fix(bug): correct multi-byte character handling in get_visual_selection (#126)

Author: YousefHadder

lua/markdown-plus/utils.lua

@@ -105,6 +105,34 @@ function M.debug_print(...)
   end
 end
 
+---Get the byte index of the last byte of a multi-byte character
+---When vim.fn.getpos() returns a column position for a multi-byte character,
+---it returns the byte index of the FIRST byte of that character.
+---This function adjusts it to return the byte index of the LAST byte.
+---@param line string The line content
+---@param byte_col number The 1-indexed byte column from getpos()
+---@return number The 1-indexed byte column of the last byte of the character
+function M.get_char_end_byte(line, byte_col)
+  if byte_col > #line then
+    return #line
+  end
+
+  -- Convert 1-indexed byte position to 0-indexed for vim.str_utfindex
+  local char_idx = vim.str_utfindex(line, byte_col - 1)
+
+  -- Get the byte index of the next character (0-indexed)
+  local success, next_byte = pcall(vim.str_byteindex, line, char_idx + 1)
+
+  if success and next_byte then
+    -- next_byte is 0-indexed and points to the start of next char
+    -- We want the last byte of current char, which is next_byte (in 1-indexed terms)
+    return next_byte
+  else
+    -- We're at the last character or beyond, return line length
+    return #line
+  end
+end
+
 ---Get visual selection range
 ---@param include_col? boolean Whether to include column info (default: true)
 ---@return {start_row: number, end_row: number, start_col?: number, end_col?: number}
@@ -134,6 +162,13 @@ function M.get_visual_selection(include_col)
       start_col = 1
       local end_line = vim.api.nvim_buf_get_lines(0, end_row - 1, end_row, false)[1] or ""
       end_col = #end_line
+    else
+      -- For character-wise (v) and block-wise (<C-v>) visual modes,
+      -- adjust end_col to handle multi-byte characters
+      -- getpos() returns the byte position of the first byte of a multi-byte character
+      -- We need the byte position of the last byte for proper text extraction
+      local end_line = vim.api.nvim_buf_get_lines(0, end_row - 1, end_row, false)[1] or ""
+      end_col = M.get_char_end_byte(end_line, end_col)
     end
 
     if include_col then
@@ -154,17 +189,25 @@ function M.get_visual_selection(include_col)
     local start_pos = vim.fn.getpos("'<")
     local end_pos = vim.fn.getpos("'>")
 
+    local start_col = start_pos[3]
+    local end_col = end_pos[3]
+
+    -- Adjust end_col for multi-byte characters in previous visual selection
+    local end_row = end_pos[2]
+    local end_line = vim.api.nvim_buf_get_lines(0, end_row - 1, end_row, false)[1] or ""
+    end_col = M.get_char_end_byte(end_line, end_col)
+
     if include_col then
       return {
         start_row = start_pos[2],
-        start_col = start_pos[3],
-        end_row = end_pos[2],
-        end_col = end_pos[3],
+        start_col = start_col,
+        end_row = end_row,
+        end_col = end_col,
       }
     else
       return {
         start_row = start_pos[2],
-        end_row = end_pos[2],
+        end_row = end_row,
       }
     end
   end

spec/markdown-plus/utils_spec.lua

@@ -102,4 +102,110 @@ describe("markdown-plus utils", function()
       vim.api.nvim_buf_delete(buf, { force = true })
     end)
   end)
+
+  describe("get_char_end_byte", function()
+    it("returns correct byte index for ASCII characters", function()
+      local line = "hello world"
+      -- ASCII characters are 1 byte each, so end byte = input byte
+      assert.are.equal(1, utils.get_char_end_byte(line, 1)) -- 'h'
+      assert.are.equal(6, utils.get_char_end_byte(line, 6)) -- ' '
+      assert.are.equal(11, utils.get_char_end_byte(line, 11)) -- 'd'
+    end)
+
+    it("returns correct byte index for multi-byte characters", function()
+      local line = "这是一段文本" -- Each Chinese character is 3 bytes in UTF-8
+      -- 这 = bytes 1-3
+      assert.are.equal(3, utils.get_char_end_byte(line, 1))
+      -- 是 = bytes 4-6
+      assert.are.equal(6, utils.get_char_end_byte(line, 4))
+      -- 一 = bytes 7-9
+      assert.are.equal(9, utils.get_char_end_byte(line, 7))
+      -- 段 = bytes 10-12
+      assert.are.equal(12, utils.get_char_end_byte(line, 10))
+      -- 文 = bytes 13-15
+      assert.are.equal(15, utils.get_char_end_byte(line, 13))
+      -- 本 = bytes 16-18 (last character)
+      assert.are.equal(18, utils.get_char_end_byte(line, 16))
+    end)
+
+    it("returns correct byte index for mixed ASCII and multi-byte", function()
+      local line = "hello 世界" -- "hello " = 6 bytes, 世=3 bytes, 界=3 bytes
+      assert.are.equal(1, utils.get_char_end_byte(line, 1)) -- 'h'
+      assert.are.equal(6, utils.get_char_end_byte(line, 6)) -- ' '
+      assert.are.equal(9, utils.get_char_end_byte(line, 7)) -- '世' (bytes 7-9)
+      assert.are.equal(12, utils.get_char_end_byte(line, 10)) -- '界' (bytes 10-12)
+    end)
+
+    it("handles byte position beyond line length", function()
+      local line = "test"
+      assert.are.equal(4, utils.get_char_end_byte(line, 100))
+    end)
+  end)
+
+  describe("get_visual_selection with multi-byte characters", function()
+    it("handles visual selection of multi-byte characters", function()
+      local buf = vim.api.nvim_create_buf(false, true)
+      vim.api.nvim_buf_set_lines(buf, 0, -1, false, { "这是一段文本" })
+      vim.api.nvim_set_current_buf(buf)
+
+      -- Select all text
+      vim.api.nvim_win_set_cursor(0, { 1, 0 })
+      vim.cmd("normal! v$")
+
+      local selection = utils.get_visual_selection()
+
+      -- Should select the entire line
+      assert.are.equal(1, selection.start_row)
+      assert.are.equal(1, selection.end_row)
+      assert.are.equal(1, selection.start_col)
+      assert.are.equal(18, selection.end_col) -- Full line is 18 bytes
+
+      -- Verify the text extraction works correctly
+      local text =
+        utils.get_text_in_range(selection.start_row, selection.start_col, selection.end_row, selection.end_col)
+      assert.are.equal("这是一段文本", text)
+
+      vim.api.nvim_buf_delete(buf, { force = true })
+    end)
+
+    it("handles partial selection of multi-byte characters", function()
+      local buf = vim.api.nvim_create_buf(false, true)
+      vim.api.nvim_buf_set_lines(buf, 0, -1, false, { "这是一段文本" })
+      vim.api.nvim_set_current_buf(buf)
+
+      -- Select first two characters (这是)
+      vim.api.nvim_win_set_cursor(0, { 1, 0 })
+      vim.cmd("normal! v")
+      vim.cmd("normal! l") -- Move to second char
+
+      local selection = utils.get_visual_selection()
+
+      -- Should get correct byte positions
+      assert.are.equal(1, selection.start_col)
+      assert.are.equal(6, selection.end_col) -- End of second char (是)
+
+      local text =
+        utils.get_text_in_range(selection.start_row, selection.start_col, selection.end_row, selection.end_col)
+      assert.are.equal("这是", text)
+
+      vim.api.nvim_buf_delete(buf, { force = true })
+    end)
+
+    it("handles mixed ASCII and multi-byte selection", function()
+      local buf = vim.api.nvim_create_buf(false, true)
+      vim.api.nvim_buf_set_lines(buf, 0, -1, false, { "hello 世界" })
+      vim.api.nvim_set_current_buf(buf)
+
+      -- Select entire line
+      vim.api.nvim_win_set_cursor(0, { 1, 0 })
+      vim.cmd("normal! v$")
+
+      local selection = utils.get_visual_selection()
+      local text =
+        utils.get_text_in_range(selection.start_row, selection.start_col, selection.end_row, selection.end_col)
+      assert.are.equal("hello 世界", text)
+
+      vim.api.nvim_buf_delete(buf, { force = true })
+    end)
+  end)
 end)