JEditorPane slow HTML parsing and GUI lockup

Question

Hi,
I use a JEditorPane to display HTML page. If the page is relatively small, everything is ok, but with big pages (about 1 Mb) it takes a huge amount of time to load (actually, to parse) the page. And the main problem is that user interface locks and doesn't even repaint unitl page is loaded. Is there any workaround to this problem? I use JDK 1.5.0_2 std. edition.

Answer 1

Try loading a 1Mb plain text file and you will find
that loading is slow. So it will only be slower with
the additional parsing required for an HTML file.Really? The following is the text loading code, which is run in a separate thread:
                    BufferedInputStream in = new BufferedInputStream(con.getInputStream());
                    BufferedReader reader = new BufferedReader(new InputStreamReader(in));
                    String line;
                    if(type.startsWith("text/plain")||type.startsWith("text/xml")) // It's not an HTML page
                        jEditorPane1.setContentType("text/plain");
                        jEditorPane1.setFont(new Font("Monospaced",Font.PLAIN,14));               
                        page=new StringBuilder();
                        Document ed_doc=jEditorPane1.getDocument();
                        while ( ( line = reader.readLine()) != null )
                            line=line.concat("\n");
                            page.append(line);
                            ed_doc.insertString(ed_doc.getLength(),line,null);
                    jEditorPane1.getDocument().putProperty(HTMLDocument.StreamDescriptionProperty, u);Yes, it takes some time to load a big page, but no, it does not block GUI repainting and does not block any events such as mouse, etc.
I'm not an HTML
expert but I believe if you have, for example, a
large table structure then the table won't be painted
until the parser knows how many columns is in the
table and how wide to make each column. It can't
determine column width until all the data is read.The files I use for testing don't have tables, etc. See this: http://www.lib.ru/ADAMS/liff.txt (Though it ends with .txt, it is html file - almost plain text with some HTML formatting tags). So, again, the problem is not that parsing is slow; the problem is that parsing blocks any gui events, and I'd like to know why and how to cope this problem. Running setPage() in separate thread doesn't change anything.