> This is precisely what I was thinking. Except that unlike you, who don't
> think these minimal headers should be distributed with the kernel, I do.
I don't.
1) Any changes would have to go through Linus; Linus doesn't want
that, and we don't want Linus' time taken with this.
2) Some people want to use old kernels ("It does what I want"), but we
still want them to compile their programs using the latest version of
this "header package" (aka "LKILP" aka ?), so that that person's
programs will work even on very recent kernels.
3) People shouldn't have to keep the whole linux source tree on their
disk just to keep an up-to-date version of the LKILP.
(However, maybe 2 and 3 are non-issues for people who use
distributions.)
I think part of the reason for our disagreement is that I expect a
greater difference between the LKILP headers and the kernel headers
than you do.
I'm thinking that anything that #include's <linux/config.h> from user
space (as most kernel headers do) is wrong: the result of gcc foo.c
should be the same regardless of my current kernel config.
I hear a lot of people talking about namespace issues, and, to me,
this implies that name changes are required. (Note that I'm largely
ignorant of the namespace requirements, so maybe I overestimate them.)
pjm.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu