It looks like @mat just committed:

Be consistent with option helpers

In D42758#975365, @michaelo wrote:

I wonder why this port not uses "USE_GITHUB": https://docs.freebsd.org/en/books/porters-handbook/makefiles/#makefile-master_sites-github

Use new upstream site rather than relying on forwarding

The problem here is that upstream expects us to either use the published tarballs or have git available to determine the version. Using the published tarballs is also preferable because the ones generated from GitHub are not indefinitely stable. I took a shot at redoing the port using upstream's published tarball and also made portclippy/portfmt happy.

I, perhaps naively, still prefer your approach in https://reviews.freebsd.org/D37005. However, I may not appreciate valid cases where /usr might not be mounted. This is certainly a step in the right direction.

A key argument for this change is that we align with what most Linux distributions are doing. Since Linux/systemd says it should be a symbolic link, some third-party software like Chromium or R runs into problems when the zoneinfo file is copied to /etc/. The arguments I've read against this change (https://twitter.com/kaevans91/status/1580652408861958144) describe (theoretical) problems when /usr isn't mounted. However, peter@ says,

"I’ve never found a case where a symlink doesn’t Just Work."

If none of our "Java people" like @glewis, or no one who has worked on certificate management like @des or @kevans has better ideas or objections by December 6, let's proceed. It's a non-default option, and it can help some enterprise users until a more robust solution comes along.

Good to see this made it into the tree. The Committer's Guide describes helpful metadata, such as Approved by: and Differential Revision: to include in your commit logs. See https://docs.freebsd.org/en/articles/committers-guide/#_include_appropriate_metadata_in_a_footer. Differential Revision: is helpful to refer a commit to review, and it also automatically closes the review after the commit. There is a commit log template you can install. See https://docs.freebsd.org/en/articles/committers-guide/#git-mini-primer-getting-started.

To test:

In D41904#970117, @jrm wrote:

Sorry for the delay. Looks good to me.

I spoke to @seanc about the cpu-microcode ports and will take maintainership but will wait for you to commit so you don't have to deal with a conflict.

Sorry for the delay. Looks good to me.

I heard back from @mat today on IRC. He's fine to proceed, given you build tested. @mat made another good point. If the man pages build fine without textproc/py-sphinx, let's remove the option and always install the man pages. Also, please apply the same change to dns/bind9-devel, unless I'm missing something and that port is different.

Let me ping mat@ on IRC. Are dns/bind918 and dns/bind9-devel the same?

Remove change that wasn't intended for this review.

I only have a few minor comments. All the poudriere testport I ran also look good.

Looks good. Don't forget about Differential Revision: https://reviews.freebsd.org/D42300. Phab isn't the best at showing spacing, but was the commit message derived from the local git hooks? It looks like there might not be two tabs after PR: and one after Approved by:.

In D42294#965110, @michaelo wrote:

In D42294#965109, @jrm wrote:

If it's easier, you can paste the full commit log in a comment.

I haven't pushed it yet, please have a look at the attached Git-formatted patch. If fine, I will try to push.

If it's easier, you can paste the full commit log in a comment.

This was just a test.

In D42294#965039, @michaelo wrote:

@jrm , I have now used:

osipovmi@deblndw011x:~/var/Projekte/freebsd/doc (michaelo)
$ git arc update HEAD

but fail to see that the commit message is actually visiable here...

It will close automatically with a line like Differential Revision: https://reviews.freebsd.org/D42293, but I just now notice it had Differential Review:.... No biggie, but you'll want to get the commit template set up to make your life (and ours) easier. :)

You should be able to commit to src. Have you tried? If so, it's probably a git hook that's preventing you. Can you show the output?

Could you also share your commit log here?

Committed.

@bofh Thanks again for confirming the poudriere results. All is clear now, and a workaround has been committed.

Thank you René. @timur, I hope you don't mind. I'll commit this to get it in before the Q4 branch.

Thanks for the heads-up. I'll update devel/freebsd-git-devtools after this lands.

I successfully built an image, did some basic tests, and didn't see any obvious problems. \o/. Hopefully, someone from RE or someone more familiar with the build system can take another look.

I added a few comments after a first pass. I'm unfamiliar with the src build system, so treat my comments/questions with that in mind. I'm going through a few tests now, and I'll add more comments here when those are done. Also, at some point, it would also be helpful to update release(7).

Ah, yes, I see you committed after I started this, but I think we were both late for the show. https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273954

Ok, thanks for letting me know. In the meantime, I tested with USE_GCC=yes and all was fine. If you can start building packages on POWER again, we can address any issues then.