The flag enables the current local directory set by ":lcd" to be saved
to views which is the current default behaviour. The option can be
removed to disable this behaviour.
closes#7435
vim-patch:8.0.1289
neovim-ruby-host is a ruby script.
neovim-node-host is a shell script.
Both don't work in cmd.exe so gem and npm provide batchfile shims.
Return the full path of these shims, cmd.exe knows better what to do with these files.
Make the Windows impl closer to Vim os_win32.c, and the Unix impl closer
to Vim os_unix.c.
Outcomes:
- Do not send negative fd to close(). ref #4806#4772#6860
- Fallback return-value is now correct in (hopefully) all cases.
- unix: check S_ISXXX instead of relying on os_open (which can fail for
irrelevant reasons). buf_write() expects NODE_WRITABLE for character
devices such as /dev/stderr. 96f834a842
Anything claiming to be an xterm gets DECSCUSR. This is the only
reasonable choice unless/until we get more reliable detection (#7490).
ref #6997closes#7550
Fix bug that checked for npm AND yarn, where we wanted npm OR yarn.
But since we call `npm` exclusively, and it's highly unlikely you have
yarn installed without npm, let's just remove the yarn check altogether.
Addresses https://github.com/neovim/node-client/issues/41
Problem: Crash when 'fileformat' is forced to "dos" and the first line in
the file is empty and does not have a CR character.
Solution: Don't check for CR before the start of the buffer.
2aa5f696b9
:checkhealth reports that remote plugins are unregistered
after running :UpdateRemotePlugins because of the backslashes in filepath.
Normalize them to forward slashes because the paths in rplugin.vim are normalized in autoload/remote/host.vim.
* vim-patch:8.0.0245
Problem: The generated zh_CN.cp936.po message file is not encoded properly.
Solution: Instead of using zh_CN.po as input, use zh_CN.UTF-8.po.
16038d50c4
* vim-patch:8.0.0248
Problem: vim_strcat() cannot handle overlapping arguments.
Solution: Use mch_memmove() instead of strcpy(). (Justin M Keyes,
closesvim/vim#1415)
45600ce8f2
Integer is a 64-bit type so using %d can produce incorrect results.
test/functional/api/highlight_spec.lua @ 35: highlight api nvim_get_hl_by_id
...W7Xi/neovim-0.2.1/test/functional/api/highlight_spec.lua:46: Expected objects to be the same.
Passed in:
(string) 'Invalid highlight id: 7671724'
Expected:
(string) 'Invalid highlight id: 30000'
This fixes an apparent difference in behavior between Lua and LuaJIT.
Lua fails to format nil:
test/functional/terminal/tui_spec.lua:381: bad argument #2 to 'format' (string expected, got nil)
Doc for UV_PROCESS_DETACHED in uv.h mentions:
> child process will still keep the parent's event loop alive unless
> the parent process calls uv_unref() on the child's process handle.
ref #3944
Problem: When setting wildoptions=tagfile the completion context is not set
correctly. (desjardins)
Solution: Check for EXPAND_TAGS_LISTFILES. (Christian Brabandt, closesvim/vim#1399)
ba47b51ff8
Problem: Completion of user defined functions is not covered by tests.
Solution: Add tests. Also test various errors of user-defined commands.
(Dominique Pelle, closesvim/vim#1413)
65c836e600
Problem: No command line completion for :cexpr, :cgetexpr, :caddexpr, etc.
Solution: Make completion work. (Yegappan Lakshmanan) Add a test.
2b2207ba69
Problem: The test for patch 8.0.0224 misses the CR characters and passes
even without the fix. (Christian Brabandt)
Solution: Use double quotes and \<CR>.
1695f99d08
Problem: When 'fileformats' is changed in a BufReadPre auto command, it
does not take effect in readfile(). (Gary Johnson)
Solution: Check the value of 'fileformats' after executing auto commands.
(Christian Brabandt)
7a2699e868
vimrc_example.vim is not relevant to Nvim. Anything worth having in
there should be made an actual default.
.gitignore:
- remove *.orig ... super annoying
vim-patch:b6e0ec6b71c4
Documentation updates.
b6e0ec6b71
vim-patch:59ee05b2f5d4
Switch travis back to precise temporarily
59ee05b2f5
vim-patch:bb543088a551
Remove ruby-dev from travis config.
bb543088a5
vim-patch:d243a2a5efd9
Do not invoke rvm reset
d243a2a5ef
vim-patch:38ed1f5bea7d
Do not run coverage with clang, currently fails.
38ed1f5bea
Nvim note: intentionally did not include `--ttyfail` since its purpose
is not clear. (And it isn't used in any Vim test files/scripts).
---
Problem: When the input or output is not a tty Vim appears to hang.
Solution: Add the --ttyfail argument. Also add the "ttyin" and "ttyout"
features to be able to check in Vim script.
2cab0e1910
Travis disabled IPv6:
[ RUN ] serverstart(), serverstop() parses endpoints correctly: FAIL
...build/neovim/neovim/test/functional/eval/server_spec.lua:83: Expected objects to be the same.
Passed in:
(table) {
[1] = '127.0.0.1:12345' }
Expected:
(table) {
[1] = '127.0.0.1:12345'
*[2] = '::1:12345' }
Change all tests to ensure a server was actually started before
expecting it to be returned from serverlist().
Memory leak exposed by new test added in #7444.
==38771==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 2400 byte(s) in 1 object(s) allocated from:
0 0x51163d in realloc (/home/travis/build/neovim/neovim/build/bin/nvim+0x51163d)
1 0xfda51c in xrealloc /home/travis/build/neovim/neovim/src/nvim/memory.c:169:15
2 0xda6802 in ga_grow /home/travis/build/neovim/neovim/src/nvim/garray.c:98:14
3 0xb67ccd in script_line_start /home/travis/build/neovim/neovim/src/nvim/ex_cmds2.c:3286:11
4 0xb62885 in getsourceline /home/travis/build/neovim/neovim/src/nvim/ex_cmds2.c:3110:5
5 0xb5fbfd in do_source /home/travis/build/neovim/neovim/src/nvim/ex_cmds2.c:2951:15
6 0xb5c484 in cmd_source /home/travis/build/neovim/neovim/src/nvim/ex_cmds2.c:2710:14
7 0xb5c5b0 in ex_source /home/travis/build/neovim/neovim/src/nvim/ex_cmds2.c:2691:3
8 0xb950b9 in do_one_cmd /home/travis/build/neovim/neovim/src/nvim/ex_docmd.c:2242:5
9 0xb737e1 in do_cmdline /home/travis/build/neovim/neovim/src/nvim/ex_docmd.c:609:20
10 0xb79ac5 in do_cmdline_cmd /home/travis/build/neovim/neovim/src/nvim/ex_docmd.c:280:10
11 0xed8a7b in exe_commands /home/travis/build/neovim/neovim/src/nvim/main.c:1684:5
12 0xec6602 in main /home/travis/build/neovim/neovim/src/nvim/main.c:521:5
13 0x2ae968600f44 in __libc_start_main /build/eglibc-SvCtMH/eglibc-2.19/csu/libc-start.c:287
... that does not have that superflous last line.
Also, remove some indeterminism for the freebsd64 tests. Partially,
those were suggested by the tests themselves, while successfull. Some of
them were added after some testing because the lookaround test would
fail on freebsd64 only.
They were only used to not show the preview window when typing "s/" or
"s//" only, in which case the previous pattern would be reused. Now the
window is shown in that case.
Problem: Profiling skips the first and last script line.
Solution: Check for BOM after setting script ID. (Lemonboy, closesvim/vim#2103,
closesvim/vim#2112) Add a test. List the trailing script lines.
67435d9983
Problem: Test_profile is a little bit flaky.
Solution: Accept a match when self and total time are the same. (James
McCoy, closesvim/vim#1972)
d21b16f3c0
With the old behavior, if a GUI makes a blocking request that requires user
interaction (like nvim_input()), it would not get any screen updates.
The client, not nvim, should decide how to handle notifications during a
pending request. If an rplugin wants to avoid async calls while a sync call is
busy, it likely wants to avoid processing async calls while another async call
also is handled as well.
This may break the expectation of some existing rplugins. For compatibility,
remote/define.vim reimplements the old behavior. Clients can opt-out by
specifying `sync=urgent`.
- Legacy hosts should be updated to use `sync=urgent`. They could add a flag
indicating which async methods are always safe to call and which must wait
until the main loop returns.
- New hosts can expose the full asyncness, they don't need to offer both
behaviors.
ref #6532
ref #1398d83868fe90
Having timeouts that are likely to fail incurs a penalty of waiting for
screen:expect() to fail, hence removing such small timeouts will speed
up the test on average.
Problem: The test for :profile is slow and does not work on MS-Windows.
Solution: Use the "-es" argument. (Dominique Pelle) Swap single and double
quotes for system()
c011a3d083
Avoid crashing or hanging when editing a file than contains ludicrously
long lines (more than 100,000,000 virtual columns).
The change is in plines_win_nofold, which is called by wrapping and folding
code. As a result, wrapping and folding may be done incorrectly when the UI is
capable of rendering more than 32000 characters at a time (tiny font).
fixes#2838
Problem: Pasting inserted text in Visual mode does not work properly.
(Matthew Malcomson)
Solution: Stop Visual mode before stuffing the inserted text. (Christian
Brabandt, from neovim #5709)
f8eb9c51e5
closes#7239
The old behaviour was to set CMAKE_INSTALL_MANDIR to /usr/local/man
when MANPREFIX wasn't defined. This caused mismatching installation
paths when the installation prefix wasn't /usr/local.
This fix explicitely checks that the prefix is /usr/local to change
the value of CMAKE_INSTALL_MANDIR, and uses the default behaviour
otherwise, as /usr/local is the exception rather than the norm
(as per man hier(7)).
closes#7283
regression by 42d892913d
- Don't need to explicitly put "-O2 -g" in RelWithDebInfo; CMake does
that already. That was left-over from 42d892913d which removed the
"Dev" custom build-type, but repurposed the logic for RelWithDebInfo.
- `if(DEFINED MIN_LOG_LEVEL)` doesn't work.
- `if(${MIN_LOG_LEVEL} MATCHES "^$")` doesn't work if -DMIN_LOG_LEVEL is
omitted.
- `if(MIN_LOG_LEVEL)` also isn't what we want: it would be true if
MIN_LOG_LEVEL=0.
Problem: Pasting in virtual edit happens in the wrong place.
Solution: Do not adjust coladd when after the end of the line (closesvim/vim#2015)
d41babef89
closes#6726
Problem: Crash with virtualedit and joining lines.
(Joshua T Corbin, Neovim #6726)
Solution: When using a mark check that coladd is valid.
9aa1569128
Runtime updates that were bundled into the otherwise NA commit:
Problem: "make proto" adds extra function prototype.
Solution: Add vim/vim#ifdef.
5162822914
`:syntax keyword` is affected by 'iskeyword'. When we aligned
'iskeyword' to that of filetype=help, colon (:) is now included.
Simplest way to deal with this is to include colon (:) in the `:syntax
keyword` directive.
Also:
- change "SUGGESTIONS" mouthful to "ADVICE"
- change "SUCCESS" to "OK"
see also #7082
Problem: Normal colors are wrong with 'termguicolors'.
Solution: Initialize to INVALCOLOR instead of zero. (Ben Jackson, closesvim/vim#1344)
0cdb72aa38
Problem: When a C preprocessor statement has two line continuations the
following line does not have the right indent. (Ken Takata)
Solution: Add the indent of the previous continuation line. (Hirohito
Higashi)
c6aa475a27
...in order to retrieve highlights.
Added test/functional/api/highlight_spec.lua
HL_NORMAL is not really a good name, since it's more like an empty attribute than the normal's one.
If one pays attention, syn_cterm_attr2entry is never called with attr=0 because it's always special cased before.
I suggest in subsequent PRs we remove the ATTR_OFF and just insert an EMPTY ATTR/RESET_ATTR/UNINITIALIZED for id 0.
Problem: When sorting zero elements a NULL pointer is passed to qsort(),
which ubsan warns for.
Solution: Don't call qsort() if there are no elements. (Dominique Pelle)
a216255a4f
Problem: Manual folds are lost when a session file has the same buffer in
two windows. (Jeansen)
Solution: Use ":edit" only once. (Christian Brabandt, closesvim/vim#1958)
4bebc9a056
Problem: When creating a session when winminheight is 2 or larger and
loading that session gives an error.
Solution: Also set winminheight before setting winheight to 1. (Rafael
Bodill, neovim vim/vim#5717)
36ae89c550
Problem: Test_help_complete sometimes fails in MS-Windows console.
Solution: Use getcompletion() instead of feedkeys() and command line
completion. (Hirohito Higashi)
9f0e423c28
Problem: Cannot set 'dictionary' to a path.
Solution: Allow for slash and backslash. Add a test (partly by Daisuke
Suzuki, closesvim/vim#1279, closesvim/vim#1284)
7554da4033
Problem: Illegal memory access when 'complete' ends in a backslash.
Solution: Check for trailing backslash. (Dominique Pelle, closesvim/vim#1478)
226c534291
Functions with FUNC_ATTR_ALWAYS_INLINE need to be defined before they
are called to work around bugs with some compiler versions. When the
body is after the use of the function, compilation will fail with
/home/niko/build/neovim/src/nvim/regexp.c: In function 'regmatch':
/home/niko/build/neovim/build/src/nvim/auto/regexp.c.generated.h:77: sorry, unimplemented: inlining failed in call to 'cstrchr': function body not available
/home/niko/build/neovim/src/nvim/regexp.c:4193: sorry, unimplemented: called from here
Problem: Argument list is not stored correctly in a session file.
(lgpasquale)
Solution: Use "$argadd" instead of "argadd". (closesvim/vim#1434)
79da563cf9
Signed-off-by: Michael Schupikov <michael@schupikov.de>
As of unibilium 1.2.1, directly manipulating unibi_var_t is deprecated.
../src/nvim/tui/tui.c: In function 'update_attrs':
../src/nvim/tui/tui.c:321:7: warning: 'i' is deprecated: use unibi_var_from_num or unibi_num_from_var instead [-Wdeprecated-declarations]
data->params[0].i = (fg >> 16) & 0xff; // red
^~~~
In file included from ../src/nvim/tui/tui.c:12:0:
/usr/include/unibilium.h:632:9: note: declared here
int i UNIBI_DEPRECATED("use unibi_var_from_num or unibi_num_from_var instead");
^
All use should go through unibi_{num,str}_from_var and
unibi_var_from_{num,str}. Wrap access of unibi_var_t behind a new
UNIBI_SET_NUM_VAR macro which uses the new functions when they're
available.
BSD Make will give preference to a BSDmakefile in the same directory
over a generic Makefile; this can be used to instruct BSD users to build
neovim with GNU Make (gmake) instead.
Otherwise, a flood of syntax errors stemming from the GNU-specific
Makefile will be displayed - which most BSD users are accustomed to, but
may confuse beginners nevertheless.
vim-patch:8.0.0300
Problem: Cannot stop diffing hidden buffers. (Daniel Hahler)
Solution: When using :diffoff! make the whole list if diffed buffers empty.
(closesvim/vim#736)
25ea054458
vim-patch:8.0.0291
Problem: Visual block insertion does not insert in all lines.
Solution: Don't bail out of insert too early. Add a test. (Christian
Brabandt, closesvim/vim#1290)
23fa81d222
vim-patch:8.0.0282
Problem: When doing a Visual selection and using "I" to go to insert mode,
CTRL-O needs to be used twice to go to Normal mode.
(Coacher)
Solution: Check for the return value of edit(). (Christian Brabandt,
closes#1290)
0b5c93a7f2
Workaround for travis issue:
https://github.com/travis-ci/travis-ci/issues/8363
Sometimes `pip3` works, sometimes not:
pyenv: pip3: command not found
The `pip3' command exists in these Python versions:
3.5
3.5.3
Tried these steps to fix the issue:
- add `python: 3.6` to top level of `.travis.yml`
- add `python3` to `addons.apt.packages` level of `.travis.yml`
- `pyenv global system 3.{4,5,6}`
- `pyenv global 3.6`
In all cases the presence or absence of `pip3` was random.
Always check for the presence of pyenv_root if pyenv is installed: if it
is not set, we don't know if it was intentional. If it wasn't
intentional, the warning is confusing (see #7176).
closes#7176
normal_redraw() usually takes care of this, but that doesn't happen
during terminal-mode.
regression by c484323dc6
steps to reproduce:
nvim -u NORC --cmd 'execute("set titlestring=" . $NVIM_LISTEN_ADDRESS) | set title | startinsert | !sleep 1' term://sh
closes#7248
vim-patch:8.0.0181
Problem: When 'cursorbind' and 'cursorcolumn' are both on, the column
highlignt in non-current windows is wrong.
Solution: Add validate_cursor(). (Masanori Misono, closesvim/vim#1372)
519d7785f4
vim-patch:8.0.0182
Problem: When 'cursorbind' and 'cursorline' are set, but 'cursorcolumn' is
not, then the cursor line highlighting is not updated. (Hirohito
Higashi)
Solution: Call redraw_later() with NOT_VALID.
e47683a091
vim-patch:8.0.0188
Problem: Using NOT_VALID for redraw_later() to update the cursor
line/column highlighting is not efficient.
Solution: Call validate_cursor() when 'cul' or 'cuc' is set.
9506cad7a1
main_loop.fast_events does not manifest as K_EVENT, because it is
processed at a different stage than main_loop.events. In order to queue
into main_loop.events, we need to go through the threadsafe
loop_schedule(), which queues into main_loop.thread_events and
eventually main_loop.fast_events. _Then_ it is safe to directly queue
into main_loop.events.
This makes it more likely that the event is treated as K_EVENT.
The "mapping" tests added in 541dde36e3 were flawed:
- Unlike op-pending mode, RPCs are _blocked_ during map-pending. So
a synchronous RPC like nvim_get_current_buf() waits until
'timeoutlen', then the mapping is canceled.
- helpers.expect() also performs a blocking RPC, so again that must not
intervene the two nvim_input() calls.
closes#6166
normal_finish_command() and normal_prepare() assume that any pending
operator needs to be finished after any subsequent key.
Set `finish_op = false` in nv_event() to indicate that the pending
operator shouldn't be finished in normal_execute().
This is how nv_visual() indicates that 'v' or 'V' in operator-pending
mode should not finish the current pending operator.
fixes#5398fixes#6166 (partially; mappings are still interrupted)
Problem: Under some circumstances, one needs to type Ctrl-N or Ctrl-P twice
to have a menu entry selected. (Lifepillar)
Solution: call ins_compl_free(). (Christian Brabandt, closesvim/vim#1411)
aed6d0b81a
Problem: When using :substitute with the "c" flag and 'cursorbind' is set
the cursor is not updated in other windows.
Solution: Call do_check_cursorbind(). (Masanori Misono)
41baa7983a
If an autoloaded function hasn't been resolved before it is used in
function(), the self dict will not be created which causes E725 when
calling the function. Since self isn't being used in
provider#stderr_collector, we can remove the dict attribute to
workaround the self dict bug[0].
Closes#7115
[0]: https://groups.google.com/d/msg/vim_dev/I7AXOyv-P4o/DzbyOxDHBgAJ
iTerm2 got its own entry in Thomas Dickey's terminfo.src on 2017-08-16.
Make sure that the new entry is handled in the same way as the old entry.
closes#7209closes#7214
When using an index returned by unibi_add_ext_str() we should
always use unibi_get_ext_str() and not rely on the index being
lower than unibi_string_begin_.
Closes#7206
- Show hint only once per session.
- provider#clipboard#Call(): prevent recursion
- provider#clear_stderr(): use has_key(), because :silent! is still
captured by :redir.
closes#7184
- :hi Normal works with -u NONE
- Makes HL_TABLE and ATTR_ENTYRY a function instead of a macro so that in can be used in gdb.
- Introduces ATTRENTRY_INIT to init attrentry_t
Problem: Plugins in start packages are sourced twice. (mseplowitz)
Solution: Use the unmodified runtime path when loading plugins (test by Ingo
Karkat, closesvim/vim#1801)
07ecfa64a1
Problem: Package directories are added to 'runtimepath' only after loading
non-package plugins.
Solution: Split off the code to add package directories to 'runtimepath'.
(Ingo Karkat, closesvim/vim#1680)
ce876aaa9a
Problem: The packadd test does not create the symlink correctly and does
not test the right thing.
Solution: Create the directory and symlink correctly.
644df41c44
Problem: When using a symbolic link, the package path will not be inserted
at the right position in 'runtimepath'. (Dugan Chen, Norio Takagi)
Solution: Resolve symbolic links when finding the right position in
'runtimepath'. (Hirohito Higashi)
2f9e575583
Problem: C indenting does not support nested namespaces that C++ 17 has.
Solution: Add check that passes double colon inside a name. (Pauli, closes
vim/vim#1214)
ca8b8d6956
start_batch_changes() doesn't avoid invoking the clipboard
once-per-line, because the loop is actually in ex_echo(), which calls
redir_write() for each message. But we've already entered
start_batch_changes() by then, so that was never the problem.
redir_write at /home/vagrant/old.neovim/build/../src/nvim/message.c:2523
msg_puts_attr_len at /home/vagrant/old.neovim/build/../src/nvim/message.c:1600
msg_outtrans_len_attr at /home/vagrant/old.neovim/build/../src/nvim/message.c:1221
ex_echo at /home/vagrant/old.neovim/build/../src/nvim/eval.c:19433
do_one_cmd at /home/vagrant/old.neovim/build/../src/nvim/ex_docmd.c:2242
Trying to defer _explicit_ clipboard updates is difficult.
:redir @+ | silent echo system('cat foo') | redir END
is essentially equivalent to:
for l in readfile('foo')
let @+ .= l
endfor
We cannot make judgements about when to ignore a script's bad decisions.
start_batch_changes() only works around the case of clipboard=unnamed,
i.e. _implicit_ clipboard updates (`:g/foo/d`). Not explicit
assignment.
redir_write():
- This is a "batch" operation which was not yet covered by
start_batch_changes()
adjust_clipboard_name():
- msg() and friends during :redir will, of course, cause redir_write()
to try to capture that message, which causes recursion.
- EMSG() here is trouble: if it interrupts :redir it is a mess.
Rather than deal with the mess, show a non-error message.
closes#7182closes#7184closes#7183
ref #6048
ref #7032
Group some options, and sort them alphabetically.
`nvim -h` should fit on one (smallish) screen.
Uncommon options don't need to be here, they live in the :help.
Closes#6748
Problem: When 'equalalways' is set and closing a window in a separate
frame, not all window sizes are adjusted. (Glacambre)
Solution: Resize all windows if the new current window is not in the same
frame as the closed window. (closesvim/vim#1707)
8eeeba8c02
To deal with SIGWINCH limitations on Windows, change some resize tests
to _shrink_ the screen width. ... But this didn't work, so still
ignoring those tests on Windows.
- Make sure that proc->in is not NULL, because nvim crashed when
starting a job with pty.
- Make sure that proc->out is not NULL, because nvim crashed when stopping
a job opened with pty.
Handling of process exit is still broken. It detects the moment when the
child process exits, then quickly stops polling for process output. It
should continue polling for output until the agent has scraped all of the
process' output. This problem is easy to notice by running a command like
"dir && exit", but even typing "exit<ENTER>" can manifest the problem --
the "t" might not appear.
winpty's Cygwin adapter handles shutdown by waiting for the agent to close
the CONOUT pipe, which it does after it has scraped the child's last
output. AFAIK, neovim doesn't do anything interesting when winpty closes
the CONOUT pipe.
Closes#7086
Problem: Storing a zero byte from a multi-byte character causes fold text
to show up wrong.
Solution: Avoid putting zero in ScreenLines. (Christian Brabandt,
closesvim/vim#1567)
c6cd8409c2
vim-patch:8.0.0290: cursor positioning wrong if wide character wraps
Problem: If a wide character doesn't fit at the end of the screen line, and
the line doesn't fit on the screen, then the cursor position may
be wrong. (anliting)
Solution: Don't skip over wide character. (Christian Brabandt, closes vim/1408)
vim-patch:8.0.0394
Problem: Tabs are not aligned when scrolling horizontally and a Tab doesn't
fit. (Axel Bender)
Solution: Handle a Tab as a not fitting character. (Christian Brabandt)
Also fix that ":redraw" does not scroll horizontally to show the
cursor. And fix the test that depended on the old behavior.
abc39ab642
Problem: Linebreak tests are old style.
Solution: Turn the tests into new style. Share utility functions. (Ozaki
Kiichi, closesvim/vim#1444)
544d3bc9f0
Problem: Display problem with 'foldcolumn' and a wide character.
(esiegerman)
Solution: Don't use "extra" but an allocated buffer. (Christian Brabandt,
closesvim/vim#1310)
6270660611
Closes#7086
Problem: Storing a zero byte from a multi-byte character causes fold text
to show up wrong.
Solution: Avoid putting zero in ScreenLines. (Christian Brabandt,
closesvim/vim#1567)
c6cd8409c2
vim-patch:8.0.0290: cursor positioning wrong if wide character wraps
Problem: If a wide character doesn't fit at the end of the screen line, and
the line doesn't fit on the screen, then the cursor position may
be wrong. (anliting)
Solution: Don't skip over wide character. (Christian Brabandt, closes vim/1408)
vim-patch:8.0.0394
Problem: Tabs are not aligned when scrolling horizontally and a Tab doesn't
fit. (Axel Bender)
Solution: Handle a Tab as a not fitting character. (Christian Brabandt)
Also fix that ":redraw" does not scroll horizontally to show the
cursor. And fix the test that depended on the old behavior.
abc39ab642
Problem: Linebreak tests are old style.
Solution: Turn the tests into new style. Share utility functions. (Ozaki
Kiichi, closesvim/vim#1444)
544d3bc9f0
Problem: Display problem with 'foldcolumn' and a wide character.
(esiegerman)
Solution: Don't use "extra" but an allocated buffer. (Christian Brabandt,
closesvim/vim#1310)
6270660611
32396b5879 add length checks to
TERMINAL_FAMILY/STARTS_WITH to ensure memcmp() wouldn't read past the
end of the string. However, "term" was copy/pasted from TERMINAL_FAMILY
so STARTS_WITH() was unnecessarily reading the, potentially NULL, term
variable.
Problem: For completion "locale -a" is executed on MS-Windows, even though
it most likely won't work.
Solution: Skip executing "locale -a" on MS-Windows. (Ken Takata)
b8f7bd68f6
Closes#4370
Explication:
In the backtrace in #4370, we see that `buf_write()` was called with
non-NULL `fname` and `sfname` arguments, but they've since _become_
NULL.
#7 0x00000000004de09d in buf_write (buf=0x1dee040, fname=0x0, fname@entry=0x1e985b0 "/home/sean/src/github.com/snczl/virta/pkg/meld/segment.go",
sfname=0x0, sfname@entry=0x1ddfa60 "segment.go", start=1, end=72, eap=eap@entry=0x7ffc6b032e60, append=0,
forceit=0, reset_changed=1, filtering=0)
at /home/travis/build/neovim/bot-ci/build/neovim/src/nvim/fileio.c:2576
This is most likely due to the code that restores those values from
`buf`, which happens just before the fatal call to `os_fileinfo`
```c
/*
* The autocommands may have changed the name of the buffer, which may
* be kept in fname, ffname and sfname.
*/
if (buf_ffname)
ffname = buf->b_ffname;
if (buf_sfname)
sfname = buf->b_sfname;
if (buf_fname_f)
fname = buf->b_ffname;
if (buf_fname_s)
fname = buf->b_sfname;
```
It's worth noting that at this point `ffname` is still non-NULL, so
it _could_ be used. However, our current code is purely more strict
than Vim in this area, which has caused us problems before (e.g.,
`getdigits()`). The commentary for `struct file_buffer` clearly
indicate that all of `b_ffname`, `b_sfname`, and `b_fname` may be
NULL:
```c
/*
* b_ffname has the full path of the file (NULL for no name).
* b_sfname is the name as the user typed it (or NULL).
* b_fname is the same as b_sfname, unless ":cd" has been done,
* then it is the same as b_ffname (NULL for no name).
*/
char_u *b_ffname; /* full path file name */
char_u *b_sfname; /* short file name */
char_u *b_fname; /* current file name */
```
Vim directly calls `stat(2)` which, although it is annotated to tell
the compiler that the path argument is non-NULL, does handle a NULL
pointer by returning a `-1` value and setting `errno` to `EFAULT`.
This satisfies Vim's check, since it treats any `-1` return from
`stat(2)` to mean the file doesn't exist (at least in this code
path).
Note that Vim's mch_stat() implementations on win32 and solaris
clearly cannot accept NULL `name`. But the codepaths that call
mch_stat will NULL `name` tend to be unix-only (eg: u_read_undo)!
Problem: ":earlier" and ":later" do not work after startup or reading the
undo file.
Solution: Use absolute time stamps instead of relative to the Vim start
time. (Christian Brabandt, Pavel Juhas, closesvim/vim#1300, closes
vim/vim#1254)
cbd4de44e8
Problem: searchpair() does not work when 'magic' is off. (Chris Paul)
Solution: Add \m in the pattern. (Christian Brabandt, closesvim/vim#1341)
6e450a5754
Problem: When 'maxfuncdepth' is set above 200 the nesting is limited to
200. (Brett Stahlman)
Solution: Allow for Ex command recursion depending on 'maxfuncdepth'.
777b30f827
Unlike the normal wildmenu, the CTRL-D wild-list is not restored by
statusline redraw. (Semantics: ^D is controlled by 'wildoptions' option,
so it's in the "wild..." family.)
TODO: externalize the c_CTRL-D wild-list.
This might be too coarse, but it passes all tests ...
A more nuanced approach might be: only skip the windows whose
statuslines are overwritten by the wildmenu.
Closes#2255Closes#7108
vim-patch:8.0.0710 N/A because of the changes in this commit.
Problem: Null pointer access reported by UBsan.
Solution: Check curwin->w_buffer is not NULL. (Yegappan Lakshmanan)
c4bfedabe0
src/nvim/ex_cmds.c have been merged by someone. while version.c remain unmerged.
This reverts commit eb40b7ec40.
The change caused this error on QuickBuild:
INFO - # test/functional/core/job_spec.lua @ 668: pty process teardown does not prevent/delay exit. #4798#4900
INFO - not ok 321 - pty process teardown does not prevent/delay exit. #4798#4900
INFO - # test/functional/core/job_spec.lua @ 668
INFO - # Failure message: ./test/functional/ui/screen.lua:302: Row 1 did not match.
INFO - # Expected:
INFO - # |* |
INFO - # |[Process exited 0] |
INFO - # | |
INFO - # | |
INFO - # | |
INFO - # |-- TERMINAL -- |
INFO - # Actual:
INFO - # |*E575: Error while reading ShaD|
INFO - # |a file: mark entry at position|
INFO - # | 92 has invalid line number |
INFO - # |Press ENTER or type command to|
INFO - # | continue |
INFO - # |-- TERMINAL -- |
INFO - #
INFO - # To print the expect() call that would assert the current screen state, use
INFO - # screen:snaphot_util(). In case of non-deterministic failures, use
INFO - # screen:redraw_debug() to show all intermediate screen states.
INFO - # stack traceback:
INFO - # ./test/functional/ui/screen.lua:302: in function 'wait'
INFO - # ./test/functional/ui/screen.lua:216: in function 'expect'
INFO - # test/functional/core/job_spec.lua:677: in function <test/functional/core/job_spec.lua:668>
- Prefer "TUI" where possible to refer to the host terminal.
- Remove obsolete tags and ancient TTY exposition.
- Establish "terminal" to consistently mean "terminal emulator" in all
Nvim documentation. This removes the need for verbose qualifiers in
tags and prose.
References #6280
References #6803
Problem: Two autocmd tests are skipped on MS-Windows.
Solution: Make the test pass on MS-Windows. Write the messages in a file
instead of getting the output of system().
e94260f358
Certain functions (e.g. feedkeys(…, 'x!') and input()) will attempt to
read from stdin, which shouldn't be expected to work during oldtests.
In Debian/Ubuntu's build environment, it explicitly can't work because
/dev/null is redirected to stdin, which causes read_error_exit() to
exit.
Running oldtests with --headless prevents nvim from setting up its input
handling, thus avoiding the problem altogether.
Reference #6794
vim-patch:8.0.0012
Problem: Typos in comments.
Solution: Change "its" to "it's". (Matthew Brener, closesvim/vim#1088)
9af4184276
vim-patch:8.0.0046
version.c: mark 8.0.0046 applied
vim-patch:8.0.0063
version.c: mark 8.0.0063 as NA patch
This delete() sometimes hangs the TSAN build. Work around it by using
a unique filename. Do it at the start instead of the end, for hygiene
(though it doesn't actually matter on CI, it helps local dev).
This TAP.lua implementation was upstreamed:
f9db364291
but it has not been released yet. It avoids hangs, and helps debug
hangs, on QuickBuild, so inline it until the next busted release.
Tests that check localized error messages need a stable locale, else
errors like this occur:
[ FAILED ] 2 tests, listed below:
[ FAILED ] ...npack/file/vim/neovim/test/functional/eval/null_spec.lua @ 29: NULL list is accepted as an empty list by writefile()
...npack/file/vim/neovim/test/functional/eval/null_spec.lua:30: Expected objects to be the same.
Passed in:
(string) '
E484: Cannot open file Xtest-functional-viml-null'
Expected:
(string) '
E484: Can't open file Xtest-functional-viml-null'
stack traceback:
...npack/file/vim/neovim/test/functional/eval/null_spec.lua:30: in function <...npack/file/vim/neovim/test/functional/eval/null_spec.lua:29>
[ FAILED ] ...k/file/vim/neovim/test/functional/ex_cmds/write_spec.lua @ 81: :write errors out correctly
...k/file/vim/neovim/test/functional/ex_cmds/write_spec.lua:97: Expected objects to be the same.
Passed in:
(string) 'Vim(write):E510: Cannot make backup file (add ! to override)'
Expected:
(string) 'Vim(write):E510: Can't make backup file (add ! to override)'
stack traceback:
...k/file/vim/neovim/test/functional/ex_cmds/write_spec.lua:97: in function <...k/file/vim/neovim/test/functional/ex_cmds/write_spec.lua:81>
10 SKIPPED TESTS
2 FAILED TESTS
-- Output to stderr:
2017/07/29 00:41:32 ERROR 31133/open_log_file:170: Logging to stderr, failed to open $NVIM_LOG_FILE: Xtest-startup-xdg-logpath/nvim/log
2017/07/29 00:41:32 WARN 31133/call_set_error:815: RPC: ch 1 was closed by the client
CMake Error at /home/shlomif/Download/unpack/file/vim/neovim/cmake/RunTests.cmake:50 (message):
Running functional tests failed with error: 1.
FAILED: CMakeFiles/functionaltest
cd /home/shlomif/Download/unpack/file/vim/neovim/build && /usr/bin/cmake -DBUSTED_PRG=/home/shlomif/Download/unpack/file/vim/neovim/.deps/usr/bin/busted -DLUA_PRG=/home/shlomif/Download/unpack/file/vim/neovim/.deps/usr/bin/luajit -DNVIM_PRG=/home/shlomif/Download/unpack/file/vim/neovim/build/bin/nvim -DWORKING_DIR=/home/shlomif/Download/unpack/file/vim/neovim -DBUSTED_OUTPUT_TYPE=nvim -DTEST_DIR=/home/shlomif/Download/unpack/file/vim/neovim/test -DBUILD_DIR=/home/shlomif/Download/unpack/file/vim/neovim/build -DTEST_TYPE=functional -DSYSTEM_NAME=Linux -P /home/shlomif/Download/unpack/file/vim/neovim/cmake/RunTests.cmake
ninja: build stopped: subcommand failed.
Makefile:102: recipe for target 'functionaltest' failed
make: *** [functionaltest] Error 1
Coverity warning is a false positive: if rbuffer_read_ptr() returns
NULL then `cnt` is zero.
Revert 76ea97c809 (which caused
the TSan build to hang often--possibly because of the missing ui_flush()).
Instead, modify out_data_append_to_screen() to check for NULL.
ref #6862
With `xdg-*` utilities CMAKE_INSTALL_PREFIX is incorrectly ignored.
Taken from [nvim-qt][1]. For some reason it only checks for !APPLE.
[1]: b26596d164/src/gui/CMakeLists.txt (L48-L55)
Problem: OPEN_CHR_FILES not defined for FreeBSD using Debian userland
files.
Solution: Check for __FreeBSD_kernel__. (James McCoy, closesvim/vim#1166)
ca291aec99
Problem: getwinvar() returns wrong Value of boolean and number options,
especially non big endian systems. (James McCoy)
Solution: Cast the pointer to long or int. (closesvim/vim#1060)
789a5c0e3d
17:25:45,363 WARN - /home/quickbuild/buildagent/workspace/root/neovim/pull-requests-automated/src/nvim/ex_getln.c: In function ‘color_cmdline’:
17:25:45,363 WARN - /home/quickbuild/buildagent/workspace/root/neovim/pull-requests-automated/src/nvim/ex_getln.c:2335:8: error: variable ‘printed_errmsg’ set but not used [-Werror=unused-but-set-variable]
17:25:45,363 WARN - bool printed_errmsg = false;
17:25:45,363 WARN - ^
17:25:45,399 WARN - cc1: all warnings being treated as errors
Problem: A string argument for function() that is not a function name
results in an error message with NULL. (Christian Brabandt)
Solution: Use the argument for the error message.
5582ef1438
The issue with debug mode was actually not cleaning up after `try_enter`:
location `&tstate` was pointing to got invalidated and received some “garbage”
(actually, values that got stored on the stack afterwards). But pointer to that
garbage was still stored in `msg_list`, so next attempt to check it resulted in
a crash.
Problem: Cannot get the number of the current quickfix or location list.
Solution: Use the current list if "nr" in "what" is zero. (Yegappan
Lakshmanan) Remove debug command from test.
890680ca63
Several people have suggested that the "by Bram" byline is misleading,
it implies that Bram is actively involved with the project. Up to now we
left it as an homage.
Bram agreed that it is misleading, and suggested a mention somewhere
other than the intro.
helpers.skip_fragile() already skips the problematic tests
on the ASan build. But the 15s timeout plus 5s 'mousetime'
cause the tests to take 1+ minutes anyways.
f04507d132
Author: Bram Moolenaar <Bram@vim.org>
Date: Sat Aug 20 15:05:39 2016 +0200
patch 7.4.2229
Problem: Startup test fails on Solaris.
Solution: Recognize a character device. (Danek Duvall)
f71d7b9ee5
Author: Bram Moolenaar <Bram@vim.org>
Date: Tue Aug 9 22:14:05 2016 +0200
patch 7.4.2189
Problem: Cannot detect encoding in a fifo.
Solution: Extend the stdin way of detecting encoding to fifo. Add a test
for detecting encoding on stdin and fifo. (Ken Takata)
Problem: The example that explains nested backreferences does not work
properly with the new regexp engine. (Harm te Hennepe)
Solution: Also save the end position when adding a state. (closesvim/vim#990)
d563883a1f
The libvte test was too agressive, and is reduced to only triggering
when it is libvte 0.36 AND a gnome or xterm terminal type is used.
Contrastingly, tmux was not on the list at all and now is.
During a preview, we can stop looking for matches after we got enough
lines for the preview buffer.
Because of this perf improvement, the 'redrawtime' test needs to be
slowed down in a different way: _long_ lines instead of just many lines.
The gchar_cursor() == NUL check is already done in ins_ctrl_o.
ins_esc changes gchar_cursor() so this if block is probably never
entered.
Issue:
Pressing CTRL-O in insert mode at the end of the line and typing
:startinsert moves the cursor 1 column back, when I expect the cursor
to remain at the end of the line
This is a regression from Vim behavior. Since at least Vim version 7.0,
Vim returns you to insert mode at the end of the line.
091e7d033c is the first bad neovim commit
Steps to reproduce using `nvim -u NORC`:
`aaaa<C-o>:startinsert<CR>`
Fixes#6962
Update a flawed match pattern for the vimCommand syntax group. To see
the effect of this fix, open a vimscript buffer,
nvim -u NONE foo.vim
configure a couple highlight groups,
:hi! vimIsCommand ctermfg=Green
:hi! vimCommand ctermfg=Red
:syntax enable
and add the following lines to the buffer:
let foo=xFoo
let bar=zBar
You'll notice the "z" in zBar is Red, while xFoo and the rest of Bar are green. This will
be the case as long as the word following `=` starts with the letter "z". This has already
been fixed upstream by adding a "\>" word boundary to the match pattern:
https://github.com/vim/vim/issues/124e271909625 (diff-86da060e2153c8ce5dc317a7b4b5a29dR27)
This particular match pattern was also mentioned in issue #5491, but in reference to a bug
that was related to the generated part of syntax/vim.vim, whereas this bug lives in the
non-generated part of the file.
The terminfo doco explicitly states that it covers the case where the
terminal is in application cursor/keypad (i.e. "keypad transmit") mode,
and not where it is in normal cursor/keypad (i.e. "keypad local") mode.
Full screen applications like nvim must switch to and from keypad
transmit mode when expecting the control sequences given by terminfo.
This is the result of malloc error handling elimination: push_current_state()
used to (not) return OK depending on whether growing garray failed or not and
this return was checked, if errorred out push_next_match() will simply return
its argument unchanged.
Now when allocations are supposed to either always succeed or crash Neovim this
check was returned, push_current_state() was stripped of its return value and
moved out of if() condition, resulting in V763.
Running this test with a mocked passwd file whose $HOME was set to
/home/jamessan/src/debian.org/pkg-vim/deb-packages/neovim/neovim-0.2.0/debian/fakehome
caused the test to fail, since the expanded result was >= 99 bytes. The
test should be reflecting the actual size of the buffer, instead of some
arbitrary other number, anwyay.
Problem: when processing cycle such as
:for pat in [' \ze*', ' \zs*']
: try
: let l = matchlist('x x', pat)
: $put ='E888 NOT detected for ' . pat
: catch
: $put ='E888 detected for ' . pat
: endtry
:endfor
`:let l = …` throwing an error causes this error to be caught after
color_cmdline attempts to get callback for highlighting next line (the one with
`$put = 'E888 NOT…`). Saving/restoring state prevents this from happening.
The main purpose of this build-type was to avoid unwanted ~/.nvimlog
files (which could get really big, and also affects performance) for
non-devs. But that is no longer necessary since the log system now
avoids non-critical logging by default (#6827).
This essentially reverts 87e5a41316
This also attempted to fix problem with cancelling input() on error by avoiding
standard error printing facilities (assumed thrown error message is the
problem), but with no luck so far.
Problem: Redraw problem when using 'incsearch'.
Solution: Save the current view when deleting characters. (Christian
Brabandt) Fix that the '" mark is set in the wrong position. Don't
change the search start when using BS.
dda933d06c
Should speed up execution without arabic characters a bit, slowing down with
arabic characters. More necessary, this allows coloring prompt without caring
about arabic shaping at the first iteration.
if (oap->regname == 0 &&
oap->motion_type != kMTLineWise &&
oap->line_count == 0 &&
!oap->use_reg_one ){
then reg is not initialised
and our call to set_clipboard will dereference NULL
}
rbuffer_read_ptr may return a null
if ptr == null && cnt == 0 && !out_data_decide_throttle(cnt)
then we would have called out_data_append_to_screen(ptr, cnt, eof)
which dereferences the null pointer.
This allows users who have per-project Ruby versions (e.g. with `rvm`)
to pin to a particular gem installation.
For example: `let g:ruby_host_prog = 'rvm system do neovim-ruby-host'`
Sometimes the `gem list` command used for finding the latest version of
the `neovim` gem prints an error, which can throw off the `split()` call
due to extra parenthesis. This locks down the split pattern to make
conflicts less likely.
* Update libuv to 1.12.0
* Update msgpack-c to 2.13
* Update LuaJIT to 2.0.5 (82151a45 for a Windows build fix)
* Update jemalloc to 4.5.0
* Update luv to 1.9.1-1
* Update gperf to 3.1
Problem: When startup test fails it's not easy to find out why.
GUI test fails with Gnome.
Solution: Add the help entry matches to a list an assert that.
Set $HOME for Gnome to create .gnome2 directory.
50fa8dd00c
This was added from the beginning to submit_pr in 775a16b0b, but I
cannot see why that is useful - in contrast, it will mess with the local
branch in case "origin" cannot be pushed to (i.e. when it points to
neovim/neovim itself).
The ':tcd' command is the first tab-specific command written to the file
and it is wrapped inside an 'if has('nvim')' block to keep the session
file compatible with Vim.
Closes#6678
Problem: When creating a bufref, then using :bwipe and :new it might get
the same memory and bufref_valid() returns true.
Solution: Add br_fnum to check the buffer number didn't change.
45e5fd135d
This fixes a test failure caused by dfaecb25f6a9a94f29a38d9f2d24a579b3dff5f
not tracking what the current visibility is and whether it matches the
current business state.
When higher layers flush the TUI layer output buffer, but there is
nothing in the buffer to flush, no longer does the TUI layer write out
unnecessary cnorm/civis sequences surrounding that nothing.
In order to generate INT64_MIN from literal values, it's necessary to
use "-0x7fffffffffffffff - 1". Using "-0x8000000000000000" causes the
value to get clamped to INT64_MAX and then negated.
If if the resolved $NVIM_LOG_FILE *and* stdpath("data")/log cannot be
created (e.g. because the XDG data directory does not exist), fall back
to .nvimlog in the current direcrtory.
- Do not delete it: may need to inspect it after tests finished.
- Avoids writing to stderr in cases where the test-local $XDG_DATA_HOME
was not created yet.
- Establish ERROR log level as "critical". Such errors are rare and will
be valuable when users encounter unusual circumstances.
- Set -DMIN_LOG_LEVEL=3 for release-type builds
Problem: float2nr() is not exactly right.
Solution: Make float2nr() more accurate. Turn test64 into a new style test.
(Hirohito Higashi, closesvim/vim#1688)
863e80b445
Problem: Using :change in between :function and :endfunction fails.
Solution: Recognize :change inside a function. (ichizok, closesvim/vim#1374)
70bcd7336f
Problem: str2nr() and str2float() do not always work with negative values.
Solution: Be more flexible about handling signs. (LemonBoy, closesvim/vim#1332)
Add more tests.
08243d26d2
Problem: On MS-Windows some float functions return a different value when
passed unusual values. strtod() doesn't work for "inf" and "nan".
Solution: Accept both results. Fix str2float() for MS-Windows. Also
reorder assert function arguments.
6247361101
This is a new convention pioneered by tmux. It does not do much for
nvim; since nvim always looks to see whether it should be making up
"setrgbf" and "setrgbb" capabilities. But it is a way for terminfo to
force this, irrespective of the hardwired list in the code, for more
terminal types. On the gripping hand, updating terminfo descriptions to
actually have "setrgbf" and "setrgbb" capabilities so that nvim never
has to try to invent them in the first place, is as good if not better
an approach for overriding what is baked into the code.
After merging +num64, the 64-bit sanitizer builds show that Vim doesn't
buffer the user from C's UB in signed arithmetic. Upstream doesn't
appear to be [interested] in fixing the issue, so suppress UBSAN until
someone decides to fix the problem.
N.B., the problem existed before but went unnoticed since the sanitizer
builds weren't being run in 32-bit mode.
[interested]: https://groups.google.com/d/topic/vim_dev/_tqf8eQy5eA/discussion
Problem: Ubsan reports errors for integer overflow.
Solution: Define macros for minimum and maximum values. Select an
expression based on the value. (Mike Williams)
7a40ea2138
Problem: Compiler warnings with older compiler and 64 bit numbers.
Solution: Add "LL" to large values. (Mike Williams)
af9c4c9b57
Equivalent change was made in ZyX's typval refactoring.
Problem: Compiler warns for loss of data.
Solution: Use size_t instead of int. (Christian Brabandt)
fef524bbff
Equivalent change had already been made when merging earlier pack
patches.
Problem: With 64 bit changes don't need three calls to sprintf().
Solution: Simplify the code, use vim_snprintf(). (Ken Takata)
bde9810d61
nvim already had the equivalent code, so only the patch number was
needed.
Problem: On MS-Windows large files (> 2Gbyte) cause problems.
Solution: Use "off_T" instead of "off_t". Use "stat_T" instead of "struct
stat". Use 64 bit system functions if available. (Ken Takata)
8767f52fbf
Only the off_T changes are relevant, since all the "struct stat" usage
is abstracted by libuv.
From observation, there are several different possible behaviours:
1. Deferred wrap like a real DEC VT. The cursor stays visible in the last
column, and CUB is calculated relative to that column.
Examples: xterm, Unicode rxvt, PuTTY, nosh console-terminal-emulator,
FreeBSD kernel's built-in emulator, Linux's built-in emulator
2. Deferred wrap like a real DEC VT. CUB is calculated relative to the last
column. But the cursor is invisible.
Examples: emulators using newer libvte
3. Non-deferred wrap. The cursor has already wrapped to the next line and CUB
does not wrap back.
Examples: cygwin, Interix
4. Non-deferred wrap that acts like deferred wrap. The cursor has already
visibly wrapped to the next line, but CUB can wrap back around the left
margin.
Examples: Konsole
5. Deferred wrap with visibly out of bounds cursor. The cursor visibly moves
outwith the screen boundaries. CUB is calculated relative to a cursor
column that has overflowed the end of the screen grid array.
Examples: iTerm2
6. Deferred wrap with invisibly out of bounds cursor. CUB is calculated
relative to a cursor column that has overflowed the end of the screen grid
array. And the cursor is invisible.
Examples: emulators using older libvte
In many cases, nvim does not have enough information to know which behaviour
the terminal will exhibit, and thus the correct amount of CUB to issue.
Partly undo 8ab08a65ba3bc9a44741a2ec9aa81fbcc77467fb. Further testing
by Enrico Ghirardi suggests limiting the non-deferred automatic wrap to
only the bottom line, whose rightmost column is not printed for iTerm.
They are now in their own nvim/tui/terminfo.c file.
Also turn the TERMINAL_FAMILY macro into a function. Use the terminfo_
prefix for its name as other parts of the program are unlikely to want
that namespace, and the prefix is already used for some other TUI
functions.
Testing by Enrico Ghirardi and review of the source indicates that
iTerm2 is a second terminal emulator that does not defer automatic wrap
at the right margin.
The example used &term which is no longer meaningful.
Fortunately, we can change this into a useful example using $TERM that also
shows how to address a common need with termguicolors at the same time.
PM...ST actually sends the string to screen's message area. Sending the
string to the status line requires a different control sequence peculiar to
screen.
Also make iTerm2 SGR 38/48 consistent.
The Interix termcap entry is missing the carriage_return capability which nvim
relies upon. And Interix is one of the few terminal emulators that does not
defer automatic wrap at the right margin, which is now accounted for when
moving the cursor left and when outputting whole lines at a time.
tmux has its own code path, now; and the tmux wrapping was not the ideal thing
to do in the first place.
Also improve the commentary on the built-in terminfo records.
The details are in the on-line help under :help true-color .
The brief precis is that nvim is (I hope.) converging with tmux and libvte.
It is taking the same approach with setrgbf and setrgbb terminfo capabilities
that it does with the Ss and Se terminfo capabilities.
The details are in the on-line help under :help cursor-shape .
The brief precis is that nvim is following the lead of tmux, and going
beyond what tmux does to make cursor shape changes work on a broad range of
terminals. This includes on tmux itself, which is no longer bypassed.
Ironically, higher layers trying to be "smart" about the terminal type
but not actually being very smart at all, makes it more difficult rather
than less to correct the TUI layer.
Note that this orphans the os_term_is_nice() function and down the road,
presuming that we do not have to revert this, that function can be removed.
It incorporates knowledge of terminal types and behaviours in the wrong place.
There are now a few built-in terminfo entries, taken either from unibilium
or ncurses terminfo, for falling back upon when there is no terminfo database
or when it is missing stuff. In an ideal world, these would be in unibilium
itself.
The ultimate fallback, for no terminfo database and no built-in terminfo
record that matches the terminal type, is now the "ansi" terminal type; so
unknown terminal types are now considered to have at minimum the basic
ECMA-48 colour, motion, and editing capabilities.
The terminfo records are just blobs, raw images of the equivalent terminfo file
created with the od command. No longer are incomplete terminfo records built
up with code. These blobs are the full, real, records; already built.
The post-processing of the terminfo record, once found, is split into the
part where we fix known errors and deficiencies in terminfo, and the part
where we add extensions that we need that terminfo does not define
capabilities for. In an ideal world, the former would be a no-op.
No part of the TUI layer apart from these is aware of terminal type or has
conditional code based upon checking environment variables at runtime. It
is all pre-calculated and written into unibilium (or the TUIData object) at
initialization time.
This is fairly aggressive about turning on 256-colour and true colour support.
This also positively decodes genuine xterm for turning on DECSLRM use, rather
than assuming that anything that says that it is xterm is actually xterm,
fixing scrolling problems with vertically split windows.
This documents 256-colour and true colour handling, cursor shapes,
and scrolling regions.
Almost all of these headings are taken from the Vim doco, so that
the :help commands that people learn are a transferable skill.
PuTTY does not implement DECLRMM or DECSLRM, but it does implement DECSTBM.
So allow using PuTTY terminal scrolling when the scroll rectangle is the
full width of the terminal.
Instead of emitting CUP in several places each with their own poor local
optimizations, funnel all cursor motion through a central place.
This central function performs the same optimization for every place that
needs to move the cursor, and implements a better set of optimizations:
* Emit CUU/CUD/CUF/CUB instad of CUP when they are likely shorter.
* Use BS and LF when they are shorter than CUB and CUD.
* Use CR for quick returns to column zero.
* If printing the next few characters is shorter than a rightwards motion,
then just write out the characters.
LuaRocks 2.3 and onwards changed the /P option to no longer include the
version number which made newer releases of LuaRocks fail when compiling
on Windows.
Reverts commit 337b6179dfCloses#6716 at the expense of not being able to use a
multi-key 'pastetoggle' manually.
Multi-key 'pastetoggle' can still be used when inserting the entire
option into the typebuffer at once (though the use here is
questionable).
Also remove those tests to do with waiting for the completion of
'pastetoggle' and mention in the documentation that 'pastetoggle'
doesn't wait for timeout.
Although in_port_t is a typedef for uint16_t, GCC in Ubuntu 12.04
complains about potential loss of data due to converting int to
uint16_t. Since we know this isn't possible, silence the warning to
avoid breaking QB until it gets upgraded to a newer Ubuntu.
Problem: The buffer that quickfix caches for performance may become
invalid. (Daniel Hahler)
Solution: Reset qf_last_bufref in qf_init_ext(). (Daniel Hahler,
closesvim/vim#1728, closesvim/vim#1676)
6dd4a53502
When using serverstart("ip.ad.d.r:") to listen on a random port, we need
to abide by getaddrinfo()'s API and pass in a NULL service, rather than
an empty string.
When given an empty string, getaddrinfo() is free to search for a
service by the given name (since the string isn't a number) which will
fail. At least FreeBSD does perform this lookup.
Showing the 'number' column in terminal buffers is a bit silly because
of 'scrollback'. But it's mostly harmless and technically works as
expected.
The least surprising thing is to leave the user's settings alone. Since
there are tradeoffs in both cases, we choose inertia.
We still disable 'relativenumber' in *terminal-mode* (as opposed to
normal-mode) because it is totally broken: the Nvim cursor (not terminal
cursor) is always on the last line.
Respect the BGE flag from terminfo rather than guessing that it is
always off. Emit DECLRMM and DECSLRM (or equivalent) to properly define
the scroll rectangle.
DECSLPP is explicitly documented as not affecting the scroll region. The
dtterm extension is not as well documented, but it is safer than not to
assume that it operates similarly.
This also eliminates a pointlessly repeated test from tui_scroll(). It
additionally uses a non-parameterized DECSTBM sequence when attempting
to reset back to whole-screen scrolling.
This limits the use of dtterm's extension to DECSLPP to only those
terminal types where it is known to be supported.
Because it can be potentially understood as genuine DECSLPP
sequence, setting the number of lines to a number larger than 25,
which of course can cause confusion (especially if it is the width
parameter that results in this) only use it on terminals that are
known to support the dtterm extension.
rxvt (Unicode) also understands dtterm's extension.
This change implicitly adds IPv6 support.
If the address contains ":", we try to use a TCP socket instead of a Unix domain
socket. Everything in front of the last occurrence of ":" is the hostname and
everything after it the port.
If the hostname lookup fails, we fall back to using a Unix domain socket.
If the port is empty ("localhost:"), a random port will be assigned.
Examples:
NVIM_LISTEN_ADDRESS=localhost:12345 -> TCP (IPv4 or IPv6), port: 12345
NVIM_LISTEN_ADDRESS=localhost: -> TCP (IPv4 or IPv6), port: random (> 1024)
NVIM_LISTEN_ADDRESS=localhost:0 -> TCP (IPv4 or IPv6), port: random (> 1024)
NVIM_LISTEN_ADDRESS=localhost -> Unix domain socket "localhost" in current dir
`blocksize` was checked against UINT_MAX after it was checked against
MAX_SWAP_PAGE_SIZE which makes it always pass the check. Better use
STATIC_ASSERT instead.
AFAIK there is no way NULL can be there, including from the line it points to.
Dunno what analyser was thinking, but dereferencing of `argv[0]` happened just
before `get_number_arg()` call: in `ascii_isdigit()` two lines above. And `idx`
cannot possibly be NULL ever, it comes from `&varname`, this could not ever give
anything, but a valid pointer.
Could not be uninitialized because `func_or_func_caller_profiling` is true only
if `do_profiling` is `YES`, and if `do_profiling` is `YES` then
`script_prof_save()` was called to initialize the variable.
Based on comments it appears that some non-printable characters intended to be
shown as `|x` (0xA0..0xFE) and some as `~x` (0x80..0x9F, 0xFF, excluding
previous). But this never happens because this is being catched by condition `c
>= 0x80` above which makes them be represented as `<A0>`. Since I find this
variant more useful and it additionally is backwards compatible (Vim does the
same thing) I just dropped dead branches.
This should better allow distributing the load among PRs, while getting
critical feedback to the submitter sooner.
First stage runs the ASAN/UBSAN/TSAN since any failures in those are
gating issues.
Second stage runs the rest of the normal builds in parallel.
Remaining stages provide lower priority feedback. The lint build runs
fast locally, so it's better to run that locally than wait on CI. The
coverage build is pretty fickle, so it is only run once all other jobs
are green.
Compile `nvim` executable against Lua if PREFER_LUA=ON.
As the testing library `nvim-test` requires LuaJIT, it is
still compiled against LuaJIT. If LuaJIT is not available,
`nvim-test` is not built.
When the end character in a range matches a different standard range
(e.g., [0-z]), the range would be incorrectly detected as the class of
the end character (CLASS_az).
Instead of using a fallthrough, immediately FAIL when the end character
doesn't match the expected range.
Although GCC now detects possibly unintentional fall through, there
rules around which the comments are detected are rather strict. In
cases where a comment isn't detected, upstream [recommends] using their
fallthrough attribute.
[recommends]: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77817#c11
The terminfo entry for linux only advertises 8 colours, but nvim tries
to make it display 16 colours anyway, resulting in erroneous SGR control
sequences for colours 8 and above. The Linux kernel terminal emulator
itself has actually understood the 256-colour control sequences since
version 4.8 and the 16-colour control sequences since version 4.9. Thus
we apply the same terminfo fixup as we apply for *xterm* and *256*, to
emit the 16-colour and 256-colour control sequences even if terminfo's
setaf and setab do not advertise them.
Fix the following issues according to rubocop:
runtime/autoload/provider/script_host.rb:2:11: C: Prefer single-quoted strings when you don't need string interpolation or special symbols.
require "neovim/ruby_provider"
^^^^^^^^^^^^^^^^^^^^^^
runtime/autoload/provider/script_host.rb:5:5: C: Prefer single-quoted strings when you don't need string interpolation or special symbols.
"Your neovim RubyGem is missing or out of date. " +
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
runtime/autoload/provider/script_host.rb:5:55: C: Use \ instead of + or << to concatenate those strings.
"Your neovim RubyGem is missing or out of date. " +
runtime/autoload/provider/script_host.rb:6:5: C: Prefer single-quoted strings when you don't need string interpolation or special symbols.
"Install the latest version using `gem install neovim`."
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This commit assumes Ruby 2.0.0+.
For CI builds unibilium is provided through msys2 packages, and
libtermkey is built from source in third-party from equalsraf/libtermkey.
In Windows we cannot read terminal input from the stdin file descriptor,
instead use libuv's uv_tty API. It should handle key input and encoding.
The UI suspend is not implemented for Windows, because the
SIGSTP/SIGCONT do not exist in windows. Currently this is a NOOP.
Closes#3902Closes#6640
As part of the refactoring in #5119, some vim_strchr() were changed to
strchr(). However, vim_strchr() behaves differently than strchr() when
c is NUL, returning NULL instead of a pointer to the NUL.
Revert the strchr() calls where it isn't known whether c is NUL, since
this causes a semantic change the surrounding code doesn't expect. In
the case of #6650, this led to a heap overrun.
Closes#6650
Needed as argument list is growing large and I absolutely do not find short
options provided by getopts being particularly readable for unfamiliar commands.
[ci skip]
Should not really change anything, but code should be more efficient by using
more optimized libc functions (memchrsub is not libc, but it uses memchr) in
place of a cycle.
Reasoning is majorly the same: check whether lua has bug or API function has
bug, but on the other side: previous commit is checking whether similar bug when
using API via msgpack RPC, this commit is checking whether another API function
used via lua bindings triggers the same bug. Should additionally give a hint
about which lua code contains a bug.
Removes dead code (enc_utf8, enc_dbcs and has_mbyte now have hardcoded values),
relies on libc implementation being more optimized. Also where previously
negative character just would never be found it is an assertion error now.
Ref #1476
Fixes:
1. Allocate space for the NUL byte.
2. Do not exclude last line from range.
3. Remove code for sandbox: it is handled earlier.
4. Fix index in new_line_transformed when converting NULs to NLs.
5. Always allocate new_line_transformed, but save allocated value.
1. Use `char *` for strings.
2. Add `const` qualifiers.
3. Add attributes and documentation.
4. Handle skipping *inside*.
5. Handle non-heredoc argument also inside: deferring this to the caller is
pointless because all callers need the same thing. Though new ex_lua caller
may live without allocations in this case, allocating nevertheless produces
cleaner code.
6. Note that all callers call script_get with `eap` and `eap->arg`. Thus second
argument is useless in practice: it is one and the same always and can be
reached through the first argument.
Fixes problem introduced by “api: Allow kObjectTypeNil to be zero without
breaking compatibility”: apparently there are clients which use metadata and
there are which aren’t. For the first that commit would not be needed, for the
second that commit misses this critical piece.
Now it checks functions also after every semicolon and closing figure brace,
possibly preceded by whitespaces (tabs and spaces). This should make messing
with declarations in macros not needed.
Reasoning: luajit is not being compiled with sanitizers, lua is. Given that
linking with sanitized libraries requires sanitizers enabled, it is needed to
either compile libnvim-test with sanitizers or link it with lua compiled without
sanitizers. Most easy way to do the latter is just use luajit which is compiled
without sanitizers (as they do not work well with luajit).
Function declarations generator is able to handle properly only the *first*
function definition that is in macros, and only if it is the first entity in the
macros. So msgpack_rpc_from_* was already really a static function, additionally
its attributes were useless. This commit switches to explicit declarations and
makes generated functions static.
Reasoning; currently INTERNAL_CALL is mostly used to determine whether it is
needed to deal with NL-used-as-NUL problem. This code is useful for nvim_… API
calls done from VimL, but not for API calls done from lua, yet lua needs to
supply something as channel_id.
During testing found the following bugs:
1. msgpack-gen.lua script is completely unprepared for Float values either in
return type or in arguments. Specifically:
1. At the time of writing relevant code FLOAT_OBJ did not exist as well as
FLOATING_OBJ, but it would be used by msgpack-gen.lua should return type
be Float. I added FLOATING_OBJ macros later because did not know that
msgpack-gen.lua uses these _OBJ macros, otherwise it would be FLOAT_OBJ.
2. msgpack-gen.lua should use .data.floating in place of .data.float. But it
did not expect that .data subattribute may have name different from
lowercased type name.
2. vim_replace_termcodes returned its argument as-is if it receives an empty
string (as well as _vim_id*() functions did). But if something in returned
argument lives in an allocated memory such action will cause double free:
once when freeing arguments, then when freeing return value. It did not cause
problems yet because msgpack bindings return empty string as {NULL, 0} and
nothing was actually allocated.
3. New code in msgpack-gen.lua popped arguments in reversed order, making lua
bindings’ signatures be different from API ones.
No tests yet, no documentation update, no :lua* stuff, no vim module.
converter.c should also work with typval_T, not Object.
Known problem: luaeval("1", {}) results in
PANIC: unprotected error in call to Lua API (attempt to index a nil value)
Ref #3823
- Update Neovim to the latest version to see if your problem persists.
-If you're using a plugin manager, comment out your plugins, then add them back
in one by one, to narrow down the cause of the issue.
-Crash reports which include a stacktrace are 10x more valuable.
-[Bisecting][git-bisect] to the cause of a regression often leads to an
immediate fix.
-Disable plugins incrementally, to narrow down the cause of the issue.
- When reporting a crash, include a stacktrace.
-[Bisect][git-bisect] to the cause of a regression, if you are able. This is _extremely_ helpful.
-Check `$NVIM_LOG_FILE`, if it exists.
- Include `cmake --system-information` for **build** issues.
Pull requests ("PRs")
---------------------
- To avoid duplicate work, create a `[WIP]` pull request as soon as possible.
- Avoid cosmetic changes to unrelated files in the same commit: noise makes
reviews take longer.
- Avoid cosmetic changes to unrelated files in the same commit.
- Use a [feature branch][git-feature-branch] instead of the master branch.
- Use a **rebase workflow** for small PRs.
- After addressing review comments, it's fine to rebase and force-push.
@@ -43,7 +43,7 @@ Pull requests ("PRs")
- Use the `ri` git alias:
```
[alias]
ri = "!sh -c 't=\"${1:-master}\"; s=\"${2:-HEAD}\"; if git merge-base --is-ancestor \"$t\" \"$s\" ; then o=\"$t\" ; else mb=\"$(git merge-base \"$t\" \"$s\")\"; if test \"x$mb\" = x ; then o=\"$t\"; else lm=\"$(git log -n1 --merges \"$t..$s\" --pretty=%H)\"; if test \"x$lm\" = x ; then o=\"$mb\"; else o=\"$lm\"; fi; fi; fi ; [ $# -gt 0 ] && shift; [ $# -gt 0 ] && shift; git rebase --interactive \"$o\" \"$@\"' -"
ri = "!sh -c 't=\"${1:-master}\"; s=\"${2:-HEAD}\"; mb=\"$(git merge-base \"$t\" \"$s\")\"; if test \"x$mb\" = x ; then o=\"$t\"; else lm=\"$(git log -n1 --merges \"$t..$s\" --pretty=%H)\"; if test \"x$lm\" = x ; then o=\"$mb\"; else o=\"$lm\"; fi; fi; test $# -gt 0 && shift; test $# -gt 0 && shift; git rebase --interactive \"$o\" \"$@\"'"
```
This avoids unnecessary rebases yet still allows you to combine related
commits, separate monolithic commits, etc.
@@ -86,16 +86,18 @@ the VCS/git logs more valuable.
### Automated builds (CI)
Each pull request must pass the automated builds ([travis CI] and [quickbuild]).
Each pull request must pass the automated builds on [travis CI], [quickbuild]
and [AppVeyor].
- CI builds are compiled with [`-Werror`][gcc-warnings], so if your PR
introduces any compiler warnings, the build will fail.
- CI builds are compiled with [`-Werror`][gcc-warnings], so compiler warnings
will fail the build.
- If any tests fail, the build will fail.
See [Building Neovim#running-tests][wiki-run-tests] to run tests locally.
Passing locally doesn't guarantee passing the CI build, because of the
different compilers and platforms tested against.
- CI runs [ASan] and other analyzers. To run valgrind locally:
`VALGRIND=1 make test`
- CI runs [ASan] and other analyzers.
- To run valgrind locally: `VALGRIND=1 make test`
- To run Clang ASan/UBSan locally: `CC=clang make CMAKE_FLAGS="-DCLANG_ASAN_UBSAN=ON"`
- The `lint` build ([#3174][3174]) checks modified lines _and their immediate
neighbors_. This is to encourage incrementally updating the legacy style to
meet our style guidelines.
@@ -112,11 +114,19 @@ QuickBuild uses this invocation:
In LaTeX, since "%" is used as the comment character, you can >
:let b:match_skip = 'r:%'
< Unfortunately, this will skip anything after "\%", an escaped "%". To
allow for this, and also "\\%" (an excaped backslash followed by the
allow for this, and also "\\%" (an escaped backslash followed by the
comment character) you can >
:let b:match_skip = 'r:\(^\|[^\\]\)\(\\\\\)*%'
<
@@ -356,7 +356,8 @@ The back reference '\'.d refers to the same thing as '\'.b:match_table[d] in
The various |:vmap|s defined in the script (%, |g%|, |[%|, |]%|, |a%|) may
have undesired effects in Select mode |Select-mode-mapping|. At least, if you
want to replace the selection with any character in "ag%[]" there will be a
pause of |'updatetime'| first.
pause of |'updatetime'| first. E.g., "yV%" would normally work linewise, but
the plugin mapping makes it characterwise.
It would be nice if "\0" were recognized as the entire pattern. That is, it
would be nice if "foo:\end\0" had the same effect as "\(foo\):\end\1".
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.
