On Thu, 2008-02-14 at 10:46 -0600, inode0 wrote:
> On Thu, Feb 14, 2008 at 9:15 AM, Sharpe, Sam J
> <[EMAIL PROTECTED]> wrote:
> >
> >  On Thu, 2008-02-14 at 08:49 -0600, Daryl Herzmann wrote:
> >  > I was puzzled to see the most recent kernel errata come out with a 
> > version
> >  > number of "2.6.18-53.1.13" compared with the last kernel of "...6", 
> > that's
> >  > 7 bumps for fixing 1 security issue?
> >
> >  That refers to the build of the RPM, not the version of the kernel.
> >  2.6.18 is the Kernel version, 53.1.13 is the RPM version number.
> >
> >
> >  When I build an RPM, I change the version every time I make an update to
> >  the RPM, whether it fixes an issue or it fixes a fix to an issue.
> >
> >  Lets say I apply a patch, so I bump the RPM version up. I then find out
> >  that the patched binary didn't work properly, so I redo it (new
> >  version). I then find out that fixing this broke some other thing, so I
> >  correct that and rebuild (new version number). I then find out I forgot
> >  to update the changelog in the .spec file, so I update that and rebuild
> >  (new version number).
> >
> >  I can quite easily believe that including one patch requires multiple
> >  RPM rebuilds and therefore a huge number of version changes in the
> >  RPM... which explains why it's 2.6.18-53.1.13 rather than 2.6.18-11 (I'm
> >  making up the 11 versions, I think there's been loads more kernel RPM
> >  releases in RHEL4.)
> 
> If you actually look at the changelog you will see exactly which fixes
> went in at each level from .7 through .13 ... no gaps, no holes, no
> oops I need to rebuild this rpm.

err, that's nice. However, if you actually *read* the changelog you
might ask: "what happened between 2.6.18-10 and 2.6.18-12... Where is
the changelog for 2.6.18-11?" or "I know what -10 fixed and I know what
-12 fixed, but what did they slip into -11 that isn't listed?"

The point I was making is that the 53.1.13 is an RPM version number, not
related to how many issues were necessarily fixed in the kernel. I did
that by imaging a number of build steps that might be required just to
fix a single issue... 

You also provide the counter example - a single version bump contains
multiple patches and therefore fixes multiple issues.

--
Sam

_______________________________________________
rhelv5-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/rhelv5-list

Reply via email to