eric/Crosspoint
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge pull request #109 from jpirnay/kosync_extension

Browse Source 
refactor: Picking up some itsthisjustin ideas for KOreader synchronisation
This commit is contained in:
jpirnay
2026-04-20 14:21:31 +02:00
committed by GitHub
parent 6f52aec95d 440d64e48e
commit 83fe041a6b
24 changed files with 493 additions and 140 deletions
Download Patch File Download Diff File Expand all files Collapse all files
lib/Epub/Epub/parsers/ChapterHtmlSlimParser.cpp
+18 -5
View File
@@ -593,7 +593,7 @@ void XMLCALL ChapterHtmlSlimParser::startElement(void* userData, const XML_Char*
(self->currentPageNextY + totalImageHeightWithSpacing > self->viewportHeight)) {
LOG_DBG("EHP", "Image page break: currentY=%d needed=%d viewportH=%d", self->currentPageNextY,
totalImageHeightWithSpacing, self->viewportHeight);
self->paragraphIndexPerPage.push_back(self->xpathParagraphIndex);
self->paragraphLutPerPage.push_back({self->lastBodyChildByteOffset, self->xpathParagraphIndex});
self->completePageFn(std::move(self->currentPage));
self->completedPageCount++;
self->currentPage.reset(new Page());
@@ -687,8 +687,16 @@ void XMLCALL ChapterHtmlSlimParser::startElement(void* userData, const XML_Char*
// check so that hidden <p> elements are still counted, matching ChapterXPathIndexer's
// counting (pure XML, no CSS). This ensures paragraph indices in the section cache LUT
// align with KOReader's crengine XPath indices.
if (self->xpathBodyDepth >= 0 && self->depth == self->xpathBodyDepth + 1 && strcmp(name, "p") == 0) {
self->xpathParagraphIndex++;
// At the same time, record the byte offset of every direct-body-child element start:
// the forward mapper's partial-parse heuristic requires the seek hint to land on a
// body-child boundary, otherwise partialBaseDepth can misidentify wrapped paragraphs.
if (self->xpathBodyDepth >= 0 && self->depth == self->xpathBodyDepth + 1) {
if (self->activeParser) {
self->lastBodyChildByteOffset = static_cast<uint32_t>(XML_GetCurrentByteIndex(self->activeParser));
}
if (strcmp(name, "p") == 0) {
self->xpathParagraphIndex++;
}
}
if (matches(name, SKIP_TAGS, NUM_SKIP_TAGS)) {
@@ -1344,6 +1352,7 @@ bool ChapterHtmlSlimParser::parseAndBuildPages() {
XML_SetUserData(parser, this);
XML_SetElementHandler(parser, startElement, endElement);
XML_SetCharacterDataHandler(parser, characterData);
activeParser = parser;
// Compute the time taken to parse and build pages
const uint32_t chapterStartTime = millis();
@@ -1354,6 +1363,7 @@ bool ChapterHtmlSlimParser::parseAndBuildPages() {
XML_StopParser(parser, XML_FALSE); // Stop any pending processing
XML_SetElementHandler(parser, nullptr, nullptr); // Clear callbacks
XML_SetCharacterDataHandler(parser, nullptr);
activeParser = nullptr;
XML_ParserFree(parser);
file.close();
return false;
@@ -1376,6 +1386,7 @@ bool ChapterHtmlSlimParser::parseAndBuildPages() {
XML_StopParser(parser, XML_FALSE); // Stop any pending processing
XML_SetElementHandler(parser, nullptr, nullptr); // Clear callbacks
XML_SetCharacterDataHandler(parser, nullptr);
activeParser = nullptr;
XML_ParserFree(parser);
file.close();
return false;
@@ -1389,6 +1400,7 @@ bool ChapterHtmlSlimParser::parseAndBuildPages() {
XML_StopParser(parser, XML_FALSE); // Stop any pending processing
XML_SetElementHandler(parser, nullptr, nullptr); // Clear callbacks
XML_SetCharacterDataHandler(parser, nullptr);
activeParser = nullptr;
XML_ParserFree(parser);
file.close();
return false;
@@ -1400,6 +1412,7 @@ bool ChapterHtmlSlimParser::parseAndBuildPages() {
XML_StopParser(parser, XML_FALSE); // Stop any pending processing
XML_SetElementHandler(parser, nullptr, nullptr); // Clear callbacks
XML_SetCharacterDataHandler(parser, nullptr);
activeParser = nullptr;
XML_ParserFree(parser);
file.close();
@@ -1410,7 +1423,7 @@ bool ChapterHtmlSlimParser::parseAndBuildPages() {
anchorData.push_back({std::move(pendingAnchorId), static_cast<uint16_t>(completedPageCount)});
pendingAnchorId.clear();
}
paragraphIndexPerPage.push_back(xpathParagraphIndex);
paragraphLutPerPage.push_back({0u, xpathParagraphIndex}); // post-parse: no byte offset available
completePageFn(std::move(currentPage));
completedPageCount++;
currentPage.reset();
@@ -1431,7 +1444,7 @@ ParsedText::LineProcessResult ChapterHtmlSlimParser::addLineToPage(std::shared_p
}
if (currentPageNextY + lineHeight > viewportHeight) {
paragraphIndexPerPage.push_back(xpathParagraphIndex);
paragraphLutPerPage.push_back({lastBodyChildByteOffset, xpathParagraphIndex});
completePageFn(std::move(currentPage));
completedPageCount++;
currentPage.reset(new Page());
lib/Epub/Epub/parsers/ChapterHtmlSlimParser.h
+20 -4
View File
@@ -90,9 +90,25 @@ class ChapterHtmlSlimParser {
// Counts <p> sibling indices (1-based, matching XPath convention) during page building.
// Stored per page in the section cache so that XPath p[N] can be resolved to a page
// without reparsing, and current page can generate an XPath without reparsing.
uint16_t xpathParagraphIndex = 0; // current <p> sibling index (1-based)
int xpathBodyDepth = -1; // depth of the <body> element (-1 = not yet seen)
std::vector<uint16_t> paragraphIndexPerPage; // <p> index at each page completion
uint16_t xpathParagraphIndex = 0; // current <p> sibling index (1-based)
int xpathBodyDepth = -1; // depth of the <body> element (-1 = not yet seen)
// Byte offset of the most recent direct-body-child element start (any tag at xpathBodyDepth+1).
// Recorded at the same depth condition that increments xpathParagraphIndex, so the stored
// offset is guaranteed to land on a body-child element boundary. This keeps the XPath forward
// mapper's partial-parse heuristic reliable for wrapped chapters: without this, the offset
// could point mid-way into a nested <div>/<section>, which confuses partialBaseDepth.
uint32_t lastBodyChildByteOffset = 0;
struct ParagraphLutEntry {
uint32_t xhtmlByteOffset; // byte offset of most recent body-child element start at page break
uint16_t paragraphIndex; // 1-based <p> index at page completion
};
std::vector<ParagraphLutEntry> paragraphLutPerPage; // deep LUT: one entry per page
// Active parser handle during parseAndBuildPages(), nullptr otherwise.
// Stored as a member so page-break sites (addLineToPage, image breaks) can call
// XML_GetCurrentByteIndex without needing the parser threaded through every call.
XML_Parser activeParser = nullptr;
// Footnote link tracking
bool insideFootnoteLink = false;
@@ -154,5 +170,5 @@ class ChapterHtmlSlimParser {
ParsedText::LineProcessResult addLineToPage(std::shared_ptr<TextBlock> line, bool lineEndsWithHyphenatedWord,
bool suppressHyphenationRetry);
const std::vector<std::pair<std::string, uint16_t>>& getAnchors() const { return anchorData; }
const std::vector<uint16_t>& getParagraphIndexPerPage() const { return paragraphIndexPerPage; }
const std::vector<ParagraphLutEntry>& getParagraphLutPerPage() const { return paragraphLutPerPage; }
};