From owner-freebsd-bugs Sun Jun 16 19:40:08 1996 Return-Path: owner-bugs Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id TAA06212 for bugs-outgoing; Sun, 16 Jun 1996 19:40:08 -0700 (PDT) Received: (from gnats@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id TAA06203; Sun, 16 Jun 1996 19:40:04 -0700 (PDT) Date: Sun, 16 Jun 1996 19:40:04 -0700 (PDT) Message-Id: <199606170240.TAA06203@freefall.freebsd.org> To: freebsd-bugs Cc: From: Bruce Evans Subject: Re: bin/1328: install silently removes target if strip is unexecutable Reply-To: Bruce Evans Sender: owner-bugs@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk The following reply was made to PR bin/1328; it has been noted by GNATS. From: Bruce Evans To: FreeBSD-gnats-submit@freebsd.org, hsu@clinet.fi Cc: Subject: Re: bin/1328: install silently removes target if strip is unexecutable Date: Mon, 17 Jun 1996 12:29:27 +1000 >>Synopsis: install silently removes target if strip is unexecutable It actually removes the target if almost any error occurs, unless the -C option is used, and the usual error for stripping unexecutables is unsilent but results in `install' exiting with status 0. >>Fix: >Files should not get removed if strip fails. No, strip might have left junk in the file. Strip shouldn't even be attempted on the final file. The -C option avoids touching the final file as a side effect. Someday all installs should be atomic. -C is atmic now except for some cases involving immutable flags (immutable files can't be renamed atomically). >If install fails, it should return non-zero exit value so that make >will stop before serious damage gets done. Yes. Bruce