sending.
When a peer leaves a group, they send a packet to the group
indicating that they're leaving. However if this packet is sent
via TCP, it gets put in a packet queue, which is then destroyed
on the rest of the group cleanup process before ever being able
to send.
This pr allows do_gc() to finish an iteration before cleaning the
group up, which allows the TCP packet queue to be emptied. However
this bug still exists on a tox_kill() event because we don't have
a chance to do another do_gc() iteration.
We really don't support it. I tried for half an hour to get some kind of
plan9 cross compilation to work, but it's not working. If anyone wants
to bring it back, they are welcome to send a PR including a CI check for
it. Until then, these 5 lines of unused code are gone.
There will be more object arrays that need to be packed. This function
takes care of NULL (creating an empty array), and putting the correct
array size and calling the per-element callback the right amount of
times.
- We no longer assert peer roles in the mod event callback
because this causes an issue with the new events implementation,
which triggers the events after all the packets from the
current tox_iterate() are processed, rather than as the
packets are received. These checks were superfluous and shouldn't
reduce code coverage.
- A moderator now sets the topic before the founder kicks him in
order to increase internal code coverage.
These are quite expensive, because they go through all events to index
in a typed array that no longer exists. Clients should index in the
union array and find the event they want themselves, or use dispatch.
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.
We allow non-null data pointers to be passed to functions
alongside 0-length data. For example when creating a data
buffer that has room for the entire packet, including ignored
header data.
This error broke a rare but legitimate case where we miss
packets during a handshake attempt and need to store empty
handshake packets in the packet array.
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".
