Add targets and scripts for web fonts.

4 files changed, 382 insertions, 1 deletions

diff --git a/scripts/force_yminmax.py b/scripts/force_yminmax.py
index 454414c..36420b1 100755
--- a/scripts/force_yminmax.py
+++ b/scripts/force_yminmax.py
@@ -1,5 +1,5 @@
 #!/usr/bin/python
-"""Post-subset changes for Roboto for Android."""
+"""Post-subset changes for Roboto."""
 
 import sys
 
diff --git a/scripts/run_web_tests.py b/scripts/run_web_tests.py
new file mode 100755
index 0000000..779abdc
--- /dev/null
+++ b/scripts/run_web_tests.py
@@ -0,0 +1,143 @@
+#!/usr/bin/python
+"""Test assumptions that Android relies on."""
+
+import glob
+import unittest
+
+from fontTools import ttLib
+from nototools import coverage
+from nototools import font_data
+
+
+def load_fonts():
+    """Load all web fonts."""
+    all_font_files = glob.glob('out/web/*.ttf')
+    all_fonts = [ttLib.TTFont(font) for font in all_font_files]
+    assert len(all_font_files) == 12
+    return all_font_files, all_fonts
+
+
+class TestVerticalMetrics(unittest.TestCase):
+    """Test the vertical metrics of fonts."""
+
+    def setUp(self):
+        _, self.fonts = load_fonts()
+
+    def test_ymin_ymax(self):
+        """Tests yMin and yMax to be equal to the old values."""
+        for font in self.fonts:
+            head_table = font['head']
+            self.assertEqual(head_table.yMin, -555)
+            self.assertEqual(head_table.yMax, 2163)
+
+    def test_other_metrics(self):
+        """Tests other vertical metrics to be equal to the old values."""
+        for font in self.fonts:
+            hhea_table = font['hhea']
+            self.assertEqual(hhea_table.descent, -500)
+            self.assertEqual(hhea_table.ascent, 1900)
+
+            os2_table = font['OS/2']
+            self.assertEqual(os2_table.sTypoDescender, -512)
+            self.assertEqual(os2_table.sTypoAscender, 1536)
+            self.assertEqual(os2_table.sTypoLineGap, 102)
+            self.assertEqual(os2_table.usWinDescent, 512)
+            self.assertEqual(os2_table.usWinAscent, 1946)
+
+
+class TestCharacterCoverage(unittest.TestCase):
+    """Tests character coverage."""
+
+    def setUp(self):
+        _, self.fonts = load_fonts()
+        self.LEGACY_PUA = frozenset({0xEE01, 0xEE02, 0xF6C3})
+
+    def test_inclusion_of_legacy_pua(self):
+        """Tests that legacy PUA characters remain in the fonts."""
+        for font in self.fonts:
+            charset = coverage.character_set(font)
+            for char in self.LEGACY_PUA:
+                self.assertIn(char, charset)
+
+    def test_non_inclusion_of_other_pua(self):
+        """Tests that there are not other PUA characters except legacy ones."""
+        for font in self.fonts:
+            charset = coverage.character_set(font)
+            pua_chars = {
+                char for char in charset
+                if 0xE000 <= char <= 0xF8FF or 0xF0000 <= char <= 0x10FFFF}
+            self.assertTrue(pua_chars <= self.LEGACY_PUA)
+
+    def test_lack_of_unassigned_chars(self):
+        """Tests that unassigned characters are not in the fonts."""
+        for font in self.fonts:
+            charset = coverage.character_set(font)
+            self.assertNotIn(0x2072, charset)
+            self.assertNotIn(0x2073, charset)
+            self.assertNotIn(0x208F, charset)
+
+
+class TestItalicAngle(unittest.TestCase):
+    """Test the italic angle of fonts."""
+
+    def setUp(self):
+        _, self.fonts = load_fonts()
+
+    def test_italic_angle(self):
+        """Tests the italic angle of fonts to be correct."""
+        for font in self.fonts:
+            post_table = font['post']
+            if 'Italic' in font_data.font_name(font):
+                expected_angle = -12.0
+            else:
+                expected_angle = 0.0
+            self.assertEqual(post_table.italicAngle, expected_angle)
+
+
+class TestDigitWidths(unittest.TestCase):
+    """Tests the width of digits."""
+
+    def setUp(self):
+        _, self.fonts = load_fonts()
+        self.digits = [
+            'zero', 'one', 'two', 'three', 'four',
+            'five', 'six', 'seven', 'eight', 'nine']
+
+    def test_digit_widths(self):
+        """Tests all decimal digits to make sure they have the same width."""
+        for font in self.fonts:
+            hmtx_table = font['hmtx']
+            widths = [hmtx_table[digit][0] for digit in self.digits]
+            self.assertEqual(len(set(widths)), 1)
+
+
+class TestHints(unittest.TestCase):
+    """Tests hints."""
+
+    def setUp(self):
+        _, self.fonts = load_fonts()
+
+        # FIXME: remove as soon as issue 100 is fixed
+        bad_glyphs = ['uniFB01', 'uniFB02', 'uniFB03',
+                      'uniFB04', 'uniFFFC', 'uni048C']
+        self.known_missing_hints = [
+            (g, 'Roboto Light Italic') for g in bad_glyphs]
+
+    def test_digit_widths(self):
+        """Tests all glyphs and makes sure non-composite ones have hints."""
+        missing_hints = []
+        for font in self.fonts:
+            glyf_table = font['glyf']
+            for glyph_name in font.getGlyphOrder():
+                glyph = glyf_table[glyph_name]
+                if glyph.numberOfContours <= 0:  # composite or empty glyph
+                    continue
+                if len(glyph.program.bytecode) <= 0:
+                    missing_hints.append(
+                        (glyph_name, font_data.font_name(font)))
+
+        self.assertTrue(missing_hints <= self.known_missing_hints)
+
+
+if __name__ == '__main__':
+    unittest.main()
diff --git a/scripts/subset_for_web.py b/scripts/subset_for_web.py
new file mode 100644
index 0000000..e100dfc
--- /dev/null
+++ b/scripts/subset_for_web.py
@@ -0,0 +1,53 @@
+#!/usr/bin/python
+#
+# Copyright 2014 Google Inc. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Subset for web fonts."""
+
+__author__ = 'roozbeh@google.com (Roozbeh Pournader)'
+
+import sys
+
+from nototools import subset
+
+
+def read_charlist(filename):
+    """Returns a list of characters read from a charset text file."""
+    with open(filename) as datafile:
+        charlist = []
+        for line in datafile:
+            if '#' in line:
+                line = line[:line.index('#')]
+            line = line.strip()
+            if not line:
+                continue
+            if line.startswith('U+'):
+                line = line[2:]
+            char = int(line, 16)
+            charlist.append(char)
+        return charlist
+
+
+def main(argv):
+    """Subset the first argument to second, dropping unused parts of the font.
+    """
+    charlist = read_charlist('res/web_charset.txt')
+    # Add private use characters for legacy reasons
+    charlist += [0xEE01, 0xEE02, 0xF6C3]
+    subset.subset_font(argv[1], argv[2], include=charlist)
+
+
+if __name__ == '__main__':
+    main(sys.argv)
diff --git a/scripts/touchup_for_web.py b/scripts/touchup_for_web.py
new file mode 100755
index 0000000..0f30728
--- /dev/null
+++ b/scripts/touchup_for_web.py
@@ -0,0 +1,185 @@
+#!/usr/bin/python
+"""Post-build web fonts changes for Roboto."""
+
+import collections
+import os
+from os import path
+import sys
+
+from fontTools import ttLib
+from nototools import font_data
+from nototools import unicode_data
+
+
+def drop_lookup(table, lookup_number):
+    """Drop a lookup from an OpenType table by number.
+
+    Actually remove pointers from features to the lookup, which should be less
+    intrusive.
+    """
+    for feature in table.table.FeatureList.FeatureRecord:
+        if lookup_number in feature.Feature.LookupListIndex:
+            feature.Feature.LookupListIndex.remove(lookup_number)
+            feature.Feature.LookupCount -= 1
+
+
+def get_font_name(font):
+    """Gets the name of the font from the name table."""
+    return font_data.get_name_records(font)[4]
+
+
+DIGITS = ['zero', 'one', 'two', 'three', 'four',
+          'five', 'six', 'seven', 'eight', 'nine']
+
+def fix_digit_widths(font):
+    """Change all digit widths in the font to be the same."""
+    hmtx_table = font['hmtx']
+    widths = [hmtx_table[digit][0] for digit in DIGITS]
+    if len(set(widths)) > 1:
+        width_counter = collections.Counter(widths)
+        most_common_width = width_counter.most_common(1)[0][0]
+        print 'Digit widths were %s.' % repr(widths)
+        print 'Setting all glyph widths to %d.' % most_common_width
+        for digit in DIGITS:
+            assert abs(hmtx_table[digit][0] - most_common_width) <= 1
+            hmtx_table[digit][0] = most_common_width
+
+
+_MAP_SPACING_TO_COMBINING = {
+    'acute': 'acutecomb',
+    'breve': 'brevenosp',
+    'caron': 'uni030C',
+    'cedilla': 'cedillanosp',
+    'circumflex': 'circumflexnosp',
+    'dieresis': 'dieresisnosp',
+    'dotaccent': 'dotnosp',
+    'grave': 'gravecomb',
+    'hungarumlaut': 'acutedblnosp',
+    'macron': 'macroncomb',
+    'ogonek': 'ogoneknosp',
+    'tilde': 'tildecomb',
+    'ring': 'ringnosp',
+    'tonos': 'acutecomb',
+    'uni02F3': 'ringsubnosp',
+}
+
+def fix_ccmp_lookup(font):
+    """Fixes the broken ccmp lookup."""
+    cmap = font_data.get_cmap(font)
+    reverse_cmap = {name: code for (code, name) in cmap.items()}
+
+    # Where we know the bad 'ccmp' is
+    ccmp_lookup = font['GSUB'].table.LookupList.Lookup[2]
+    assert ccmp_lookup.LookupType == 4
+    assert ccmp_lookup.SubTableCount == 1
+    ligatures = ccmp_lookup.SubTable[0].ligatures
+    for first_char, ligtable in ligatures.iteritems():
+        ligatures_to_delete = []
+        for index, ligature in enumerate(ligtable):
+            assert len(ligature.Component) == 1
+            component = ligature.Component[0]
+            if (component.endswith('comb')
+                or component in ['commaaccent',
+                                 'commaaccentrotate',
+                                 'ringacute']):
+                continue
+            # https://code.google.com/a/google.com/p/roboto/issues/detail?id=54
+            if first_char == 'a' and component == 'uni02BE':
+                ligatures_to_delete.append(index)
+                continue
+            char = reverse_cmap[component]
+            general_category = unicode_data.category(char)
+            if general_category != 'Mn': # not a combining mark
+                ligature.Component[0] = _MAP_SPACING_TO_COMBINING[component]
+        ligatures[first_char] = [
+            ligature for (index, ligature) in enumerate(ligtable)
+            if index not in ligatures_to_delete]
+
+
+def apply_temporary_fixes(font):
+    """Apply some temporary fixes.
+    """
+    # Make sure macStyle is correct
+    # https://code.google.com/a/google.com/p/roboto/issues/detail?id=8
+    font_name = get_font_name(font)
+    bold = ('Bold' in font_name) or ('Black' in font_name)
+    italic = 'Italic' in font_name
+    font['head'].macStyle = (italic << 1) | bold
+
+    # Mark the font free for installation, embedding, etc.
+    # https://code.google.com/a/google.com/p/roboto/issues/detail?id=29
+    os2 = font['OS/2']
+    os2.fsType = 0
+
+    # Set the font vendor to Google
+    # https://code.google.com/a/google.com/p/roboto/issues/detail?id=46
+    os2.achVendID = 'GOOG'
+
+    # Drop the lookup forming the ff ligature
+    # https://code.google.com/a/google.com/p/roboto/issues/detail?id=47
+    drop_lookup(font['GSUB'], 5)
+
+    # Correct the ccmp lookup to use combining marks instead of spacing ones
+    # https://code.google.com/a/google.com/p/roboto/issues/detail?id=48
+    fix_ccmp_lookup(font)
+
+    # Fix the digit widths
+    # https://code.google.com/a/google.com/p/roboto/issues/detail?id=49
+    fix_digit_widths(font)
+
+    # Add cmap for U+2117 SOUND RECORDING COPYRIGHT
+    # https://code.google.com/a/google.com/p/roboto/issues/detail?id=44
+    font_data.add_to_cmap(font, {0x2117: 'published'})
+
+    # Fix version number from buildnumber.txt
+    # https://code.google.com/a/google.com/p/roboto/issues/detail?id=50
+    from datetime import date
+
+    build_number_txt = path.join(
+        path.dirname(__file__), os.pardir, 'res', 'buildnumber.txt')
+    build_number = open(build_number_txt).read().strip()
+
+    version_record = 'Version 2.0%s; %d' % (build_number, date.today().year)
+
+    for record in font['name'].names:
+        if record.nameID == 5:
+            if record.platformID == 1 and record.platEncID == 0:  # MacRoman
+                record.string = version_record
+            elif record.platformID == 3 and record.platEncID == 1:
+                # Windows UCS-2
+                record.string = version_record.encode('UTF-16BE')
+            else:
+                assert False
+
+
+def apply_web_specific_fixes(font):
+    """Apply fixes needed for web fonts."""
+    # Set ascent, descent, and lineGap values to Android K values
+    hhea = font['hhea']
+    hhea.ascent = 1900
+    hhea.descent = -500
+    hhea.lineGap = 0
+
+    os2 = font['OS/2']
+    os2.sTypoAscender = 1536
+    os2.sTypoDescender = -512
+    os2.sTypoLineGap = 102
+    os2.usWinAscent = 1946
+    os2.usWinDescent = 512
+
+
+def correct_font(source_font_name, target_font_name):
+    """Corrects metrics and other meta information."""
+    font = ttLib.TTFont(source_font_name)
+#    apply_temporary_fixes(font)
+    apply_web_specific_fixes(font)
+    font.save(target_font_name)
+
+
+def main(argv):
+    """Correct the font specified in the command line."""
+    correct_font(argv[1], argv[2])
+
+
+if __name__ == "__main__":
+    main(sys.argv)