Class org.mozilla.javascript.MemberBox cannot access class sun.net.www.protocol.http.HttpURLConnection

[Coooz]

Hi,

Referring to https://github.com/nextgenhealthcare/connect/issues/6254,
I have the following code that checks whether a REST API interface is available and which logs the exact same error (except my unnamed module is @30c31dd7).

function check_http(url) {
	//create url
	var url_check = new java.net.URL(url);
	
	//open connection to url
	var conn = url_check.openConnection();
	try{
		conn.connect();
		logger.info("HTTP Server " + url + " available.");
	}
	catch(err){
		logger.error("HTTP Server " + url + " NOT available.");
		logger.error(err.toString());
	}
	
	return conn;
}

The "Add a VM flag" - to mcserver.options and mcservice.options - workaround did not work, and I am not entirely sure what else to do to solve the issue. If someone can help me along here then that would be greatly appreciated. I use Mirth Connect 4.4.2.

Thanks in advance!

[rjkroll]

Because of Java 9 and its access restrictions, and the fact Mirth uses Rhino, there are issues when trying to access an internal Java class that is not meant to be directly access by external code. Java recommends that you try using java.net.HttpUrlConnection and its API to get around this.

EDIT: This solution was generated by AI after I ran through the function and error: https://www.perplexity.ai/search/here-is-my-code-in-mirth-conne-MWZBQtutR4O6K2Upzf6NfQ

You could maybe try this:

function check_http(url, timeout) {
    var url_check = new java.net.URL(url);
    var conn = null;
    
    try {
        conn = url_check.openConnection();
        
        if (conn instanceof java.net.HttpURLConnection) {
            conn.setConnectTimeout(timeout || 5000); // 5 seconds default timeout
            conn.connect();
            logger.info("HTTP Server " + url + " available.");
            return true;
        } else {
            logger.error("URL " + url + " is not an HTTP URL.");
            return false;
        }
    } catch (err) {
        if (err instanceof java.net.ConnectException) {
            logger.error("HTTP Server " + url + " connection refused.");
        } else if (err instanceof java.net.UnknownHostException) {
            logger.error("HTTP Server " + url + " unknown host.");
        } else {
            logger.error("HTTP Server " + url + " NOT available.");
        }
        logger.error(err.toString());
        return false;
    } finally {
        if (conn && conn instanceof java.net.HttpURLConnection) {
            conn.disconnect();
        }
    }
}

Along with the VMOptions possibly of :

--add-opens java.base/sun.net.www.protocol.http=ALL-UNNAMED
--add-opens java.base/sun.net.www.protocol.https=ALL-UNNAMED

[pacmano1]

@rjkroll a few of your posts are looking a bit chatGPT/AI generated. That's fine if course since using tools at our disposal is always smart - but maybe attribute your solution as such if that is true or perhaps post a direct link to the chatGPT discussion. That can be helpful in the sense it shows how people work through problems using AI.

[rjkroll]

Sorry, you are correct and I do use AI to see if what possible solutions may be, and I know some may not work right away, but just trying to point people in the right direction. In this case, I just pasted in the function and error and what I posted was what it suggested. Would be interesting to see if it actually works, but here is the link to the conversation and I will update the original post.

https://www.perplexity.ai/search/here-is-my-code-in-mirth-conne-MWZBQtutR4O6K2Upzf6NfQ

[pacmano1]

No need to apologize. I would say cutting and pasting into an AI tool and pasting its answer here and representing that as your own and more importantly without testing it isn't really the greatest way to help folks. If you test it and it works that is a big help to folks AI or not.

[Coooz]

@rjkroll Yes, this works absolutely fine. The vmoptions you specified are indeed necessary. Thank you!

[odoodo]

An issue was already filed for this bug here

It seems to be fixed in the latest rhino version.

I worked around it by using the Apache http client instead.