The existence of /tmp is required by the FHS standard, so I think it shouldbe safe to assume it exists.Sorry to be unclear. I mean "safe" in the sense of security. A predictable *file*name in a world writable directory is unsafe because it can be used for a symlink attack to overwrite arbitrary files. For a *directory* this is less the case, because in this script the 'mkdir' will fail if the dir alreadyexists and "set -e" will terminate the script.
Ah, cool! :-) I think we're good, then!Well, as for mustang, I've done some rather substantial changes to the packaging, with the inclusion of the test data in a separate package, so it would probably be useful if someone could review these changes.
Otherwise, I am quite happy with the state of affairs. I received a very nice email from mustang's upstream author, he is happy about the packaging, and is working on a new and improved version. Hopefully it will be released under a free license!
Cheers, Morten
Attachment:
PGP.sig
Description: This is a digitally signed message part