On 07/14/2010 02:33 PM, Ken Moffat wrote: > > If you weren't using the package users hint, the problem would not > arise. Therefore we can assume that allowing the kernel's headers > to be installed is the well-tested choice. > You are absolutely right, but I would not characterize it as a problem, but something to be considered and about which to make a decision. In fact, it's the reason I posted. Using only the instructions in the LFS book, this header file would be replaced three times--which may be the answer to the questions. It's first installed by the linux ABI headers in Ch. 6 and then by GlibC in Ch. 6. Then it would be replaced again when the kernel was compiled and installed. So maybe the answer is just to use the kernel version. > In the sense that LFS is "your distro, your rules" you are, of course, > free to do things differently. I've no idea *which* packages in a > fully-built system use these particular headers (probably, not much > in LFS itself). Taking a look at the two sets of headers, they aren't > *that* different, so it is possible that both *might* work. > Duh! Occam's Razor again!!! My bad. I ran <diff> on the two files and if volume is the measure, there's a whole lot more to the kernelt version. But your suggestion is great. I could patch one of the files, and, if it were overwritten by the kernel installation, replace it with the patched version.
Thanks, Ken Dan -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
