That's great data, thanks!
I tried thin but it performed noticably worse and wasn't much faster but it probably scales better on multicore systems.

@zirias
This specific patch is a positive change so thank you however it would be nicer if we could avoid having local patches in the first place since you're one of a handful of people that works on libressl support.

Can we please stop adding sporadic libressl patches to the tree? Upstream and backport.

This should be a flavor rather than default?

@eduardo
Did you ever get around doing some build tests?

Ahh, I managed to missed that looking at shells. Thanks for the clarification!

@cy
Perhaps time to switch repo to https://github.com/ksh93/ksh
@brooks Submit it upstream if applicable?

Sites are sorted by protocol (HTTPS --> HTTP) and overall connectivity

I think we should try to aim for using LTO as much as possible so we should probably add a toggle later on. At least in texlab's case it clearly said that it wanted unwind instead of abort so by doing a full run we should get more data if it's "that simple" to determine. Other than I think it's it looks like a good proposal.

@mikael asked for some performance data and here's a small sample from my slow build box (Intel G3220 with 12Gb of RAM) with random sample of ports:

No, don't shoehorn this into one commit. This should be in a separate commit and all of this would preferably be committed after 2022-10-30 (maintainer timeout) unless I can get a portmgr approval to commit all pending fixups for ports in tree.
PRs: 267067, 267068, 267069, 267070, 267072, 267089, 267092, 267101, 267118, 267103, 267104, 267076

@tcberner
I would recommend make configure too to make sure WRKSRC doesn't get changed and breaks the port

Unrelated changes, (ports) should be split into separate commits.

Yes, set https://archive.apache.org/ as MASTER_SITES

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=225342 might also be of interest