strange hang on wscons

[Chavdar Ivanov <ci4ic4%gmail.com@localhost>]

Hi,

On a -current system, when running wip/neovim-git on the console or
any of the enabled terminals, I get a hang - the screen becomes
entirely black and no longer responds, I cannot blindly quit nvim
anymore. If I ssh to another system and attach with gdb to the
process, I get:

...
0x00007b6def844d5a in _sys___kevent50 () from /usr/lib/libc.so.12
(gdb) thread apply all bt

Thread 2 (LWP 8876 of process 8876):
#0  0x00007b6def844d5a in _sys___kevent50 () from /usr/lib/libc.so.12
#1  0x00007b6df0c079d8 in __kevent50 () from /usr/lib/libpthread.so.1
#2  0x00007b6df281d78e in uv__io_poll (loop=loop@entry=0x960600
<main_loop>, timeout=<optimized out>) at src/unix/kqueue.c:214
#3  0x00007b6df280fc87 in uv_run (loop=loop@entry=0x960600
<main_loop>, mode=UV_RUN_ONCE) at src/unix/core.c:385
#4  0x00000000004ca70c in loop_poll_events (loop=0x960600 <main_loop>,
ms=ms@entry=-1) at
/usr/pkgsrc/wip/neovim-git/work/neovim/src/nvim/event/loop.c:62
#5  0x00000000005912b0 in inbuf_poll (ms=ms@entry=-1,
events=events@entry=0x7b6df36f8000) at
/usr/pkgsrc/wip/neovim-git/work/neovim/src/nvim/os/input.c:403
#6  0x000000000059170c in os_inchar (buf=buf@entry=0x0,
maxlen=maxlen@entry=0, ms=ms@entry=-1,
tb_change_cnt=tb_change_cnt@entry=0, events=0x7b6df36f8000) at
/usr/pkgsrc/wip/neovim-git/work/neovim/src/nvim/os/input.c:128
#7  0x00000000006059d3 in state_enter (s=s@entry=0x7f7ffff81370) at
/usr/pkgsrc/wip/neovim-git/work/neovim/src/nvim/state.c:54
#8  0x0000000000569af9 in normal_enter (cmdwin=cmdwin@entry=false,
noexmode=noexmode@entry=false) at
/usr/pkgsrc/wip/neovim-git/work/neovim/src/nvim/normal.c:463
#9  0x00000000005380d1 in main (argc=<optimized out>, argv=<optimized
out>) at /usr/pkgsrc/wip/neovim-git/work/neovim/src/nvim/main.c:553

Thread 1 (LWP 8779 of process 8876):
#0  0x00007b6def844d5a in _sys___kevent50 () from /usr/lib/libc.so.12
#1  0x00007b6df0c079d8 in __kevent50 () from /usr/lib/libpthread.so.1
#2  0x00007b6df281d78e in uv__io_poll (loop=loop@entry=0x7b6deedff9c0,
timeout=<optimized out>) at src/unix/kqueue.c:214
#3  0x00007b6df280fc87 in uv_run (loop=loop@entry=0x7b6deedff9c0,
mode=UV_RUN_ONCE) at src/unix/core.c:385
#4  0x00000000004ca70c in loop_poll_events
(loop=loop@entry=0x7b6deedff9c0, ms=ms@entry=-1) at
/usr/pkgsrc/wip/neovim-git/work/neovim/src/nvim/event/loop.c:62
#5  0x000000000062125c in tui_main (bridge=0x7b6df360e000,
ui=0x7b6df36b4140) at
/usr/pkgsrc/wip/neovim-git/work/neovim/src/nvim/tui/tui.c:451
#6  0x0000000000626646 in ui_thread_run (data=<optimized out>) at
/usr/pkgsrc/wip/neovim-git/work/neovim/src/nvim/ui_bridge.c:104
#7  0x00007b6df0c0bee0 in ?? () from /usr/lib/libpthread.so.1
#8  0x00007b6def892390 in ?? () from /usr/lib/libc.so.12
#9  0x0000000000400000 in ?? ()
#10 0x00007b6def200000 in ?? ()
#11 0x0000001003a0efff in ?? ()
#12 0x00007b6def0000c0 in ?? ()
#13 0x00000000001fff40 in ?? ()
#14 0x0000000000000000 in ?? ()

If I now kill the nvim process, the screen remains stuck; I cannot
blindly logout and then login again for instance; perhaps there is a
way to recover the console I am not aware of, but for now I just
reboot...

The same program runs fine under X or if I ssh to the machine from
elsewhere. On the other hand, it happens the same way if I ssh from
the console to some other system with neovim installed and run it.

Chavdar

-- 
----