illuaminate does not handle Java files, for obvious reasons. In order to
get around that, we have a series of stub files within /doc/stub which
mirrored the Java ones. While this works, it has a few problems:
- The link to source code does not work - it just links to the stub
file.
- There's no guarantee that documentation remains consistent with the
Java code. This change found several methods which were incorrectly
documented beforehand.
We now replace this with a custom Java doclet[1], which extracts doc
comments from @LuaFunction annotated methods and generates stub-files
from them. These also contain a @source annotation, which allows us to
correctly link them back to the original Java code.
There's some issues with this which have yet to be fixed. However, I
don't think any of them are major blockers right now:
- The custom doclet relies on Java 9 - I think it's /technically/
possible to do this on Java 8, but the API is significantly uglier.
This means that we need to run javadoc on a separate JVM.
This is possible, and it works locally and on CI, but is definitely
not a nice approach.
- illuaminate now requires the doc stubs to be generated in order for
the linter to pass, which does make running the linter locally much
harder (especially given the above bullet point).
We could notionally include the generated stubs (or at least a cut
down version of them) in the repo, but I'm not 100% sure about that.
[1]: https://docs.oracle.com/javase/9/docs/api/jdk/javadoc/doclet/package-summary.html
- Refer to this as "data" rather than "metadata". I'm still not sure
where the meta came from - blame OpenPeripheral I guess.
- Likewise, use getItemDetail within inventory methods, rather than
getItemMeta.
- Refactor common data-getting code into one class. This means that
turtle.getItemDetail, turtle.inspect and commands.getBlockInfo all
use the same code.
- turtle.getItemDetail now accepts a second "detailed" parameter which
will include the full metadata (#471, #452).
- Tags are now only included in the detailed list. This is a breaking
change, however should only affect one version (1.89.x) and I'm not
convinced that the previous behaviour was safe.
This allows for configuring the size of computers and pocket computers,
as well as the max size of monitors.
There's several limitations with the current implementation, but it's
still "good enough" for an initial release:
- Turtles cannot be resized.
- GUIs do not scale themselves, so "large" sizes will not render within
the default resolution.
This exposes a basic peripheral for any tile entity which does not have methods
already registered. We currently provide the following methods:
- Inventories: size, list, getItemMeta, pushItems, pullItems.
- Energy storage: getEnergy, getEnergyCapacity
- Fluid tanks: tanks(), pushFluid, pullFluid.
These methods are currently experimental - it must be enabled through
`experimental.generic_peripherals`. While this is an initial step towards
implementing #452, but is by no means complete.
Well, mostly. We currently don't do recipe serializers as I'm a little
too lazy. For items, blocks and TE types this does make registration
nicer - we've some helper functions which help reduce duplication.
Some types (containers, TEs, etc..) are a little less nice, as we now
must define the registry object (i.e. the WhateverType<?>) in a separate
class to the class it constructs. However, it's probably a worthwhile
price to pay.
No clue how we're going to do this for the dynamic peripheral system
if/when that ships, but this is a good first stage.
Like the Java APIs, this relies on stub files, so we can't link to the
implementation which is a bit of a shame. However, it's a good first
step.
I'm really not very good at this modding lark am I? I've done a basic
search for other missing methods, and can't see anything, but goodness
knows.
Fixes#480
Translations for French
Translations for French
Co-authored-by: hds <hds536jhmk@gmail.com>
Co-authored-by: Anavrins <xanavrins@gmail.com>
Co-authored-by: AxelFontarive <afontarive@gmail.com>
When calling .flip(), we limit the size of the buffer. However, this
limit is not reset when writing the next time, which means we get
out-of-bounds errors, even if the buffer is /technically/ big enough.
Clearing the buffer before drawing (rather than just resetting the
position) is enough to fix this.
Fixes#476 (and closes#477, which is a duplicate)
We never added back replacing of ${version} strings, which means that CC
was reporting incorrect version numbers in _HOST, the user agent and
network versions. This meant we would allow connections even on
mismatched versions (#464).
We shift all version handling into ComputerCraftAPI(Impl) - this now
relies on Forge code, so we don't want to run it in emulators.
- Strip any gui._.config options. These haven't been used since 1.12
and while they may return, it doesn't seem worth it right now.
- Fix a couple of typos in the English translations.
- Import from https://i18n.tweaked.cc. There's definitely some problems
with the import - empty translations are still included, so we write
a script to strip them.
This is simply exposed as a table from tag -> true. While this is less
natural than an array, it allows for easy esting of whether a tag is
present.
Closes#461
- Use texture over texture2D - the latter was deprecated in GLSL 1.30.
- Cache the tbo buffer - this saves an allocation when monitors update.
Closes#455. While the rest of the PR has some nice changes, it
performs signlificantly worse on my system.
This moves monitor networking into its own packet, rather than serialising
using NBT. This allows us to be more flexible with how monitors are
serialised.
We now compress terminal data using gzip. This reduces the packet size
of a max-sized-monitor from ~25kb to as little as 100b.
On my test set of images (what I would consider to be the extreme end of
the "reasonable" case), we have packets from 1.4kb bytes up to 12kb,
with a mean of 6kb. Even in the worst case, this is a 2x reduction in
packet size.
While this is a fantastic win for the common case, it is not abuse-proof.
One can create a terminal with high entropy (and so uncompressible). This
will still be close to the original packet size.
In order to prevent any other abuse, we also limit the amount of monitor
data a client can possibly receive to 1MB (configurable).
timetout, max_upload, max_download and max_websocket_message may now be
configured on a domain-by-domain basis. This uses the same system that
we use for the block/allow-list from before:
Example:
[[http.rules]]
host = "*"
action = "allow"
max_upload = 4194304
max_download = 16777216
timeout = 30000
This registers IPeripheral as a capability. As a result, all (Minecraft
facing) functionality operates using LazyOptional<_>s instead.
Peripheral providers should now return a LazyOptional<IPeripheral> too.
Hopefully this will allow custom peripherals to mark themselves as
invalid (say, because a dependency has changed).
While peripheral providers are somewhat redundant, they still have their
usages. If a peripheral is applied to a large number of blocks (for
instance, all inventories) then using capabilities does incur some
memory overhead.
We also make the following changes based on the above:
- Remove the default implementation for IWiredElement, migrating the
definition to a common "Capabilities" class.
- Remove IPeripheralTile - we'll exclusively use capabilities now.
Absurdly this is the most complex change, as all TEs needed to be
migrated too.
I'm not 100% sure of the correctness of this changes so far - I've
tested it pretty well, but blocks with more complex peripheral logic
(wired/wireless modems and turtles) are still a little messy.
- Remove the "command block" peripheral provider, attaching a
capability instead.
When creating a peripheral or custom Lua object, one must implement two
methods:
- getMethodNames(): String[] - Returns the name of the methods
- callMethod(int, ...): Object[] - Invokes the method using an index in
the above array.
This has a couple of problems:
- It's somewhat unwieldy to use - you need to keep track of array
indices, which leads to ugly code.
- Functions which yield (for instance, those which run on the main
thread) are blocking. This means we need to spawn new threads for
each CC-side yield.
We replace this system with a few changes:
- @LuaFunction annotation: One may annotate a public instance method
with this annotation. This then exposes a peripheral/lua object
method.
Furthermore, this method can accept and return a variety of types,
which often makes functions cleaner (e.g. can return an int rather
than an Object[], and specify and int argument rather than
Object[]).
- MethodResult: Instead of returning an Object[] and having blocking
yields, functions return a MethodResult. This either contains an
immediate return, or an instruction to yield with some continuation
to resume with.
MethodResult is then interpreted by the Lua runtime (i.e. Cobalt),
rather than our weird bodgey hacks before. This means we no longer
spawn new threads when yielding within CC.
- Methods accept IArguments instead of a raw Object array. This has a
few benefits:
- Consistent argument handling - people no longer need to use
ArgumentHelper (as it doesn't exist!), or even be aware of its
existence - you're rather forced into using it.
- More efficient code in some cases. We provide a Cobalt-specific
implementation of IArguments, which avoids the boxing/unboxing when
handling numbers and binary strings.
- cc.pretty.pretty now accepts two additional options:
- function_args: Show function arguments
- function_source: Show where functions are defined.
- Expose the two options as lua.* settings (defaulting function_args to
true, and function_source to false).
These are then used in the Lua REPL.
Closes#361
- Use jacoco for Java-side coverage. Our Java coverage is /terrible
(~10%), as we only really test the core libraries. Still a good thing
to track for regressions though.
- mcfly now tracks Lua side coverage. This works in several stages:
- Replace loadfile to include the whole path
- Add a debug hook which just tracks filename->(lines->count). This
is then submitted to the Java test runner.
- On test completion, we emit a luacov.report.out file.
As the debug hook is inserted by mcfly, this does not include any
computer startup (such as loading apis, or the root of bios.lua),
despite they're executed.
This would be possible to do (for instance, inject a custom header
into bios.lua). However, we're not actually testing any of the
behaviour of startup (aside from "does it not crash"), so I'm not
sure whether to include it or not. Something I'll most likely
re-evaluate.
`local varname = value` results in `varname` being inaccessible in
the next REPL input. This is often unintended and can lead to confusing
behaviour. We produce a warning when this occurs.
This uses the system described in #409, to render monitors in a more
efficient manner.
Each monitor is backed by a texture buffer object (TBO) which contains
a relatively compact encoding of the terminal state. This is then
rendered using a shader, which consumes the TBO and uses it to index
into main font texture.
As we're transmitting significantly less data to the GPU (only 3 bytes
per character), this effectively reduces any update lag to 0. FPS appears
to be up by a small fraction (10-15fps on my machine, to ~110), possibly
as we're now only drawing a single quad (though doing much more work in
the shader).
On my laptop, with its Intel integrated graphics card, I'm able to draw
120 full-sized monitors (with an effective resolution of 3972 x 2330) at
a consistent 60fps. Updates still cause a slight spike, but we always
remain above 30fps - a significant improvement over VBOs, where updates
would go off the chart.
Many thanks to @Lignum and @Lemmmy for devising this scheme, and helping
test and review it! ♥
Unknown blit colours, such as " " will be translated to black for the
background or white for the foreground. This restores the behaviour from
before #412.
- Write to a PacketBuffer instead of generating an NBT tag. This is
then converted to an NBT byte array when we send across the network.
- Pack background/foreground colours into a single byte.
This derives from some work I did back in 2017, and some of the changes
made/planned in #409. However, this patch does not change how terminals
are represented, it simply makes the transfer more compact.
This makes the patch incredibly small (100 lines!), but also limited in
what improvements it can make compared with #409. We send 26626 bytes
for a full-sized monitor. While a 2x improvement over the previous 58558
bytes, there's a lot of room for improvement.
This functions the same as shell.run, but does not tokenise the
arguments. This allows us to pass command line arguments through to
another program without having to re-quote them.
Closes#417
- Remove the parenthesis around the text (so it's now
"Computer ID: 12"), rather than "(Computer ID: 12").
- Show the tooltip if the computer has an ID and no label (as well as
when in advanced mode).
- Lint references to unknown fields of modules, excluding the keys and
colours modules. This caught several silly errors in our stub files,
but nothing else.
- Lint on using unknown globals. This highlighted a couple of really
silly mistakes. Fixes#427.
- Add documentation for fs.attributes, fs.getCapacity and pocket, as
they were not defined before.
Co-authored-by: JackMacWindows <jackmacwindowslinux@gmail.com>
