Red Hat Bugzilla – Bug 436083
cp -i -f should act sensibly
Last modified: 2008-03-05 07:03:26 EST
Description of problem: If both -i and -f options (in that order) are specified to cp then cp should not act interactively. Conversely, if -f and -i options (in that order) are specified then -f should be ignored. Version-Release number of selected component (if applicable): coreutils-6.9-9 How reproducible: Always Steps to Reproduce: 1. unalias cp 2. touch a b 3. cp -i -f a b 4. cp -f -i a b Actual results: 3. prompts 4. prompts Expected results: 3. no prompt 4. prompt Additional info: In a sense -i and -f are complimentary. That is, -i means ask, and -f means don't ask. Often users will alias cp='cp -i' to avoid clobbering files. But then when they really want to overwrite files unconditionally they will use 'cp -f' which will translate to 'cp -i -f'. While it is entirely possible to use '\cp -f' that assumes that the user is fully conscious of any cp alias in effect. It would make a lot more sense to have both -f and -i use the same internal flag, rather than distinct "unlink_dest_after_failed_open" and "interactive", and thereby the final -i/-f would be the flag that endures.
from cp info page: `-f' `--force' ... This option is independent of the `--interactive' or `-i' option: neither cancels the effect of the other. So IMO this options work as designed and this behaviour is not a bug. POSIX specification (defines this behaviour): http://www.opengroup.org/onlinepubs/009695399/utilities/cp.html Upstream report with answer: http://www.mail-archive.com/bug-coreutils@gnu.org/msg11297.html I'm not going to break POSIX definitions and to create fork of cp in this case, therefore closing NOTABUG.