Many thanks to Lem for managing to reproduce it. It was actually an easy
bug bug to spot on second look, but having a reliable way to verify was
super helpful.
Fixes#1338
While it is a really nice library, it ends up being a bit overkill for
our needs - we don't need config syncing or anything. By NIHing our own,
we can drop one dependency and ease the updating burden a little.
Closes#1296
Mostly in prep for 1.19.4.
- Update to Loom 1.1.
- Simplifies our handling of remapped configurations a little.
- Removes the need for a fake fabric.mod.json in the API jar.
For reasons I don't quite understand, this required us to bump the
Fabric API version. Otherwise interfaces are not injected.
- Update to Rollup 3.0.
- Do NOT update NullAway: It now correctly checks @Nullable fields in
inherited classes. This is good, but also a pain as Minecraft is a
little over-eager in where it puts @Nullable.
We define a tag which allows specifying which blocks can be used. Right
now this is is just cauldrons and hives, as they have "placing into"
semantics.
Closes#1305. Many thanks to Lindsay-Needs-Sleep for their initial work
on this!
Fixes#1008. I believe also fixes#854.
- We now write to a "ids.json.new" file, then move that on top of the
original ids.json file instead.
- Use FileChannel.force to ensure the new file is properly flushed to
disk. I can't really guarantee this'll work with the later
Files.move, but it's better than not doing it!
Closes#1346.
See the discussion in #1352 - Netty uses the system one by default,
so no sense creating our own.
Also make sure we through the HTTP error every time, not just on the
first failure. Otherwise we get cryptic connection dropped errors.
Given an input like f(x), which is both a valid statement and
expression, both parsers would accept the whole input. However, this was
treated the same as both parsers rejecting the input, resulting in a
crash when trying to print the error.
We now return immediately when any parser accepts the input.
Fixes#1354
Translations for Italian
Translations for French
Co-authored-by: Alessandro <ale.proto00@gmail.com>
Co-authored-by: RomanPlayer22 <romansulc13122006@gmail.com>
Co-authored-by: chesiren <chesiren63@gmail.com>
If someone had a recursive table (created with an IIFE), then we'd throw
an error inside reserialize. We now catch this error and silently drop
the value.
I'm not thrilled by this behaviour - there's an argument we should
return false instead - but it's consistent with what we currently do.
Closes#1337.
Historically, the VBO was an order of magnitude slower than the TBO
renderer. However, as of fccca22d3f, the
difference is much smaller now. While TBOs /are/ still faster, this only
has a measurable impact in extreme stress tests, and so isn't worth the
occasional issues which occur.
I'm still keeping the code around for now: I'm incredibly fond of it,
even three years later. I may end up re-evaluating this the next time
Minecraft's rendering code changes :D.
This also adds a line to the debug screen showing the current monitor
renderer, though maybe less useful now that almost everyone will be
using VBOs!
Previously it was possible to access all methods of the multishell
redirect by calling term.current(). This is definitely not intended, as
it leaks all sorts of internals to the user.
Also bump illuaminate - the new version is about twice as fast on my
machine.
Woops!
- Fix the REPL not printing values, as exception.try didn't return
values. It did originally, and then I tried to simplify it >_>
- Change repl_exprs to run an expression and program parser in
parallel, rather than handling the parallelism on the grammar side -
that has a few shift/reduce conflicts which result in bad parse
errors.
- Bump Cobalt to 0.6.0. We now track both line and column numbers of
each bytecode instruction, allowing us to map an error to a concrete
position.
- `loadfile` (and similar functions) now use the full path, rather than
the file name. Cobalt truncates this to 30 characters (rather than
the previous 60) so this should be less noisy.
- The shell, edit and Lua REPL now display the corresponding source
code alongside an error.
Note this is incredibly limited right now - it won't cope with errors
which cross coroutine boundaries. Supporting this is on the roadmap,
but requires some careful API design.
While it's a nice immersive interaction, it's far too easy to trigger by
accident. Dying/undying can now only be done via crafting and the
cauldron.
Closes#1321
It's much less aggressive than dunking it in a cauldron, so won't damage
any of your precious electronics.
I had this idea back in 2017 (dan200/ComputerCraft#230). Can't believe
it took me almost 6 years to implement.
In older versions we just used a hard-coded list of items and
superclasses. This was somewhat ugly, and so in 1.19.3 I tried to make
this code more generic.
However, this has a lot of unintended consequences - for instance
turtles can now throw ender pearls, which is definitely not intended!
By using a tag, we can emulate the old behaviour, while still allowing
modders and pack devs to add additional items if needed.
- Encode the DFA as a virtual machine (identical to lrgrep) rather than
compiling it to a series of Lua functions. While this is a little
slower and uglier, it's much more space efficient, shaving off 16Kb.
- Minimise the DFA properly. This only shaves off a few states, but
every little helps.
- Run the error handling code from a non-reduced parser stack. This was
incredibly nasty to get right (and positions are still not correctly
handled), but it fixes several broken error messages.
- Move modem recipes out of the usage section.
- Add missing argument names to BinaryWriableHandle.write. Illuaminate
really should catch this, but for now I did a grep and couldn't find
any more instances of this.
- Add several (internal) modules for lexing and parsing Lua code. These
allow us to provide (hopefully) higher quality error messages than
Lua's built-in messages.
- `shell.run`, `edit` and `lua` now use this parser when fed invalid
code. This allows us to provide better syntax errors, while not
having any impact on the happy path.
Note this does not affect any other mechanism for loading code
(`load`, `require`, `dofile`).
There's still a lot of work to do here in improving error message
quality, but hopefully this provides a good starting point.
- Document the thread safety of DetailRegistry a little better.
- Turtles now duplicate their inventory to the "previous
inventory" (now called inventorySnapshot) immediately, rather than
when the block is ticked.
This is slightly more resource intensive, but I don't think it's so
bad we need to worry.
- As this snapshot is now always up-to-date, we can read it from the
computer thread. Given the item is immutable, it's safe to read NBT
from it.
_Technically_ this is not safe under the Java memory model, but in
practice I don't think we'll observe the wrong value.
Closes#1306
Just ran[^1] over the codebase. Turns out we'd duplicated one of the
changelog entries entirely - I suspect due to a version merge gone
wrong!
[^1]: https://github.com/crate-ci/typos/
When the peripheral is attached, we add the computer to the map and
queue the actual disk to be mounted next tick. This avoids the
thread-safety issues with mutating the item (and creating disk ids) that
might be caused by doing it on the computer thread.
The mount is now also managed separately to the MediaStack, as that was
meant to be an immutable snapshot of the item!
Fixes#1282
This gives us slightly better guarantees that the config has actually
been loaded. This, along with a FCAP bump, fixes this config option
not doing anything on Fabric.
I originally put cc.import in a separate directory from the main
modules. This means that programs must extend the package path in order
to import these modules.
However, this ends up being a mixed blessing: while it makes it much
harder for users to accidentally require user code, it also means we
can't expose a public interface which wraps a private module.
Instead, cc.import now lives on the main package path, but lives under
the cc.internal namespace and is not documented anywhere. Hopefully this
should be enough of a clue that one shouldn't use it :p.
I was having issues where dropped items would clip into blocks when
dropped, and then phase upwards through the turtle instead. This makes
things a little more consistent with dispenser behaviour.
Instead of taking control of the breaking logic in all cases, we now
only do so when we have both a cable and modem. This allows us to fall
back to default vanilla behaviour and so correctly drop the modem/cable
item.
We now call getLevel() when attaching the peripheral, so need the
position to be available immediately. Fixes#1274.
I /think/ the entity should always be present, as peripherals are
only created on startup or when calling pocket.equipBack, both of which
require a player.[^1]
I suspect this was a little broken before (the level wouldn't be
available if a modem received a message before the position had
been set), but that would be much rarer.
I'm not 100% convinced about the thread-safety of this code (the writes
to level may not be immediately visible on other threads), so may need
to think about that.
[^1]: Note that when peripherals come to be /attached/ they may no
longer have a player (as there's a gap between turning a computer on and
it actually starting). However, the level/position will have been
initialised by then, so this isn't a problem.
Use `InteractionResult.sidedSuccess` / `.CONSUME` where applicable instead of `.SUCCESS`. This prevents the server from sending an additional swing animation packet to the client. Normally this isn't a problem, since the client will de-duplicate swing packets if they are within the animation duration of the currently playing swing; however, when connected to a server with a high ping the packet is sent after the animation is already finished on the client, resulting in a duplicate animation.
We now perform movement translations before rendering the label, rather
than afterwards. This means the label moves smoothly(ish), rather than
jumping from block to block.
I've been meaning to fix this for over 6 years, and just kept
forgetting.
Previously ILuaContext.executeMainThreadTask worked by running
ILuaContext.issueMainThreadTask, pulling task_complete events, and then
returning the results.
While this makes the implementation simple, it means that the task's
results were converted into Lua values (in order to queue the event) and
then back into Java ones (when the event was pulled), before eventually
being converted into Lua once more.
Not only is this inefficient, as roundtripping isn't lossless, you
couldn't return functions or rich objects from main thread functions
(see https://github.com/dan200/ComputerCraft/issues/125).
We now store the return value on the Java side and then return that when
the receiving the task_complete event - the event no longer carries the
result. Note this does not affect methods using issueMainThreadTask!
- Fix client classes not being included in Forge.
- Only remap Nettty's HTTP classes, not all of them. This feels a
little more error prone - maybe we should jar-in-jar this in the
future.
- Use the correct refmaps on Forge.
- Prevent the Fabric jar pulling in some other mods.
Closes#1247
- Correctly handle FileOperationExceptions for the root mount.
- Remove some checks from MountWrapper: Mount/WritableMount should do
these already!
- Normalise file paths, always using a '/'.
- Separate FileMount into separate FileMount and WritableFileMount
classes. This separates the (relatively simple) read-only code from
the (soon to be even more complex) read/write code.
It also allows you to create read-only mounts which don't bother with
filesystem accounting, which is nice.
- Make openForWrite/openForAppend always return a SeekableFileHandle.
Appendable files still cannot be seeked within, but that check is now
done on the FS side.
- Refactor the various mount tests to live in test contract interfaces,
allowing us to reuse them between mounts.
- Clean up our error handling a little better. (Most) file-specific code
has been moved to FileMount, and ArchiveMount-derived classes now
throw correct path-localised exceptions.
Lots of minor changes, but nothing too nasty - just tedious.
Known bugs/issues:
- REI and JEI haven't been updated at the time of writing, so our usage
of their APIs may be incompatible.
- Crash when opening the config UI in Fabric, as forgeconfigapi-port
hasn't been updated yet.
Will hold off on doing a release until those mods have updated.
- Flip http.websocket and http.websocketAsync docs (fixes#1244)
- Fix http.request queuing a http_failure event with no URL when
passing a malformed URL
- Fix http.websocketAsync not queuing websocket_failure events on
immediate failure.
In classic squid tradition: 20% code, and 80% test logic.
Closes#962. Alas, whoever reported this has deleted their account, so
they can't even be happy about it :(.
