fix(multibyte): handle utf-8 multibyte characters in text manipulation operations (#146)

Author: YousefHadder

lua/markdown-plus/footnotes/insertion.lua

@@ -123,7 +123,10 @@ function M.insert_footnote()
     -- Insert reference after the character the cursor is on
     local line = vim.api.nvim_buf_get_lines(bufnr, row - 1, row, false)[1]
     local reference = "[^" .. id .. "]"
-    local new_line = line:sub(1, col + 1) .. reference .. line:sub(col + 2)
+
+    -- Use UTF-8 safe split to handle multibyte characters correctly
+    local before, after = utils.split_after_cursor(line, col)
+    local new_line = before .. reference .. after
     vim.api.nvim_buf_set_lines(bufnr, row - 1, row, false, { new_line })
 
     -- If definition already exists, just insert the reference and notify

lua/markdown-plus/format/init.lua

@@ -280,7 +280,7 @@ end
 function M.get_word_boundaries()
   local cursor = utils.get_cursor()
   local row = cursor[1]
-  local col = cursor[2]
+  local col = cursor[2] -- 0-indexed byte offset
   local line = utils.get_current_line()
 
   -- Define what characters are considered word boundaries (stop points)
@@ -305,38 +305,62 @@ function M.get_word_boundaries()
     return false
   end
 
-  -- Find word start
-  local word_start = col
-  while word_start > 0 do
-    local char = line:sub(word_start, word_start)
+  -- Convert byte offset to character index for iteration
+  local char_idx = vim.fn.charidx(line, col)
+  if char_idx < 0 then
+    char_idx = 0
+  end
+
+  -- Get total character count
+  local total_chars = vim.fn.strcharlen(line)
+
+  -- Find word start (iterate backwards by character)
+  local word_start_char = char_idx
+  while word_start_char > 0 do
+    local char = vim.fn.strcharpart(line, word_start_char, 1)
     if is_word_boundary(char) then
-      word_start = word_start + 1
+      word_start_char = word_start_char + 1
       break
     end
-    word_start = word_start - 1
+    word_start_char = word_start_char - 1
   end
-  if word_start == 0 then
-    word_start = 1
+  if word_start_char < 0 then
+    word_start_char = 0
   end
 
-  -- Find word end
-  local word_end = col + 1
-  while word_end <= #line do
-    local char = line:sub(word_end, word_end)
+  -- Find word end (iterate forwards by character)
+  local word_end_char = char_idx + 1
+  while word_end_char < total_chars do
+    local char = vim.fn.strcharpart(line, word_end_char, 1)
     if is_word_boundary(char) then
-      word_end = word_end - 1
+      word_end_char = word_end_char - 1
       break
     end
-    word_end = word_end + 1
+    word_end_char = word_end_char + 1
+  end
+  if word_end_char >= total_chars then
+    word_end_char = total_chars - 1
   end
-  if word_end > #line then
-    word_end = #line
+  if word_end_char < 0 then
+    word_end_char = 0
+  end
+
+  -- Convert character indices back to byte positions (1-indexed for get_text_in_range)
+  local start_byte = vim.fn.byteidx(line, word_start_char)
+  if start_byte == -1 then
+    start_byte = 0
+  end
+  local end_byte = vim.fn.byteidx(line, word_end_char + 1)
+  if end_byte == -1 then
+    end_byte = #line
+  else
+    end_byte = end_byte - 1
   end
 
   return {
     row = row,
-    start_col = word_start,
-    end_col = word_end,
+    start_col = start_byte + 1, -- Convert to 1-indexed
+    end_col = end_byte + 1, -- Convert to 1-indexed
   }
 end
 

lua/markdown-plus/images/init.lua

@@ -189,11 +189,13 @@ function M.insert_image()
   local line = utils.get_current_line()
   local col = cursor[2]
 
-  local new_line = line:sub(1, col) .. image .. line:sub(col + 1)
+  -- Use UTF-8 safe split to handle multibyte characters correctly
+  local before, after = utils.split_after_cursor(line, col)
+  local new_line = before .. image .. after
   utils.set_line(cursor[1], new_line)
 
   -- Move cursor after the image
-  utils.set_cursor(cursor[1], col + #image)
+  utils.set_cursor(cursor[1], #before + #image)
 
   utils.notify("Image inserted")
 end

lua/markdown-plus/links/init.lua

@@ -175,11 +175,13 @@ function M.insert_link()
   local line = utils.get_current_line()
   local col = cursor[2]
 
-  local new_line = line:sub(1, col) .. link .. line:sub(col + 1)
+  -- Use UTF-8 safe split to handle multibyte characters correctly
+  local before, after = utils.split_after_cursor(line, col)
+  local new_line = before .. link .. after
   utils.set_line(cursor[1], new_line)
 
   -- Move cursor after the link
-  utils.set_cursor(cursor[1], col + #link)
+  utils.set_cursor(cursor[1], #before + #link)
 
   utils.notify("Link inserted")
 end

lua/markdown-plus/list/handlers.lua

@@ -65,8 +65,8 @@ function M.handle_enter()
 
     if not list_info then
       -- Not in a list at all, simulate default Enter behavior
-      local line_before = current_line:sub(1, col)
-      local line_after = current_line:sub(col + 1)
+      -- Use UTF-8 safe split to handle multibyte characters correctly
+      local line_before, line_after = utils.split_at_cursor(current_line, col)
 
       utils.set_line(row, line_before)
       utils.insert_line(row + 1, line_after)
@@ -102,8 +102,8 @@ function M.handle_enter()
 
   if should_split then
     -- Split content at cursor position
-    local content_before = current_line:sub(1, col)
-    local content_after = current_line:sub(col + 1)
+    -- Use UTF-8 safe split to handle multibyte characters correctly
+    local content_before, content_after = utils.split_at_cursor(current_line, col)
 
     -- Update current line with content before cursor
     utils.set_line(row, content_before)
@@ -142,8 +142,8 @@ function M.continue_list_content()
 
   if not list_info then
     -- Not in a list, simulate default Enter behavior
-    local line_before = current_line:sub(1, col)
-    local line_after = current_line:sub(col + 1)
+    -- Use UTF-8 safe split to handle multibyte characters correctly
+    local line_before, line_after = utils.split_at_cursor(current_line, col)
 
     utils.set_line(row, line_before)
     utils.insert_line(row + 1, line_after)
@@ -155,8 +155,8 @@ function M.continue_list_content()
   local marker_end = shared.get_content_start_col(list_info)
 
   -- Split line at cursor
-  local line_before = current_line:sub(1, col)
-  local line_after = current_line:sub(col + 1)
+  -- Use UTF-8 safe split to handle multibyte characters correctly
+  local line_before, line_after = utils.split_at_cursor(current_line, col)
 
   -- Update current line
   utils.set_line(row, line_before)
@@ -179,9 +179,11 @@ function M.handle_tab()
     local cursor = utils.get_cursor()
     local row, col = cursor[1], cursor[2]
     local indent = string.rep(" ", vim.bo.shiftwidth or 2)
-    local new_line = current_line:sub(1, col) .. indent .. current_line:sub(col + 1)
+    -- Use UTF-8 safe split to handle multibyte characters correctly
+    local before, after = utils.split_after_cursor(current_line, col)
+    local new_line = before .. indent .. after
     utils.set_line(row, new_line)
-    utils.set_cursor(row, col + #indent)
+    utils.set_cursor(row, #before + #indent)
     return
   end
 

lua/markdown-plus/utils.lua

@@ -105,6 +105,74 @@ function M.debug_print(...)
   end
 end
 
+---Split a line at a byte column position, ensuring proper UTF-8 character boundaries.
+---Splits BEFORE the character at the cursor position (character at cursor goes to 'after').
+---Use this for line splitting operations (e.g., Enter key behavior).
+---@param line string The line content
+---@param byte_col number 0-indexed byte column (from nvim_win_get_cursor)
+---@return string before Text before the cursor position
+---@return string after Text from cursor position onwards (including character at cursor)
+function M.split_at_cursor(line, byte_col)
+  if #line == 0 then
+    return "", ""
+  end
+
+  -- Handle cursor past end of line
+  if byte_col >= #line then
+    return line, ""
+  end
+
+  -- Convert 0-indexed byte offset to character index
+  local char_idx = vim.fn.charidx(line, byte_col)
+  if char_idx < 0 then
+    -- Should not happen if byte_col < #line, but be safe
+    return line, ""
+  end
+
+  -- Get byte position for start of current character
+  local curr_char_byte = vim.fn.byteidx(line, char_idx)
+  if curr_char_byte == -1 or curr_char_byte >= #line then
+    -- Past end of line
+    return line, ""
+  end
+
+  return line:sub(1, curr_char_byte), line:sub(curr_char_byte + 1)
+end
+
+---Split a line after the character at cursor position, ensuring proper UTF-8 character boundaries.
+---Splits AFTER the character at the cursor position (character at cursor goes to 'before').
+---Use this for insertion operations (e.g., inserting footnotes, links, images after current char).
+---@param line string The line content
+---@param byte_col number 0-indexed byte column (from nvim_win_get_cursor)
+---@return string before Text up to and including the character at cursor
+---@return string after Text after the character at cursor
+function M.split_after_cursor(line, byte_col)
+  if #line == 0 then
+    return "", ""
+  end
+
+  -- Handle cursor past end of line
+  if byte_col >= #line then
+    return line, ""
+  end
+
+  -- Convert 0-indexed byte offset to character index
+  local char_idx = vim.fn.charidx(line, byte_col)
+  if char_idx < 0 then
+    -- Should not happen if byte_col < #line, but be safe
+    return line, ""
+  end
+
+  -- Get byte position for start of next character
+  local next_char_byte = vim.fn.byteidx(line, char_idx + 1)
+  if next_char_byte == -1 then
+    -- char_idx is at or past last character, split at end
+    return line, ""
+  end
+
+  return line:sub(1, next_char_byte), line:sub(next_char_byte + 1)
+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.