Looking at the code that performs jar names filtering in StandardJarScanner in 8.0.5, 7.0.53.

This is in view of
a) backporting it to Tomcat 6 to get rid of ecj-xxx.jar names in the lists of ignored files there,
b) documenting what filename patterns are supported here.

Essentially, the code calls "Matcher.matchPath(toSkip, jarName)" (see StandardJarScanFilter#check() in 8.0.5) where jarName is just a file name.

This is an overkill. The matchPath() method is designed to work with Ant-style patterns matching against file paths. This method handles file names hierarchy (directory names and "**" directory name pattern).  The actual matching of the names is performed by Matcher.match(String pattern, String str, boolean caseSensitive), which is a rather more simple API.

Thus the proposal is to get rid of Matcher.matchPath() and call Matcher.match() directly. This allows to remove a lot of unneeded code.

For documentation, the file name patterns supported by Matcher.match() according to its javadoc are as follows:
 * The pattern may contain two special characters:<br>
 * '*' means zero or more characters<br>
 * '?' means one and only one character