As per #10628 (comment)
the real fix will be in 0.64.

This reverts commits:

Revert "linkers: Add a representation for the Apple AR Linker"
ee16f01.

Revert "backends/ninja: run `ranlib -c $out` when using the apple ar"
77e589c.

Revert "tests: Test extern'd globals on MacOS with the Apple Archiver"
bcb382b.

Fixes #10695

install_mode can include the setuid bit, which has the special property
(mentioned in the set_mode logic for minstall itself) of needing to come
last, because it "will get wiped by chmod" (or at least chown).

In fact, it's not just chown that wipes setuid, but other changes as
well, such as the file contents. This is not an issue for install_data /
custom_target, but for compiled outputs, we run depfixer to handle
rpaths. This may or may not cause edits to the binary, depending on
whether we have a build rpath to wipe, or an install rpath to add. (We
also may run `strip`, but that external program already has its own mode
restoration logic.)

Fix this by switching the order of operations around, so that setting
the permissions happens last.

Fixes void-linux/void-packages#38682

In a bunch of places we need to list various types of libraries
including custom_target outputs, and it gets very long. Use a common
T.Union for this.

These make no sense as single strings.

e.g. ldc -- the compiler needs to process args before consuming them.

Fixes #10693

This reverts commit 99ad11b.

This breaks working setups in conda. It is either wrong or incomplete
and thus cannot be used.

Fixes #10737

Need to link the qt4 install so it's in PATH.

The returned GResource is transfer-none, since the generated function
basically calls g_static_resource_get_resource(). It should not be
unreffed. Causes an abort on Debian:

GLib-GIO:ERROR:../../../gio/gresource.c:1451:g_static_resource_fini: assertion failed: (g_atomic_int_get (&resource->ref_count) >= 2)