tools(bootstrap_sanctoral): fix round 1 (F6) -- stale entry-count comments

Three comments still said "322" after the previous commit's regeneration
moved the total to 327: the date-form comment (rewritten to not carry a
count at all, since it is about a FORM, not a quantity); the explicit-class
comment (327, plus recorded that the 6 entries carrying it changed
membership -- 12 September dropped it, 13 January gained it -- even though
the count of 6 itself did not move); the name.pl comment (327 total, 5
missing it, 322 carrying it -- was "5 of the source's 322 entries",
conflating the new total with the old one).

1 files changed, 18 insertions, 8 deletions

diff --git a/tools/bootstrap_sanctoral.ml b/tools/bootstrap_sanctoral.ml
index 4135518..4868c6c 100644
--- a/tools/bootstrap_sanctoral.ml
+++ b/tools/bootstrap_sanctoral.ml
@@ -94,8 +94,10 @@ let parse_slug sec =
   | Ok s -> s
   | Error e -> die "section [%s]: invalid slug: %s" sec.name e
 
-(* All 322 entries use plain MM-DD (spec §4.1); anything else is a date form
-   the mapping does not cover. *)
+(* Every entry uses plain MM-DD (spec §4.1; 327 of them as of the
+   2026-08-12 ef-rebootstrap, up from 322 -- this rule itself is about the
+   date FORM, not a count, so it does not go stale as entries are added or
+   removed); anything else is a date form the mapping does not cover. *)
 let parse_date sec =
   let raw = field sec "date" in
   match String.split_on_char '-' raw with
@@ -140,10 +142,15 @@ let parse_colour sec =
 (* class = lord|bvm|saint -> Subject.t; absent -> Subject.Saint (decision 1:
    Celebration.make's kernel default is Subject.Temporal, correct for the
    temporal cycle and wrong for every sanctoral entry -- overridden here,
-   never left to the default). Only 6 of 322 entries carry an explicit
-   class in the source, and all 6 are "lord" as of this bootstrap; bvm/saint
-   are mapped in case a future lectio update adds one, but "temporal" is
-   rejected -- no sanctoral entry is the temporal cycle's own office. *)
+   never left to the default). Only 6 of 327 entries carry an explicit
+   class in the source (STALE DENOMINATOR CORRECTED, ef-rebootstrap fix
+   round 1, F6 -- was "6 of 322"; the count of 6 itself did not move, but
+   WHICH six did: the 2026-08-12 re-bootstrap dropped `class = lord` from
+   `most-holy-name-of-mary` and added it to
+   `commemoration-of-the-baptism-of-the-lord`, a straight swap, net zero),
+   and all 6 are "lord" as of this bootstrap; bvm/saint are mapped in case
+   a future lectio update adds one, but "temporal" is rejected -- no
+   sanctoral entry is the temporal cycle's own office. *)
 let parse_subject sec =
   match field_opt sec "class" with
   | None -> Subject.Saint
@@ -156,8 +163,11 @@ let parse_subject sec =
       | None -> die "section [%s]: unknown class %S" sec.name raw)
 
 (* name.en is always required; name.pl is not -- the 2026-08-12 re-bootstrap
-   (RG16(a) blast-radius task) found 5 of the source's 322 entries missing it
-   outright (agnes-secundo, boniface-martyr, eusebius-confessor, evaristus,
+   (RG16(a) blast-radius task) found 5 of the source's 327 entries missing
+   it outright, 322 still carrying it (CORRECTED, fix round 1, F6: was "5
+   of the source's 322 entries" -- the source has 327, of which 5 lack
+   name.pl and 322 have it; the 322 is a coincidence of the OLD total, not
+   the new one) (agnes-secundo, boniface-martyr, eusebius-confessor, evaristus,
    theodore -- all 5 either newly added or newly un-hidden by lectio's own
    generator fix, never previously reachable through this tool at all, since
    every entry the OLD ini shipped happened to carry both languages). Prior