On Sat, Mar 15, 2008 at 08:01:46PM -0400, Derrick Brashear wrote: > On Sat, Mar 15, 2008 at 6:22 PM, Axel Thimm <[EMAIL PROTECTED]> wrote: > > On Wed, Feb 13, 2008 at 04:49:48PM -0800, Darren Patterson wrote: > > > After building new kmod packages for the latest RHEL5 kernel I discovered > > > that rpm is very unhappy with the naming convention. Yum refuses to > > > install the package claiming that kmod-openafs-1.4.5-2.2.6.18_53.1.4.el5 > > is > > > newer than kmod-openafs-1.4.5-1.2.6.18_53.1.13.el5. To install this > > > package I have to use rpm with "--force". > > > > > > Red Hat updates their kernels in the above fashion regularly so I > > > anticipate this problem will continue to crop up in the not too distant > > > future. > > > > > > Before I create my own SRPM to work around this issue, are there plans to > > > deal with this? > > > > IMHO the kmod is flawed and what you see is the missing > > uname-r-in-name part of it. > > YHO doesn't help that he's trying to use an older minor revision of > the RPMs he's using with a newer minor revision of modules; The kernel > version is entirely irrelevant. > > foo-1-(`uname -r`) would still be older than foo-2-(`uname -r`).
No, look closer and you'll see that he has two different uname-r, not the same. And by the very construction of the uname-r-in-name scheme this comparison cannot and shouldn't be done. And "1" and "2" would be placed in the version/release field, where they belong, not the name part as it is implied above, but this will still make them upgrade-path separate packages. In the kmdl world these would be name - vr openafs-kmdl-2.6.18_53.1.4.el5 - 1.4.5-2 and openafs-kmdl-2.6.18_53.1.13.el5 - 1.4.5-1 As you see neither rpm, nor yum, nor smart/apt and any other depsolver will consider to compare these two. The kmdl scheme works w/o special treatment for each depsolver. Do take some time investigating the above writeup in the Fedora wiki. Many people have found it quite valuable and even if not everybody converts to kmdls at the very leats the uname-r-in-name part has become undisputable by now. Even by the former kmod authors. -- Axel.Thimm at ATrpms.net
pgp5ceC9AvKRT.pgp
Description: PGP signature
