How to generate and use the generated file - ocaml

Question

Is it possible, with ocp-build, to do the following actions:
Compile a generator.
Call the generator to generate a file.
Compile the project with the generated file.
So far, I tried this:
(generator.ocp)
begin library "error_gen"
sort = true
files = [ "error_gen.ml" ]
requires = [str]
end
(generated.ocp)
begin library "error_code"
sort = true
files = [
"error_code.ml" (
pp = [ "./_obuild/error_gen/error_gen.byte" ]
pp_requires = [ "error_gen:byte" ]
)
]
requires = ["error_gen"]
end
(and the main.ocp)
begin program "main"
sort = true
files = []
requires = ["error_code" "parser"]
end
It complains with this message:
Error: in project "error_code", the source filename
"src/generated/error_code.ml" does not exist
I saw that some supports exist for version file generation, such as in the project ocp-indent
line 46.
"indentVersion.ml" (ocp2ml) (* auto-generated by ocp-build *)
Any helps greatly appreciated, thanks.

Answer 1

In branch "next" of github.com/OCamlPro/ocp-build, you will find a version of ocp-build that might solve your issue:
begin library "error_code"
sort = true
files = [ "error_code.ml" ]
build_rules = [
"error_code.ml" (
(* the files that are needed to call the command *)
sources = [ "%{error_gen_FULL_DST_DIR}%/error_gen.byte" ]
(* the commands to be executed, in the sub-directory of the library
each command has the syntax: { "COMMAND" "ARG1" "ARG2" ... }
*)
commands = [
{ "%{error_gen_FULL_DST_DIR}%/error_gen.byte" }
]
)
]
requires = ["error_gen"]
end
This is, for example, used in wxOCaml:
https://github.com/OCamlPro/ocplib-wxOCaml/blob/next-ocpbuild/configure.ocp
Commands can be post-fixed with options:
{ "configure" } (chdir = "subdirectory") (* to execute in a sub-directory *)
{ "cat" "toto" } (stdout = "new_toto") (* to copy the stdout in "new_toto" *)