[ksh93-integration-discuss] SIGTTOU bug in b72 ksh93?

[Shawn Walker]

On 15/09/2007, David Powell <David.Powell at sun.com> wrote:
> Shawn Walker wrote:
>  > Greetings,
>  >
>  > Clean install of SXDE b72:
>  >
>  > I've discovered if I:
>  > 1) set ksh93 as my login shell
>  >
>  > 2) Open a gnome-terminal
>  >
>  > 3) bldenv -d opensolaris.sh
>  >
>  > 4) clear
>  >
>  > I get:
>  >
>  > [1] + Stopped(SIGTTOU)         clear
>
>   I'm seeing this on build 73.  My steps to reproduce are:
>
>   (in xterm)
>   1) ksh93
>   2) clear
>
>   Trussing ksh93 indeed shows it receiving a SIGCLD indicating
>   its child was stopped with a SIGTTOU:
>
>     Received signal #18, SIGCLD [caught]
>       siginfo: SIGCLD CLD_STOPPED pid=1825 status=0x001B
>     Received signal #18, SIGCLD [caught]
>       siginfo: SIGCLD CLD_CONTINUED pid=1825 status=0x0019
>
>   Instead of just:
>
>     Received signal #18, SIGCLD [caught]
>       siginfo: SIGCLD CLD_EXITED pid=1830 status=0x0000
>
>   which I get when I run some other command under ksh93 that writes to
>   the terminal or when I run clear under ksh88.
>
>   Curiously, both 'clear' and 'tput clear' are sent SIGTTOU, though
>   'tput clear' doesn't reproduce the problem.
>
>   Additionally, trussing ksh93 with -f causes the problem to go away.
>   Smells like a race to me.

I agree, because I notice if I do it really fast repeatedly, it
happens *almost* every time, but not every time.

-- 
Shawn Walker, Software and Systems Analyst
binarycrusader at gmail.com - http://binarycrusader.blogspot.com/

"Beware of bugs in the above code; I have only proved it correct, not
tried it. " --Donald Knuth