I'm fine with these changes, even though they break the API. For my purposes, it isn't a big deal, especially with the version bump.

Thanks for the add!

@imp re the GENERIC linkage, note the original code only references GENERIC. However, a different location might be more appropriate, given this file is normally installed in /boot.

Given that there's no remote binary currently built, and that there's precedent for using .MADE, per https://svnweb.freebsd.org/base?view=revision&revision=337928, this appears to be the best solution.

FWIW, since lib80211 does not have an MK_WIRELESS exclusion, I don't think it makes sense to require it, but I do think it makes sense to ship with that library instead of ifconfig.

It is not clear from the diff, does this change include deleting etc/defaults/* as well?

In D16785#357262, @brd wrote:

In D16785#357233, @will wrote:

Sorta weird to set an empty FILES=, perhaps bluetooth.device.conf should be moved first, so this file can be moved along with deleting the etc/defaults directory altogether.

We do it in many places already, is it worth worrying about for a few days until the other move happens? The bluetooth move is a relatively bigger move so I was waiting a bit for that.

@rgrimes A lot of your questions would be resolved by reviewing how CONFS works in share/mk/bsd.confs.mk and its dependencies.

This looks good to me.

Sorta weird to set an empty FILES=, perhaps bluetooth.device.conf should be moved first, so this file can be moved along with deleting the etc/defaults directory altogether.

In D16742#356017, @0mp wrote:

Alright, those are fair reasons, @will.

I guess that the main reason of creating a manual page for beinstall.sh is having it show up search browsers and manpage browsers like http://man.freebsd.org/. Having a manpage for a command makes it more discoverable.

If you find this argument strong enough to install this manpage then we might consider committing these changes. Otherwise we should probably abandon this revision.

Hi Mateusz,

Issues using LINKS, on cursory inspection.

LGTM.

Assuming:

• The purpose of this particular change is to move only those things that would move to sbin/init. (Answer to my earlier question.)
• You've tested this specific change. (I'm sure I've tested a similar version...)

Summarizing discussion:

• chpass is the program that implements chsh which modifies /etc/shells, so it makes sense to install with this program.
• However, one could, strictly speaking, choose not to install chpass on a system, either because it will never update the shells list, or they edit it through other means.
• Additionally, /etc/shells is read by getusershell(3) in libc, which _is_ required by some programs, in particular su(1). So it could make more sense to install this file with libc instead.

This seems like a bug that should be fixed in install(1), not here. It's unclear from the description what the actual bug is.

Looks good to me, assuming you've tested this.

Hi, just a quick question.

Dropping this for lack of interest.

Hi, @mat experience also bit me, running FreeBSD-head (last update 12/18/2017). The pkg package itself is 1.10.3_1. I do not get a REMOVED section. I only have the FreeBSD repository configured.

@jhb Any further feedback? Does this look ready for HEAD?

On amd64, this now requires v2 minidumps (generated since 2010).

• libkvm: require v2 minidumps for walk_pages on amd64.

• libkvm: fix minor nits.

• kvm: pass up better error messages.

Committed.

Committed last summer in several parts.