Add .notdef glyph, etc. / dedicated glyph for U+0075

xkcd-script/generator/pt4_additional_sources.py

@@ -154,6 +154,8 @@ def extract_symbol(arr, y0, y1, x0, x1, name, exclude=None):
 
 # Each entry is (output_name, source_filename_stem).
 EXTRAS = [
+    ('notdef', '1913_i_2x__notdef'),            # .notdef fallback glyph source
+    ('square', '2251_alignment_chart_2x__square'),  # □ U+25A1 source
     ('eszett', 'eszett'),                       # ß U+00DF / ẞ U+1E9E source
     ('lambda', '1145_sky_color_2x__lambda'),    # λ U+03BB source
     ('tau', '2520_symbols_2x__tau'),            # τ U+03C4 source

xkcd-script/generator/pt5_svg_to_font.py

@@ -317,6 +317,11 @@ def charname(char):
 c = font.createMappedChar(32)
 c.width = 256
 
+c = font.createChar(0x0000, '.null')   # U+0000: null, zero-width; required by OpenType
+c.width = 256
+c = font.createChar(0x000D, 'nonmarkingreturn')  # U+000D: carriage return, zero-width; required by OpenType
+c.width = 256
+
 
 # ---------------------------------------------------------------------------
 # Glyphs imported from xkcd comic images
@@ -391,6 +396,23 @@ def _import_comic_glyph(font, name, svg_path, target_top, weight_delta=0):
     return g
 
 
+# .notdef is shown for any codepoint the font doesn't cover; OpenType requires it.
+# Hand-drawn source: xkcd #1913 (I), a sketchy question-mark-in-a-box.
+_notdef_svg = os.path.join(_COMIC_CHARS_DIR, 'notdef.svg')
+_notdef_src = _import_comic_glyph(font, 'notdef', _notdef_svg, target_top=font.ascent, weight_delta=20)
+_bb = _notdef_src.boundingBox()
+_notdef_src.transform(psMat.translate(0, -_bb[1]))
+_bb = _notdef_src.boundingBox()
+if _bb[3] > 0:
+    _notdef_src.transform(psMat.scale(font.ascent / _bb[3]))
+    _notdef_src.width = int(round(_notdef_src.boundingBox()[2] + 20))
+c = font.createChar(-1, '.notdef')
+c.clear()
+for _cont in _notdef_src.foreground:
+    c.foreground += _cont
+c.width = _notdef_src.width
+
+
 # Greek letters vectorised by pt4 from xkcd comic images.
 # Each entry: (svg_name, unicode_cp, ref_char_for_height, baseline_snap)
 # baseline_snap=True  → translate so bb[1]=0 (letters that sit on the baseline).
@@ -592,6 +614,23 @@ def _import_comic_glyph(font, name, svg_path, target_top, weight_delta=0):
     _ch.width = _g.width
 
 
+# □ U+25A1 WHITE SQUARE — hand-drawn source from extras/square.png.
+_square_svg = os.path.join(_COMIC_CHARS_DIR, 'square.svg')
+_target_top_sq = font['H'].boundingBox()[3]
+_square_src = _import_comic_glyph(font, 'square', _square_svg, target_top=_target_top_sq, weight_delta=45)
+_bb = _square_src.boundingBox()
+_square_src.transform(psMat.translate(0, -_bb[1]))
+_bb = _square_src.boundingBox()
+if _bb[3] > 0:
+    _square_src.transform(psMat.scale(_target_top_sq / _bb[3]))
+    _square_src.width = int(round(_square_src.boundingBox()[2] + 20))
+_ch = font.createMappedChar(0x25A1)
+_ch.clear()
+for _cont in _square_src.foreground:
+    _ch.foreground += _cont
+_ch.width = _square_src.width
+
+
 # ---------------------------------------------------------------------------
 # Save
 # ---------------------------------------------------------------------------

xkcd-script/generator/pt6_derived_chars.py

@@ -212,11 +212,33 @@ def _accented(cp, base_name, mark_name, gap=20, x_adj=0):
 # Glyph aliases and re-uses
 # ---------------------------------------------------------------------------
 
+# U+20DE COMBINING ENCLOSING SQUARE — zero-width mark sized and positioned to
+# enclose '?' with a small margin.  GPOS would be needed for full generality.
+_sq = font[0x25A1]
+_sq_bb = _sq.boundingBox()
+_sq_cx = (_sq_bb[0] + _sq_bb[2]) / 2
+_sq_h = _sq_bb[3] - _sq_bb[1]   # sq_bb[1] == 0 after baseline snap
+_q_bb = font[ord('?')].boundingBox()
+_q_adv = font[ord('?')].width
+_margin = 20
+_scale_y = (_q_bb[3] - _q_bb[1] + 2 * _margin) / _sq_h
+_x_offset = -_q_adv / 2 - _sq_cx
+_y_offset = _q_bb[1] - _margin
+c = font.createMappedChar(0x20DE)
+c.addReference(_sq.glyphname, psMat.compose(
+    psMat.scale(1, _scale_y),
+    psMat.translate(_x_offset, _y_offset),
+))
+c.width = 0
+
 # Vertical pipe: re-use the I glyph (same stroke, same weight).
+c = font.createChar(-1, 'I.sansserif')
+c.addReference('I')
+c.width = font['I'].width
 pipe = font.createMappedChar(ord('|'))
 pipe.clear()
-pipe.addReference('I')
-pipe.width = font['I'].width
+pipe.addReference('I.sansserif', psMat.compose(psMat.scale(1, 1.3), psMat.translate(0, -0.2 * font.ascent)))
+pipe.width = font['I.sansserif'].width
 
 
 

xkcd-script/samples/gen_charmap.py

@@ -12,6 +12,7 @@
 """
 import os
 import re
+import sys
 import unicodedata
 
 import matplotlib
@@ -43,19 +44,57 @@
     (0x0180, 0x0250, "Latin Extended-B"),
     (0x0300, 0x0370, "Combining Diacritical Marks"),
     (0x0370, 0x0400, "Greek and Coptic"),
-    (0x2018, 0x2040, "General Punctuation"),
+    (0x1E00, 0x1F00, "Latin Extended Additional"),
+    (0x2000, 0x2070, "General Punctuation"),
     (0x2190, 0x2200, "Arrows"),
     (0x2200, 0x2300, "Mathematical Operators"),
 ]
 
 COLS = 16
 COMBINING_CATS = {'Mn', 'Mc', 'Me'}
 
-# Any font glyphs not covered by the defined blocks go in a catch-all block.
+# Explicit ordered list of codepoints that fall outside all named blocks above.
+# Order here controls layout in the "Non-Latin / Other" charmap table.
+# If the font contains a glyph outside every named block that is not listed
+# here, the script errors and tells you to add it.
+# Use None to reserve a slot (renders as a blank cell) so that removing a
+# character doesn't shift subsequent entries and cause a noisy table diff.
+EXTRAS_ORDER = [
+    0x025B,   # ɛ  LATIN SMALL LETTER OPEN E
+    0x1F382,  # 🎂  BIRTHDAY CAKE
+    0x20DE,   # ⃞  COMBINING ENCLOSING SQUARE
+    0x25A1,   # □  WHITE SQUARE
+]
+
 block_covered = set()
 for start, end, _ in BLOCKS:
     block_covered.update(range(start, end))
-extras = sorted(cp for cp in present if cp not in block_covered)
+
+# Control/format characters (Cc, Cf) have no visual representation; skip them.
+def _is_invisible(cp):
+    try:
+        return unicodedata.category(chr(cp)) in ('Cc', 'Cf')
+    except (ValueError, OverflowError):
+        return False
+
+extras_in_block = [cp for cp in EXTRAS_ORDER if cp is not None and cp in block_covered]
+if extras_in_block:
+    lines = [f"  U+{cp:04X}  {unicodedata.name(chr(cp), '(unknown)')}" for cp in extras_in_block]
+    raise ValueError(
+        "EXTRAS_ORDER contains codepoints already covered by a named block:\n"
+        + "\n".join(lines)
+    )
+
+extras_cps = set(cp for cp in EXTRAS_ORDER if cp is not None)
+uncovered = sorted(cp for cp in present if cp not in block_covered and cp not in extras_cps and not _is_invisible(cp))
+if uncovered:
+    lines = [f"  U+{cp:04X}  {unicodedata.name(chr(cp), '(unknown)')}" for cp in uncovered]
+    raise ValueError(
+        "Font contains codepoints not in any named block or EXTRAS_ORDER.\n"
+        "Add them to EXTRAS_ORDER in gen_charmap.py:\n" + "\n".join(lines)
+    )
+
+extras = list(EXTRAS_ORDER)
 if extras:
     BLOCKS = list(BLOCKS) + [(None, None, "Non-Latin / Other")]
 
@@ -134,14 +173,18 @@ def render_block(label, rows):
         y_bottom = y_top - CELL_H
         y_center = (y_top + y_bottom) / 2
 
-        hex_tag = f'U+{cps[0]:04X}' if cps else ''
+        first_cp = next((c for c in cps if c is not None), None)
+        hex_tag = f'U+{first_cp:04X}' if first_cp is not None else ''
         ax.text(LABEL_W - 0.05, y_center, hex_tag,
                 ha='right', va='center', fontproperties=fp_tiny, color='#555555')
 
         for col, cp in enumerate(cps):
             x_left   = LABEL_W + col * CELL_W
             x_center = x_left + CELL_W / 2
 
+            if cp is None:
+                continue
+
             try:
                 cat = unicodedata.category(chr(cp))
             except (ValueError, OverflowError):