Fix potential NPE in AndroidExecutorImpl
Always wait for the startup latch to ensure the background thread's Handler is initialised before using it.
Closes#665.
See merge request !319
Refactor ValidationManager and fix some bugs
The validation manager uses a pattern of throwing an exception from within a transaction, then trying to ensure that the transaction gets committed anyway by setting it complete in a finally block. This is a bad idea - if the intentional exception isn't thrown but a later method throws an unintentional exception, the changes up to that point will be committed instead of rolled back.
I reworked the control flow to remove this pattern, and in the process realised that the validation manager was doing a lot of work to provide objects to attach to MessageStateChangedEvents, which are only used by tests these days. So I fixed that and a couple of other issues.
The VALID state is no longer needed and has been removed.
Closes#619
See merge request !313
Fetch RSS feeds via Tor
This patch replaces jsocks with our own minimal SOCKS 5 implementation, which is compatible with Android's OpenSSL hacks (see discussion on #599 for the horrifying details). This allows us to use OkHttp over Tor to fetch RSS feeds.
It turns out that SOCKS 5 without authentication is a really simple protocol: https://tools.ietf.org/html/rfc1928Closes#599.
See merge request !308
Before the introducee sends her ACK,
she derives a master key from the ephemeral shared secret as before.
Two nonces and a MAC key are then derived from the master key.
The local introducee signs one of the nonces and calculates a MAC
over her own identity public key, ephemeral public key,
transport properties and timestamp.
The local introducee includes the signature and MAC in her ACK.
On receiving the remote introducee's ACK,
the local introducee verifies the signature and MAC.
Should the verification fail, an ABORT is sent to the introducer and
the remote introducee that was added as inactive is deleted again.
Comments and reblogs need to depend on the post they refer to.
Since message dependencies are limited to one group,
the post and also the comments need to be wrapped
when commented on or reblogged to another blog.
For this reason, in addition to comments, two new wrapping message types
are introduced. They retain all data of the original messages and allow
for reconstruction and signature verification.
This commit breaks backwards compatibility with old blog posts.
It removes the content type, title and parent ID from the post
message structure.
Use Briar's IoUtils.copy(), not H2's IOUtils.copy()
Our implementation closes both streams, H2's implementation leaves them open.
Closes#614.
See merge request !293
Blog controller thread safety
This patch removes the mutable list of posts from the blog controller to make it thread-safe, and adds a cache of message bodies to speed up reloads.
Closes#555.
See merge request !276
Group strings and remove unused resources
Before I grouped the strings, I removed unused resources as well to not group things that are not being used anymore.
The grouping tries to follow a normal usage lifecycle of the app where the user works their way down the navigation drawer and explores all features in one area.
Closes#342
See merge request !277
Update Tor assets if they're older than the APK
Extract the Tor binary, GeoIP database and config file from the APK if they haven't been extracted since the APK was last updated.
On the Galaxy Nexus, skipping extraction of the binary if it's already up to date shaves about 1.5 seconds off the Tor plugin's startup time.
Closes#582.
Depends on !272.
See merge request !273
Don't connect to Tor if it's already running
For some time now we've had a reliable way of shutting down the Tor process (the __OwningControllerProcess command line argument combined with the TAKEOWNERSHIP command), but TorPlugin#start() still assumes that Tor may already be running. This allows another app to bind the Tor control and SOCKS ports and collect confidential data from Briar (#578). It also allows two Briar instances running on the same device to try to communicate with the same Tor process, which prevents proper shutdown (#572).
This patch prevents the Tor plugin from starting unless it's able to start its own Tor process with the expected control and SOCKS ports. If two Briar instances are running on the same device, only one of them will be able to use Tor. The other should fail to start its Tor plugin and then function normally without Tor access, including normal shutdown.
Fixes#572, #578. Open another ticket if you want two Briar instances on the same device to have their own Tor processes. :-)
See merge request !272
When we clear a list in onPause(), the proper behavior is that the empty state
message is shown, because the list is indeed empty.
However, we will reload the list content again in onResume(),
so the fix chosen in this commit is to force showing the progress bar
right after clearing the list.
Closes#576
UI for sharing blogs
Not posting any screenshots, because the UI is the same as for forums.
This does not yet offer the possibility to unsubscribe from blogs again. Should be done in a different MR as this one is big enough already.
Closes#498, #497
See merge request !263
This commit refactors the code for sharing forums,
so it can be used for sharing blogs as well.
It does not yet include code for responding to blog invitations.
* Implemented in briar-core as a `ScheduledExecutorService`
that gets started when the app starts
* The briar-api has a `FeedManager` interface
that the UI can use to register and unregister feeds
* In this first iteration, feeds are fetched via HTTP(S), not Tor
Closes#484
Torsten Groteandakwizgran