Subject: Re: syslog_r (Re: CVS commit: src/lib/libc)
To: None <firstname.lastname@example.org>
From: Christos Zoulas <email@example.com>
Date: 10/26/2006 18:04:52
In article <20061026174938.GB6101@britannica.bec.de>,
Joerg Sonnenberger <firstname.lastname@example.org> wrote:
>On Thu, Oct 26, 2006 at 12:53:59PM -0400, Christos Zoulas wrote:
>> The ssp library.
>Beside all other parts of this discussion, please *don't* make it a
>separate library. It doesn't add any value and just introduces a
>dependency loop when the system itself should be built with ssp. Merge
>it into libc where it belongs.
>This doesn't change the situation for the kernel at all, since libkern
>has to provide a different initialiser and fault handler anyway.
The major reason gnu supplies a separate ssp library is because
they also replace the following functions with versions that check
for stack overflows:
memcpy(3), memmove(3), mempcpy(3), memset(3),
stpcpy(3), strcat(3), strcpy(3), strncat(3), strncpy(3),
vsnprintf(3), vsprintf(3), snprintf(3), sprintf(3),
We have to decide what we want to do. Do we want to supply separate
ssp implementations for those or not?