Plain-text ruby annotation controls no longer draw a stray glyph
Japanese text copied with ruby annotations in plain-text form carries the interlinear annotation controls U+FFF9–FFFB. They now draw nothing and add no width, so an annotated line no longer picks up a stray space or an unrelated kanji.
Japanese text that carries ruby annotations in plain-text form includes the interlinear annotation controls — U+FFF9 anchor, U+FFFA separator and U+FFFB terminator. They mark where a reading sits above the base text and are zero-width, but Unicode deliberately leaves them out of its default-ignorable set, so the renderer used to look for a glyph anyway: it substituted a space, or in some fonts an unrelated kanji.
They now draw nothing and add no width to the measured line. Only the three controls are skipped — the annotation text they bracket still renders normally, as ordinary inline text, so 日<U+FFF9>本<U+FFFA>やま<U+FFFB>語 reads 日本やま語 with no gap and no stray character.
This applies to every entry point — POST /api/v1/md, templates through /api/v1/render, the canvas endpoint and the web app's converters. Nothing to change on your side: re-render any document where you saw an unexpected gap or stray character and it will come out clean.