Also include Makefile update

In D17380#370866, @jkoshy_users.sourceforge.net wrote:

We would also need to check for 64bit-ness prior to using the e_hdr64 field.

I did not check the list but the approach sounds good to me

Only amd64 and i386 are handled at the moment

We discovered corrupt relocation info in BE mips64 modules and @sbruno bisected to this change - see PR 231790.

Seems reasonable to me.

OK - this is not a dependency for importing libevent 2.1.8 into vendor/libevent (and using it as libntpevent), so I'll proceed with that independent of this review.

In D10123#209648, @adrian wrote:

hi!

libevent2 is .. very big compared to the footprint of libevent1, even the base. If someone wants to do it as a separate pass and show we can get the size down to libevent1 size, then great. :)

Approved assuming @kib's updates are included.

In D17221#367105, @mmokhi wrote:

while we are here I'd like to mention this relevant review which maybe some people have opinions on as well 😅

rP480026

The variable substitution/evaluation is done at the end so it works both ways; my natural tendency would be to put them in the other order but this was the one used in the first instance of this approach in review D16813 / rP478120 by @mmokhi.

PR 230209