Get data directory value at app startup (Android)

[MarkusPettersson98]

Remove APP_PATH from mullvad-paths on Android since it should not be a constant value. Instead, it is passed down from the Android app startup. As it turns out, it is really ever used for pointing to the RPC socket in use.

---

This change is 

[linear]

DES-993 Provide data dir path to daemon during initialization

[Rawa]

@albin-mullvad mentioned that we also should update the README.md since it describes the "hard-coded" path.

See:
https://github.com/mullvad/mullvadvpn-app?tab=readme-ov-file#file-paths-used-by-mullvad-vpn-app

Data Directory (usually, /data/data/net.mullvad.mullvadvpn/)
https://developer.android.com/reference/android/content/Context#getDataDir()

Files Directory (usually, /data/data/net.mullvad.mullvadvpn/files)
https://developer.android.com/reference/android/content/Context#getFilesDir()

Cache Directory (usually, /data/data/net.mullvad.mullvadvpn/cache)
https://developer.android.com/reference/android/content/Context#getCacheDir()

Reviewed 16 of 16 files at r1, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @MarkusPettersson98)

---

mullvad-jni/src/lib.rs line 244 at r1 (raw file):

        data_dir,
        cache_dir,
        files_dir,

Just a small remark, should these be in the same order as above? data, files, cache.
or does it conflict with how it works on other platforms?

[Rawa]

The paths might be different depending on the device, but this is in broad terms for the normal use case correct. However, if the user make use of work profiles or similar features the paths can be different.

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @MarkusPettersson98)

[MarkusPettersson98]

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @Rawa)

---

mullvad-jni/src/lib.rs line 244 at r1 (raw file):

Previously, Rawa (David Göransson) wrote…

Just a small remark, should these be in the same order as above? data, files, cache.
or does it conflict with how it works on other platforms?

They might as well 😊

[dlon]

Reviewed 15 of 16 files at r1, 1 of 1 files at r2, all commit messages.
Reviewable status: all files reviewed, 2 unresolved discussions (waiting on @MarkusPettersson98 and @Rawa)

---

mullvad-daemon/src/management_interface.rs line 1042 at r2 (raw file):

                server_abort_rx.into_future().await;
            },
            &rpc_socket_path,

Nit: I think & can be removed.

---

mullvad-paths/src/lib.rs line 75 at r2 (raw file):

}

#[cfg(not(target_os = "android"))]

Do we not only ever use mullvad_paths for get_rpc_socket_path? Could we not conditionally compile all of this code and just define the socket path in the one module where it's called on Android?

[Rawa]

Reviewed 9 of 9 files at r3, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @MarkusPettersson98)

[MarkusPettersson98]

Reviewable status: 17 of 22 files reviewed, 1 unresolved discussion (waiting on @dlon and @Rawa)

---

mullvad-daemon/src/management_interface.rs line 1042 at r2 (raw file):

Previously, dlon (David Lönnhager) wrote…

Nit: I think & can be removed.

Done

---

mullvad-paths/src/lib.rs line 75 at r2 (raw file):

Previously, dlon (David Lönnhager) wrote…

Do we not only ever use mullvad_paths for get_rpc_socket_path? Could we not conditionally compile all of this code and just define the socket path in the one module where it's called on Android?

Solved this by not compiling mullvad-paths for Android at all. Instead, the path to the RPC socket is passed from the Android client, which means that the daemon does not have to care about it at all 😊

[Rawa]

Reviewed 5 of 5 files at r4, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @dlon)

[dlon]

Reviewed 7 of 9 files at r3, 5 of 5 files at r4, 2 of 2 files at r5, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved

---

mullvad-paths/src/lib.rs line 75 at r2 (raw file):

Previously, MarkusPettersson98 (Markus Pettersson) wrote…

Solved this by not compiling mullvad-paths for Android at all. Instead, the path to the RPC socket is passed from the Android client, which means that the daemon does not have to care about it at all 😊

Nice!

[dlon]

Reviewed 7 of 9 files at r3, 5 of 5 files at r4, 2 of 2 files at r5, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved

---

README.md line 357 at r5 (raw file):

| macOS | `/Library/Caches/mullvad-vpn/` |
| Windows | `C:\ProgramData\Mullvad VPN\cache` |
| Android | [`getFilesDir()`](https://developer.android.com/reference/android/content/Context#getFilesDir())  |

Is filesDir correct? I think we use the cache dir in this case?

[Rawa]

Reviewed all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @MarkusPettersson98)

---

README.md line 357 at r5 (raw file):

Previously, dlon (David Lönnhager) wrote…

Is filesDir correct? I think we use the cache dir in this case?

Nope, getCacheDir() should be the one

[Rawa]

Besides the getCacheDir remark it lgtm

Reviewed 1 of 1 files at r6, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved

---

README.md line 357 at r5 (raw file):

Previously, Rawa (David Göransson) wrote…

Nope, getCacheDir() should be the one

Since @MarkusPettersson98 is on vacay, I pushed a fix for this.

[dlon]

Reviewed 1 of 1 files at r6, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved