Java Language What the classpath means: how searches work


The purpose of the classpath is to tell a JVM where to find classes and other resources. The meaning of the classpath and the search process are intertwined.

The classpath is a form of search path which specifies a sequence of locations to look for resources. In a standard classpath, these places are either, a directory in the host file system, a JAR file or a ZIP file. In each cases, the location is the root of a namespace that will be searched.

The standard procedure for searching for a class on the classpath is as follows:

  1. Map the class name to a relative classfile pathname RP. The mapping for class names to class filenames is described elsewhere.

  2. For each entry E in the classpath:

    • If the entry is a filesystem directory:
      • Resolve RP relative to E to give an absolute pathname AP.
      • Test if AP is a path for an existing file.
      • If yes, load the class from that file
    • If the entry is a JAR or ZIP file:
      • Lookup RP in the JAR / ZIP file index.
      • If the corresponding JAR / ZIP file entry exists, load the class from that entry.

The procedure for searching for a resource on the classpath depends on whether the resource path is absolute or relative. For an absolute resource path, the procedure is as above. For a relative resource path resolved using Class.getResource or Class.getResourceAsStream, the path for the classes package is prepended prior to searching.

(Note these are the procedures implemented by the standard Java classloaders. A custom classloader might perform the search differently.)