1 |
On Mon, 2021-09-27 at 19:14 +0000, Robin H. Johnson wrote: |
2 |
> I wanted to break the prior thread to discuss the root issue. |
3 |
> |
4 |
> We have some set of packages (A) which collectively depend on one or |
5 |
> more kernel options being set in specific ways, and the options need to |
6 |
> REMAIN set if you want the packages to continue work. |
7 |
> |
8 |
> There are also a subset of packages (B), usually kernel modules themselves |
9 |
> that will outright fail to compile if specific options are/are not set. |
10 |
> |
11 |
> Can we consider moving the checks for set A somewhere else, such that we |
12 |
> don't check the kernel config during package compile & install time, but |
13 |
> only check it later? This also meaningfully resolves that cases where |
14 |
> the system that has package building isn't where the packages are being |
15 |
> used. |
16 |
> |
17 |
|
18 |
I'm not sure if I understand you correctly but if you mean not doing |
19 |
checks before compiling/installing, then I have to disagree. There is |
20 |
value in knowing about this kind of problems early (hey, that's why we |
21 |
have pkg_pretend in the first place!) |
22 |
|
23 |
There's certainly value in knowing 'I need to rebuild my kernel' early |
24 |
vs learning only after you've spent significant time waiting for some |
25 |
package to build. |
26 |
|
27 |
-- |
28 |
Best regards, |
29 |
Michał Górny |