xref: /lsquic/CHANGELOG

2020-12-09
    - 2.26.0
    - [OPTIMIZATION] Adjust packet reordering threshold when spurious losses
      are detected.
    - [API] Pass pointer to local sockaddr to ea_get_ssl_ctx() callback.

2020-12-04
    - 2.25.0
    - [API, FEATURE] Add es_delay_onclose option to delay on_close until all
      data is ACKed.  Use new function lsquic_stream_has_unacked_data() to
      learn whether peer acknowledged all data written to stream.
    - [API] Add optional on_reset() stream callback to get notifications
      when RESET or STOP_SENDING frames are received.
    - [BUGFIX] On STOP_SENDING, make conn tickable is _writeable_, not
      readable.

2020-11-24
    - 2.24.5
    - [FEATURE] Improve Delayed ACKs extension and turn it on by default.
    - Limit receive history to a finite amount of memory.

2020-11-18
    - 2.24.4
    - [BUGFIX] Check whether ECN counts are set in ACK struct before using
      them.
    - [BUGFIX] Calculate TLP timer correctly when only one packet is in
      flight.
    - [BUGFIX] Min RTO delay is 200 milliseconds, not 1 second.
    - [BUGFIX] Memory leak in QPACK decoder handler: discard hset when
      necessary.
    - Allow retired and drained CIDs to be reused after a timeout.

2020-11-11
    - 2.24.3
    - [BUGFIX] Get rough RTT estimate on receipt of Handshake packet.
      This prevents BBR on the client from miscalculating pacing rate,
      slowing down sending of ACK packets.
    - [BUGFIX] Packets sent during handshake are app-limited.
    - [BUGFIX] Bandwidth sampler starts in app-limited mode.
    - [BUGFIX] Memory leak: free QPACK handler context in stream dtor.
    - Logging improvements.

2020-11-05
    - 2.24.2
    - [BUGFIX] Allow peer to migrate when its SCID is zero-length.
    - [BUGFIX] PADDING size calculation: only one Short packet can be
      coalesced.  (This should have been part of the fix in 2.24.1).
    - Abort connect if received NEW_CONNECTION_ID but current DCID is
      zero-length.
    - Improve log messages

2020-11-04
    - 2.24.1
    - [API] Allow use of ea_get_ssl_ctx() on the client (optional).  PR #186.
    - [BUGFIX] Expand datagram with ack-eliciting Initial to 1200 bytes
      after connection promotion.
    - [BUGFIX] Discard CRYPTO frames from lower encryption levels after
      connection promotion.
    - [BUGFIX] Cancel path response if path could not be initialized.

2020-10-28
    - 2.24.0
    - [FEATURE] QUIC and HTTP/3 Internet Draft 31 support.  Drop ID-30
      and ID-31 support.
    - [BUGFIX] Divide-by-zero in newly enabled conn stats code when no
      packets were sent.
    - [BUGFIX] Memory leak in gQUIC client when server hello cannot be
      parsed.
    - [BUGFIX] Server Initial packet size calculation.
    - Log user-agent and CONN_CLOSE reason when peer reports error.
    - Example programs: Specify ALPN for echo and md5 clients and servers
      (issue #184).
    - Example programs: Don't add "QUIC_" prefix to lines in keylog file
      (issue #185).
    - http_server: Fix fd leak in preadv mode; fix preadv() usage when
      reading from disk.

2020-10-22
    - 2.23.3
    - [BUGFIX] Update packetization threshold when writing to stream
      after packet size is reduced following an RTO.

2020-10-21
    - 2.23.2
    - Add QPACK stats collection and experimentation mode, see the new
      es_qpack_experiment setting.
    - Log busy connection stats every second using the new "conn-stats"
      log module.
    - Log about skipping only once.
    - Update HTTP/3 greased frame type formula.
    - Use ls-qpack v2.2.1.

2020-10-13
    - 2.23.1
    - [FEATURE] IETF Client 0-RTT support.
    - [BUGFIX] Do not schedule MTU probe on first tick.
    - [BUGFIX] Parsing DATAGRAM frame.
    - [BUGFIX] If push promise fails, do not invoke hset destructor.
    - [BUGFIX] Client: When connections are IDed by port number, check DCID.
      Fixes issue #176.
    - [BUGFIX] Regression introduced in 2.22.0: use correct number of
      PNSs for IETF mini conn during promotion.
    - Revert the 2.22.1 lsquic_is_valid_hs_packet change.  All that was
      necessary is a change to the way we call it in lsquic_engine.  No
      change to the function itself is required.

2020-10-08
    - 2.22.1
    - [BUGFIX] Function that checks validity of handshake packets.

2020-10-07
    - 2.22.0
    - [FEATURE] Extensible HTTP Priorities (HTTP/3 only).
    - [FEATURE] Add conn context to packet-out memory interface (PR #175).
    - [BUGFIX] gQUIC proof generation: allocate buffer big enough for
      signature (issue #173).
    - [BUGFIX] Make library thread-safe: drop use of global variables
      (issue #133, issue #167).
    - [BUGFIX] Deactivate only *recent* HQ frame, not any HQ frame.
    - [BUGFIX] gQUIC server: associate compressed cert with SSL_CTX,
      instead of keeping them in a separate hash, potentially leading
      to mismatches.
    - [BUGFIX] Stream data discard infinite loop: break on FIN.
    - cmake: add install target via -DCMAKE_INSTALL_PREFIX (PR #171).
    - Support randomized packet number to begin a connection.
    - Mini and full IETF connection size optimization.
    - http_client: specify HTTP priorities based on stream conditions.

2020-09-29
    - 2.21.0
    - [FEATURE] QUIC and HTTP/3 Internet Draft 31 support.
    - [API] Let user generate Souce Connection IDs.
    - [FEATURE] Allow building lsquic as shared library.
    - [OPTIMIZATION] Receive history: use a single contiguous memory
      block for everything.
    - Deprecate QUIC versions ID-27 and ID-30.

2020-09-25
    - 2.20.2
    - [BUGFIX] Memory leak: free pushed promise when refcnt is zero.
    - [BUGFIX] Memory leak in IETF full conn dtor: cleanup closed IDs sets.

2020-09-23
    - 2.20.1
    - [BUGFIX] Typo in new "validate peer addr by DCID" code.  It is
      a benign bug (works either way), but better to fix it.
    - Simplify Stream Priority Iterator (SPI).
    - Minor documentation updates.

2020-09-15
    - 2.20.0
    - [FEATURE] QUIC and HTTP/3 Internet Draft 30 support.
    - [FEATURE] Unreliable Datagram Extension support.
    - [FEATURE] Adaptive congestion controller.
    - [BUGFIX] Do not send MAX_STREAM_DATA frames on crypto streams.
    - [BUGFIX] Fail with CRYPTO_BUFFER_EXCEEDED when too much CRYPTO
      data comes in.
    - [BUGFIX] Spin bit is now strictly per path; value is reset on
      DCID change.
    - [BUGFIX] Check that max value of max_streams_uni and
      max_streams_bidi TPs is 2^60.
    - [BUGFIX] Close IETF mini conn immediately if crypto session
      cannot be initialized.
    - Deprecate ID-28 (no browser uses it): it's no longer in the
      default versions list.
    - New programs duck_server and duck_client that implement the
      experimental siduck-00 protocol.  They quack!
    - IETF crypto streams: don't limit ourselves from sending.
    - Command-line programs: turn off QL loss bits if -G is used, as
      Wireshark cannot decrypt QUIC packets when this extension is used.
    - Turn all h3 framing unit tests back on.
    - Fix malo initialization when compiled in no-pool mode.

2020-09-08
    - 2.19.10
    - [FEATURE] Add lsquic_stream_pwritev().  This function allows one to
      reduce the number of system calls required to read a file from disk
      by using lsquic_stream_pwritev() together with preadv(2).
    - [BUGFIX] When stream is reset, it is writeable -- let user collect
      the error.
    - [BUGFIX] Calculate correct conn flow control if reading ends early.
    - [BUGFIX] Remove stream from read and write queues on internal
      shutdown.  This is a regression introduced in 2.19.7.
    - [BUGFIX] Swapped arguments in IETF RESET_FRAME generation.
    - Turn off mini conn history when compiling with Visual Studio; this
      allows the project to compile on Windows again.
    - http_client: Add -3 flag to stop reading from streams early; code
      cleanup.
    - Don't use -Werror.

2020-09-02
    - 2.19.8
    - [FEATURE] Update the timestamp extension to latest version.
    - [FEATURE] Cope with appearance of ECN blackholes.
    - [OPTIMIZATION] return packno offset and size when header is generated.
    - [BUGFIX] ignore old ACK frames in mini conns.
    - [BUGFIX] Mark initial server path as initialized.
    - [BUGFIX] Do not merge ACK if ECN counts do not match.
    - Turn incoming packet number history in mini conn back on.
    - Record mini conn event history again when compiled in debug mode.
    - IETF mini conn: log when ACK is queued.
    - Clean up and refactor code in several places.

2020-08-26
    - 2.19.7
    - Handle ECT-CE event: issue a loss event.
    - Log the fact that we ignore SETTINGS_MAX_HEADER_LIST_SIZE.
    - Use Max Push ID in GOAWAY frame to cancel promises.
    - Add support for HTTP/3 CANCEL_PUSH frame.
    - lsquic_stream_is_pushed: streams without headers are never pushed.
    - [BUGFIX] Regression in lsquic_stream_shutdown_internal: now it shuts down.
    - Improve logic whether to generate CONNECTION_CLOSE.

2020-08-20
    - 2.19.6
    - Don't process incoming ECN marks if ECN is not enabled.
    - Schedule ACK when incoming packet is marked with CE.

2020-08-11
    - 2.19.5
    - [BUGFIX] Generate frame record when moving an ACK from one buffered
      packet to another.

2020-08-06
    - 2.19.4
    - [BUGFIX] Do not return an oversize MTU probe to connection twice.
    - [FEATURE] Delayed Acks updated to latest draft.  Still experimental.
    - Minor code cleanup in IETF full connection.

2020-08-04
    - 2.19.3
    - [BUGFIX] Regression in 2.19.1 that breaks Q050

2020-07-30
    - 2.19.2
    - [BUGFIX] Do not reduce PLPMTU size by network overhead.
    - [BUGFIX] Windows build.

2020-07-29
    - 2.19.1
    - [FEATURE] DPLPMTUD support.  IETF connections now search for the
      maximum packet size, improving throughput.
    - [DEBUG] Record event in stream history when on_close() is called
      in dtor.

2020-07-22
    - 2.18.2
    - [BUGFIX] Send prediction: lone path challenges do not get squeezed out
    - Fix crash in http_client: now -K and -B can be used simultaneously

2020-07-14
    - 2.18.1
    - [FEATURE] Implement the "QUIC bit grease" extension.
    - [BUGFIX] Selecting CID used for logging on client.
    - [BUGFIX] Header protection assertion.
    - [BUGFIX] Server: enable SSL key logging when cert lookup callback
      is not set.
    - Remove some dead code.

2020-07-06
    - 2.18.0
    - [API] Rename "0-RTT" to "session resumption." In IETF QUIC, "0-RTT"
      always refers to early data, meaning a request that the server can
      reply to in the very first return flight.  A more appropriate name
      for what we support on the client site is "session resumption," which
      is standard TLS terminology.  Later, when we add support for 0-RTT
      (early data), we can use the 0-RTT terminology again, this time in
      proper context.
    - [BUGFIX] Do not set certificate callback if ea_lookup_cert is NULL.
    - [BUGFIX] Make connection tickable when it's marked as closed.
    - [BUGFIX] Fail certificate lookup if SNI is not present in HTTP mode.
    - Several documentation fixes and improvements.
    - Minor code cleanup.

2020-06-24
    - 2.17.2
    - [BUGFIX] Infinite loop in stream: advance read offset when discarding
      data.
    - [OPTIMIZATION] Header protection: only initialize cipher once.
    - [OPTIMIZATION] Batch header protection application.

2020-06-18
    - 2.17.1
    - [FEATURE] QUIC and HTTP/3 Internet Draft 29 support.
    - [BUGFIX] Check that scheduled packets are also sendable when
      calculating a connection's "tickable" property.
    - [BUGFIX] Don't count scheduled packets as in-flight when pacer is
      checked on tick.
    - gQUIC: delay calling on_new for pushed stream until headers are
      available.
    - Allow nested calls to lsquic_engine_connect().

2020-06-15
    - 2.16.3
    - [OPTIMIZATION] Stash up to two reordered packets in IETF mini conn
      instead of dropping them.
    - [BUGFIX] Crash: check decrypt context before using it.  This regression
      was introduced in 2.16.2.

2020-06-12
    - 2.16.2
    - [BUGFIX] ID-28: do not use TLS middlebox compatibility mode in
      ClientHello.  This change requires using a newer version of BoringSSL.
    - [BUGFIX] Free connections in Advisory Tick Time Queue in engine dtor.
    - [BUGFIX] IETF QUIC client: narrow migration check to a single path.
    - [BUGFIX] NULL dereference: set function pointers for alarm for path
      challenges 2 and 3.
    - [BUGFIX] HTTP/3 headers may be followed immediately by trailers.
    - [BUGFIX] Log messages when SCID changes.

2020-06-09
    - 2.16.1
    - [FEATURE] Use "no-progress timeout" after which connection is closed.
    - [BUGFIX] Select new SCID when current SCID is retired.
    - [BUGFIX] Don't warn about dropped Initial packet sequence gaps during
      mini/full handoff.
    - [BUGFIX] Send correct conn error when HTTP/3 frame is truncated.
    - [BUGFIX] Mini conn: consider amplification when deciding to return
      TICK_SEND.
    - [BUGFIX] Don't double-count tag length in amplification logic.
    - [BUGFIX] Don't squeeze out lone path challenges.
    - [BUGFIX] Log messages dealing with scheduled packet queue squeezing.
    - [BUGFIX] don't wipe current path if no path challenge responses
      come back.
    - [BUGFIX] When path is reset, don't lose path_id which is used for
      logging.
    - Downgrade flow control violations to info log level from warnings.
    - Fix connection cap extra check, avoid checks in nested calls.
    - Fix some unit tests when extra checks are enabled.
    - Use ls-hpack 2.2.1.
    - Turn off unconditional extra checks for IETF clients.
    - Extra checks: don't verify sent size of hello packets.  Client
      changes DCID length and this check will fail.

2020-06-03
    - 2.16.0
    - [API] Use lsxpack_header v206.
    - [FEATURE] Windows supported.
    - [BUGFIX] Fix uninitialized variable use in client (regression in
      2.15.0).
    - Use ls-hpack 2.2.0.
    - Use ls-qpack 2.2.0.
    - Sample programs: fix the way maximum number of packets is
      calculated.
    - Remove some dead code.

2020-05-27
    - 2.15.0
    - [FEATURE] QUIC and HTTP/3 Internet Draft 28 support.
    - [BUGFIX] Ignore Retry packets after other packets are decrypted
      successfully.
    - [BUGFIX] Transport parameter decoding: CID no longer has 4-byte
      length minimum.
    - http_client: fix and optimize lsxpack_header allocator.
    - Drop support for Internet Draft 25.

2020-05-19
    - 2.14.8
    - Support Android.
    - Rerrange tree: move command-line examples into bin/ and unit
      tests into tests/ from test/unittests/.

2020-05-12
    - 2.14.7
    - [BUGFIX] ALPN-to-version mapping: do not skip h3-Q050.
    - [BUGFIX] Frame reader: skip headers if target stream is closed.

2020-05-06
    - 2.14.6
    - [BUGFIX] Fix amplification mitigation in 0-RTT case.
    - [BUGFIX] IETF mini connection should not tickable if cannot send
      a packet due to amplification.
    - [BUGFIX] Fail if active_connection_id_limit TP is smaller than 2.
    - [BUGFIX] Qlog server certificates for IETF QUIC connections.
    - [BUGFIX] Uninitialized struct padding usage in tokgen (benign).
    - [BUGFIX] Incorrect argument to shi_lookup() (benign).

2020-04-29
    - 2.14.5
    - [BUGFIX] In coalesced datagram, ignore packets whose CID does not match.
    - [BUGFIX] Frame reader: skip headers if target stream is not found.
    - [BUGFIX] Log message in QPACK decoder handler.

2020-04-24
    - 2.14.4
    - [BUGFIX] Heed es_rw_once for pushed HTTP/3 streams.
    - [BUGFIX] IETF client: set correct flags on bidirectional streams.
    - [BUGFIX] Generate Cancel Stream QPACK instructions for abandoned
      streams.
    - [BUGFIX] Do not call header callbacks after stream is closed.
    - Use ls-qpack 2.1.1

2020-04-15
    - 2.14.3
    - [BUGFIX] gQUIC: pass correct stream to hsi_create_header_set() callback.
    - [BUGFIX] Use ls-hpack 2.1.1
    - Improve stream code readability.
    - Use ls-qpack 2.0.5

2020-04-08
    - 2.14.2
    - [BUGFIX] Use ls-qpack 2.0.4
    - [BUGFIX] Honor max packet size on the client and when path changes.
    - http_server: fix prepare_decode() function.

2020-04-07
    - 2.14.1
    - [BUGFIX] Place connections on tickable queue when sending is reenabled.
    - [BUGFIX] A connection is tickable if it has unsent packets.
    - [BUGFIX] Heed peer's max_packet_size transport parameter.

2020-03-30
    - 2.14.0
    - [API] Use lsxpack_header structure to send HTTP headers.
    - [OPTIMIZATION] nocopy's readable_bytes() function.
    - http_server: fix typo in error message
    - Use ls-hpack 2.1.0.
    - Use ls-qpack 2.0.0.

2020-03-23
    - 2.13.3
    - [BUGFIX] ACK ping-pong: TIMESTAMP frame is not to be acked.

2020-03-13
    - 2.13.2
    - [BUGFIX] Use of new lsxpack_header API's hsi_prepare_decode().

2020-03-12
    - 2.13.1
    - [API] Use lsxpack_header structure to process incoming headers.
    - [BUGFIX] Fix assertion when zero-padding Initial packet.
    - [BUGFIX] Use idle timeout before we learn of peer's value.
    - Use ls-hpack 2.0.1 -- has lsxpack_header changes.
    - Use ls-qpack 0.14.0 -- new, common with ls-hpack, XXH seed (not
      used yet).
    - Code cleanup: prefix exported functions with "lsquic_".

2020-03-02
    - 2.12.0
    - [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.

2020-02-24
    - 2.11.1
    - [FEATURE] QUIC and HTTP/3 Internet Draft 27 support.
    - [FEATURE] Add experimental delayed ACKs extension.
    - Drop support for Internet Draft 24.
    - Code cleanup.

2020-02-14
    - 2.10.6
    - [BUGFIX] HTTP/3 framing: don't misinterpret rare occurence as error.
    - [BUGFIX] Send gap warning due to missing poisoned packet.
    - Stream unit test for scenario in issue #106.

2020-02-13
    - 2.10.5
    - [BUGFIX] BBR: call cci_sent() with correct arguments and at correct
      time.
    - Refactor transport parameters module.
    - Minor code cleanup.

2020-02-11
    - 2.10.4
    - [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).
    - Code cleanup: remove unnecessary #includes.

2020-01-31
    - 2.10.3
    - [BUGFIX] Cancel path responses and challenges on old path when
      switching to new path
    - Logging network path information.

2020-01-30
    - 2.10.2
    - [BUGFIX] Do not delay ACKs for Initial and Handshake packets.
    - [BUGFIX] Send PATH_CHALLENGE if path changed before mini conn
      promotion.
    - Logging improvements.
    - http_client: discard data faster.

2020-01-29
    - 2.10.1
    - [BUGFIX] Coalesced packets could get longer than normal packet
      size.
    - Add spin bit configuration option es_spin (-o spin=[01]).
    - Disable spin bit in 1/16 of connections.
    - Improve logging a bit.

2020-01-28
    - 2.10.0
    - [FEATURE] QUIC and HTTP/3 Internet Draft 25 support.
    - [API] Drop support for ID-23.
    - [BUGFIX] Set key phase bit on outgoing packets correctly.
    - Code cleanup.

2020-01-20
    - 2.9.0
    - [API] Drop support for Q039.
    - Improve ACK-queuing logic.  Send an ACK once in a while if
      peer keeps on sending non-ack-eliciting packets.
    - Improve Alt-Svc string: Q050 and later are not included in
      the old-style "quic" string.
    - Send stateless resets if connection could not be promoted.
    - Schedule MAX_DATA if needed when DATA_BLOCKED is received.
    - Use ls-qpack 0.11.2 -- needed for server push optimization.
    - Code cleanup: handle some error cases, improve logging.

2020-01-16
    - 2.8.9
    - [BUGFIX] Use ls-qpack 0.11.1
    - [OPTIMIZATION] Generate random bytes in batches.
    - Change loss_bits transport parameter ID to 0x1057 following
      latest draft.
    - Randomize period with which PINGs are sent to elicit ACKs.
    - Some refactoring and code cleanup.

2020-01-14
    - 2.8.8
    - [BUGFIX] Invalid read when parsing IETF transport parameters
      (this was benign).
    - [OPTIMIZATION] Frame bundling when using buffered packets in
      IETF QUIC: a) flush QPACK decoder stream and b) include ACKs
      in opportunistic fashion.
    - Fix HTTP/3 framing unit test.
    - Code cleanup.

2020-01-09
    - 2.8.7
    - [BUGFIX] Initial packet size check for IETF mini conn applies to
      UDP payload, not QUIC packet.
    - Support old and new school loss_bits transport parameter.
    - Use Q run length of 64 as suggested in the loss bits Draft.
    - Undo square wave count when packet is delayed.
    - Code cleanup; minor fixes.

2020-01-06
    - 2.8.5
    - [HTTP3] Verify number of bytes in incoming DATA frames against
      content-length.
    - [HTTP3] Stop issuing streams credits if peer stops opening QPACK
      decoder window.  This addresses a potential attack whereby client
      can cause the server to keep allocating memory.  See Security
      Considerations in the QPACK draft.
    - [BUGFIX] Mini conn: don't shorten max packet size for Q050 and later.
    - [BUGFIX] Init IETF connection flow controller using correct setting.
    - [BUGFIX] Fix unintended sign extension when removing header protection.
    - Code cleanup and minor fixes.

2019-12-30
    - 2.8.1
    - [FEATURE] Use occasional packet number gaps to detect optimistic
      ACK attacks.
    - [BUGFIX] Q050 client: all packet numbers are in the App PNS.
    - [OPTIMIZATION] Merge multi-range ACK frames, not just single-range
      ACK frames.
    - IETF QUIC: use RTT estimate in ack timeout calculation.
    - IETF handshake: abort conn when unexpected errors occur.
    - Use PING rather than MAX_DATA frames to elicit ACKs from peer.
    - Server: enforce 1200 byte Initial minimum packet size.
    - [CLEANUP] Remove code to disable gQUIC crypto.
    - [CLEANUP] Remove n_timestamps from ACK info struct.
    - Optimize driver: reuse previous ancillary message when possible.

2019-12-23
    - 2.8.0
    - [FEATURE] Add support for Q050.
    - [OPTIMIZATION] Reduce mallocs in gQUIC handshake.
    - [BUGFIX] Disable redo of failed STREAM frame insertion with debug
      logging.

2019-12-18
    - 2.7.3
    - [DEBUG] Further dedup next advisory tick messages when reason is
      the same.
    - [BUGFIX] Update size of `a` array in TP struct.  Fixes (benign)
      GitHub bug #94.
    - Use Cubic by default again instead of BBR, as it delivers more
      consistent performance.

2019-12-11
    - 2.7.2
    - [BUGFIX] Send controller: update scheduled bytes when DCID length
      changes (IETF client).
    - [BUGFIX] Drop alarm check from sanity test.  It no longer works now
      that we use loss chains.
    - [PORTABILITY] Fix build on Alpine Linux.
    - [PORTABILITY] Fix build using XCode.
    - Client initial DCID length: use RAND_bytes() instead of rand(3).
    - Add unit tests for connection min heap.
    - [DEBUG] Log CID in gQUIC handshake module
    - [DEBUG] Turn on extra checks for IETF client send controller.
    - [DEBUG] Dedup next advisory tick messages when reason is IDLE timer.
    - [DEBUG] QPACK decoder handler: log header error code.

2019-12-05
    - 2.7.1
    - [BUGFIX] client: don't call ignore_init() in middle of batch send.
      ignore_init() makes an assumption that the send controller has access
      to all outgoing packets.  This change wraps a few IETF full connection
      methods to delay calling ignore_init() until the engine returns all
      outgoing packets that were batched.
    - [BUGFIX] set errno to EAGAIN if sendmmsg() can't send all of them.
      This needs to be done because the value of errno may be lost on
      some platforms.
    - [BUGFIX] Typo that set all bits in sm_qflags lead to crashes.
    - [BUGFIX] Do not cancel header block processing after failure, as
      QPACK releases the reference in that case.
    - [CLEANUP] IETF encrypt: replace assert(0) with a warning.
    - Several small improvements to the test server.

2019-11-27
    - 2.7.0
    - [API, FEATURE] Close connection immediately when ea_packets_out()
      fails with errno != EAGAIN.  The API change is that errno is now
      examined.  Make sure to set it if using something other than
      sendmsg() to send packets.
    - [CLEANUP] Immediate close logic in IETF full conn.
    - [CLEANUP] Fix bogus warning about uninitialized `pair' variable.

2019-11-22
    - 2.6.7
    - [FEATURE] Implement the QL extension (offered by default).
    - [BUGFIX] Abort when encountering unexpected HTTP/3 frames.
    - [BUGFIX] Acknowledge (QPACK) HTTP/3 trailers correctly.
    - [DEBUG] Turn on debug message for next advisory tick.

2019-11-20
    - 2.6.6
    - [BUGFIX] Using HTTP/3 to HTTP/1.x converter.
    - [BUGFIX] Truncate log messages instead of throwing them away.

2019-11-15
    - 2.6.5
    - [BUGFIX] High priority buffered packet queue length.
    - [BUGFIX] Rain time calculation: max_ack_delay is in milliseconds.

2019-11-12
    - 2.6.3
    - [BUGFIX] Close DATA frames with empty payload correctly.

2019-11-11
    - 2.6.2
    - [BUGFIX] SCID!=ODCID rule applies to Retry packets, not regular
      packets.
    - [BUGFIX] Zero-RTT: BoringSSL no longer flips read/write secrets.
    - [BUGFIX] Truncate ACK frame rather instead of aborting IETF
      connection.
    - [BUGFIX] Client: don't send duplicate reset tokens.
    - [BUGFIX] Remove invalid assertion in H3 framing code.
    - Silence a warning in send ctl by restructuring switch() statement.

2019-11-08
    - 2.6.1
    - [BUGFIX] set retry token on all resubmitted packets.
    - Event log: log sent packet flags.  In particular, this allows one
      to see whether token was sent.
    - Don't migrate client if es_allow_migration is false.

2019-11-07
    - 2.6.0
    - [FEATURE] QUIC and HTTP/3 Internet Draft 24 support

2019-11-07
    - 2.5.2
    - [BUGFIX] argument order to gQUIC client constructor.  Regression
      introduced in 2.5.0.
    - [BUGFIX] split buffered packet error recovery: destroy the correct
      packet.

2019-11-04
    - 2.5.1
    - [BUGFIX] Fix double-free when emptying a packet number space.
    - [BUGFIX] http_server: fix md5sum handler: handle EOF correctly.
    - [BUGFIX] Use random values in bits 4 and 5 of the first byte of
      verneg packets (regression introduced in 2.5.0).
    - [OPTIMIZATION] Don't compile in expensive attq checks by default.
    - [OPTIMIZATION] http_server: compile regexes only once.

2019-10-31
    - 2.5.0
    - [API] lsquic_engine_connect() can now be passed QUIC version to use.
    - [OPTIMIZATION] Queue opportunistic ACKs if there is data to be sent.
    - [BUGFIX] Don't evict streams from priority iterator if there is
      only one queue.
    - [OPTIMIZATION, BUGFIX] Several other optimizations and bug fixes.
    - Use ls-qpack v0.10.7.

2019-10-24
    - 2.4.10
    - [BUGFIX] IETF QUIC server: fix uninitialized variable use.
    - [BUGFIX] make sure TLSv1.3 is not disabled in SSL object.
    - [BUGFIX] Use issuer name and serial number to cache certs (SKID
      values are not unique).
    - [BUGFIX] Always set the idle alarm in IETF connection so that it
      can time out.
    - Use ls-qpack v0.10.6, as it silences some warnings.

2019-10-21
    - 2.4.8
    - [OPTIMIZATION, BUGFIX] Use ls-qpack v0.10.5.

2019-10-15
    - 2.4.7
    - Add echo client and server to the distibution.
    - Add MD5 client and server to the distibution.
    - Fix http_client: check command-line arguments better, prevent crash.
    - Fix IETF conn: can_write_ack() should only care about APP PNS.
    - Client: delay stream creation until handshake succeds.
    - Reset HTTP stream whose write end is closed prematurely.
    - Fix tickable(): mirror behavior of tick() wrt buffered packets.
    - Log reason why engine is tickable.

2019-10-11
    - 2.4.6
    - Minor code cleanup and logging improvements.
    - Server and client programs: include library version (e.g. 2.4.6)
      into `server' and `user-agent' headers.

2019-10-08
    - 2.4.5
    - [OPTIMIZATION]: flush encoder stream only when necessary.
    - [BUGFIX] Use ls-qpack v0.10.2 for new API -- and for a bug fix.
    - [BUGFIX] Typo in IETF conn SETTINGS writer.
    - Use latest BoringSSL.

2019-10-08
    - 2.4.4
    - [API] Add lsquic_alpn2ver() to aid parsing Alt-Svc header.
    - [BUGFIX] NULL dereference when H3 frame header would be split.
    - [BUGFIX] Do not close fixed-size H3 frame prematurely.
    - [BUGFIX] Allow PING frames in IETF mini conn.
    - [BUGFIX] Mini conns: don't send any packets after receiving
      CONNECTION_CLOSE.
    - [BUGFIX] Client migration: reserve slot for DCID from transport params.
    - [BUGFIX] Allow max_early_data_size=0 -- early_data might not be there.
    - [BUGFIX] Use an invalid stream number to reset BPT cache (zero is now a
      valid stream number).
    - [SPEC] Use FINAL_SIZE_ERROR when FIN mismatch is detected.
    - [OPTIMIZATION] Closed connection only gets one chance to send packets.
    - [OPTIMIZATION] Flush headers stream before packetizing stream data.
    - [OPTIMIZATION] process QPACK encoder STREAM frames immediately.
    - Update ls-qpack to v0.10.1.

2019-09-30
    - 2.4.3
    - Add GQUIC versions to the list of h3 ALPNs for Alt-Svc header.
      For example, h3-Q043.  Chrome will switch to using this format at
      some point in the future.
    - [BUGFIX] Send correct value in max_streams_uni transport param
    - [SPEC] Abort IETF QUIC connection if max_early_data_size is not set
      to 0xFFFFFFFF

2019-09-23
    - 2.4.2
    - [BUGFIX] H3 framing: fix zero-byte write when space is available
    - [BUGFIX] Don't send STREAM frame when incoming unidirectgional stream
      is closed
    - [BUGFIX] Cancel all pending writes by stream reset by a GOAWAY
    - [BUGFIX] Fix use-after-free in IETF full conn
    - [OPTIMIZATION] Wait for session tickets for two seconds and then drop
      SSL object and crypto streams.

2019-09-18
    - 2.4.0
    - [FEATURE] QUIC and HTTP/3 Internet Draft 23 support

2019-09-13
    - 2.3.1
    - [BUGFIX] Fix memory leaks
    - [BUGFIX] Fix unit tests

2019-09-12
    - 2.3.0
    - [FEATURE] BBR congestion control is on by default
    - [BUGFIX] BBR app-limited logic
    - [BUGFIX] Fix uninitialized warnings in IETF
    - [BUGFIX] Update ls-qpack to v0.9.14
    - [CLEANUP] Code cleanup

2019-09-11
    - 2.2.0
    - [FEATURE] Server code is included in the library
    - [FEATURE] IETF QUIC and HTTP/3 Support (ID-22)

2019-05-13
    - 1.21.2
    - [OPTIMIZATION] HPACK: use history to improve compression performance

2019-05-06
    - 1.21.1
    - [BUGFIX] If FIN or RST not received, don't delay stream destruction.
    - [OPTIMIZATION] Speed up HPACK encoder by using same hash value to
      search static and dynamic tables.

2019-04-12
    - 1.21.0
    - [FEATURE] Add qlog log module.

2019-04-01
    - 1.20.0
    - [FEATURE] Add support for Q046.

2019-03-19
    - 1.19.6
    - [BUGFIX] Ensure that Largest Observed does not decrease in ACKs we
      send.  ACK frames placed in packets in buffered queues (optimization
      introduced in 1.17.15) can be preceded by an ACK frame generated later.
      In this case, the older ACK frame should not be sent out, as Chromium-
      based servers flags decrease in the ACK frame's Largest Observed value
      as an error.

2019-03-05
    - 1.19.5
    - [BUGFIX] Use correct public key from PUBS based on KEXS index.
    - [BUGFIX] Check flags before dispatching writes, avoiding assert.
    - [BUGFIX] Set :scheme to "https" (instead of "HTTP").

2019-02-25
    - 1.19.4
    - [BUGFIX] Check buffer bounds when looking up version in 0-RTT blob.
    - [BUGFIX] http_client: don't fetch 0-rtt info if handshake failed.
    - Log number of pacer calls at DEBUG, rather than NOTICE, level.

2019-02-18
    - 1.19.3
    - [BUGFIX] Q044: don't encode packet number in 6 bytes.  Six-byte
      packet number encoding does not exist in Q044.  This fixes a
      regression introduced in '[BUGFIX] Buffered packets can contain
      ACK frames' -- we need to keep QUIC version in mind when selecting
      the longest possible packet number encoding used for the buffered
      packet that carries the ACK.
    - [BUGFIX] Do not increase CWND when timeout occurs.
    - http_client: support setting handshake timeout on command line.
      Use -o handshake_to=timeout.
    - http_client: use -k to connect UDP socket to pick up ICMP errors.
    - http_client: allow pathless mode, when only handshake is performed
      without issuing any requests.  This can be done by simply not
      specifying a -p flag on the command line.

2019-02-11
    - 1.19.2
    - [BUGFIX] Begin negotiation with version provided in 0-RTT info.
    - [BUGFIX] Version checking in zero_rtt deserialize function.

2019-02-04
    - 1.19.1
    - [BUGFIX] Fix Windows build.

2019-02-04
    - 1.19.0
    - [FEATURE, API Change] 0-RTT support.  Add function to export 0-RTT
      information; it can be supplied to a subsequent connect() call.
    - [FEATURE] Add -0 flag to http_client to exercise 0-RTT support.
    - [BUGFIX] Resuscitate the Windows build.
    - [BUGFIX] Send HTTP settings (max header list size) if necessary.
    - [BUGFIX] Buffered packets can contain ACK frames.
    - [BUGFIX] Make packet writeable once all STREAM frames are elided.
    - [BUGFIX] Fix potential null dereference when realloc fails.
    - cmake: simplify build configuration.

2019-01-28
    - 1.18.0
    - [API Change] Can specify clock granularity in engine settings.
    - [BUGFIX] Pacer uses fixed clock granularity.  Since the change on
      2018-04-09, it is not necessary to try to approximate the next tick
      time in the pacer: it can use fix clock granularity specified by
      the user.
    - [BUGFIX] Do not tick constantly before handshake is done.
    - [BUGFIX] Do not exit tick prematurely: reuse packet with ACK.  Even
      if we cannot allocate *more* packets, we could still be able to write
      to one already allocated.
    - [BUGFIX] Do not schedule pacer if there are no lost packets.

2019-01-17
    - 1.17.15
    - [BUGFIX] http_client: make sure only one read per on_read() callback
      is performed in the header conversion bypass (-B) mode.
    - http_client: with -E, assign random priority when stream is created.
    - [OPTIMIZATION] On immediate write, place an ACK frame into the first
      buffered packet if an ACK is queued.  This reduces the number of
      standalone ACK packets.
    - [OPTIMIZATION] Allow placing more than one STREAM frame from the same
      stream into an outgoing packet.  This change minimizes the number of
      buffered packets required to store several small HTTP messages by
      virtue of allowing more than one STREAM frame from HEADERS stream in
      the same packet.
    - [OPTIMIZATION] Flush headers when writing to buffered packets.  This
      causes the headers to be written to the same buffered packet queue,
      thereby improving packet utilization, especially for small HTTP
      messages.

2019-01-16
    - 1.17.14
    - [FEATURE] http_client can now collect stats: time to connect, TTFB,
      time to download, requests per seconds, and bandwidth.  See -t flag.
    - [BUGFIX] http_client: -B, -K, and -r can be used at the same time.
    - [BUGFIX] http_client: do not display cert chain before each request.
      Now this is only done once per connection, if the handshake is
      successful and -a option is given.
    - [BUGFIX] Do not wait to dispatch on_close() after reading.  If a
      stream is scheduled to be closed after on_read() callback returns,
      close it immediately instead of waiting until the end of the tick.
      If client creates new request from on_close() event, they will be
      processed in the same tick instead of waiting for the next one.
    - [BUGFIX] HEADERS stream is critical: always treat it with highest
      priority.

2019-01-10
    - 1.17.12
    - [FEATURE] http_client can now issue parallel requests in the context
      of a single connection.  See -w option.

2019-01-03
    - 1.17.11
    - Fix strict aliasing warning in optimized compilation.

2018-12-27
    - 1.17.10
    - Fix the example program to be able to use parallel connections
      again.  (See the -n argument.)

2018-12-18
    - 1.17.9
    - [BUGFIX] Engine: reduce minimum batch size from 256 to 4

2018-12-10
    - 1.17.8
    - [BUGFIX] Fix compilation on FreeBSD and 32-bit Linux

2018-12-03
    - 1.17.7
    - [BUGFIX] Do not unset PING alarm before ringing expired alarms.
      This prevented PING from ever being sent.

2018-11-29
    - 1.17.6
    - Add failsafe: resume sending packets after some time

      The change puts a 1-second limit on the amount of time the engine
      will not send packets after some packets are delayed.  This makes
      the library robust in case the user does not unblock the engine
      explicitly using lsquic_engine_send_unsent_packets() call.

    - [BUGFIX] Handle corner cases in send controller when packets are
      a) delayed or b) dropped during repackaging.
    - [BUGFIX] Memory leak: destroy buffered packets during controller
      cleanup.

2018-11-16
    - 1.17.3
    - [BUGFIX] Do not send STOP_WAITING frames when using Q044

2018-10-19
    - 1.17.2
    - [BUGFIX] Memory leak in test_frame_rw unit test.
    - [BUGFIX] Parsing packets with short IETF header: do not overwrite
      flags.  (Only a problem in unit tests -- benign otherwise.)

2018-10-16
    - 1.17.0
    - [API Change] Packet out Memory Interface (PMI) update:
      - Split PMI pool return into pmi_release and pmi_return
      - PMI callbacks take peer_ctx and is_ipv6 arguments
    - [BUGFIX] Fix use-after-free when certificate is updated
    - Silence gcc warning in optimized mode by performing useless
      initialization
    - cmake: use the standard variable CMAKE_BUILD_TYPE instead of
      DEVEL_MODE

2018-10-03
    - 1.16.0
    - [API Change] Add lsquic_conn_n_avail_streams()
    - [BUGFIX] only dispatch crypto stream read events if WANT_READ is on

2018-09-27
    - 1.15.0
    - [API Change] Add LSCONN_ST_PEER_GOING_AWAY to the list of conn statuses
    - [BUGFIX] free uncompressed headers correctly when error occurs

2018-09-12
    - 1.14.3
    - [BUGFIX] Do not abort conn on STREAM frame for a reset stream
    - [BUGFIX] Drop packets that would become empty due to repackaging.
      Packets on the scheduled queue may be marked for repackaging.
      Frames such as ACK frame that are never resent are removed from
      repackaged packets.  We must check that the newly repackaged packet
      would not be empty.  If it would be, it is destroyed instead and
      the next packet on the scheduled queue is used.  Note that this
      change only affects the logic to return the next packet to be sent.
      Lost packets that are being rescheduled are already processed in
      this fashion.
    - Byteswap CID before logging it - this makes it match Chrome CIDs.
      (Except Q035, which is the last little-endian GQUIC version we
      support.)

2018-09-06
    - 1.14.0
    - [API Change] Disable packet sending if full batch cannot be sent
      If lsquic_packets_out_f() cannot send the whole batch, disable
      packet sending until lsquic_engine_send_unsent_packets() is called.
    - [BUGFIX] Handle case when STREAM frame does not fit.
    - [BUGFIX] Always allow incoming STREAM frames to overlap.  Peers
      may send overlapping STREAM frames even if using versions older
      than Q043.
    - Custom header set fixes:
      - set "FIN reached" flag when custom header with FIN flag is
        claimed;
      - do not return custom header set for a reset stream.

2018-08-27

    - 1.13.0
    - [FEATURE, API Change] Add ability to create custom header set
      objects via callbacks.  This avoids reading and re-parsing
      headers from the stream.

2018-08-27

    - 1.12.4
    - Fix memory leak when engine is destroyed
    - Fix memory leak in http_client
    - Fix gcc warning in unit tests

2018-08-22

    - 1.12.3
    - [BUGFIX] Fix duplicate STREAM frame detection

2018-08-20

    - 1.12.2
    - [BUGFIX] Update count of scheduled bytes when adjusting size of
      an already-scheduled packet.
    - Emit info instead of warning messages when stream is used in
      unexpected ways.

2018-08-17

    - 1.12.0
    - [FEATURE, API Change] Add support for certificate verification

2018-08-16

    - 1.11.1
    - [BUGFIX] Fix STOP_WAITING frame processing in the new Q044 code

2018-08-15

    - 1.11.0
    - [FEATURE] Add support for Q044.

2018-08-09

    - 1.10.2
    - [BUGFIX] Don't go over limit when creating delayed streams

2018-07-10

    - 1.10.1
    - [BUGFIX]  process connections after each batch of packets is read
      This avoids a problem of accumulating a very large list of packets
      (possible when speeds are high and socket's receive buffer is large)
      and processing it all at once.
    - If glibc is older than 2.17, link with rt.  This is necessary for
      clock_getres(2).
    - Add version macros to lsquic.h; remove unnecessary includes.

2018-06-13

    - [BUGFIX] allow multiple parallel connections by default

      Use the original method of tracking connections by CIDs by default.
      If zero-sized CID support is turned on, connections are tracked by
      the address.  A new connection is not created if another connection
      is using the same network address

2018-05-30

    - [FEATURE, API CHANGE] Support zero-sized CIDs in received packets

2018-05-24

    - Close connection properly when packet encryption fails

2018-05-23

    - [BUGFIX] Do not produce packet sequence gaps due to delayed packets

2018-05-21

    - [API Change] Add optional callback to call when handshake is done
    - [API Change, BUGFIX] After send failure, wait until transport available

2018-05-18

    - [API] Expose useful lsquic_ver2str[] in lsquic.h
    - [BUGFIX] Do not produce packet sequence gaps due to STREAM frame elision
    - Improve checks of number of incoming streams limit and associated
      error reporting.
    - [BUGFIX] Command-line option `-6` now works correctly.

2018-05-16

    - [FEATURE] DNS resolution
    - [BUGFIX] Frame insertion mis-ID as overlap instead of dup
    - http_client: fix priority range generated by -E flag

2018-05-09

    - [FEATURE] Add support for Q043.
    - Support for versions Q037, Q038, Q041, and Q042 has been removed.
    - Fix typo in debug message.
    - Fix code indentation.
    - Add /* fallthru */ comment to stop newer gcc from complaining.
    - Logger: fix compilation of optimized Windows build.

2018-05-04

    - [FEATURE] Add support for Q042.
    - Remove comment: MSPC is obsolete (no code changes)
    - Prog: use lsquic_str2ver() when processing -o version flag
    - Remove unused CTIM and SRBF transport parameters
    - Disable QUIC versions Q037 and Q038 by default
    - Fix Windows compilation by including compat header file in
      lshpack.c
    - Address warnings produced by newer versions of gcc
    - Future-proof: turn off -Werror

2018-05-02

    - [BUGFIX] Make lsquic_conn_quic_version() available
    - Switch to using ls-hpack 1.1
    - [BUGFIX] Do not ignore stream resets after receiving FIN

2018-04-27

    - HPACK: do not allow header block to end with table size update.

2018-04-25

    - [BUGFIX] Do not create gap in sent packnos when squeezing delayed
      packets.
    - [BUGFIX] sendctl checks for all unacked bytes, not just retx bytes.
    - [BUGFIX] connections with blocked scheduled packets are not tickable
      for sending.
    - [BUGFIX] Conn is tickable if it wants to send a connection-level
      frame.

2018-04-23

    - Fix busy loop: tickable must make progress.  When connection is
      self-reporting as tickable, it must make progress when ticked.  There
      are two issues:
        1. If there are buffered packets, the connection is only tickable if
           they can be sent out.
        2. A connection is tickable if there are streams on the servicing
           queue.  When the tick occurs, we must service the stream
           independent of whether any packets are sent.
    - Fix assertion in pacer which can be incorrect under some
      conditions.
    - cmake: do not turn on address sanitizer if in Travis.

2018-04-20

    - [BUGFIX] Fix bug in lsquic_engine_connect() exposed by yesterday's
      changes.

2018-04-19

    - [BUGFIX] Add connection to Tickable Queue on stream write
    - cmake: use MSVC variable instead of trying to detect
    - engine: improve connection incref/decref logging
    - stream: don't ignore errors that may occur on triggered flush
    - connection: remove obsolete method
    - engine: indicate connection as tickable if previous call went
      over threshold

2018-04-09

    [API Change, OPTIMIZATION] Only process conns that need to be processed

    The API is simplified: do not expose the user code to several
    queues.  A "connection queue" is now an internal concept.
    The user processes connections using the single function
    lsquic_engine_process_conns().  When this function is called,
    only those connections are processed that need to be processed.
    A connection needs to be processed when:

        1. New incoming packets have been fed to the connection.
        2. User wants to read from a stream that is readable.
        3. User wants to write to a stream that is writeable.
        4. There are buffered packets that can be sent out.  (This
           means that the user wrote to a stream outside of the
           lsquic library callback.)
        5. A control frame (such as BLOCKED) needs to be sent out.
        6. A stream needs to be serviced or delayed stream needs to
           be created.
        7. An alarm rings.
        8. Pacer timer expires.

    To achieve this, the library places the connections into two
    priority queues (min heaps):

        1. Tickable Queue; and
        2. Advisory Tick Time queue (ATTQ).

    Each time lsquic_engine_process_conns() is called, the Tickable
    Queue is emptied.  After the connections have been ticked, they are
    queried again: if a connection is not being closed, it is placed
    either in the Tickable Queue if it is ready to be ticked again or
    it is placed in the Advisory Tick Time Queue.  It is assumed that
    a connection always has at least one timer set (the idle alarm).

    The connections in the Tickable Queue are arranged in the least
    recently ticked order.  This lets connections that have been quiet
    longer to get their packets scheduled first.

    This change means that the library no longer needs to be ticked
    periodically.  The user code can query the library when is the
    next tick event and schedule it exactly.  When connections are
    processed, only the tickable connections are processed, not *all*
    the connections.  When there are no tick events, it means that no
    timer event is necessary -- only the file descriptor READ event
    is active.

    The following are improvements and simplifications that have
    been triggered:

        - Queue of connections with incoming packets is gone.
        - "Pending Read/Write Events" Queue is gone (along with its
          history and progress checks).  This queue has become the
          Tickable Queue.
        - The connection hash no longer needs to track the connection
          insertion order.

2018-04-02

    - [FEATURE] Windows support

    - Reduce stack use -- outgoing packet batch is now allocated on the heap.

2018-03-09

    - [OPTIMIZATION] Merge series of ACKs if possible

      Parsed single-range ACK frames (that is the majority of frames) are
      saved in the connection and their processing is deferred until the
      connection is ticked.  If several ACKs come in a series between
      adjacent ticks, we check whether the latest ACK is a strict superset
      of the saved ACK.  If it is, the older ACK is not processed.

      If ACK frames can be merged, they are merged and only one of them is
      either processed or saved.

    - [OPTIMIZATION] Speed up ACK verification by simplifying send history.

      Never generate a gap in the sent packet number sequence.  This reduces
      the send history to a single number instead of potentially a series of
      packet ranges and thereby speeds up ACK verification.

      By default, detecting a gap in the send history is not fatal: only a
      single warning is generated per connection.  The connection can continue
      to operate even if the ACK verification code is not able to detect some
      inconsistencies.

    - [OPTIMIZATION] Rearrange the lsquic_send_ctl struct

      The first part of struct lsquic_send_ctl now consists of members that
      are used in lsquic_send_ctl_got_ack() (in the absense of packet loss,
      which is the normal case).  To speed up reads and writes, we no longer
      try to save space by using 8- and 16-bit integers.  Use regular integer
      width for everything.

    - [OPTIMIZATION] Cache size of sent packet.

    - [OPTIMIZATION] Keep track of the largest ACKed in packet_out

      Instead of parsing our own ACK frames when packet has been acked,
      use the value saved in the packet_out structure when the ACK frame
      was generated.

    - [OPTIMIZATION] Take RTT sampling conditional out of ACK loop

    - [OPTIMIZATION] ACK processing: only call clock_gettime() if needed

    - [OPTIMIZATION] Several code-level optimizations to ACK processing.

    - Fix: http_client: fix -I flag; switch assert() to abort()

2018-02-26
    - [API Change] lsquic_engine_connect() returns pointer to the connection
      object.
    - [API Change] Add lsquic_conn_get_engine() to get engine object from
      connection object.
    - [API Change] Add lsquic_conn_status() to query connection status.
    - [API Change] Add add lsquic_conn_set_ctx().
    - [API Change] Add new timestamp format, e.g. 2017-03-21 13:43:46.671345
    - [OPTIMIZATION] Process handshake STREAM frames as soon as packet
      arrives.
    - [OPTIMIZATION] Do not compile expensive send controller sanity check
      by default.
    - [OPTIMIZATION] Add fast path to gquic_be_gen_reg_pkt_header.
    - [OPTIMIZATION] Only make squeeze function call if necessary.
    - [OPTIMIZATION] Speed up Q039 ACK frame parsing.
    - [OPTIMIZATION] Fit most used elements of packet_out into first 64 bytes.
    - [OPTIMIZATION] Keep track of scheduled bytes instead of calculating.
    - [OPTIMIZATION] Prefetch next unacked packet when processing ACK.
    - [OPTIMIZATION] Leverage fact that ACK ranges and unacked list are.
      ordered.
    - [OPTIMIZATION] Reduce function pointer use for STREAM frame generation
    - Fix: reset incoming streams that arrive after we send GOAWAY.
    - Fix: delay client on_new_conn() call until connection is fully set up.
    - Fixes to buffered packets logic: splitting, STREAM frame elision.
    - Fix: do not dispatch on_write callback if no packets are available.
    - Fix WINDOW_UPDATE send and resend logic.
    - Fix STREAM frame extension code.
    - Fix: Drop unflushed data when stream is reset.
    - Switch to tracking CWND using bytes rather than packets.
    - Fix TCP friendly adjustment in cubic.
    - Fix: do not generate invalid STOP_WAITING frames during high packet
      loss.
    - Pacer fixes.

2017-12-18

    - Fix: better follow cubic curve after idle period
    - Fix: add missing parts to outgoing packet splitting code
    - Fix: compilation using gcc 4.8.4

2017-10-31

    - Add APIs.txt -- describes LSQUIC APIs

2017-10-31

    - [API Change] Sendfile-like functionality is gone.  The stream no
      longer opens files and deals with file descriptors.  (Among other
      things, this makes the code more portable.)  Three writing functions
      are provided:

        lsquic_stream_write
        lsquic_stream_writev
        lsquic_stream_writef    (NEW)

      lsquic_stream_writef() is given an abstract reader that has function
      pointers for size() and read() functions which the user can implement.
      This is the most flexible way.  lsquic_stream_write() and
      lsquic_stream_writev() are now both implemented as wrappers around
      lsquic_stream_writef().

    - [OPTIMIZATION] When writing to stream, be it within or without the
      on_write() callback, place data directly into packet buffer,
      bypassing auxiliary data structures.  This reduces amount of memory
      required, for the amount of data that can be written is limited
      by the congestion window.

      To support writes outside the on_write() callback, we keep N
      outgoing packet buffers per connection which can be written to
      by any stream.  One half of these are reserved for the highest
      priority stream(s), the other half for all other streams.  This way,
      low-priority streams cannot write instead of high-priority streams
      and, on the other hand, low-priority streams get a chance to send
      their packets out.

      The algorithm is as follows:

      - When user writes to stream outside of the callback:
        - If this is the highest priority stream, place it onto the
          reserved N/2 queue or fail.
            (The actual size of this queue is dynamic -- MAX(N/2, CWND) --
             rather than N/2, allowing high-priority streams to write as
             much as can be sent.)
        - If the stream is not the highest priority, try to place the
          data onto the reserved N/2 queue or fail.
      - When tick occurs *and* more packets can be scheduled:
        - Transfer packets from the high N/2 queue to the scheduled
          queue.
        - If more scheduling is allowed:
          - Call on_write callbacks for highest-priority streams,
            placing resulting packets directly onto the scheduled queue.
        - If more scheduling is allowed:
          - Transfer packets from the low N/2 queue to the scheduled
            queue.
        - If more scheduling is allowed:
          - Call on_write callbacks for non-highest-priority streams,
            placing resulting packets directly onto the scheduled queue

      The number N is currently 20, but it could be varied based on
      resource usage.

    - If stream is created due to incoming headers, make headers readable
      from on_new.

    - Outgoing packets are no longer marked non-writeable to prevent placing
      more than one STREAM frame from the same stream into a single packet.
      This property is maintained via code flow and an explicit check.
      Packets for stream data are allocated using a special function.

    - STREAM frame elision is cheaper, as we only perform it if a reset
      stream has outgoing packets referencing it.

    - lsquic_packet_out_t is smaller, as stream_rec elements are now
      inside a union.

2017-10-12

    - Do not send RST_STREAM when stream is closed for reading
    - Raise maximum header size from 4K to 64K
    - Check header name and value lengths against maximum imposed by HPACK
    - Fix NULL dereference in stream flow controller

2017-10-09

    - Hide handshake implementation behind a set of function pointers
    - Use monotonically increasing clock
    - Make sure that retx delay is not larger than the max of 60 seconds

2017-09-29

    - A few fixes to code and README

2017-09-28

    - Add support for Q041; drop support for Q040

2017-09-27

    - Fix CMakeLists.txt: BoringSSL include and lib was mixed up

2017-09-26

    - Add support for Mac OS
    - Add support for Raspberry Pi
    - Fix BoringSSL compilation: include <openssl/hmac.h> explicitly

2017-09-22

    - Initial release