martinrobinson

I've just finished an initial version of a mxj/Jitter class which renders html in a Jitter matrix.

http://miajo.co.uk/?MaxMSP:mxj_jithtmlrender

It needs Cobra/Lobo to be installed (see the README in the distribution) and the help file uses jhno's jsscrollbar.

I made it for relatively simple text formatting but it seems fairly robust (if slow) with more complex pages (see the Wikipedia example in the help file). I'd be keen for some feedback if there are any serious anomalies (there are quite a few kludges in there to get it to work). I know for example, that text color tags in the tag don't change the body text color (have a look at jhno's homepage for proof) and CSS generally works but not with background images set using CSS (see my miajo homepage for proof of that).

I'd be especially grateful to know/if how it works on XP (I don't have access to an XP machine with Max/MSP having moved jobs back to Mac-centricity!)

I've just finished an initial version of a mxj/Jitter class which renders html in a Jitter matrix. 

Available here:
http://miajo.co.uk/?MaxMSP:mxj_jithtmlrender

Cobra/Lobo:
http://html.xamjwg.org/cobra.jsp

jhno:
http://www.delicateear.com/maxmsp.html

I made it for relatively simple text formatting but it seems fairly robust (if slow) with more complex pages (see the Wikipedia example in the help file). I'd be keen for some feedback if there are any serious anomalies (there are quite a few kludges in there to get it to work). I know for example, that text color tags in the  tag don't change the body text color (have a look at jhno's homepage for proof) and CSS generally works but not with background images set using CSS (see my miajo homepage for proof of that). 

rendering-html-in-jitter

I'm just embarking on this same journey, so i'm glad i'm not the first!

https://cycling74.com/forums/index.php?t=msg&th=12639]

 where the possibility of using javax.swing.text.html was discussed, which seems like a good idea (fewer dependencies the better i reckon) - did you try this and find it not to work? Or is the Cobra way just much better?

I'm stuck no because the c74 wiki seems to be doing White Screen of Death so i can't get jhno, but i'll check out the other bits in while i wait for it to come back (ah, i see it's only the help file that needs jhno- i've been using jit.cellblock for scrolly lists recently - not bad!)

I see there was an older thread here [https://cycling74.com/forums/index.php?t=msg&th=12639] where the possibility of using javax.swing.text.html was discussed, which seems like a good idea (fewer dependencies the better i reckon) - did you try this and find it not to work? Or is the Cobra way just much better?

I'm stuck no because the c74 wiki seems to be doing White Screen of Death so i can't get jhno, but i'll check out the other bits in while i wait for it to come back (ah, i see it's only the help file  that needs jhno- i've been using jit.cellblock for scrolly lists recently - not bad!)

> I'm just embarking on this same journey, so i'm glad i'm not the first!

> possibility of using javax.swing.text.html was discussed, which seems like

> a good idea (fewer dependencies the better i reckon) - did you try this and

> find it not to work? Or is the Cobra way just much better?

> I'm stuck no because the c74 wiki seems to be doing White Screen of Death

> so i can't get jhno, but i'll check out the other bits in while i wait for

> it to come back (ah, i see it's only the help file that needs jhno- i've

> been using jit.cellblock for scrolly lists recently - not bad!)

>
> Hi Martin
>
> I'm just embarking on this same journey, so i'm glad i'm not the first!
>
> I see there was an older thread here [
> https://cycling74.com/forums/index.php?t=msg&th=12639] where the
> possibility of using javax.swing.text.html was discussed, which seems like
> a good idea (fewer dependencies the better i reckon) - did you try this and
> find it not to work? Or is the Cobra way just much better?
>
> I'm stuck no because the c74 wiki seems to be doing White Screen of Death
> so i can't get jhno, but i'll check out the other bits in while i wait for
> it to come back (ah, i see it's only the help file  that needs jhno- i've
> been using jit.cellblock for scrolly lists recently - not bad!)
>
> More soon i hope
>
> jonny B
>
>

I've got it working well for me now but with a couple of issues i was wondering if you had come across.

I seem to be getting crashes when quitting if my html objects (on jit.gl.videoplanes flying about in 3D). I noticed there was no notifyDelete() function - i've added one but i'm not sure i'm releasing everything properly - routine below)

The other problem i have is that most websites now use UTF-8 encoding, but this only seems to handle LATIN-1 (ISO 8859-1) properly - have you come across a way of using different text encodings? Or do i have to convert everything?

Thanks for the great start though, i'll probably work these things out eventually & will post back here if i do.

if (rendererContext != null) { rendererContext.close(); }

if (htmlPanel != null) { htmlPanel.removeAll(); }

public void notifyDeleted() {
 if (bi != null) { bi.flush(); }
 if (rendererContext != null) { rendererContext.close(); }
 if (htmlPanel != null) { htmlPanel.removeAll(); }
}

Fixed the encoding (the hard way, i think)

My version of getDocumentFromUrl follows.

//---------------------------------------

    private void getDocumentFromUrl(String uri) throws Exception {

        URLConnection connection = url.openConnection();

        InputStream in = connection.getInputStream();

        // A Reader should be created with the correct charset,

        // which may be obtained from the Content-Type header

        String encoding = connection.getContentEncoding();

            // take a peek in the file then

            byte[] buff = new byte[512];

            int rc = in.read(buff);

                // fish out charset from meta tag

                String s = new String(buff);

                Pattern p = Pattern.compile("",

                        Pattern.CASE_INSENSITIVE | Pattern.UNICODE_CASE | Pattern.MULTILINE | Pattern.DOTALL);

                Matcher m = p.matcher(s);

                if (m.find()) {

                    if (m.groupCount() > 1) {

                        encoding = m.group(1);

                    }

            if (encoding == null) {    // no idea?

                encoding = "UTF-8";

        reader = new InputStreamReader(in, encoding);

        // InputSourceImpl constructor with URI recommended

        // so the renderer can resolve page component URLs.

        InputSource is = new InputSourceImpl(reader, uri);

        // This example does not perform incremental rendering.

        document = builder.parse(is);

	private void getDocumentFromUrl(String uri) throws Exception {
		URL url = new URL(uri);
		URLConnection connection = url.openConnection();
		InputStream in = connection.getInputStream();

		// A Reader should be created with the correct charset,
		// which may be obtained from the Content-Type header
		// of an HTTP response.
		Reader reader;
		String encoding = connection.getContentEncoding();
		if (encoding == null) {
			// take a peek in the file then
			byte[] buff = new byte[512];
			int rc = in.read(buff);
			if (rc > 0) {
				// fish out charset from meta tag
				String s = new String(buff);
				Pattern p = Pattern.compile("",
						Pattern.CASE_INSENSITIVE | Pattern.UNICODE_CASE | Pattern.MULTILINE | Pattern.DOTALL);
				Matcher m = p.matcher(s);
				if (m.find()) {
					if (m.groupCount() > 1) {
						encoding = m.group(1);
					}
				}
			}
			if (encoding == null) {	// no idea?
				encoding = "UTF-8";
			}
		}
		reader = new InputStreamReader(in, encoding);

		// InputSourceImpl constructor with URI recommended
		// so the renderer can resolve page component URLs.
		InputSource is = new InputSourceImpl(reader, uri);

		// This example does not perform incremental rendering.
		document = builder.parse(is);

I found the memory leak - at the end of outputImage() you need

Not sure why this works, but it does - this gets called on every scroll tick so gets through memory quite quickly (kaboom!)

(i better send you my updated version - will try...)

rendering HTML in Jitter