flutter_flutter/libs/minikin/FontCollection.cpp
Roozbeh Pournader 16ed4ef07e Try putting combining marks in the same font run
Indic combining marks, when combined with a common character such as
a hyphen or a dotted circle, used to get rendered in a different font
due to the greedy algorithm used in determining runs, which resulted
in the base character and the combining mark getting rendered in
separate font runs, resulting in a dotted circle appearing in phrases
such as "100-ാം" (0031 0030 0030 002D 0D3E 0D02).

This change makes combining marks change the font run of the base
character if the base character is supported in the same font as the
combining mark, similar to the support for emoji modifiers and the
combining keycap.

Bug: 25036888
Bug: 24535344
Change-Id: I8e2798e8ecb8efaf723a0fd02c05c6fbdef8b365
2016-03-30 17:48:34 -07:00

424 lines
16 KiB
C++

/*
* Copyright (C) 2013 The Android Open Source Project
*
* 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.
*/
// #define VERBOSE_DEBUG
#define LOG_TAG "Minikin"
#include <cutils/log.h>
#include <algorithm>
#include "unicode/unistr.h"
#include "unicode/unorm2.h"
#include "FontLanguage.h"
#include "FontLanguageListCache.h"
#include "MinikinInternal.h"
#include <minikin/FontCollection.h>
using std::vector;
namespace android {
template <typename T>
static inline T max(T a, T b) {
return a>b ? a : b;
}
uint32_t FontCollection::sNextId = 0;
FontCollection::FontCollection(const vector<FontFamily*>& typefaces) :
mMaxChar(0) {
AutoMutex _l(gMinikinLock);
mId = sNextId++;
vector<uint32_t> lastChar;
size_t nTypefaces = typefaces.size();
#ifdef VERBOSE_DEBUG
ALOGD("nTypefaces = %zd\n", nTypefaces);
#endif
const FontStyle defaultStyle;
for (size_t i = 0; i < nTypefaces; i++) {
FontFamily* family = typefaces[i];
MinikinFont* typeface = family->getClosestMatch(defaultStyle).font;
if (typeface == NULL) {
continue;
}
family->RefLocked();
const SparseBitSet* coverage = family->getCoverage();
if (coverage == nullptr) {
family->UnrefLocked();
continue;
}
mFamilies.push_back(family); // emplace_back would be better
if (family->hasVSTable()) {
mVSFamilyVec.push_back(family);
}
mMaxChar = max(mMaxChar, coverage->length());
lastChar.push_back(coverage->nextSetBit(0));
}
nTypefaces = mFamilies.size();
LOG_ALWAYS_FATAL_IF(nTypefaces == 0,
"Font collection must have at least one valid typeface");
size_t nPages = (mMaxChar + kPageMask) >> kLogCharsPerPage;
size_t offset = 0;
// TODO: Use variation selector map for mRanges construction.
// A font can have a glyph for a base code point and variation selector pair but no glyph for
// the base code point without variation selector. The family won't be listed in the range in
// this case.
for (size_t i = 0; i < nPages; i++) {
Range dummy;
mRanges.push_back(dummy);
Range* range = &mRanges.back();
#ifdef VERBOSE_DEBUG
ALOGD("i=%zd: range start = %zd\n", i, offset);
#endif
range->start = offset;
for (size_t j = 0; j < nTypefaces; j++) {
if (lastChar[j] < (i + 1) << kLogCharsPerPage) {
FontFamily* family = mFamilies[j];
mFamilyVec.push_back(family);
offset++;
uint32_t nextChar = family->getCoverage()->nextSetBit((i + 1) << kLogCharsPerPage);
#ifdef VERBOSE_DEBUG
ALOGD("nextChar = %d (j = %zd)\n", nextChar, j);
#endif
lastChar[j] = nextChar;
}
}
range->end = offset;
}
}
FontCollection::~FontCollection() {
for (size_t i = 0; i < mFamilies.size(); i++) {
mFamilies[i]->UnrefLocked();
}
}
// Special scores for the font fallback.
const uint32_t kUnsupportedFontScore = 0;
const uint32_t kFirstFontScore = UINT32_MAX;
// Calculates a font score.
// The score of the font family is based on three subscores.
// - Coverage Score: How well the font family covers the given character or variation sequence.
// - Language Score: How well the font family is appropriate for the language.
// - Variant Score: Whether the font family matches the variant. Note that this variant is not the
// one in BCP47. This is our own font variant (e.g., elegant, compact).
//
// Then, there is a priority for these three subscores as follow:
// Coverage Score > Language Score > Variant Score
// The returned score reflects this priority order.
//
// Note that there are two special scores.
// - kUnsupportedFontScore: When the font family doesn't support the variation sequence or even its
// base character.
// - kFirstFontScore: When the font is the first font family in the collection and it supports the
// given character or variation sequence.
uint32_t FontCollection::calcFamilyScore(uint32_t ch, uint32_t vs, int variant, uint32_t langListId,
FontFamily* fontFamily) const {
const uint32_t coverageScore = calcCoverageScore(ch, vs, fontFamily);
if (coverageScore == kFirstFontScore || coverageScore == kUnsupportedFontScore) {
// No need to calculate other scores.
return coverageScore;
}
const uint32_t languageScore = calcLanguageMatchingScore(langListId, *fontFamily);
const uint32_t variantScore = calcVariantMatchingScore(variant, *fontFamily);
// Subscores are encoded into 31 bits representation to meet the subscore priority.
// The highest 2 bits are for coverage score, then following 28 bits are for language score,
// then the last 1 bit is for variant score.
return coverageScore << 29 | languageScore << 1 | variantScore;
}
// Calculates a font score based on variation sequence coverage.
// - Returns kUnsupportedFontScore if the font doesn't support the variation sequence or its base
// character.
// - Returns kFirstFontScore if the font family is the first font family in the collection and it
// supports the given character or variation sequence.
// - Returns 3 if the font family supports the variation sequence.
// - Returns 2 if the vs is a color variation selector (U+FE0F) and if the font is an emoji font.
// - Returns 2 if the vs is a text variation selector (U+FE0E) and if the font is not an emoji font.
// - Returns 1 if the variation selector is not specified or if the font family only supports the
// variation sequence's base character.
uint32_t FontCollection::calcCoverageScore(uint32_t ch, uint32_t vs, FontFamily* fontFamily) const {
const bool hasVSGlyph = (vs != 0) && fontFamily->hasVariationSelector(ch, vs);
if (!hasVSGlyph && !fontFamily->getCoverage()->get(ch)) {
// The font doesn't support either variation sequence or even the base character.
return kUnsupportedFontScore;
}
if ((vs == 0 || hasVSGlyph) && mFamilies[0] == fontFamily) {
// If the first font family supports the given character or variation sequence, always use
// it.
return kFirstFontScore;
}
if (vs == 0) {
return 1;
}
if (hasVSGlyph) {
return 3;
}
if (vs == 0xFE0F || vs == 0xFE0E) {
// TODO use all language in the list.
const FontLanguage lang = FontLanguageListCache::getById(fontFamily->langId())[0];
const bool hasEmojiFlag = lang.hasEmojiFlag();
if (vs == 0xFE0F) {
return hasEmojiFlag ? 2 : 1;
} else { // vs == 0xFE0E
return hasEmojiFlag ? 1 : 2;
}
}
return 1;
}
// Calculates font scores based on the script matching and primary langauge matching.
//
// If the font's script doesn't support the requested script, the font gets a score of 0. If the
// font's script supports the requested script and the font has the same primary language as the
// requested one, the font gets a score of 2. If the font's script supports the requested script
// but the primary language is different from the requested one, the font gets a score of 1.
//
// If two languages in the requested list have the same language score, the font matching with
// higher priority language gets a higher score. For example, in the case the user requested
// language list is "ja-Jpan,en-Latn". The score of for the font of "ja-Jpan" gets a higher score
// than the font of "en-Latn".
//
// To achieve the above two conditions, the language score is determined as follows:
// LanguageScore = s(0) * 3^(m - 1) + s(1) * 3^(m - 2) + ... + s(m - 2) * 3 + s(m - 1)
// Here, m is the maximum number of languages to be compared, and s(i) is the i-th language's
// matching score. The possible values of s(i) are 0, 1 and 2.
uint32_t FontCollection::calcLanguageMatchingScore(
uint32_t userLangListId, const FontFamily& fontFamily) {
const FontLanguages& langList = FontLanguageListCache::getById(userLangListId);
// TODO use all language in the list.
FontLanguage fontLanguage = FontLanguageListCache::getById(fontFamily.langId())[0];
const size_t maxCompareNum = std::min(langList.size(), FONT_LANGUAGES_LIMIT);
uint32_t score = fontLanguage.getScoreFor(langList[0]); // maxCompareNum can't be zero.
for (size_t i = 1; i < maxCompareNum; ++i) {
score = score * 3u + fontLanguage.getScoreFor(langList[i]);
}
return score;
}
// Calculates a font score based on variant ("compact" or "elegant") matching.
// - Returns 1 if the font doesn't have variant or the variant matches with the text style.
// - No score if the font has a variant but it doesn't match with the text style.
uint32_t FontCollection::calcVariantMatchingScore(int variant, const FontFamily& fontFamily) {
return (fontFamily.variant() == 0 || fontFamily.variant() == variant) ? 1 : 0;
}
// Implement heuristic for choosing best-match font. Here are the rules:
// 1. If first font in the collection has the character, it wins.
// 2. Calculate a score for the font family. See comments in calcFamilyScore for the detail.
// 3. Highest score wins, with ties resolved to the first font.
// This method never returns nullptr.
FontFamily* FontCollection::getFamilyForChar(uint32_t ch, uint32_t vs,
uint32_t langListId, int variant) const {
if (ch >= mMaxChar) {
return mFamilies[0];
}
const std::vector<FontFamily*>* familyVec = &mFamilyVec;
Range range = mRanges[ch >> kLogCharsPerPage];
std::vector<FontFamily*> familyVecForVS;
if (vs != 0) {
// If variation selector is specified, need to search for both the variation sequence and
// its base codepoint. Compute the union vector of them.
familyVecForVS = mVSFamilyVec;
familyVecForVS.insert(familyVecForVS.end(),
mFamilyVec.begin() + range.start, mFamilyVec.begin() + range.end);
std::sort(familyVecForVS.begin(), familyVecForVS.end());
auto last = std::unique(familyVecForVS.begin(), familyVecForVS.end());
familyVecForVS.erase(last, familyVecForVS.end());
familyVec = &familyVecForVS;
range = { 0, familyVecForVS.size() };
}
#ifdef VERBOSE_DEBUG
ALOGD("querying range %zd:%zd\n", range.start, range.end);
#endif
FontFamily* bestFamily = nullptr;
uint32_t bestScore = kUnsupportedFontScore;
for (size_t i = range.start; i < range.end; i++) {
FontFamily* family = (*familyVec)[i];
const uint32_t score = calcFamilyScore(ch, vs, variant, langListId, family);
if (score == kFirstFontScore) {
// If the first font family supports the given character or variation sequence, always
// use it.
return family;
}
if (score > bestScore) {
bestScore = score;
bestFamily = family;
}
}
if (bestFamily == nullptr) {
UErrorCode errorCode = U_ZERO_ERROR;
const UNormalizer2* normalizer = unorm2_getNFDInstance(&errorCode);
if (U_SUCCESS(errorCode)) {
UChar decomposed[4];
int len = unorm2_getRawDecomposition(normalizer, ch, decomposed, 4, &errorCode);
if (U_SUCCESS(errorCode) && len > 0) {
int off = 0;
U16_NEXT_UNSAFE(decomposed, off, ch);
return getFamilyForChar(ch, vs, langListId, variant);
}
}
bestFamily = mFamilies[0];
}
return bestFamily;
}
const uint32_t NBSP = 0xa0;
const uint32_t ZWJ = 0x200c;
const uint32_t ZWNJ = 0x200d;
const uint32_t HYPHEN = 0x2010;
const uint32_t NB_HYPHEN = 0x2011;
// Characters where we want to continue using existing font run instead of
// recomputing the best match in the fallback list.
static const uint32_t stickyWhitelist[] = { '!', ',', '-', '.', ':', ';', '?', NBSP, ZWJ, ZWNJ,
HYPHEN, NB_HYPHEN };
static bool isStickyWhitelisted(uint32_t c) {
for (size_t i = 0; i < sizeof(stickyWhitelist) / sizeof(stickyWhitelist[0]); i++) {
if (stickyWhitelist[i] == c) return true;
}
return false;
}
static bool isVariationSelector(uint32_t c) {
return (0xFE00 <= c && c <= 0xFE0F) || (0xE0100 <= c && c <= 0xE01EF);
}
bool FontCollection::hasVariationSelector(uint32_t baseCodepoint,
uint32_t variationSelector) const {
if (!isVariationSelector(variationSelector)) {
return false;
}
if (baseCodepoint >= mMaxChar) {
return false;
}
if (variationSelector == 0) {
return false;
}
// Currently mRanges can not be used here since it isn't aware of the variation sequence.
for (size_t i = 0; i < mVSFamilyVec.size(); i++) {
AutoMutex _l(gMinikinLock);
if (mVSFamilyVec[i]->hasVariationSelector(baseCodepoint, variationSelector)) {
return true;
}
}
return false;
}
void FontCollection::itemize(const uint16_t *string, size_t string_size, FontStyle style,
vector<Run>* result) const {
const uint32_t langListId = style.getLanguageListId();
int variant = style.getVariant();
FontFamily* lastFamily = NULL;
Run* run = NULL;
if (string_size == 0) {
return;
}
const uint32_t kEndOfString = 0xFFFFFFFF;
uint32_t nextCh = 0;
uint32_t prevCh = 0;
size_t nextUtf16Pos = 0;
size_t readLength = 0;
U16_NEXT(string, readLength, string_size, nextCh);
do {
const uint32_t ch = nextCh;
const size_t utf16Pos = nextUtf16Pos;
nextUtf16Pos = readLength;
if (readLength < string_size) {
U16_NEXT(string, readLength, string_size, nextCh);
} else {
nextCh = kEndOfString;
}
bool shouldContinueRun = false;
if (lastFamily != nullptr) {
if (isStickyWhitelisted(ch)) {
// Continue using existing font as long as it has coverage and is whitelisted
shouldContinueRun = lastFamily->getCoverage()->get(ch);
} else if (isVariationSelector(ch)) {
// Always continue if the character is a variation selector.
shouldContinueRun = true;
}
}
if (!shouldContinueRun) {
FontFamily* family = getFamilyForChar(ch, isVariationSelector(nextCh) ? nextCh : 0,
langListId, variant);
if (utf16Pos == 0 || family != lastFamily) {
size_t start = utf16Pos;
// Workaround for combining marks and emoji modifiers until we implement
// per-cluster font selection: if a combining mark or an emoji modifier is found in
// a different font that also supports the previous character, attach previous
// character to the new run. U+20E3 COMBINING ENCLOSING KEYCAP, used in emoji, is
// handled properly by this since it's a combining mark too.
if (utf16Pos != 0 &&
((U_GET_GC_MASK(ch) & U_GC_M_MASK) != 0 ||
(isEmojiModifier(ch) && isEmojiBase(prevCh))) &&
family && family->getCoverage()->get(prevCh)) {
const size_t prevChLength = U16_LENGTH(prevCh);
run->end -= prevChLength;
if (run->start == run->end) {
result->pop_back();
}
start -= prevChLength;
}
Run dummy;
result->push_back(dummy);
run = &result->back();
run->fakedFont = family->getClosestMatch(style);
lastFamily = family;
run->start = start;
}
}
prevCh = ch;
run->end = nextUtf16Pos; // exclusive
} while (nextCh != kEndOfString);
}
MinikinFont* FontCollection::baseFont(FontStyle style) {
return baseFontFaked(style).font;
}
FakedFont FontCollection::baseFontFaked(FontStyle style) {
return mFamilies[0]->getClosestMatch(style);
}
uint32_t FontCollection::getId() const {
return mId;
}
} // namespace android