Path: utzoo!utgpu!water!watmath!clyde!rutgers!cmcl2!brl-adm!adm!MISS026%ECNCDC.BITNET@CUNYVM.CUNY.EDU From: MISS026%ECNCDC.BITNET@CUNYVM.CUNY.EDU (GREENY) Newsgroups: comp.unix.wizards Subject: re: a solution to the RM insanity Message-ID: <10767@brl-adm.ARPA> Date: 10 Dec 87 18:12:15 GMT Sender: news@brl-adm.ARPA Lines: 30 I have been sitting back lately and reading many a message on how to deal with the problems that RM * can cause, and it occured to me that if one has a problem with not knowing how a command will react with certainty (such as the use of RM $* over a network [where $ represents an alphanumeric character of a group to files to kill (e.g. rm a*)]) then one SHOULD NOT MAKE USE OF SUCH A COMMAND. It is much safer to simply type the entire file name and make use of the rm -i option to be certain that the system actually received what you typed. As a rule I have always steered away from RM * or RM $* after one bad experience with it. Usually I will send myself a note (I use the rs system for this purpose...) to remind myself to delete the file. Then at the end of the week I compile all of these filenames and their ABSOLUTE pathname into a file, and have a shell script remove each file for me. If I perform this action over a network then I make certain to use the rm -i command to counter against the possibility of noise bursts on the line altering a filename. In summary, I believe that there are a number of valid ways in which to deal with the RM * problem, and my above solution is only one of a multitude. However, the number of messages on the supposed 'problem' is getting to be a little boring to read. How about we move onto something else? :-) bye for now but not for long... Greeny Bitnet: Miss026@ecncdc Internet: Miss026%ecncdc.bitnet@cunyvm.cuny.edu <- I think...this is new Disclaimer: What? Who? Me? Nope...not me...you *MUST* have the wrong hacker!!