Remove stray boilerplate that is no longer relevant

Rebase after rP486510 and rP486511

Remove workaround for LLD on armv6/armv7, split to separate issue

In D18411#391857, @jbeich wrote:

If armv7 bustage isn't a regression from 0.36.0 update then it should be in a separate bug/review.

Considering.

Have you tried -Wl,-z,notext (a la rP484698)?

Second item tried after CFLAGS+=-fPIC; replace the above error output's can't create dynamic relocation R_ARM_V4BX with unrecognised reloc 40; same addresses as well.

The ARM SIMD and NEON assembler bits fail to link when LLD is the linker, even with 0.34.0. No mitigations (not even CFLAGS+=-fPIC) save for possibly forcing BFD use work. Last successful armv7 build on my machine was 120083.

Workaround for static ccache in poudriere. /ccache hierarchy isn't in $PATH.

Fix conditional, especially when running make config.

Mailman creates the config file upon first run, so specifying @sample in plist is not necessary. Furthermore such a directive doesn't play the nicest with autoplist.

Update to 3.2.0. Still QAing on my end, but at least the port-plumbing looks somewhat sane now.

make distribution no longer crashes with this.

Refresh based on current ports tree revision and use PY_FLAVOR for dependencies

All of these fail the checksum in the fetch phase primarily due to file sizes for the mbox patches changing from those set in distinfo.

Note that I'm not a committer (yet), so I can't commit this myself.

Use BINARY_ALIAS. How did I miss this part of the handbook?

Remove stray commented line again.

Rebase on current revision. gsed and gnustat patches are still needed to build properly; upstream claimed fixed but tested against a docker environment with CircleCI, not poudriere.

Remove stray/leftover commented Makefile target

This commit is possibly incomplete. I get a make[4]: "/usr/src/lib/geom/Makefile.inc" line 15: Could not find Makefile.classes immediately after the new geom objdir is created; make completely aborts here. Makefile.classes indeed does not exist in the new directory.

If a -L/usr/lib was missing, then linking would have also failed at -lz and a few other spots. This is without regard to whether this patch exists or not.

3.4 isn't affected by this last time I tested. In fact, upstream only patched 2.7, 3.6, 3.7beta and master; I manually backported the 3.6 patch to 3.5.

Regenerate with devel/arcanist

Regenerate with devel/arcanist

Regenerate with devel/arcanist

Regenerate with devel/arcanist and rebase on svn copy

Regenerate using git diff -U9999

Regenerate using git diff -U9999. Also remove superfluous USES python version.

Regenerate with git diff -U9999

I'm not so sure about having the mailman2 (mail/mailman) origin. Mailman 3 is a complete re-architecture, re-write, etc and thus has absolutely nothing in common with the old mailman apart from the name.