fix(bufwrite.c): handle invalid byte sequences #37363

Problem:

bw_rest was used as an extra buffer to save incomplete byte sequences
between calls to buf_write_bytes. Besides being unnecessarily
complicated, this introduced a number of issues:

1) The bytes stored in bw_rest could still be there at the end of
writing the file, never having been written, thus losing some of the
file content on write.

2) bw_rest was not cleared out after the "checking_conversion" phase,
leaving them to affect the written file content during the writing
phase, corrupting the file.

3) bw_rest could contain extra bytes that need to be written to the
output buffer during a buf_write_convert call, potentially before any
bytes are consumed. But some conversions are in-place, without a
separate output buffer. Writing bytes from bw_rest to the "output"
buffer actually overwrote bytes from the input buffer before they were
read, corrupting the data to be written.

4) The extra bytes in bw_rest that need to be written to the conversion
output buffer were not originally accounted for in the size calculation
for the output buffer, causing a buffer overflow (previously fixed in
Vim patch 9.1.2028).

Solution:

Rather than maintaining a separate buffer, the unconverted bytes at the
end of the buffer can just be shifted to the beginning of the buffer,
and the buffer size updated. This requires a bit of refactoring, and
buf_write_convert and buf_write_convert_with_iconv need to report the
number of bytes they consumed so that buf_write_bytes can handle the
remaining bytes.

Following conversion, bw_buf can be checked for any remaining bytes.
Leftover bytes in this case result in a conversion error, which is
better than silently dropping them.

A short section of dead code was removed from buf_write_convert, for
converting a non-UTF-8 buffer to UTF-8. Neovim buffers are always UTF-8.

A few additional tests for iconv conversions have been added. Vim's
iconv tests are disabled in Neovim because they use unsupported values
for 'encoding'.

test/functional/ex_cmds/write_spec.lua

@@ -11,6 +11,7 @@ local api = n.api
 local skip = t.skip
 local is_os = t.is_os
 local is_ci = t.is_ci
+local read_file = t.read_file
 
 local fname = 'Xtest-functional-ex_cmds-write'
 local fname_bak = fname .. '~'
@@ -181,6 +182,99 @@ describe(':write', function()
     vim.uv.fs_symlink(fname_bak .. ('/xxxxx'):rep(20), fname)
     eq("Vim(write):E166: Can't open linked file for writing", pcall_err(command, 'write!'))
   end)
+
+  it('fails converting a trailing incomplete sequence', function()
+    -- From https://github.com/neovim/neovim/issues/36990, an invalid UTF-8 sequence at the end of
+    -- the file during conversion testing can overwrite the rest of the file during the real
+    -- conversion.
+
+    api.nvim_buf_set_lines(0, 0, 1, true, { 'line 1', 'line 2', 'aaabbb\235\128' })
+    command('set noendofline nofixendofline')
+
+    eq(
+      "Vim(write):E513: Write error, conversion failed in line 3 (make 'fenc' empty to override)",
+      pcall_err(command, 'write ++enc=latin1 ' .. fname)
+    )
+  end)
+
+  it('converts to latin1 with an invalid sequence at buffer boundary', function()
+    -- From https://github.com/neovim/neovim/issues/36990, an invalid UTF-8 sequence that falls
+    -- right at the end of the 8 KiB buffer used for encoding conversions causes subsequent data to
+    -- be overwritten.
+
+    local content = string.rep('a', 1024 * 8 - 1) .. '\251' .. string.rep('b', 20)
+    api.nvim_buf_set_lines(0, 0, 1, true, { content })
+    command('set noendofline nofixendofline fenc=latin1')
+    command('write ' .. fname)
+
+    local tail = string.sub(read_file(fname) or '', -10)
+    eq('bbbbbbbbbb', tail)
+  end)
+
+  it('converts to CP1251 with iconv', function()
+    api.nvim_buf_set_lines(
+      0,
+      0,
+      1,
+      true,
+      { 'Привет, мир!', 'Это простой тест.' }
+    )
+    command('write ++enc=cp1251 ++ff=unix ' .. fname)
+
+    eq(
+      '\207\240\232\226\229\242, \236\232\240!\n'
+        .. '\221\242\238 \239\240\238\241\242\238\233 \242\229\241\242.\n',
+      read_file(fname)
+    )
+  end)
+
+  it('converts to GB18030 with iconv', function()
+    api.nvim_buf_set_lines(0, 0, 1, true, { '你好，世界！', '这是一个测试。' })
+    command('write ++enc=gb18030 ++ff=unix ' .. fname)
+
+    eq(
+      '\196\227\186\195\163\172\202\192\189\231\163\161\n'
+        .. '\213\226\202\199\210\187\184\246\178\226\202\212\161\163\n',
+      read_file(fname)
+    )
+  end)
+
+  it('converts to Shift_JIS with iconv', function()
+    api.nvim_buf_set_lines(
+      0,
+      0,
+      1,
+      true,
+      { 'こんにちは、世界！', 'これはテストです。' }
+    )
+    command('write ++enc=sjis ++ff=unix ' .. fname)
+
+    eq(
+      '\130\177\130\241\130\201\130\191\130\205\129A\144\162\138E\129I\n'
+        .. '\130\177\130\234\130\205\131e\131X\131g\130\197\130\183\129B\n',
+      read_file(fname)
+    )
+  end)
+
+  it('fails converting an illegal sequence with iconv', function()
+    api.nvim_buf_set_lines(0, 0, 1, true, { 'line 1', 'aaa\128bbb' })
+
+    eq(
+      "Vim(write):E513: Write error, conversion failed (make 'fenc' empty to override)",
+      pcall_err(command, 'write ++enc=cp1251 ' .. fname)
+    )
+  end)
+
+  it('handles a multi-byte sequence crossing the buffer boundary converting with iconv', function()
+    local content = string.rep('a', 1024 * 8 - 1) .. 'Дbbbbb'
+    api.nvim_buf_set_lines(0, 0, 1, true, { content })
+    -- Skip the backup so we're testing the "checking" phase also.
+    command('set nowritebackup')
+    command('write ++enc=cp1251 ++ff=unix ' .. fname)
+
+    local expected = string.rep('a', 1024 * 8 - 1) .. '\196bbbbb\n'
+    eq(expected, read_file(fname))
+  end)
 end)
 
 describe(':update', function()