ggCheysson 1.1.0

Accepted on CRAN 2026-10-01: https://CRAN.R-project.org/package=ggCheysson. This update fixes
the 1.0.1 check ERROR on Intel macOS, corrects and renames palettes (see Breaking changes), and
adds the experimental scale_cheysson().

ggCheysson 1.1.0

Breaking changes

• Fixed a palette-naming collision bug. Palette names were built from
  Album + Qty (e.g. "1880_07"), but Qty is not a unique plate identifier - it's a
  colors/pattern-element count that happened to repeat across genuinely different plates. 4 such
  collisions silently overwrote 5 of the original 25 palettes during data extraction. Names are
  now built from the real, unique plate identifier (RumseyListNo's decimal suffix) instead:
  19 of the 20 previously-shipped palettes have a new name, and all 5 previously-missing palettes
  are now included - cheysson_palettes/cheysson_patterns now have 25 palettes (was 20),
  with 134 pattern specifications (was 83). See the new cheysson_labels dataset (below) to map
  an old name to its new one.

• Migration aids for the rename: an old (1.0.1) palette name that no longer exists now errors
  with the palette's new name, e.g. "1883_06" -> "1883_30", instead of just "not found".
  Watch out for "1880_07": it is still a valid name, but now refers to a different
  palette (grouped, 1880 plate 7). The palette called "1880_07" in 1.0.1 is now "1880_21".
  Using "1880_07" gives a once-per-session message saying so. ("1906_06" is the only old name
  that still means the same palette.)

• Palette order is now consistent: sequential palettes are stored from low to high (light to
  dark) and diverging palettes end to end, so reverse = TRUE means the same for every palette.
  Previously palettes kept RJ Andrews' swatch order, so 1891_19, 1891_25 and 1900_28 ran
  dark to light, and 1883_31 put its hatched patterns at the extremes; these are now
  reordered. Category and grouped palettes are unchanged.

• When a sequential or diverging palette has more elements than needed, cheysson_pal(n = ),
  cheysson_pattern(n = ) and all the scale_*_cheysson() scales now pick elements spread
  over the whole palette, keeping both ends, instead of the first n - which could drop one
  end of a diverging palette entirely (e.g. 1883_21 with 5 classes).

New features

• New (experimental) scale_cheysson() applies a whole color-and-pattern palette - fill, hatch
  type, line colors and angle - with one +, in place of four or five separate
  scale_*_cheysson() calls; aes_cheysson(x) maps a variable to all of those aesthetics.
  See the new vignette "Combining Colors and Patterns". The individual scales remain.

• Two-color crosshatches: the crosshatches in 1883_30 and 1886_17 draw their two sets of
  lines in different colors (red and blue; orange and blue-grey). These are now stored as
  pattern_fill2, returned by cheysson_pattern_params(param = "pattern_fill2"), and applied by
  the new scale_pattern_fill2_cheysson(); map ggpattern's pattern_fill2 aesthetic to use them.

• Added cheysson_labels, a new dataset mapping every palette's current name to its previous
  (pre-fix) name, RJ Andrews' original Advent-calendar label, Tom Shanley's Observable notebook
  ID, and the underlying David Rumsey catalog number - useful for looking up a palette by
  whichever naming scheme you encountered it in.

• Added cheysson_name(), a lookup function that translates a palette label from any of the
  cheysson_labels naming schemes (Andrews' label, Shanley's ID, a Rumsey catalog number, the
  old pre-fix name, or an advent day) into the current palette name, for direct use in calls like
  scale_color_cheysson(cheysson_name("Dec.01-1883.21")). When a label is ambiguous (only
  possible via the old naming scheme, which had 4 real collisions), it warns and lists every
  match; pass advent_day to disambiguate without a warning, or to assert a specific palette.

Bug fixes

• Vignettes now render figures with ragg. This fixes a segfault on CRAN's Intel macOS check
  machines, where the Quartz png() device crashes drawing ggpattern's masks. Also, set
  fig.showtext = TRUE, so showtext text is drawn at the device's real dpi and figures look the
  same in the package vignettes and on the pkgdown site.

• Missing values in a mapped pattern aesthetic no longer make ggpattern fail: the pattern
  scales now give missing data no hatching (pattern = "none") by default.

• The 1883 category palette of Advent day 22 is named 1883_30: the source data's Rumsey
  number for it (12512.013) was a typo for 12514.030, the 1883 album's plate 30, as checked
  against the original plate. cheysson_labels keeps RJ Andrews' and Tom Shanley's labels for
  it (Dec.22-1881.13, category12512013), which were built from the bad number.

• Fixed 15 missing pattern elements across 6 palettes (1882_04, 1883_07, 1886_04,
  1886_07, 1887_06, 1900_06, in their old names): an SVG-parsing bug silently dropped
  hatch-line patterns whose coordinates relied on SVG's implicit default of 0 for an omitted
  x1/y1/x2/y2 attribute.

• Fixed two scale_pattern_*_cheysson() bugs: scale_pattern_type_cheysson() targeted the
  wrong ggpattern aesthetic (pattern_type instead of pattern) and so never actually varied
  the rendered pattern shape by category; scale_pattern_fill_cheysson() read the wrong
  parameter (fill instead of pattern_fill) and always returned "transparent" regardless of
  palette.

• Fixed theme_cheysson()'s/theme_cheysson_map()'s plot.title (set in CheyssonTitle)
  rendering smaller than intended, the same underlying issue as the 1.0.1 axis/legend-title fix
  below but previously missed for the plot title itself. The correction is now generalized
  (cheysson_font_size_adjust()) to all four Cheysson display fonts instead of hardcoded to one.

Other changes

• The help topic for scale_color_cheysson()/scale_fill_cheysson() is now
  ?scale_color_cheysson (the name scale_cheysson belongs to the new function).

• Added real (rendered) pattern examples to the README's Pattern Support section, which
  previously had none.

• theme_cheysson_map() text sizing redesigned: plot.title, plot.subtitle,
  legend.title, and legend.text now scale from base_size using a deliberate poster-style
  hierarchy (a much larger title relative to the rest, in the spirit of Guerry's own maps) instead
  of the modest ratios theme_cheysson() uses; plot.caption is now styled too (previously
  unstyled and tiny). base_size alone now controls this - no need for the
  theme(plot.title = element_text(size = ...)) overrides the guerry-maps vignette previously
  needed on every map. Sizes are calibrated at true scale (a ~40-character title just fits an
  8in-wide figure), and titles are now centered on the whole plot rather than the panel, so a
  side legend no longer pushes a long title off the left edge.