I guess we did not observe a problem here in the OpenSSL 3 experiments because the deprecation checks only activate when OPENSSL_API_COMPAT is set.

Looks like a sensible approach for 12.x

anyhow, I think it's reasonable to commit this version and update if/when we change factor to disallow non-integer args

When it's time to merge this I'd suggest squashing in the OPENSSL_API_COMPAT addition so that usr.bin/factor/Makefile does not change on stable/13, but we retain the "cherry picked from" line.

Don't forget to add a Fixes: trailer upon commit

Run: https://cirrus-ci.com/build/6750220826968064

also add arm64-llvm16 job

(Copied from the GitHub pull request) So far I am aiming for minimal diffs, and plan to make another pass over everything later on including updates to use OpenSSL 3 APIs. In particular I don't want to delete SSL_library_init(); if other pre-OpenSSL-1.1 code remains (I haven't checked whether or not it does, yet).

For reference, in current OpenSSL 1.1:

Can you please provide the author info as you'd like it to be, for git commit --author=...

In D39806#916606, @markj wrote:

According to our general compatibility policy I'd presume to be able to use 13.0 kgdb to debug 13.x vmcores. If so, that precludes MFCing the kernel patch, no matter which approach is taken.

Userland cpuset_t increase has been reapplied.

IMO for privatelib we can leave SHLIB_MAJOR alone

AFAICT "real" refers to it being time of day, rather than not-cached.

I'm not sure if this one needs to be the non-get version, could be getnanotime?

Oops @carlavilla just missed your review before pushing.

Combine new note into existing block

I think @jrm is right, this can just go in the existing box:

In this case, I *think* you can just go ahead even though I don't have a doc bit either.

(You can commit to doc/ as well with approval.)

This was added in 8da7e46fae2f in fact. Perhaps it was by accident?

This LGTM. I suspect "All OK" might have intentionally had no default (to force the admin to pause and confirm) but IMO it really does not have a benefit.

yes let's add the man page in here too

IMO bsd.compiler.mk (enabling for GCC) can be committed separately

IMO this shouldn't sort first, perhaps it should go along with Sponsored by:

Are we aware of any actual Linux binary use cases for argc=0?