fix(cli): make the flag surface systematic

An audit probed every flag against every command rather than reading --help,
and found three inconsistencies.

--rite reached only `day` and `readings`. main.ml's own comment gave the
reason -- "has not been widened to a second rite in this task" -- a scope note
that had outlived its task and hardened into apparent design. View.of_days was
already polymorphic over the rite's type parameters, so widening was plumbing,
not library work: `temporal`, `rubrics`, `emit`, `table`, `render` and
`publish` now all take it.

`temporal` was the sharpest case. Its refusal said "--rite has no effect on
`temporal`", which was false: the EF has Septuagesima and Passiontide, the OF
neither, and every EF slug is ef-prefixed, so the flag would change nearly
every line. A message claiming no effect where the effect is total is exactly
what the audit set out to find.

Three ways of naming a year (positional, --year, --from/--to) now cross-accept
additively; naming two that disagree is a usage error rather than one silently
winning.

emit --format csv gained a rite-dependent header: EF's 16 columns are
unchanged, OF gets a 17th "second" between "first" and "gospel". An earlier
task had recorded RFC 4180 as permanently blocking this; the rule constrains
one file, not a family of them.

`rubrics` keeps refusing --lang/--raw, now with its reason. An intermediate
version accepted them by adding a name column, which changed the default from
six tab-separated fields to seven and broke both existing consumers and the
byte-identical-EF rule. The asymmetry is real but principled: the row is a
date, a slug, a source keyword, two booleans and a preface key, so there is
nothing to translate and nothing to strip. `easter` refuses --rite for the
same kind of reason -- its six anchors sit at identical Easter offsets in both
rites.

EF output verified byte-identical to 8590338 across day, readings, rubrics,
temporal and easter for 2026, 1583 and 9999, and across all five emit formats.

1 files changed, 41 insertions, 22 deletions

diff --git a/lib/render/emit_csv.ml b/lib/render/emit_csv.ml
index e04211f..ff11a81 100644
--- a/lib/render/emit_csv.ml
+++ b/lib/render/emit_csv.ml
@@ -23,26 +23,42 @@ let s v k = match get v k with Some (T.Str x) -> x | _ -> ""
    Machine formats still carry [slug] alongside it, so a script keeps the
    stable key and a human reads the name.
 
-   W5: deliberately NOT widened with a "second" column, unlike
+   W5 left this deliberately NOT widened with a "second" column, unlike
    [Emit_xml]/[Emit_ics] (each a repeated-element/free-text format that
-   grows for free). A CSV's header row is a CONTRACT with every row under
+   grows for free): a CSV's header row is a CONTRACT with every row under
    it (RFC 4180 sec. 2: "each field in the header... should contain the
    same number of fields as the records"), so there is no way to add a
    column only on the days that carry a [Second] reading -- a real
-   three-citation day would either need an ALWAYS-present, mostly-empty
-   "second" column (widening every EF row forever for a case EF can never
-   hit) or genuinely variable-width rows (not CSV). Neither is a change to
-   make speculatively: [bin/main.ml]'s [reject_rite_for] refuses [--rite]
-   on `emit` entirely for any rite but EF, so no [Second] reading can reach
-   this module today, and EF's own [citations] is always exactly
-   [[First; Gospel]] ({!Rite_ef}'s own [citation_shapes]) -- there is
-   nothing to widen FOR yet. When OF is admitted here, this is the one
-   emitter that needs a deliberate, disclosed breaking change (a new
-   column, a version bump, or a documented "third reading silently
-   dropped" trade-off), not a silent one. *)
-let columns =
-  [ "date"; "rite"; "season"; "season_name"; "week"; "slug"; "name"; "weekday";
-    "rank"; "rank_name"; "colour"; "colour_name"; "subject"; "first"; "gospel"; "comms" ]
+   three-citation day needs either an ALWAYS-present, mostly-empty
+   "second" column or genuinely variable-width rows (not CSV). At the
+   time, [bin/main.ml]'s [reject_rite_for] refused [--rite] on `emit`
+   entirely for any rite but EF, so no [Second] reading could reach this
+   module, and this was recorded as the deliberate, disclosed breaking
+   change the day OF was admitted here would need -- "a new column, a
+   version bump, or a documented trade-off, not a silent one".
+
+   Fix 1 (cli-flags-report, 2026-08-27): that day is this one. The
+   "always-present, mostly-empty column" option is taken, gated on
+   [rite] (read from [v] itself, exactly as [row]'s own [~rite] already
+   is) rather than threaded as a new parameter: EF's [rite = "ef"] keeps
+   the ORIGINAL 16-column header and rows, byte for byte -- no [Second]
+   reading has ever reached EF's own [citations]
+   ({!Rite_ef}'s own [citation_shapes] is still exactly
+   [[First; Gospel]]), so [rite = "ef"]'s branch below is provably
+   unreachable-different, not merely untested-different. Any OTHER rite
+   (OF today) gets a 17th column, "second", spliced between "first" and
+   "gospel" -- the same position [Emit_xml]/[Emit_ics] already use for
+   it -- reading [s d "second"], which [get]'s own [None -> ""] fallback
+   already resolves to the empty string on every day that has no [Second]
+   part (a feria, feast or memorial; OLM 1981 Praenotanda n. 66.1/n. 69.1),
+   so nothing here needs to know in advance which OF days do carry one. *)
+let columns ~rite =
+  let base =
+    [ "date"; "rite"; "season"; "season_name"; "week"; "slug"; "name"; "weekday"; "rank"; "rank_name";
+      "colour"; "colour_name"; "subject"; "first" ]
+  in
+  let tail = [ "gospel"; "comms" ] in
+  if rite = "ef" then base @ tail else base @ [ "second" ] @ tail
 
 let row ~rite d =
   let comms =
@@ -50,17 +66,20 @@ let row ~rite d =
     | Some (T.List l) -> String.concat " " (List.map (fun c -> s c "slug") l)
     | _ -> ""
   in
-  String.concat ","
-    (List.map escape_field
-       [ s d "iso"; rite; s d "season"; s d "season_name"; s d "week"; s d "slug";
-         s d "name"; s d "weekday"; s d "rank"; s d "rank_name"; s d "colour";
-         s d "colour_name"; s d "subject"; s d "first"; s d "gospel"; comms ])
+  let base =
+    [ s d "iso"; rite; s d "season"; s d "season_name"; s d "week"; s d "slug"; s d "name";
+      s d "weekday"; s d "rank"; s d "rank_name"; s d "colour"; s d "colour_name"; s d "subject";
+      s d "first" ]
+  in
+  let tail = [ s d "gospel"; comms ] in
+  let fields = if rite = "ef" then base @ tail else base @ [ s d "second" ] @ tail in
+  String.concat "," (List.map escape_field fields)
 
 let year v =
   let rite = s v "rite" in
   let days = match get v "days" with Some (T.List l) -> l | _ -> [] in
   let b = Buffer.create (64 * 400) in
-  Buffer.add_string b (String.concat "," columns);
+  Buffer.add_string b (String.concat "," (columns ~rite));
   Buffer.add_char b '\n';
   List.iter (fun d -> Buffer.add_string b (row ~rite d); Buffer.add_char b '\n') days;
   Buffer.contents b