On 8.6.2014, at 16.14, Jilles Tjoelker <[email protected]> wrote: > On Fri, Jun 06, 2014 at 02:33:59PM +1000, John Marshall wrote: >> On Thu, 05 Jun 2014, 13:16 +0000, FreeBSD Security Advisories wrote: > >>> Corrected: > >>> 2014-06-05 12:33:23 UTC (releng/9.2, 9.2-RELEASE-p8) > >>> VI. Correction details > >>> Branch/path Revision >>> ------------------------------------------------------------------------- > >>> releng/9.2/ r267104 > >> I've just src-upgraded a system and expected to see OpenSSL version >> 0.9.8za at the end of it all. I checked the patches and the OpenSSL >> version number wasn't touched. Is this an expected outcome? > >> rwsrv04> uname -v; openssl version >> FreeBSD 9.2-RELEASE-p8 #0 r267130: Fri Jun 6 12:43:09 AEST 2014... >> OpenSSL 0.9.8y 5 Feb 2013 > >> rwsrv04> ls -l /usr/lib/libssl.so.6 >> -r--r--r-- 1 root wheel 304808 6 Jun 13:31 /usr/lib/libssl.so.6 > >> I understand that it was the FreeBSD distribution that was patched and >> not the OpenSSL distribution, but having the operating system and >> applications reporting a "vulnerable" version of OpenSSL isn't >> reassuring to other folks. > > Yes, this is expected and common practice. > > Perhaps the version number should instead be removed in head given that > it is not updated for security patches anyway. > > -- > Jilles Tjoelker
I strongly disagree. There has to be a version number so that no one has to guess what is base version of the software used. Instead I’d look into incorporating the patch level information that is now in ‘uname -r’ (for example '10.0-RELEASE-p5’) to various version strings in the world binaries. -Kimmo
signature.asc
Description: Message signed with OpenPGP using GPGMail
