| From: David Korn <dgk@research.att.com>
| To: werner@suse.de
| Subject:  Re: Re: [ast-users] Crash with SIGPIPE trap...
|
| OK, now I get what is causing this.  Here is a simpler fix
| that I think should work.  If the stream is in a pool, then
| this just purges the top level so that it doesn't write to
| that level.
|
--- src/cmd/ksh93/sh/fault.c
+++ src/cmd/ksh93/sh/fault.c	2009-08-24 15:52:21.770401660 +0200
@@ -413,6 +413,9 @@ void	sh_chktrap(void)
 			sh.sigflag[sig] &= ~SH_SIGTRAP;
 			if(trap=sh.st.trapcom[sig])
 			{
+				Sfio_t *fp;
+				if (sig==SIGPIPE && (fp=sfpool((Sfio_t*)0,sh.outpool,SF_WRITE)))
+					    sfpurge(fp);
 				sh.oldexit = SH_EXITSIG|sig;
 				sh_trap(trap,0);
 			}