Problem: The cursor shape is changed to indicate when it is behind an
unfocused floating window (since a2b92a5e). This behavior
cannot be controlled by a floating window that doesn't want
to dim the cursor.
Solution: Assign a zindex-offset of 50 to the zindex of the current
window. To not dim the cursor when creating a floating window
on top of the current window one can assign the zindex
accordingly.
Problem: vim.ui_attach() is unable to display streamed shell output,
and will display it as individual messages.
Unwanted newlines in "shell_ret" message.
Solution: Treat the "shell_*" kinds as non-fast and set msg_show->append
for the streamed stdout/err messages.
Remove leading newline from (translated) message with
ext_messages, remove trailing newline altogether.
Problem:
a2b92a5efb is regressive if the
ui_mode_idx set by ui_cursor_shape differs from what ui_flush later sets
it to. (e.g: if mode changes between both calls)
Solution:
Don't overwrite ui_mode_idx to obscure the cursor so that we don't
forget what shape was last set by ui_cursor_shape.
May not be needed, but does fix a potential regression. (and makes it
easier to set ui_mode_idx to something other than cursor_get_mode_idx's
retval in other places, if that's ever wanted)
Problem:
There is no indication of when cursor is "behind" an unfocused, floating window.
Solution:
Set underline cursor to indicate when an unfocused floatwin is over the active cursor.
Problem: There are still ways to run into textlock errors with
vim.ui_attach callbacks trying to display a UI event.
Solution: Disregard textlock again during vim.ui_attach() callbacks
(also when scheduled). Partially revert 3277dc3b; avoiding
to flush while textlock is set is still helpful.
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: UI callbacks disregard textlock unnecessarily (since d909de2).
Solution: In case the UI is flushed during textlock, make sure Nvim does
not flush UI callbacks that are expected to change text. Instead
assume postponing callbacks during textlock is insignificant,
and will safely happen soon after.
These are not needed after #35129 but making uncrustify still play nice
with them was a bit tricky.
Unfortunately `uncrustify --update-config-with-doc` breaks strings
with backslashes. This issue has been reported upstream,
and in the meanwhile auto-update on every single run has been disabled.
Problem: When 'winminheight' is zero and the window height is set to
zero, the actual height is clamped whereas the stored config
value is not. Reciprocal window configuration through
nvim_win_get_config() then results in an error.
Solution: Also clamp the stored dimensions in the window config.
Problem: ext_messages is implemented to mimic the message grid
implementation w.r.t. scrolling messages, clearing scrolled
messages, hit-enter-prompts and replacing a previous message.
Meanwhile, an ext_messages UI may not be implemented in a way
where these events are wanted. Moreover, correctness of these
events even assuming a "scrolled message" implementation
depends on fragile "currently visible messages" global state,
which already isn't correct after a previous message was
supposed to have been overwritten (because that should not only
happen when `msg_scroll == false`).
Solution: - No longer attempt to keep track of the currently visible
messages: remove the `msg_ext(_history)_visible` variables.
UIs may remove messages pre-emptively (timer based), or never
show messages that don't fit a certain area in the first place.
- No longer emit the `msg(_history)_clear` events to clear
"scrolled" messages. This opens up the `msg_clear` event to
be emitted when messages should actually be cleared (e.g.
when the screen is cleared). May also be useful to emit before
the first message in an event loop cycle as a hint to the UI
that it is a new batch of messages (vim._extui currently
schedules an event to determine that).
- Set `replace_last` explicitly at the few callsites that want
this to be set to true to replace an incomplete status message.
- Don't store a "keep" message to be re-emitted.
Problem: Cmdline buffer is not cleared for a new message (since c973c7ae),
resulting in an incorrect spill indicator. When the cmdline
buffer is cleared, "msg_row" is not invalidated, resulting in
an error. The extui module is untested.
Return value of `vim.ui_attach()->callback` is undocumented.
Solution: Clear the cmdline buffer for the first message in an event
loop iteration. Ensure msg_row passed as end_row does not
exceed buffer length.
Add `messages_spec2.lua` to test the extui module, keeping in
mind that test coverage will greatly increase if this UI is made
the default. As such, only tests for specific extui functionality
unlikely to be covered by tests leveraging the current message grid.
Document the return value of `vim.ui_attach()->callback`, it seems
to make sense, and is also used to suppress remote UI events in
`messages_spec2.lua`.
Problem: Since 48e2a736, prompt messages are handled by an actual
active cmdline, resulting in `State` no longer being equal
to `MODE_CONFIRM` which is used in some places. E.g. to
specify the current `mode()` or to re-emit a confirm message.
Solution: Replace `MODE_CONFIRM` with a new `MODE_CMDLINE` sub-mode when
`ccline.one_key/mouse_used` is set. Use it to avoid clearing
mouse_used prompt messages, and to re-emit one_key messages
(when ext_messages is inactive, for which this is unnecessary).
Problem:
Cursor is visible in "hidden" floating window.
Solution:
Hide cursor when curwin is a hidden floating window.
Show cursor after returning to a normal (non-hidden) window.
Problem: We allow setting 'cmdheight' to 0 with ext_messages enabled
since b72931e7. Enabling ext_messages with vim.ui_attach()
implicitly sets 'cmdheight' to 0 for BWC. When non-zero
'cmdheight' is wanted, this behavior make it unnecessarily
hard to keep track of the user configured value.
Solution: Add set_cmdheight to vim.ui_attach() opts table that can be
set to false to avoid setting 'cmdheight' to 0.
Problem: Error messages that cause a vim.ui_attach() namespace to
detach are not visible in the message history. Decoration
provider and vim.ui_attach error messages are dissimilar.
Solution: Emit vim.ui_attach() errors as an actual message in addition
to logging it. Adjust error message format.
Problem: Assert hit related to message kind, which is reset after a
ext_messages UI is forcibly detached, so the assertion is
expectedly false. Assert hit related to message grid variables
after an ext_messages UI attaches while message grid is scrolled.
Solution: Don't check message kind assertion if no ext_messages UI is
attached. Flush message grid when first/last ext_messages UI
attaches/detaches.
Problem: Prompts are emitted as messages events, where cmdline events
are more appropriate. The user input is also emitted as
message events in fast context, so cannot be displayed with
vim.ui_attach().
Solution: Prompt for user input through cmdline prompts.
Problem: Option metadata like list of valid values for an option and
option flags are not listed in the `options.lua` file and are instead
manually defined in C, which means option metadata is split between
several places.
Solution: Put metadata such as list of valid values for an option and
option flags in `options.lua`, and autogenerate the corresponding C
variables and enums.
Supersedes #28659
Co-authored-by: glepnir <glephunter@gmail.com>
Problem: No longer able to show prompt messages with vim.ui_attach().
Solution: Do not execute callback in fast context for prompt message
kinds. These events must be safe to show the incoming message
so the event itself serves to indicate that the message
should be shown immediately.
Problem: Lua callbacks for "msg_show" events with vim.ui_attach() are
executed when it is not safe.
Solution: Disallow non-fast API calls for "msg_show" event callbacks.
Automatically detach callback after excessive errors.
Make sure fast APIs do not modify Nvim state.
Problem: Highlight group id is not propagated to the end of the message call
stack, where ext_messages are emitted.
Solution: Refactor message functions to pass along highlight group id
instead of attr id.
In the api_info() output:
:new|put =map(filter(api_info().functions, '!has_key(v:val,''deprecated_since'')'), 'v:val')
...
{'return_type': 'ArrayOf(Integer, 2)', 'name': 'nvim_win_get_position', 'method': v:true, 'parameters': [['Window', 'window']], 'since': 1}
The `ArrayOf(Integer, 2)` return type didn't break clients when we added
it, which is evidence that clients don't use the `return_type` field,
thus renaming Dictionary => Dict in api_info() is not (in practice)
a breaking change.
Problem:
The default builtin UI client does not declare its client info. This
reduces discoverability and makes it difficult for plugins to identify
the UI.
Solution:
- Call nvim_set_client_info after attaching, as recommended by `:help dev-ui`.
- Also set the "pid" field.
- Also change `ui_active()` to return a count. Not directly relevant to
this commit, but will be useful later.
The ui_rgb_attached function determines if any UI is attached which
supports RGB (truecolor). We determine if the TUI supports RGB via the
'termguicolors' option which is checked at the beginning of this
function. If the TUI does not support RGB ('termguicolors' is unset), we
check to see if any _other_ UI is attached which supports RGB.
Normally, the TUI's "rgb" flag and the 'termguicolors' option are the
same. However, they may differ during startup when the "rgb" flag is set
by tui/tui.c to indicate to the core that the terminal emulator supports
truecolor. The 'termguicolors' option is not actually set until
_defaults.lua runs.
Problem: Calling :redraw from vim.ui_attach() callback results in
recursive cmdline/message events.
Solution: Avoid recursiveness where possible and replace global "call_buf"
with separate, temporary buffers for each event so that when a Lua
callback for one event fires another event, that does not result
in invalid memory access.
Problem: Enabling ext_messages claims to set 'cmdheight' to zero, but
only does so for the current tabpage.
Solution: Set stored 'cmdheight' value to zero for all tabpages.
Also close Nvim instance before removing log file, otherwise the Nvim
instance will still write to the log file.
Also adjust log level in libuv_process_spawn(). Ref #27660
Problem: Cursor line is unconcealed when pressing 'r' in Normal mode
when 'concealcursor' contains 'n' but not 'i'.
Solution: Don't check conceal when pressing 'r' in Normal mode.
Vim doesn't have this problem because it doesn't call redrawWinline() in
conceal_check_cursor_line() and instead sets a global variable.
Just some basic spring cleaning.
In the distant past, not all UI:s where remote UI:s. They still aren't,
but both of the "UI" and "UIData" structs are now only for remote UI:s.
Thus join them as "RemoteUI".
and for return value of nlua_exec/nlua_call_ref, as this uses
the same family of functions.
NB: the handling of luaref:s is a bit of a mess.
add api_luarefs_free_XX functions as a stop-gap as refactoring
luarefs is a can of worms for another PR:s.
as a minor feature/bug-fix, nvim_buf_call and nvim_win_call now preserves
arbitrary return values.
When an embedded Nvim instance changes its current directory a "chdir"
UI event is emitted. Attached UIs can use this information however they
wish. In the TUI it is used to synchronize the cwd of the TUI process
with the cwd of the embedded Nvim process.
Remove `export` pramgas from defs headers as it causes IWYU to believe
that the definitions from the defs headers comes from main header, which
is not what we really want.
grid_put_linebuf() used grid_clear() internally to handle clearing the
left part of a rightleft line. By reimplementing that internally, we
can instead use grid_put_linebuf() as the implementation of
grid_clear(), which in the end is a net reduction of code as grid_fill()
is used for nothing else.
win_draw_end: Implement "draw_margin" on a per-row basis which closer reflects how
terminals work. Also use the magic mirror for 'rightleft'
