XSD download fails with "Server returned HTTP response code: 403"

[Gama11]

<?xml version="1.0" encoding="utf-8"?>
<project 
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:noNamespaceSchemaLocation="https://lime.software/xsd/project-1.0.0.xsd">

</project>

Causes the following error output in the XML output channel:

[Info  - 20:11:06] Feb 22, 2021 08:11:06 org.eclipse.lemminx.uriresolver.CacheResourcesManager lambda$downloadResource$0()
Message: Downloading https://lime.software/xsd/project-1.0.0.xsd to C:\Users\<User>\.lemminx\cache\https\lime.software\xsd\project-1.0.0.xsd...
[Error - 20:11:06] Feb 22, 2021 08:11:06 org.eclipse.lemminx.uriresolver.CacheResourcesManager lambda$downloadResource$0()
Message: Error while downloading https://lime.software/xsd/project-1.0.0.xsd to C:\Users\<User>\.lemminx\cache\https\lime.software\xsd\project-1.0.0.xsd : [java.io.IOException] Server returned HTTP response code: 403 for URL: https://lime.software/xsd/project-1.0.0.xsd
[Info  - 20:11:06] Feb 22, 2021 08:11:06 org.eclipse.lemminx.uriresolver.CacheResourcesManager getResource()
Message: Ignored unavailable schema URI: https://lime.software/xsd/project-1.0.0.xsd

The 403 error response is very curious, as I have no issues viewing https://lime.software/xsd/project-1.0.0.xsd in my browser.
curl https://lime.software/xsd/project-1.0.0.xsd also has no complaints.

---

VSCode 1.53.2, Windows 10 20H2, vscode-xml 0.15.0

[datho7561]

Hello! I can't seem to reproduce the issue. I have a few questions that might help to pin down the issue. Does this happen every time you open the document, or just once? Are you behind a proxy for internet access? Are you running the binary or the Java version of the server?

[Gama11]

Does this happen every time you open the document, or just once?

If I close and reopen the document, it doesn't seem to attempt the download again, it only logs "Ignored unavailable schema...". Reloading VSCode triggers the entire output channel sequence from above.

Are you behind a proxy for internet access?

No.

Are you running the binary or the Java version of the server?

Whatever the default is I guess? I just installed it from the VSCode marketplace and haven't configured anything in settings.json.

[datho7561]

Okay. Thanks for the information. I can't seem to reproduce the issue. Are you using the default settings for vscode-xml? Would you be able to search for "LemMinX Server info:" in the XML Support output, then copy and paste the list of infos?

[Gama11]

Are you using the default settings for vscode-xml?

Yes, as I said.

Would you be able to search for "LemMinX Server info:" in the XML Support output, then copy and paste the list of infos?

LemMinX Server info:
 - Version : 0.15.0
 - Java : C:\Program Files\Amazon Corretto\jdk1.8.0_265\jre
 - VM Version : 25.265-b01
 - Git cb5fba5 - [maven-release-plugin] prepare release 0.15.0

[datho7561]

Okay. Thank you!

[datho7561]

I'm able to reproduce now.

[datho7561]

Under Java 11, I didn't get the bug, but under Java 8 in Windows and Linux I get the same thing.

[Gama11]

That's interesting. I'm not even sure how it ends up using that JDK 8, the one in my PATH is a JDK 14. If I set "xml.java.home" to that it indeed works.

[angelozerr]

Problem comes from with URLConnection use https://github.com/eclipse/lemminx/blob/a54317a7490f35a3fd34a2acd70f94fc5ce289e8/org.eclipse.lemminx/src/main/java/org/eclipse/lemminx/uriresolver/CacheResourcesManager.java#L169

The problem can be reproduced easily with a simple main:

package test;

import java.net.URL;
import java.net.URLConnection;

public class Main {

	public static void main(String[] args) throws Exception {
		String actualURI = "https://lime.software/xsd/project-1.0.0.xsd";
		URL url = new URL(actualURI);
		URLConnection conn = url.openConnection();
		conn.getInputStream();
	}
}

If you execute this code in Java 11 it's working although in Java 1.8 it throws a 403 error.

It's seems that it's a user agent request matter.

In Java 8, the user agent is Java/1.8 which throws a 403 error, although in Java 11 the user agent is Java/11 and it's working. It seems that 1. is not accepted. (Why I don't know?)

You can experiment that with the following code executed in Java 1.8:

package test;

import java.net.URL;
import java.net.URLConnection;

public class Main {

	public static void main(String[] args) throws Exception {
		String actualURI = "https://lime.software/xsd/project-1.0.0.xsd";
		URL url = new URL(actualURI);
		URLConnection conn = url.openConnection();
		conn.addRequestProperty("User-Agent", "Java/1.8"); // throws a 403
		conn.getInputStream();
	}
}

should throw a 403, although the following code should work:

package test;

import java.net.URL;
import java.net.URLConnection;

public class Main {

	public static void main(String[] args) throws Exception {
		String actualURI = "https://lime.software/xsd/project-1.0.0.xsd";
		URL url = new URL(actualURI);
		URLConnection conn = url.openConnection();
		conn.addRequestProperty("User-Agent", "Java/11"); // should work
		conn.getInputStream();
	}
}

A fix could catch the 403 error and recreate the connection by forcing the User Agent like conn.addRequestProperty("User-Agent", "LemMinX");

[fbricon]

No don't catch the 403. always set the user agent to LemMinX. That'll take care of downloading files hosted on cloudflare managed servers

[rgrunber]

curl -A "Java/1.8" -I https://lime.software/xsd/project-1.0.0.xsd
.. 403

:|