Not showing all undolevels

[debugloop]

Not showing all undolevels

[debugloop]

Not sure how that would happen, this plugin uses the complete :undolist. Can you compare the output of that with what your seeing on both plugins? Are the state numbers in telescope-undo continuous and without gaps too?

Also, I'm going to need a reproducible case where the two diverge. Can you provide a series of inputs to conjure a file? I.e. iline one<esc>oline two<esc>yypA1<esc>...

However, I can't imagine how this happens, as both plugins use the undolist built in and just render that.

[soer9459]

Took some screenshots.

Heres the last state in undotree and telescopeundo compared:

It seems that all states appear, but they dont appear in correct order on telescopeundo. I'm assuming this is due to the ascending sort order in my ts config that screws with the branch alignment

[soer9459]

I can reproduce by creating a new file, and inputting "hello" in the first line.
I then yank the line and paste over and over (~350 times)

Undotree goes to 0, and telescope-undo only goes to 121

[debugloop]

I'm on vacation and can test it out myself right now. Some further things you can try:

• At the bottom, your screenshots show that the buffer is not modifiable -> that will lead to telescope-undo exiting without doing anything further. Why is your buffer not modifiable? Maybe this is some strange behaviour from having the undo tree focused?
• If you have branches, it is correct that you don't have some states in linear order. They should be all accounted for inside a subtree.
• Did you compare to :undolist as well?
• Can you find an undostate at position 50 when you do your 350 undostates example? You'd have to sneak in a world in your hello example at one of the early positions that you can search for.
• Do you have some sort of result limiter as part of your telescope config? It's possible that something like this exists and has been configured. It's not a nvim/telescope default though.

[soer9459]

• Can you find an undostate at position 50 when you do your 350 undostates example? You'd have to sneak in a world in your hello example at one of the early positions that you can search for.
• Do you have some sort of result limiter as part of your telescope config? It's possible that something like this exists and has been configured. It's not a nvim/telescope default though.

Arh, yes i can. It is able to find previous states, that otherwise don't show in the list.
It might be relateed to this max list length item:
nvim-telescope/telescope.nvim#2779

Corresponds to about 250 lines as well

Heres the bottom of the list:

And here's after searching for world. States that were previously hidden are now visible:

I don't have anything to limit the list configured in telescope, but that may be the issue then

Heres my config

return {
    'nvim-telescope/telescope.nvim',
    dependencies = {
        'nvim-lua/plenary.nvim',
        'nvim-telescope/telescope-file-browser.nvim',
        'debugloop/telescope-undo.nvim',
        {
            'nvim-telescope/telescope-fzf-native.nvim',
            build = 'make',
        },
    },
    config = function()
        local ts = require('telescope')
        local ts_undo = require('telescope-undo.actions')
        ts.setup {
            defaults = {
                borderchars = { '─', '│', '─', '│', '┌', '┐', '┘', '└' },
                layout_strategy = 'horizontal',
                sorting_strategy = 'ascending',
                layout_config = {
                    prompt_position = 'top',
                    width = 0.8,
                    height = 0.8,
                    preview_width = 0.5,
                },
            },
            pickers = {
                find_files = {
                    find_command = {
                        'fd',
                        '--type',
                        'f',
                        '-H',
                        '--strip-cwd-prefix',
                    }
                }
            },
            extensions = {
                file_browser = {
                    hijack_netrw = true,
                    hide_parent_dir = true,
                    hidden = {
                        file_browser = false,
                        folder_browser = false,
                    },
                    display_stat = { },
                    mappings = {
                        ['n'] = {
                            ['-'] = ts.extensions.file_browser.actions.goto_parent_dir,
                            ['<BS>'] = ts.extensions.file_browser.actions.goto_parent_dir,
                            -- KEYBIND TO OPEN SELECTED FILES
                        },
                    }
                },
                undo = {
                    mappings = {
                        i = {
                            ['<cr>'] = ts_undo.restore,
                            ['<C-cr>'] = ts_undo.restore,
                            ['<C-y>'] = ts_undo.yank_deletions,
                            ['<C-r>'] = ts_undo.yank_additions,
                        },
                        n = {
                            ['y'] = ts_undo.yank_additions,
                            ['r'] = ts_undo.yank_deletions,
                            ['<cr>'] = ts_undo.restore,
                        },
                    },
                }
            }
        }
        ts.load_extension('file_browser')
        ts.load_extension('fzf')
        ts.load_extension('undo')
    end,
}

[debugloop]

Seems you found the culprit in that other issue, yes 👍🏻 I read through your config, looking good, nothing out of the ordinary. I assume my nvim would behave the same way if I could check it right now.

Thanks for reporting and thus documenting this here, but I'm going to close this for now until there's something actionable :)

I know once noticed stuff like this can be nagging, but in all honesty: There should not be a scenario where you need more that those 250ish undo states since your last git commit 😅 Especially since searching still works, too ;)

[soer9459]

Seems you found the culprit in that other issue, yes 👍🏻 I read through your config, looking good, nothing out of the ordinary. I assume my nvim would behave the same way if I could check it right now.

Thanks for reporting and thus documenting this here, but I'm going to close this for now until there's something actionable :)

I know once noticed stuff like this can be nagging, but in all honesty: There should not be a scenario where you need more that those 250ish undo states since your last git commit 😅 Especially since searching still works, too ;)

Yes, that seems to be it. I don't think it'll ever be a "real" issue, but sometimes you notice something that nags you anyway,even if irrational :)

Thank you for the quick replies