Update DAITA to v2 on Windows

[dlon]

This PR updates Windows DAITA to v2 by reintroducing wireguard-go.

WIP

• Sign and pack DLLs
• Update build instructions
• Make sure multihop handles network changes correctly. wgRebindTunnelSocket seems to affect all peers rather than only the entry.
• Remove maybenot_machines (only from installer, since wgnt hasn't yet been removed)
• Update changelog
• Push tag, verify that dlls are signed
• Set DAITA platform to wg-go for windows (currently not included)
• Prepare build server before merging
• Benchmarks

Out of scope for this PR

• Remove wireguard-nt code for DAITA
• Use the multihop implementation of wireguard-go instead of WireguardNT

---

This change is 

[MarkusPettersson98]

Reviewed 6 of 13 files at r1, 6 of 10 files at r3, 2 of 3 files at r4, 1 of 4 files at r6, 7 of 9 files at r7, 1 of 1 files at r8.
Reviewable status: 18 of 20 files reviewed, 2 unresolved discussions (waiting on @dlon)

---

talpid-wireguard/src/wireguard_go/mod.rs line 230 at r7 (raw file):

impl WgGoTunnel {
    #[cfg(all(not(target_os = "android"), unix))]

⛏️ Change to #cfg[any(target_os = "linux", target_os = "macos")]

Code quote:

#[cfg(all(not(target_os = "android"), unix))]

---

wireguard-go-rs/src/lib.rs line 148 at r8 (raw file):

            handle: code,
            assigned_name: assigned_name.to_owned(),
            // SAFETY: The LUID is guaranteed to be intialized by wgTurnOn

⛏️ // SAFETY: wgTurnOn succeeded and the LUID is guaranteed to be intialized by wgTurnOn.

Code quote:

// SAFETY: The LUID is guaranteed to be intialized by wgTurnOn

[MarkusPettersson98]

Reviewed 2 of 3 files at r9, 1 of 1 files at r10, all commit messages.
Reviewable status: 19 of 20 files reviewed, 4 unresolved discussions (waiting on @dlon)

---

wireguard-go-rs/build.rs line 147 at r10 (raw file):

            go_build.env("CC", "zig cc -target aarch64-windows");
        }
    }

⛏️ Please document that we use zig as a C-compiler, and therefore we explicitly set CC for cgo. Maybe we should even state that we use it instead of MinGW, but this might not be the correct place for that kind of documentation.

Code quote:

    let target_arch = target_arch()?;
    match target_arch {
        Arch::Amd64 => {
            go_build.env("CC", "zig cc -target x86_64-windows");
        }
        Arch::Arm64 => {
            go_build.env("CC", "zig cc -target aarch64-windows");
        }
    }

---

wireguard-go-rs/build.rs line 293 at r10 (raw file):

                dest = dest.to_str().unwrap()
            )
        })?;

I don't think we want to embed the whole path in the error message. It was done for debugging purposes. But maybe logging the filenames could be nice?

Code quote:

        fs::copy(&src, &dest).with_context(|| {
            format!(
                "Failed to copy {src} to {dest}",
                src = src.to_str().unwrap(),
                dest = dest.to_str().unwrap()
            )
        })?;

---

wireguard-go-rs/build.rs line 418 at r10 (raw file):

    for export in libwg_exports {
        writeln!(file, "\t{export}").context("Failed to output exported function")?;
    }

⛏️ Skip the intermediary libwg_exports

    for path in [
        "./libwg/libwg.go",
        "./libwg/libwg_windows.go",
        "./libwg/libwg_daita.go",
    ] {
        for export in gather_exports(path).context("Failed to find exports")? {
            writeln!(file, "\t{export}").context("Failed to output exported function")?;
        }
    }

Code quote:

    let mut libwg_exports = vec![];
    for path in &[
        "./libwg/libwg.go",
        "./libwg/libwg_windows.go",
        "./libwg/libwg_daita.go",
    ] {
        libwg_exports.extend(gather_exports(path).context("Failed to find exports")?);
    }

    for export in libwg_exports {
        writeln!(file, "\t{export}").context("Failed to output exported function")?;
    }

---

wireguard-go-rs/build.rs line 28 at r9 (raw file):

        Os::Linux => build_linux_static_lib(&out_dir)?,
        Os::MacOs => build_macos_static_lib(&out_dir)?,
        Os::Android => build_android_dynamic_lib()?,

⛏️ build_android_dynamic_lib should probably also take out_dir as an argument (just to be consistent)

Code quote:

Os::Android => build_android_dynamic_lib()?,

[MarkusPettersson98]

Reviewed 1 of 10 files at r3.
Reviewable status: all files reviewed, 5 unresolved discussions (waiting on @dlon)

---

wireguard-go-rs/libwg/libwg_windows.go line 4 at r10 (raw file):

 *
 * Copyright (C) 2017-2019 Jason A. Donenfeld <Jason@zx2c4.com>. All Rights Reserved.
 * Copyright (C) 2024 Mullvad VPN AB. All Rights Reserved.

⛏️ Copyright year should be updated!

Code quote:

 * Copyright (C) 2024 Mullvad VPN AB. All Rights Reserved.

[dlon]

Reviewable status: 16 of 28 files reviewed, 1 unresolved discussion (waiting on @MarkusPettersson98)

---

wireguard-go-rs/build.rs line 293 at r10 (raw file):

Previously, MarkusPettersson98 (Markus Pettersson) wrote…

I don't think we want to embed the whole path in the error message. It was done for debugging purposes. But maybe logging the filenames could be nice?

Done.

[MarkusPettersson98]

Reviewed 12 of 12 files at r11, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved

[MarkusPettersson98]

Reviewed 1 of 1 files at r13, 2 of 2 files at r14, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved

[MarkusPettersson98]

Reviewed 2 of 2 files at r15, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved