For anonymous scripts, defer the creation of script items until an attempt to access a script-local
variable is made. This dramatically reduces the number of script items created when using lots of
vim.cmd and nvim_exec especially.
This will mean <SID> usage fails until a script-local variable access is first made.
Based on #13143 (and #11507) with changes:
- Omit script_type_E. Use sn_name == NULL to determine anon items.
- Keep SID_STR. Used by anon :source for .lua files (no item).
- Show SID in get_scriptname output (:verbose set).
- Factor item creation into new_script_item.
- Leave sc_seq = 0 (anon scripts don't re-use the same item when re-sourced).
- Add tests for anon :source.
Co-authored-by: Vikram Pal <vikrampal659@gmail.com>
Co-authored-by: Justin M. Keyes <justinkz@gmail.com>
Adds the following API functions.
- nvim_buf_set_mark(buf, name, line, col)
* Set marks in a buffer.
- nvim_buf_del_mark(buf, name)
* Delete a mark that belongs to buffer.
- nvim_del_mark(name)
* Delete a global mark.
- nvim_get_mark(name)
* Get a global mark.
Tests:
- Adds test to all the new api functions, and adds more for the existing
nvim_buf_get_mark.
* Tests include failure cases.
Documentation:
- Adds documentation for all the new functions, and improves the
existing fucntion docs.
Do not copy a lot of lua strings (dict keys) to just strequal() them
Just compare them directly to a dedicated hash function.
feat(generators): HASHY McHASHFACE
Now remove the addition of "start/*" packages in 'packpath' as
explicit items in 'runtimepath'. This avoids 'runtimepath' from becoming
very long when using a lot of plugins as packages.
To get the effective search path as a list, use |nvim_list_runtime_paths()|
Problem:
Anonymous :source (no args) and nvim_exec() don't support Vimscript line continuations.
Solution:
Factor out the concat logic into concat_continued_line() and a
CONCAT_CONTINUED_LINES macro for simple concatenations where lines are
fetched individually.
Closes#14807
Iterating over PARAM_COUNT is wrong as PARAM_COUNT also counts the last
element of the options array, which has a NULL fullname in order to
signal the end of the array.
- The previous commit lost information in the tests. Instead, add some
more "normalization" substitutions in pcall_err(), so that the general
shape of the stacktrace is included in the asserted text.
- Eliminate contains(), it is redundant with matches()
- We already find ourselves renaming nvim_execute_lua in tests and
scripts, which suggests "exec" is the verb we actually want.
- Add "exec" verb to `:help dev-api`.
- Rename `meth_pcall`.
- Make `pcall_err` raise an error if the function does not fail.
- Add `vim.pesc()` to treat a string as literal where a Lua pattern is
expected.
- Always load files when cleaning up jumplist.
- For Shada: avoids writing duplicate entries, which happens when you read
from a shada file with duplicate entries (merging the jumplist while
writing sometimes produces duplicate entries, bug?) and then write right
away (i.e.: without any `:jumps`, `getjumplist()`, or any jump movement,
that is: nothing that calls `cleanup_jumplist` with `loadfiles == true`).
- For Context: avoids non-idempotent behavior for the same reason (i.e.:
first call to `shada_encode_jumps` does not remove duplicate entries).
- Do not set pcmark when dumping jumplist for Context.
- Retrieving current Context shouldn't add an entry to the jumplist
(which will be removed by a subsequent `cleanup_jumplist` anyway, i.e.:
tail entry matching current position), just act like `getjumplist` for
instance.
luassert uses 3 by default, which is often not enough.
Instead of documenting how to increase it, let's use a more fitting
(sane) default of 100 levels.
- All "chunks" in a paste-stream should form a single undo-block. Side
effect of 7a85792884 was to create an undo-block for each chunk.
- Also: remove old :redraw force logic, irrelevant after 7a85792884.
- Introduce TRY_WRAP() until we have an *architectural* solution.
- TODO: bfredl idea: prepare error-handling at "top level" (nv_event).
- nvim_paste(): Revert luaeval() hack (see parent commit).
- With TRY_WRAP() in nvim_put(), 'nomodifiable' error now correctly
"bubbles up".
- nvim_paste(): Marshal through luaeval() instead of nvim_execute_lua()
because the latter seems to hide some errors.
- Handle 'nomodifiable' in `nvim_put()` explicitly.
- Require explicit `false` from `vim.paste()` in order to "cancel",
otherwise assume true ("continue").
- Show error only once per "paste stream".
- Drain remaining chunks until phase=3.
- Lay groundwork for "cancel".
- Constrain semantics of "cancel" to mean "client must stop"; it is
unrelated to presence of error(s).
Fixes strange behavior where sometimes the buffer contents of a series
of paste chunks (vim._paste) would be out-of-order.
Now the tui_spec.lua screen-tests are much more reliable. But they still
sometimes fail because of off-by-one cursor (caused by "typeahead race"
resulting in wrong mode; fixed later in this patch-series).
Rationale: the purpose of nvim_execute_lua is to simply call lua code with lua
values. If a lua function expects a floating point value, it should be enough
to specify a float as argument to nvim_execute_lua.
However, make sure to preserve the existing roundtripping behavior of
API values when using `vim.api` functions. This is covered by existing
lua/api_spec.lua tests.
The test.functional.helpers and test.unit.helpers modules now include
all of the public functions from test.helpers, so there is no need to
separately require('test.helpers').
Before now, Nvim always degrades UI capabilities to the lowest-common
denominator. For example, if any connected UI has `ext_messages=false`
then `ext_messages=true` requested by any other connected UI is ignored.
Now `nvim_ui_attach()` supports `override=true`, which flips the
behavior: if any UI requests an `ext_*` UI capability then the
capability is enabled (and the legacy behavior is disabled).
Legacy UIs will be broken while a `override=true` UI is connected, but
it's useful for debugging: you can type into the TUI and observe the UI
events from another connected (UI) client. And the legacy UI will
"recover" after the `override=true` UI disconnects.
Example using pynvim:
>>> n.ui_attach(2048, 2048, rgb=True, override=True, ext_multigrid=True, ext_messages=True, ext_popupmenu=True)
>>> while True: n.next_message();
CA_COMMAND_BUSY in nv_event() was carried over from Vim nv_cursorhold()
(ref: e5165bae11). It prevents :startinsert from working during a RPC
call, so remove it.
Helped-by: glacambre <me@r4>
closes#7254
Loading existing files into a buffer is non-trivial and requires a window.
Creating an unnamed emtpy buffer is trivial and safe though, thus worth a
special case.
Change nvim_buf_set_option to use aucmd_prepbuf. This is necessary
to allow some options to be set on a not yet displayed buffer, such
as 'buftype' option.
vim-patch:7.4.1858: Add BLN_NEW to enforce buflist_new creating new buffer
Namespaces is a lightweight concept that should be used to group
objects for purposes of bulk operations and introspection. This is
initially used for highlights and virtual text in buffers, and is
planned to also be used for extended marks. There is no plan use them
for privileges or isolation, neither to introduce nanespace-level
options.
msgpack_rpc_to_object (called by handle_request .. msgpack_rpc_to_array)
always NUL-terminates API Strings.
But handle_request .. msgpack_rpc_get_handler_for operates on a raw
msgpack_object, before preparation.
Add ext_newgrid and ext_hlstate extensions. These use predefined
highlights and line-segment based updates, for efficiency and
simplicity.. The ext_hlstate extension in addition allows semantic
identification of builtin and syntax highlights.
Reimplement the old char-based updates in the remote UI layer, for
compatibility. For the moment, this is still the default. The bulitin
TUI uses the new line-based protocol.
cmdline uses curwin cursor position when ext_cmdline is active.
- Return VimL errors instead of generic errors for:
- nvim_call_function
- nvim_call_dict_function
- Fix tests which were silently broken before this change.
This violates #6150 where we agreed not to translate API errors. But
that can be fixed later.
The `internal` param is difficult to explain, and will rarely be
anything but `true`. To avoid it, use a hack: check if the resolved
dict value starts with "function(".
Make `:verbose set ...` show when an option was last modified by an
API client or Lua script/chunk. In the case of an API client, the
channel ID is displayed.
