Hi, > However, pgrep/pkill does not take a > parameter to list how to use the command, so usage() should always > exit with EXIT_USAGE.
in fact, there *is* such a command line parameter: pgrep -?. The problem here is that the arbitrarily chosen "-?" and the behaviour of getopt(3) upon encountering unknown commandline parameters overlap. The "case '?'" is reached in both cases - having explicitly specified -? or some other invalid character. However, getopt(3) sets optopt to the character that it actually found in the input in case of an error, so that's how we can make a decision. The exit(2) call in usage() never needed a change. New patch attached. Regards, Jan
#! /bin/sh /usr/share/dpatch/dpatch-run
## 55_pgrep_usage_exitcode.dpatch by <[EMAIL PROTECTED]>
##
## DP: Distinguish between invalid commandline parameters and '-?'.
@DPATCH@
--- procps-3.2.7.orig/pgrep.c 2007-08-06 16:35:30.000000000 +0200
+++ procps-3.2.7/pgrep.c 2007-08-06 16:35:06.000000000 +0200
@@ -678,7 +678,7 @@
// case 'z': // Solaris: match by zone ID
// break;
case '?':
- usage (opt);
+ usage (optopt?optopt:opt);
break;
}
}
signature.asc
Description: Digital signature
