fix: wrong buffer encoding breaks accepting suggestions

[traysh]

These changes ensure that the text edits are applied using the correct encoding for the current buffer, improving compatibility with different file encodings.

The changes made in the suggestion.lua file are as follows:

1. Dynamic Encoding Retrieval:

   • The code now dynamically retrieves the file encoding of the current buffer using vim.api.nvim_get_option_value('fileencoding', { buf = bufnr }).
   • This replaces the previous hardcoded "utf-16" encoding.

2. Buffer Number Handling:

   • The current buffer number is stored in the bufnr variable using vim.api.nvim_get_current_buf().
   • This buffer number is then used to retrieve the encoding and apply text edits.

[tsukimizake]

I surely want this merged. Without this patch, it always shows this error when completing Japanese texts.

[xebecnan]

Also want this merged. It fixes the issue #344.

[Kaiser-Yang]

This merge needed for the all non-anscii files whose encoding is not utf-16.