kernel/modules-closure.sh: fix the modules closure#66235
kernel/modules-closure.sh: fix the modules closure#66235xaverdh wants to merge 1 commit intoNixOS:masterfrom
Conversation
|
Can you add a comment in the code explaining what that regex does? And the commit message could explain why the change was necessary and e.g. what new parts the regex doesn't include/newly includes. |
bf85c5b to
0500276
Compare
|
Should I include a reference to the source at kernel.org in the commit message? |
0500276 to
ab3fd7c
Compare
|
Also I had to touch that line some more, to place the comment somewhere sane. While doing that, I noticed, that I don't understand how the exit status test is supposed to work (even in the original version). I think its supposed to cover the case where |
|
I opened a separate pr which should fix the check for missing modules as well. |
Prior to this commit the modules closure could reference nonsensical pieces of data in to form of module parameters. These were treated as seperate modules. "modprobe --show-depends" outputs a line of the form "insmod module_name module_args", we are only interested in the module name, so we extract it with sed.
ab3fd7c to
d40ba64
Compare
|
Ping |
|
Looks good to me, though I haven't tested it. How does this PR compare to the other one? |
|
The second one
|
|
Thank you for your contributions. This has been automatically marked as stale because it has had no activity for 180 days. If this is still important to you, we ask that you leave a comment below. Your comment can be as simple as "still important to me". This lets people see that at least one person still cares about this. Someone will have to do this at most twice a year if there is no other activity. Here are suggestions that might help resolve this more quickly:
|
|
closed in favour of #92081 |
Prior to this commit the modules closure could reference nonsensical
pieces of data in to form of module parameters. These were treated as
separate modules.
Motivation for this change
resolving #57421
Things done
I tested this on my own device, but since this touches a core part, it should probably get some real testing.