bug(statuscolumn): Adjacent folds not displaying correctly with foldopen

[gczcn]

Did you check docs and existing issues?

• I have read all the snacks.nvim docs
• I have updated the plugin to the latest version before submitting this issue
• I have searched the existing issues of snacks.nvim
• I have searched the existing issues of plugins related to this issue

Neovim version (nvim -v)

0.11.0-dev-1930+gb2fa51bf15

Operating system/version

macOS 15.3.1

Describe the bug

When there are two adjacent folds, the part that determines whether a line is the beginning of a fold may not work properly. This is because when two folds are adjacent, the foldlevel of the previous line of the latter fold may be the same as that of the previous line, causing this line to be considered a line in a fold.

a simple example:

line 361 does not display the foldopen symbol.

Steps To Reproduce

1. Ensure the statuscolumn section in the configuration has the open option set to true
2. Create a lua file and enter the following content, making sure there are adjacent folds in the file

if true then
  print('true')
end
if false then
  print('false')
end

3. Problem occurs

Expected Behavior

Correctly display foldopen symbols for adjacent folds.

Repro

vim.o.foldlevel = 99
vim.o.foldmethod = 'expr'
vim.o.foldexpr = "v:lua.vim.treesitter.foldexpr()"
vim.env.LAZY_STDPATH = ".repro"
load(vim.fn.system("curl -s https://raw.githubusercontent.com/folke/lazy.nvim/main/bootstrap.lua"))()

require("lazy.minit").repro({
  spec = {
    {
      "folke/snacks.nvim",
      opts = {
        statuscolumn = {
          folds = {
            open = true,
          },
        },
      },
    },

    -- install the treesitter parser for lua
    {
      "nvim-treesitter/nvim-treesitter",
      opts = {
        ensure_installed = { 'lua' },
      },
    },
  },
})

-- Example
if true then
  -- ...
end
if false then
  -- ...
end

[dpetka2001]

That's because vim.fn.foldlevel() which is used by the statuscolumn returns the same number of level folding for those mappings.

vim.keymap.set("n", "t", function()    <-- foldlevel is some number here
  fadskfjas
end)
vim.keymap.set("n", "s", function()    <-- foldlevel here is the same as the previous
  fjasiofaja
end)

The opened markers show up only at lines where the foldlevel is not the same between the current and previous line.

[dpetka2001]

Actually, if we change the line here

snacks.nvim/lua/snacks/statuscolumn.lua

Line 147 in bc0630e

elseif config.folds.open and vim.fn.foldlevel(lnum) > vim.fn.foldlevel(lnum - 1) then

with the following elseif config.folds.open and vim.treesitter.foldexpr(vim.v.lnum):sub(1, 1) == ">" then I believe it behaves more to how you describe.

[gczcn]

Actually, if we change the line here

snacks.nvim/lua/snacks/statuscolumn.lua

Line 147 in bc0630e

elseif config.folds.open and vim.fn.foldlevel(lnum) > vim.fn.foldlevel(lnum - 1) then
with the following elseif config.folds.open and vim.treesitter.foldexpr(vim.v.lnum):sub(1, 1) == ">" then I believe it behaves more to how you describe.

For me, this solved the problem. However, as described in this issue, using vim.treesitter.foldexpr() will only allow folds provided by treesitter.

[gczcn]

Sorry, I just closed this question by accident.

[dpetka2001]

I see, you're right. I wasn't aware of that issue. Then probably my PR which essentially reverts the change that was made for that issue you mentioned will be denied. I'm not aware of a better solution in this case.

Hopefully, maintainer will have a better idea with regards to this. Maybe it's possible maybe not. But vim.fn.foldlevel() only returns the level which for adjacent foldings will be the same.

I'll close my PR and wait for maintainer's input.

[gczcn]

The statuscol.nvim plugin can handle this situation correctly, but I can't understand its code.

[dpetka2001]

I pushed a change which somewhat takes into consideration vim.lsp.foldexpr() as well, but there's a small caveat (which I mention in my PR).

I'm sure maintainer will have a more proper answer with regards to that.