refine code

Author: dummy-index

xkcd-script/generator/pt5_svg_to_font.py

@@ -18,6 +18,9 @@
 import parse
 import unicodedata
 
+SPACE = 0
+RSPACE = 20
+
 fnames = sorted(glob.glob('../generated/characters/char_*.svg'))
 
 characters = []
@@ -237,9 +240,8 @@ def pad_glyph(c):
     # Put horizontal padding around the glyph. I choose a number here that looks reasonable,
     # there are far more sophisticated means of doing this (like looking at the original image,
     # and calculating how much space there should be).
-    space = 0
-    rspace = 20
-    c.font['_pad_space'].width = space + rspace
+    space = SPACE
+    rspace = RSPACE
     bbox = c.boundingBox()
     if c.glyphname in list('gjpqy'):
         # Recalculate the bounding box by excluding the tail of the glyph
@@ -332,6 +334,7 @@ def charname(char):
 # Information to be conveyed to the next stage.
 # I wanted to use font.persistent, but it causes an error. Instead, I use a dummy glyph.
 font.createChar(-1, '_pad_space')
+font['_pad_space'].width = SPACE + RSPACE
 
 # Per-character size scaling applied after changeWeight, to fine-tune individual glyphs
 # that end up slightly too large despite correct stroke weight.

xkcd-script/generator/pt7_font_properties.py

@@ -114,7 +114,7 @@ def getkern(left, right):
         return None
 
     a = font['_pad_space'].width
-    a = max(a - 20, 0)
+    a = a - 20
 
     # The same combination will be overwritten, so the one written last will take effect.
     # autoKern looks at the outline, so even if you change the padding, it absorbs all of it.
@@ -124,6 +124,8 @@ def getkern(left, right):
     kern(60+a, ['s'], set(lower) - {'i', 'j', 'f', 't', 'x'}, onlyCloser=True, damper=0.75) # loosen by damper
     # Overwrite sf and st. (From experience, it is often just right to adopt the larger of the two
     # separation required by the glyphs on the left and right.)
+    # The horizontal bars of 'r' and 'f' get caught between the dot and the stem of 'i', causing
+    # unexpected behavior, so 'i' and 'j' are excluded.
     kern(80+a, set(lower) - {'i', 'j'}, ['f', 't'], onlyCloser=True, damper=0.75)
     kern(90+a, set(lower) - {'i', 'j'}, ['x'], onlyCloser=True, damper=0.75) # kx is fine, fx is tight
     kern(80+a, ['x'], set(lower) - {'i', 'j'}, onlyCloser=True, damper=0.75)
@@ -135,29 +137,29 @@ def getkern(left, right):
     kern(100+a + diff_ro_re, ['r'], ['e'], onlyCloser=True, damper=0.75)
     # including uppercase
     # Set *Y altogether first: CY, OY, etc. will have appropriate values set in the latter part.
-    kern(105, roman, ['Y', 'T'], onlyCloser=True, damper=0.75)
-    kern(100, caps, ['f'], onlyCloser=True, damper=0.75)
+    kern(105+a, roman, ['Y', 'T'], onlyCloser=True, damper=0.75)
+    kern(100+a, caps, ['f'], onlyCloser=True, damper=0.75)
     # F/E are separated from T/J so they can use a tighter target gap.
-    kern(110, ['F'], set(roman) - {'j'}, onlyCloser=True, damper=0.75) # keep FO≈-60
+    kern(110+a, ['F'], set(roman) - {'j'}, onlyCloser=True, damper=0.75) # keep FO≈-60
     # Since F and z mesh together and the kerning becomes too large,
     # reuse the kerning value of one of the round letterforms.
     diff_Fo_Fz = getkern('F', 'o') - getkern('F', 'z')
-    kern(110 + int(diff_Fo_Fz / 0.75), ['F'], ['z'], onlyCloser=True, damper=0.75)
-    kern(90, ['E'], set(roman) - {'j'}, onlyCloser=True, damper=0.75) # keep ES≈-30
-    kern(45, ['E'], ['V'], onlyCloser=True, touch=True)
-    kern(115, ['T', 'J'], set(roman) - {'j'}, onlyCloser=True, damper=0.75) # keep Tr≈-105
-    kern(105, ['Y'], set(roman) - {'j'}, onlyCloser=True, damper=0.75)
-    kern(85, ['V'], caps, onlyCloser=True, damper=0.75)
+    kern(110+a + int(diff_Fo_Fz / 0.75), ['F'], ['z'], onlyCloser=True, damper=0.75)
+    kern(90+a, ['E'], set(roman) - {'j'}, onlyCloser=True, damper=0.75) # keep ES≈-30
+    kern(45+a, ['E'], ['V'], onlyCloser=True, touch=True)
+    kern(115+a, ['T', 'J'], set(roman) - {'j'}, onlyCloser=True, damper=0.75) # keep Tr≈-105
+    kern(105+a, ['Y'], set(roman) - {'j'}, onlyCloser=True, damper=0.75)
+    kern(85+a, ['V'], caps, onlyCloser=True, damper=0.75)
     # C: loosen from the default (was too tight for Cj).
     # Compared to E, the lower curve of C tends to come close to the next character,
     # but this is considered an intentional design.
-    kern(60, ['C'], set(roman) - {'j'}, onlyCloser=True, damper=0.75) # keep CK≈-15
-    kern(25, ['C'], ['V'], onlyCloser=True, touch=True)
-    kern(60, ['O'], set(roman) - {'j'}, onlyCloser=True, damper=0.75) # loosen
-    kern(100, ['P'], set(roman) - {'j'}, onlyCloser=True, damper=0.75)
+    kern(60+a, ['C'], set(roman) - {'j'}, onlyCloser=True, damper=0.75) # keep CK≈-15
+    kern(25+a, ['C'], ['V'], onlyCloser=True, touch=True)
+    kern(60+a, ['O'], set(roman) - {'j'}, onlyCloser=True, damper=0.75) # loosen
+    kern(100+a, ['P'], set(roman) - {'j'}, onlyCloser=True, damper=0.75)
     diff_Po_Pe = getkern('P', 'o') - getkern('P', 'e')
-    kern(100 + int(diff_Po_Pe / 0.75), ['P'], ['e'], onlyCloser=True, damper=0.75)
-    kern(35, ['L'], set(roman) - {'j'}, onlyCloser=True, touch=True)
+    kern(100+a + int(diff_Po_Pe / 0.75), ['P'], ['e'], onlyCloser=True, damper=0.75)
+    kern(35+a, ['L'], set(roman) - {'j'}, onlyCloser=True, touch=True)
 
 
 autokern(font)