static void Test03() {
const char* s = "a < b > c <> d <";
HtmlPullParser parser(s, str::Len(s));
HtmlToken* t = parser.Next();
utassert(t && t->IsText() && str::EqNIx(t->s, t->sLen, "a "));
t = parser.Next();
utassert(t && t->IsText() && str::EqNIx(t->s, t->sLen, "< b > c "));
t = parser.Next();
utassert(t && t->IsText() && str::EqNIx(t->s, t->sLen, "<> d "));
t = parser.Next();
utassert(t && t->IsError() && HtmlToken::UnclosedTag == t->error);
t = parser.Next();
utassert(!t);
}
static char *GetTextContent(HtmlPullParser& parser)
{
HtmlToken *tok = parser.Next();
if (!tok || !tok->IsText())
return NULL;
return ResolveHtmlEntities(tok->s, tok->sLen);
}
DocTocItem *MobiEngineImpl::GetTocTree()
{
if (!tocReparsePoint)
return NULL;
EbookTocItem *root = NULL;
ScopedMem<WCHAR> itemText;
ScopedMem<WCHAR> itemLink;
int itemLevel = 0;
int idCounter = 0;
// there doesn't seem to be a standard for Mobi ToCs, so we try to
// determine the author's intentions by looking at commonly used tags
HtmlPullParser parser(tocReparsePoint, str::Len(tocReparsePoint));
HtmlToken *tok;
while ((tok = parser.Next()) && !tok->IsError()) {
if (itemLink && tok->IsText()) {
ScopedMem<WCHAR> linkText(str::conv::FromHtmlUtf8(tok->s, tok->sLen));
if (itemText)
itemText.Set(str::Join(itemText, L" ", linkText));
else
itemText.Set(linkText.StealData());
}
else if (!tok->IsTag())
continue;
else if (Tag_Mbp_Pagebreak == tok->tag)
break;
else if (!itemLink && tok->IsStartTag() && Tag_A == tok->tag) {
AttrInfo *attr = tok->GetAttrByName("filepos");
if (!attr)
attr = tok->GetAttrByName("href");
if (attr)
itemLink.Set(str::conv::FromHtmlUtf8(attr->val, attr->valLen));
}
else if (itemLink && tok->IsEndTag() && Tag_A == tok->tag) {
PageDestination *dest = NULL;
if (!itemText) {
itemLink.Set(NULL);
continue;
}
if (IsExternalUrl(itemLink))
dest = new SimpleDest2(0, RectD(), itemLink.StealData());
else
dest = GetNamedDest(itemLink);
EbookTocItem *item = new EbookTocItem(itemText.StealData(), dest);
item->id = ++idCounter;
item->open = itemLevel <= 2;
AppendTocItem(root, item, itemLevel);
itemLink.Set(NULL);
}
else if (Tag_Blockquote == tok->tag || Tag_Ul == tok->tag || Tag_Ol == tok->tag) {
if (tok->IsStartTag())
itemLevel++;
else if (tok->IsEndTag() && itemLevel > 0)
itemLevel--;
}
}
return root;
}
static void Test02() {
const char* s = "<p>Last paragraph";
HtmlPullParser parser(s, str::Len(s));
HtmlToken* t = parser.Next();
utassert(t && t->IsTag() && t->IsStartTag() && Tag_P == t->tag);
t = parser.Next();
utassert(t && t->IsText() && str::EqNIx(t->s, t->sLen, "Last paragraph"));
}
// Parse s in place i.e. we assume we can modify it. Must be 0-terminated.
// The caller owns the memory for s.
HtmlElement *HtmlParser::ParseInPlace(char *s, UINT codepage)
{
if (this->html)
Reset();
this->html = s;
this->codepage = codepage;
HtmlPullParser parser(s, strlen(s));
HtmlToken *tok;
while ((tok = parser.Next())) {
char *tag = (char *)tok->s;
if (tok->IsError()) {
errorContext = tag;
switch (tok->error) {
case HtmlToken::UnclosedTag: return ParseError(ErrParsingElementName);
case HtmlToken::InvalidTag: return ParseError(ErrParsingClosingElement);
default: return ParseError(ErrParsingElement);
}
}
if (!tok->IsTag()) {
// ignore text content
assert(tok->IsText());
continue;
}
char *tagEnd = tag + tok->nLen;
if (!tok->IsEndTag()) {
// note: call tok->NextAttr() before zero-terminating names and values
AttrInfo *attr = tok->NextAttr();
*tagEnd = '\0';
StartTag(tag);
while (attr) {
char *name = (char *)attr->name;
char *nameEnd = name + attr->nameLen;
char *value = (char *)attr->val;
char *valueEnd = value + attr->valLen;
attr = tok->NextAttr();
*nameEnd = *valueEnd = '\0';
AppendAttr(name, value);
}
}
if (!tok->IsStartTag() || IsTagSelfClosing(tok->tag)) {
*tagEnd = '\0';
CloseTag(tag);
}
}
return rootElement;
}
static WCHAR *ExtractHtmlText(EpubDoc *doc)
{
size_t len;
const char *data = doc->GetTextData(&len);
str::Str<char> text(len / 2);
HtmlPullParser p(data, len);
HtmlToken *t;
Vec<HtmlTag> tagNesting;
while ((t = p.Next()) != NULL && !t->IsError()) {
if (t->IsText() && !tagNesting.Contains(Tag_Head) && !tagNesting.Contains(Tag_Script) && !tagNesting.Contains(Tag_Style)) {
// trim whitespace (TODO: also normalize within text?)
while (t->sLen > 0 && str::IsWs(t->s[0])) {
t->s++;
t->sLen--;
}
while (t->sLen > 0 && str::IsWs(t->s[t->sLen-1]))
t->sLen--;
if (t->sLen > 0) {
text.AppendAndFree(ResolveHtmlEntities(t->s, t->sLen));
text.Append(' ');
}
}
else if (t->IsStartTag()) {
// TODO: force-close tags similar to HtmlFormatter.cpp's AutoCloseOnOpen?
if (!IsTagSelfClosing(t->tag))
tagNesting.Append(t->tag);
}
else if (t->IsEndTag()) {
if (!IsInlineTag(t->tag) && text.Size() > 0 && text.Last() == ' ') {
text.Pop();
text.Append("\r\n");
}
// when closing a tag, if the top tag doesn't match but
// there are only potentially self-closing tags on the
// stack between the matching tag, we pop all of them
if (tagNesting.Contains(t->tag)) {
while (tagNesting.Last() != t->tag)
tagNesting.Pop();
}
if (tagNesting.Count() > 0 && tagNesting.Last() == t->tag)
tagNesting.Pop();
}
}
return str::conv::FromUtf8(text.Get());
}
DocTocItem *Fb2EngineImpl::GetTocTree()
{
EbookTocItem *root = NULL;
ScopedMem<WCHAR> itemText;
int titleCount = 0;
bool inTitle = false;
int level = 0;
size_t xmlLen;
const char *xmlData = doc->GetTextData(&xmlLen);
HtmlPullParser parser(xmlData, xmlLen);
HtmlToken *tok;
while ((tok = parser.Next()) && !tok->IsError()) {
if (tok->IsStartTag() && Tag_Section == tok->tag)
level++;
else if (tok->IsEndTag() && Tag_Section == tok->tag && level > 0)
level--;
else if (tok->IsStartTag() && Tag_Title == tok->tag) {
inTitle = true;
titleCount++;
}
else if (tok->IsEndTag() && Tag_Title == tok->tag) {
if (itemText)
str::NormalizeWS(itemText);
if (!str::IsEmpty(itemText.Get())) {
ScopedMem<WCHAR> name(str::Format(TEXT(FB2_TOC_ENTRY_MARK) L"%d", titleCount));
PageDestination *dest = GetNamedDest(name);
EbookTocItem *item = new EbookTocItem(itemText.StealData(), dest);
item->id = titleCount;
item->open = level <= 2;
AppendTocItem(root, item, level);
}
inTitle = false;
}
else if (inTitle && tok->IsText()) {
ScopedMem<WCHAR> text(str::conv::FromHtmlUtf8(tok->s, tok->sLen));
if (str::IsEmpty(itemText.Get()))
itemText.Set(text.StealData());
else
itemText.Set(str::Join(itemText, L" ", text));
}
}
return root;
}
