Other (squashed) commits:

fix(tui): emit ui_send output atomically with the frame

Problem:
tui_ui_send() writes directly to the TTY, bypassing the output buffer.
Sequences sent via nvim_ui_send() (e.g. kitty multiple-cursors, or
visual-dot-repeat) always arrive in a separate TTY write from the frame
they were computed for. This manifests as "tearing", or e.g. in the case
of multicursor the terminal renders text with stale cursor overlays.

Solution:
- tui_ui_send(): while a frame is being assembled (pending invalid
  regions or buffered output), buffer instead of writing directly.
  - Out-of-frame sends (tty queries, clear-on-disable) still write
    immediately.
- mcursor.lua: emit the terminal-cursor update at the end of the redraw
  cycle (`on_end`, when screen positions are final) instead of
  vim.schedule().

Problem:
The magic globals `it`, `describe`, etc., are more trouble than they are
worth.

- Hooking into `after_each` requires `getfenv()` hacks.
- They confuse luals/emmylua, because the top-level `.luarc.json` isn't
  merged with `test/.luarc.json` (apparently a luals limitation?)
- They totally defeat discoverability because the user just has to
  "know" about the various magic symbols.

So they harm DX, which means they serve no purpose at all.

Solution:
- Expose the test API from `testutil`, so tests can call `t.it()`,
  `t.describe()`, etc., in the conventional way.
- Drop `getfenv()` hacks.
- Drop the `setfenv()` injection in `load_chunk`.
- Drop `test/_meta.lua`.

Problem:  'showcmd' buffer is being populated for :norm commands, which
          can result in a recursive uv_run() when called from a msg_show
          vim.ui_attach callback for a shell message.

Solution: The 'showcmd' buffer is never displayed while executing a
          :normal command so prevent unnecessary work, avoiding the crash.

Problem: nvim_win_set_config may merge configs despite failing to configure a
split, and without applying necessary side-effects (like setting style=minimal
options). Plus, autocommands may apply a different config after the merge,
causing side-effects to apply for an outdated config.

Solution: merge configs last, only on success. Include fields only relevant to
splits. Properly set _cmdline_offset for splits.

Maybe better to disallow _cmdline_offset for splits instead, as the pum is
relative to cmdline_row anyway? (I didn't want to change behaviour too much)

Also use expect_unchanged in an unrelated test to quash a warning.

Problem:  Newlines emitted with ext_messages intended to position
          the message/prompt on the message grid.
Solution: Don't emit these newlines with ext_messages, followup to 4260f73e.

Problem:  Logic determining messages belonging to the last command to
          show with "g<" does not flush pending messages. This can
          result in clearing the temporary message history before a
          message still belonging to the previous command was emitted.
Solution: Flush pending messages when marking the end of messages
          belonging to previous command.

Problem:  Fast context for msg_show event inhibits vim.ui_attach from
          displaying a stream of messages from a single command.

Solution: Remove fast context from msg_show events emitted as a result
          of explicit API/command calls. The fast context was originally
          introduced to prevent issues with internal messages.

Problem:  popup: there are some position logic bugs
Solution: Refactor position logic and fix a few bugs
          (Girish Palya).

This change does the following:

- Simplified and rewrote horizontal positioning logic (was overly
  complex).
- Split horizontal and vertical positioning into separate functions.
- Fixed missing truncation marker (e.g. `>`) when items were truncated
  and `pummaxwidth` was not set.
- Fixed occasional extra space being added to menu items.
- Update tests

closes: vim/vim#18441

https://github.com/vim/vim/commit/e3ed5584ed7fad7070d49ddfba4f41bc2b22ca97

Cherry-pick pum_display_{rtl,ltr}_text() changes from patch 9.1.1835.

Co-authored-by: Girish Palya <girishji@gmail.com>

Solution:  A window marked for ext_cmdline usage is still referenced
           after it is already closed and freed.
Solution:  Unset cmdline_win when the window is closed.

Problem:  Hardcoded highlight IDs for ext_messages/cmdline output need
          to be adjusted everytime a builtin highlight group is added.
Solution: Store a global map of default highlights through nvim_get_hl()
          and fetch missing (custom) highlight groups through synIDattr().
          Use more compact formatting for screen:expect().