chore: release

[github-actions]

🤖 New release

• fluke: 0.1.1 -> 0.1.2
• fluke-buffet: 0.2.0 -> 0.3.0
• fluke-io-uring-async: 0.1.0 -> 0.1.1
• fluke-h2-parse: 0.1.1 -> 0.2.0
• fluke-hpack: 0.3.1 -> 0.3.2
• httpwg: 0.1.1 -> 0.2.0
• httpwg-macros: 0.1.0 -> 0.1.1
• fluke-httpwg-server: 0.1.0
• httpwg-cli: 0.1.0

Changelog

fluke

0.1.2 - 2024-08-14

Added

• Error out if we write too few / too many response bytes
• Move BodyWriteMode into H1Encoder

Other

• 16K buffers are enough methinks
• add lut for fun
• Do a meaningful order since they get sorted by criterion anyway
• Try an itoa/pool strategy
• add std_fmt/buffet variant
• Make benchmark more realistic: we need it as a to issue the I/O call
• Fix benchmark
• Try perf?
• rename benches
• Well, benchmarks are sometimes disappointing but they tell the truth
• naming
• Hopefully that groups them on codspeed?
• Run on ubuntu-latest, profile format_content_length with itoa with buffet and alloc
• Add content_length formatting benches
• Add format_status_code benchmarks
• Try with both HTTP 1.1 and HTTP/2
• Upgrade dependencies
• Okay the takeaway is that tasks spawned when a runtime is shutting down immediately get dropped>
• Well we pass h2spec strict!
• well, h2spec and httpwg-cli agree on almost everything
• Introduce fluke-httpwg-server so we can run httpwg over TCP
• Start using httpwg-cli against real servers
• All tests pass!
• Change Headers to a VecDeque<(Piece, Piece)>, closes #193
• okay, a few more adjustments
• Pass more tests
• welp, need more validation
• Yay we're properly returning 400x
• Ready to send back 400s
• Start working on returning 400 from fluke
• Make 'static part of IntoHalves
• Fully read request body before sending a response
• race condition in sends_multiple_window_update_frames_increasing_flow_control_window_above_max_on_stream, we can fix that
• tests build again
• refactor settings sending, add more helpers, etc.
• Add 6.5.3 specs
• Introduce SettingsError, fix error code when receiving invalid initial window size
• WIP settings parsing
• 6.5 tests
• Make 6.1 tests pass
• 6.1 tests
• 5.1.2 tests
• verify_stream_close might be wrong
• Add rfc 9113 section 5.1 specs

fluke-buffet

0.3.0 - 2024-08-14

Added

• Try going back to stable with LocalKey and no nightly features

Fixed

• fix tests

Other

• Try an itoa/pool strategy
• add std_fmt/buffet variant
• Fix benchmark
• Try perf?
• Well, benchmarks are sometimes disappointing but they tell the truth
• mhh
• Run on ubuntu-latest, profile format_content_length with itoa with buffet and alloc
• More async cancellation cleanups
• Upgrade hyper + fix cancellation in test rig
• Upgrade nix
• Upgrade dependencies
• Okay the takeaway is that tasks spawned when a runtime is shutting down immediately get dropped>
• shutting down is hard
• Debug tests
• grahhhh why would you answer in multiple headers grr
• Start using httpwg-cli against real servers
• okay, a few more adjustments
• Make 'static part of IntoHalves
• Some section 8 tests (acceptable headers)
• More 4.2 tests
• Clean up 4_1 tests

fluke-io-uring-async

0.1.1 - 2024-08-14

Other

• Upgrade hyper + fix cancellation in test rig
• Upgrade dependencies
• yep that's what's happening
• Okay the takeaway is that tasks spawned when a runtime is shutting down immediately get dropped>

fluke-h2-parse

0.2.0 - 2024-08-14

Other

• Try perf?
• Upgrade dependencies
• Address Rust 1.80.1 warnings
• grahhhh why would you answer in multiple headers grr
• Start using httpwg-cli against real servers
• More Section 8 tests
• Make test invalid_header_block_fragment more readable
• race condition in sends_multiple_window_update_frames_increasing_flow_control_window_above_max_on_stream, we can fix that
• refactor settings sending, add more helpers, etc.
• Introduce SettingsError, fix error code when receiving invalid initial window size
• WIP settings parsing
• 6.5 tests
• 5.1.2 tests
• 5.1 specs compile
• Add rfc 9113 section 5.1 specs
• 4.3 tests
• Clean up 4_1 tests

fluke-hpack

0.3.2 - 2024-08-14

Other

• Upgrade dependencies
• Yay we're properly returning 400x
• Ready to send back 400s
• Some section 8 tests (acceptable headers)

httpwg

0.2.0 - 2024-08-14

Other

• Upgrade hyper + fix cancellation in test rig
• Upgrade dependencies
• debug shutdowns
• shutting down is hard
• Debug tests
• Well we pass h2spec strict!
• well, h2spec and httpwg-cli agree on almost everything
• Introduce fluke-httpwg-server so we can run httpwg over TCP
• Yey it runs
• grahhhh why would you answer in multiple headers grr
• Start using httpwg-cli against real servers
• All tests pass!
• Change Headers to a VecDeque<(Piece, Piece)>, closes #193
• Ey test code was wrong
• Start working on returning 400 from fluke
• wip goaway tests
• Connect tests
• More section 8 tests
• More Section 8 tests
• Make 'static part of IntoHalves
• A bunch of failing section 8 tests
• Some section 8 tests (acceptable headers)
• RFC 9113 Section 8
• Add section 7 tests
• Move 3.4 tests to _3 file
• Make test invalid_header_block_fragment more readable
• Nove section 5 tests all to the same file
• race condition in sends_multiple_window_update_frames_increasing_flow_control_window_above_max_on_stream, we can fix that
• tests build again
• Fix settings (partly)
• section 6 tests compile
• refactor settings sending, add more helpers, etc.
• Add 6.8
• ping specs
• Add 6.5.3 specs
• Introduce SettingsError, fix error code when receiving invalid initial window size
• WIP settings parsing
• 6.5 tests
• 6.4 tests
• more 6.x tests
• add 6.2 cases (in 6.1 file for now)
• 6.1 tests
• Send unknown frame for right stream_id
• Section 5.5
• 5.4.1 test cases
• Section 5.3
• 5.1.2 tests
• 5.1.1 tests
• Take care of clippy warnings
• Remove some warnings
• All 5.1 tests pass
• verify_stream_close might be wrong
• Fix some tests
• 5.1 specs compile
• Add rfc 9113 section 5.1 specs
• rename 3.4 tests to proper section
• 4.3 tests
• More 4.2 tests
• More 4.1 tests
• clean up tests some more
• Clean up 4_1 tests
• Add a spec to 4_1_frame_format, cf. #175

httpwg-macros

0.1.1 - 2024-08-14

Other

• Well we pass h2spec strict!
• Change Headers to a VecDeque<(Piece, Piece)>, closes #193
• Ey test code was wrong
• Start working on returning 400 from fluke
• wip goaway tests
• Connect tests
• More section 8 tests
• More Section 8 tests
• Make 'static part of IntoHalves
• A bunch of failing section 8 tests
• Some section 8 tests (acceptable headers)
• RFC 9113 Section 8
• Add section 7 tests
• Move 3.4 tests to _3 file
• Make test invalid_header_block_fragment more readable
• Nove section 5 tests all to the same file
• refactor settings sending, add more helpers, etc.
• Add 6.8
• ping specs
• Add 6.5.3 specs
• WIP settings parsing
• 6.5 tests
• 6.4 tests
• more 6.x tests
• add 6.2 cases (in 6.1 file for now)
• 6.1 tests
• Section 5.5
• 5.4.1 test cases
• Section 5.3
• 5.1.2 tests
• 5.1.1 tests
• 5.1 specs compile
• Add rfc 9113 section 5.1 specs
• rename 3.4 tests to proper section
• 4.3 tests
• More 4.2 tests
• More 4.1 tests
• Clean up 4_1 tests
• Add a spec to 4_1_frame_format, cf. #175
• release

fluke-httpwg-server

0.1.0 - 2024-08-14

Other

• debug shutdowns
• well, h2spec and httpwg-cli agree on almost everything
• Introduce fluke-httpwg-server so we can run httpwg over TCP

httpwg-cli

0.1.0 - 2024-08-14

Other

• More async cancellation cleanups
• Make httpwg-cli build on macOS
• Upgrade dependencies
• Okay the takeaway is that tasks spawned when a runtime is shutting down immediately get dropped>
• Run httpwg tests over TCP in CI
• Everything passes
• Run httpwg-cli tests in parallel
• grahhhh why would you answer in multiple headers grr
• Start using httpwg-cli against real servers

---

This PR was generated with release-plz.