XWIKI-19270: Add support for performing the PDF export using a browser running...

XWIKI-19270: Add support for performing the PDF export using a browser running in a Docker container
* Read smaller chunks from the PDF "stream" in order to avoid a BufferOverflow exception raised by the Tyrus (WebSocket) client used to communicate with the headless Chrome browser.

xwiki-platform-core/xwiki-platform-export/xwiki-platform-export-pdf/xwiki-platform-export-pdf-default/src/main/java/org/xwiki/export/pdf/internal/chrome/PrintToPDFInputStream.java

@@ -35,6 +35,11 @@
  */
 public class PrintToPDFInputStream extends InputStream
 {
+    /**
+     * Read chunks of 1MB.
+     */
+    private static final int BUFFER_SIZE = 1 << 20;
+
     private IO io;
 
     private String stream;
@@ -84,7 +89,7 @@ private byte[] readBuffer()
             return new byte[] {};
         }
 
-        Read read = this.io.read(this.stream);
+        Read read = this.io.read(this.stream, null, BUFFER_SIZE);
         this.finished = read.getEof() == Boolean.TRUE;
         if (read.getBase64Encoded() == Boolean.TRUE) {
             return Base64.getDecoder().decode(read.getData());