Fix dependency generators sometimes dying with SIGPIPE

If a dependency generator dies while we're still writing to its stdin,
it causes us to die rather randomly. Typically happens with fake
dependency generators that don't actually bother reading their input
and/or write anything back. This is almost certainly the ghost failure
we've occasionally seen in the test-suite too (rpm-software-management#2470).

Resurrect the explicit SIG_IGN that got removed apparently in commit
375a6b5. The matter was further
confused by NSPR (which we once relied on) always setting SIGPIPE
to SIG_IGN.

Suggested-by: Michael Schroeder <mls@suse.de>

Fixes: rpm-software-management#2949

build/rpmfc.c

@@ -295,6 +295,8 @@ static int getOutputFrom(ARGV_t argv,
 	return -1;
     }
 
+    void *oldhandler = signal(SIGPIPE, SIG_IGN);
+
     child = fork();
     if (child < 0) {
 	rpmlog(RPMLOG_ERR, _("Couldn't fork %s: %s\n"),
@@ -416,6 +418,8 @@ static int getOutputFrom(ARGV_t argv,
 	reaped = waitpid(child, &status, 0);
     } while (reaped == -1 && errno == EINTR);
 
+    signal(SIGPIPE, oldhandler);
+
     /*
      * It's not allowed to call WIFEXITED or WEXITSTATUS if waitpid return -1.
      * Note, all bits set, since -1 == 0xFFFFFFFF