[new release] eio_main, eio_luv, eio_linux and eio (0.8)

[talex5]

Effect-based direct-style IO mainloop for OCaml

• Project page: https://github.com/ocaml-multicore/eio
• Documentation: https://ocaml-multicore.github.io/eio/

CHANGES:

New features:

• Add Eio.Net.run_server (@bikallem @talex5 net: add run_server to run eio servers ocaml-multicore/eio#408).
  Runs an accept loop in one or more domains, with cancellation and graceful shutdown,
  and an optional maximum number of concurrent connections.

• Add Buf_read.BE and LE parsers (@Cjen1 Add Buf_read.BE/LE parsers ocaml-multicore/eio#399).
  Parse numbers in various binary formats.

• Add Eio.Buf_read.uint8 (@talex5 Buf_write: Fix BE.uint48 and LE.uint48. ocaml-multicore/eio#418).

Performance:

• Make Eio.Condition lock-free (@talex5 Make Eio.Condition lock-free ocaml-multicore/eio#397 Make Eio_linux.wakeup signal-safe ocaml-multicore/eio#381).
  In addition to being faster, this allows using conditions in signal handlers.

• Make Eio.Semaphore lock-free (@talex5 @polytypic Make Eio.Semaphore lock-free ocaml-multicore/eio#398).

• Make Eio.Stream lock-free when the capacity is zero (@talex5 Add cancellable lock-free synchronous channel ocaml-multicore/eio#413 Improve stream benchmark ocaml-multicore/eio#411).

• Make Eio.Promise lock-free (@talex5 Make Eio.Promise lock-free ocaml-multicore/eio#401).

Bug fixes:

• eio_linux: call Uring.submit as needed (@talex5 @bikallem eio_linux: call submit as needed ocaml-multicore/eio#428).
  Previously, we could fail to submit a job promptly because the SQE queue was full.

• Fix luv signals (@haesbaert Fix luv signals (issue #400) ocaml-multicore/eio#412).
  libuv automatically retries polling if it gets EINTR, without giving OCaml signal handlers a chance to run.

• eio_luv: fix some resource leaks (@talex5 @patricoferris eio_luv: fix some resource leaks ocaml-multicore/eio#421).

• eio_luv: fix "unavailable signal" error on Windows (@talex5 Windows: fix "unavailable signal" error ocaml-multicore/eio#420, reported by @nojb).

• Fix Buf_write.BE.uint48 and LE.uint48 (@adatario Buf_write: Fix BE.uint48 and LE.uint48. ocaml-multicore/eio#418).

Documentation:

• Add example programs (@talex5 Documentation improvements ocaml-multicore/eio#389).

• Update network examples to use run_server (@talex5 Update network examples to use run_server ocaml-multicore/eio#417).

• Add a warning to the tutorial about Fiber.first (@talex5 Add a warning to the tutorial about Fiber.first ocaml-multicore/eio#394).

• Clarify the epoch used for Eio.Time.now (@bikallem time: clarify Eio.Time.now is UTC ocaml-multicore/eio#395).

• Describe secure_random as an infinite source (@patricoferris Describe secure_random as an infinite source ocaml-multicore/eio#426).

• Update README for OCaml 5 release (@talex5 Update README for OCaml 5.0.0~rc1 ocaml-multicore/eio#384 Update README for OCaml 5 release ocaml-multicore/eio#391 Simply Dockerfile now OCaml 5 is out ocaml-multicore/eio#393).

Other changes:

• Delay setting SIGPIPE handler until the run function is called (@talex5 Windows: fix "unavailable signal" error ocaml-multicore/eio#420).

• Remove debug-level logging (@talex5 Remove debug-level logging ocaml-multicore/eio#403).

• eio-luv: improve process.md test (@smondet eio-luv: improve process.md test ocaml-multicore/eio#414).

• Update to Dune 3 (@talex5 Update to Dune 3 ocaml-multicore/eio#410).

• Remove test dependency on Astring (@talex5 Remove test dependency on Astring ocaml-multicore/eio#402 Fix off-by-one in fuzz test ocaml-multicore/eio#404).

• Simplify cancellation logic (@talex5 Simplify cancellation logic ocaml-multicore/eio#396).

• time: Mtime.Spand.to_s has been deprecated in mtime 2.0.0 (@bikallem time: Mtime.Spand.to_s has been deprecated in mtime 2.0.0 ocaml-multicore/eio#385).

[talex5]

• The s390x failure is due to a bug in luv (Write fails on s390x (size_t vs uint) aantron/luv#144).
• The macOS tests are are expected to fail, due to the sandbox blocking the loopback interface.

[raphael-proust]

If the macos failures are expected, we might want to add x-ci-accept-failures : [ … ].

Alternatively, we could condition the @runtest with "@runtest" {with-test & …}. This would allow dependencies to be tested in the CI I think.

[talex5]

OK, I added x-ci-accept-failures: ["macos-homebrew"]. I think changing the project itself would be wrong, since the tests do work on macos in general, just not in this CI.

I also investigated why the Lwt_eio revdep tests failed. I believe this is due to ocsigen/lwt#981, and is not a bug in Eio.

[talex5]

I'm going to make another release with work-arounds for some of the build problems.