4.4 repeated failures in CentOS 7

[CdeMills]

Hello,

I have built two times octave 4.4 from checked out sources in CentOS7. The
resulting binary, in gui mode, usually quickly fails with signal 6 or 11.

Compilation options: ./configure --prefix=${HOME}/usr/local/OCTAVE/4.4.0
--with-qt=5 --disable-jit

execution: strace -f -o signal6 octave --gui 2>&1

One exemple of failure:
30139 <... write resumed> )             = 8
30131 <... write resumed> )             = 27
30139 futex(0x7fb3f80021bc, FUTEX_WAIT_PRIVATE, 1, NULL <unfinished ...>
30131 write(2, "'\n", 2)                = 2
30131 rt_sigprocmask(SIG_UNBLOCK, [ABRT], NULL, 8) = 0
30131 tgkill(30131, 30131, SIGABRT)     = 0
30131 --- SIGABRT {si_signo=SIGABRT, si_code=SI_TKILL, si_pid=30131,
si_uid=1261} ---
30131 futex(0x7fb426a42c40, FUTEX_WAKE_PRIVATE, 2147483647) = 0
30131 write(2, "fatal: caught signal ", 21) = 21
30131 write(2, "Abandon", 7 <unfinished ...>
[...]
30149 +++ killed by SIGABRT (core dumped) +++
30148 +++ killed by SIGABRT (core dumped) +++
30140 +++ killed by SIGABRT (core dumped) +++
30139 +++ killed by SIGABRT (core dumped) +++
30137 +++ killed by SIGABRT (core dumped) +++
30172 +++ killed by SIGABRT (core dumped) +++
30136 +++ killed by SIGABRT (core dumped) +++
30135 +++ killed by SIGABRT (core dumped) +++
30134 +++ killed by SIGABRT (core dumped) +++
30131 +++ killed by SIGABRT (core dumped) +++
30130 <... wait4 resumed> [{WIFSIGNALED(s) && WTERMSIG(s) == SIGABRT &&
WCOREDUMP(s)}], 0, NULL) = 30131
30130 --- SIGCHLD {si_signo=SIGCHLD, si_code=CLD_DUMPED, si_pid=30131,
si_uid=1261, si_status=SIGABRT, si_utime=168, si_stime=113} ---
30130 write(2, "octave exited with signal ", 26) = 26
30130 write(2, "6", 1)                  = 1
30130 write(2, "\n", 1)                 = 1
30130 exit_group(0)                     = ?
30130 +++ exited with 0 +++

Anybody got the same behavior ? The console mode does not exhibit all those
issues.

Regards

Pascal



--
Sent from: http://octave.1599824.n4.nabble.com/Octave-Maintainers-f1638794.html