Some terminals don't report which buttons are involved in some mouse
events. For example, the urxvt protocol
(http://www.huge-man-linux.net/man7/urxvt.html section "Mouse
reporting") does not report which button has been released.
In this case libtermkey reports button 0
(http://www.leonerd.org.uk/code/libtermkey/doc/termkey_interpret_mouse.3.html)
Up to now, forward_mouse_event did not handle button==0.
On press events there is not much we can do, and we keep the
current behavior which is dropping the event. But on drag-and-release
events we can compensate by remembering the last button pressed.
fixes#3182 for urxvt
fixes#5400
ref #6725
fsync() is very slow on some systems. And since the parent commit, Nvim
is smarter about flushing files at certain times (e.g. CursorHold),
regardless of whether 'fsync' is enabled. So it's less risky to disable
'fsync'.
Profiling showed slow (2-4s) :write and :quit caused by fsync():
:quit
shada_write_file(NULL, false);
:write + fsync
0 0x00007f72da567b2d in fsync () at ../sysdeps/unix/syscall-template.S:84
1 0x0000000000638970 in uv__fs_fsync (req=<optimized out>) at /home/vagrant/neovim/.deps/build/src/libuv/src/unix/fs.c:150
2 uv__fs_work (w=<optimized out>) at /home/vagrant/neovim/.deps/build/src/libuv/src/unix/fs.c:953
3 0x0000000000639a70 in uv_fs_fsync (loop=<optimized out>, req=<optimized out>, file=41, cb=0x7f72da567b2d <fsync+45>)
at /home/vagrant/neovim/.deps/build/src/libuv/src/unix/fs.c:1094
4 0x0000000000573694 in os_fsync (fd=41) at ../src/nvim/os/fs.c:631
5 0x00000000004ec9dc in buf_write (buf=<optimized out>, fname=<optimized out>, sfname=<optimized out>, start=1, end=1997, eap=0x7fffc864c570,
append=<optimized out>, forceit=<optimized out>, reset_changed=<optimized out>, filtering=<optimized out>) at ../src/nvim/fileio.c:3387
6 0x00000000004b44ff in do_write (eap=0x7fffc864c570) at ../src/nvim/ex_cmds.c:1745
...
:write + nofsync
0 0x00007f72da567b2d in fsync () at ../sysdeps/unix/syscall-template.S:84
1 0x0000000000638970 in uv__fs_fsync (req=<optimized out>) at /home/vagrant/neovim/.deps/build/src/libuv/src/unix/fs.c:150
2 uv__fs_work (w=<optimized out>) at /home/vagrant/neovim/.deps/build/src/libuv/src/unix/fs.c:953
3 0x0000000000639a70 in uv_fs_fsync (loop=<optimized out>, req=<optimized out>, file=36, cb=0x7f72da567b2d <fsync+45>)
at /home/vagrant/neovim/.deps/build/src/libuv/src/unix/fs.c:1094
4 0x0000000000573694 in os_fsync (fd=36) at ../src/nvim/os/fs.c:631
5 0x0000000000528f5a in mf_sync (mfp=0x7f72d8968d00, flags=5) at ../src/nvim/memfile.c:466
6 0x000000000052d569 in ml_preserve (buf=0x7f72d890f000, message=0) at ../src/nvim/memline.c:1659
7 0x00000000004ebadf in buf_write (buf=<optimized out>, fname=<optimized out>, sfname=<optimized out>, start=1, end=1997, eap=0x7fffc864c570,
append=<optimized out>, forceit=<optimized out>, reset_changed=<optimized out>, filtering=<optimized out>) at ../src/nvim/fileio.c:3071
8 0x00000000004b44ff in do_write (eap=0x7fffc864c570) at ../src/nvim/ex_cmds.c:1745
...
Vim has the 'swapsync' option which we removed in 62d137ce09.
Instead let 'fsync' control swapfile-fsync.
These cases ALWAYS force fsync (ignoring 'fsync' option):
- Idle (CursorHold).
- Exit caused by deadly signal.
- SIGPWR signal.
- Explicit :preserve command.
Avoid a hot loop in retry(), there's no need to retry more than 50/s.
Also use luv.sleep() to implement sleep() instead of spinning the
event-loop, so events are not silently discarded.
1. Don't check elapsed time in children_kill_cb(), it's already implied
by the start-time of the timer itself.
2. Restart timer from children_kill_cb() for PTY jobs, to send SIGKILL
after SIGTERM. There is an edge case where SIGKILL might follow
SIGTERM too quickly, if jobstop() is called near the 2-second timer
window. But this edge case is not worth code complication.
Before f31c26f1af the timer was used to try SIGTERM *and* SIGKILL, so
a repeating timer was needed. After f31c26f1af process_stop() sends
SIGTERM immediately, and the timer only sends SIGKILL.
So we don't need a repeating timer.
- Simplifies the logic: don't need to call uv_timer_stop() explicitly.
- Avoids a problem: if process_stop() is called more than once in the
2-second window, the first on_process_exit() would call
uv_timer_stop() which stops the timer for all stopped processes.
children_kill_cb() is racey. One obvious problem is that
process_close_handles() is *queued* by on_process_exit(), so when
children_kill_cb() is invoked, the dead process might still be in the
`loop->children` list. If the OS already reclaimed the dead PID, Nvim
may try to SIGKILL it.
Avoid that by checking `proc->status`.
Vim doesn't have this problem because it doesn't attempt to kill
processes that ignored SIGTERM after a timeout.
closes#8269
Fixes#6890 by reading from the Windows console input buffer after
stdin has been closed.
Vim defines HAVE_DUP for Windows and does the close-dup dance[1]:
close(0);
dup(2);
which always fails, then falls back to reading from the Windows console
input buffer[2].
[1] e7499ddc33/src/fileio.c (L2397-L2398)
[2] e7499ddc33/src/os_win32.c (L1703-L1714)
* Reading from stdin on Windows is fixed in the same way as it was in
#8267.
* The file_read function was returning without filling the
destination buffer when it was called with a non-blocking file
descriptor.
It's only appropriate to set CMAKE_MAKE_PROGRAM to gmake when we're
using the "Unix Makefiles" generator. On QB, the nodes have Ninja
available and will use it, which means CMAKE_GENERATOR is set to
"Ninja". Setting CMAKE_MAKE_PROGRAM was forcing the build to use gmake
instead of ninja, which was causing the build failure.
