1 |
Joseph Jezak wrote: |
2 |
> I don't have a problem with the "cp crap". It's obvious what it |
3 |
> does and make install doesn't work on all architectures. "make |
4 |
> install" isn't as explicit and may introduce other issues. |
5 |
> |
6 |
> -Joe |
7 |
|
8 |
Well, alrighty then. I hadn't figured "make install" would change its |
9 |
behavior so much on different platforms. Fortunately, I'd only commented |
10 |
out that stuff in a couple of handbooks, so I'll go back and revert that |
11 |
change. |
12 |
|
13 |
I did, however, have to fix some instances where the directions referred |
14 |
to cping vmmlinuz, but the *-bootloader files talked about vmlinux. |
15 |
That'll be the only thing to remain changed. |
16 |
|
17 |
What about just leaving it for x86+amd64? the "make install" copies |
18 |
everything properly for me whether or not /boot is a separate partition, |
19 |
and since I actually have the hardware (a whole 2 arches ;)), I can say |
20 |
with certainty it causes no issues. |
21 |
|
22 |
I'm always surprised when people find out there's a "make install" |
23 |
command to begin with -- it seems like it's one of the best-kept secrets |
24 |
of kernel compilation, right up there with "make gconfig". It doesn't |
25 |
cause problems on these two arches, and it's nicer than cping a buncha |
26 |
stuff and manually naming it. I'd like to keep it in for x86+amd64. |
27 |
Thoughts? |