## Problem
With 'winfixbuf' enabled, `nvim_win_set_buf` and `nvim_set_current_buf` fail
even if targeting the already-current buffer.
vim.wo.winfixbuf = true
vim.api.nvim_win_set_buf(0, 0)
vim.api.nvim_set_current_buf(0)
Solution:
Check for this condition.
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: Not possible to only set a "redraw later" type with
nvim__redraw, which seems to be desired for the
treesitter highlighter.
Solution: Do not update the screen when "flush" is explicitly set to
false and only redraw later types are present. In that case,
do not call ui_flush() either.
Problem: nvim__redraw's "range" marks a buffer range for redraw, and subsequently
flushes the UI without updating the windows containing that buffer.
Solution: Implicitly update the screen, unless specified otherwise.
Only update the screen with the last call of the treesitter
on_changedtree() callback.
Problem: Documentation for "hl_group" in nvim_buf_set_extmark() is
unclear. "hl_group" in nvim_echo() does not accept
highlight group id.
Solution: Move documentation for highlight group name/id to first
mention of hl_group. Update nvim_echo() to accept highlight
group id.
Problem:
The way option scopes currently work is inflexible and does not allow for nested
option scopes or easily finding the value of an option at any arbitrary scope
without having to do long handwritten switch-case statements like in
`get_varp()`. `.indir` is also confusing and redundant since option indices for
each scope can be autogenerated.
Solution:
Expand option scopes in such a way that an option can support any amount of
scopes using a set of scope flags, similarly to how it's already done for option
types. Also make options contain information about its index at each scope it
supports. This allows for massively simplifying `get_varp()` and
`get_varp_scope()` in the future by just using a struct for options at each
scope. This would be done by creating a table that stores the offset of an
option's variable at a scope by using the option's index at that scope as a key.
This PR also autogenerates enums for option indices at each scope to remove the
need for `.indir` entirely, and also to allow easily iterating over options all
options that support any scope.
Ref: #29314
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.
Problem: too many strlen() calls in register.c
Solution: refactor code, add string_T struct to keep track
of string lengths (John Marriott)
closes: vim/vim#1595279f6ffd388
Co-authored-by: John Marriott <basilisk@internode.on.net>
Problem: When using nvim_paste in a mapping during a macro recording,
both the mapping and the paste are recorded, causing the paste
to be performed twice when replaying the macro.
Solution: Only record a paste when it is from RPC.
Unfortunately this means there is no way for a script to make a recorded
paste. A way to enable that can be discussed later if there is need.
Before this PR, the behavior of nvim_paste is:
- When vim.paste() returns false, return false to the client, but treat
following chunks normally (i.e. rely on the client cancelling the
paste as expected).
- When vim.paste() throws an error, still return true to the client, but
drain the following chunks in the stream without calling vim.paste().
There are two problems with such behavior:
- When vim.paste() errors, the client is still supposed to send the
remaining chunks of the stream, even though they do nothing.
- Having different code paths for two uncommon but similar situations
complicates maintenance.
This PR makes both the cancel case and the error case return false to
the client and drain the remaining chunks of the stream, which, apart
from sharing the same code path, is beneficial whether the client checks
the return value of nvim_paste or not:
- If the client checks the return value, it can avoid sending the
following chunks needlessly after an error.
- If the client doesn't check the return value, chunks following a
cancelled chunk won't be pasted on the server regardless, which leads
to less confusing behavior.
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.
- Fixes 'autoindent' being applied during redo.
- Makes redoing a large paste significantly faster.
- Stores pasted text in the register being recorded.
Fix#28561
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.
Problem:
- "process" is often used as a verb (`multiqueue_process_events`), which
is ambiguous for cases where it's used as a topic.
- The documented naming convention for processes is "proc".
- `:help dev-name-common`
- Shorter is better, when it doesn't harm readability or
discoverability.
Solution:
Rename "process" => "proc" in all C symbols and module names.
Since paste data is handled via a separate channel, the data processed via `input_buffer` is typically just explicit keys as typed in by the user. Therefore it should be fine to use `memmove()` to always put the remaining data in front when refilling the buffer.
Problem:
NetBSD's libc already has a function by the same name.
Solution:
Rename popcount to xpopcount and add #if defined(__NetBSD__) to
prefer NetBSD's own implementation. This fixes#28983.
Problem:
`vim.rpcnotify(0)` and `rpcnotify(0)` are documented as follows:
If {channel} is 0, the event is broadcast to all channels.
But that's not actually true. Channels must call `nvim_subscribe` to
receive "broadcast" events, so it's actually "multicast".
- Assuming there is a use-case for "broadcast", the current model adds
an extra step for broadcasting: all channels need to "subscribe".
- The presence of `nvim_subscribe` is a source of confusion for users,
because its name implies something more generally useful than what it
does.
Presumably the use-case of `nvim_subscribe` is to avoid "noise" on RPC
channels not expected a broadcast notification, and potentially an error
if the channel client reports an unknown event.
Solution:
- Deprecate `nvim_subscribe`/`nvim_unsubscribe`.
- If applications want to multicast, they can keep their own multicast
list. Or they can use `nvim_list_chans()` and `nvim_get_chan_info()`
to enumerate and filter the clients they want to target.
- Always send "broadcast" events to ALL channels. Don't require channels
to "subscribe" to receive broadcasts. This matches the documented
behavior of `rpcnotify()`.
Experimental and subject to future changes.
Add a way to redraw certain elements that are not redrawn while Nvim is waiting
for input, or currently have no API to do so. This API covers all that can be
done with the :redraw* commands, in addition to the following new features:
- Immediately move the cursor to a (non-current) window.
- Target a specific window or buffer to mark for redraw.
- Mark a buffer range for redraw (replaces nvim__buf_redraw_range()).
- Redraw the 'statuscolumn'.
Instead of painfully messing with timing to determine if queries were
reparsed, we can simply keep a counter next to the call to ts_query_new
Also memoization had a hidden dependency on the garbage collection of
the the key, a hash value which never is kept around in memory. this was
done intentionally as the hash does not capture all relevant state for the
query (external included files) even if actual query objects still
would be reachable in memory. To make the test fully deterministic in
CI, we explicitly control GC.
TODO:
FUNC_API_REMOTE_ONLY APIs such as `nvim_ui_*` cannot (yet) be used in
`nvim_exec_lua`. We can change FUNC_API_REMOTE_ONLY to allow
Vimscript/Lua to pass an explicit `channel_id`. #28437
Problem: Dialog for file changed outside of Vim not tested.
Solution: Add a test. Move FileChangedShell test. Add 'L' flag to
feedkeys().
5e66b42aae
Co-authored-by: Bram Moolenaar <Bram@vim.org>
Problem: `set_string_option_direct()` contains a separate codepath specifically for setting string options. Not only is that unnecessary code duplication, but it's also limited to only string options.
Solution: Replace `set_string_option_direct()` with `set_option_direct()` which calls `set_option()` under the hood. This reduces code duplication and allows directly setting an option of any type.
Problem: assertion failure in nvim_create_buf if buflist_new autocommands open
a swapfile when "scratch" is set.
Solution: block autocommands when setting up the buffer; fire them later
instead.
Note that, unlike buflist_new, I don't check if autocommands aborted script
processing; the buffer is already created and configured at that point, so might
as well return the handle anyway.
Rather than repeat try_{start,end} and {un}block_autocmds for each relevant
operation, just do it at the start and near the end. This means that, if
TermResponse fires from unblock_autocmds for whatever reason, it can see the
buffer in an already configured state if we didn't bail due to an error (plus
it's probably a bit cleaner this way).
Then we can just load metadata in C as a single msgpack blob. Which also
can be used directly as binarly data, instead of first unpacking all the
functions and ui_events metadata to immediately pack it again, which was
a bit of a silly walk (and one extra usecase of `msgpack_rpc_from_object`
which will get yak shaved in the next PR)
Problem:
The documentation flow (`gen_vimdoc.py`) has several issues:
- it's not very versatile
- depends on doxygen
- doesn't work well with Lua code as it requires an awkward filter script to convert it into pseudo-C.
- The intermediate XML files and filters makes it too much like a rube goldberg machine.
Solution:
Re-implement the flow using Lua, LPEG and treesitter.
- `gen_vimdoc.py` is now replaced with `gen_vimdoc.lua` and replicates a portion of the logic.
- `lua2dox.lua` is gone!
- No more XML files.
- Doxygen is now longer used and instead we now use:
- LPEG for comment parsing (see `scripts/luacats_grammar.lua` and `scripts/cdoc_grammar.lua`).
- LPEG for C parsing (see `scripts/cdoc_parser.lua`)
- Lua patterns for Lua parsing (see `scripts/luacats_parser.lua`).
- Treesitter for Markdown parsing (see `scripts/text_utils.lua`).
- The generated `runtime/doc/*.mpack` files have been removed.
- `scripts/gen_eval_files.lua` now instead uses `scripts/cdoc_parser.lua` directly.
- Text wrapping is implemented in `scripts/text_utils.lua` and appears to produce more consistent results (the main contributer to the diff of this change).
To align the output of `nvim_get_hl` with its documentation -- which
points to `nvim_set_hl`, remove mentions of the keys `foreground`,
`background` and `special`.
The long keys are are still supported (via fallback checks inside
`dict2hlattrs`), but the `fg`, `bg` and `sp` keys are preferenced.
As only a few API functions make use of explicit freeing of the return
value, make it opt-in instead. The arena is always present under the
hood, so `Arena *arena` arg now doesn't mean anything other than getting
access to this arena. Also it is in principle possible to return an
allocated value while still using the arena as scratch space for other
stuff (unlikely, but there no reason to not allow it).
The way ml_replace_buf is implemented makes it unfriendly for
being used in a loop: every call allocates a scratch buffer for putting
the line into the "dirty" state. This then immediately needs to be freed
as the next ml_replace_buf and/or ml_append_buf call will flush that buffer.
It's better to later pay the price of allocating the scratch buffer only if
the line is being immediately edited (likely when using the API to only
change one line) with an extra memcpy, than allocating that buffer
multiple times every time the API is called.
Of course, a separate xmalloc/xfree cycle for each time the dirty line
changes is unwanted to begin with. But fixing that is a later refactor.
