Text laid out top to bottom, column by column, right to left - vertical writing, which
canvas has no writing-mode for at all. Each character is its own Text, positioned by
layoutVertical; this is real per-glyph work, not a flag, which is why the README
schedules it separately from left-to-right/right-to-left (see mwg/i18n).
What this does not do: tate-chu-yoko (shrinking a short latin/digit run to sit upright
within one column instead of rotating whole characters), and it does not reposition
punctuation to a cell's corner the way professionally typeset Japanese does. Both are
real refinements a native reader would want on top of this, not included here.
Text laid out top to bottom, column by column, right to left - vertical writing, which canvas has no writing-mode for at all. Each character is its own
Text, positioned bylayoutVertical; this is real per-glyph work, not a flag, which is why the README schedules it separately from left-to-right/right-to-left (seemwg/i18n).What this does not do: tate-chu-yoko (shrinking a short latin/digit run to sit upright within one column instead of rotating whole characters), and it does not reposition punctuation to a cell's corner the way professionally typeset Japanese does. Both are real refinements a native reader would want on top of this, not included here.
Example