dubbed "magic throws". The parallel API is now aware of whether it is
being pcalled or not, and thus able to decide whether to wrap the error
into an exception or not:
- Add a new `cc.internal.tiny_require` module. This is a tiny
reimplementation of require, for use in our global APIs.
- Add a new (global, in the debug registry) `cc_try_barrier` function.
This acts as a marker function, and is used to store additional
information about the current coroutine.
Currently this stores the parent coroutine (used to walk the full call
stack) and a cache of whether any `pcall`-like function is on the
stack.
Both `parallel` and `cc.internal.exception.try` add this function to
the root of the call stack.
- When an error occurs within `parallel`, we walk up the call stack,
using `cc_try_barrier` to traverse up the parent coroutine's stack
too. If we do not find any `pcall`-like functions, then we know the
error is never intercepted by user code, and so its safe to throw a
full exception.
This allows shift+clicking a pocket computer on to a lectern. These
computers can be right clicked, opening the no-term computer GUI.
Terminal contents is rendered in-world, and broadcast to everyone in
range.
- Add a new lectern PocketHolder.
- Refactor some of the `PocketItemComputer` code to allow ticking pocket
computers from a non-player/entity source.
- Add a new model for pocket computers. This requires several new
textures (somewhat mirroring the item ones), which is a little
unfortunate, but looks much better than reusing the map renderer or
item form.
In 94ad6dab0e5b8d9eb65467dd1b635d708ce58b53, we changed it so typing
characters outside of CC's codepage were replaced with '?' rather than
ignored. This can be quite annoying for non-European users (where latin1
isn't very helpful!), so it makes sense to revert this change.
See discussion in #860 for more context.
I removed this in fc834cd97fe941a192e40962ac3bb27be102ce09, way back in
late 2024. Looks like it's been updating in the meantime and I hadn't
noticed, so add it back.
I've simplified the code a little bit, to make use of our NeoForge's new
capability system, but otherwise it's almost exactly the same :D.
- Store the filter alongside the coroutine rather than in a separate
table (like we do in multishell).
- Remove the redudant (I think!) second loop that checks for dead
coroutines. We already check for dead coroutines in the main loop.
- Rename some variables to be a bit more consistent. This makes this
commit look noisier than it is. Sorry!
It's not actually safe to reuse this, as we need to recompute the
internal StackedContents each time the inventory changes, otherwise
ShapelessRecipe.matches will continue to return true, even if the actual
inventory doesn't include the required items.
Fixes#2094
Oh, this is so broken, and really has been since the 1.13 update, if not
earlier.
- Fix call to isUnobstructed using the bounding box of the
*destination* block rather than the turtle. This is almost always
air, so the box is empty.
- Because the above check has been wrong for so many years, we now
significantly relax the "can push" checks for entities. We now allow
pushing entities in any direction.
We also remove the "isUnobstructed" check for the destination entity
pos. This causes problems (if two entities are standing on a turtle,
they'll obstruct each other), and given pistons don't perform such a
check, I don't think we need it.
- Also do a bit of cleanup around air/liquid checks. We often ended up
reading the block state multiple times, which is a little ugly.
I'd originally kept this around to prevent mods that use CC internals,
but fa2140d00ba667902d361b4873804dd97dde3a19 has probably broken those
anyway, so let's not worry too much.
- Remove ContainerData.open.
- Change PlatformHelper.openMenu to take a separate display name and
MenuConstructor, rather than a MenuProvider. This makes the interface
slightly easier to use in the common case, where we want to use
lambdas instead.
- Check whether the computer is a command computer before registering
the capability.
- Add tests to check what is/isn't a peripheral. See also #2020, where
we forgot to register a peripheral on NeoForge 1.21.1.
Fixes#2070.
This isn't required in vanilla, as the command computer is a
GameMasterBlock, and so isn't placeable in the first place.
*However*, this is a problem with Create contraptions — with those it's
possible to "place" a command computer complete with NBT. We override
onlyOpCanSetNbt to prevent this [^1].
[^1]: 7a7993deb8/src/main/java/com/simibubi/create/foundation/utility/NBTProcessors.java (L179)
This adds a new "computercraft:storage_capacity" component to items (and
"Capacity" NBT tag to BEs), that overrides the capacity for the given
item.
Fixes#1814
This is where vanilla will read the sprites from in future versions, so
means we have a consistent layout between versions.
Also move the turtle "selected slot" texture to a sprite sheet. It would
be good to do more of these in the future (e.g. printer progress, maybe
bits of printouts).
Sorry to resource pack artists for causing trouble again.
We currently need to pass a whole bunch of arguments to a ServerComputer
in order to construct it, and if we implement #1814, this will get a
whole lot worse. Instead, we now pass most parameters (computer id,
family, label, term size, components) via a separate Properties class,
much like Minecraft does for blocks and items.
I'm not wild about the design of the API here, but I think it's a step
in the right direction.
We now convert uncode characters from "char" and "paste" events to CC's
charset[^1], rather than just leaving them unconverted. This means you
can paste in special characters like "♠" or "🮙" and they will be
converted correctly. Characters outside that range will be replaced with
"?", as before.
It would be nice to make this a bi-directional mapping, and do this for
Lua methods too (e.g. os.setComputerLabel). However, that has much wider
ramifications (and more likelyhood of breaking something), so avoiding
that for now.
- Remove the generic "queue event" client->server message, and replace
it with separate char/terminate/paste messages. This allows us to
delete a chunk of code (all the NBT<->Object conversion), and makes
server-side validation of events possible.
- Fix os.setComputerLabel accepting the section sign — this is treated
as special by Minecraft's formatting code. Sorry, no fun allowed.
- Convert paste/char codepoints to CC's charset. Sadly MC's char hook
splits the codepoint into surrogate pairs, which we *don't* attempt
to reconstruct, so you can't currently use unicode input for block
characters — you can paste them though!
[^1]: I'm referring this to the "terminal charset" within the code. I've
flip-flopped between "CraftOS", "terminal", "ComputerCraft", but feel
especially great.
This abstraction never made much sense on InputHandler, as we only leave
the default methods on ServerComputer.
We now add a new class (ComputerEvents), which has a series of *static*
methods, that can queue an event on a ComputerEvents.Receiver object.
This is a bit of an odd indirection (why not just make them instance
methods on Receiver?!), but I don't really want those methods leaking
everywhere.
This occurs when syncing the server config to the client. Ideally we'd
not hit this code path in the first place, but unfortunately there's no
way to tell where the config file comes from.
Fixes#2065
As part of this, we also rewrite some of the turtle placing code, and
how it uses the turtle_can_use tag:
Minecraft 1.21 cleaned up the item/block clicking code a little bit,
splitting Block.use into Block.useItemOn and Block.useWithoutItem. The
first of these is pretty much exactly what we wanted in the first place,
so the tag was kinda redundant and we commented it out in the 1.21
update.
This was never meant to be a long-term fix, but time has gone by anyway.
We now check that tag, and call useWithoutItem() if present —
effectively restoring the previous behaviour.
Fixes#2011
When "placing" the item (e.g. hoeing soil), we were using the tool item,
rather than the passed stack. This was introduced in
9a914e75c4a4e122c16aac5d010059d28b475b5e, so never made it into a
release.
