FYI: PkgBase installs older kernel and (slightly) newer world (at least in my example)

The Official PkgBase-build-based system install I'm now using
is based on the vintage range of main [so: 15]:

    • Fri, 04 Oct 2024
        . . .
        • git: d0af970e3ce4 - main - sysv shm: Fix SHM_REMAP flag value Brooks Davis 
        • git: 77fd02501f09 - main - src.conf.5: regen after diff3 changes Brooks Davis 
        • git: 74f6ec6fe3cf - main - lib{c,sys}: .note.GNU-stack in syscall stubs Brooks Davis

The kernel is reported to be based on d0af970e3ce4 above:

# strings /boot/kernel.GENERIC-NODEBUG/kernel | grep main-n
@(#)FreeBSD 15.0-CURRENT main-n272659-d0af970e3ce4 GENERIC-NODEBUG
FreeBSD 15.0-CURRENT main-n272659-d0af970e3ce4 GENERIC-NODEBUG

That d0af970e3ce4 supplied the most recent file in the /usr/src/sys/
it installed.

But the most recent source in /usr/src/ that it installed is
the source supplied by 74f6ec6fe3cf .

(I've little evidence about how big of a time frame difference
there can be. Here it is something like a couple of hours.)

So it appears that /usr/src/ ( excluding /usr/src/sys/ ) and
/usr/src/sys/ as supplied by PkgBase need not be from the same
checkout from the src repository: /usr/src/ can be newer than
/usr/src/sys/ .

My expectation is that the kernel was actually built based on
d0af970e3ce4 ( src/ and src/sys/ ) instead of being based on a
spliced src/ and src/sys/ tree with src/ being newer. As far
as I can tell, a spliced alternative when ether the kernel or
the world was built was historically never the basis for an
official build.

If building one of kernel/world ever involves files from both
/usr/src/ and from /usr/src/sys/ with a source tree that is
spliced, the result is not match to any just-one commit in
the repository.

Normally newer kernels support same or older worlds, not newer
worlds, because the newer world can depend on newer kernel
functionality/structure.

Thus, it appears that, if one is having odd problems with a
PkgBase system update that looks like the world and kernel
might be mismatched, it is something to consider/investigate.
Establishing the commit that the world build is based on is
not as simple as for the kernel's build.

(No claim my context has such a problem. I was looking around
for other reasons and happened to notice that the mismatch
is still the case in this modern example.)

===
Mark Millard
marklmi at yahoo.com