Re: [gentoo-security] Security concerns and portage versioning

[Jeremy Bowers]

Calum wrote:
 > The more choice the better, IMHO.

Not true. You already *have* maximal choice; assuming you want to mimize 
emerges, you are required to pick and choose exactly which updates you 
want to emerge, by hand, with no automation support.

If you're not going to reduce this choice to some degree by providing 
some bundling, then there's no point continuing; we already have the 
system that provides maximal flexibility.

Some of us would like to be able to "volunteer" for portage to have a 
little less flexibility and automatically do the minimal security 
updates, even though that technically reduces our choices (by making 
some things harder since they go "against the grain"). Endless (and 
ultimately futile) categorization of the security issue types reduces 
the value of "emerge security" by requiring the user to make almost as 
many choices about what they care about as the original state we are in now.

Calum, I'm sure I'm not going to convince you of this, but I think it 
should be said in general: I don't see any reason to force the admin to 
choose between *which* security fixes need to be applied. What 
security-concious admin is honestly going to update only *remote-root* 
and leave *local-root* alone, when a "local-root" exploit + a 
(potentially unknown) remote shell exploit (or potentially even weaker 
exploits) constitutes a remote-root exploit?

The -sX solution seems like a good idea to me. No real need to 
differentiate between the reasons, IMHO.

--
gentoo-security@g.o mailing list