void EbookController::ExtractPageAnchors()
{
if (pageAnchorIds || pageAnchorIdxs) {
CrashIf(!pageAnchorIds || !pageAnchorIdxs);
return;
}
pageAnchorIds = new WStrVec();
pageAnchorIdxs = new Vec<int>();
ScopedMem<WCHAR> epubPagePath;
int fb2TitleCount = 0;
size_t len;
const char *data = doc.GetHtmlData(len);
HtmlPullParser parser(data, len);
HtmlToken *tok;
while ((tok = parser.Next()) != nullptr && !tok->IsError()) {
if (!tok->IsStartTag() && !tok->IsEmptyElementEndTag())
continue;
AttrInfo *attr = tok->GetAttrByName("id");
if (!attr && Tag_A == tok->tag && doc.Type() != Doc_Fb2)
attr = tok->GetAttrByName("name");
if (attr) {
ScopedMem<WCHAR> id(str::conv::FromUtf8(attr->val, attr->valLen));
pageAnchorIds->Append(str::Format(L"%s#%s", epubPagePath ? epubPagePath : L"", id.Get()));
pageAnchorIdxs->Append((int)(tok->GetReparsePoint() - parser.Start()));
}
// update EPUB page paths and create an anchor per chapter
if (Tag_Pagebreak == tok->tag &&
(attr = tok->GetAttrByName("page_path")) != nullptr &&
str::StartsWith(attr->val + attr->valLen, "\" page_marker />")) {
CrashIf(doc.Type() != Doc_Epub);
epubPagePath.Set(str::conv::FromUtf8(attr->val, attr->valLen));
pageAnchorIds->Append(str::Dup(epubPagePath));
pageAnchorIdxs->Append((int)(tok->GetReparsePoint() - parser.Start()));
}
// create FB2 title anchors (cf. Fb2Doc::ParseToc)
if (Tag_Title == tok->tag && tok->IsStartTag() && Doc_Fb2 == doc.Type()) {
ScopedMem<WCHAR> id(str::Format(TEXT(FB2_TOC_ENTRY_MARK) L"%d", ++fb2TitleCount));
pageAnchorIds->Append(id.StealData());
pageAnchorIdxs->Append((int)(tok->GetReparsePoint() - parser.Start()));
}
}
}
// Return the next parsed page. Returns NULL if finished parsing.
// For simplicity of implementation, we parse xml text node or
// xml element at a time. This might cause a creation of one
// or more pages, which we remeber and send to the caller
// if we detect accumulated pages.
HtmlPage *HtmlFormatter::Next(bool skipEmptyPages)
{
for (;;)
{
// send out all pages accumulated so far
while (pagesToSend.Count() > 0) {
HtmlPage *ret = pagesToSend.At(0);
pagesToSend.RemoveAt(0);
pageCount++;
if (skipEmptyPages && IsEmptyPage(ret))
delete ret;
else
return ret;
}
// we can call ourselves recursively to send outstanding
// pages after parsing has finished so this is to detect
// that case and really end parsing
if (finishedParsing)
return NULL;
HtmlToken *t = htmlParser->Next();
if (!t || t->IsError())
break;
currReparseIdx = t->GetReparsePoint() - htmlParser->Start();
CrashIf(!ValidReparseIdx(currReparseIdx, htmlParser));
if (t->IsTag())
HandleHtmlTag(t);
else if (!IgnoreText())
HandleText(t);
}
// force layout of the last line
AutoCloseTags(tagNesting.Count());
FlushCurrLine(true);
UpdateLinkBboxes(currPage);
pagesToSend.Append(currPage);
currPage = NULL;
// call ourselves recursively to return accumulated pages
finishedParsing = true;
return Next();
}
