Janestreet

[maurer]

Motivation for this change

OCaml has switched its primary macro system in OCaml 4.02 to PPX (from camlp4).
Jane Street, who produces one of the OCaml basis libraries, took this opportunity to change their build system, rip out all of their camlp4 stuff, and replace it with PPX.
New libraries and programs which depend on the new PPX style macros are being released, so we need a way to support them.

Things done

• Tested using sandboxing
  (nix.useSandbox on NixOS,
  or option build-use-sandbox in nix.conf
  on non-NixOS)
• Built on platform(s)
  • NixOS
  • OS X
  • Linux
• Tested compilation of all pkgs that depend on this change using nix-shell -p nox --run "nox-review wip"
• Tested execution of all binary files (usually in ./result/bin/)
• Fits CONTRIBUTING.md.

---

To deal with older compilers and packages, I did a few things in addition to packaging the new code:

• Each package which has transitioned to PPX has been renamed to packagename_p4, allowing its last camlp4 release to be referenced explicitly
• Each package which transitioned to PPX had an if statement added which will diversion to packagename_p4 if the compiler is too old, so OCaml 4.0.0, 3.1, etc will still be able to access e.g. "sexplib" rather than "sexplib_p4"
• Existing packages which referenced a transitioning package were pointed at the _p4 version.

[mention-bot]

@maurer, thanks for your PR! By analyzing the annotation information on this pull request, we identified @ts468 and @ericbmerritt to be potential reviewers

[maurer]

@vbgl Tagging you as another potential reviewer, feel free to ignore if you're not interested.

[thufschmitt]

This description doesn't seem to correspond to the ppx

[thufschmitt]

Wow, that's a huge work !

I added a few minor annotations, otherwise it's a big 👍

[aske]

Right now I'm working on moving ocaml packages from all-packages.nix to make it look more like current haskell system; would be nice to see this PR merged first so I can deal with merge conflicts sooner ( the opposite way is fine too :) ).

[joachifm]

Is the darwin failure reported by travis legit and if so is it a regression?

[vbgl]

I can reproduce the error on darwin: ppx_core won’t build with the following message.

+ xargs stat -f '%d:%i:%m' < .compiler-libs.common.files > .compiler-libs.common.stats
stat: cannot read file system information for '%d:%i:%m': No such file or directory

I have no idea where that comes from.

[maurer]

I don't have easy access to a darwin machine, but ppx_core as a piece of software does work on OSX outside of this packagaing.

@joachifm It's not a regression since ppx_core is not in nixpkgs yet, but should probably be fixed before merging since I know the packages I'm adding/updating work on OSX when installed via OPAM.

[maurer]

https://github.com/janestreet/js-build-tools/blob/7ff35bb6451f241bb1a643f73f73d9d55fd36d94/ocamlbuild_goodies/jane_street_ocamlbuild_goodies.ml#L66-L74

It looks like they are expecting some BSD-like version of stat for which -f means something else, that is normally present on OSX. With Darwin + nix, it may be getting a more Linuxish stat.

I can't test on Darwin easily, but does someone who can want to test removing | "Darwin" from that top match and see if it fixes things? @vbgl

[vbgl]

Good catch. The following patch fixes the build.

--- a/ocamlbuild_goodies/jane_street_ocamlbuild_goodies.ml
+++ b/ocamlbuild_goodies/jane_street_ocamlbuild_goodies.ml
@@ -65,7 +65,7 @@ let track_external_deps = function

     let stat, md5sum =
       match run_and_read "uname" |> String.trim with
-      | "Darwin" ->
+      | "FreeBSD" | "NetBSD" ->
         (S [A "stat"; A "-f"; A "%d:%i:%m"],
          A "md5")
       | _ ->

[maurer]

I've applied the patch to fix the darwin build. Anything else to be done?