* Fix Appveyor CI build
The problem was that when running commands in powershell, any stderr
output is treated as an "exception", stopping the build even if the exit
code of the command was good.
* Add workarounds to get Cirrus CI to pass
Removing the cache isn't great, but the build is still only 5-7 minues,
so it's not terrible either.
Disabling the tcp_relay test also sucks. It passes locally, so we're
likely getting caught in a firewall or something like it somewhere.
The default stack size for musl-libc is 128kb. Therefore we should try to keep stack
allocations well below this limit in order to avoid stack overflows.
* Don't overwrite the CMAKE_MODULE_PATH
* Allow linking libsodium statically with MSVC
* Allow finding libsodium the normal way on MSVC
* Allow using pthreads4w for pthreads on MSVC
* Fall back to find_package if pkg_find_module fails
* Don't pass incompatible compile flags to MSVC
* Also try to find Opus and libvpx using their canonical names
* Support building using conan
* Allow pkg_use_module to take a list of libraries to look for
* Build for Windows on Appveyor using conan
We no longer allow `int a, b;`. In the few cases where we used it, we
instead better
* limit the scope of the identifier (e.g. in a for-init-decl)
* split the line and have 2 separate declarators, because the
identifiers designate different types of things (e.g. friend numbers
and group numbers).
This check puts all of our code in a C++ anonymous namespace, which is
effectively making all functions `static`. This allows the compiler to
determine that a function is unused, so we can delete it.
* Use-after-free because we free network before dht in one case.
* Various unchecked allocs in tests (not so important).
* We used to not check whether ping arrays were actually allocated in DHT.
* `ping_kill` and `ping_array_kill` used to crash when passing NULL.
Also:
* Added an assert in all public API functions to ensure tox isn't NULL.
The error message you get from that is a bit nicer than "Segmentation
fault" when clients (or our tests) do things wrong.
* Decreased the sleep time in iterate_all_wait from 20ms to 5ms.
Everything seems to still work with 5ms, and this greatly decreases
the amount of time spent per test run, making oomer run much faster.
If you have UINT32_MAX friends, then adding one more friend will cause an
overflow of the friend list (wrap to 0) and result in all friends being
deleted. This subsequently results in a null pointer dereference when
we're trying to add one friend to the deleted friend list.
* Use fully static build for the bootstrap daemon.
* Store a sha256sum of the binary in the repo.
* Updated documentation for it.
* Add support for fully static build in cmake.
* Enable the docker build on every PR, so we catch changes to the
checksum. I realise this is adding toil, but having the checksum is
valuable for security of released binaries.
