1 |
On Saturday 17 February 2007, Brian Harring wrote: |
2 |
> Security impact is from a pkg potentially dragging along old libs; if |
3 |
> you've got a stable pkg that gets an update once every blue moon, it |
4 |
> can hold onto the lib for a *long* time while still using the lib; |
5 |
> thus if a vuln. in the lib, said pkg still is screwed. |
6 |
|
7 |
umm, no ... the package itself is updated against the new copy while the old |
8 |
copy exists for other packages that have already been built |
9 |
|
10 |
> Other angle is someone intentionally forcing usage of a known bad |
11 |
> library that is still dangling. Corner case, but doable. |
12 |
|
13 |
as i said, this is the "invalid" syntax: |
14 |
... /usr/lib/libfoo.so.3 ... |
15 |
|
16 |
besides, this is not a real concern ... if a user is purposefully relinking |
17 |
against files because it has a security issue, they have the ability to do a |
18 |
lot more than any bug exposed in the library |
19 |
|
20 |
> Bit curious how this is going to behave if via linked in libs, new loc |
21 |
> and old get loaded alongside... |
22 |
|
23 |
this would require multiple libraries to be involved in the equation and the |
24 |
answer is undefined behavior which most certainly result in runtime |
25 |
failures ... |
26 |
|
27 |
besides, just like the gcc-3.3 -> gcc-3.4 transition, if you dont run |
28 |
revdep-rebuild and things are breaking, it's your own fault |
29 |
-mike |