/* ==========================================================================
   print.css — shared print layer for all ops tools
   --------------------------------------------------------------------------
   Extracted from pars-cards.html. These numbers were arrived at by trial
   against Chrome's print engine. Do not "clean them up" without printing.

   THE CONTRACT
     - Screen UI lives inside .app
     - Print output is written into #printArea
     - Each printed page is one .sheet element
     - A .sheet sets --rh (row height) and --fs (font size) inline, from
       the values RSFit.fitRows() returns. See assets/fit.js.
     - A sheet is two columns read down-then-across (SPEC §4): .cols holds
       two .col cells, each with its own .ptable. Column one takes
       fit.split rows; a ragged split repeats the group header in column
       two with "(cont.)" AND its own Cook/Mgr checks — each half of a
       split group is checked where it lives (Adam, 2026-08-07, amending
       the one-check-per-group lock for the split case).
     - The footer (.sheet-foot) pins to the page bottom, not to the last
       row. A short sheet leaves white space in the middle; that is the
       decision that makes the 0.55in row cap look deliberate.

   WHY THESE NUMBERS
     0.5in margins  Chrome triggers shrink-to-fit at tighter margins, which
                    silently rescales the whole sheet. 0.5in is the safe floor.
                    Users must also set Margins=Default and Scale=100.
     9.92in height  The usable box inside 0.5in margins on Letter, minus a
                    hair of slack. Paired with overflow:hidden so a sheet is
                    structurally incapable of spilling onto a second page.
                    Content overruns get clipped instead of silently
                    reflowing — a visible failure beats an invisible one.
     no flexbox     Chrome's print engine collapses flex layout here. Tables
                    and block layout only inside .sheet.
   ========================================================================== */

#printArea { display: none; }

@media print {
  .app { display: none !important; }
  #printArea { display: block !important; }
  body { background: #FFFFFF; }
}

@page { size: letter portrait; margin: 0.5in; }

/* ---------- sheet ---------- */

.sheet {
  page-break-after: always;
  color: var(--ink);
  --rh: 44px;
  --fs: 15px;
  height: 9.92in;
  overflow: hidden;
  position: relative; /* .sheet-foot pins to the bottom of this box */
}
.sheet:last-child { page-break-after: auto; }

/* The tape header is gone (2026-08-07): it was background color, which
   Chrome doesn't print by default — the wall photos show it as plain bold
   text. Borders and weight only (locked). Its 22px display size carries
   over to the title; the double rule replaces the tape as the header's
   weight. */
.sheet-head {
  display: flex;
  justify-content: space-between;
  align-items: flex-end;
  margin-bottom: 6px;
  border-bottom: 3px double var(--ink);
  padding-bottom: 4px;
}
/* THE TYPE DECISION (1.3, settled deliberately, not inherited)
   Constraint first: the printed sheet must render identically offline,
   from file://, and on a phone — no webfont participates in print layout,
   because a font that fails to load reflows the sheet silently (SPEC:
   zero runtime dependencies; the Google Fonts link is screen-UI only).
   Within that constraint, the reference posture — condensed, heavy,
   uppercase station name over clean sans body — is reached with what's
   installed: Arial Narrow (Windows/macOS/iOS), Roboto Condensed
   (Android), Calibri then Arial as wide-but-present fallbacks. Condensed
   is also what the measured column budget wants: full 15.5px items in a
   1.70in column beat 13px items in a wider face — density where it
   counts. A display webfont (Oswald) for the station name alone was
   considered and declined: it buys posture the wall never printed (the
   tape header photographed as plain bold text) at the price of ~30KB
   inlined in the standalone build. If the printed title ever needs more
   weight than Arial Narrow bold, that's Adam's call to fund, not a
   quiet stack change. */
.sheet-title {
  font-family: "Arial Narrow", "Roboto Condensed", Calibri, "Helvetica Neue", Arial, sans-serif;
  font-size: 22px;
  font-weight: 700;
  text-transform: uppercase;
  letter-spacing: 0.02em;
}
/* Pars version stamp (SPEC §4): which generation of pars this sheet was
   printed from — not the shift DATE. Reads as a stamp under the station
   name: small, letterspaced, no competition with the title. Restored in
   1.5; the wall sheets carry it as "PARS GUIDE · v 2026-08-06". */
.sheet-stamp {
  font-size: 9px;
  letter-spacing: 0.14em;
  margin-top: 2px;
}
.sheet-meta { font-size: 10px; }
/* the exception convention, stated on the sheet itself (SPEC §8): a
   dash instead of initials means addressed, see notes — same reasoning
   as the 86 line. Lives under the DATE/SHIFT line so the head budget
   doesn't grow. */
.sheet-legend {
  font-size: 8.5px;
  color: var(--steel);
  text-align: right;
  margin-top: 4px;
}

/* Empty circle for the SHIFT marker — border only, prints without
   background graphics. */
.shift-o {
  display: inline-block;
  width: 10px;
  height: 10px;
  border: 1.5px solid var(--ink);
  border-radius: 50%;
  vertical-align: -1px;
  margin: 0 2px 0 6px;
}

/* ---------- two columns ---------- */
/* Real table display, not flex — Chrome's print engine collapses flex
   inside the sheet (see header note), tables survive. Column budget from
   SPEC §4, closes at exactly 3.60in: item 1.70 · par 0.40 · where 1.50.
   Letter inside 0.5in margins is 7.5in: 3.60 + 0.30 gap + 3.60. */

.cols {
  display: table;
  table-layout: fixed;
  width: 100%;
  border-spacing: 0;
}
.col {
  display: table-cell;
  vertical-align: top;
  width: 3.60in;
}
.col + .col {
  width: 3.90in;
  padding-left: 0.30in; /* the column gap */
}

/* ---------- printed table ---------- */

.ptable { width: 100%; border-collapse: collapse; table-layout: fixed; }

/* Column widths — the verified budget, exactly 3.60in (SPEC §4).
   table-layout: fixed makes the browser honor them instead of measuring
   content, which is what lets the fit math trust that only the two known
   `where` strings wrap.

   THE TYPE THAT MAKES THE BUDGET TRUE (measured in Chromium via canvas,
   2026-08-07, against every real string in pars-backup-grid.json; cell
   content = column minus 2×5px padding):
     widest one-line where, "1/6 pan · on line + under":
       Arial Narrow 13px = 115.9px · Calibri 13px = 129.9px — fits 134px
       Segoe UI 15.5px = 170.4px — wraps, which is what the first print
       showed: half the sheet on two lines
     widest item in the reference data (22 chars) at bold 15.5px:
       Arial Narrow = 139.9px · Calibri = 150.8px — fits 153.2px
   On a generic sans fallback the worst couple of strings wrap again;
   render.js measures every cell for real and budgets the lines, so a
   wide fallback costs row height, never clipped content. Do not widen
   the padding without re-measuring. The stack matches .sheet-title's —
   the decision and its reasoning live beside that rule. render.js
   mirrors this stack and these widths; scripts/test.js asserts the
   mirror. */
.ptable {
  font-family: "Arial Narrow", "Roboto Condensed", Calibri, "Helvetica Neue", Arial, sans-serif;
}
.ptable .c-item  { width: 1.70in; }
.ptable .c-par   { width: 0.40in; }
.ptable .c-where { width: 1.50in; }
/* run checklist columns (5.1): task 3.20 · check 0.40, same 3.60in
   close. Realistic task names run well under 3.20in at 15.5px — wraps
   are budgeted by the same measurement path regardless. */
.ptable .c-task  { width: 3.20in; }
.ptable .c-check { width: 0.40in; }
/* a checkbox per task (Adam's paper call, 5.1) — the same 18px box the
   pars group rows use, centered in its cell */
.ptable td.check { text-align: center; }
.ptable td.check .cbox { margin-left: 0; }

/* ---------- rotation matrix (5.2) ---------- */
/* Task column 2.5in, occurrence columns split the remaining 5.0in
   (7 × 0.71 / 5 × 1.00 / 12 × 0.42 — widths set inline per sheet).
   The grid IS the borders: column separators + the row rules the
   family already draws. An empty cell is the point — a gap must read
   as neglect at arm's length, so nothing decorates it. */
.mtable .c-mtask { width: 2.5in; }
.mtable th.c-mcol {
  text-align: center;
  padding: 0 2px 5px;
  font-size: 8.5px;
  letter-spacing: 0.06em;
}
.mtable td.mcell {
  border-left: 1px solid var(--line-soft);
  text-align: center;
}

.ptable th {
  text-align: left;
  font-size: 11px;
  text-transform: uppercase;
  letter-spacing: 0.1em;
  color: var(--ink);
  border-bottom: 2px solid var(--ink);
  padding: 0 5px 5px; /* 5px sides — part of the measured budget above */
  font-weight: 600;
}
/* The spanning-header rules (Mgr initials > Day/Night) were removed with
   the feature: the initials columns were cut for doing no work on paper. */

.ptable td {
  border-bottom: 1px solid var(--line-soft);
  padding: 0 5px 4px; /* 5px sides — part of the measured budget above */
  height: var(--rh);
  font-size: var(--fs);
  vertical-align: bottom;
}

/* ---------- row tracking ---------- */
/* Restored 1.4: the wall sheets had a heavier rule every fifth row and
   the two-column rewrite dropped it — the paper test showed the eye
   losing the row crossing from item to where. render.js counts item
   rows only, restarting at each group, and tags every fifth with
   .rule5. The left rule fences each group's items so a column reads as
   blocks. Borders only — neither cue depends on background graphics. */
.ptable tr.rule5 td { border-bottom: 1.5px solid #8A8578; }
.ptable td.item { border-left: 1px solid var(--line-soft); }

.ptable .item { font-weight: 600; }
/* item name and par get the legibility budget — par prints as a number.
   An item with NO par prints a write-in rule (locked, §5 amendment
   2026-08-07): Salad printed with an entirely blank par column because
   its pars only ever existed in marker on the wall, and a par cell with
   nowhere to write degrades the sheet into a location list. The rule
   sits where a printed number sits, so mixed rows stay aligned. */
.ptable td.par { text-align: center; font-weight: 600; }
.ptable td.par .wline { width: 85%; margin: 0 auto; }
/* where is reference detail, not the reason the sheet exists. Fixed 13px:
   --fs is pinned at 15.5 by the row-height floor (see fit.js), and 13px
   is the size the measured budget above clears. */
.ptable .where { color: #2A2823; font-size: 13px; word-wrap: break-word; }

/* ---------- group header rows ---------- */
/* Checks live here (locked): COOK and MGR on the header row, item rows
   carry no check column — that width is already spent. A split group
   carries the pair on both halves, including the (cont.) header (Adam,
   2026-08-07): the cook working the right column checks where they
   stand. A group row is one row in the fit budget like any other. */

.ptable tr.grow td {
  border-bottom: 2px solid var(--ink);
  font-size: 11px;
}
.ptable tr.grow .gname {
  text-transform: uppercase;
  letter-spacing: 0.06em;
  font-weight: 700;
}
/* the "(cont.)" marker on a repeated header stays lowercase and unbold —
   it's an aside, not a second group */
.ptable tr.grow .gname .cont { text-transform: none; font-weight: 400; }
.ptable tr.grow .gchecks {
  text-align: right;
  white-space: nowrap;
  letter-spacing: 0.04em;
}

/* empty checkbox — border only, prints without background graphics.
   18px ≈ 0.19in, settled by Adam after seeing both extremes on paper:
   the original 11px box was about a marker tip and a half — a check
   fills it into a blob — and the 26px correction read too heavy. The
   box is a target, not a container; a checkmark overflowing it is
   normal on the wall sheets. Fits any row height, costs no capacity. */
.cbox {
  display: inline-block;
  width: 18px;
  height: 18px;
  border: 1.5px solid var(--ink);
  vertical-align: middle;
  margin-left: 4px;
}

/* ---------- footer ---------- */
/* Pinned to the page bottom, not the last row: a short sheet leaves its
   white space in the middle, which reads as deliberate. Budgeted at 80px
   in fit.js (was 48 — the first paper print showed pen-sized bands a
   marker can't use; Adam, option B, 2026-08-07). Each write-in line gets
   a ~22px clear band above its rule; with the text line itself that's
   ~0.35in of marker room. 2 + 22 + 14 + 24 + 14 = 76px, 4px slack. */

.sheet-foot {
  position: absolute;
  left: 0;
  right: 0;
  bottom: 0;
  border-top: 2px solid var(--ink);
}
.foot-86 { font-size: 11px; padding-top: 22px; }
/* the checklist's notes band occupies the same slot in the same budget */
.foot-notes { font-size: 11px; padding-top: 22px; }
.foot-sign {
  font-size: 10px;
  text-transform: uppercase;
  letter-spacing: 0.08em;
  padding-top: 24px;
}

/* ---------- write-in lines ---------- */
/* Anything filled in by hand on the line prints as a dry-erase rule,
   never as data. Data that changes on the floor does not belong in toner. */

.wline { display: block; border-bottom: 1.5px solid var(--ink); height: 4px; }
/* inline variant for header/footer write-ins (DATE, 86, sign-off) —
   width set where it's used */
.wline.inline { display: inline-block; height: 10px; vertical-align: baseline; }
/* the .ini and .parcell write-in rules left with their columns: initials
   were cut, and pars print as numbers */
.ptable td.wide .wline { width: 96%; }