PacketBuffer accepts bigger packets than it should

[mzero]

While it is hard to discern, there is a limit in CoreMIDI on the size of a packet in a MIDIPacketList. The docs only say:

The maximum size of a packet list is 65536 bytes. Large sysex messages must be sent in smaller packet lists.

Note also that the length field of MIDIPacket is only a UInt16. The actual packet size limit is smaller than that, after subtracting out the MIDIPacket and MIDIPacketList overheads from the max size of the list.

---

PacketBuffer will attempt to create a MIDIPacketList with a single MIDIPacket, large enough to accommodate however big a data object is passed in. If used with very large messages (think 80k sysex!) - PacketBuffer will build the structures, but the resulting thing will be corrupted, as it can't really hold a message that big.

PacketBuffer should at least make it clear in the docs that there is a limit on size.

---

How did I find this? I'm sending 80k sysexs (to control a synth) from a WebMIDI app.... and in Firefox on Mac it crashes the browser completely and instantly! I don't know who to blame for the crash... but the 80k byte array gets passed:

JavaScript -> Firefox/Gekko's C++ code -> Firefox/Gekkot's Rust crate midir_impl -> Boodlnagg's Rust midir package -> coremidi

[mzero]

Bug filed with midir: Boddlnagg/midir#165
Bug filed with FireFox: https://bugzilla.mozilla.org/show_bug.cgi?id=1945967

[mzero]

Ah, further investigation implicates this code (which appears twice in PacketBuffer):

        let current_packet_ptr = unsafe {
            MIDIPacketListAdd(
                packet_list_ptr,
                storage.capacity() as u64,
                current_packet_ptr,
                timestamp,
                data.len() as u64,
                data.as_ptr(),
            )
        };
        let current_packet_offset = unsafe {
            (current_packet_ptr as *const u8).offset_from(packet_list_ptr as *const u8) as usize
        };

If the data is too big to fit, then MIDIPacketListAdd returns null (as per CoreMIDI header files)... this will cause the computation of current_packet_offset to be some giant value.

[Boddlnagg]

The same problem exists with the call to MIDIEventListAdd in EventBuffer::push ...