Problem: no error check for neg values for 'messagesopt'
(after v9.1.0908)
Solution: add additional error checks and tests (h-east)
closes: vim/vim#1618765be834c30
Nvim's getdigits() checks for overflow, so the code change isn't needed.
Co-authored-by: h-east <h.east.727@gmail.com>
Problem: Variable name for 'messagesopt' doesn't match short name
(after v9.1.0908)
Solution: Change p_meo to p_mopt. Add more details to docs.
closes: vim/vim#161828cc43daee1
Problem: 'messagesopt' does not check max wait time
(after v9.1.0908)
Solution: Check for max wait value
(Shougo Matsushita)
closes: vim/vim#16183d9e9f89e0f
Co-authored-by: Shougo Matsushita <Shougo.Matsu@gmail.com>
Problem: :sleep! not hiding the cursor is an arbitrary difference from
Vim without obvious justification, and Vim's behavior isn't
easily achievable in Nvim.
Solution: Make :sleep! hide the cursor while sleeping.
Ref:
6a01b3fcc3b5c0ade437
Problem:
vim.json.encode escapes every slash in string values (for example in
file paths), and is not optional. Use-case is for preventing HTML
injections (eg. injecting `</script>` closing tag); in the context of
Nvim this is rarely useful.
Solution:
- Add a `escape_slash` flag to `vim.json.encode`.
- Defaults to `false`. (This is a "breaking" change, but more like
a bug fix.)
Problem: When completion leader changes, text that might be reinserted
immediately after is deleted. This unnecessarily affects
extmarks. #31387 restored the original extmarks but that
prevents end_right_gravity marks from growing.
Solution: Avoid deleting leader text that will be reinserted.
Problem: In an autocommand Lua callback whether `args.file` is expanded
depends on whether `expand('<afile>')` has been called.
Solution: Always use the unexpanded file name for `args.file`.
Related to #31306 and vim/vim#16106. This doesn't provide `sfname`, but
at least makes `args.file` have a consistent value.
Problem: Marks that are properly restored by the splice associated with
an undo edit, are unnecessarily pushed to the undo header. This
results in incorrect mark tracking in the "copy_only"
save/restore completion path.
Solution: Avoid pushing left gravity marks at the beginning of the range,
and right gravity marks at the end of the range to the undo
header.
Problem: potential overflow in spell_soundfold_wsal()
Solution: Protect wres from buffer overflow, by checking the
length (Zdenek Dohnal)
Error: OVERRUN (CWE-119):
vim91/src/spell.c:3819: cond_const: Checking "reslen < 254" implies that
"reslen" is 254 on the false branch.
vim91/src/spell.c:3833: incr: Incrementing "reslen". The value of "reslen"
is now 255.
vim91/src/spell.c:3792: overrun-local: Overrunning array "wres" of 254
4-byte elements at element index 254 (byte offset 1019) using index
"reslen - 1" (which evaluates to 254).
3789| {
3790| // rule with '<' is used
3791|-> if (reslen > 0 && ws != NULL && *ws != NUL
3792| && (wres[reslen - 1] == c
3793| || wres[reslen - 1] == *ws))
Error: OVERRUN (CWE-119):
vim91/src/spell.c:3819: cond_const: Checking "reslen < 254" implies that
"reslen" is 254 on the false branch.
vim91/src/spell.c:3833: overrun-local: Overrunning array "wres" of 254
4-byte elements at element index 254 (byte offset 1019) using index
"reslen++" (which evaluates to 254).
3831| {
3832| if (c != NUL)
3833|-> wres[reslen++] = c;
3834| mch_memmove(word, word + i + 1,
3835| sizeof(int) * (wordlen -
(i + 1) + 1));
related: vim/vim#1616339a94d2048
Co-authored-by: Zdenek Dohnal <zdohnal@redhat.com>
Problem: Missing information in CompleteDone event
Solution: add complete_word and complete_type to v:event dict
(glepnir)
closes: vim/vim#161531c5a120a70
In order to run the marktree unit test in release mode, the test functions need to be available even when NDEBUG is defined.
Keep the body of marktree_check a nop during release builds, which limits the usefulness of the testing, but at least lets the tests run.
In order to run unittests with a release build, we need the test
functions to be accessible when NDEBUG is defined. Moving the functions
into the test fixture ensures they are available and only available for
use by the unit tests.
Problem:
- nvim_subscribe, nvim_unsubscribe were deprecated in
aec4938a21 but this wasn't set in the
API metadata.
- The function annotations
```
FUNC_API_SINCE(1) FUNC_API_REMOTE_ONLY FUNC_API_DEPRECATED_SINCE(13)
```
cause this test to fail:
```
RUN T3 api metadata functions are compatible with old metadata or have new level: 3.00 ms ERR
test/functional/api/version_spec.lua:135: function vim_subscribe was removed but exists in level 0 which nvim should be compatible with
stack traceback:
test/functional/api/version_spec.lua:135: in function <test/functional/api/version_spec.lua:128>
```
Solution:
- Set the API metadata.
- Rearrange the annotations so that FUNC_API_DEPRECATED_SINCE is 2nd:
```
FUNC_API_SINCE(1) FUNC_API_DEPRECATED_SINCE(13) FUNC_API_REMOTE_ONLY
```
In particular, make the distinction and interaction between "noinsert"
and "noselect" clearer as it was very confusing before.
closes: vim/vim#16148768728b487
Co-authored-by: dundargoc <gocdundar@gmail.com>
Problem: completion list wrong after v9.1.0891
Solution: update compl_mach_array after leader change
(glepnir)
compl_shown_match update not correct after refactoring in v9.1.0891
Unfortunately, this regressed what item is selected after leader change.
So generate compl_match_array before updating compl_shown_match range,
and split generate compl_match_array into range match_head
fixes: https://github.com/vim/vim/issues/16128
closes: https://github.com/vim/vim/pull/16129a49c077a88
Problem: building the completion list array is inefficient
Solution: refactor and improve ins_compl_build_pum() func
(glepnir)
current time complexity is O(n^2). I guess garray is not used here to save memory and avoid efficiency
is caused by heap memory allocation. A simple way is to add an extra pointer as a single linked list
to store the matching compl_T, and traverse this single linked list to generate compl_match_array.
The time complexity is O(n x m). The worst case is m=n, but we can still get a little improvement.
Because the if condition does not need to be run at one time. This should be a good solution for now.
Later we may be able to complete it in O(lgn) time. But this requires more reconstruction. So this is
the first step.
closes: #1612580b662009c
Problem:
gen_declarations.lua is complex and has duplicate logic with
c_grammar.lua
Solution:
Move all lpeg logic to c_grammar.lua and refactor gen_declarations.lua.
Problem: the max value of 'cmdheight' is limited by other tabpages
Solution: Limit the maximum value of 'cmdheight' to the current tabpage only.
(Milly)
The Help says that cmdheight is local to the tab page, but says nothing
about the maximum value depending on the state of all tab pages. Users
may wonder why they can't increase cmdheight when there are still rows
available on the current tab page. This PR changes the behavior of
cmdheight so that its maximum value depends only on the state of the
current tab page.
Also, since magic numbers were embedded in various places with the
minimum value of cmdheight being 1, we defined a constant to make it
easier to understand.
closes: vim/vim#161312cddf0e85a
Cherry-pick Test_cmdheight_not_changed() from patch 9.0.0187.
Co-authored-by: Milly <milly.ca@gmail.com>
Problem: ins_compl_add() has too many args
Solution: refactor it and use an int array instead of 2 separate int
args (glepnir)
closes: vim/vim#160625c66e23c62
Co-authored-by: glepnir <glephunter@gmail.com>
Problem: %! item not allowed for 'rulerformat'
(yatinlala)
Solution: also allow to use %! for rulerformat option
(Yegappan Lakshmanan)
fixes: vim/vim#16091closes: vim/vim#16118ac023e8baa
Co-authored-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Enabling private DEC mode 2031 tells the terminal to notify Nvim
whenever the OS theme changes (i.e. light mode to dark mode or vice
versa) or the terminal emulator's palette changes. When we receive one
of these notifications we query the terminal color's background color
again to see if it has changed and update the value of 'background' if
it has.
We only do this though if the user has not explicitly set the value of
'bg' themselves. The help text is updated slightly to hint to users that
they probably shouldn't set this value: on modern terminal emulators
Nvim is able to completely determine this automatically.
Problem: Possible unnecessary redraw after adding/deleting lines.
Solution: Check b_mod_set before using b_mod_xlines to avoid using stale
b_mod_xlines value (zeertzjq).
closes: vim/vim#161249f25a3a237
Problem: Hit double grid_line_start() assert when redrawing from
ext_messages msg_ruler event.
Solution: Do not start() batched grid calls when win_redr_ruler() will not
puts() anything.
Problem: The name `scope` is often used to refer to option flags because
`OPT_LOCAL` and `OPT_GLOBAL` are often used to determine the option
scope. This leads to the name `req_scope` being used for actual option
scopes instead.
Solution: Since the end-goal is to remove `OPT_LOCAL` and `OPT_GLOBAL`
entirely and replace them with `OptScope`, rename `OptScope` variables
to `scope` and the old scope flag variables to `opt_flags`.
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:
`validate_option_value_args()` returns `OK` even if option name is
invalid or if option doesn't have the supported scope, which leads to
Neovim still trying to erroneously get/set the option in those cases,
which can lead to an assertion failure when `option_has_scope()` is
invoked. This issue miraculously doesn't exist in release builds since
the assertion is skipped and `(get/set)_option_value_for` returns if
there is an error set, but that is not the intended location for that
error to be caught.
Solution:
Make `validate_option_value_args()` return `FAIL` if there is an error
set, which causes the API option functions to return early instead of
trying to get/set an invalid option.
The newer UCRT runtime has native support for UTF-8, including forcing
it as the active codepage even before `main()` is called. This means
the c runtime will properly convert windows WCHAR:s into UTF-8 bytes,
as early as the argv/argc params to `main()` . Whereas MSVCRT does not
support this reliably and required us to use `wmain()`.
Only MSVC supports using manifest files directly as source files.
The solution for other Windows toolchains is to use a .rc file.
