On Sun, Jan 20, 2002 at 08:52:55PM -0500, Michael P. Soulier wrote: > On 20/01/02 Colin Watson did speaketh: > > Well, quite, but the question you snipped was about why the maintainer > > didn't notice it. Although I occasionally test packages I build on, say, > > my workstation at work, I'll probably only do that if I think the change > > is particularly risky. > > Should testing not imply a proper regression test? I think so.
I'm sure the base-passwd maintainer would appreciate a regression test suite if you wrote one and sent it to him. (In other words: in theory, yes, but someone has to do the work.) > > The problem with this, IMHO, is that it encourages people to set 'umask > > 022' routinely so that people in the same group can't write to files in > > their home directory. They then often forget to set 'umask 002' when > > working on shared files. > > > > I think one group per user combined with setgid directories is a much > > more practical solution, but I realize that this seems to be a religious > > issue. > > Who sets group write permission on shared files?? I would _never_ do > that. It only encourages clobbering each other's changes. That's what > version control systems are for. This is a matter of preference, and version control systems are not always appropriate. I think I mentioned the religion thing somewhere too. -- Colin Watson [EMAIL PROTECTED]
