The Lua modules that make up vim.lua are embedded as raw source files into the
nvim binary. These sources are loaded by the Lua runtime on startuptime. We can
pre-compile these sources into Lua bytecode before embedding them into the
binary, which minimizes the size of the binary and improves startuptime.
Fixes `q` in more pager, where `:highlight` can be quit out of with a
single `q` keystroke, while in `:lua print(vim.inspect(vim))` it just
scrolls down a page.
Problem: Giving error messages is not flexible.
Solution: Add semsg(). Change argument from "char_u *" to "char *", also
for msg() and get rid of most MSG macros. (Ozaki Kiichi, closes
vim/vim#3302) Also make emsg() accept a "char *" argument. Get rid of
an enormous number of type casts.
f9e3e09fdc
Fixes#15147 and fixes#15497. Also sketch "subdir" caching. Currently
this only caches whether an rtp entry has a "lua/" subdir but we could
consider cache other subdirs potentially or even "lua/mybigplugin/"
possibly.
Note: the async_leftpad test doesn't actually fail on master, at least
not deterministically (even when disabling the fast_breakcheck
throttling). It's still useful as a regression test for further changes
and included as such.
* add vim.json.encode and vim.json.decode
* use vim.NIL instead of cjson.null
* resolve strict-prototypes warnings
* The following benchmark shows an approximately 2.5x (750 ms vs 300 ms) improvement in deserialization performance over
vim.fn.json_decode on a medium package.json
```lua
local uv = vim.loop
local function readfile(path)
return
end
local json_url = "https://raw.githubusercontent.com/rust-analyzer/rust-analyzer/b24c8d5c89ee93d1172b4127564f5da3b0c88dad/editors/code/package.json"
io.popen(string.format('curl -v -f -L -O %q &> /dev/null', json_url))
local json_string = io.open('package.json'):read '*a'
uv.update_time()
local start = uv.hrtime()
for i = 1,1000 do
vim.fn.json_decode(json_string)
end
uv.update_time()
print(string.format("Deserialization time vim.fn.json_decode: %s ms", (uv.hrtime() - start) * (1e-6)))
uv.update_time()
local start = uv.hrtime()
for i = 1,1000 do
vim.json.decode(json_string)
end
uv.update_time()
print(string.format("Deserialization time vim.json.decode: %s ms", (uv.hrtime() - start) * (1e-6)))
```
Co-Authored-By: Björn Linse <bjorn.linse@gmail.com>
Analogous to nodejs's `on('data', …)` interface, here on_key is the "add
listener" interface.
ref 3ccdbc570d#12536
BREAKING_CHANGE: vim.register_keystroke_callback() is now an error.
fixes#15524
Note: this is obviously a quickfix. A scalabe solution will
involve being able to specify a _list_ of modules to be
put into packages.preload, without needing to manually copypasta
a blurb of C code. Perhaps even involving bytecode for
static builds (to speedup initialization)
Note: the reason for removing them is not that there after this refactor
is no use of them, but rather that having them available is an
anti-pattern: they manange an _extra_ heap allocation which has
nothing to do with the functionality of the map itself (khash
manages the real buffers internally). In case there happens to
be a reason to allocate the map structure itself later, this
should be made explicit using xcalloc/xfree calls.
* feat(api): add lua C bindings for xdiff
* chore: opt.hunk_lines -> opt.result_type
opt.on_hunk now takes precedence over opt.result_type
* chore: fix indents
Fix indents
* chore: change how priv is managed
Assign priv NULL and unconditionally apply XFREE_CLEAR to it when
finished.
Problem: Function call functions have too many arguments.
Solution: Pass values in a funcexe_T struct.
c6538bcc1c
Use FUNCEXE_INIT to initialize funcexe_T instances.
call_callback() and other Vim listener related stuff is N/A.
lua is used as part of implementation for more core features. As an
example, every user keypress will invoke a lua function to check for
keypress handlers (regardless if they are registered or not). Thus not
starting lua until it is first used doesn't make much sense anymore.
nlua_enter was also needed due to the earlier stateful &rtp
translation, which by now have been made stateless.
When a keymap is set from lua currently verbose message says
it's set from line 1. That's incorrect because we don't really know when
it was set. So until proper :verbose support isn't added for sourceing
lua it shouldn't say where it was set at.