xmllint doesn't work with https - warning: failed to load external entity

Question

xmllint does work fine with http://somesite.xml

But it doesn't work with https://somesite.xml

xmllint https://somesite.xml
warning: failed to load external entity "https://somesite.xml"

Dobbin

I seem to be having the same problem. – Fixity

The same. xmllint doesn't works with DTD on HTTPS. – Fincher

If the schema you're using a remote DTD the only workaround I've been able to figure out was to download the DTD locally and replace the schema file with something like: schemaLocation="file:///tmp/my-path.xsd" – Rennes

"xmllint doesn't support HTTPS and probably never will." according to gitlab.gnome.org/GNOME/libxml2/-/issues/160 – Pantsuit

Answer 1

As a workaround, you could use another utility like curl or wget to download the file first, then pipe it to xmllint.

curl --silent "https://somesite.xml" | xmllint -

Notes:

• Use - ("hyphen/minus") for xmllint's filename argument to get its XML input from the standard input stream instead of from a file or URL.
• You might want to use --silent (-s) to suppress curl progress/error messages, to prevent those from being parsed by xmllint.
• Quotes might be required around the URL if it contains special characters.

This should work for xmllint's XML input over HTTPS, but not sure about a DTD or schema; you might need to download that to a local file first, using a separate curl or wget command.

Granthem

This workaround does not really work for XML schema validation with schemas that have imports using HTTPS in their schemaLocation. Only one single schema file can be provided using the --schema parameter. The only workaround for that would be to download all imported schemas and modify their schemaLocation to relative URLs. – Pantsuit

Answer 2

It is not directly a replacement for xmllint as a command-line tool, but the VS Code XML plugin supports automatic downloading of XML schemas via HTTPS.

Pantsuit