Effects: introduce primitive caml_cps_trampoline

[vouillon]

This is a better name for the primitive that sets up a trampoline to switch to CPS.

[hhugo]

I think this make the CI fail because wasmoo will "resolve" caml_cps_trampoline to caml_callback

[hhugo]

The following diff is missing from the PR. It would be nice if we could check that import/export are coherent instead of relying on bad error reporting such as

File "compiler/tests-wasm_of_ocaml/dune", line 2, characters 8-12:
2 |  (names gh38 gh46 gh107 gh112 gh1904)
            ^^^^
(cd _build/default/compiler/tests-wasm_of_ocaml && ../../.bin/node gh38.bc.wasm.js)
node:internal/process/promises:394
    triggerUncaughtException(err, true /* fromPromise */);
    ^

[TypeError: WebAssembly.instantiate(): Import #110 "effect": module is not an object or function]

 (func $caml_is_continuation (param (ref eq)) (result i32)))
(@if (= effects "cps")
(@then
-   (import "effect" "caml_trampoline"
-      (func $caml_trampoline (param (ref eq) (ref eq)) (result (ref eq))))
+   (import "effect" "caml_cps_trampoline"
+      (func $caml_cps_trampoline (param (ref eq) (ref eq)) (result (ref eq))))
))


@@ -460,14 +460,14 @@
(@then
   (func $caml_callback_1 (export "caml_callback_1")
      (param $f (ref eq)) (param $x (ref eq)) (result (ref eq))
-      (return_call $caml_trampoline
+      (return_call $caml_cps_trampoline
         (local.get $f)
         (array.new_fixed $block 2 (ref.i31 (i32.const 0)) (local.get $x))))

   (func (export "caml_callback_2")
      (param $f (ref eq)) (param $x (ref eq)) (param $y (ref eq))
      (result (ref eq))
-      (return_call $caml_trampoline
+      (return_call $caml_cps_trampoline
         (local.get $f)
         (array.new_fixed $block 3 (ref.i31 (i32.const 0))
           (local.get $x) (local.get $y))))

[vouillon]

The following diff is missing from the PR.

That's because this code was added in master since the PR was made.

It would be nice if we could check that import/export are coherent instead of relying on bad error reporting such as
[...]

See #1931

[hhugo]

included in #1915