Very large number of buffer overflows exist in cscope <=15.5. Most deal with user controlled factors (enviorment variables, filenames), but theoretically, if someone inserted a carefully managed a #include statement on a largely distributed source (the kernel, firefox, open office, etc), they could overflow the buffer on a large number of computers.
vim or emacs herd, please check if we are really vulnerable (seems to be an old problem) and provide fixed ebuilds in case that we are, thank you.
vim/emacs teams: please advise
Can you provide a pointer to the list of vulnerabilities? I'm not sure what you're asking -- do you want us to do a code audit?
No, was asking if you could provide some insight on that problem, like if you know about a patch or a new version that we could bump to. The closest thing we have to a patch would be in : http://www.us.debian.org/security/2006/dsa-1064
It is my opinion that our port is vulnerable. cscope-15.5-r5.ebuild includes several patches but none of them address the 30+ potential buffer overflows the debian patch at http://security.debian.org/pool/updates/main/c/cscope/cscope_15.5-1.1sarge1.diff.gz addresses.
mkennedy, since you are in the emacs herd and said that we are probably vulnerable, could you please provide a fixed revbump?
revbumped to cscope-5.15-r6.ebuild w/ the following: src_unpack() { unpack ${A} # ~30 buffer overflows fix: Gentoo Bug #133829, patch developed by # the Debian Security Team (thanks to those guys), CVE-2004-2541, # Moritz Muehlenhoff. The Debian patch also includes the tempfile # fix (previously ${PN}-${PV}-tempfile.patch) epatch ${P}-debian-security.patch
arches please test and stable cscope-5.15-r6, thanks
cscope-15.5-r6 stable on ppc64
alpha stable.
ppc stable
amd64 stable.
sparc stable.
x86 done *~_~*
stable on hppa
GLSA 200606-10 arm, ia64, mips, s390 don't forget to mark stable to benifit from the GLSA.