Better global

[sumneko]

I'm going to do some work for global variable distinguishing environments. At present, my idea is as follows:

---@global global -- the table will be merged into current globals. the second `global` can be omitted
local t = {
    X = 1,
    Y = 2,
    Z = 3,
}

print(X) -- `X` is `t.X`

---@global module1, module2 -- the table will be merged into environment `module1` and `module2`
local t = {
    X = 1,
    Y = 2,
    Z = 3,
}

print(X) -- `x` is undefined

---@environment module1, global -- set the current environment can visit `module1` and `global`
print(X) -- `X` is `t.X`

---@environment module1
print(X) -- `X` is `t.X` and `print` is undefined

---

EDIT:

---@partial class _G -- the table will be merged into class `_G`. the default environment is class `_G`
local t = {
    X = 1,
    Y = 2,
    Z = 3,
}

print(X) -- `X` is `t.X`

---@class module1
local t = {
    X = 1,
    Y = 2,
    Z = 3,
}

print(X) -- `x` is undefined

---@environment module1, _G -- set the current environment can visit class `module1` and `_G`
print(X) -- `X` is `t.X`

---@environment module1
print(X) -- `X` is `t.X` and `print` is undefined

[ZegWe]

Hello, thanks for your work on this project, is there any progress on this issue recently?

[emmericp]

What's the best way to handle environments until we get first class support for them?

I'm currently thinking about writing a plugin that just emits setlocal <field> <nil> with an @type doc class annotation for every field in my environment class (which is rather small anyways) in the files that make use of that environment

[sarahsturgeon]

I love this 👍