Add the breve mark and dedicated forms of that

xkcd-script/generator/pt6_derived_chars.py

@@ -152,6 +152,16 @@ def _make_cedilla(font, cp, base_name, gap=8):
     c.addReference('comma', _place_below(font, base_name, 'comma', gap))
 
 
+def _make_macron_below(font, cp, base_name, gap=15):
+    if cp in _SKIP_CPS:
+        return
+    c = font.createMappedChar(cp)
+    c.clear()
+    c.addReference(base_name)
+    c.width = font[base_name].width
+    c.addReference('_macron_below_mark', _place_below(font, base_name, '_macron_below_mark', gap))
+
+
 def _make_dot_below(font, cp, base_name, gap=15):
     if cp in _SKIP_CPS:
         return
@@ -254,6 +264,10 @@ def _accented(cp, base_name, mark_name, gap=20, x_adj=0):
 # Single dot below: same dot shape, used for Vietnamese dot-below characters.
 _dot_below_mark = make_mark(font, '_dot_below_mark', extract_top_contours(font, 0x00DC, 2)[:1])
 
+# Macron below: same stroke as the macron above, used for Semitic transliteration.
+_macron_below_mark = make_mark(font, '_macron_below_mark', extract_top_contours(font, 0x0112, 1))
+_macron_below_mark.changeWeight(20)
+
 
 # Macron: topmost stroke from Ē, with weight correction.
 _macron_mark = make_mark(font, '_macron_mark', extract_top_contours(font, 0x0112, 1))
@@ -272,6 +286,14 @@ def _accented(cp, base_name, mark_name, gap=20, x_adj=0):
 # Tilde.
 _tilde_mark = _make_weighted_mark(font, 'asciitilde', _mark_scale, 35, '_tilde_mark')
 
+# Breve: parenleft rotated 90° CCW — the arc gives the right bowl shape for ˘.
+_breve_mark = _make_weighted_mark(font, 'parenleft', _mark_scale, 35, '_breve_mark')
+_breve_mark.transform(psMat.rotate(math.radians(90)))
+_bb = _breve_mark.boundingBox()
+_breve_mark.transform(psMat.translate(-(_bb[0] + _bb[2]) / 2, 0))
+_breve_mark.transform(psMat.scale(0.5, 1))
+_breve_mark.transform(psMat.translate(-220, 0))
+
 # --- Marks composed from existing mark glyphs ---
 
 # dotlessi (U+0131): i without the dot, so í etc. don't stack dot + acute.
@@ -316,6 +338,7 @@ def _accented(cp, base_name, mark_name, gap=20, x_adj=0):
     (0x030A, _ring_mark),
     (0x030B, _double_acute_mark),
     (0x030C, _caron_mark),
+    (0x0306, _breve_mark),   # ̆  combining breve
 ]:
     c = font.createMappedChar(cp)
     c.clear()
@@ -324,6 +347,28 @@ def _accented(cp, base_name, mark_name, gap=20, x_adj=0):
     c.addReference(mark.glyphname, psMat.translate(0, dy))
     c.width = 0
 
+# Below combining marks share the macron-below shape at different vertical positions.
+_descender_bottom = font['p'].boundingBox()[1]
+_mb_bb = font['_macron_below_mark'].boundingBox()
+
+# U+0331 ◌̱  COMBINING MACRON BELOW — below the descender.
+_c0331 = font.createMappedChar(0x0331)
+_c0331.clear()
+_c0331.addReference('_macron_below_mark', psMat.translate(0, _descender_bottom - _combining_gap - _mb_bb[3]))
+_c0331.width = 0
+
+# U+0332 ◌̲  COMBINING LOW LINE — just below the baseline (underline position).
+_c0332 = font.createMappedChar(0x0332)
+_c0332.clear()
+_c0332.addReference('_macron_below_mark', psMat.translate(0, -_combining_gap - _mb_bb[3]))
+_c0332.width = 0
+
+# U+0320 ◌̠  COMBINING MINUS SIGN BELOW — halfway between baseline and descender.
+_c0320 = font.createMappedChar(0x0320)
+_c0320.clear()
+_c0320.addReference('_macron_below_mark', psMat.translate(0, _descender_bottom // 2 - _combining_gap - _mb_bb[3]))
+_c0320.width = 0
+
 
 # ---------------------------------------------------------------------------
 # Accented character tables
@@ -365,6 +410,17 @@ def _accented(cp, base_name, mark_name, gap=20, x_adj=0):
 for cp, base in [(0x00E5, 'a'), (0x016E, 'U'), (0x016F, 'u')]:
     _make_accented(font, cp, base, '_ring_above_mark')
 
+# Breve: Ă Ĕ Ğ Ĭ Ŏ Ŭ / ă ĕ ğ ĭ ŏ ŭ  (Romanian, Turkish, Belarusian, Esperanto, transliteration)
+for cp, base in [
+    (0x0102, 'A'), (0x0114, 'E'), (0x011E, 'G'), (0x012C, 'I'), (0x014E, 'O'), (0x016C, 'U'),
+]:
+    _make_accented(font, cp, base, '_breve_mark', gap=20)
+for cp, base in [
+    (0x0103, 'a'), (0x0115, 'e'), (0x011F, 'g'), (0x014F, 'o'), (0x016D, 'u'),
+]:
+    _make_accented(font, cp, base, '_breve_mark', gap=8)
+_make_accented(font, 0x012D, 'dotlessi', '_breve_mark', gap=40)  # ĭ — dotless to avoid dot+breve stack
+
 # Ď Ť (uppercase): caron above, like other uppercase caron letters.
 for cp, base in [(0x010E, 'D'), (0x0164, 'T')]:
     _make_accented(font, cp, base, '_caron_mark')
@@ -627,6 +683,14 @@ def _make_eng(font, cp, base_name, comma_name, x_frac=0.88, x_offset=0, y_offset
 ]:
     _make_cedilla(font, cp, base)
 
+# Comma below: Ș ș Ț ț  (Romanian — U+0219/U+021B are the canonical forms;
+# U+015F/U+0163 cedilla variants already exist above)
+for cp, base in [
+    (0x0218, 'S'), (0x0219, 's'),
+    (0x021A, 'T'), (0x021B, 't'),
+]:
+    _make_cedilla(font, cp, base)
+
 # Macron: Ā Ī Ō Ū / ā ī ō ū ē  (Ē skipped — hand-drawn)
 for cp, base in [
     (0x0100, 'A'), (0x012A, 'I'), (0x014C, 'O'), (0x016A, 'U'), (0x0112, 'E'),
@@ -649,6 +713,14 @@ def _make_eng(font, cp, base_name, comma_name, x_frac=0.88, x_offset=0, y_offset
 ]:
     _make_dot_below(font, cp, base)
 
+# Macron below: Ḏ Ḻ Ṉ Ṟ Ṯ Ẕ / ḏ ḻ ṉ ṟ ṯ ẕ  (Semitic transliteration)
+for cp, base in [
+    (0x1E0E, 'D'), (0x1E3A, 'L'), (0x1E48, 'N'), (0x1E5E, 'R'), (0x1E6E, 'T'), (0x1E94, 'Z'),
+    (0x1E0F, 'd'), (0x1E3B, 'l'), (0x1E49, 'n'), (0x1E6F, 't'), (0x1E95, 'z'),
+]:
+    _make_macron_below(font, cp, base)
+_make_macron_below(font, 0x1E5F, 'r', gap=45)  # ṟ — r sits low, needs extra gap
+
 # ij U+0133 / IJ U+0132: Dutch IJ digraph ligatures.
 # Position so the ink edges have the same gap as adjacent letters would after kerning (~40 units).
 for cp, left, right in [(0x0133, 'i', 'j'), (0x0132, 'I', 'J')]:

xkcd-script/generator/pt7_font_properties.py

@@ -17,7 +17,21 @@
 # ---------------------------------------------------------------------------
 
 def _base_char(c):
-    """Return the base letter for an accented character (e.g. É → E)."""
+    """Return the base letter for an accented character (e.g. É → E).
+
+    Characters without a Unicode canonical decomposition (e.g. ø, ł) are
+    handled via the manual table below.
+    """
+    _no_decomp = {
+        'ø': 'o', 'Ø': 'O',
+        'ł': 'l', 'Ł': 'L',
+        'đ': 'd', 'Đ': 'D',
+        'ħ': 'h', 'Ħ': 'H',
+        'ŧ': 't', 'Ŧ': 'T',
+        'ð': 'd', 'Ð': 'D',
+    }
+    if c in _no_decomp:
+        return _no_decomp[c]
     decomp = unicodedata.decomposition(c)
     if decomp and not decomp.startswith('<'):
         return chr(int(decomp.split()[0], 16))