Re: gzip zgrep-signal failure on Fedora 15 in an Emacs shell window

[Paul Eggert]

On 11/02/11 14:49, Eric Blake wrote:
> Another portability concern - with newer ksh, and allowed by POSIX,
> a child that exits due to a signal causes $? to be 269, not 141.

Thanks.  Plus, POSIX doesn't require SIGPIPE to be 13, so the
141 is unportable for that reason too.  I pushed this:

* tests/zgrep-signal: Don't assume exit status 141 on PIPE signal.
Problem reported by Eric Blake in
<http://lists.gnu.org/archive/html/bug-gzip/2011-11/msg00007.html>.
diff --git a/tests/zgrep-signal b/tests/zgrep-signal
index e19293a..3260875 100644
--- a/tests/zgrep-signal
+++ b/tests/zgrep-signal
@@ -43,13 +43,14 @@ write_to_dangling_pipe () {
   ) | : || framework_failure_
 }

-st=$(write_to_dangling_pipe cat f.gz f.gz)
-test "$st" = 141 || framework_failure_ 'signal handling busted on this host'
+signal_status=$(write_to_dangling_pipe cat f.gz f.gz)
+test 128 -lt $signal_status ||
+  framework_failure_ 'signal handling busted on this host'

 fail=0

 st=$(write_to_dangling_pipe zgrep a f.gz f.gz)

-test "$st" = 141 || fail=1
+test $st = $signal_status || fail=1

 Exit $fail