Le jeu. 10 avr. 2025 à 12:14, Alexander Kanavin via lists.openembedded.org <alex.kanavin=gmail.com@lists.openembedded.org> a écrit :

On Wed, 9 Apr 2025 at 12:40, Jeroen Hofstee via lists.openembedded.org
<oe=myspectrum.nl@lists.openembedded.org> wrote:
> For completeness, this is not some theoretical issue, always calling if up
> directly after a network device is discovered does cause issues and is
> against spec as well.
>
> The sed stuff isn't that complicated, is it?

I beg to differ. The code is basically unreadable (unless you want to
write an elaborate comment with example input->output transformations
for every single line it), and that in itself is reason not to take
it. Any software solution must not only solve the problem, it must
also be written such that other humans can understand and change it.

I think I'd prefer a patch to busybox; even if upstream rejects or
ignores it, we still have something that is easier to take care of.

FWIW, I agree with Alexander here.

A patch (even a local one) will be easier to maintain in the long run.