Table page break bug and PDFBOX character bug

This is a reduced testcase which shows two things:

 - The repeated table header is broken on pagebreak
 - The arrow symbol is not correctly resolved from the unicode
character.

openhtmltopdf-examples/src/main/java/com/openhtmltopdf/testcases/RepeatedTableSample.java

@@ -0,0 +1,33 @@
+package com.openhtmltopdf.testcases;
+
+import com.openhtmltopdf.bidi.support.ICUBidiReorderer;
+import com.openhtmltopdf.bidi.support.ICUBidiSplitter;
+import com.openhtmltopdf.pdfboxout.PdfBoxRenderer;
+import com.openhtmltopdf.pdfboxout.PdfRendererBuilder;
+import com.openhtmltopdf.pdfboxout.PdfRendererBuilder.TextDirection;
+import org.apache.pdfbox.io.IOUtils;
+import org.apache.pdfbox.util.Charsets;
+
+import java.io.FileOutputStream;
+
+public class RepeatedTableSample {
+	public static void main(String[] args) throws Exception {
+		byte[] htmlBytes = IOUtils.toByteArray(RepeatedTableSample.class
+				.getResourceAsStream("/testcases/RepeatedTableSample.html"));
+		String html = new String(htmlBytes, Charsets.UTF_8);
+		FileOutputStream outputStream = new FileOutputStream("RepeatedTableSample.pdf");
+		PdfRendererBuilder builder = new PdfRendererBuilder();
+		builder.useBidiSplitter(new ICUBidiSplitter.ICUBidiSplitterFactory());
+		builder.useBidiReorderer(new ICUBidiReorderer());
+		builder.defaultTextDirection(TextDirection.LTR);
+		builder.withHtmlContent(html, "");
+		builder.toStream(outputStream);
+		PdfBoxRenderer pdfBoxRenderer = builder.buildPdfRenderer();
+		// Add Noto Sans Font
+		// pdfBoxRenderer.getFontResolver().addFont(tempFile.getPath(), null);
+		pdfBoxRenderer.layout();
+		pdfBoxRenderer.createPDF();
+		outputStream.close();
+
+	}
+}