the need for safe dynamic string libraries

[Dag-Erling Smørgrav]

Luigi Rizzo <rizzo at iet.unipi.it> writes:
> "Dag-Erling Smørgrav" <des at des.no> writes:
> > Restoring sbuf_printf() to what it was would not change the API, but
> > the semantics would be different in certain cases.
> doesn't seem a big deal, we already have diffent behaviour in
> kernel vs userland for certain functions (e.g. printf() itself,
> if nothing else the format specifiers are different in some cases).

There are parts of CAM that are shared between the kernel and userland
(which is why libsbuf was created in the first place)...  OTOH, I don't
think they use the sbuf API in such a way that they would be affected.

DES
-- 
Dag-Erling Smørgrav - des at des.no