Christos Zoulas wrote:

> I think that chpass should obey the /etc/ptmp lock.
> [...] 
> The problem with editing the password file and creating a temp file
> in /tmp is data loss from concurrent edits. How does this scheme
> prevent this? Let's say I am editing the password file using vipw.
> You run chpass; I save my edit; your changes are lost.

Did chpass(1) use the /etc/ptmp lock prior to the changes mentioned
in the original mail?

-e.

-- 
Elad Efrat