| 1 |
On Thursday 30 September 2004 21:55, Ned Ludd wrote: |
| 2 |
> On Thu, 2004-09-30 at 15:21, Stephen P. Becker wrote: |
| 3 |
> > ...which is the reason why we have sandbox. FEATURES="sandbox" |
| 4 |
> > causes an emerge to terminate immediately with an access violation if |
| 5 |
> > it attempts to touch system files before the build is complete. |
| 6 |
> > |
| 7 |
> > I'm not saying you are wrong by the way, I'm just pointing out that |
| 8 |
> > stuff like this has been thought of before, so portage has safety |
| 9 |
> > nets accordingly. |
| 10 |
|
| 11 |
Sandbox should never ever be regarded as a security measure. It isn't. It |
| 12 |
is almost trivial to subvert the sandbox. The reason for it's |
| 13 |
effectiveness is solely that it's purpose is to protect against |
| 14 |
accidental installing outside of the destination directory and so |
| 15 |
subverting the package management (in short protecting against bad |
| 16 |
makefiles and ebuilds). It IS NOT SECURE. |
| 17 |
|
| 18 |
Paul |
| 19 |
|
| 20 |
-- |
| 21 |
Paul de Vrieze |
| 22 |
Gentoo Developer |
| 23 |
Mail: pauldv@g.o |
| 24 |
Homepage: http://www.devrieze.net |
Archives