Path: utzoo!utgpu!watmath!clyde!att!rutgers!apple!bionet!lear From: lear@NET.BIO.NET (Eliot Lear) Newsgroups: comp.mail.uucp Subject: Re: Return paths (was Re: Another example why not to re-route) Message-ID:Date: 3 Dec 88 10:41:49 GMT References: <913@tank.uchicago.edu> <351@lakart.UUCP> Organization: Natl Computer Resource for Mol. Biology Lines: 18 To: dg@lakart.UUCP The problem is that .UUCP is a magic cookie for certain sendmail configurations, and they do not handle the case correctly. It is not uncommon to find a rule that does something like the following: given user%host.uucp@host, convert to host!user@host. This gives mailers ample opportunity to screw up precedence of ! and @. Aside from the rabidity of Chip's tone, I would agree that sendmail.cfs can be abused. This is mostly due to a lack of understanding of the consequences that occur when changes are made. It is often said that the CF creeping featurism. HOWEVER. The above isn't a good enough reason to dismiss the CF. The problem is that we exist in a heterogenious environment, and the configurability is necessary. -- Eliot Lear [lear@net.bio.net]