snark13/Xteink-X4-crosspoint-reader
1
0
Fork 0
mirror of https://github.com/daveallie/crosspoint-reader.git synced 2026-02-06 15:47:39 +03:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

refactor: Simplify break information handling in buildExplicitBreakInfos and enhance clarity in liangBreakIndexes

Browse Source
This commit is contained in:
Arthur Tazhitdinov 2026-01-14 19:17:29 +05:00
parent 5dab3d4fcf
commit 04a084f6c8
2 changed files with 11 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
lib/Epub/Epub/hyphenation/Hyphenator.cpp
View File

@ -48,28 +48,6 @@ std::vector<Hyphenator::BreakInfo> buildExplicitBreakInfos(const std::vector<Cod
breaks.push_back({cps[i + 1].byteOffset, isSoftHyphen(cp)});
}
if (breaks.empty()) {
return breaks;
}
// Sort by byte offset so we can deduplicate sequential markers in-place.
std::sort(breaks.begin(), breaks.end(), [](const Hyphenator::BreakInfo& lhs, const Hyphenator::BreakInfo& rhs) {
return lhs.byteOffset < rhs.byteOffset;
});
// Deduplicate in-place: merge entries at same offset while retaining "needs hyphen" flag.
size_t writePos = 0;
for (size_t readPos = 1; readPos < breaks.size(); ++readPos) {
if (breaks[readPos].byteOffset == breaks[writePos].byteOffset) {
// Merge: explicit hyphen wins over soft hyphen at same offset.
breaks[writePos].requiresInsertedHyphen =
breaks[writePos].requiresInsertedHyphen || breaks[readPos].requiresInsertedHyphen;
} else {
breaks[++writePos] = breaks[readPos];
}
}
breaks.resize(writePos + 1);
return breaks;
}

15
lib/Epub/Epub/hyphenation/LiangHyphenation.cpp
View File

@ -354,36 +354,43 @@ std::vector<size_t> liangBreakIndexes(const std::vector<CodepointInfo>& cps,
return {};
}
// Liang scores: one entry per augmented char (leading/trailing dots included).
std::vector<uint8_t> scores(augmented.charCount(), 0);
// Walk every starting character position and stream bytes through the trie.
for (size_t charStart = 0; charStart < augmented.charByteOffsets.size(); ++charStart) {
size_t byteStart = augmented.charByteOffsets[charStart];
const size_t byteStart = augmented.charByteOffsets[charStart];
AutomatonState state = root;
for (size_t cursor = byteStart; cursor < augmented.bytes.size(); ++cursor) {
AutomatonState next;
if (!transition(automaton, state, augmented.bytes[cursor], next)) {
break;
break; // No more matches for this prefix.
}
state = next;
if (state.levels && state.levelsLen > 0) {
size_t offset = 0;
// Each packed byte stores the byte-distance delta and the Liang level digit.
for (size_t i = 0; i < state.levelsLen; ++i) {
const uint8_t packed = state.levels[i];
const size_t dist = static_cast<size_t>(packed / 10);
const uint8_t level = static_cast<uint8_t>(packed % 10);
offset += dist;
const size_t splitByte = byteStart + offset;
if (splitByte >= augmented.byteToCharIndex.size()) {
continue;
}
const int32_t boundary = augmented.byteToCharIndex[splitByte];
if (boundary < 0) {
continue;
continue; // Mid-codepoint byte, wait for the next one.
}
if (boundary < 2 || boundary + 2 > static_cast<int32_t>(augmented.charCount())) {
continue;
continue; // Skip splits that land in the leading/trailing sentinels.
}
const size_t idx = static_cast<size_t>(boundary);
if (idx >= scores.size()) {
continue;