Support all architectures of FreeBSD

[ripienaar]

While I doubt this supports all places FreeBSD runs, this replicates
a commit done in the upstream package that makes this usable on 32bit
FreeBSD

Signed-off-by: R.I.Pienaar [email protected]

[ripienaar]

/cc @smortex

[smortex]

@ripienaar Thanks!

As far as I can see, this repository and the upstream one only differ by a branch unix_race_condition that has never been merged to master nor upstream (a single commit: 4433868).

Maybe it's worth pulling all the commits from upstream if this work is still in progress, otherwise relying on the upstream repository directly instead of this one may make more sense? What is the status of this fork @aelsabbahy?

[ripienaar]

This dependency is being dropped in the next major release, so I am trying to do the least - with the least risk - to get you unblocked @smortex

[ripienaar]

In the v4 branch the dependency is gone goss-org/goss#597

[smortex]

Ok! Then for me the easiest is to mark the FreeBSD port of choria as ONLY_FOR_ARCHS=amd64 for now, and then attempt to remove this with the next release when the dependency is gone.

Thanks!

[ripienaar]

We could get the next choria release - which wold be before next goss major release - to have this fix, you can use mod replace feature in the mean time once this is merged to get your somewhere @smortex

[aelsabbahy]

Had to dig into this a bit to remember the history behind it.

So, goss is actually pulling the dependency from the unix_race_condition branch (pointing to the commit directly). It was supposed to be a temporary measure until upstream fixed it.

I submitted a PR a long time ago to upstream mitchellh#23 but it looks like it never got any attention, so I've just been using the forked/branch this whole time.

For reference, here's the goss bug that triggered all this: goss-org/goss#267

Maybe the proper fix is to create a goss branch and merge this PR and the unix_race_condition branch into it?

Or I can just rebase the unix_race_condition branch on the latest upstream master?

[ripienaar]

I can do my PR against that branch which is probably least effort. Will do in the morning unless you really feel like merging in upstream changes

[aelsabbahy]

Created a goss branch that points to the same commit as unix_race_condition

Feel free to submit the PR against the new branch.

[ripienaar]

PR updated to goss branch

[aelsabbahy]

Looks good, thanks!