Features:
c7d84c5550 PR #4980 Full `:ruby` support!
c74ce334f2 PR #4624 timers: timer_start(), timer_stop()
b8e6f04e69 PR #5205 `:CheckHealth` command
47a15d0256 PR #4865 file: Add buffered reading and writing
*Much* faster shada file reading (important for startup time).
71b3e20d0f PR #4723 jobstart() learned 'rpc'
jobs and RPC channel IDs share the same "namespace".
jobstart() can starts RPC channels, which allows scripts to handle
'stderr' on a RPC channel, like a typical non-RPC job.
jobpid()/jobstop() work on RPC "jobs".
Deprecates rpcstart().
4dc4efc36f PR #4449 man.vim rewrite
`:Man` command is enabled by default.
New features: completion, window handling, better parsing, and more.
8a4e5b4bc2 PR #4697 capture() function (renamed to execute())
Supports nesting, including nested :redir.
ae6db26b09 PR #5050 'rplugin manifest: default to XDG dir'
a1682281f4 PR #5214 Restore ":browse oldfiles".
1f7304b846 Better handling of mouse-clicks on concealed chars.
5ea4d58a1b PR #5026 terminal: Ensure b:term_title always has a value
c002310787 tui: Assume 256 colors in most cases.
a2ecbc2cc0 PR #4929 Always resize the :terminal
a59330d6fc PR #4925 api_info()
a160590e40 PR #4813 allow setting cwd in jobstart(), termopen()
74f6460181 PR #4633: support "special" highlight (undercurl)
5a5ef1c222 PR #3450 mouse: Implement horizontal scroll.
Windows support:
All PRs now build on Appveyor targeting win32 and win64!
Numerous fixes!
Fixes:
e9061117a5 PR #4646 Prevent data loss for process output streams
7fa1baf44e PR #4798 'process.c: Fix block in teardown'
c10fe010f1 Prevent endless loop in printdigraph(). (#5215)
add41dca98 PR #5192 timers: Avoid crash after processing events
006f9c0c9c PR #5195 Set the default value for 'packpath'
6da7d6890c PR #5025 Restore double click
d622e9c416 readfile(): Less-disruptive readonly check.
Fixes an issue where nvim unnecessarily "touched" open files.
fe6ec75725 PR #4964 Handle very long $XDG_DATA_DIRS.
895f712df8 option: Do not expand options in XDG vars.
1d8a076157 server_init: Handle server_address_new() failure.
be531aba77 PR #5042 Fix v:register for clipboard=unnamed,unnamedplus
204f557a11 PR #4984 'Trigger TabNewEntered with <CTRL-W>T'
1e93e24f5e PR #4851 synIDattr(): Return RRGGBB value for `fg#`.
Changes:
acc5d08b37 PR #4690 'termguicolors' option enables "true color".
NVIM_TUI_ENABLE_TRUE_COLOR is now ignored.
:silent does not silence this message, even :redir does not consume it.
But execute() _does_ consume it, which interferes with the current
implementation of health.vim. It's prudent to avoid it in any case, even
if the implementation of health.vim changes in the future.
We can add this later if it is proven necessary, but it should not be
because:
1. User can run a subset of checkers via `:CheckHealth plugin1, ...,`
2. Healthcheck is a very rare operation. Optimizing it is not worth the
code/API complexity.
To healthcheck the "foo" plugin:
:CheckHealth foo
To healthcheck the "foo" and "bar" plugins:
:CheckHealth foo bar
To run all auto-discovered healthchecks:
:CheckHealth
- Use execute() instead of redir
- Fixed logic on suboptimal pyenv/virtualenv checks.
- Move system calls from strings to lists. Fixes#5218
- Add highlighting
- Automatically discover health checkers
- Add tests
Helped-by: Shougo Matsushita <Shougo.Matsu@gmail.com>
Helped-by: Tommy Allen <tommy@esdf.io>
Closes#4932
* sub_joining_lines: Optimization for :%s/\n//
* sub_grow_buf: Allocation of buffer to contain replacement text
* sub_parse_flags: Parse {flags} from :s command into subflags_T
Although this doesn't reduce do_sub's size enough to satisfy lint, it
covers the more straightforward pieces.
Problem: Equivalence classes are not properly tested.
Solution: Add tests for multi-byte and latin1. Fix an error. (Owen Leibman)
22e421549d
src/regexp.c changes weren't applied because they're specific to EBCDIC
handling, which has been dropped from nvim.
The latin1-specific tests were also removed since neovim intends to
remove the ability to have 'encoding' set to anything other than utf8.
Problem: getcompletion(.., 'dir') returns a match with trailing "*" when
there are no matches. (Chdiza)
Solution: Return an empty list when there are no matches. Add a trailing
slash to directories. (Yegappan Lakshmanan) Add tests for no
matches. (closesvim/vim#947)
b56195ed00
Reported in #4955, get_past_head() is supposed to return a pointer
after the head of the path (/ in UNIX, c:\ in Windows) but the windows
case was removed.
Removed the Mac reference in the comment, since there no special
handling for Mac.
vim-patch:0
- Weird tab+space combination used for alignment. All spaces now
- Added back <C-T> mapping (somehow we missed that completely)
- Fixed mistake that <Plug>(Man) opens in a new tab. Also added note at
top on how the window is chosen/opened.
- Clarified q local mapping
- Removed section that shows an example autocmd to add desired folding
style.
- Removed random line in `usr_12.txt` about `<Leader>` and backslash.
- :Man supports completion, not auto-completion.
Closes#5171
Calling printdiagraph() with msg_silent != 0 can result in an endless
loop because the loop condition never changes, if msg_col is never
changed.
To fix this, calculate the number of iterations before the loop, which
is always smaller than list_width.
In 3b12bb225a, ":oldfiles" was taught to
behave like Vim's ":browse oldfiles" if ":oldfiles!" was used. However,
this conflates the use of ! for abandoning a modified buffer with
choosing one file out of a list of oldfiles.
Now that ":browse" is supported again, ":browse oldfiles" will allow the
user to select an old file, while still complaining if that would cause
a modified buffer to be abandoned. ":browse oldfiles!" will just
abandon the buffer, as expected.
Without this the "cd scripts/.." might change to another dir (since
CDPATH is looked at before a local path), and then NEOVIM_SOURCE_DIR
might end up being "/somewhere/else\n/somewhere/else" (since the "cd"
prints the dir already in that case).
Closes https://github.com/neovim/neovim/pull/5213.
With -count, if the first argument is a number, it is made available
with <count>. Problem is, there is always a default count it is impossible
to tell whether the user set it.
Since v:count and v:count1 still work with 'keywordprg', -count is
unnecessary. But 'keywordprg' still calls ':Man' with a count prefixed.
So it must still accept a count in the line number position, but not consume
the first argument. This is done with -range.
Fixes#5202.
Temporary change to avoid frequent hangs on Travis macOS/OSX builds.
Hang does not occur on Quickbuild OSX (Yosemite) build.
Reverting e9061117a5 avoids the hang, but causes
more serious regressions on many more systems.
Note that the job_spec hang only happens with the gcc-4.9 Travis OSX build.
References #5002
References #5029
Build the default CMake target now that helptag generation is
working again.
For build artifacts create a zip file with an instalation of
Neovim (generated by cpack).
Per #2471, some path handling functions hardcode the UNIX path
separator '/' causing them to fail in Windows.
When BLACKSLASH_IN_FILENAME is set we may have to check against
psepc and psepcN instead of PATHSEP or use vim_ispathsep_nocolon().
As noted in “:help 'packpath'”, the default value is supposed to be the
same as that for 'runtimepath'. This was missed in the original port of
the packages functionality from Vim.
Closes#5193
Regression from #5168. Also changed the Man command's nargs to '+' so
that man#open_page does not need to handle 0 arguments, because that
will never occur.
- Use the default buffer text provided by before_each (avoids extra steps and
makes the tests more consistent with each other)
- Indent
- Adjust help doc
77135447e0 introduced:
if (!newfile) {
return FAIL;
}
which changed the semantics of the un-braced `else` in the
`#ifndef UNIX` block immediately above it.
This commit restores the semantics of Vim. Until now it mostly worked by
accident, but on Windows it would mean that opening a directory would
show "[Permission Denied]".
In the (!read_buffer && !read_stdin) case, always set `perm` for all
platforms. This also means we no longer need to set `perm` in the case
of (fd < 0) for non-Unix.
Wrap up uv_translate_sys_error and fallbacks into a new function
os_translate_sys_error(). In windows a copy of the original
uv_translate_sys_error() was imported from libuv.
Using /MT was causing issues when building luarocks, revert it, use the
dynammic runtime and generate release DLLs for the dependencies.
Some refactoring was required because for linking cmake looks for the
import libraries (.lib) but on runtime executables we need the .dll files
to be in the same folder.
The DLLs are placed in the bin/ folder in order for nvim.exe to run
during the build and tests. The install target installs the DLLs with
the nvim binary - uses GetPrerequisites to find runtime DLLs.
Some minor issues that required adjustments:
- [MSVC] FindMsgpack.cmake now looks for msgpack_import.lib instead of
msgpack.lib
- The lua-client fails to find libuv.lib, instead it looks for uv.lib,
added second copy of the file to the install command.
- [MSVC] CMAKE_BUILD_TYPE affects the output paths, default to Release.
Part of these changes are credited to @jasonwilliams200OK who fixed the
third-party recipes to consistently use the same build type.
- Smart autocomplete. It's automatically sorted, filtered for duplicates
and even formats the candidates based on what is needed. For example,
`:Man 1 printf<TAB>` will show the pages that are in section 1m as
'page(sect)' to let you know they are in a more specific section.
- Instead of trying to unset $MANPAGER we use the -P flag to set the
pager to cat
- Always use the section arg '-s', it makes the code much simpler
(see comment in s:man-args).
- A manpage name starting with '-' is invalid. It's fine for sections
because of the use of '-s'.
- The tagstack is an actual stack now, makes it much simpler.
- By using v:count and v:count1, the plugin can explicitly check whether
the user set a count, instead of relying on a default value (0) that
is actually a real manpage section.
- Extraction of a manpage reference is much more simple. No giant long
complicated regexes. Now, the plugin lets `man` handle the actual
validation. We merely extract the section and page. Syntax regexes are
a bit more specific though to prevent highlighting everything.
- Multilingual support in the syntax file. Removed the cruft that was only
relevent to vim. Also simplified and improved many of the regexes.
- Using shellescape when sending the page and sect as arguments
- In general, the code flow is much more obvious.
- man#get_page has been split up into smaller functions with explicit
responsibilties
- ':help' behavior in opening splits and manpages
- Comments explaining anything that needs explaining and isn't
immediately obvious.
- If a manpage has already been loaded but if it were to reloaded at the
current width which is the same as the width at which it was loaded at
previously, it is not reloaded.
- Use substitute to remove the backspaced instead of `col -b`, as the
latter doesn't work with other languages.
- Open paths to manpages
- It uses cWORD instead of cword to get the manpage under the cursor, this
helps with files that do not have (,) in iskeyword. It also means the
plugin does not set iskeyword locally anymore.
- <Plug>(Man) mapping for easy remapping
- Switched to single quotes wherever possible.
- Updated docs in $VIMRUNTIME/doc/filetype.txt (still need to update
user-manual)
- Always call tolower on section name. See comment in
s:extract_page_and_sect_fpage
- Formatting/consistency cleanup
- Automatically map q to ':q<CR>' when invoked as $MANPAGER
- It also fully supports being used as $MANPAGER. Setting the name and
stuff automatically.
- Split up the setlocals into multiple lines for easier readability
- Better detection of errors by redirecting stderr to /dev/null. If an
error occured, stdout will be empty.
- Functions return [sect, page] not [page, sect]. Makes more sense with
how man takes the arguments as sect and then page.
- Pretty prints errors on a single line.
- If no section is given, automatically finds the correct section for
the buffer name. It also gets the correct page. See the comment in
s:get_page
- If $MANWIDTH is not set, do not assign directly to $MANWIDTH because
then $MANWIDTH will always stay set to the same value as we only use
winwidth(0) when the global $MANWIDTH is empty. Instead we set it
locally for the command.
- Maintainer notes on all files.
Fixes#4162Fixes#4200Closes#4944
Regression by 4a138137f7. That commit mentions
a "possible race condition" but the cost isn't worth the (unexplained) gain.
Problem: The justify plugin has to be copied or sourced to be used.
Solution: Turn it into a package.
2946d0236d
Justification for changes that do not carry over
- `Filelist` does not exist in neovim
- neovim does not have a `README.txt` in `runtime/macros`
- Build for MinGW x86/x86_64. Move build scripts out of the yml file into
separate batch files.
- The MinGW builds use MSYS to get runtime dependencies, but they do not
link against the POSIX adaptation layer.
- For now only build the nvim.exe binary, but not the helptags.
Problem: "\%1l^#.*" does not match on a line starting with "#".
Solution: Do not clear the start-of-line flag. (Christian Brabandt)
7c29f38781
Helped-by: jamessan
Helped-by: mhinz
vim-patch:7.4.2011
Problem: It is not easy to get a list of command arguments.
Solution: Add getcompletion(). (Yegappan Lakshmanan)
aa4d73235b
vim-patch:7.4.2012
Problem: Test for getcompletion() does not pass on all systems.
Solution: Only test what is supported.
0d3e24be56
vim-patch:7.4.2066
Problem: getcompletion() not well tested.
Solution: Add more testing.
c1fb763184
Problem: The shellmenu plugin has to be copied or sourced to be used.
Solution: Turn it into a package.
fead3ac9a3
Justification for changes that do not carry over
- `Filelist` does not exist in neovim
- neovim does not have a `README.txt` in `runtime/macros`
Problem: The swapmous plugin has to be copied or sourced to be used.
Solution: Turn it into the swapmouse package.
e101204906
Justification for changes that do not carry over
- `Filelist` does not exist in neovim
- neovim does not have a `README.txt` in `runtime/macros`
vim-patch:7.4.2013
Problem: Using "noinsert" in 'completeopt' breaks redo.
Solution: Set compl_curr_match. (Shougo, closesvim/vim#874)
67081e5061
vim-patch:7.4.2014
Problem: Using "noinsert" in 'completeopt' does not insert match.
Solution: Set compl_enter_selects. (Shougo, closes#875)
32b808a4bd
Problem: Sticky type checking is more annoying than useful.
Solution: Remove the error for changing a variable type.
f6f32c38bf
Note: There are a bunch of other changes to eval.txt that I believe are
N/A and not related to this patch.
Otherwise, busted may run a different interpreter than the one we
detected, without the capabilities we expect. (For instance, we might
have detected the luajit interpreter, but busted might run the lua
interpreter, without the ffi module.)
Change implementation to compare sequences of bytes instead of C
strings.
The former implementation would return "false positives" in these cases:
"\0a" and "\0b": C strings are NUL-terminated, so it would compare two
empty strings.
"a" and "ab": If both strings aren't the same length, it would
compare only up to the length of the shorter one.
Fixes#5046.
Factor out a helper function to set b:term_title, and call it during
terminal initialization as well, to set the initial title to the term://
URL.
This makes it much easier to use b:term_title in a statusline, with just
`setlocal statusline=%{b:term_title}`, rather than needing an expression
to handle the unset case.
It is a wrong thing to do, this makes valid variable values be treated
incorrectly: in
XDG_DATA_HOME='/home/$foo/.local/share'
`$foo` should be treated literally and not expanded to `foo` environment
variable value.
Also makes option_expand not try to expand too long strings even if these too
long strings are default values. Previously it thought that default values
should always be expanded. Also does not try to expand NULL should it be the
default value just in case.
Fixes#4961
Problem: Recognizing <sid> does not work when the language is Turkish.
(Christian Brabandt)
Solution: Use MB_STNICMP() instead of STNICMP().
e266d6d664
Note: Added new test
Problem: On MS-Windows the package directory may be added at the end
because of forward/backward slash differences. (Matthew
Desjardins)
Solution: Ignore slash differences.
4c5717ed8a
Problem: When using packages an "after" directory cannot be used.
Solution: Add the "after" directory of the package to 'runtimepath' if it
exists.
a570244531
Updated runtime files.
03413f4416
Ignore changes to
* doc/Makefile, doc/help.txt: Related to Vim's version8 documentation
* doc/gui_x11.txt, doc/todo.txt, doc/vim.1, gvim.desktop, vim.desktop:
Irrelevant to Neovim
* doc/quickref.txt, doc/options.txt: As of yet unported 'emoji'
* doc/tags, syntax/vim.vim: Generated at build time
Problem: For plugins in packages, plugin authors need to take care of all
dependencies.
Solution: When loading "start" packages and for :packloadall, first add all
directories to 'runtimepath' before sourcing plugins.
49b2732644
Updated runtime files.
7db8f6f4f8
Ignore changes to
* doc/tags: Generated at build time
* doc/channel.txt, doc/todo.txt: Irrelevant to Neovim
* pack/dist/opt/matchit/doc/matchit.txt: matchit is enabled by default,
so description of how to enable isn't needed.
Problem: The matchit plugin needs to be copied to be used.
Solution: Put the matchit plugin in an optional package.
aedfcbe1e6
Ignore changes to
* Filelist, src/Makefile: Irrelevant to NeoVim
* runtime/vimrc_example.vim, runtime/macros/*, runtime/pack/*: matchit
is enabled by default in Neovim.
Updated runtime files.
4f3f668c84
Ignore changes to
* doc/channel.txt: Channel related docs
* doc/tags: Generated at build time
* doc/todo.txt: Irrelevant to Neovim
Updated runtime files.
77cdfd1038
Ignore changes to:
* doc/channel.txt, doc/eval.txt: Channel related docs
* doc/options.txt: GUI related docs
* doc/tags: Generated at build time
* doc/todo.txt: Irrelevant for Neovim
Update runtime files.
5f148ec0b5
Ignore changes to
* doc/channel.txt, doc/eval.txt: Channel related docs
* doc/tags: Generated at build time
* doc/todo.txt: Irrelevant to Neovim
Update runtime files.
328da0dcb7
Ignore changes to
* doc/channel.txt, doc/eval.txt: Channel related docs
* doc/tags: Generated at build time
* doc/todo.txt: Irrelevant to Neovim
Updated runtime files
dae8d21dd2
Ignore changes to
* doc/eval.txt: Channel related docs
* doc/help.txt, doc/index.txt, doc/os_390.txt: Removal of obsolete features, which
already happened in Neovim
* doc/tags: Generated at build time
* doc/todo.txt, doc/version5.txt: Irrelevant to Neovim
This commit doesn't change any behavior, only moves the init out of main.c We
_could_ move some initialization from tui.c:terminfo_start to an earlier phase,
in order to avoid mis-reporting 't_Co' during startup. But this will be messy,
and gains very little: TERM=linux works "good enough" as long as we correct t_Co
in tui.c:terminfo_start (c5b02d5a7).
Closes#3428
References #4999
The Linux "virtual consoles" available on Alt-F1...Alt-F7 (i.e.
tty1-tty7) support only 8 colors (actually, it's 16 colors when counted
together with "bold/bright" attribute) and 8 background colors (those in
some cases can be upped to 16 too, by using "blink" attribute - but this
might be more risky, in case some legacy consoles really show it as
blinking? I'm not sure about that.) This limit is buried deep in kernel
sources for default tty drivers. Trying to use the Neovim's default 256
colors in this case gives totally bad colors, breaking all color schemes
and sometimes rendering parts of the text invisible. A simple change
enables code paths for handling 8/16 colors, which are still present in
Neovim codebase.
A single line was deleted from `ex_drop()` in 1a91000 when fixing clint
warnings causing the `:drop` command to not work correctly if the buffer
is not already open in a window.
Fixes#4981
Only update some entries that are already in `version.c`. Mercilessly stolen from https://github.com/neovim/neovim/pull/4634. At least one possible contributor got confused by it not being as-up-do-date-as-it-coul-be(tm). We shouldn't have that.
1005, 1010: :smile
1039: small Build
1058, 1073, 1079, 1097: alloc
1555, 1556, 1573: Makefile
1560, 1579: channel
Also adds one exception to linter rules:
typedef struct {
kvec_t(Object) stack;
} EncodedData;
is completely valid (from the style guide point of view) code.
It appears that used msgpack library is not able to parse back message created
by msgpack_rpc_from_object() if nesting level is too high, so log_server_msg now
cares about msgpack_unpack_next() return value. Also error message from
server_notifications_spec.lua is not readable if something is wrong (though at
least now it does not crash when parsing deeply nested structures).
log_server_msg() in the test reports
[msgpack-rpc] nvim -> client(1) [error] "parse error"
This removes some stack overflows in new test regarding deeply nested variables.
Now in place of crashing vim_to_object/msgpack_rpc_from_object/etc it crashes
clear_tv with stack overflow.
This ought to prevent stack overflow, but I do not see this actually working:
*lua* code crashes with stack overflow when trying to deserialize msgpack from
Neovim, Neovim is fine even if nesting level is increased 100x (though test
becomes very slow); not sure how recursive function may survive this. So it
looks like there are currently only two positive effects:
1. NULL lists are returned as empty (#4596).
2. Functional tests are slightly more fast. Very slightly. Checked for Release
build for test/functional/eval tests because benchmarking of debug mode is
not very useful.
Actual value on FreeBSD is -31, UV_EMLINK was obtained from
/usr/include/asm-generic/errno-base.h (there EMLINK is defined as 31 there).
This may actually be something else, but I do not think so as “Too many links”
description also fits in. [Man page][1] agrees with me, search for `[EMLINK]`
([linux man page][2] also specifies ELOOP explicitly in a similar section).
[1]: https://www.freebsd.org/cgi/man.cgi?query=open&sektion=2
[2]: http://man7.org/linux/man-pages/man3/open.3p.html
Previously there was file descriptor leak, not detected by sanitizers. Now it is
file descriptor leak with a small memory leak which is detected by ASAN what
fails one of the tests (actually, “ShaDa support code leaves .tmp.z in-place
when there is error in original ShaDa and it has .tmp.a … .tmp.x”, but error is
reported at the next test because leaks are not detected until Neovim exit and
Neovim exit happens when clear()/reset() is called which happens in before_each
only).
One of the reasons is that O_RDONLY is zero, which makes checking whether file
is opened read- or write-only harder. It is not guaranteed that on other system
O_WRONLY will not be zero (e.g. because file can only be opened in read-write
mode).
Moves low-level functions handling to os/fs.c. Adds file.c with a proxy
interface.
Target: while leaving syscalls handling is os.c (partially handled by libuv),
add buffering for reading and writing to file.c.
add tests for synIDattr() with [fg|bg|sp]#
add tests for synIDattr and various #RGB colors
synIDattr: test for ui_rgb_attached()
test: fix tests for synIDattr fg/bg/sp
This was more trouble than it is worth:
- remove_directory fails if doc/ is not owned by the user
- some devs build in-tree, then deleting doc/ breaks the build
- `make install` isn't affected by the stale files at all: the tags are
built before install-time
So, reverting this change means only that devs who use a build/
directory will need to delete build/runtime/doc/ on the occasion that we
rename a doc file.
If a help file is renamed, stale help files in the build workspace will
cause duplicate tags (which causes the build to fail). To avoid this,
always delete build/runtime/doc/ before building helptags.
Tighten our jargon, avoid mentioning "msgpack" everywhere we mention the
RPC API. Prefer "RPC API" when speaking about the remote API.
Though it's conceivable that we may one day support some protocol other
than msgpack, that isn't relevant to most of our discussion of the API,
including this document.
The file name msgpack_rpc.txt is preserved to avoid totally breaking
URLs.
Problem: Can only get the directory of the current window.
Solution: Add window and tab arguments to getcwd() and haslocaldir().
(Thinca, Hirohito Higashi)
c970330676
menu_spec.lua yanks to the clipboard, but never pastes from it. This
can leave a child xsel process waiting around for something to paste the
content, causing the test process to hang.
Since the test isn't explicitly trying to exercise the clipboard, simply
use the default register.
https://neovim.io/develop/style-guide.xml#Horizontal_Whitespace
Note that this errors out for e.g.
if (has_mbyte) mb_copy_char((const char_u **)(&f), &t); \
(found in macros.h:151) or
#define ECMD_SET_HELP 0x02 /* set b_help flag of (new) buffer before
(found in ex_cmds.h:11) (note `(new) buffer`). I left this as-is because these
pieces of code are already caught by another rule (braces near if and `/*`-style
comments). Other false positives I found were:
1. `FUNC_ATTR_NONNULL_ARG(1) FUNC_ATTR_WARN_UNUSED_RESULT`: rejected by
requiring type name to start with `[a-zA-Z_]` (this makes `1` not be
incorrectly recognized as a type).
2. `#define FOO(var) (var)`: rejected by creating `cast_line`.
3. `(expr) * (expr)`: rejected by constructing regexp so that unary operators
require no spaces after them.
4. `int f(void) FUNC_ATTR_PURE`: rejected by requiring line to start with
a space.
5. `"." STR(NVIM_VERSION_PATCH) NVIM_VERSION_PRERELEASE`: not rejected, such
thing should be rare and it would be easy to get rid of the false positive by
e.g. breaking line.
Struct literals like `(MyStruct) { 4, 2 }` are not checked:
1. I am not sure whether they are under this rule at all (this is not a cast).
2. It would be hard to get rid of false positives (like the example with `if`
above, but now for *valid* `if() {}`s).
When skipping these test blocks they may error out:
Error → test/functional/shell/viml_system_spec.lua @ 154
system() with output containing NULs setup
./test/functional/helpers.lua:75: attempt to index upvalue 'session' (a nil value)
stack traceback:
./test/functional/helpers.lua:75: in function 'request'
./test/functional/helpers.lua:166: in function 'nvim_feed'
./test/functional/helpers.lua:195: in function 'feed'
test/functional/shell/viml_system_spec.lua:14: in function <test/functional/shell/viml_system_spec.lua:13>
Error → test/functional/shell/viml_system_spec.lua @ 155
system() with output containing NULs teardown
./test/functional/helpers.lua:75: attempt to index upvalue 'session' (a nil value)
stack traceback:
./test/functional/helpers.lua:75: in function 'eval'
test/functional/shell/viml_system_spec.lua:21: in function <test/functional/shell/viml_system_spec.lua:20>
It is otherwise impossible to determine which test failed sanitizer/valgrind
check. test/functional/helpers.lua module return was changed so that tests which
do not provide after_each function to get new check will automatically fail.
Problem: Autocommands triggered by quickfix cannot always get the current
title value.
Solution: Call qf_fill_buffer() later. (Christian Brabandt)
6920c72d4d
Helped by @mhinz
Updated runtime files.
e0fa3742ea
Ignore changes to
* doc/channel.txt: Channel related docs
* doc/netbeans.txt, doc/os_dos.txt, doc/todo.txt: Not relevant to Neovim
* doc/tags: Generated at build time
Update runtime files.
38a55639d6
Ignore changes to
* doc/channel.txt, doc/eval.txt, doc/various.txt: Channel related docs
* doc/tags: Generated at build time
* doc/todo.txt, doc/vi_diff.txt: Not relevant to neovim
Updated runtime files.
cbebd4879c
Ignore changes to
* doc/channel.txt, doc/eval.txt: Channel related docs
* doc/tags: Generated at build time
* doc/todo.txt: Not relevant to Neovim
* ftplugin/man.vim: Change already present in autoload/man.vim
Update runtime files.
681baaf4a4
Ignore changes to
* doc/channel.txt, doc/eval.txt, doc/usr_41.txt: Channel related docs
* doc/tags: Generated at build time
* doc/todo.txt: Not relevant to Neovim
Updated runtime files and translations.
5e9b2fa9bb
Ignore changes to
* doc/tags: generated at build time
* doc/develop.txt, doc/todo.txt, doc/netbeans.txt, doc/vim-ja.UTF-8.1,
doc/xxd-ja.UTF-8.1, lang/menu_*: Not applicable to Neovim
* doc/editing.txt: Crypt related
* doc/change.txt, doc/insert.txt, doc/various.txt: Removal of ex_extra
tags, which already happened in Neovim
* doc/vim-ja.UTF-8.1, doc/xxd-ja.UTF-8.1
This avoids the issue of nvim started daemons causing mountpoints to be
unmountable. This is currently the only place in runtime/ where this
calling convention occurred.
Signed-off-by: Aleksa Sarai <cyphar@cyphar.com>
Add both a test for cwd=/ and cwd=/tmp/nvim.XXXXX, to make sure that we
don't have regressions in cwd handling.
Signed-off-by: Aleksa Sarai <cyphar@cyphar.com>
Update a few runtime files.
705ada1aff
Ignored changes to
* doc/eval.txt, all json related documentation
* doc/if_mzsch.txt
* doc/tags, generated at build time
Update runtime files
13d5aeef56
Ignored changes to
* doc/develop.txt, since they were all in the "Coding Style"
section, which is completely different between Vim and Neovim.
* doc/tags, doc/todo.txt
* syntax/vim.vim, generated at build time
Processes in vim are always started in the current directory, which
causes issues when the process is a daemon and the current directory is
a mountpoint. Fix this by adding an option to set the cwd of the new
process with jobstart(). In addition, fix termopen() so that it actually
uses the cwd option from the dict (it couldn't previously set the cwd
value due to dead code).
Signed-off-by: Aleksa Sarai <cyphar@cyphar.com>
Specify that the ${GENERATED_HELP_TAGS} "command" (output) depends on
`helptags` so that it always regenerates the doc/ tags. (cmake "targets"
always run, whereas "commands" are contingent on their dependencies. But
we don't define doc/ dependencies because they are circular.)
Problem: 'langmap' applies to the first character typed in Select mode.
(David Watson)
Solution: Check for SELECTMODE. (Christian Brabandt, closes#572)
Add the 'x' flag to feedkeys().
25281634cd
Single point of truth (SPOT): Do not sprinkle "contributing" guidelines
across various resources; CONTRIBUTING.md is the authority.
DRY: Do not repeat guidelines in CONTRIBUTING.md which are covered by
ISSUE_TEMPLATE.md
Update help files.
85084ef1e9
Ignored changes to:
* doc/editing.txt related to encryption
* doc/eval.txt related to perleval()
* doc/if_mzsch.txt
* doc/tags
* doc/todo.txt
If a user specifies both {window} and {tab}, `getcwd()`/`haslocaldir()`
are using "tab" as the scope that should be reported. However, it
should be using "window" as the scope, within the specified tab page.
The initial implementation for `:tcd` always used `curtab` to find the
specified window. This would result in either inaccurate information or
an unexpected error (e.g., when there are more windows in the
user-specified tab page vs. the current tab page).
The directories table contains the names of the expected directory names
for varying scopes of the :cd tests. Using named indexes, instead of
numbered, makes the test more readable.
Build wcwd/tcwd and wlwd/tlwd on top of the reworked cwd/lwd functions.
This will allow for easier testing of `getcwd()`/`haslocaldir()` in
arbitrary windows and/or tab pages.
`event_teardown` is there from 974752c, by aktau. It was introduced with
`init_homedir` and `event_init`. Then both were removed by justinmk in
99a9161bac (`init_homedir`) and
49c5689f45 (`event_init`), but `event_teardown`
was not removed. Now this may cause a crash. More details in #4852.
Closes#4852
For string() it looks like no optimization, sometimes performance is even worse.
Since it was designed to avoid heap allocations in clear_tv let’s see whether it
will make any difference once clear_tv uses typval_encode to avoid stack
overflow in the disabled test.
Changes:
1. Linter finds no errors now.
2. Most of macros changed to `static inline … FUNC_ATTR_ALWAYS_INLINE` functions
(that was the purpose: they are easier to debug).
3. Queue is now not a pair of void* pointers, but a struct with two QUEUE
pointers, next and prev. This should not affect anything, except that _QUEUE
private macros can really be avoided without reducing readability and they do
not need any casts.
Current name is inappropriate for the following reasons:
1. It is often masked by local `loop` variables.
2. It cannot be searched for. There are many `loop` variables where `loop` is
some local variable. There are many cases when “loop” word is used in
a comment.
3. It is in any case bad idea to use a generic name as a name of the global
variable. Best if global has module prefix: this is why it is in `main.h`:
`main_loop` both stands for “a main loop” and “a loop defined in `main.*`”.
Since I have no idea how to list every occurrence of this variable method used
to rename it is “remove it from globals.h, try to compile, fix errors”. Thus if
some occurrence was hidden under false `#if` branch it was not replaced.
If a user has multiple remotes set for neovim/neovim, then
find_get_remote was returning 'remote1\nremote2\n', which breaks
anything trying to use it. Since we're just using this remote to fetch
from, any one will do.
This also checks the major/min version only for expected return codes.
With pyenv, you might get the following (return code 127):
pyenv: python3.4: command not found
The `python3.4' command exists in these Python versions:
3.4.3
3.4.3/envs/tmp-3.4.3-eElS6Y
tmp-3.4.3-eElS6Y
Compiling with macro -DEXITFREE opens a code path on which the event
loop is used after it was teared down, because not all close events
were processed yet.
nvim blocking can be tested with "nvim +te +'!xclip' +qa"
By closing all handles for a pty process, we unblock the event loop if
the process has not terminated yet.
The TUI can be enabled/disabled at build time with -DFEAT_TUI, default is ON for
UNIX, and OFF for non UNIX. When off, Neovim prints a message to stderr, along
with a list of the server endpoints.
Updated runtime files.
acb4f221c7
Ignored changes to:
* doc/eval.txt since alloc_fail isn't relevant for neovim
* doc/index.txt for unmerge :smile command
* doc/tags, syntax/vim.vim since they're generated at build time
* doc/todo.txt
Problem: Need several lines to verify a command produces an error.
Solution: Add assert_fails(). (suggested by Nikolay Pavlov)
Make the quickfix alloc test actually work.
a260b87d9d
Problem: No test for skipping over code that isn't evaluated.
Solution: Add a test with code that would fail when not skipped.
ea8c219ca8
Also add Test 91 from 7.4.1157 to test_viml.vim.
Problem: When a spell file has single letter compounding creating
suggestions takes an awful long time.
Solution: Add th eNOCOMPOUNDSUGS flag.
7b877b3605
NA runtime update (original commit message below) since Neovim doesn't
have runtime/spell/.
Update to newer English spell files,
but without the COMPOUND rules that make suggestions very slow.
4d1c0a4441
Problem: New style tests are executed in arbitrary order.
Solution: Sort the test function names. (Hirohito Higashi)
Fix the quickfix test that depended on the order.
cfc0a350a9
The following upstream patches that modified the original test have reached
master: 7.4.798, 7.4.818, 7.4.883, 7.4.977. The changes are incorporated into
the migrated test.
Also improve readability and some string delimiters.
Vim creates these scripts in test1 depending on what build features
are enabled so that tests that use these features are skiped if
necessary. Because Neovim only has one type of build (and the features
+eval, +windows, and +multi-byte are enabled in this build) they are
not necessary.
Problem: When there is a backslash in an option ":set -=" doesn't work.
Solution: Handle a backslash better. (Jacob Niehus) Add a new test, merge
in old test.
8f79acdf7e
Problem: Warning for unused var with tiny features. (Tony Mechelynck)
Solution: Add #ifdef. Use vim_snprintf(). Reduce number of statements.
c71982b239
By default Neovim searched a Luajit instalation and linked against
the luajit library.
In practice Neovim only requires luajit to run the unit tests. All other
targets only require lua and the correct lua modules. This commit:
1. Remove the strict dependency on Luajit
2. Makes the unittest target depend on the lua 'ffi' module.
If the module is not available the target is not enabled
and a message is displayed.
Problem: Only terminals with up to 256 colors work properly.
Solution: Use the 256 color behavior for all terminals with 256 or more
colors. (Robert de Bath)
fa03fd6c4a
It's acceptable for “git describe --tags --exact-match …” to fail, since
all runtime updates commits are untagged. All that matters is that we
get a tag when one exists.
Therefore, ignore the failure status of the git describe call, relying
on the captured output instead.
patch 7.4.1728
Problem: The help for functions require a space after the "(".
Solution: Make CTRL-] on a function name ignore the arguments. (Hirohito
Higashi)
81edd171a9
...and small refactorings:
- Style changes
- Variable renames
- Changes in variable scope
This change won't bring new problems (undefined behavior) as `memcpy` is
already being used in the function.
The only data loss should be, if a process forked a child that keeps
sending data after the parent terminated.
While not in teardown mode we could keep reading child data, but then
`:!cmd` would block after `cmd` exited. In teardown mode we want to exit
nvim so we cannot keep reading child data.
The whole stream buffer is now put on screen at once instead of only
data up to the last newline. This has some advantages:
* RBuffer cannot wrap around, so we never forget to output second
half of the buffer.
* Stream data is not delayed anymore, because we don't have to wait for
a newline.
This works by remembering the last used screen column.
Updated runtime files.
fa7353428f
Missing in runtime/doc: if_tcl.txt, tags, todo.txt, version5.txt. Some
other changes related to binary nrformats were already merged.
Problem: Using "q!" when there is a modified hidden buffer does not unload
the current buffer, resulting in the need to abandon it again.
Solution: When using "q!" unload the current buffer when needed. (Yasuhiro
Matsumoto, Hirohito Higashi)
027387f70c
This was not a problem locally, but would often/sometimes/etc. (YMMV) fail on QB
and/or travis. This seems to fix it. Quoting @justinmk: "I have a feeling this
is just a bug in the bracketed paste special-cases in the existing code".
This was noticed due to a user issue (#4750) when building Neovim 0.1.4
via ports. The crux of the issue is that we did not detect the
be64toh() macro, because there is no endian.h on FreeBSD (along with
several other BSDs). So we were defaulting to our builtin version of
be64toh(). However, it appears that sys/endian.h was being picked up by
an include (likely msgpack.h) and so be64toh() was actually defined and
corrupting our definition of it.
So the answer here was to use the correct include file in our check, and
export that information in the config.h. Then we use that information
to include the right header in shada.c.
This fixes#4750.
Problem: ":argadd" without argument is supposed to add the current buffer
name to the arglist.
Solution: Make it work as documented. (Coot, closesvim/vim#577)
2faa29f896
Problem: Using ":argadd" when there are no arguments results in the second
argument to be the current one. (Yegappan Lakshmanan)
Solution: Correct the w_arg_idx value.
a24f0a550f
There are a total of 5 shell scripts in the Neovim source tree.
All but runtime\macros\less.sh had warnings/errors when run through
Shellcheck (http://www.shellcheck.net/).
This commit fixes all warnings/errors and also changes the shebang to
"#!/bin/sh" when possible (this was not possible for vim-patch.sh
because it uses many bashisms).
The shellcheck errors that were fixed are:
SC2068: Double quote array expansions to avoid re-splitting elements.
SC2086: Double quote to prevent globbing and word splitting.
SC2124: Assigning an array to a string! Assign as array, or use *
instead of @ to concatenate
SC2155: Declare and assign separately to avoid masking return values.
The character class %s also matches a newline in lua, that's not really what we
want here. It works in the other cases in this function, so I left them, but
the final gsub should preserve newlines.
TODO: Only works at startup (i.e., in the user's init.vim/vimrc/--cmd),
but it should probably work at any time.
---
patch 7.4.1799
Problem: 'guicolors' is a confusing option name.
Solution: Use 'termguicolors' instead. (Hirohito Higashi)
61be73bb0f
patch 7.4.1806
Problem: 'termguicolors' option missing from the options window.
Solution: Add the entry.
8e3d1b6326
patch 7.4.1808
Problem: Using wrong feature name to check for 'termguicolors'.
Solution: Use the right feature name. (Ken Takata)
8a24b794b8
patch 7.4.1809
Problem: Using wrong short option name for 'termguicolors'.
Solution: Use the option name.
868cfc19bb
Problem: Cannot measure elapsed time.
Solution: Add reltimefloat().
79c2c881bb
Applied manually. None of the ifdef's applies anymore, and proftime_T was
changed into an uint64_T, so the function profile_float to convert proftime_T to
float is not needed in nvim.
Problem: When a symbolic link points to a file in the root directory, the
swapfile is not correct.
Solution: Do not try getting the full name of a file in the root directory.
(Milly, closesvim/vim#501)
e3303cb081
This was already fixed in Neovim by c708061.
Problem: When making a change while need_wait_return is set there is a two
second delay.
Solution: Do not assume the ATTENTION prompt was given when need_wait_return
was set already.
b01f357791
Problem: Editing a URL, which netrw should handle, doesn't work.
Solution: Avoid changing slashes to backslashes. (Yasuhiro Matsumoto)
b4f6a46b01
Cherry-picked from https://github.com/neovim/neovim/pull/810, rebased.
Problem: g-CTRL-G shows the word count, but there is no way to get the word
count in a script.
Solution: Add the wordcount() function. (Christian Brabandt)
ed767a2073
Problem: The column is not restored properly when the matchparen plugin is
used in Insert mode and the cursor is after the end of the line.
Solution: Set the curswant flag. (Christian Brabandt). Also fix
highlighting the match of the character before the cursor.
c21d67e33c
vim-patch:7.4.1562 is N/A because do_helptags(..) no longer has a dirname
argument and dirname is explicitly allocated in the method body, so it must be
freed.
Helped-by: oni-link
It's possible that the first test encounters a temp directory with files
in it, due to a previous test causing the first test to fail. Instead,
let's clean up before and after the test to make sure the temp area is
pristine before and after the test.
While trying to debug an issue, I discovered that the tests for illegal
arguments depended on the prior suite having run and started a session.
Let's remove that unintentional dependency by starting our own session
before each test.
Could not find when it was fixed in the vim source, it originates in 7.001, and
is fixed as of today, but my git-fu did not discern a relevant patch. I don't
think it matters much.
Updated runtime files.
256972a984
Missing files in runtime/doc: todo.txt, tags. Patch to runtime/doc/syntax.txt
was applied manually in part, for no discernible reason.
Update runtime files.
d042dc825c
Missing in runtime/doc: hangulin.txt, tags, todo.txt. The changes to options.txt
do not apply for nvim. man.vim is very different in nvim, some changes applied
manually, others discarded.
Update runtime files.
e392eb41f8
Files runtime/doc/tags and runtime/doc/todo.txt did not exist. Ignored
runtime/syntax/vim.vim. One change in runtime/doc/windows.txt had already been
applied.
Updated runtime files.
b4ff518d95
Missing files: runtime/doc/tags, runtime/doc/todo.txt. Changes to
runtime/doc/if_pyth.txt, runtime/doc/options.txt and runtime/doc/quickref.txt
did not aply. Excluded runtime/syntax/vim.vim.
Update runtime files.
a0f849ee40
Missing files runtime/doc/tags and runtime/doc/todo.txt. Excluded
runtime/syntax/vim.vim, since we diverged quite a bit from vim in this file.
Problem: When completing a shell command, directories in the current
directory are not listed.
Solution: When "." is not in $PATH also look in the current directory for
directories.
b5971141df
Most of it applied manually.
If libintl is not available, LC_MESSAGES is not defined. For now fallback to
using LC_CTYPE.
Neovim and Vim have diverged significantly in ex_cmds2.c concerning this logic.
In other locations the fallback is actually LC_COLLATE, but in this case Vim
calls get_mess_env() (which in turn falls back to LC_CTYPE).
In Neovim get_mess_env() is only available with libint. This means we are not
completely consistent with Vim when handling LC_ environment variables and do
not build against libintl.
Problem: Triggering OptionSet from setwinvar() isn't tested.
Solution: Add a test. (Christian Brabandt)
74b738d414
Mark 7.4.889 as merged
Done in
9bd8fcde1e
Problem: When leaving the command line window with CTRL-C while a
completion menu is displayed the menu isn't removed.
Solution: Force a screen update. (Hirohito Higashi)
5f1fea28f5
Applied manually.
This change effectively disables history for lines inserted using this method.
Not a big problem since it does not work for them in Vim in first place.
Also solves a bug(?): ex_window() run while in :append mode opens search history
in Vim for some reason. Now it opens empty cmdline window.
When backupcopy=auto buf_write assumes backupcopy=yes when the file is a
hard/symbolic link. However this check was guarded by a UNIX ifdef. The
check itself is portable and the guard can be removed.
Added a couple tests to check the behaviour of bkc=auto and bkc=no
with a symbolic link.
Reported in #4525
Problem: When "syntax manual" was used switching between buffers removes
the highlighting.
Solution: Set the syntax option without changing the value. (Anton
Lindqvist)
885f24fbca
Patch applied cleanly to the nvim sources except for version.c.
Problem: Having 'autochdir' set during startup and using diff mode doesn't
work. (Axel Bender)
Solution: Don't use 'autochdir' while still starting up. (Christian
Brabandt)
6bd364e084
The Vim function `haslocaldir()` would crash if the users called it with
the two arguments `-1, -1`. Now it returns `0` in that case.
The coverity issue was complaining about a NULL dereference, but there
can never be a case where the pointer `tp` is NULL and being
dereferenced. An assertion has been put in place to satisfy coverity.
Furthermore the functions themselves have been cleaned up. First of all
the documentation comment for the different scopes has been extended and
a macro for the minimum scope has been introduced. In both functions any
time a scope is used as a range (e.g. in a loop) macros instead of
actuals scopes are used, that makes the functions more robust if new
scopes are added.
Second, in the implementation of `getcwd()` there was a superfluous
loop, it has been removed completely. I also changed all `goto end` to
plaing `return` statements by moving the allocation of `cwd` down, that
way there is no need for `goto` anymore.
Problem: With 'rightleft' and concealing the cursor may move to the wrong
position.
Solution: Compute the column differently when 'rightleft' is set. (Hirohito
Higashi)
e39b3d9fb4
Problem: The fix in patch 7.3.192 is not tested.
Solution: Add a test, one for each regexp engine. (Elias Diem)
96c664af27
The patch was applied to 044_099_regexp_multibyte_magic_spec.lua as
these two legacy tests (44 and 99) were merged together (and
de-duplicated) in Neovim.
test_cursor_func.vim hangs at the call to
disable_char_avail_for_testing(). The test does not actually need this
function (and it correctly fails if the fix from 7.4.1300 is reverted).
Given that disable_char_avail_for_testing is a gigantic hack, if we can
avoid it let's do so.
Problem: An Ex range gets adjusted for folded lines even when the range is
not using line numbers.
Solution: Only adjust line numbers for folding. (Christian Brabandt)
a3306958dc
7.4.797:
- Modified function redraw_asap was removed in
e0e41b30c6,
together with its only caller check_termcode.
- Grepping for the following regexps did not yield any result in
nvim/src and its subdirectories: rows.*\*.*Columns (lines modified
often), ScreenLinesC\[r\] (one of the sources of the bug),
msg_scrolled followed by grepping for NORMAL (another line that was
changed for the bug)
- The out-of-bound access was on the array *screenlineC[MAX_MCO] (see
the diff). I grepped for MAX_MCO an checked every array of that length
for out-of-bounds access in its scope (I did not check for called
functions, e.g. utfc_ptr2char(p, u8cc) where u8cc was of length
MAX_MCO). I did not find any.
- The code for drawing on the screen was moved to the TUI, which was
newly written.
7.4.733:
- test_listchars was converted to a lua test in
https://github.com/neovim/neovim/pull/2492
- Since no file is sourced anymore, the behavior of the lua test should
not depend on the setting of ff
- Code from Vim source.
- Removed the check for 'guioptions'
- mouse_spec.lua: test <ScrollWheelLeft> and <ScrollWheelRight>
- Move horizontal scroll logic to mouse.c
- Remove 'gui_' from the function names
- Renamed variables to be more specific (as opposed to generic p, w).
- Marked some functions as `static`
Problem: When there is any error Vim will use a non-zero exit code.
Solution: When using ":silent!" do not set the exit code. (Yasuhiro
Matsumoto)
8b778d5599
Problem: More problems reported by coverity.
Solution: Avoid the warnings. (Christian Brabandt)
cde8854730
Applied manually. Files that do not exst anymore: gui.c gui_w16.c gui_w32.c
if_xcmdsrv.c os_unix.c
Problem: Vim leaks memory, when 'wildignore' filters out all matches.
Solution: Free the files array when it becomes empty.
7b256fe744
The only nontrivial part of 7.4.871 missing (renamings of variables are in
another commit; freeing *files after 0 matches was already there, just FAIL was
not returned in that case)
New ex commands: 'tcd', 'tchdir'
Changed Vimscript functions: 'haslocaldir', 'getcwd'
The ex-commands ':tcd' and ':tchdir' are the tab-local equivalents of
':lcd' and ':lchdir'. There are no new Vimscript functions introduced,
instead the functions 'haslocaldir' and 'getcwd' take in optional
arguments. See the documentation for details
Since there is now different levels of local directory a simple boolean
at source level is no longer sufficient; a new enumeration type is used
for the scope-level from now on.
The documentation has been accommodated for these new commands and
functional tests have been written to test the feature.
Problem: Vim can create a directory but not delete it.
Solution: Add an argument to delete() to make it possible to delete a
directory, also recursively.
da440d21a6
Problem: Beeping when running the tests.
Solution: Fix 41 beeps. (Roland Eggner)
901e58c243
Applied cleanly, except for files test29.in, test4.in, test61.in, test82.in,
test83.in, test90.in, test95.in, which were all converted to lua tests, and
version.c.
Some beeps remain, but this is much better.
vim-patch:fc39ecf (committed in aae7e8b)
vim-patch:541f92d (committed in cc52060)
vim-patch:88774fd (committed in 896f548)
This commit message forces vim-patch.sh to recognize these patches
which were committed with only 6 chars (vim-patch.sh requires 7).
vim-patch:35e7594
35e7594dd4
"Add missing test files from 7.4.634 to the repository."
The discrepancy between the expected getpos() result of the old test
[0, 15, 2, 0] and the converted test [0, 3, 2, 0] is just a matter of how
the buffer is constructed: in the old Vim test the buffer has a bunch of
junk at the top.
The central purpose of the test is to verify that the getpost("'a") does
*not* return [0, 0, 0, 0].
Adds two undocumented v: variables: _null_list and _null_dict because I do not
know a reproducible way to get such lists (though I think I heard about this)
and dictionaries (do not remember hearing about them). NULL strings are obtained
using $XXX_UNEXISTENT_VAR_XXX.
Fixes crash in json_encode($XXX_UNEXISTENT_VAR_XXX). Other added tests worked
fine before this commit.
Also removes a note regarding the performance of `empty(long_list)` vs
`len(long_list) == 0` because this has nothing to do with the actual state:
first checks that list is not NULL and pointer to its first element is also not
NULL, second gets length by comparing list with NULL and falls back to
`tv->vval.v_list->lv_len` if not. `len(long_list)` *may* still be *slightly*
slower, but the slow down has nothing to do with the length of the list, is
hardly noticeable and depends on how good compiler is at inlining and what
exactly have author of the plugin written (I mean `len(long_list) == 0` vs
`empty(long_list)` vs `!len(long_list)`).
Compiler used by one VM in QuickBuild has found a number of false positives.
Everything is fine on travis.
List of failures:
From [QuickBuild][1], build [7429][2]:
14:38:19,945 WARN - /home/quickbuild/buildagent/workspace/root/neovim/pull-requests-automated/src/nvim/eval.c: In function ‘assert_bool’:
14:38:19,945 WARN - /home/quickbuild/buildagent/workspace/root/neovim/pull-requests-automated/src/nvim/eval.c:7551:40: error: comparison between signed and unsigned integer expressions [-Werror=sign-compare]
14:38:20,058 WARN - cc1: all warnings being treated as errors
. This is not making much sense (7551:40 is `!=` in `{SpecialVarValue} !=
({bool}?{SpecialVarValue}:{SpecialVarValue})`), but this error is present.
---
Also fail from [build][3] [4930][4]:
15:47:00,853 WARN - /home/quickbuild/buildagent/workspace/root/neovim/pull-requests-automated/src/nvim/eval/encode.c: In function ‘encode_read_from_list’:
15:47:00,853 WARN - /home/quickbuild/buildagent/workspace/root/neovim/pull-requests-automated/src/nvim/eval/encode.c:258:30: error: conversion to ‘char’ from ‘int’ may alter its value [-Werror=conversion]
, pointing to `:` in `{char} = ({char} == {const} ? {const} : {char})` where
`{const}` is character constant like `'\n'`. I have no idea where exactly it saw
conversion, so simply casted everything to (char).
---
[Build][5] error:
08:32:03,472 WARN - /home/quickbuild/buildagent/workspace/root/neovim/pull-requests-automated/src/nvim/eval.c: In function ‘tv_equal’:
08:32:03,472 WARN - /home/quickbuild/buildagent/workspace/root/neovim/pull-requests-automated/src/nvim/eval.c:5077:1: error: control reaches end of non-void function [-Werror=return-type]
---
Build [4949][7]:
11:28:00,578 WARN - /home/quickbuild/buildagent/workspace/root/neovim/pull-requests-automated/src/nvim/eval.c: In function ‘f_type’:
11:28:00,578 WARN - /home/quickbuild/buildagent/workspace/root/neovim/pull-requests-automated/src/nvim/eval.c:16085:24: error: ‘n’ may be used uninitialized in this function [-Werror=uninitialized]
11:28:00,581 WARN - /home/quickbuild/buildagent/workspace/root/neovim/pull-requests-automated/src/nvim/eval.c: In function ‘f_empty’:
11:28:00,581 WARN - /home/quickbuild/buildagent/workspace/root/neovim/pull-requests-automated/src/nvim/eval.c:8505:24: error: ‘n’ may be used uninitialized in this function [-Werror=uninitialized]
[1]: http://neovim-qb.szakmeister.net/wicket/page?5-1.ILinkListener-content-buildTab-panel-masterStep-body-children-0-step-body-children-2-body-children-3-step-body-children-0-step-body-children-0-step-head-logLink
[2]: http://neovim-qb.szakmeister.net/build/4929
[3]: http://neovim-qb.szakmeister.net/build/4930
[4]: http://neovim-qb.szakmeister.net/wicket/page?1-1.ILinkListener-content-buildTab-panel-masterStep-body-children-0-step-body-children-1-body-children-3-step-body-children-0-step-body-children-0-step-head-logLink
[5]: http://neovim-qb.szakmeister.net/build/4948/step_status
[7]: http://neovim-qb.szakmeister.net/build/4949
Note: currently they are both *dumped*, but parsing them produces an error. This
is inappropriate: variables should either be skipped with error message when
dumping or should be read back properly.
It also appears that I did not have test for “has wrong variable value type”
error, so nothing got removed from errors_spec.
Special dictionaries representing map are created when encountering duplicate
key or when key is empty or contains NUL.
Also checks that values are separated by a comma/colon properly.
Reasons:
- One does not have to do `s[len] = NUL` to work with these functions if they do
not need to replace the whole string: thus `s` may be const.
- One does not have to save/restore p_cpo to work with them.
A wrong format specifier and not enough allocated memory for
sourcing_name could lead to a heap corruption.
Original patch by Rui Abreu Ferreira (@equalsraf)
Fixes#4582
Having a tags file in the calling directory of make test would make this test
fail, so disable tag file completion for it. Disable all other options except the
current buffer, to, applying the principle of least surprise.
Problem: test_viml isn't run.
Solution: Include change in makefile.
f49e240c2d
Applied cleanly except for version.c. Needs to be added to the build system.
Problem: Test 49 is slow and doesn't work on MS-Windows.
Solution: Start moving parts of test 49 to test_viml.
c06624661a
Applied cleanly except Makefile changes. File src/nvim/testdir/runtest.vim was
configured to set nocp for all tests. This option is not in nvim anymore, so it
was removed.
Problem: Crash when evaluating the pattern of ":catch" causes an error.
(Dominique Pelle)
Solution: Block error messages at this point.
768ce2435a
Applied manually. Could not reproduce the crash both using vim and nvim,
therefore could not device a test.
Should be merged anyways, since it's in vim?
To get an UTF-8 character, utf_ptr2char() is used.
But this function can read more than maxlen bytes, if an incomplete
byte sequence is used(first byte specifies a length > maxlen).
Asterisk should be preceded by a space except the following cases:
1. `foo[*idx_ptr]`
2. `++*foo` (though I would prefer to write this as `++(*foo)`)
3. `(*foo)->bar`, `if (*p)`
4. `char **foo`
5. `/* comment */`
6. `!*p`
Note: `[^ (*/![]` is followed by `+` to make error position easier to find. It
is not needed to find location of the problem.
Luv is a simple lua binding to libuv, which is now used by neovim lua client.
The bundled luv installation a bit different from other dependencies in that it
is installed two times:
- The "BuildLuv.cmake" script downloads and installs a static version of luv
using its normal cmake build script. This static version will be used later.
- Luv default rockspec is replaced with the alternate under the "rockspecs"
directory(the alternate rockspec plays nicer with neovim build system)
- The alternate rockspec is used to build/install the lua module and make it
available to lua scripts.
In #3027 we deferred the "missing spell file" prompt until VimEnter, but
we were sending the wrong argument (should be "en", not "en_us").
This also caused redundant prompts if user answers "No" to the download
prompt invoked by the SpellFileMissing event.
Closes#3966Closes#4406
- Always auto-create spell/ directory, don't ask.
- Don't ask where to put .spl file if only 1 choice exists.
- Always download .sug file, don't ask.
- Use blackhole register for :delete and :g//d.
- Formatting: expand tabs.
Make get_vim_sources fetch the whole repository (it's not THAT big) so we can
pick up all the patches. The regexp didn't pick up the NA patches if there was a
comma after NA, so I extended it (I allowed arbitray things after NA, so maybe
someone can write a comment or so, should not lead to confusion).
-1 is index past the end, and -2 is the index of the last element.
This eliminates the need for include_start/include_end.
Allow the handling of out-of-bounds to be configurable.
- Create `mf_free_fnames()` that frees and nullifies `mf_[f]fname`
- Create `mf_set_fnames()` to set the `mf_fname` and the `mf_ffname`
altoghether
- Have `mf_do_open` return a bool to indicate success so that calles
don't have to check `memfile_T::mf_fd` (file descriptor)
- Inline `mf_write_block`
This is needed as long as we support cmake older than 2.8.12. When we
bump the minimum version to 2.8.12, we can use target_compile_options()
and add_compile_options() instead.
Closes#4389
-NaN doesn't exist in the IEEE 754 spec, it is a hardware-specific detail
abstracted away by luajit(and not by lua or nvim), so there's no need to test
it. Normalize all tests that involve -nan so the suite will be compatible with
both Lua and Luajit.
The hexadecimal notation is a Luajit extension which is not compatible with Lua
5.1. While Lua 5.2 does support hexadecimal sequences, it is better to target
Lua 5.1 for maximum compatibility with Luajit(which has fully compatible with
5.1 API/ABI).
Parameter should_free, indicating that the caller has to free the
returned pointer from get_exception_string(), is not set to true if
type == ET_ERROR.
Problem: When pasting on the command line line breaks result in literal
<CR> characters. This makes pasting a long file name difficult.
Solution: Skip the characters.
6f62fed349
- Define a collection of legal characters when parsing page and section
in `s:parse_page_and_section()` instead of relying on 'iskeyword',
which is unreliable.
- Allow non-numeric section names (e.g., `3c`).
- Simplify argument handling in `man#get_page()` to accommodate
non-numeric section names.
Fixes#4165.
New feature: `VTermState->mode.bracketpaste`
Enabled by default, but note that `vterm_state_reset()` disables it.
03981def6b
References #3476
New feature: `vterm_state_set_unrecognised_fallbacks`
acf7f19713
Oh, and terminal reflow works now.
Closes#2514 (but not #3864, that's a bit more tricky)
The previous defaults were including the nvim suffix, causing it to
apear twice in the final paths.
kXDGDataHome and kXDGConfigHome are now set as %LOCALAPPDATA%,
kXDGCacheHome is set as $TEMP.
In Windows there is no distinction between configuration and data
storage, but we don't want to place all files under the same path.
get_xdg_home() now appends a different path suffix for kXDGDataHome.
- Configuration files are saved under %LOCALAPPDATA%\nvim
- Data files are saved under %LOCALAPPDATA%\nvim-data
Problem: Using searchpair() with a skip expression that uses syntax
highlighting sometimes doesn't work. (David Fishburn)
Solution: Reset next_match_idx. (Christian Brabandt)
6773a348da
Prior to this change, type errors were silently ignored. They're explicit now.
setqflist(list, action, title)
setloclist(win, list, action, title)
"list" (required) must be a list.
"action" (optional) must a string.
"title" (optional) must a string or number that gets converted to a string.
An error is thrown otherwise.
Problem: Typo in test goes unnoticed.
Solution: Fix the typo. Give error for wrong arguments to cursor().
(partly by Hirohito Higashi) Add a test for cursor().
24c4d539ee
Problem: searchpos() always starts searching in the first column, which is
not what some people expect. (Brett Stahlman)
Solution: Add the 'z' flag: start at the specified column.
ad4d8a192a
Problem: When using slices there is a mixup of variable name and namespace.
Solution: Recognize variables that can't be a namespace. (Hirohito Higashi)
9bbf63dbf8
All syntastic users experienced this problem:
E685: Internal error: get_tv_string_buf()
It's reproducable with:
:call setloclist(0, [''])
So, not given optional arguments to setloclist() lead to some fields not
inizilied and the code took the wrong branches.
Add an extra argument to these functions to set w:quickfix_title.
This is a modified version of a patch from vim_dev. Discussion here:
https://groups.google.com/forum/#!topic/vim_dev/X7VVPd4Do5s
Credits go to Christian "chrisbra" Brabandt and Daniel "blueyed" Hahler.
After 87a49405b0, terminal_open() is not
nested by default. The default "term://" handler depended on that, but
it should instead explicitly raise TermOpen.
References #4306
Change POROJECT_NAME to 'nvim', and use it as the gettext
domain name. The *.mo files, previously installed as
$runtime/lang/xx/LC_MESSAGES/nvim.mo, are now installed as
$prefix/locale/xx/LC_MESSAGES/nvim.mo.
Adds support for:
- api:vim_input("<D-a>")
- ":nnoremap <C-D-S-...>" and permutations thereof
UIs must capture the modifier and send it as "<D-...>" to vim_input().
Note: Before this commit, any arbitrary ":nnoremap <{foo}-{bar}>"
mapping could already be invoked with feedkeys("\<{foo}-{bar}>"). This
commit supports "D-" as a modifier that can be combined with "C-", "A-",
"S-" in any order.
For non-GUI (terminal) support, user must:
:set <D-a>={CSI sequence}
then send the {CSI sequence} from their terminal. But this does not work
yet (regression #2204).
Closes#2190
Problem: Crash when dragging with the mouse.
Solution: Add safety check for NULL pointer. Check mouse position for valid
value. (Hirohito Higashi)
294a7e55b0
---
see: "Crash while mouse-selecting in two-buffer mode"
https://github.com/vim/vim/issues/486Fix#3704
Problem: MS-Windows: When 'encoding' differs from the current code page,
expandinig wildcards may cause illegal memory access.
Solution: Allocate a longer buffer. (Ken Takata)
7314efd87d
Problem: $HOME in `=$HOME . '/.vimrc'` is expanded too early.
Solution: Skip over `=expr` when expanding environment names.
be83b73ddb
---
To reproduce:
```sh
nvim -u NONE -c 'e `=$HOME . "/.vimrc"`'
```
Problem: When expanding `=expr` on the command line and encountering an
error, the command is executed anyway.
Solution: Bail out when an error is detected.
3f188935ec
<abuf> from the TermClose event now returns the correct buffer number.
Prior to this change it would always return the buffer number of the current
buffer, which is obviously wrong in an async environment.
Problem: May get into an invalid state when using getchar() in an
expression mapping.
Solution: Anticipate mod_mask to change. (idea by Yukihiro Nakadaira)
2455c4ede8
Problem: User may yank or put using the register being recorded in.
Solution: Add the recording register in the message. (Christian Brabandt,
closesvim/vim#470)
a0ed84a268
* Calling "vim-patch.sh -p" on a checked-out branch already created with
"-p" will re-use the branch and append commits.
* Fetch upstream/master before checking out branch on first call of "-p".
* Reverted creation of commit in submit step ("-s") to previous behavior:
Create an empty commit with correct commit message when "-p" is called.
* Submitting a pull request with "-s" will create a correct pull request
message even if multiple patches have been ported in one single branch
with "-p".
When calling "vim-patch.sh -s" on a checked-out branch created with
"vim-patch.sh -p", create commit from staged changes, push to origin,
create pull request (using hub), and clean up patch files.
Problem: test_utf8 fails on MS-Windows when executed with gvim.
Solution: Use the insert flag on feedkeys() to put the string before the
":" that was already read when checking for available chars.
6040256d8b
---
feedkeys() is not used by legacy test utf8.
Problem: Custom command line completion does not work for a command
containing digits.
Solution: Skip over the digits. (suggested by Yasuhiro Matsumoto)
23d1b62746
Problem: Completing the longest match doesn't work properly with multi-byte
characters.
Solution: When using multi-byte characters use another way to find the
longest match. (Hirohito Higashi)
4f8fa1633c
Problem: When a BufLeave autocommand changes folding in a way it syncs
undo, undo can be corrupted.
Solution: Prevent undo sync. (Jacob Niehus)
e7d1376b63
---
see: "[bug] [patch] Setting foldmethod in WinLeave autocommand can corrupt undo state"
https://groups.google.com/d/msg/vim_dev/xF5uMLb1vwY/Jn4RglosDgAJ
Problem: C indenting is wrong below a "case (foo):" because it is
recognized as a C++ base class construct. Issue #38.
Solution: Check for the case keyword.
d1b15dec4d
Problem: The OptionSet autocommands are not triggered from setwinvar().
Solution: Do not use switch_win() when not needed. (Hirohito Higashi)
ba117c23df
Problem: Indentation of array initializer is wrong.
Solution: Avoid that calling find_start_rawstring() changes the position
returned by find_start_comment(), add a test. (Hirohito Higashi)
089af18d1f
The test was split into several blocks reusing the same input file. As it is
complicated to send text in different encodings and with control characters
from the test suite to nvim and back the results are written to a temp file
and loaded into the test from there.
Some parts of the test depend on gzip(1). They are skipped if gzip is not
available.
Some `:write` and `:edit` commands produce messages and "hit enter" prompts
that had to be treated with an extra `feed('<C-L>')`. In the original test
file this was not neccessary because it was `:source!`ed.
- `syntax_on` is documented. Rather than introduce a new undocumented
VimL global `g:syntax_off`, use a module-local flag.
- Rename "maybe" functions to follow style guidelines (use standard
module prefix)
Problem: More side effects of ":set all&" are missing. (Björn Linse)
Solution: Call didset_options() and add didset_options2() to collect more
side effects to take care of. Still not everything...
e68c25c677
Problem: Resetting 'encoding' when doing ":set all&" causes problems.
(Bjorn Linse) Display is not updated.
Solution: Do not reset 'encoding'. Do a full redraw.
b341dda575
---
":set all&" does not reset 'encoding' in neovim.
Problem: When doing an upwards search without wildcards the search fails if
the initial directory doesn't exist.
Solution: Fix the non-wildcard case. (Stefan Kempf)
c79a5452ac
If the build directory path has symlinks in it, 'make functionaltest'
fails at shada_spec.lua:177 because readme_fname has symlink but
nvim resolves the symlink when writing it into the shada file.
Problem: Crash when using "syn keyword x c". (Dominique Pelle)
Solution: Initialize the keyword tabble. (Raymond Ko, PR 397)
670acbc70f
---
see: "fix crash due to hash table not being initialized"
https://github.com/vim/vim/pull/397
To reproduce:
```sh
nvim -u NONE -c ow -c 'sy keyword x c'
```
Problem: Block-mode replace works characterwise instead of blockwise after
column 147. (Issue #422)
Solution: Set Visual mode. (Christian Brabandt)
10ad1d90da
Problem: "gv" after paste selects one character less if 'selection' is
"exclusive".
Solution: Increment the end position. (Christian Brabandt)
d29c6fea94
Problem: ml_get error when using "p" in a Visual selection in the last
line.
Solution: Change the behavior at the last line. (Yukihiro Nakadaira)
d009e86826
Problem: Dragging the current tab with the mouse doesn't work properly.
Solution: Take the current tabpage index into account. (Hirohito Higashi)
4a4b821085
Problem: On MS-Windows vimgrep over arglist doesn't work (Issue 361)
Solution: Only escape backslashes in ## expansion when it is not used as the
path separator. (James McCoy)
6e8d3b0d41
---
Fix a lint error for the closing parenthesis.
NOLINT does not help to do this.
Some toolchains apparently set _FORTIFY_SOURCE=2 in internal header
files. Include <string.h> (which in turn should include such internal
header files) before checking the value of _FORTIFY_SOURCE to catch
that.
Fixes#4183.
Problem: "zt" still doesn't work well with filler lines. (Gary Johnson)
Solution: Check for filler lines above the cursor. (Christian Brabandt)
a09a2c5857
Problem: Comparing utf-8 sequences does not handle different byte sizes
correctly.
Solution: Get the byte size of each character. (Dominique Pelle)
f6470c288c
Problem: 'smarttab' is also effective when 'paste' is enabled. (Alexander
Monakov)
Solution: Disable 'smarttab' when 'paste' is set. (Christian Brabandt)
Do the same for 'expandtab'.
54f018cd59
Problem: gettabvar() doesn't work after Vim start. (Szymon Wrozynski)
Solution: Handle first window in tab still being NULL. (Christian Brabandt)
7e47d1ac6a
Problem: When 'conceallevel' is 1 and quitting the command-line window with
CTRL-C the first character ':' is erased.
Solution: Reset 'conceallevel' in the command-line window. (Hirohito
Higashi)
fa67fbe6b8
Problem: When removing from 'path' and then adding, a comma may go missing.
(Malcolm Rowe)
Solution: Fix the check for P_ONECOMMA. (closes#471)
174674743d
Features:
ef66249 tabline: Add %[] atom to the tabline, for random commands on click
f338ea7 job control: implement jobpid() to get PID of job
d0d5d17 job control: add 'detach' option to jobstart
7ad3f07 Add support for binary numbers
Fixes:
291495a regexp_nfa.c: Speed up find_match_text()
317d5ca input: Do not set high-bit; preserve ALT modifier.
3b7c409 shell: Unquote &shell* options before using them
Notable changes:
49b06a8 encoding: Always use "utf-8" as default for &encoding
79a6983 ui: revert "gui_running" hack
Other changes:
b4b4536 version: semver.org compliance
c6aa716 reproducible builds: Stop using __{DATE,TIME}__
46bd3c0 clipboard: Check $DISPLAY. Prefer xsel. #4150f6ecd12 job control: don't kill PTY processes on exit
49f0417 clipboard: Detach clipboard helper, so contents is kept after nvim exit
38435e8 python: Add missing I/O methods to RedirectStream
d26b01d eval: Use better error messages when failing to dump values
62d137c Remove swapsync.
Problem: Cannot specify a vertical split when loading a buffer for a
quickfix command.
Solution: Add the "vsplit" value to 'switchbuf'. (Brook Hong)
a594d77ffc
Problem: On MS-Windows, when 'autochdir' is set, diff mode with files in
different directories does not work. (Axel Bender)
Solution: Remember the current directory and use it where needed. (Christian
Brabandt)
d87c36ea5e
---
see: "autochdir + encoding=utf8 messes up diff"
https://groups.google.com/d/msg/vim_dev/QrE4Y2LMJR8/uxigns5KGxYJ
By default, find_library() searches all directories for one possible
name and then looks for the next name. To make sure we're building
against the same headers and libraries, look for all names in a
directory before moving to the next one.
In 33bc332, version constraints were added to pkg_search_module(), but
that only affects the set of directories searched by
find_library()/find_path().
Once the header directory is found, parse the version from
version_master.h so it can be checked by the find_package() call in the
root CMakeLists.txt.
Problem: gettext() is called too often.
Solution: Do not call gettext() for messages until they are actually used.
(idea by Yasuhiro Matsumoto)
77354e78a8
Problem: When "$" is in 'cpo' the popup menu isn't undrawn correctly.
(Issue 166)
Solution: When using the popup menu remove the "$".
478c46e50f
NOTE: To reproduce in nvim:
nvim -u NONE -c 'exe "norm iaaa iabbbbbb acc" | norm yyp' -c 'set cpo+=$'
Problem: Fold can't be opened after ":move". (Ein Brown)
Solution: Delete the folding information and update it afterwards.
(Christian Brabandt)
d5f6933d5c
The test involves argument and buffer numbers. Therefore it was necessary to
use a custom testing session to ensure that the initial buffer corresponds to
an argument.
Problem: When 'M' is in the 'cpo' option then selecting a text object in
parenthesis does not work correctly.
Solution: Keep 'M' in 'cpo' when finding a match. (Hirohito Higashi)
438b64ab71
Declare dependency in terms of directory, rather than individual doc
files to avoid target dependency cycles. This still maintains install
targets at doc file level.
Plugins (YCM, dispatch.vim) use gui_running to decide behavior; so do
some colorschemes. Up to now, nvim lied about gui_running for the
benefit of colorschemes and the detriment of all _other_ plugins that
check this condition. That's counterproductive: a user employs at most
_one_ colorscheme but may use many other plugins which expect the Vim
legacy semantics of gui_running.
Moreover, colorschemes usually don't _need_ to check gui_running: they
can instead set cterm{fg,bg} and gui{fg,bg} in the same :highlight call.
It is reasonable for users who want "true color" to modify their
colorscheme once (or request upstream to do so) in order to avoid
running into quirks in any other plugins.
Closes#2782
Problem: Invalid memory access in file_pat_to_reg_pat().
Solution: Use vim_isspace() instead of checking for a space only. (Dominique
Pelle)
2288afed42
Problem: The ruler shows "Bot" even when there are only filler lines
missing. (Gary Johnson)
Solution: Use "All" when the first line and one filler line are visible.
29bc9db36e
Problem: Repeating a change in Visual mode does not work as expected.
(Urtica Dioica)
Solution: Make redo in Visual mode work better. (Christian Brabandt)
31b259bf95
This will avoid confusion if additional `it()` blocks are added.
(`setup()` only runs once per `describe()` block, whereas `before_each()`
runs before each `it()`).
Problem: When using += with ":set" a trailing comma is not recognized.
(Issue 365)
Solution: Don't add a second comma. Add a test. (partly by Christian
Brabandt)
a7b7b1cef9
Problem: Coverity warns for uninitialized variables. Only one is an actual
problem.
Solution: Move the conditions. Don't use endpos if handling an error.
7ae4fbca55
Problem: CTRL-A and CTRL-X do not work properly with blockwise visual
selection if there is a mix of Tab and spaces.
Solution: Add OP_NR_ADD and OP_NR_SUB. (Hirohito Higashi)
d79e55016c
Problem: Using "." to repeat CTRL-A in Visual mode increments the wrong
numbers.
Solution: Append right size to the redo buffer. (Ozaki Kiichi)
e1edc1caba
Problem: Cannot use the "dll" options on MS-Windows.
Solution: Support the options on all platforms. Use the built-in name as
the default, so that it's clear what Vim is looking for.
25e4fcde76
Problem: Using CTRL-A in a line without a number moves the cursor. May
cause a crash when at the start of the line. (Urtica Dioica)
Solution: Do not move the cursor if no number was changed.
3ec3261980
Problem: Using CTRL-A in Visual mode does not work well. (Gary Johnson)
Solution: Make it increment all numbers in the Visual area. (Christian
Brabandt)
3a304b2382
Problem: Search does not handle multi-byte character at the start position
correctly.
Solution: Take byte size of character into account. (Yukihiro Nakadaira)
5f1e68b7bc
is_executable_in_path() searches for executables in $PATH, but on
Windows executable files have extensions available in the environment
var $PATHEXT. This commit changes is_executable_in_path() to append
those extensions to the filename.
This patch diverges from standard Vim, in that Vim only checked for
the given filename if it already has an extensions. This one always
checks for the given filename.
Problem: ":cnext" may jump to the wrong column when setting
'virtualedit=all' (cs86661)
Solution: Reset the coladd field. (Hirohito Higashi)
b8c890035e
Problem: The entries added by matchaddpos() are returned by getmatches()
but can't be set with setmatches(). (Lcd)
Solution: Fix setmatches(). (Christian Brabandt)
0fce425772
Wrapping it will send it to the terminal "directly", which might change the
cursor of another pane, e.g. when starting Neovim and changing to
another pane directly.
tmux has a terminfo extension (Ss/Se) to handle and translate the
DECSCUSR sequences internally. This can be controlled through
`terminal-overrides`, but does not appear to be able to handle the
uncommon sequences for Konsole.
Problem: Illegal memory access when using :copen and :cclose.
Solution: Avoid that curbuf is invalid. (suggestion by Justin M. Keyes)
Add a test.
62ef797496
Problem: ":tabmove" does not work as documented.
Solution: Make it work consistently. Update documentation and add tests.
(Hirohito Higashi)
40ce3a4e1f
More updated runtime files.
83d1b19015
Left out:
- doc/tags
- doc/todo.txt
- runtime/tutor/de.*
- runtime/syntax/vim.vim that seems to have already been
applied/autogenerated without the has(...) tests
Problem: test_listlbr_utf8 sometimes fails.
Solution: Don't use a literal multibyte character but <C-V>uXXXX. Do not
dump the screen highlighting. (Christian Brabandt, closes#518)
1c57fe8b94
Problem: When using 'colorcolumn' and there is a sign with a fullwidth
character the highlighting is wrong. (Andrew Stewart)
Solution: Only increment vcol when in the right state. (Christian Brabandt)
32a214e78d
Problem: Appending in Visual mode with 'linebreak' set does not work
properly. Also when 'selection' is "exclusive". (Ingo Karkat)
Solution: Recalculate virtual columns. (Christian Brabandt)
74db34cc91
Problem: Various problems with locked and fixed lists and dictionaries.
Solution: Disallow changing locked items, fix a crash, add tests. (Olaf
Dabrunz)
9bc174b69d
As the work of this migration did take quite some time, some changes where
made in master after this branch was started. These are ported to the new
test file.
The old test_eval.in file did a `:wq` on line 168. The following 60 lines
where not used and there was no expected output for them in test_eval.ok.
This test code is now used in several test cases in the new test file.
While in the `while` loop at line 603 of function searchit(), memory
address ptr+matchpos is always valid. The strlen() check should not be
necessary to verify this.
Also added a check to prevent reading a line after the end of the
buffer.
An extra test in commit 0a116c828d was
introduced, to check for end of string with a call to strlen(). This was
necessary, because an incorrect length for invalid byte sequences was
used to step through the string. This slowed down find_match_text()
compared to vim's version.
To speed up things, the extra check was removed and a sequence length
of 1 for invalid byte sequences is used.
Fixes issue #3486
Problem: In a string "\U" only takes 4 digits, while after CTRL-V U eight
digits can be used.
Solution: Make "\U" also take eight digits. (Christian Brabandt)
acc39888cd
Problem: Searching for a character matches an illegal byte and causes
invalid memory access. (Dominique Pelle)
Solution: Do not match an invalid byte when search for a character in a
string. Fix equivalence classes using negative numbers, which
result in illegal bytes.
d82a2a990b
Problem: When there are illegal utf-8 characters the old regexp engine may
go past the end of a string.
Solution: Only advance to the end of the string. (Dominique Pelle)
0e462411ca
libmsgpack was the old C++ library provided by msgpack-c. The C library
is libmsgpackc.
The C++ support became header-only, but there was a bug
(msgpack/msgpack-c#395) wherein using msgpack-c's CMake build system
would only install libmsgpack instead of libmsgpackc.
Searching for both libraries, but preferring libmsgpackc, allows for
building against older msgpack-c releases and prepares for the upcoming
msgpack-c release which fixes the aforementioned issues.
Signed-off-by: James McCoy <jamessan@jamessan.com>
Problem: Autocommands triggered by quickfix cannot get the current title
value.
Solution: Set w:quickfix_title earlier. (Yannick)
Also move the check for a title into the function.
81278efadf
Problem: glob() and globpath() cannot include links to non-existing files.
(Charles Campbell)
Solution: Add an argument to include all links with glob(). (James McCoy)
Also for globpath().
a245bc79b4
Using `:cscope find s <cword>` with the cursor on a very short word like
`key` does not output the first line on the screen correctly: Output is
`Cscope tag: keyrd>` instead of
`Cscope tag: key`.
To fix this, clear the screen line after the first line was printed.
Variable tbuf going out of scope leaks the storage it points to.
We don't have to use the copy tbuf of a match. Because matches are always in
ctags style, we can operate on them directly.
Dereferencing a pointer that might be null(ptag) when calling strlen().
False positive. A match always contains a tab, so ptag is never null.
Because matches are always in ctags style, we can rewrite the code to
not use strtok().
Problem: 'foldcolumn' may be set such that it fills the whole window, not
leaving space for text.
Solution: Reduce the foldcolumn width when there is not sufficient room.
(idea by Christian Brabandt)
1c93429c48
This is consistent with VIM behavior. When showing a visual selection,
VIM only extends it past the last character if eol is in listchars (even
if nolist is set).
Unix systems must have this header but Windows does not have it at all.
Since src/nvim/os/unix_defs.h includes <unistd.h> without the guard
in order to avoid including this in the numerous places we would
need <unistd.h> on Unix we just include src/nvim/os/os.h which will pull
in <unistd.h> for us.
`RedirectStream` is used to redirect `stdout` and `stderr`, but are
missing certain I/O methods available on other file-like objects.
This causes external plugins (like `colorama`) to crash.
Inheriting from `io.IOBase` adds an abstract implementation of these
methods, which will at least keep the python code running.
Fixes#4045
When using the Visual Studio generator don't use get_target_property for
custom command, because it returns unexpanded VS variables (e.g.
$(Configuration)) within the result as part of the LOCATION path.
The single case where this is a problem is for getting the output path for
nvim-test, instead we use a path here.
This header is required by POSIX for the constants (O_RDONLY, etc.)
but we were only including it on Unix systems as a side effect of
including <unistd.h>.
Problem: When running new style tests the output is not visible.
Solution: Add the testdir/messages file and show it. Update the list of
test names.
096c8bb40d
- ROOT_UID's comment is misleading, as it's always defined to 0.
- SEEK_{SET,END} & O_NOFOLLOW should already be defined on Unix-like
systems in <stdio.h> and <fcntl.h>, respectively. In any case,
neither of those #ifdef blocks should be in the middle of source files.
- The S_IS{LNK,DIR,...} macros should only be undefined on Windows.
Problem: When using the 'c' flag of ":substitute" and selecting "a" or "l"
at the prompt the flags are not remembered for ":&&". (Ingo
Karkat)
Solution: Save the flag values and restore them. (Hirohito Higashi)
cad2fc9935
Problem: Invalid memory access when there are illegal bytes.
Solution: Get the length from the text, not from the character. (Dominique
Pelle)
2186ffa2c7
Problem: When 'list' is set Visual mode does not highlight anything in
empty lines. (mgaleski)
Solution: Check the value of lcs_eol in another place. (Christian Brabandt)
d59c099120
Mentioned here:
https://github.com/neovim/neovim/pull/3985#issuecomment-170663426
HAVE_BE64TOH is defined in `config/config.h', which is included by
`vim.h'. Since the HAVE_BE64TOH check in `shada.c' is evaluated before
`vim.h' is included, it always evaluates to false, meaning that
be64toh() in shada.c is always used instead of the one in <endian.h>.
Moving the HAVE_BE64TOH block to after where `vim.h' is included seems
to fix the issue.
Helped by @Shougo.
vim-patch:7.4.955
vim-patch:7.4.974
vim-patch:7.4.975
vim-patch:7.4.989
Port upstream vim patches 955, 974, 975 and 989. Mark patches
964, 968, 970, and 971, and 982 as NA. Update patch list to 1022.
patch 7.4.955
Problem: Vim doesn't recognize .pl6 and .pod6 files.
Solution: Recognize them as perl6 and pod6. (Mike Eve)
patch 7.4.974
Problem: When using :diffsplit the cursor jumps to the first line.
Solution: Put the cursor on the line related to where the cursor was before
the split.
patch 7.4.975
Problem: Using ":sort" on a very big file sometimes causes text to be
corrupted. (John Beckett)
Solution: Copy the line into a buffer before calling ml_append().
patch 7.4.989
Problem: Leaking memory when hash_add() fails. Coverity error 99126.
Solution: When hash_add() fails free the memory.
778 marked as not NA as it will be needed once vim patch 754 is merged
Marked as NA:
964 test 87 was deleted
968 tests 86/87 were deleted
970 guarded by: `# if defined(FEAT_GUI_GTK) || defined(PROTO`
and is inside a function that no longer exists
971 function table already sorted correctly
982 marked as NA because Neovim tests are only specified in exactly one location
vim-patch:7.4.569
vim-patch:7.4.573
Helped-by: @glts https://github.com/neovim/neovim/pull/2621
Problem: Having CTRL-C interrupt or not does not check the mode of the
mapping. (Ingo Karkat)
Solution: Use a bitmask with the map mode. (Christian Brabandt)
651863c94a
Problem: Mapping CTRL-C in Visual mode doesn't work. (Ingo Karkat)
Solution: Call get_real_state() instead of using State directly.
5000869712
In Vim, at least the constant `I_PUSH` is used from <stropts.h>, but
Neovim doesn't seem to use anything from said header.
Besides that, POSIX.1-2008[1] marks this header as obsolescent, and
there don't seem to be many platforms that even have it.
[1]: http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/stropts.h.html#tag_13_52_11
In Windows the separator character in the PATH environment is ';' instead
of ':'. Add a new define ENV_SEPCHAR to be used instead of hardcoding
the character literal.
Background: Vim internally prefers to represent ALT/META chords as
single-byte keys, by setting the high bit of the key byte.
extract_modifiers() _discards_ the meta/alt modifier, but we need it for
libvterm and libtermkey.
Closes#2440Closes#3727Closes#2017
References #2277
References #2254https://github.com/neovim/neovim/issues/2017#issuecomment-140423557
> We [not libtermkey] are setting the high bit for some reason
https://github.com/neovim/neovim/issues/176#issuecomment-77834715
> libvtermkey requires the leading esc to parse alt/meta
https://github.com/neovim/neovim/pull/3246#issuecomment-136328450
> A program could do better than the current logic on some terminals, by
> asking for pure 8bit mode (S8C1T) and then immediately querying the
> mode again. If the result comes back as an 8bit single-byte CSI, then
> it can presume the mode setting was successful, and now the ESC prefix
> byte won't be seen in multibyte sequences; only as an Alt- prefix or
> a real Escape key. On such a terminal, it could therefore avoid
> needing to use that waiting timeout.
The Visual Studio cmake generator creates subdirectories inside
the build path for different build configuration. But this breaks some of our
cmake scripts, like the help tags installer, that assume the targets are
built in that location. Updated CMakeLists.txt to remove extra paths.
- Check if MIN_LOG_LEVEL value is a number 0-3, default to
INFO (1) or ignore it in Release mode
- When TRAVIS_CI_BUILD is ON the default is DEBUG (0)
- Add local.mk.example
The block is only compiled if the system is NOT WIN32, but the closing
parenthesis for the statement opened outside the block is closed inside
the block.
Problem: Running "make newtests" in src/testdir has no output.
Solution: List the messages file when a test fails. (Christian Brabandt)
Update the list of tests.
e7893a4088
Problem: message from assert_false() does not look nice.
Solution: Handle missing sourcing_name. Use right number of spaces. (Watiko)
Don't use line number if it's zero.
cbfe32953a
Problem: When a test script navigates to another buffer the .res file is
created with the wrong name.
Solution: Use the "testname" for the .res file. (Damien)
de0ad40cb3
Problem: New style testing is incomplete.
Solution: Add the runtest script to the list of distributed files.
Add the new functions to the function overview.
Rename the functions to match Vim function style.
Move undolevels testing into a new style test script.
683fa185a4
Problem: Writing tests for Vim script is hard.
Solution: Add assertEqual(), assertFalse() and assertTrue() functions. Add
the v:errors variable. Add the runtest script. Add a first new
style test script.
43345546ae
Otherwise the dynamic library is built also and find_library will prefer
that over the static one. That results in linking against the dynamic
library which will not be found after install.
This code:
8b3c399b6d/third-party/CMakeLists.txt (L130)
should prevent the above problem, but it doesn't hurt to be explicit.
- Remove mention of "build-issues" and "runtime-issues" sections from
the "Troubleshooting" page; they're already mentioned at the top of
the section: "Before reporting an issue, see the ...".
- As of [1], clint-ignored-files.txt isn't used anymore.
[1]: 57eaefbb23
We use `git describe` to stamp pre-release versions (dev builds). But
`git describe` uses the result of the most-recent tag (the current
_release_ version)--so we must munge it with the _next_ (i.e.
unreleased) version.
Also fix non-git builds: do not invoke git_describe(NVIM_VERSION_MEDIUM)
if we're not in a git tree, else it gets the dummy value
"HEAD-HASH-NOTFOUND".
Example :version output in non-git build:
NVIM 0.1.2-dev
Example :version output in git build:
NVIM v0.1.2-176-g9c3c2b5
The nested line was the else-branch of an if-then-else block that dealt
with cryptography, but after commit
85338fe1d5 (Remove cryptography) removed
the if-then part, the indentation of this line was not adjusted.
Without this compilation fails due to a missing symbol: SRWLOCK in libuv
headers. _WIN32_WINNT defines the Windows header version that is to be used,
and it seems libuv requires this version. See
b471b33da8
Examples:
let g:SR = [[]]
call add(g:SR[0], g:SR)
wshada
" E952: Unable to dump variable g:SR: container references itself in index 0, index 0
let g:F = {'_TYPE': v:msgpack_types.map, '_VAL': [[{'abc': 1}, function("tr")]]}
wshada
" E951: Error while dumping variable g:F, key {'abc': 1} at index 0 from special map, key '': attempt to dump function reference
" (no msgpack#string available)
" E951: Error while dumping variable g:F, key {="abc": 1} at index 0 from special map, key '': attempt to dump function reference
" (msgpack#string available)
let g:F = {'_TYPE': v:msgpack_types.map, '_VAL': [[g:SR, function("tr")]]}
wshada
" E951: Error while dumping variable g:F, key [[[[{E724@0}]]]] at index 0 from special map, index 1: attempt to dump function reference
call msgpackdump([g:SR])
" E952: Unable to dump msgpackdump() argument, index 0: container references itself in index 0, index 0
Not tested yet.
Since uid_t is not defined on Windows use uv_uid_t instead.
We now use uv_uid_t everywhere except one place in src/nvim/main.c
which is wrapped in a #ifdef UNIX check.
The do_path_expand() function is still using the unix_expandpath
variant from Vim. For Windows it should behave as the old
dos_expandpath() function. This commit adds an ifdef to ignore
EW_ICASE flag in this function, otherwise all matches fail on Windows
if wildignorecase is set.
Problem: The # register is not writable, it cannot be restored after
jumping around.
Solution: Make the # register writable. (Marcin Szamotulski)
3b3a9498d1
Problem: It's a bit clumsy to execute a command on a list of matches.
Solution: Add the ":ldo", ":lfdo", ":cdo" and ":cfdo" commands. (Yegappan
Lakshmanan)
aa23b37942
- Nvim has no concept of "huge" or "normal" features: the overwhelming
majority of features are compiled in by default
- Nvim does not link to X (X11), so doesn't support setting things via
~/.Xresources, among many other things
We do not support every Unix flavor in existence so note
that here.
We do not support Windows 95 or NT. Update the list to
the Windows versions we do support.
We only support OS X not classic Mac OS so note that too.
The UI_CALL macro uses variadic macros and relies on the non standard
GCC extension. MSVC already applies similar behavior for the regular
__VA_ARGS__ removing the comma.
In MSVC UI_CALL can be implemented by calling UI_CALL_MORE directly.
When converting a msgpack object to a String object, strings (and byte
arrays) with length 0 are handled as errors. This is fixed by
always using the msgpack data pointer as a valid pointer. For a NULL
pointer there is nothing to copy.
Test by @snoe
Fixes#3844
Problem: test_eval fails when the clipboard feature is missing.
Solution: Skip part of the test. Reduce the text used.
4ac163ae5f
Problem: test_eval fails on Mac.
Solution: Use the * register instead of the + register. (Jun Takimoto)
e08dd4e49e
Any patch may contain mixed encodings, so we must process them as byte
arrays. E.g. with stock `sed` on OS X patch
8a94d873aa8c753a8522ea86a049bdf2abd0c507 causes this error:
sed: RE error: illegal byte sequence
To avoid that, set LC_ALL=C.
Also remove redundant *.patch creation from review_pr().
Problem: The search highlighting and match highlighting replaces the
cursorline highlighting, this doesn't look good.
Solution: Combine the highlighting. (Yasuhiro Matsumoto)
09deeb7c94
Problem: When a FileReadPost autocommand moves the cursor inside a line it
gets moved back.
Solution: When checking whether an autocommand moved the cursor store the
column as well. (Christian Brabandt)
eab316bdf9
Problem: 'linebreak' does not work properly with multi-byte characters.
Solution: Compute the pointer offset with mb_head_off(). (Yasuhiro
Matsumoto)
76feaf1bfe
Problem: Warning for shadowing a variable.
Solution: Rename off to mb_off. (Kazunobu Kuriyama)
4df702999d
Problem: If no NL or CR is found in the first block of a file then the
'fileformat' may be set to "mac". (Issue 77)
Solution: Check if a CR was found. (eswald)
05eb612ff3
Problem: 'formatexpr' is evaluated too often.
Solution: Only invoke it when beyond the 'textwidth' column, as it is
documented. (James McCoy)
0f8dd840fc
Problem: When splitting the window in a BufAdd autocommand while still in
the first, empty buffer the window count is wrong.
Solution: Do not reset b_nwindows to zero and don't increment it.
8da9bbfd02
Problem: After deleting characters in Insert mode such that lines are
joined undo does not work properly. (issue 324)
Solution: Use Insstart instead of Insstart_orig. (Christian Brabandt)
c3bbad085c
Windows does not have setenv(), instead the _putenv_s() function is used - added
a function check and fatal errors. Implemented os_setenv() for Windows.
Vim supports the original putenv() function if no alternative is available.
Neovim only supports systems where safer alternatives exist, so the check for
putenv() was removed from config/CMakeLists.txt.
Currently,
```
if bool then
--stuff
elseif bool2 then
--morestuff
elseif bool3 then
--more stuff
else
--fail
end
```
Would get indented out strangely when using =. Now it behaves correctly.
File names starting with periods are perfectly acceptable on Windows
file systems. The only place where this is not acceptable is on
MS-DOS FAT file systems which only support 8.3 file names.
See here:
https://msdn.microsoft.com/en-us/library/windows/desktop/aa365247%28v=vs.85%29.aspx
Since Neovim does not support MS-DOS or 8.3 file names (#605)
we can drop this codepath.
It was not compiling anyways since we do not define WIN3264.
@equalsraf: I took a look at the Vim source pstrcmp() is actually used in the
dos_expandpath(misc1.c). The only difference between the UNIX and WIN32 version
of the functions is the _cdecl call convention annotation - the body of the
function is identical. Neovim kept the comment from the Unix function but not
from the Windows variant. Seems to me its safe to use the same function for
both - and just correct the comment.
If nvim is built from a non-tagged commit, the truncated commit hash is
already appended to the main version string (e.g., "NVIM v0.1.0-83-g959f260 ..."),
making the "Commit:" field redundant.
Regarding the truncated hash length: we don't have nearly enough commits
to worry about collisions, and probably won't ever, so the default
length should be fine.
The tests would leave the following test files in the root directory:
Xtest-functional-plugin-shada.shada
Xtest-functional-plugin-shada.shada.tmp.f
Clean them up in teardown().
So far luacheck's rockspec specified only the git protocol. Hence people
behind firewalls/proxies, that block port 9814, had trouble fetching this
dependency via luarocks.
The latest commit updated the rockspec to use either git or https. Thus common
workarounds like this are not needed anymore:
git config --global url."https://".insteadOf git://
References #3769.
Problem: The default conceal character is documented to be a space but it's
initially a dash. (Christian Brabandt)
Solution: Make the intial value a space.
4a42710695
When building for X86 the CMake check_library_exists always fails to find
functions from the Win32 API due to name mangling conventions. The convention
for API functions is __stdcall and the CMake test code assumes __cdecl. Since
these are libraries from the Windows API we can simply link against the
libraries without checking for the functions.
Problem: Test fails when the autochdir feature is not available. Test
output contains the test script.
Solution: Check for the autochdir feature. (Kazunobu Kuriyama) Only write
the relevant test output.
d113a80c77
Problem: It is not possible for a plugin to adjust to a changed setting.
Solution: Add the OptionSet autocommand event. (Christian Brabandt)
537443018d
refs https://github.com/neovim/neovim/pull/2124#discussion_r26107174
Unlike Travis, `make test` currently only runs functional tests.
This can cause confusion since one might (understandably) think that `make
test` runs unit tests too, which it doesn't.
The `oldtest` target is still left out because it's quite slow and
Travis already runs it.
Regarding the individual items in the header:
`Vim - Vi improved by Bram Moolenar`
Bram Moolenar is already mentioned throughout the documentation, as
well as the intro screen.
`:help uganda`
It's already shown to all users who don't use `shortmess+=I` upon
starting nvim, and is already placed prominently in help.txt, i.e.,
`:help` run with no arguments.
`:help credits`
Already mentioned near the top of help.txt.
`README.md`
Already mentioned in develop.txt.
In Windows we can't assume errno will be set by calls to os_* functions,
instead the return value from os_* functions can be used. This commit fixes two
occurences for os_open().
1. EFBIG is replaced with UV_EFBIG and checked against the return from os_open().
2. EOVERFLOW does not have a corresponding libuv constant, and is not defined
by open() in Windows - disabled this case with a UNIX guard, and check the return
value against -EOVERFLOW (libuv errors are negative errno values in Unix).
From #3473, shada.c used errno constants (e.g. ENOENT) to check the return
of os_open(), but in Windows the return from libuv functions is not -errno.
Instead use libuv error constants (e.g UV_ENOENT) for error checks.
Replaced old unit tests for errno with libuv error codes UV_ENOENT
and UV_EEXIST (for os_open and os_getperms).
Added libuv include path to test/includes compiler calls - needed
to get hold of libuv headers.
Previously os_getperms() returned -1 for any error condition, it
now returns the libuv error code (as returned by os_stat()). This
allows checking for error conditions without relying on errno
(which not available in Windows).
The only case where the errno value from os_getperms() was being used
was in readfile() to set the new-file flag - replaced the errno check
with UV_ENOENT.
Instead of returning bool from os_stat return the actual libuv return code.
This function is static and used internally in nvim/os/fs.c it should not
impact the rest of the API. This is a first step to change other fs functions.
In windows libuv does not return -errno, instead it uses negative
error codes e.g. UV_ENOENT. This commit changes the comments in os_*
functions to reflect this.
It's not documentation in the same sense as the majority of files in
runtime/doc, so is of little use to most users and should probably not
be installed alongside the "real" documentation.
It may be full of good ideas, but it's also full of things which are no
longer applicable to Nvim, such as references to platforms we no longer
support (e.g., MS-DOS), Vi compatibility, Autoconf, the Perl interface,
etc.
If someone is looking for things to fix or improve, the GitHub issue
tracker should prove useful, and, unlike todo.txt, the issues are
generally much more relevant to Nvim. Besides all that, removing
todo.txt makes porting runtime patches a bit easier.
refs #2911, discussed briefly in #2608 and #2553
All `{not in Vi}' tags have been removed, so the first sentence is no
longer true. As for the second sentence, the header ("Differences
between Vim and Vi") is already self-explanatory, so it can be removed
as well.
[RFC] doc/develop.txt pruning
Reviewed-by: Felipe Morales < hel DOT sheep AT gmail DOT com >
Reviewed-by: Marco Hinz < mh DOT codebro AT gmail DOT com >
These are more like notes than proper documentation, so remove it as
they're not of much use unless you don't know what you're looking for.
Many of these are somewhat useful, but many also duplicate what's
already documented elsewhere. In any case, vi_diff.txt is the last
place that assorted tips'n'tricks should be located (we already have
tips.txt!!!).
In regards to Nvim, it's very outdated; for an updated list of such
arguments refer to the manual page or `nvim --help`. In all other
regards, it's of little to no use to end-users.
Note: it looks like viminfo files do not store search direction intentionally.
After reading viminfo file search direction was considered to be “forward”.
Note 2: all files created on earlier Neovim version will automatically receive
“forward” direction.
Fixes#3580
This change adds switch cases for K_FOCUSGAINED and K_FOCUSLOST to the
input handling functions in ex_getln.c and terminal.c. The handling is
identical to what's found in edit.c (just calling apply_autocmds).
If one enters cmdline-mode by feeding `:` and sends a focuslost event (by
leaving the window for example) the text `<FocusLost>` will be inserted
into the command line. There is similar behaviour in terminal mode. This
patch corrects this behavior to fire the apropriate autocmd instead.
Fixes#3714
Problem: On some systems automatically adding the missing EOL causes
problems. Setting 'binary' has too many side effects.
Solution: Add the 'fixeol' option, default on. (Pavel Samarkin)
34d72d4b6c
Problem: Text deleted by "dit" depends on indent of closing tag.
(Jan Parthey)
Solution: Do not adjust oap->end in do_pending_operator(). (Christian
Brabandt)
b6c2735c56
NA patches list:
652: xxd changes xxd is not included in neovim
744: if_ruby and if_perl legacy tests
750: if_perl build problem
751: compile options
756: if_perl support
757: term.c changes
759: if_lua support
761: term.c changes
762: term.c changes
763: if_lua support
766: term.c changes
770: term.c changes
772: if_mzsch support
802: duplicated test
809: revert of 802
829: GUI
837: Windows with if_sniff
840: GUI
846: CONTRIBUTING.md
850: CONTRIBUTING.md
851: Win32 console
852: Win32 console
855: GTK GUI
861: GTK GUI
862: GTK GUI
867: Windows Fix
869: Windows GUI
872: CI service
874: Windows GUI
875: README.md
876: Windows GUI
880: CI service
884: CI service
886: Windows GUI
890: if_python3
894: vimrun.exe
899: README.md
900: README.md
902: Win32 console
906: Windows problem
907: Dynamic loading
908: Windows build error
910: if_python
911: terminal
913: hangul input GUI only
917: hangulin.c changes
918: if_python
919: if_lua, if_python, if_python3
920: ruby stuff
921: Windows
Reviewed-by: Michael Reed <Pyrohh@users.noreply.github.com>
Two new keys have been added to key_name_entry in keymap.c:
`FocusGained` and `FocusLost`.
Two cases have been added to the key handing switch in edit.c each
applying their respective autocmds.
In normal.c two functions have been added alongside nv_cursorhold doing
a similar job of applying the autocmd for the appropriate key.
tui/input.c has a new function `handle_focus_event` which eats either of
the control sequences for focus gained or lost. This function is checked
before handle_bracketed_paste and handle_forced_escape.
tui.c registers neovim as able to receive these control sequences in
terminfo_start and undoes that in terminfo_stop.
Closes#2302
As can be seen in [1], the ifdefs are leftovers from Vim, where tons of
features are optional. This is not the case in Nvim, so the macros are
of little use. Besides, ex_ni is already used in a bunch of places,
just not consistently.
The ex_language macro was left untouched as the inclusion of
``:language'' actually is determined at compile time, unlike the other
commands which _always_ print "this command is not available in this
version".
[1]: https://github.com/vim/vim/blob/master/src/ex_docmd.c#L18-L497
Reviewed-by: Justin M. Keyes <justinkz@gmail.com>
When running out of memory, garbage collection would free lists and
dictionaries that are not yet referenced. This would later on lead to
a use-after-free for these objects.
Releated to vim-patch:7.4.916, so also updating version.c.
Patch by @ZyX-I
"vimrc" refers to all files that are used to configure Neovim. The main
configuration file is init.vim nowadays.
All nvimrc references that are left refer to a local ".nvimrc" which is read
if 'exrc' is set. ".ngvimrc" references were completely wiped.
Closes#3552.
- change approach for test 1: screen:expect() instead of assert()
- use execute() instead of command()
- 2 new tests that check none and wrong input for :oldfiles!
Helped-by: @fwalch
Helped-by: @tarruda
Helper-by: @justinmk
Problem: Can't match "%>80v" properly for multi-byte characters.
Solution: Multiply the character number by the maximum number of bytes in a
character. (Yasuhiro Matsumoto)
4f36dc3bf7
Libuv and LuaJIT like to hide the actual compilation and linking
commands behind nice text. This change makes them spit out the actual
command line to help us with debugging issues that people are seeing.
Currently the logfile (":profile start {logfile}") only gets written when Vim
exits. This new command allows to dump the log immediately without exiting.
Problem: "zr" and "zm" do not take a count.
Solution: Implement the count, restrict the fold level to the maximum
nesting depth. (Marcin Szamotulski)
7d2757a472
From the documentation itself:
:[range]o[pen] Works like |:visual|: end Ex mode.
{Vi: start editing in open mode}
...
Vim does not support open mode, since it's not really useful. For
those situations where ":open" would start open mode Vim will leave Ex
mode, which allows executing the same commands, but updates the whole
screen instead of only one line.
Part of the reason behind this is to make removing vi_diff.txt easier,
although it's also because :open is not too useful.
Helped-by: @fdinoff
Helped-by: @dsummersl
Helped-by: @mhinz
Helped-by: @justinmk
When building in a git repo:
- If HEAD corresponds to an annotated tag, (i.e. git_get_exact_tag()
returns truthy) the current build is considered a "release" build:
NVIM_VERSION_MEDIUM is directly assigned the tagged version name,
and NVIM_VERSION_* defines are ignored.
- If HEAD is not a tagged release, then NVIM_VERSION_MEDIUM is
directly assigned the result of `git describe`.
If git (or the repo) is not available:
- The NVIM_VERSION_* defines are used to define NVIM_VERSION_MEDIUM.
Sample outputs for `nvim --version` and `nvim +version`:
Building with git @ non-tagged commit e66df14:
NVIM v0.1.0-1-ge66df14 (compiled Nov 1 2015 19:10:30)
Commit: e66df148f9401be17adab324a6e41d927aae20b3
Building with git @ v0.1.1 tag:
NVIM v0.1.1 (compiled Nov 1 2015 19:03:52)
[no "Commit:" line]
Building this commit _not_ in a git repo:
NVIM 0.1.0-dev (compiled Nov 1 2015 19:16:11)
[no "Commit:" line]
[](https://gitter.im/neovim/neovim?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge)
This provides the code necessary to configure vim's YCM plugin to provide C semantic support (completion, go-to-definition, etc) for the Neovim project.
This provides the code necessary to configure vim's YCM plugin to provide C semantic support (completion, go-to-definition, etc) for developers working on the Neovim project.
let [classname_candidate,class_candidate_namespace] =phpcomplete#GetCallChainReturnType(classname_candidate,class_candidate_namespace,class_candidate_imports,methodstack)
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
