rsyntaxtree 1.8.1 → 1.8.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9b9513e8d667af90de0a12e427eae11d0089f6a969ed1a47714ca30401001821
-  data.tar.gz: '0458b65b7f9ad1f495d344963b41b801b7c26562b0256abe48885756a1481bef'
+  metadata.gz: 232efc991d608eb2af9ff19d6816d9f22f9bc4a1006506276507c1b8705f34ec
+  data.tar.gz: da8e1ab08b470d76d47809b89ae4fd5b7bf0ef7519847ef3bc7e2955814f5521
 SHA512:
-  metadata.gz: d1738b00df2a2dbe5cafba5d52b19c44149472912c2859c8183384582770be0b54c0c4e96688fe4c0a808969c96096504956b7ce7dffc76bd3c682ee28124cf6
-  data.tar.gz: a8363b3b3e6b11f345603f020e36766275f740ab0f80af4ecb4be3f9160dd7860d4a26c4cdd2788a8b02e27d4c99d1147bd9843200cdd41cf48f87b6ce373e34
+  metadata.gz: 5af80d2af48070f3733c21f19cb543da6a2f443b4e4e2fb4fa2f5a0760584cedcfac5e3b86067623f5f24ba63cfe3267e69692f5ddbc44b3c9fdf35f1de34ace
+  data.tar.gz: 3d391279ae3b669f87f6c62cfa6b2f44cb607a6f5c9d0adbfd37e147cd9a9d2a8c4f998a3e508c423886a4adfe56610e5041c838ffcd40d47f958e96b45d9d49

data/CHANGELOG.md

@@ -1,5 +1,33 @@
 # Changelog
 
+## [1.8.2] - 2026-08
+
+### Changed
+- `tidy: high` now compresses as far as its name promises. The level-balance
+  floor added in 1.8.0, which keeps a pair of siblings from being tucked
+  narrower than the level below it, applied to `medium` and `high` alike and
+  bound first in nearly every tree: across the 75 gallery examples the two
+  modes produced identical figures 70 times. The floor now applies to `medium`
+  only, leaving `high` free to trade even branch angles for width — 29 of the
+  75 examples now differ, by up to 19%. `off`, `low` and `medium` are
+  unchanged; `high`, like `symmetric` at the other end of the scale, is for
+  figures that ask for it.
+- Mathematical alphanumerics (U+1D400–, such as the little *v* of *v*P) are
+  named in the family chains. Neither Noto Sans nor Noto Serif covers the
+  block, so the glyphs came from whatever the machine offered: Noto Sans Math
+  on Alpine, DejaVu Serif on Debian/Ubuntu, STIX Two Math on macOS. The serif
+  style now asks for a serif source first, so a serif tree no longer shows a
+  sans *v*. The Docker images install `font-dejavu` for it.
+
+### Fixed
+- The CLI printed parse errors to stdout and exited 0, so a script generating
+  figures in bulk could not tell a rejected input from a drawn one. Errors go
+  to stderr and the exit status is 1.
+- Dropped `Noto Sans Mono SemiCondensed` from the mono chain. It is a width
+  style of the variable Noto Sans Mono rather than a family of its own, and
+  resolved nowhere on macOS, Debian/Ubuntu or Alpine; the chain fell through
+  to `Noto Sans Mono`, which it now names directly.
+
 ## [1.8.1] - 2026-08
 
 ### Fixed
@@ -7,11 +35,13 @@
   Scripts outside Latin and CJK were left to the system's generic font
   fallback, which is not the same font on every machine: on Alpine the Arabic
   block was claimed by Noto Sans Math, which has the glyphs but no joining
-  rules, while on Debian the same text fell to DejaVu Sans. The family chains
-  now name the scripts the gallery covers — Arabic (`Noto Sans Arabic`, with
-  `Noto Naskh Arabic` for the serif style), Hebrew, Devanagari, Thai and
-  Khmer — so the same input renders the same way everywhere. Gallery example
-  067 (Arabic) was affected and has been regenerated.
+  rules, while on Debian/Ubuntu the same text fell to DejaVu Sans. The family
+  chains now name the scripts the gallery covers — Arabic (`Noto Sans Arabic`,
+  with `Noto Naskh Arabic` for the serif style), Hebrew, Devanagari, Thai and
+  Khmer — so machines that have those fonts installed produce the same shapes.
+  Mathematical alphanumerics (U+1D400–) are not named yet and still vary by
+  environment. Gallery example 067 (Arabic) was affected and has been
+  regenerated.
 - Emoji were measured with one font and drawn with another where a colour
   emoji font was installed: Pango selects `Noto Color Emoji`, but the
   librsvg/Cairo pipeline does not rasterise its bitmap glyphs, so the drawing

data/CITATION.cff

@@ -11,7 +11,7 @@ repository-code: "https://github.com/yohasebe/rsyntaxtree"
 # Concept DOI: always resolves to the latest archived version on Zenodo
 doi: "10.5281/zenodo.21916150"
 license: MIT
-version: 1.8.1
+version: 1.8.2
 date-released: 2026-08-15
 keywords:
   - linguistics

data/Dockerfile

@@ -10,7 +10,8 @@ RUN apk update && \
     apk add --no-cache -t .build-packages --no-cache build-base curl-dev wget gcompat && \
     apk add --no-cache font-noto font-noto-cjk font-noto-cjk-extra \
         font-noto-arabic font-noto-naskh-arabic font-noto-hebrew \
-        font-noto-devanagari font-noto-thai font-noto-khmer
+        font-noto-devanagari font-noto-thai font-noto-khmer \
+        font-dejavu
 
 # Noto Sans Math (pulled in by font-noto) claims the Arabic block but carries no
 # joining rules, and fontconfig ranks it above Noto Sans Arabic — Arabic then

data/README.md

@@ -112,20 +112,19 @@ See [RSyntaxTree Example Gallery](https://yohasebe.github.io/rsyntaxtree/example
 
 ## System Fonts
 
-RSyntaxTree resolves fonts by family name through fontconfig (measurement via Pango, rendering via librsvg), so the fonts must be installed on the system that generates the images. Recommended:
+RSyntaxTree resolves fonts by family name through fontconfig (measurement via Pango, rendering via librsvg), so the fonts must be installed on the machine that generates the images. As of v1.8.0 the gem no longer bundles font files; they were not used at runtime.
 
-- Debian/Ubuntu: `apt install fonts-noto-core fonts-noto-cjk fonts-noto-color-emoji`
-- Alpine: `apk add font-noto font-noto-cjk font-noto-cjk-extra font-noto-emoji` (the `-extra` package carries Noto Serif CJK)
-- macOS: install [Noto Sans/Serif](https://fonts.google.com/noto) (and Noto Sans/Serif JP for CJK); otherwise the system falls back to available fonts
+- Debian/Ubuntu: `apt install fonts-noto-core fonts-noto-cjk`
+- Alpine: `apk add font-noto font-noto-cjk font-noto-cjk-extra` (the `-extra` package carries Noto Serif CJK)
+- macOS: `brew install --cask font-noto-sans font-noto-serif font-noto-sans-mono font-noto-sans-jp font-noto-serif-jp font-noto-sans-cjk font-noto-serif-cjk font-noto-sans-mono-cjk-jp` (macOS itself ships Noto only for a few rare scripts)
 
-Any script supported by an installed font renders correctly — the family chains fall back from Noto Sans/Serif to the JP and CJK variants, so Japanese, Chinese (simplified and traditional) and Korean are covered by the Noto CJK package. As of v1.8.0 the gem no longer bundles font files (they were not used at runtime).
+Arabic, Hebrew, Devanagari, Thai and Khmer need their own packages, and emoji need the **monochrome** [Noto Emoji](https://fonts.google.com/noto/specimen/Noto+Emoji) — colour emoji fonts are not drawn by this pipeline:
 
-Scripts outside Latin and CJK need their own Noto package, or they render as tofu boxes. Since v1.8.0 the gem names these families explicitly (`Noto Sans Arabic`, `Noto Naskh Arabic`, `Noto Sans/Serif Hebrew`, `Devanagari`, `Thai`, `Khmer`) instead of leaving them to the system fallback, so install them to get the intended shapes:
-
-- Debian/Ubuntu: `apt install fonts-noto-core` already covers them; on minimal images add `fonts-noto` for the full set
+- Debian/Ubuntu: `fonts-noto-core` covers the scripts; on minimal images add `fonts-noto` for the full set
 - Alpine: `apk add font-noto-arabic font-noto-naskh-arabic font-noto-hebrew font-noto-devanagari font-noto-thai font-noto-khmer`
+- macOS: the same families are available as `font-noto-*` casks
 
-Emoji need the **monochrome** [Noto Emoji](https://fonts.google.com/noto/specimen/Noto+Emoji); colour emoji fonts (`NotoColorEmoji`, shipped by Alpine's `font-noto-emoji` and Debian's `fonts-noto-color-emoji`) are measured but not drawn by the librsvg/Cairo pipeline, so emoji then fall back to whatever outline font happens to cover them.
+See [Fonts](https://yohasebe.github.io/rsyntaxtree/documentation#install-fonts-for-svg) in the documentation for which family is used for what, how to substitute your own, and the macOS caveats.
 
 ## Installation
 

data/bin/rsyntaxtree

@@ -173,8 +173,8 @@ data = RSyntaxTree::FormatConverter.to_bracket(data) if data
 begin
   RSyntaxTree::RSGenerator.check_data(data)
 rescue RSTError => e
-  puts e
-  exit
+  warn e
+  exit 1
 end
 
 begin
@@ -188,10 +188,10 @@ begin
   output = rsg.send(draw_method)
   File.binwrite(filepath, output)
 rescue RSTError => e
-  puts e
-  exit
+  warn e
+  exit 1
 rescue StandardError => e
-  p e
-  puts "Error: something unexpected occurred"
-  exit
+  warn "Error: something unexpected occurred"
+  warn "#{e.class}: #{e.message}"
+  exit 1
 end

data/lib/rsyntaxtree/base_graph.rb

@@ -376,7 +376,8 @@ module RSyntaxTree
     # being compressed narrower than the level right below it (an inversion
     # that reads as a needle-thin top over flat lower branches); branch
     # angles then stay comparable between adjacent levels. 1.0 = a pair is
-    # never tighter than the widest pair among its own children.
+    # never tighter than the widest pair among its own children. Applied in
+    # the :ordered nest mode only; see tidy_compress.
     TIDY_LEVEL_BALANCE = 1.0
 
     # Base multiplier for the minimum clearance between adjacent subtrees.
@@ -554,16 +555,21 @@ module RSyntaxTree
             end
           end
 
-          # Level-balance floor (nest mode only): keep the pair at least a
+          # Level-balance floor (:ordered only): keep the pair at least a
           # fraction of the widest child-level spread found inside its own
           # two subtrees. Nesting can tuck an upper pair far narrower than
           # the level right below it — an inversion that reads as a
-          # needle-thin top over flat lower branches. Without nesting the
-          # leaf-span guard already prevents such inversions, and applying
-          # the floor there would only forfeit legitimate compression.
+          # needle-thin top over flat lower branches.
+          #
+          # :none does not need it, since the leaf-span guard already
+          # prevents such inversions. :free deliberately goes without: it is
+          # the end of the scale where density outranks even branch angles,
+          # and with the floor in place it compressed no further than
+          # :ordered on all but a handful of trees.
+          #
           # Compression is a negative delta (the right subtree moves left),
           # so the floor bounds delta from below.
-          if @tidy_nest != :none && delta.negative?
+          if @tidy_nest == :ordered && delta.negative?
             dist = (right_child.horizontal_indent + right_child.content_width / 2.0) -
                    (left_child.horizontal_indent + left_child.content_width / 2.0)
             below = [left_child, right_child].map { |c| child_spread(c) }.compact.max

data/lib/rsyntaxtree/utils.rb

@@ -29,18 +29,36 @@ SCRIPT_FAMILIES_SANS = ["Noto Sans Arabic", "Noto Sans Hebrew", "Noto Sans Devan
 # Arabic behind it for systems that ship one but not the other.
 SCRIPT_FAMILIES_SERIF = ["Noto Naskh Arabic", "Noto Sans Arabic", "Noto Serif Hebrew", "Noto Serif Devanagari", "Noto Serif Thai", "Noto Serif Khmer"].freeze
 
+# Monochrome emoji faces only. Colour emoji fonts carry their glyphs as bitmap
+# or COLR tables, which Pango happily measures but librsvg does not draw, so a
+# colour font in this list would give a figure whose emoji are blank or blobbed.
+#
+# This holds only where Pango resolves through fontconfig. On macOS it goes
+# through CoreText, which answers every emoji codepoint with Apple Color Emoji
+# whatever the chain asks for; emoji figures have to be generated elsewhere.
 EMOJI_FAMILIES = ["OpenMoji", "OpenMoji Color", "OpenMoji Black", "Noto Emoji"].freeze
 
+# Mathematical alphanumerics (U+1D400–) label heads such as the little v of vP.
+# Neither Noto Sans nor Noto Serif covers the block, so without these entries
+# the glyphs came from whatever the machine happened to offer — Noto Sans Math
+# on Alpine, DejaVu Serif on Debian/Ubuntu, STIX Two Math on macOS. Noto Sans
+# Math is the only Noto face that covers the block and it is a sans design, so
+# the serif style asks for a serif source first. Listed after the script
+# families: Noto Sans Math also claims the Arabic block without joining rules,
+# and must never be reached before Noto Sans Arabic.
+MATH_FAMILIES_SANS = ["Noto Sans Math"].freeze
+MATH_FAMILIES_SERIF = ["DejaVu Serif", "Noto Sans Math"].freeze
+
 FONT_FAMILIES = {
-  sans:  (["Noto Sans", "Noto Sans JP", "Noto Sans CJK JP"] + SCRIPT_FAMILIES_SANS + EMOJI_FAMILIES + ["sans-serif"]).freeze,
-  serif: (["Noto Serif", "Noto Serif JP", "Noto Serif CJK JP"] + SCRIPT_FAMILIES_SERIF + EMOJI_FAMILIES + ["serif"]).freeze,
+  sans:  (["Noto Sans", "Noto Sans JP", "Noto Sans CJK JP"] + SCRIPT_FAMILIES_SANS + MATH_FAMILIES_SANS + EMOJI_FAMILIES + ["sans-serif"]).freeze,
+  serif: (["Noto Serif", "Noto Serif JP", "Noto Serif CJK JP"] + SCRIPT_FAMILIES_SERIF + MATH_FAMILIES_SERIF + EMOJI_FAMILIES + ["serif"]).freeze,
   # Noto ships no monospaced faces for these scripts, so the mono style borrows
   # the proportional ones rather than dropping to the generic fallback.
-  mono:  (["Noto Sans Mono SemiCondensed", "Noto Sans Mono", "Noto Sans JP", "Noto Sans Mono CJK JP"] + SCRIPT_FAMILIES_SANS + EMOJI_FAMILIES + ["monospace"]).freeze,
+  mono:  (["Noto Sans Mono", "Noto Sans JP", "Noto Sans Mono CJK JP"] + SCRIPT_FAMILIES_SANS + MATH_FAMILIES_SANS + EMOJI_FAMILIES + ["monospace"]).freeze,
   # The cjk style puts a full-coverage CJK family first, for text that mixes
   # Han, Hangul and kana. Latin falls back to the same Noto faces the sans
   # style uses. (Before 1.8.0 this was WQY Zen Hei.)
-  cjk:   (["Noto Sans CJK JP", "Noto Sans", "Noto Sans JP"] + SCRIPT_FAMILIES_SANS + EMOJI_FAMILIES + ["sans-serif"]).freeze
+  cjk:   (["Noto Sans CJK JP", "Noto Sans", "Noto Sans JP"] + SCRIPT_FAMILIES_SANS + MATH_FAMILIES_SANS + EMOJI_FAMILIES + ["sans-serif"]).freeze
 }.freeze
 
 module FontFamily

data/lib/rsyntaxtree/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module RSyntaxTree
-  VERSION = "1.8.1"
+  VERSION = "1.8.2"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: rsyntaxtree
 version: !ruby/object:Gem::Version
-  version: 1.8.1
+  version: 1.8.2
 platform: ruby
 authors:
 - Yoichiro Hasebe