Add header pitch and leading to clin_row_height() (#117)

[mstackhouse]

Closes #117. The last of the three vertical controls #97 asked for.

Measuring settled the design question you left open

You asked whether a header height would reach the multi-line case, and suspected not. It doesn't. Reading what Word actually receives for a 3-line header cell:

lever	emits	reaches a 3-line cell?
clin_row_height(body = 15.35)	nothing on the header — stays line=240, trHeight=796/auto	no
header height, rule = "atleast"	trHeight=260/atLeast	no — a floor, the cell still grows past it
header height, rule = "exact"	trHeight=260/exact	only by clipping
line_spacing(space = 0.75)	w:line=180	yes

So both ship, and the docs say which one to reach for. header bounds the header rows the way body bounds the body rows; header_leading closes the gap between the lines inside a cell, which is what a wrapped arm label needs.

clin_row_height(body = 15.35, title = 11.4, footnote = 11.4,
                header = 13, header_leading = 0.75)

Verified together: hdr line=180, hdr trH=260/atLeast, body trH=307/atLeast — three independent pitches, none interfering.

Leading is a multiple, not a length

The one wrinkle. flextable's line_spacing() takes a multiple of single spacing and Word writes it that way — space = 0.75 becomes w:line=180 against a single-spacing 240, with no lineRule. There is no points form to convert to without hand-writing w:lineRule="exact", which would mean going around flextable into officer's XML.

So header_leading is a multiple, unit does not apply to it, and it is named and documented accordingly rather than pretending to be a length like its neighbours.

It has to be deferred, and your styler is why

Applied at verb-call time it would be undone: your clinify_defaults.R calls flextable::line_spacing(x, space = 1, part = "all") in all three styling functions. So it is applied in finish_table_(), after the option styling function — the same precedence rule as the rest of the per-table config.

There is a test that stands that styler up and checks the leading survives at 0.75 rather than being reset to 1.

Verification

• 2,156 tests pass, zero snapshot changes.
• devtools::check() — 0 errors, 0 warnings, and the unable to verify current time NOTE, which is the no-network note this repo's cran-comments.md documents as local-only; it comes and goes between runs on this machine.
• Lints on R/row_height.R: 4, unchanged from development.
• Assertions read w:line, w:trHeight and hRule out of word/document.xml, so they check what Word receives rather than in-memory fields.
• One pre-existing assertion on the "nothing asked for" message was updated, since it now names all five arguments.

Note

Thank you for the header line-structure check — the finding that 9 of 30 tables stack their header lines differently from the reference is exactly the kind of thing that justifies this being an API lever rather than a per-project workaround. Worth saying that header_leading changes the leading, not where the lines break: if 14-1.02 is putting p-value on header line 5 where the reference has line 6, that is a wrapping/level-count difference rather than a spacing one, and this won't move it. Happy to look at that separately if you file it.

DESCRIPTION left at 0.4.0 with the entry under # clinify (development version), as with #114/#115/#116.

🤖 Generated with Claude Code