Merge pull request #24 from dougfelt/master

change builder tooling to support emoji sequences

Makefile

@@ -12,7 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-
 EMOJI = NotoColorEmoji
 font: $(EMOJI).ttf
 
@@ -20,16 +19,32 @@ CFLAGS = -std=c99 -Wall -Wextra `pkg-config --cflags --libs cairo`
 LDFLAGS = `pkg-config --libs cairo`
 PNGQUANTDIR := third_party/pngquant
 PNGQUANT := $(PNGQUANTDIR)/pngquant
-PNGQUANTFLAGS = --speed 1 --skip-if-larger --ext '.png' --force
+PNGQUANTFLAGS = --speed 1 --skip-if-larger --force
 
-$(PNGQUANT):
+# zopflipng is better (about 10%) but much slower.  it will be used if
-	$(MAKE) -C $(PNGQUANTDIR)
+# present.  pass ZOPFLIPNG= as an arg to make to use optipng instead.
 
-waveflag: waveflag.c
+ZOPFLIPNG = zopflipng
-	$(CC) $< -o $@ $(CFLAGS) $(LDFLAGS)
+OPTIPNG = optipng
+
+EMOJI_BUILDER = third_party/color_emoji/emoji_builder.py
+ADD_GLYPHS = third_party/color_emoji/add_glyphs.py
+PUA_ADDER = map_pua_emoji.py
+VS_ADDER = add_vs_cmap.py # from nototools
+
+EMOJI_SRC_DIR := png/128
+FLAGS_SRC_DIR := third_party/region-flags/png
+
+BUILD_DIR := build
+EMOJI_DIR := $(BUILD_DIR)/emoji
+FLAGS_DIR := $(BUILD_DIR)/flags
+RESIZED_FLAGS_DIR := $(BUILD_DIR)/resized_flags
+RENAMED_FLAGS_DIR := $(BUILD_DIR)/renamed_flags
+QUANTIZED_DIR := $(BUILD_DIR)/quantized_pngs
+COMPRESSED_DIR := $(BUILD_DIR)/compressed_pngs
 
 LIMITED_FLAGS = CN DE ES FR GB IT JP KR RU US
-FLAGS = AD AE AF AG AI AL AM AO AR AS AT AU AW AX AZ \
+SELECTED_FLAGS = AD AE AF AG AI AL AM AO AR AS AT AU AW AX AZ \
 	BA BB BD BE BF BG BH BI BJ BM BN BO BR BS BT BW BY BZ \
 	CA CC CD CF CG CH CI CK CL CM CN CO CR CU CV CW CX CY CZ \
 	DE DJ DK DM DO DZ \
@@ -54,51 +69,127 @@ FLAGS = AD AE AF AG AI AL AM AO AR AS AT AU AW AX AZ \
 	WS \
 	YE \
 	ZA ZM ZW
+ALL_FLAGS = $(basename $(notdir $(wildcard $(FLAGS_SRC_DIR)/*.png)))
 
-FLAGS_SRC_DIR = third_party/region-flags/png
+FLAGS = $(SELECTED_FLAGS)
-FLAGS_DIR = ./flags
 
-GLYPH_NAMES := $(shell ./flag_glyph_name.py $(FLAGS))
+FLAG_NAMES = $(FLAGS:%=%.png)
-WAVED_FLAGS := $(foreach flag,$(FLAGS),$(FLAGS_DIR)/$(flag).png)
+FLAG_FILES = $(addprefix $(FLAGS_DIR)/, $(FLAG_NAMES))
-PNG128_FLAGS := $(foreach glyph_name,$(GLYPH_NAMES),$(addprefix ./png/128/emoji_$(glyph_name),.png))
+RESIZED_FLAG_FILES = $(addprefix $(RESIZED_FLAGS_DIR)/, $(FLAG_NAMES))
 
-$(FLAGS_DIR)/%.png: $(FLAGS_SRC_DIR)/%.png ./waveflag $(PNGQUANT)
+FLAG_GLYPH_NAMES = $(shell ./flag_glyph_name.py $(FLAGS))
-	mkdir -p $(FLAGS_DIR)
+RENAMED_FLAG_NAMES = $(FLAG_GLYPH_NAMES:%=emoji_%.png)
-	./waveflag "$<" "$@"
+RENAMED_FLAG_FILES = $(addprefix $(RENAMED_FLAGS_DIR)/, $(RENAMED_FLAG_NAMES))
-	optipng -quiet -o7 "$@"
-	$(PNGQUANT) $(PNGQUANTFLAGS) "$@"
 
-flag-symlinks: $(WAVED_FLAGS)
+EMOJI_NAMES = $(notdir $(wildcard $(EMOJI_SRC_DIR)/emoji_u*.png))
-	$(subst ^, ,                                \
+EMOJI_FILES= $(addprefix $(EMOJI_DIR)/,$(EMOJI_NAMES)))
-	  $(join                                    \
-	    $(FLAGS:%=ln^-fs^../../flags/%.png^),   \
-	    $(GLYPH_NAMES:%=./png/128/emoji_%.png;) \
-	   )                                        \
-	 )
 
-$(PNG128_FLAGS): flag-symlinks
+ALL_NAMES = $(EMOJI_NAMES) $(RENAMED_FLAG_NAMES)
 
-EMOJI_PNG128 = ./png/128/emoji_u
+ALL_QUANTIZED_FILES = $(addprefix $(QUANTIZED_DIR)/, $(ALL_NAMES))
+ALL_COMPRESSED_FILES = $(addprefix $(COMPRESSED_DIR)/, $(ALL_NAMES))
 
-EMOJI_BUILDER = third_party/color_emoji/emoji_builder.py
+# tool checks
-ADD_GLYPHS = third_party/color_emoji/add_glyphs.py
+ifeq (,$(shell which $(ZOPFLIPNG)))
-PUA_ADDER = map_pua_emoji.py
+  ifeq (,$(wildcard $(ZOPFLIPNG)))
-VS_ADDER = add_vs_cmap.py
+    MISSING_ZOPFLI = fail
-ifeq (, $(shell which $(VS_ADDER)))
+  endif
-  $(error "$(VS_ADDER) not in path, run setup.py in nototools")
 endif
 
-%.ttx: %.ttx.tmpl $(ADD_GLYPHS) $(UNI) $(PNG128_FLAGS)
+ifeq (,$(shell which $(OPTIPNG)))
-	python $(ADD_GLYPHS) "$<" "$@" "$(EMOJI_PNG128)"
+  ifeq (,$(wildcard $(OPTIPNG)))
+    MISSING_OPTIPNG = fail
+  endif
+endif
+
+ifeq (, $(shell which $(VS_ADDER)))
+  MISSING_ADDER = fail
+endif
+
+
+emoji: $(EMOJI_FILES)
+
+flags: $(FLAG_FILES)
+
+resized_flags: $(RESIZED_FLAG_FILES)
+
+renamed_flags: $(RENAMED_FLAG_FILES)
+
+quantized: $(ALL_QUANTIZED_FILES)
+
+compressed: $(ALL_COMPRESSED_FILES)
+
+check_compress_tool:
+ifdef MISSING_ZOPFLI
+  ifdef MISSING_OPTIPNG
+	$(error "neither $(ZOPFLIPNG) nor $(OPTIPNG) is available")
+  else
+	@echo "using $(OPTIPNG)"
+  endif
+else
+	@echo "using $(ZOPFLIPNG)"
+endif
+
+check_vs_adder:
+ifdef MISSING_ADDER
+	$(error "$(VS_ADDER) not in path, run setup.py in nototools")
+endif
+
+
+$(EMOJI_DIR) $(FLAGS_DIR) $(RESIZED_FLAGS_DIR) $(RENAMED_FLAGS_DIR) $(QUANTIZED_DIR) $(COMPRESSED_DIR):
+	mkdir -p "$@"
+
+$(PNGQUANT):
+	$(MAKE) -C $(PNGQUANTDIR)
+
+waveflag: waveflag.c
+	$(CC) $< -o $@ $(CFLAGS) $(LDFLAGS)
+
+$(EMOJI_DIR)/%.png: $(EMOJI_SRC_DIR)/%.png | $(EMOJI_DIR)
+	echo "emoji $< $@"
+	@convert -extent 136x128 -gravity center -background none "$<" "$@"
+
+$(FLAGS_DIR)/%.png: $(FLAGS_SRC_DIR)/%.png ./waveflag $(PNGQUANT) | $(FLAGS_DIR)
+	@./waveflag "$<" "$@"
+
+$(RESIZED_FLAGS_DIR)/%.png: $(FLAGS_DIR)/%.png | $(RESIZED_FLAGS_DIR)
+	@convert -extent 136x128 -gravity center -background none "$<" "$@"
+
+flag-symlinks: $(RESIZED_FLAG_FILES) | $(RENAMED_FLAGS_DIR)
+	@$(subst ^, ,                                  \
+	  $(join                                       \
+	    $(FLAGS:%=ln^-fs^../resized_flags/%.png^), \
+	    $(RENAMED_FLAG_FILES:%=%; )                \
+	   )                                           \
+	 )
+
+$(RENAMED_FLAG_FILES): flag-symlinks
+
+$(QUANTIZED_DIR)/%.png: $(RENAMED_FLAGS_DIR)/%.png $(PNGQUANT) | $(QUANTIZED_DIR)
+	$(PNGQUANT) $(PNGQUANTFLAGS) -o "$@" "$<"
+
+$(QUANTIZED_DIR)/%.png: $(EMOJI_DIR)/%.png $(PNGQUANT) | $(QUANTIZED_DIR)
+	$(PNGQUANT) $(PNGQUANTFLAGS) -o "$@" "$<"
+
+$(COMPRESSED_DIR)/%.png: $(QUANTIZED_DIR)/%.png | check_compress_tool $(COMPRESSED_DIR)
+ifdef MISSING_ZOPFLI
+	$(OPTIPNG) -quiet -o7 -clobber -force -out "$@" "$<"
+else
+	$(ZOPFLIPNG) -y "$<" "$@" 2> /dev/null
+endif
+
+
+%.ttx: %.ttx.tmpl $(ADD_GLYPHS) $(ALL_COMPRESSED_FILES)
+	@python $(ADD_GLYPHS) "$<" "$@" "$(COMPRESSED_DIR)/emoji_u"
 
 %.ttf: %.ttx
 	@rm -f "$@"
 	ttx "$<"
 
 $(EMOJI).ttf: $(EMOJI).tmpl.ttf $(EMOJI_BUILDER) $(PUA_ADDER) \
-  $(EMOJI_PNG128)*.png $(PNG128_FLAGS)
+	$(ALL_COMPRESSED_FILES) | check_vs_adder
-	python $(EMOJI_BUILDER) -V $< "$@" $(EMOJI_PNG128)
+	@python $(EMOJI_BUILDER) -V $< "$@" "$(COMPRESSED_DIR)/emoji_u"
-	python $(PUA_ADDER) "$@" "$@-with-pua"
+	@python $(PUA_ADDER) "$@" "$@-with-pua"
 	$(VS_ADDER) --dstdir '.' -o "$@-with-pua-varsel" "$@-with-pua"
 	mv "$@-with-pua-varsel" "$@"
 	rm "$@-with-pua"
@@ -106,5 +197,10 @@ $(EMOJI).ttf: $(EMOJI).tmpl.ttf $(EMOJI_BUILDER) $(PUA_ADDER) \
 clean:
 	rm -f $(EMOJI).ttf $(EMOJI).tmpl.ttf $(EMOJI).tmpl.ttx
 	rm -f waveflag
-	rm -rf $(FLAGS_DIR)
+	rm -rf $(BUILD_DIR)
-	rm -f `find -type l -name "*.png"`
+
+.SECONDARY: $(EMOJI_FILES) $(FLAG_FILES) $(RESIZED_FLAG_FILES) $(RENAMED_FLAG_FILES) \
+  $(ALL_QUANTIZED_FILES) $(ALL_COMPRESSED_FILES)
+
+.PHONY:	clean flags emoji renamed_flags quantized compressed check_compress_tool
+

NotoColorEmoji.tmpl.ttx.tmpl

@@ -4,12 +4,16 @@
   <GlyphOrder>
     <!-- The 'id' attribute is only for humans; it is ignored when parsed. -->
     <GlyphID id="0" name=".notdef"/>
+    <GlyphID id="1" name="null"/>
+    <GlyphID id="2" name="nonmarkingreturn"/>
+    <GlyphID id="3" name="space"/>
+    <GlyphID id="4" name="u200D"/>
   </GlyphOrder>
 
   <head>
     <!-- Most of this table will be recalculated by the compiler -->
     <tableVersion value="1.0"/>
-    <fontRevision value="1.22"/>
+    <fontRevision value="1.23"/>
     <checkSumAdjustment value="0x4d5a161a"/>
     <magicNumber value="0x5f0f3cf5"/>
     <flags value="00000000 00001011"/>
@@ -119,12 +123,19 @@
 
   <hmtx>
     <mtx name=".notdef" width="2550" lsb="0"/>
+    <mtx name="null" width="0" lsb="0"/>
+    <mtx name="nonmarkingreturn" width="2550" lsb="0"/>
+    <mtx name="space" width="2550" lsb="0"/>
+    <mtx name="u200D" width="0" lsb="0"/>
   </hmtx>
 
   <cmap>
     <tableVersion version="0"/>
     <cmap_format_12 platformID="3" platEncID="10" language="0" format="12" reserved="0" length="1" nGroups="1">
       <map code="0x0" name=".notdef"/><!-- &lt;control> -->
+      <map code="0xd" name="nonmarkingreturn"/>
+      <map code="0x20" name="space"/>
+      <map code="0x200d" name="u200D"/>
     </cmap_format_12>
   </cmap>
 
@@ -145,7 +156,7 @@
       Noto Color Emoji
     </namerecord>
     <namerecord nameID="5" platformID="3" platEncID="1" langID="0x409">
-      Version 1.22
+      Version 1.23
     </namerecord>
     <namerecord nameID="6" platformID="3" platEncID="1" langID="0x409">
       NotoColorEmoji

generate_emoji_placeholders.py

@@ -0,0 +1,95 @@
+import os
+from os import path
+import subprocess
+
+OUTPUT_DIR = '/tmp/placeholder_emoji'
+
+def generate_image(name, text):
+  print name, text.replace('\n', '_')
+  subprocess.check_call(
+      ['convert', '-size', '100x100', 'label:%s' % text,
+       '%s/%s' % (OUTPUT_DIR, name)])
+
+def is_color_patch(cp):
+  return cp >= 0x1f3fb and cp <= 0x1f3ff
+
+def has_color_patch(values):
+  for v in values:
+    if is_color_patch(v):
+      return True
+  return False
+
+def regional_to_ascii(cp):
+  return unichr(ord('A') + cp - 0x1f1e6)
+
+def is_flag_sequence(values):
+  if len(values) != 2:
+    return False
+  for v in values:
+    v -= 0x1f1e6
+    if v < 0 or v > 25:
+      return False
+  return True
+
+def is_keycap_sequence(values):
+  return len(values) == 2 and values[1] == 0x20e3
+
+def get_keycap_text(values):
+  return '-%c-' % unichr(values[0]) # convert gags on '['
+
+char_map = {
+    0x1f468: 'M',
+    0x1f469: 'W',
+    0x1f466: 'B',
+    0x1f467: 'G',
+    0x2764: 'H', # heavy black heart, no var sel
+    0x1f48b: 'K', # kiss mark
+    0x200D: '-', # zwj placeholder
+    0xfe0f: '-', # variation selector placeholder
+    0x1f441: 'I', # Eye
+    0x1f5e8: 'W', # 'witness' (left speech bubble)
+}
+
+def get_combining_text(values):
+  chars = []
+  for v in values:
+    char = char_map.get(v, None)
+    if not char:
+      return None
+    if char != '-':
+      chars.append(char)
+  return ''.join(chars)
+
+
+if not path.isdir(OUTPUT_DIR):
+  os.makedirs(OUTPUT_DIR)
+
+with open('sequences.txt', 'r') as f:
+  for seq in f:
+    seq = seq.strip()
+    text = None
+    values = [int(code, 16) for code in seq.split('_')]
+    if len(values) == 1:
+      val = values[0]
+      text = '%04X' % val # ensure upper case format
+    elif is_flag_sequence(values):
+      text = ''.join(regional_to_ascii(cp) for cp in values)
+    elif has_color_patch(values):
+      print 'skipping color patch sequence %s' % seq
+    elif is_keycap_sequence(values):
+      text = get_keycap_text(values)
+    else:
+      text = get_combining_text(values)
+      if not text:
+        print 'missing %s' % seq
+
+    if text:
+      if len(text) > 3:
+        if len(text) == 4:
+          hi = text[:2]
+          lo = text[2:]
+        else:
+          hi = text[:-3]
+          lo = text[-3:]
+        text = '%s\n%s' % (hi, lo)
+      generate_image('emoji_u%s.png' % seq, text)

third_party/color_emoji/add_glyphs.py

@@ -1,6 +1,6 @@
 #!/usr/bin/env python
 
-import glob, os, sys
+import collections, glob, os, sys
 from fontTools import ttx
 from fontTools.ttLib.tables import otTables
 from png import PNG
@@ -10,11 +10,34 @@ sys.path.append(
 import add_emoji_gsub
 
 
+def is_vs(cp):
+        return cp >= 0xfe00 and cp <= 0xfe0f
+
+def codes_to_string(codes):
+	if "_" in codes:
+		pieces = codes.split ("_")
+		string = "".join ([unichr (int (code, 16)) for code in pieces])
+	else:
+          try:
+		string = unichr (int (codes, 16))
+          except:
+            raise ValueError("uh-oh, no unichr for '%s'" % codes)
+        return string
+
+
+def glyph_sequence(string):
+        # sequence of names of glyphs that form a ligature
+        # variation selectors are stripped
+        return ["u%04X" % ord(char) for char in string if not is_vs(ord(char))]
+
+
 def glyph_name(string):
+        # name of a ligature
+        # includes variation selectors when present
 	return "_".join (["u%04X" % ord (char) for char in string])
 
 
-def add_ligature (font, string):
+def add_ligature (font, seq, name):
 	if 'GSUB' not in font:
 		ligature_subst = otTables.LigatureSubst()
 		ligature_subst.ligatures = {}
@@ -34,17 +57,27 @@ def add_ligature (font, string):
 	ligatures = lookup.SubTable[0].ligatures
 
 	lig = otTables.Ligature()
-	lig.CompCount = len(string)
+	lig.CompCount = len(seq)
-	lig.Component = [glyph_name(ch) for ch in string[1:]]
+	lig.Component = seq[1:]
-	lig.LigGlyph = glyph_name(string)
+	lig.LigGlyph = name
 
-	first = glyph_name(string[0])
+	first = seq[0]
 	try:
 		ligatures[first].append(lig)
 	except KeyError:
 		ligatures[first] = [lig]
 
 
+# Ligating sequences for emoji that already have a defined codepoint,
+# to match the sequences for the related emoji with no codepoint.
+# The key is the name of the glyph with the codepoint, the value is the
+# name of the sequence in filename form.
+EXTRA_SEQUENCES = {
+    'u1F46A': '1F468_200D_1F469_200D_1F466', # MWB
+    'u1F491': '1F469_200D_2764_FE0F_200D_1F468', # WHM
+    'u1F48F': '1F469_200D_2764_FE0F_200D_1F48B_200D_1F468', # WHKM
+}
+
 if len (sys.argv) < 4:
 	print >>sys.stderr, """
 Usage:
@@ -65,23 +98,22 @@ table and the first GSUB lookup (if existing) are modified.
 
 in_file = sys.argv[1]
 out_file = sys.argv[2]
-img_prefix = sys.argv[3]
+img_prefixen = sys.argv[3:]
 del sys.argv
 
 font = ttx.TTFont()
 font.importXML (in_file)
 
 img_files = {}
-glb = "%s*.png" % img_prefix
+for img_prefix in img_prefixen:
-print "Looking for images matching '%s'." % glb
+        glb = "%s*.png" % img_prefix
-for img_file in glob.glob (glb):
+        print "Looking for images matching '%s'." % glb
-	codes = img_file[len (img_prefix):-4]
+        for img_file in glob.glob (glb):
-	if "_" in codes:
+        	codes = img_file[len (img_prefix):-4]
-		pieces = codes.split ("_")
+                u = codes_to_string(codes)
-		u = "".join ([unichr (int (code, 16)) for code in pieces])
+                if u in img_files:
-	else:
+                        print 'overwriting %s with %s' % (img_files[u], imag_file)
-		u = unichr (int (codes, 16))
+        	img_files[u] = img_file
-	img_files[u] = img_file
 if not img_files:
 	raise Exception ("No image files found in '%s'." % glb)
 
@@ -98,20 +130,72 @@ h = font['hmtx'].metrics
 img_pairs = img_files.items ()
 img_pairs.sort (key=lambda pair: (len (pair[0]), pair[0]))
 
+glyph_names = set()
+ligatures = {}
+
+def add_lig_sequence(ligatures, seq, n):
+        # Assume sequences with ZWJ are emoji 'ligatures' and rtl order
+        # is also valid.  Internal permutations, though, no.
+        # We associate a sequence with a filename.  We can overwrite the
+        # sequence with a different filename later.
+        tseq = tuple(seq)
+        if tseq in ligatures:
+                print 'lig sequence %s, replace %s with %s' % (
+                    tseq, ligatures[tseq], n)
+        ligatures[tseq] = n
+        if 'u200D' in seq:
+                rev_seq = seq[:]
+                rev_seq.reverse()
+                trseq = tuple(rev_seq)
+                # if trseq in ligatures:
+                #        print 'rev lig sequence %s, replace %s with %s' % (
+                #        trseq, ligatures[trseq], n)
+                ligatures[trseq] = n
+
+
 for (u, filename) in img_pairs:
-	print "Adding glyph for U+%s" % ",".join (["%04X" % ord (char) for char in u])
+	# print "Adding glyph for U+%s" % ",".join (["%04X" % ord (char) for char in u])
 	n = glyph_name (u)
+        glyph_names.add(n)
+
 	g.append (n)
 	for char in u:
-		if char not in c:
+                cp = ord(char)
+		if cp not in c and not is_vs(cp):
 			name = glyph_name (char)
-			c[ord (char)] = name
+			c[cp] = name
 			if len (u) > 1:
 				h[name] = [0, 0]
 	(img_width, img_height) = PNG (filename).get_size ()
 	advance = int (round ((float (ascent+descent) * img_width / img_height)))
 	h[n] = [advance, 0]
 	if len (u) > 1:
-		add_ligature (font, u)
+                seq = glyph_sequence(u)
+                add_lig_sequence(ligatures, seq, n)
+
+for n in EXTRA_SEQUENCES:
+        if n in glyph_names:
+                seq = glyph_sequence(codes_to_string(EXTRA_SEQUENCES[n]))
+                add_lig_sequence(ligatures, seq, n)
+        else:
+                print 'extras: no glyph for %s' % n
+
+
+keyed_ligatures = collections.defaultdict(list)
+for k, v in ligatures.iteritems():
+        first = k[0]
+        keyed_ligatures[first].append((k, v))
+
+for base in sorted(keyed_ligatures):
+        pairs = keyed_ligatures[base]
+        # print 'base %s has %d sequences' % (base, len(pairs))
+
+        # Sort longest first, this ensures longer sequences with common prefixes
+        # are handled before shorter ones.  It would be better to have multiple
+        # lookups, most likely.
+        pairs.sort(key = lambda pair: (len(pair[0]), pair[0]), reverse=True)
+        for seq, name in pairs:
+                # print seq, name
+                add_ligature(font, seq, name)
 
 font.saveXML (out_file)

third_party/color_emoji/emoji_builder.py

@@ -20,7 +20,8 @@
 
 import sys, struct, StringIO
 from png import PNG
-
+import os
+from os import path
 
 def get_glyph_name_from_gsub (string, font, cmap_dict):
 	ligatures = font['GSUB'].table.LookupList.Lookup[0].SubTable[0].ligatures
@@ -83,6 +84,7 @@ class CBDT:
 		write_func = self.image_write_func (image_format)
 		for glyph in glyphs:
 			img_file = glyph_filenames[glyph]
+                        # print 'writing data for glyph %s' % path.basename(img_file)
 			offset = self.tell ()
 			write_func (PNG (img_file))
 			self.glyph_maps.append (GlyphMap (glyph, offset, image_format))
@@ -107,7 +109,11 @@ class CBDT:
 		line_height = (ascent + descent) * y_ppem / float (upem)
 		line_ascent = ascent * y_ppem / float (upem)
 		y_bearing = int (round (line_ascent - .5 * (line_height - height)))
+                # fudge y_bearing if calculations are a bit off
+                if y_bearing == 128:
+                  y_bearing = 127
 		advance = width
+                # print "small glyph metrics h: %d w: %d" % (height, width)
 		# smallGlyphMetrics
 		# Type	Name
 		# BYTE	height
@@ -115,10 +121,14 @@ class CBDT:
 		# CHAR	BearingX
 		# CHAR	BearingY
 		# BYTE	Advance
-		self.write (struct.pack ("BBbbB",
+                try:
+                        self.write (struct.pack ("BBbbB",
 					 height, width,
 					 x_bearing, y_bearing,
 					 advance))
+                except Exception as e:
+                  raise ValueError("%s, h: %d w: %d x: %d y: %d %d a:" % (
+                      e, height, width, x_bearing, y_bearing, advance))
 
 	def write_format1 (self, png):
 
@@ -437,8 +447,10 @@ By default they are dropped.
 	eblc.write_header ()
 	eblc.start_strikes (len (img_prefixes))
 
-	for img_prefix in img_prefixes:
+        def is_vs(cp):
+                return cp >= 0xfe00 and cp <= 0xfe0f
 
+	for img_prefix in img_prefixes:
 		print
 
 		img_files = {}
@@ -448,9 +460,14 @@ By default they are dropped.
 			codes = img_file[len (img_prefix):-4]
 			if "_" in codes:
 				pieces = codes.split ("_")
-				uchars = "".join ([unichr (int (code, 16)) for code in pieces])
+                                cps = [int(code, 16) for code in pieces]
+				uchars = "".join ([unichr(cp) for cp in cps if not is_vs(cp)])
 			else:
-				uchars = unichr (int (codes, 16))
+                                cp = int(codes, 16)
+                                if is_vs(cp):
+                                        print "ignoring unexpected vs input %04x" % cp
+                                        continue
+				uchars = unichr(cp)
 			img_files[uchars] = img_file
 		if not img_files:
 			raise Exception ("No image files found in '%s'." % glb)
@@ -460,14 +477,19 @@ By default they are dropped.
 		advance = width = height = 0
 		for uchars, img_file in img_files.items ():
 			if len (uchars) == 1:
-				glyph_name = unicode_cmap.cmap[ord (uchars)]
+                                try:
+                                        glyph_name = unicode_cmap.cmap[ord (uchars)]
+                                except:
+                                        print "no cmap entry for %x" % ord(uchars)
+                                        raise ValueError("%x" % ord(uchars))
 			else:
 				glyph_name = get_glyph_name_from_gsub (uchars, font, unicode_cmap.cmap)
 			glyph_id = font.getGlyphID (glyph_name)
 			glyph_imgs[glyph_id] = img_file
 			if "verbose" in options:
 				uchars_name = ",".join (["%04X" % ord (char) for char in uchars])
-				print "Matched U+%s: id=%d name=%s image=%s" % (uchars_name, glyph_id, glyph_name, img_file)
+				# print "Matched U+%s: id=%d name=%s image=%s" % (
+                                #    uchars_name, glyph_id, glyph_name, img_file)
 
 			advance += glyph_metrics[glyph_name][0]
 			w, h = PNG (img_file).get_size ()
@@ -476,7 +498,7 @@ By default they are dropped.
 
 		glyphs = sorted (glyph_imgs.keys ())
 		if not glyphs:
-			raise Exception ("No common characteres found between font and '%s'." % glb)
+			raise Exception ("No common characters found between font and '%s'." % glb)
 		print "Embedding images for %d glyphs for this strike." % len (glyphs)
 
 		advance, width, height = (div (x, len (glyphs)) for x in (advance, width, height))