On Saturday 16 July 2005 10:38 pm, Marius Mauch wrote:
> On Sat, 16 Jul 2005 20:22:29 -0400
>
> Mike Frysinger <[EMAIL PROTECTED]> wrote:
> > On Saturday 16 July 2005 01:03 pm, Marius Mauch wrote:
> > > On Fri, 15 Jul 2005 21:34:09 -0400
> > >
> > > Mike Frysinger <[EMAIL PROTECTED]> wrote:
> > > > we could care less what users do with /etc/profile.d ... the
> > > > point is that *only* users should use /etc/profile.d ... we dont
> > > > want random Gentoo developer Foo installing some Bar.sh
> > > > into /etc/profile.d with package app-crap/FooBar
> > >
> > > Would the following in /et/profile be a solution to this problem?
> > >
> > > for x in $( < /etc/profile.d/.default); do
> > >   source "/etc/profile.d/${x}"
> > > done
> > >
> > > That way devs could install stuff there, but it would only be run if
> > > users added it to the .default file.
> >
> > that kind of limits the intuitiveness of profile.d ... plus, not like
> > they couldnt just do 'echo blah >> /etc/profile.d/.default' at the
> > end of pkg_postinst or something
>
> They could do the same to /etc/profile, no?

yeah could which is why we could just do a QA smackdown on package maintainers 
who utilize /etc/profile.d ...

a quick grep shows that the following packages mention /etc/profile.d for some 
reason or another:
dev-util/aegis (but it seems to correctly remove support)
x11-base/xorg-x11 (no idea what it's trying to do with /etc/profile.d/xprint*)
app-shells/bash-completion
app-shells/tcsh
-mike
-- 
[email protected] mailing list

Reply via email to