- [BUGFIX] min RTO delay is 200 ms , not 1 second.
- [BUGFIX] calculate TLP timer correctly when only one packet is in
flight.
- [BUGFIX] discard CRYPTO frames from lower levels after connection
promotion.
- [BUGFIX] cancel path response if path could not be initialized.
- [BUGFIX] if push promise fails, do not invoke hset destructor.
Need to call SSL_set_quic_early_data_context() to enable early data
in session tickets issued to the client. This change was introduced
in when we switched to newer BoringSSL.
This is a regression introduced when switched to the new BoringSSL API
in. In the new APIs, read and write secrets are installed separately
and at different times. The previous logic that checked context
initialization can no longer be used. The new guard differentiates
between read and write secrets.
In ID-28, using TLS middlebox compatibility mode is forbidden and
we need a later version of BoringSSL for it to do the right thing.
This also means we had to update our code, because BoringSSL's QUIC
API has changed.
From the spec:
Prior to validating the client address, servers MUST NOT send more
than three times as many bytes as the number of bytes they have
received. This limits the magnitude of any amplification attack that
can be mounted using spoofed source addresses. In determining this
limit, servers only count the size of successfully processed packets.
Because connections' tickability depends on the ability of the engine
to send packets (ENPUB_CAN_SEND is checked), this property should be
recalculated when sending is reenabled in the engine via
lsquic_engine_send_unsent_packets()
- [FEATURE] QUIC timestamps extension.
- [API] New: ea_alpn that is used when not in HTTP mode.
- [BUGFIX] SNI is mandatory only for HTTP/3 and gQUIC.
- [BUGFIX] Benign double-free -- issue #110.
- [BUGFIX] Printing of transport parameters.
- [FEATURE] QUIC and HTTP/3 Internet Draft 27 support.
- [FEATURE] Add experimental delayed ACKs extension.
- Drop support for Internet Draft 24.
- Code cleanup.
- [BUGFIX] Send HANDSHAKE_DONE only after Finished is received.
- [BUGFIX] Don't treat garbage UDP padding as library error; ignore
it instead.
- [BUGFIX] Fix compilation on FreeBSD (missing header).
