Ability to jump to the first and the last hunk

[dtcyganok]

Is your feature request related to a problem? Please describe.
It's pretty common to have the comprehensive navigation through a set of items. Usually it includes:

• prev item
• next item
• first item
• last item

Many plugins' authors follow that pattern. For instance unimpaired.vim and ale.

Describe the solution you'd like
My suggestion is to add the api functions first_hunk and last_hunk.

Describe alternatives you've considered
To write my own functions first_hunk and last_hunk.

[lewis6991]

I would do gg]c for the first hunk and G[c for the last

[dtcyganok]

Thanks, i already tried that. And it doesn't work correctly, if the first hunk starts from the first line. Similarly, the situation with the last hunk is the same

[dtcyganok]

I wrote those functions, but i'm not sure that's correct:

local gitsigns = require('gitsigns')

local function opts()
  return {
    buffer = bufnr,
    unique = true,
  }
end

vim.keymap.set('n', '[G', function()
  local hunks = gitsigns.get_hunks()
  local hunk = hunks[1]
  local line = math.min(hunk.added.start, hunk.removed.start)
  vim.api.nvim_win_set_cursor(0, { line, 0 })
end, opts())

vim.keymap.set('n', ']G', function()
  local hunks = gitsigns.get_hunks()
  local hunk = hunks[#hunks]
  local line = math.max(hunk.added.start, hunk.removed.start)
  vim.api.nvim_win_set_cursor(0, { line, 0 })
end, opts())

[lewis6991]

It's close. The buffer indices are in hunk.added. hunk.removed is what you're comparing too.

Try hunk.added.start + hunk.added.count for the end of the hunk.

[ribru17]

What about G]c for first hunk and gg[c for last hunk? I tested this and it works when hunks are at the first and last lines thanks to the wrap-around functionality. For the last hunk if you want your cursor to be at the beginning of the changes then it would be gg[c[c]c.

[dtcyganok]

@ribru17 gg[c doesn't work properly if the first line has been changed, as i mentioned before. There is the bug with wrap option. I've already created the related issue

[lewis6991]

Done in #976