[new release] obatcher (1.0)

[koonwen]

A Framework for building Batched Concurrent Data Structures

• Project page: https://github.com/koonwen/obatcher
• Documentation: https://koonwen.github.io/obatcher

CHANGES:

- Initial release.

[shonfeder]

Hello! Thanks for publishing :)

I'm seeing one test failure that looks relevant:

#=== ERROR while compiling obatcher.1.0 =======================================#
# context              2.0.10 | linux/x86_64 | ocaml-base-compiler.5.2.0 | pinned(https://github.com/koonwen/obatcher/releases/download/v1.0/obatcher-1.0.tbz)
# path                 ~/.opam/5.2/.opam-switch/build/obatcher.1.0
# command              ~/.opam/opam-init/hooks/sandbox.sh build dune build -p obatcher -j 39 @install @runtest
# exit-code            1
# env-file             ~/.opam/log/obatcher-43-5f1023.env
# output-file          ~/.opam/log/obatcher-43-5f1023.out
### output ###
# (cd _build/default && /home/opam/.opam/5.2/bin/ocamlopt.opt -w -40 -g -I test/.test_skiplist.eobjs/byte -I test/.test_skiplist.eobjs/native -I /home/opam/.opam/5.2/lib/alcotest -I /home/opam/.opam/5.2/lib/astring -I /home/opam/.opam/5.2/lib/backoff -I /home/opam/.opam/5.2/lib/bytes -I /home/opam/.opam/5.2/lib/cmdliner -I /home/opam/.opam/5.2/lib/containers -I /home/opam/.opam/5.2/lib/containers/monomorphic -I /home/opam/.opam/5.2/lib/either -I /home/opam/.opam/5.2/lib/fmt -I /home/opam/.opam/5.2/lib/logs -I /home/opam/.opam/5.2/lib/multicore-magic -I /home/opam/.opam/5.2/lib/multicore-magic/__private__/multicore_magic_atomic_array_unboxed5 -I /home/opam/.opam/5.2/lib/ocaml/threads -I /home/opam/.opam/5.2/lib/ocaml/unix -I /home/opam/.opam/5.2/lib/ounit2 -I /home/opam/.opam/5.2/lib/ounit2/advanced -I /home/opam/.opam/5.2/lib/picos -I /home/opam/.opam/5.2/lib/picos/__private__/picos_bootstrap -I /home/opam/.opam/5.2/lib/picos/__private__/picos_ocaml5 -I /home/opam/.opam/5.2/lib/picos/thread -I /home/opam/.opam/5.2/lib/picos_aux/htbl -I /home/opam/.opam/5.2/lib/picos_aux/mpscq -I /home/opam/.opam/5.2/lib/picos_mux/fifo -I /home/opam/.opam/5.2/lib/picos_mux/random -I /home/opam/.opam/5.2/lib/picos_mux/thread -I /home/opam/.opam/5.2/lib/picos_std/event -I /home/opam/.opam/5.2/lib/picos_std/structured -I /home/opam/.opam/5.2/lib/qcheck -I /home/opam/.opam/5.2/lib/qcheck-alcotest -I /home/opam/.opam/5.2/lib/qcheck-core -I /home/opam/.opam/5.2/lib/qcheck-core/runner -I /home/opam/.opam/5.2/lib/qcheck-ounit -I /home/opam/.opam/5.2/lib/re -I /home/opam/.opam/5.2/lib/seq -I /home/opam/.opam/5.2/lib/stdlib-shims -I /home/opam/.opam/5.2/lib/thread-local-storage -I /home/opam/.opam/5.2/lib/uuidm -I examples/ds/.ds.objs/byte -I examples/ds/.ds.objs/native -I src/.obatcher.objs/byte -I src/.obatcher.objs/native -intf-suffix .ml -no-alias-deps -open Dune__exe -o test/.test_skiplist.eobjs/native/dune__exe__Test_obatcher.cmx -c -impl test/test_obatcher.ml)
# File "test/test_obatcher.ml", line 189, characters 17-42:
# 189 |   let internal = Mock_Service.get_internal t in
#                        ^^^^^^^^^^^^^^^^^^^^^^^^^
# Alert unsafe: Mock_Service.get_internal
# Should not be used typically by application
# 
# File "test/test_obatcher.ml", line 200, characters 17-42:
# 200 |   let internal = Mock_Service.get_internal t in
#                        ^^^^^^^^^^^^^^^^^^^^^^^^^
# Alert unsafe: Mock_Service.get_internal
# Should not be used typically by application
# 
# File "test/test_obatcher.ml", line 239, characters 19-44:
# 239 |     let internal = Mock_Service.get_internal mock_service in
#                          ^^^^^^^^^^^^^^^^^^^^^^^^^
# Alert unsafe: Mock_Service.get_internal
# Should not be used typically by application
# File "test/dune", line 2, characters 22-35:
# 2 |  (names test_skiplist test_obatcher)
#                           ^^^^^^^^^^^^^
# (cd _build/default/test && ./test_obatcher.exe)
# qcheck random seed: 541542795
# Testing Obatcher.
# This run has ID `735F0845-B729-474D-815A-31186C139C5E`.
# test_obatcher.exe: internal error, uncaught exception:
#                    Unix.Unix_error(Unix.EEXIST, "symlink", "/home/opam/.opam/5.2/.opam-switch/build/obatcher.1.0/_build/default/test/_build/_tests/latest")
#                    
# Fatal error: exception Alcotest__Core.Make(M).Test_error
# (cd _build/default/test && ./test_skiplist.exe)
# qcheck random seed: 408376870
# Testing Batched data structures.
# This run has ID `B4CBD931-84F8-4AF4-A820-EF3DF41D4D08`.
#  ...                skiplist          0   inserts return true on member lookup.
[OK]                skiplist          0   inserts return true on member lookup.
#  ...                skiplist          1   member lookup not vacuously true.
[OK]                skiplist          1   member lookup not vacuously true.
#  ...                skiplist          2   duplicate elements are not inserted.
[OK]                skiplist          2   duplicate elements are not inserted.
# The full test results are available in `/home/opam/.opam/5.2/.opam-switch/build/obatcher.1.0/_build/default/test/_build/_tests/B4CBD931-84F8-4AF4-A820-EF3DF41D4D08`.
# Test Successful in 2.000s. 3 tests run.

The alert warning is coming from https://github.com/koonwen/obatcher/blob/3899dabce1c969b175e85db842309fa428ba14e6/src/obatcher.mli#L51

IIUC, you can disable or demote it's severity as per https://ocaml.org/manual/5.2/alerts.html by configuring the alert flag accordingly in the dune env for your test: https://dune.readthedocs.io/en/stable/faq.html#how-to-turn-specific-errors-into-warnings

The error looks like a (possibly flaky) test failure?

[koonwen]

@shonfeder, thanks for having a look at this. None of my tests use symlinking, Think the error might be coming from the test framework alcotest or qcheck instead. Is this a blocker for merging?

[avsm]

Thanks! You may want to announce this on https://discuss.ocaml.org, where we have a Community category and an announce tag for this purpose.

[shonfeder]

@shonfeder, thanks for having a look at this. None of my tests use symlinking, Think the error might be coming from the test framework alcotest or qcheck instead. Is this a blocker for merging?

Turns out that ci failure was some flaky failure in the testing infrastructure. (We have an issue open to track this kind of thing, but I didn't recognize it at the time.).

Glad to see it cleared up after I reran it last night. Thanks to Anil for merging!