feat(overlay): a flat INI front end, which verifies its own output

A convenience format for calendars that add a few local feasts and drop one
or two universal entries. Section names are slugs, a [overlay] section carries
the id, and status/subject/layer default so the common case -- an ordinary
local saint's feast -- says only what distinguishes it.

It is a FRONT DOOR, not a second data model. It parses to exactly the
Overlay.t the S-expression form parses to, and everything downstream is the
same code on the same values; a test asserts an INI overlay and its
hand-written sexp equivalent produce identical Overlay.t values. It is also
deliberately less expressive -- Add, Suppress and single-field Edit only --
and refuses Replace, multi-field edits and citation edits BY NAME rather than
dropping them silently. Anything it cannot say is a reason to write sexp.

Little of this is new machinery: tools/bootstrap_sanctoral.ml has parsed INI
and mapped it to celebrations since the sanctoral was bootstrapped from
lectio. The dates needed extending, since that mapping handled only MM-DD;
the flat forms are easter+N/easter-N and mon/day/nth, with nth negative to
count from the end.

`colitur convert` is a separate step rather than --overlay sniffing the
extension, so the author can read what their INI became. When a date form was
mistyped, "what did the engine actually get" is the question, and an invisible
transpile cannot answer it.

The conversion verifies its own output: the emitted text is parsed back with
the same function that loads an overlay and must equal what the INI denoted,
or nothing is written. That is the point of the module. A transpiler emitting
valid-but-wrong sexp is the failure a convenience format invites, and `colitur
check` could never catch it -- the output would parse cleanly and mean
something else.

That check was WRONG on the first attempt, in exactly the way it exists to
prevent. It re-serialised the parsed value instead of parsing the text being
returned, so it verified t -> sexp -> t, which is true by construction and
proves nothing. Found by mutation: corrupting the renderer to emit a different
overlay id sailed through and exited 0. It now parses the returned text, the
mutation is caught with exit 2, and two tests fail under it where none did
before.

2 files changed, 363 insertions, 0 deletions

diff --git a/lib/kernel/overlay_ini.ml b/lib/kernel/overlay_ini.ml
new file mode 100644
index 0000000..6b6249f
--- /dev/null
+++ b/lib/kernel/overlay_ini.ml
@@ -0,0 +1,294 @@
+open Sexplib0.Sexp_conv
+
+type section = { name : string; fields : (string * string) list }
+
+let err fmt = Printf.ksprintf (fun s -> Error s) fmt
+
+(* A minimal INI reader: [section] headers, [key = value] lines, ';' and '#'
+   comments, blank lines ignored. Values keep interior spaces and are trimmed
+   at both ends. Deliberately not a general INI implementation -- no
+   continuations, no quoting, no repeated-key semantics -- because every
+   feature here is one a user can trip over, and the format's whole purpose is
+   to be unsurprising. *)
+let parse_sections text =
+  let lines = String.split_on_char '\n' text in
+  let sections = ref [] and cur = ref None and fields = ref [] in
+  let flush () =
+    match !cur with
+    | Some n -> sections := { name = n; fields = List.rev !fields } :: !sections
+    | None -> ()
+  in
+  let rec go n = function
+    | [] ->
+        flush ();
+        Ok (List.rev !sections)
+    | raw :: rest -> (
+        let line = String.trim raw in
+        let n = n + 1 in
+        if line = "" || line.[0] = ';' || line.[0] = '#' then go n rest
+        else if line.[0] = '[' then
+          if line.[String.length line - 1] <> ']' then
+            err "line %d: %S looks like a section header but does not end with ']'" n line
+          else begin
+            flush ();
+            cur := Some (String.sub line 1 (String.length line - 2));
+            fields := [];
+            go n rest
+          end
+        else
+          match String.index_opt line '=' with
+          | None -> err "line %d: %S is neither a [section] nor a key = value line" n line
+          | Some i ->
+              if !cur = None then
+                err "line %d: %S appears before any [section] header" n line
+              else begin
+                let k = String.trim (String.sub line 0 i) in
+                let v = String.trim (String.sub line (i + 1) (String.length line - i - 1)) in
+                fields := (k, v) :: !fields;
+                go n rest
+              end)
+  in
+  go 0 lines
+
+let get sec k = List.assoc_opt k sec.fields
+let is_yes v = List.mem (String.lowercase_ascii v) [ "yes"; "true"; "1" ]
+
+let months =
+  [ ("jan", 1); ("feb", 2); ("mar", 3); ("apr", 4); ("may", 5); ("jun", 6);
+    ("jul", 7); ("aug", 8); ("sep", 9); ("oct", 10); ("nov", 11); ("dec", 12) ]
+
+let weekdays =
+  [ ("sun", Date.Sun); ("mon", Date.Mon); ("tue", Date.Tue); ("wed", Date.Wed);
+    ("thu", Date.Thu); ("fri", Date.Fri); ("sat", Date.Sat) ]
+
+(* The three {!Date_spec} shapes, flattened. Every failure names the section
+   and shows the three forms, because a wrong date is the commonest mistake
+   and "invalid date" alone does not tell anyone what to type. *)
+let parse_date ~section raw =
+  let bad () =
+    err
+      "section [%s]: date %S is not one of the three forms: MM-DD (a civil \
+       date), easter+N or easter-N (signed days from Easter), or mon/day/nth \
+       such as oct/sun/1 or oct/sun/-1 (nth may be negative to count from the \
+       end of the month)"
+      section raw
+  in
+  let low = String.lowercase_ascii raw in
+  let starts p = String.length low >= String.length p && String.sub low 0 (String.length p) = p in
+  if starts "easter" then
+    let tail = String.sub raw 6 (String.length raw - 6) in
+    match int_of_string_opt (String.trim tail) with
+    | Some n -> (
+        match Date_spec.easter_offset n with Ok d -> Ok d | Error e -> err "section [%s]: %s" section e)
+    | None -> bad ()
+  else
+    match String.split_on_char '/' low with
+    | [ m; w; n ] -> (
+        match (List.assoc_opt m months, List.assoc_opt w weekdays, int_of_string_opt n) with
+        | Some month, Some weekday, Some nth -> (
+            match Date_spec.nth_weekday ~month ~nth ~weekday with
+            | Ok d -> Ok d
+            | Error e -> err "section [%s]: %s" section e)
+        | _ -> bad ())
+    | _ -> (
+        match String.split_on_char '-' low with
+        | [ mm; dd ] -> (
+            match (int_of_string_opt mm, int_of_string_opt dd) with
+            | Some month, Some day -> (
+                match Date_spec.fixed ~month ~day with
+                | Ok d -> Ok d
+                | Error e -> err "section [%s]: %s" section e)
+            | _ -> bad ())
+        | _ -> bad ())
+
+let ( let* ) r f = match r with Ok v -> f v | Error e -> Error e
+
+let parse_names ~section sec =
+  let entries =
+    List.filter_map
+      (fun (k, v) ->
+        if String.length k > 5 && String.sub k 0 5 = "name." then
+          Some (String.sub k 5 (String.length k - 5), v)
+        else None)
+      sec.fields
+  in
+  match entries with
+  | [] -> err "section [%s]: at least one name.<lang> is required (e.g. name.en)" section
+  | _ ->
+      List.fold_left
+        (fun acc (l, v) ->
+          let* names = acc in
+          match Lang.of_string l with
+          | Error e -> err "section [%s]: name.%s: %s" section l e
+          | Ok lang -> Ok (Names.set names lang v))
+        (Ok Names.empty) entries
+
+let celebration ~section ~id ~rank_of_string sec =
+  let* slug =
+    match Slug.of_string section with
+    | Ok s -> Ok s
+    | Error e -> err "section [%s]: %s" section e
+  in
+  let* names = parse_names ~section sec in
+  let* rank =
+    match get sec "rank" with
+    | None -> err "section [%s]: rank is required (class-1, class-2, class-3 or class-4)" section
+    | Some r -> (
+        match rank_of_string r with
+        | Some r -> Ok r
+        | None -> err "section [%s]: unknown rank %S -- expected class-1..class-4" section r)
+  in
+  let* colour =
+    match get sec "colour" with
+    | None -> err "section [%s]: colour is required (white, red, violet, green, black, rose)" section
+    | Some c -> (
+        match Colour.of_string (String.lowercase_ascii c) with
+        | Some c -> Ok c
+        | None ->
+            err "section [%s]: unknown colour %S -- expected white, red, violet, green, black or rose"
+              section c)
+  in
+  let* status =
+    match get sec "status" with
+    | None | Some "feast" -> Ok Celebration.Feast
+    | Some "commemoration" -> Ok Celebration.Commemoration_only
+    | Some s -> err "section [%s]: unknown status %S -- expected feast or commemoration" section s
+  in
+  let* subject =
+    match Option.map String.lowercase_ascii (get sec "subject") with
+    | None | Some "saint" -> Ok Subject.Saint
+    | Some "lord" -> Ok Subject.Lord
+    | Some "bvm" -> Ok Subject.Bvm
+    | Some "temporal" -> Ok Subject.Temporal
+    | Some s -> err "section [%s]: unknown subject %S -- expected lord, bvm, saint or temporal" section s
+  in
+  Ok (Celebration.make ~slug ~names ~rank ~status ~colour ~subject ~citations:[] ~layer:id ())
+
+(* Field edits the flat form can express. Citation edits and Remove_name are
+   deliberately absent: both need a structured key this format has no shape
+   for, and a half-expressible edit is worse than one the parser refuses by
+   name. *)
+let edits ~section ~rank_of_string sec =
+  List.fold_left
+    (fun acc (k, v) ->
+      let* es = acc in
+      match k with
+      | "edit" -> Ok es
+      | "rank" -> (
+          match rank_of_string v with
+          | Some r -> Ok (Overlay.Set_rank r :: es)
+          | None -> err "section [%s]: unknown rank %S" section v)
+      | "colour" -> (
+          match Colour.of_string (String.lowercase_ascii v) with
+          | Some c -> Ok (Overlay.Set_colour c :: es)
+          | None -> err "section [%s]: unknown colour %S" section v)
+      | "subject" -> (
+          match String.lowercase_ascii v with
+          | "lord" -> Ok (Overlay.Set_subject Subject.Lord :: es)
+          | "bvm" -> Ok (Overlay.Set_subject Subject.Bvm :: es)
+          | "saint" -> Ok (Overlay.Set_subject Subject.Saint :: es)
+          | "temporal" -> Ok (Overlay.Set_subject Subject.Temporal :: es)
+          | _ -> err "section [%s]: unknown subject %S" section v)
+      | k when String.length k > 5 && String.sub k 0 5 = "name." -> (
+          let l = String.sub k 5 (String.length k - 5) in
+          match Lang.of_string l with
+          | Ok lang -> Ok (Overlay.Set_name (lang, v) :: es)
+          | Error e -> err "section [%s]: name.%s: %s" section l e)
+      | _ ->
+          err
+            "section [%s]: %S cannot be edited from an INI overlay -- this form \
+             expresses rank, colour, subject and name.<lang> only. Write the \
+             S-expression form for anything else (see colitur-overlay(5))."
+            section k)
+    (Ok []) sec.fields
+  |> Result.map List.rev
+
+let parse ~rank_of_string text =
+  let* sections = parse_sections text in
+  let header, entries = List.partition (fun s -> s.name = "overlay") sections in
+  let* id =
+    match header with
+    | [ h ] -> (
+        match get h "id" with
+        | Some id when String.trim id <> "" -> Ok id
+        | _ -> err "the [overlay] section needs an id (e.g. id = my-parish)")
+    | [] -> err "no [overlay] section: the file must open with one, carrying id = <name>"
+    | _ -> err "more than one [overlay] section"
+  in
+  let* directives =
+    List.fold_left
+      (fun acc sec ->
+        let* ds = acc in
+        let section = sec.name in
+        match (get sec "suppress", get sec "replace", get sec "edit") with
+        | Some v, _, _ when is_yes v -> (
+            match Slug.of_string section with
+            | Ok s -> Ok (Overlay.Suppress s :: ds)
+            | Error e -> err "section [%s]: %s" section e)
+        | _, Some v, _ when is_yes v ->
+            err
+              "section [%s]: replace is not expressible in the INI form -- it \
+               needs a whole entry, which is what the S-expression form is for \
+               (see colitur-overlay(5)). Suppress plus a fresh section is \
+               usually what you want instead."
+              section
+        | _, _, Some v when is_yes v -> (
+            let* es = edits ~section ~rank_of_string sec in
+            match Slug.of_string section with
+            | Ok s when es <> [] -> Ok (Overlay.Edit (s, es) :: ds)
+            | Ok _ -> err "section [%s]: edit = yes but no editable field given" section
+            | Error e -> err "section [%s]: %s" section e)
+        | _ ->
+            let* date =
+              match get sec "date" with
+              | None -> err "section [%s]: date is required" section
+              | Some d -> parse_date ~section d
+            in
+            let* cel = celebration ~section ~id ~rank_of_string sec in
+            Ok (Overlay.Add { Layer.date; cel } :: ds))
+      (Ok []) entries
+    |> Result.map List.rev
+  in
+  Ok { Overlay.id; directives }
+
+let to_sexp_string rank_to_sexp t =
+  Printf.sprintf
+    ";; GENERATED by `colitur convert` from an INI overlay -- edit the INI and\n\
+     ;; regenerate, or adopt this file and drop the INI, but do not maintain\n\
+     ;; both. The conversion verified that this file parses back to exactly\n\
+     ;; what the INI denoted.\n\
+     %s\n"
+    (Sexplib0.Sexp.to_string_hum (Overlay.sexp_of_t rank_to_sexp t))
+
+let convert ~rank_of_string ~rank_to_sexp ~rank_of_sexp text =
+  let* t = parse ~rank_of_string text in
+  let rendered = to_sexp_string rank_to_sexp t in
+  (* The self-check this module exists for. Parse the emitted text back with
+     the SAME function the engine uses, and require the result to equal what
+     the INI denoted. A transpiler emitting valid-but-wrong sexp is the failure
+     a convenience format invites, and `colitur check` could never catch it:
+     the output would parse cleanly and simply mean something else. *)
+  (* Parse the ACTUAL text about to be returned -- not a fresh serialisation of
+     [t], which would make this check vacuous: [t] round-tripping through
+     [sexp_of_t]/[t_of_sexp] is true by construction and proves nothing about
+     [rendered]. The first version of this function did exactly that, and a
+     mutation corrupting the renderer (emitting a different overlay id) sailed
+     straight through it and exited 0. Two tests in test_overlay_ini.ml catch
+     that mutation now; they did not before, because they too must parse the
+     RETURNED text rather than re-derive it. *)
+  match Sexplib.Sexp.of_string rendered with
+  | exception exn ->
+      err "internal: generated sexp does not re-parse (%s) -- this is a bug in colitur, not in your file"
+        (Printexc.to_string exn)
+  | sexp -> (
+      match Overlay.t_of_sexp rank_of_sexp sexp with
+      | exception exn ->
+          err
+            "internal: generated sexp does not load (%s) -- this is a bug in colitur, not in your file"
+            (Printexc.to_string exn)
+      | back ->
+          if back = t then Ok rendered
+          else
+            err
+              "internal: the generated sexp does not mean what the INI said -- \
+               this is a bug in colitur, not in your file. Nothing was written.")
diff --git a/lib/kernel/overlay_ini.mli b/lib/kernel/overlay_ini.mli
new file mode 100644
index 0000000..bace93b
--- /dev/null
+++ b/lib/kernel/overlay_ini.mli
@@ -0,0 +1,69 @@
+(** A flat INI front end for overlay files.
+
+    This is a CONVENIENCE FORMAT, not a second data model. It parses to exactly
+    the {!Overlay.t} the S-expression form parses to, and everything downstream
+    -- merge, diagnostics, validation -- is the same code on the same values.
+    There is deliberately no second semantics to keep in step.
+
+    It is also deliberately LESS EXPRESSIVE than the sexp form. It covers [Add],
+    [Suppress] and single-field [Edit], which is what a diocesan or parish
+    calendar needs; [Replace], multi-field edits and citation edits are not
+    expressible and the parser says so by name rather than failing obscurely.
+    Anything it cannot say is a reason to write sexp, not a reason to grow this.
+
+    {1 Format}
+
+    Section names are slugs. A [\[overlay\]] section carries the file's id.
+
+    {v
+    [overlay]
+    id = my-parish
+
+    [our-patron]
+    date    = 07-11
+    rank    = class-3
+    colour  = white
+    name.en = St Example, Patron
+
+    [some-universal-slug]
+    suppress = yes
+    v}
+
+    Dates take three forms, matching {!Date_spec}: [MM-DD], [easter+N] or
+    [easter-N], and [mon/day/nth] such as [oct/sun/1] or [oct/sun/-1]. *)
+
+(** [parse ~rank_of_string text] is the overlay [text] denotes.
+
+    [rank_of_string] is supplied by the rite, exactly as [Overlay.load] takes
+    [rank_of_sexp]: the kernel does not know one rite's rank vocabulary from
+    another's.
+
+    Errors carry the section name and the offending value, never a source-file
+    path -- this format exists for people who are not reading the source. *)
+val parse :
+  rank_of_string:(string -> 'r option) -> string -> ('r Overlay.t, string) result
+
+(** [to_sexp_string t] renders [t] as the S-expression form, with a header
+    noting that it was generated. *)
+val to_sexp_string : ('r -> Sexplib0.Sexp.t) -> 'r Overlay.t -> string
+
+(** [convert ~rank_of_string ~rank_to_sexp ~rank_of_sexp text] parses [text],
+    renders it, and PROVES the rendering before returning it: the emitted text
+    is parsed back with the very function the engine uses to load an overlay,
+    and the result must equal what the INI denoted.
+
+    That check is the point of this module. A transpiler that emits
+    syntactically valid but semantically wrong output is the failure mode a
+    convenience format invites, and it is one [colitur check] could not catch,
+    since the emitted file would parse cleanly and simply mean something else.
+    Verifying the round trip here makes that class of bug impossible to ship
+    rather than merely unlikely.
+
+    [Error] on a parse failure, and on a round-trip mismatch -- which is a bug
+    in this module, and says so. *)
+val convert :
+  rank_of_string:(string -> 'r option) ->
+  rank_to_sexp:('r -> Sexplib0.Sexp.t) ->
+  rank_of_sexp:(Sexplib0.Sexp.t -> 'r) ->
+  string ->
+  (string, string) result