[new release] ppx_deriving_yaml (0.2.2)

[patricoferris]

Yaml PPX Deriver

• Project page: https://github.com/patricoferris/ppx_deriving_yaml

CHANGES:

• Embed errors in the AST (Better error handling patricoferris/ppx_deriving_yaml#51, @patricoferris and special thanks to @panglesd
  for the detailed issue in Embed errors in the AST instead of raising patricoferris/ppx_deriving_yaml#48)

[mseri]

Are the test failures on bytecode-only compilers expected?
https://opam.ci.ocaml.org/github/ocaml/opam-repository/commit/39d4a56c83e6d31ded687f77734bf5ef184271d3

[patricoferris]

Hmm thanks for the info @mseri -- I didn't realise that's what distinguished those compilers. I tried locally with something like:

opam switch create 5.1.0-bytes --packages=ocaml-variants.5.1.0+options,ocaml-option-bytecode-only

And then trying to install and run the tests, but can't make cmdliner happy for some reason. Just posting here in case this has come up before maybe ?

➜  ppx_deriving_yaml git:(main) opam install alcotest
Constructing initial basis...
Number of 0-1 knapsack inequalities = 649
Constructing conflict graph...
Conflict graph has 255 + 103 = 358 vertices
The following actions will be performed:
  ∗ install   cmdliner 1.2.0 [required by alcotest]
  ∗ install   uutf     1.0.3 [required by alcotest]
  ↻ recompile fmt      0.9.0 [uses cmdliner]
  ∗ install   alcotest 1.7.0
  ↻ recompile logs     0.7.0 [uses cmdliner]
  ↻ recompile bos      0.2.1 [uses fmt]
  ↻ recompile yaml     3.2.0 [uses bos]
===== ∗ 3   ↻ 4 =====
Do you want to continue? [Y/n] Y

<><> Processing actions <><><><><><><><><><><><><><><><><><><><><><><><><><>  🐫 
⬇ retrieved bos.0.2.1  (cached)
⬇ retrieved cmdliner.1.2.0  (cached)
⬇ retrieved logs.0.7.0  (cached)
⬇ retrieved fmt.0.9.0  (cached)
⬇ retrieved uutf.1.0.3  (cached)
⬇ retrieved alcotest.1.7.0  (cached)
⬇ retrieved yaml.3.2.0  (cached)
⊘ removed   yaml.3.2.0
⊘ removed   bos.0.2.1
⊘ removed   logs.0.7.0
⊘ removed   fmt.0.9.0
∗ installed cmdliner.1.2.0
[ERROR] The compilation of uutf.1.0.3 failed at "ocaml pkg/pkg.ml build --dev-pkg false --with-cmdliner true".
[ERROR] The compilation of fmt.0.9.0 failed at "ocaml pkg/pkg.ml build --dev-pkg false --with-base-unix true --with-cmdliner true".

#=== ERROR while compiling uutf.1.0.3 =========================================#
# context     2.1.3 | macos/arm64 | ocaml-option-bytecode-only.1 ocaml-variants.5.1.0+options | https://opam.ocaml.org#e0d5a913
# path        ~/.opam/5.1.0-bytes/.opam-switch/build/uutf.1.0.3
# command     ~/.opam/opam-init/hooks/sandbox.sh build ocaml pkg/pkg.ml build --dev-pkg false --with-cmdliner true
# exit-code   1
# env-file    ~/.opam/log/uutf-70702-d41ee3.env
# output-file ~/.opam/log/uutf-70702-d41ee3.out
### output ###
# [...]
# ocamlfind ocamldep -package unix -package cmdliner -modules test/utftrip.ml > test/utftrip.ml.depends
# ocamlfind ocamlc -c -g -bin-annot -safe-string -package unix -package cmdliner -I test -I src -o test/utftrip.cmo test/utftrip.ml
# + ocamlfind ocamlc -c -g -bin-annot -safe-string -package unix -package cmdliner -I test -I src -o test/utftrip.cmo test/utftrip.ml
# File "test/utftrip.ml", line 1:
# Error: /Users/patrickferris/.opam/5.1.0-bytes/lib/cmdliner/cmdliner.cmi
#        is not a compiled interface for this version of OCaml.
# It seems to be for an older version of OCaml.
# Command exited with code 2.
# pkg.ml: [ERROR] cmd ['ocamlbuild' '-use-ocamlfind' '-classic-display' '-j' '4' '-tag' 'debug'
#      '-build-dir' '_build' 'opam' 'pkg/META' 'CHANGES.md' 'LICENSE.md'
#      'README.md' 'src/uutf.cma' 'src/uutf.cmi' 'src/uutf.mli'
#      'test/utftrip.byte']: exited with 10


#=== ERROR while compiling fmt.0.9.0 ==========================================#
# context     2.1.3 | macos/arm64 | ocaml-option-bytecode-only.1 ocaml-variants.5.1.0+options | https://opam.ocaml.org#e0d5a913
# path        ~/.opam/5.1.0-bytes/.opam-switch/build/fmt.0.9.0
# command     ~/.opam/opam-init/hooks/sandbox.sh build ocaml pkg/pkg.ml build --dev-pkg false --with-base-unix true --with-cmdliner true
# exit-code   1
# env-file    ~/.opam/log/fmt-70702-a9dd12.env
# output-file ~/.opam/log/fmt-70702-a9dd12.out
### output ###
# [...]
# ocamlfind ocamlc -c -g -bin-annot -safe-string -package cmdliner -I src -I test -o src/fmt_cli.cmi src/fmt_cli.mli
# + ocamlfind ocamlc -c -g -bin-annot -safe-string -package cmdliner -I src -I test -o src/fmt_cli.cmi src/fmt_cli.mli
# File "src/fmt_cli.mli", line 1:
# Error: /Users/patrickferris/.opam/5.1.0-bytes/lib/cmdliner/cmdliner.cmi
#        is not a compiled interface for this version of OCaml.
# It seems to be for an older version of OCaml.
# Command exited with code 2.
# pkg.ml: [ERROR] cmd ['ocamlbuild' '-use-ocamlfind' '-classic-display' '-j' '4' '-tag' 'debug'
#      '-build-dir' '_build' 'opam' 'pkg/META' 'CHANGES.md' 'LICENSE.md'
#      'README.md' 'src/fmt.cma' 'src/fmt.cmi' 'src/fmt.mli' 'src/fmt_tty.cma'
#      'src/fmt_tty.cmi' 'src/fmt_tty.mli' 'src/fmt_cli.cma' 'src/fmt_cli.cmi'
#      'src/fmt_cli.mli' 'src/fmt_top.cma' 'src/fmt_tty_top_init.ml']: exited with 10



<><> Error report <><><><><><><><><><><><><><><><><><><><><><><><><><><><><>  🐫 
┌─ The following actions failed
│ λ build fmt  0.9.0
│ λ build uutf 1.0.3
└─ 
┌─ The following changes have been performed (the rest was aborted)
│ ⊘ remove  bos      0.2.1
│ ⊘ remove  fmt      0.9.0
│ ⊘ remove  logs     0.7.0
│ ⊘ remove  yaml     3.2.0
│ ∗ install cmdliner 1.2.0
└─ 

[patricoferris]

Blergh! Nevermind, a rogue ~/.opam/4.14.0/bin was in my path and that was probably providing an ocamlopt and then later something was using ocamlc from ~/.opam/5.1.0-bytes/bin or something along those lines. Sorry for the noise.

[patricoferris]

Hmm yeh this reproduces the problem -- something up with bytecode only toplevels getting ppxes. You see it just using ocaml and #require "ppx_deriving_{yaml,yojson, etc.}" and then trying to use it. Same in utop as well.

[mseri]

Suggested change

	]
	]
	conflicts: [
	"ocaml-option-bytecode-only"
	]

For the moment we could solve it this way and release. What do you think?

[patricoferris]

Even if it is only the MDX based tests that are failing, I wonder if I can do some dune magic to disable them if it detects byte code. Thanks for the help so far, I'll get back to you shortly :))

[patricoferris]

Thanks for the patience @mseri -- I've re-released 0.2.2 over at #25072 where hopefully I have some dune checks to disable byte code MDX tests. Thanks for the help in getting a good release into opam, let's hope the tests are all green !