Require explicit seeds for `Simulation` and `SimulatedEnvironment` to
ensure reproducible test results.
Also:
- Drop packets in `FakeUdpSocket` when the receive queue is full to
prevent unbounded memory growth during stress tests.
- Improve synchronization in `Simulation::run_until` by adding a timeout
to the barrier wait, preventing hangs if a runner is unregistered.
Introduces a new testing support library 'testing/support' that provides
a clean, modular, and fully deterministic environment for testing
toxcore components.
Replaces the use of the platform-specific `Network_Addr` in the public
`Network_Funcs` interface with the platform-independent `IP_Port`
struct, allowing higher-level abstractions (like simulated networks for
testing) to interact with the network layer without needing to depend on
or mock low-level OS socket headers.
It makes no sense to include it in the public API as clients can't make
any meaningful use of it via public API, it can only be used if one also
includes other internal/private headers that we don't install.
It's used only in the testing code, which has access to the internal
headers.
Fixes#2739, at least to some degree. I decided against moving things to
a separate `tox_testing.h` and leaving only things in `tox_private.h`
that we are fine with clients using, as otherwise `tox_lock()` /
`tox_unlock()` would have to be moved out of `tox_private.h` to
somewhere else, but `tox_private.h` actually sounds like the right place
for them, naming-wise. So perhaps it's fine if we have things in
`tox_private.h` that we don't want clients to use.
Sparse checks it. This is neater than using a struct, which has some
slightly weird syntax at times. This also reduces the risk of someone
adding another struct member.
These are more convenient and safer than the manual vtables we have in
the fuzzer support code. We can override individual member functions,
and C++ will take care of correctly casting and offsetting this-pointers
when needed.
So we don't need to write so many edge case tests ourselves for things
like parsers, which really don't need those manual tests, as long as we
can check for some properties like "can output the parsed data and it'll
be the same as the input".
Also flip some callback asserts, because they can be reached by fuzzing
eventually.
Also update the bootstrapd checksum, since the alpine image changed a
bit.
If the `recvbuf` network function returns 0 all the time, that means
there is never any data available on the TCP socket. This change makes
it so there is a random amount of data available on the TCP socket.
This invalidates the bootstrap fuzzer corpus.
It can't happen in almost every reality, except when the RNG is fairly
broken and doesn't add 2 fake DHT friends on startup. Still, this code
should be defensive and never index outside `num_friends` elements.
Ideally this would be able to reach some of the events, so we can write
code to respond to those events, but so far only the friend request
event actually happens.
These help creating fuzzer fixtures with non-trivially constructed
objects and takes care of cleaning them up afterwards so the fuzzer code
can focus on the system under test.
iphydf