| Age | Commit message (Collapse) | Author |
|---|
|
This updates Noto Sans Symbols to version 1.05, which includes the
Georgian lari symbol, and cuts the subsetted fonts slightly wider to
include the symbol at U+20BE.
The Unicode proposal for the character, accepted at the last UTC
meeting is at http://www.unicode.org/L2/L2014/14161-lari-proposal.pdf
Bug: 17554382
Change-Id: Iff45edcec594881014d4b82e0bce18f6e85cf701
|
|
Update NotoColorEmoji and the subsetted version to version 1.21,
which compresses its internal PNG files using pngquant from
https://github.com/pornel/pngquant. This is a lossy compression, but
the resulting images appear the same to my naked eye.
This saves 941kB on the extended font profile and 44kB on the
non-extended profile.
Bug: 17377714
Bug: 15348046
Change-Id: Ic363285a4b5ed19d07efa83b92637179874c730a
|
|
According to the OpenType layout specifications, Khmer and Lao fonts
need to support DOTTED CIRCLE and ZWSP, while Khmer addtionally needs
to support ZWNJ and ZWJ. Otherwise, vowels without base letters
would be displyed without a dotted circle, among other potential
issues (https://code.google.com/p/noto/issues/detail?id=4).
This adds glyphs from Noto Sans LGC for ZWJ, ZWNJ, and DOTTED CIRCLE
to the Khmer fonts, and ZWSP and DOTTED CIRCLE to the Lao fonts.
Bug: 13017773
Bug: 17445029
Change-Id: Id22878aad1f44493b3a927af9b9be26446a6d47f
|
|
Several issues have been fixed, most importantly the shape of U+5605
(https://code.google.com/p/noto/issues/detail?id=38) and the position
of U+FF0C FULLWIDTH COMMA
(https://code.google.com/p/noto/issues/detail?id=54) for Traditional
Chinese. Mappings have also been added to cmap tables for various
CJK Compatibility characters with canonical decompositions, which
should fix Chrome issues with the characters.
Bug: 17416668
Bug: 17245353
Bug: 15816880
Change-Id: I930d8719ff6c743a973c578d20a4fed8b3a8ecae
|
|
|
|
A slightly larger subset of Noto Sans Symbols is included on all
devices, to complete support for WGL4 and a few more frequently used
characters (the impact on font file size is 2kB).
A larger subset of Noto Sans Symbols is included on devices on the
extended font profile to also complete the support for the ARIB
character set, and as much as possible, the following Unicode blocks
(chosen based on frequency of characters on the web): Combining
Diacritical Marks for Symbols, Letterlike Symbols, Mathematical
Operators, Miscellaneous Technical, Control Pictures, Optical
Character Recognition, Geometric Shapes, Miscellaneous Symbols,
Supplemental Arrows-A, and Supplemental Mathematical Operators.
A subset of NotoSansJP is added for the devices on the extended font
profile, to match KLP character coverage and complete support for the
following Unicode blocks (as much as available in NotoSansJP): CJK
Symbols and Punctuation, Hiragana, Katakana, Katakana Phonetic
Extensions, Enclosed CJK Letters and Months, CJK Compatibility, and
CJK Unified Ideographs Extension B, C, and D.
The total impact on font file size is 296kB for the extended font
profile and 2kB on the non-extended one.
Bug: 4673396
Bug: 13680438
Bug: 16314977
Bug: 16856221
Bug: 17208391
Change-Id: I1f3fb87f3828151c6b439b63a73abc1eac591212
|
|
This is a temporary "easy" workaround for some of the cases that text
appears too crowded: sometimes too wide, sometimes too tall. The
correct fix would be language- and script-dependent UI parameters, or
generally just more free space in various UI elements.
Bug: 16145857
Change-Id: I38607b22b248476fb522bcc880760d994b11c39c
|
|
This shrinks the fonts, specially the NotoColorEmoji-Subsetted, to
about half their previous size!
Before:
-rw-r--r-- 1 behdad eng 3454256 Aug 14 19:42 NotoColorEmoji-Subsetted.ttf
-rw-r--r-- 1 behdad eng 6251476 Aug 14 19:42 NotoColorEmoji.ttf
After:
-rw-r--r-- 1 behdad eng 1428096 Aug 20 12:31 NotoColorEmoji-Subsetted.ttf
-rw-r--r-- 1 behdad eng 3250680 Aug 20 12:31 NotoColorEmoji.ttf
Understanding what's happening is key. First, these are reasons for why
I include a 128px and a 64px in the font to begin with:
* Because for small font sizes, it's faster to scale down from a 64px
image than the 128px, since the full-size image has to be
decompressed from PNG before it can be scaled,
* If the scaler is low-quality and does bilinear interpolation only,
then scaling down by more than a factor of two can produce aliasing
artifacts.
At this point, we believe that just shipping one strike might be good
enough. This is to be tested on devices.
I also want to make sure we understand why the Subsetted font shrinks by
more than half! There are two sets of images in these fonts: the emoji,
and the flags.
* The flags are waved, so they are not very regular, and for those,
the 128px images are bigger than the 64px files.
* For plain emoji, they are very simple clipart images that compress
very well in PNG (RLE within and across runs). So we don't expect
the 64px images to be significantly smaller than the 128px ones.
Plus, the 128px images were created from SVG, and have generally one
pixel of antialiasing around them. But then we shrank those down by
a factor of two using convert(1), and this has produces *two* pixels
of antialiasing, because, I suppose, convert has used a filter too
sophisticated for what we needed.
In retrospect, we should have generated the 64px strike directly from
SVG. Or have made sure that a simple box filter is used for the down
scaling. At any rate, for this reason, the 64px emoji compress worse
than the 128px ones, hence the sizes.
If we decided that we need to keep the 64px, we can at least consider
regenerating them from SVG so they compress better.
Bug: 15348046
Change-Id: I685dcf7e9786b602bf9a35ca77e7851c6759a08f
|
|
The fonts are passed through Noto's identity subsetter at
https://code.google.com/p/noto/source/browse/nototools/subset.py?r=bdd10822706b
which keeps the same functionality, but compresses tables, removes
unreachable glyphs, converts post table to version 3.0, etc.
The build footprint of the extended font profile is reduced by 352KB. For
the non-extended font profile, it is reduced by 298KB.
Bug: 15348046
Change-Id: I66777a0efeb98da6007091548a6ee847ecb4ea1f
|
|
New images are used for U+26C4, U+26F3, U+2708, U+1F356, U+1F357, U+1F378,
U+1F385, U+1F38E, U+1F418, U+1F426, U+1F440, U+1F442, U+1F443, U+1F444,
U+1F445, U+1F452, U+1F466, U+1F467, U+1F468, U+1F469, U+1F46E, U+1F46F,
U+1F470, U+1F471, U+1F472, U+1F473, U+1F474, U+1F475, U+1F476, U+1F477,
U+1F478, U+1F47C, U+1F481, U+1F483, U+1F485, U+1F486, U+1F487, U+1F48F,
U+1F491, U+1F499, U+1F49A, U+1F49B, U+1F49C, U+1F4A0, U+1F4AA, U+1F4E1,
U+1F52B, U+1F5FE, U+1F603, U+1F606, U+1F608, U+1F630, U+1F631, and U+1F635.
All flags (including the ten old flags) are updated to use the better flag
images from https://github.com/googlei18n/region-flags based on Wikipedia
public domain flags. Also, flags have been added for AX, CC, EU, and PS.
To save size, the subsetted font, used on non-extended profiles, has the
same character coverage as KLP.
Impact on footprint: increased by 132KB on non-extended font profiles (due
to updated images, added digits and GSUB, etc), decreased by 64KB on
extended font profiles (due to smaller flag images).
Bug: 11724363
Bug: 16237162
Bug: 16239371
Bug: 17018909
Change-Id: I702625b7afa57225f922397f09d0f4095561563e
|
|
Gujarati and Gurmukhi fonts now support Common Indic Number Forms
(U+A830..A839) that are needed for the scripts. All three sets have been
cleaned up to remove unreachable glyphs, fix glyphs that go outside
Android's vertical bounds, and fix missing extrema. Footprint is reduced by
24KB on the extended font profile and by 13KB on the non-extended one.
Bug: 15348046
Change-Id: I82ba120485cde96829a7565feebfc0eb9b3a7d67
(cherry picked from commit 0867743b629d5a8ab274ae6727b04a4d4c8f8fff)
|
|
The flags are based on public domain flags made available by The World
Factbook website. They are supported using standard two-character sequences
of Regional Indicator Symbols through GSUB substitutions. Not all
territories are included yet.
Additionally, characters for ASCII digits and number sign have been added
and GSUB substitutions for two-character sequences combining those and the
combining keycap character (U+20E3) have been added.
The new font is only included in the EXTENDED_FONT_FOOTPRINT profile. File
size is increased by 2993KB.
Bug: 7557244
Change-Id: I37aada4538213c3722e0cf44f102534a66c7ffd6
|
|
U+303D PART ALTERNATION MARK is the only character remaining in the black
and white AndroidEmoji font that is used. This makes it possible for
AndroidEmoji to be removed.
Additionally, the information in the name table is corrected and improved,
and the post table has been converted to version 3.0 to reduce font size.
Altogether, file size is reduced by 6KB.
Bug: 15777351
Change-Id: Ic868d9f56dd286c17df7371b46808a2cd23e4921
|
|
The Noto Sans Japanese font will be replaced by Motoya in all builds.
Change-Id: I65a1d99d738c67798f4ea3a3e1aeb5a649d97ccc
|
|
|
|
The new Noto Sans Symbols subsetted font adds all characters from
DroidSansFallback coverage that are already supported in the complete Noto
Sans Symbols font but are not available in Roboto. This change opens up the
opportunity to subset DroidSansFallback to remove all characters already
supported in Roboto or the new Symbols subset, and remove the fallback font
from the extended font profile altogether.
Change-Id: I450eaaa5c40f72ddfa68104bfef72dd2edcee9ee
|
|
This would make it possible to further subset Droid Sans Fallback to remove
characters already supported higher in the font fallback chain.
Change-Id: I9ddcfff145cdeb23b8dd58dffef7acc49ae2248e
|
|
|
|
The last change was breaking the build.
Change-Id: I796949eb96c3cf14b964baaf3a2b7271f7fe97f9
|
|
|
|
* commit '01ccd2b6e1ffe4c8baaf77581be8ee5e517c611e':
Update NotoSansSymbols-Subsetted [DO NOT MERGE]
|
|
This is to save space on the devices with limited space. The following
fonts are moved: Canadian Aboriginal, Cherokee, Gujarati, Gurmukhi, Sinhala,
and Yi.
Change-Id: Ibd632df7d8f0bc8d331247345b06b7f5afd73c35
|
|
The nameID 4 (Full font name) is required to load the fonts on a desktop
Java VM. This change adds the nameID as "Noto Sans Symbols Regular".
This font is cherrypicked from bce7af402915f1703fdd13b097df4ab91befbb3f
and also contains some other fixes.
Change-Id: I67d37f387776e667f45794d6d0372ec35d61b4d3
|
|
The mismatch between the Latin baseline and Telugu and Kannada baselines was
too large. The new fonts shift that baseline significantly down (although
still not completely aligned, which is achievable considering Android's
limit of -555 units) and also vertically compress the below baseline glyphs
to some degree, resulting in a more consistent look. Fixes
https://code.google.com/a/google.com/p/noto-alpha/issues/detail?id=160.
Change-Id: I78a5272628bc0f181b116826c6561804c00edd9b
|
|
In the Malayalam fonts, TT-TT-LA was fixed to produce <TT-TTA, candrakkala,
LA> instead of the previous <TTA, candrakkala, TT-LA>
(https://code.google.com/a/google.com/p/noto-alpha/issues/detail?id=186).
In the Symbols font, U+20BC MANAT SIGN is added for the Azerbaijani currency
(https://code.google.com/a/google.com/p/noto-alpha/issues/detail?id=237) and
some geometrical symbols (U+25AB, 25FB..25FE, 2600, 26AA..26AB, 2757,
2934..2935, 2B05..2B07, 2B1B..2B1C, 2B50, 2B55) are added to open the way
for removal of the monochrome AndroidEmoji.ttf.
Bug: 10246852
Bug: 15777351
Change-Id: I91e46d23a39fa5d75bc752854b4b1558aa0ebc9a
|
|
Fixes the vertical versions of white lenticular and square brackets
(U+3016..3017 and U+301A..301B) and thickens center horizontal line for
U+5451, among other fixes.
Bug: 15816724
Bug: 15758882
Change-Id: I2bdf270e8bd01aa3b1c5839295ae61d4fab33745
|
|
Braille support is added through expanding the subsetted Noto Sans
Symbols font.
Bug: 15708701
Change-Id: I14db765bc2c58e86d35ac8f0bd667439bd086a2d
|
|
...also did a major cleanup of the logic here to reduce copypaste
clutter.
Bug: 15569561
Change-Id: I74685a8431c13cc69e94af21c7c298f42a9a0309
|
|
Sync fonts with upstream updates. Most important fixes are formation of
Shrii ligature in Tamil (https://code.google.com/p/noto/issues/detail?id=23)
and several small fixes for Myanamar. The changes for Khmer and Lao are
minimal.
Change-Id: I6991947040847bf3af750538fc0b010a7ae7fb7a
|
|
This commit represents a mass update of the Noto Fonts project, including
support for a number of new scripts, and a number of fonts being moved from
Droid to Noto versions.
Change-Id: Ic95f5da355da583d03a2688cd51f341f3be740ec
|
|
Bug: 11177194
Change-Id: Ia789d7ca37a6a8a2888956dfcc7f19cd0343a1a6
|
|
Added the following characters:
U+2611 BALLOT BOX WITH CHECK
U+2614 UMBRELLA WITH RAIN DROPS
U+2615 HOT BEVERAGE
Bug: 11029931
Change-Id: I36ca239ae2b4acb673aeebf9e017dddab4ea70ca
|
|
- Added 1/64 (6.25%) padding, half on each left and right of glyphs.
- Restore Standard Variants glyphs that are not provided by other
fonts in the fallback chain:
U+231A WATCH
U+231B HOURGLASS
U+2601 CLOUD
U+263A WHITE SMILING FACE
U+2648 ARIES
U+2649 TAURUS
U+264A GEMINI
U+264B CANCER
U+264C LEO
U+264D VIRGO
U+264E LIBRA
U+264F SCORPIUS
U+2650 SAGITTARIUS
U+2651 CAPRICORN
U+2652 AQUARIUS
U+2653 PISCES
U+267B BLACK UNIVERSAL RECYCLING SYMBOL
U+267F WHEELCHAIR SYMBOL
U+2693 ANCHOR
U+26A0 WARNING SIGN
U+26A1 HIGH VOLTAGE SIGN
U+26BD SOCCER BALL
U+26BE BASEBALL
U+26C4 SNOWMAN WITHOUT SNOW
U+26C5 SUN BEHIND CLOUD
U+26D4 NO ENTRY
U+26EA CHURCH
U+26F2 FOUNTAIN
U+26F3 FLAG IN HOLE
U+26F5 SAILBOAT
U+26FA TENT
U+26FD FUEL PUMP
U+1F004 MAHJONG TILE RED DRAGON
U+1F17F NEGATIVE SQUARED LATIN CAPITAL LETTER P
U+1F21A SQUARED CJK UNIFIED IDEOGRAPH-7121
U+1F22F SQUARED CJK UNIFIED IDEOGRAPH-6307
Bug: 10144988
Bug: 10310267
Change-Id: Ia619d6a594ea2a0382eb3146d639df014ecabeec
|
|
Bug: 10144988
Change-Id: Iab39e7e1043ad00bd00fc518823db29bce05b7d4
|
|
Ok, it is actually *very* stripped down and contains only:
U+20BA TURKISH LIRA SIGN
U+22EE VERTICAL ELLIPSIS
U+266B BEAMED EIGHTH NOTES
Bug: 10639804
Change-Id: I5c408fe16cf834664b063ca430111ac37e39eb9e
|
|
Bug: 7672029
Change-Id: Ice3b742404b4ce3358a244cfb610921da79f1bf8
|
|
Bug: 7672029
Change-Id: Ib1e921d95a16e9e5d2c1e4209815335027af86a8
|
|
This commit adds the 20130214 versions of the Noto fonts for Bengali,
Kannada, and Telugu. All scripts come in normal and bold weights, and
in UI variants (more compact vertical metrics) as well as normal.
Change-Id: I17ea54398b4b0c87146fa9bac9419058ac2d02a9
|
|
This change renames the NotoSansMalayalam and NotoSansMalayalamUI fonts
to the NotoSansMalayalam-Regular (and similarly for UI), to be
consistent with the naming scheme for the Noto fonts in general.
Change-Id: I83e219b8727f9df7d19a2e354469c9ec6c59956f
|
|
Noto Sans Malayalam is the first of the new Noto series of fonts to be
included in Android. This is the 0.7 version of the font.
Please refer to the upstream project page for more details:
https://code.google.com/p/noto/
Change-Id: Iadd2b5ab335a06e293ffa037d20fd6797a07eb17
|
|
|
