public void reconstruct_formatting_elements()
{
// tests attributes and multi b
string h = "<p><b class=one>One <i>Two <b>Three</p><p>Hello</p>";
HtmlDocument doc = HtmlDocument.Parse(h);
Assert.Equal(TextUtil.CompressWhitespace("<p><b class=\"one\">One <i>Two <b>Three</b></i></b></p>\n<p><b class=\"one\"><i><b>Hello</b></i></b></p>"),
TextUtil.CompressWhitespace(doc.Body.InnerHtml));
}
public void reconstruct_formatting_elements_in_table()
{
// tests that tables get formatting markers -- the <b> applies outside the table and does not leak in,
// and the <i> inside the table and does not leak out.
string h = "<p><b>One</p> <table><tr><td><p><i>Three<p>Four</i></td></tr></table> <p>Five</p>";
HtmlDocument doc = HtmlDocument.Parse(h);
string want = "<p><b>One</b></p>\n" +
"<b> \n" +
" <table>\n" +
" <tbody>\n" +
" <tr>\n" +
" <td><p><i>Three</i></p><p><i>Four</i></p></td>\n" +
" </tr>\n" +
" </tbody>\n" +
" </table> <p>Five</p></b>";
Assert.Equal(TextUtil.CompressWhitespace(want), TextUtil.CompressWhitespace(doc.Body.InnerHtml));
}
public void handles_unclosed_formatting_elements()
{
// whatwg: formatting elements get collected and applied, but excess elements are thrown away
string h = "<!DOCTYPE html>\n" +
"<p><b class=x><b class=x><b><b class=x><b class=x><b>X\n" +
"<p>X\n" +
"<p><b><b class=x><b>X\n" +
"<p></b></b></b></b></b></b>X";
HtmlDocument doc = HtmlDocument.Parse(h);
string want = "<!DOCTYPE html>\n" +
"<html>\n" +
"<head></head>\n" +
"<body>\n" +
"<p><b class=\"x\"><b class=\"x\"><b><b class=\"x\"><b class=\"x\"><b>X </b></b></b></b></b></b></p>\n" +
"<p><b class=\"x\"><b><b class=\"x\"><b class=\"x\"><b>X </b></b></b></b></b></p>\n" +
"<p><b class=\"x\"><b><b class=\"x\"><b class=\"x\"><b><b><b class=\"x\"><b>X </b></b></b></b></b></b></b></b></p>\n" +
"<p>X</p>\n" +
"</body>\n" +
"</html>";
Assert.Equal(TextUtil.CompressWhitespace(want), TextUtil.CompressWhitespace(doc.InnerHtml));
}
