Problem: Invoking mark_adjust() when adding a new line below the last line
is pointless.
Solution: Skip calling mark_adjust() when appending below the last line.
82faa259cc
Inherited signal mask may block SIGCHLD, which causes libuv to hang at
epoll_wait.
Closes#5230
Helped-by: Nicolas Hillegeer <nicolas@hillegeer.com>
Helped-by: John Szakmeister <john@szakmeister.net>
Note: the #pragma gymnastics are a workaround for broken system headers on
macOS.
signal.h:
int sigaddset(sigset_t *, int);
#define sigaddset(set, signo) (*(set) |= __sigbits(signo), 0)
sys/_types/_sigset.h:
typedef __darwin_sigset_t sigset_t;
sys/_types.h:
typedef __uint32_t __darwin_sigset_t; /* [???] signal set */
sigset_t is defined as unsigned int, but the sigaddset() ORs it with an int,
mixing the types. So GCC generates a sign-conversion warning:
sig.c:9:13: warning: implicit conversion changes signedness: 'int' to 'unsigned int' [-Wsign-conversion]
(*(&s) |= __sigbits((sigset_t) 20), 0);
~~ ^~~~~~~~~~~~~~~~~~~~~~~~
1 warning generated.
System headers are normally ignored when the compiler generates warnings:
https://gcc.gnu.org/onlinedocs/cpp/System-Headers.html
> GCC gives code found in system headers special treatment. All warnings,
> other than those generated by ‘#warning’ (see Diagnostics), are suppressed
> while GCC is processing a system header. Macros defined in a system header
> are immune to a few warnings wherever they are expanded. This immunity is
> granted on an ad-hoc basis, when we find that a warning generates lots of
> false positives because of code in macros defined in system headers.
Instead of the #pragma workaround, we could cast the sigset_t pointer:
# if defined(__APPLE__)
sigaddset((int *)&mask, SIGCHLD);
# else
sigaddset(&mask, SIGCHLD);
# endif
but that could break if the headers are later fixed.
Problem: 'cscopequickfix' option does not accept new value "a".
Solution: Adjust list of command characters. (Ken Takata)
6d20e17544
All changes applied manually. Definition of `CSQF_CMDS` was moved to
`option_defs.h` in nvim.
Problem: Cscope interface does not support finding assignments.
Solution: Add the "a" command. (ppettina, closesvim/vim#882)
b12e7ef956
All changes applied manually.
When USE_ICONV is defined, iconv.h references various errno constants,
but errno.h is only being included when HAVE_ICONV_H is not defined.
This causes build failures on at least GNU/Hurd.
For any script--not just `:global` commands--there is no reason to
update the system clipboard until the script is finished, so disable it
during do_cmdline().
Before this change, 'clipboard=unnamedplus' causes scripted editing to
be extremely slow (e.g. `:normal` in a while-loop).
Closes#3534
* 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.
- Links to Search by default
screen.c: Combine CursorLine with QuickFixLine
- HLF_QFL takes priority over HLF_CUL
docs: Updated to mention QuickFixLine
runtime: Added QuickFixLine to nvimHLGroup
tests: QuickFixLine highlight
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
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.
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
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.
