is an unofficial continuation of ProbeJS for KubeJS versions abandoned by official ProbeJS.
- You need to install ProbeJS Legacy, obviously.
- Enter a singleplayer world. Use
/probejs dump
to export datas gathered by ProbeJS Legacy. For first-time users, this will be done automatically when you join the world. - Open your game instance folder using VSCode.
- Enjoy!
Note0: type @item
or @block_tag
to see snippets generated.
Note1: game instance folder is where you can see the kubejs
folder directly.
Note2: Other IDEs with support for TypeScript typing (.d.ts) should also work.
(Compared to ProbeJS)
-
Simulated
Internal
- ProbeJS Legacy will now generate a
Internal
namespace, allowing users to migrate from ProbeJS 6 or ProbeJS Legacy 3 easily. - Example:
Internal.ItemStack
andInternal.ItemStack_
- ProbeJS Legacy will now generate a
-
Better Param Naming
- Param names that conflict with TypeScript will now be renamed in a better way.
- Param names in Method/Constructor, if in
arg123
format, will be renamed to a new name generated from its type.
Example:
"mayUseItemAt"(arg0: $BlockPos$$Type, arg1: $Direction$$Type, arg2: $ItemStack$$Type): boolean
will be renamed to
"mayUseItemAt"(blockPos0: $BlockPos$$Type, direction1: $Direction$$Type, itemStack2: $ItemStack$$Type): boolean
-
Better
$Class<T>
- ProbeJS will transform
$Class<T>
intotypeof T
, but ignoring many conditions where this should not be done. ProbeJS Legacy makes it more correct. - Some basic classes, for example
java.lang.String
andjava.lang.Character
. can now be dumped. - The
T
in$Class<T>
is now actually usable for type hint.
- ProbeJS will transform
-
better
instanceof
- ProbeJS Legacy can now properly support
obj instanceof $SomeClass
, where$SomeClass
is a class loaded viajava(...)
orJava.loadClass(...)
.
- ProbeJS Legacy can now properly support
Example:
const $BoatItem = java("some.path.to.BoatItem")
let item = Item.of("boat").item; // here the type of `item` is `$Item`
if (item instanceof $BoatItem) {
// here the type of `item` can be inferred to `$BoatItem`
}
-
Field Beaning
- Some fields will be transformed into getter+setter, in order to provide better support for TypeWrapper
- Consider a
obj
with a field like this:block: $Block
, andobj.block = "minecraft:stone"
is a valid syntax in KubeJS. This had no typing support before, but now ProbeJS Legacy can support it properly.
-
Typing files sharing
- By extracting common parts in dumped classes, the total typing files size can be much smaller than before.
- Another benefits is that TS server can now reuse common typings between 3 script types, improving IDE performance.
-
Better comment formatting
- blank lines at the front/back will now be ignored
- comments will be formatted as one line if there's only one valid line
-
Side only methods (marked by
@OnlyIn
) will have comments telling you that it's server-only or client-only. -
xxx.d.ts
file with too many modules in it will now be split into smaller files automatically, providing better IDE performance -
Method/Field in declaration files are now naturally sorted.
-
gloabl
resolving: values you put intoglobal
can be resolved now. Configure resolving depth via'global' Resolving Depth
config. -
And many internal changes. This Readme will certainly be too long if I write them all down.
For 1.16.5, ProbeJS Legacy will record all posted event. If you're unable to find some common events, try triggering it before dumping data. Example: block.place
event is triggered by placing a block.
For 1.20.1, all registered KJS events will be dumped automatically, no need for triggering.
ProbeJS Legacy will record all Forge events, regardless of whether they have been posted or not. This behaviour can be altered via Class Scanner
config.
ProbeJS Legacy will use multiple approaches for collecting Java classes.
These classes will be recorded:
- Classes loaded manually, aka those loaded via
java(...)
orJava.loadClass(...)
- Classes referenced by JS bindings, for example
AABB
andText
. - Classes references by known classes, for example
Item.class
referenced byItemStack.class
. - Registry objects, for example
BoatItem
for vanilla boats. By default this will skip non-vanilla objects, and can be altered viaRegistry Object Filter
config. - Classes from certain mods. See
Mods with forced Full Scanning
config. - Classes read from cache file. The cache file is saved to
<your instance folder>/.probe/classes.txt
. - Classes found by class scanner. See
Class Scanner
config.
Initially, all I want to do is just find some sort of doc or helper for KubeJS script writing, so that I won't need to feel such pain in the arse when developing Enlightend6Expert(https://www.curseforge.com/minecraft/modpacks/enlightened-6-expert) .
But after some research, I realized that 1.16.5, despite being one of the hottest modded versions, didn't even have a proper KubeJS helper mod.
So there it is, a ProbeJS fork for 1.16.5.
This is my first Java project and first Minecraft Mod project, and to be honest, ProbeJS is too tough for a coding noob like me. Thankfully I made it.
This project is licensed under LGPLv3, the same as the original ProbeJS.
Auto-completion snippets for Items, Blocks, Fluids, Entities and Tags:
Auto-completion, type-hinting for most of the functions and classes: