package mlt_parser

  1. Overview
  2. Docs
package mlt_parser
Legend:
Page
Library
Module
Module type
Parameter
Class
Class type
Source

Module Mlt_parserSource

Code for parsing toplevel expect test files

Sourcetype chunk = {
  1. part : string option;
    (*

    The part the chunk is in, None if it's not in any part.

    *)
  2. phrases : Ppxlib.toplevel_phrase list;
  3. expectation : Expect_test_matcher.Fmt.t Expect_test_matcher.Cst.t Expect_test_common.Expectation.t;
  4. phrases_loc : Ppxlib.Location.t;
}
Sourceval split_chunks : fname:string -> allow_output_patterns:bool -> Ppxlib.toplevel_phrase list -> chunk list * (Ppxlib.toplevel_phrase list * Ppxlib.position * string option) option

Recursively parses toplevel phrases (i.e., contiguous units of code separated by ;;) into "chunks", one chunk per %%expect statement.

For example if the mlt contents are:

  let x = 1 + 1;;

  printf "%d" x + 2;;

  [%%expect {|
 - : int: 4
 |}];;

  print_string "f" ^ "o" ^ "o";;

  [%%expect {|
 - : string: "foo"
 |}];;

  print_string 3 + 3 + 3;;

then you'd have two chunks, where the first has two phrases ("x = 1 + 1" and "printf "%d" x + 2") and an expectation.body of ": int 4". The second chunk would have just the one phrase.

"print_line 3 + 3 + 3" is not part of a chunk because there is no expectation following it, so instead it is returned as trailing_code, which is just a list of toplevel phrases with some position metadata.

"part" refers to @@@part "foo" statements, which are arbitrary section breaks. Each chunk, and the trailing code, belongs to a part (which is just the empty string "" if none has been specified).

Sourcetype mlt_block =
  1. | Org of string
  2. | Expect of string
  3. | Code of string
Sourceval sexp_of_mlt_block : mlt_block -> Sexplib0.Sexp.t
Sourceval mlt_block_of_sexp : Sexplib0.Sexp.t -> mlt_block
Sourceval parse : Ppxlib.toplevel_phrase list -> contents:string -> mlt_block list

Takes a list of toplevel phrases and the raw string they're embedded in and returns a list of labeled blocks, so that for instance the following raw toplevel code:

  [%%org {|
    Here comes a very /simple/ example.
  |}];;

  1 + 1;;
  [%%expect {|
  - : int: 2
  |}];;

is parsed into its constituent parts:

  [
    (Org "Here comes a very /simple/ example.");
    (Code "1 + 1");
    (Expect "- : int: 2")
  ]

Note that we only care about these three kinds of element (org blocks, expect blocks, and regular OCaml code blocks); everything else -- including toplevel comments -- is silently discarded.