在Java中,你可以使用相同的API但使用不同的URL协议加载各种资源:

file:///tmp.txt
http://127.0.0.1:8080/a.properties
jar:http://www.foo.com/bar/baz.jar!/COM/foo/Quux.class

这很好地将资源的实际加载与需要资源的应用程序分离开来,而且由于URL只是一个字符串,资源加载也非常容易配置。

是否存在使用当前类加载器加载资源的协议? 这与Jar协议类似,只是我不需要知道资源来自哪个Jar文件或类文件夹。

当然,我可以使用Class.getResourceAsStream(“a.xml”)做到这一点,但这需要我使用不同的API,因此需要对现有代码进行更改。我希望能够在所有可以为资源指定URL的地方使用它,只需更新属性文件即可。


当前回答

当然,使用注册URLStreamHandlers的解决方案是最正确的,但有时需要最简单的解决方案。所以,我使用下面的方法:

/**
 * Opens a local file or remote resource represented by given path.
 * Supports protocols:
 * <ul>
 * <li>"file": file:///path/to/file/in/filesystem</li>
 * <li>"http" or "https": http://host/path/to/resource - gzipped resources are supported also</li>
 * <li>"classpath": classpath:path/to/resource</li>
 * </ul>
 *
 * @param path An URI-formatted path that points to resource to be loaded
 * @return Appropriate implementation of {@link InputStream}
 * @throws IOException in any case is stream cannot be opened
 */
public static InputStream getInputStreamFromPath(String path) throws IOException {
    InputStream is;
    String protocol = path.replaceFirst("^(\\w+):.+$", "$1").toLowerCase();
    switch (protocol) {
        case "http":
        case "https":
            HttpURLConnection connection = (HttpURLConnection) new URL(path).openConnection();
            int code = connection.getResponseCode();
            if (code >= 400) throw new IOException("Server returned error code #" + code);
            is = connection.getInputStream();
            String contentEncoding = connection.getContentEncoding();
            if (contentEncoding != null && contentEncoding.equalsIgnoreCase("gzip"))
                is = new GZIPInputStream(is);
            break;
        case "file":
            is = new URL(path).openStream();
            break;
        case "classpath":
            is = Thread.currentThread().getContextClassLoader().getResourceAsStream(path.replaceFirst("^\\w+:", ""));
            break;
        default:
            throw new IOException("Missed or unsupported protocol in path '" + path + "'");
    }
    return is;
}

其他回答

URL url = getClass().getClassLoader().getResource("someresource.xxx");

这样就行了。

我认为这是值得回答的问题——如果您正在使用Spring,那么您已经有了这个功能

Resource firstResource =
    context.getResource("http://www.google.fi/");
Resource anotherResource =
    context.getResource("classpath:some/resource/path/myTemplate.txt");

就像春季文档中解释的,斯卡夫曼在评论中指出的那样。

我试图避免URL类,而是依赖URI。因此,对于需要URL的事情,我想做Spring资源查找,我做以下工作:

public static URL toURL(URI u, ClassLoader loader) throws MalformedURLException {
    if ("classpath".equals(u.getScheme())) {
        String path = u.getPath();
        if (path.startsWith("/")){
            path = path.substring("/".length());
        }
        return loader.getResource(path);
    }
    else if (u.getScheme() == null && u.getPath() != null) {
        //Assume that its a file.
        return new File(u.getPath()).toURI().toURL();
    }
    else {
        return u.toURL();
    }
}

要创建一个URI,可以使用URI.create(..)。这种方式也更好,因为您控制了将执行资源查找的ClassLoader。

我注意到其他一些答案试图将URL解析为字符串以检测该方案。我认为最好是传递URI并使用它来解析。

事实上,我在不久前提出了一个问题,要求Spring Source将他们的资源代码从核心中分离出来,这样你就不需要所有其他Spring的东西了。

如果你在类路径中有tomcat,它就像这样简单:

TomcatURLStreamHandlerFactory.register();

这将为“war”和“classpath”协议注册处理程序。

当然,使用注册URLStreamHandlers的解决方案是最正确的,但有时需要最简单的解决方案。所以,我使用下面的方法:

/**
 * Opens a local file or remote resource represented by given path.
 * Supports protocols:
 * <ul>
 * <li>"file": file:///path/to/file/in/filesystem</li>
 * <li>"http" or "https": http://host/path/to/resource - gzipped resources are supported also</li>
 * <li>"classpath": classpath:path/to/resource</li>
 * </ul>
 *
 * @param path An URI-formatted path that points to resource to be loaded
 * @return Appropriate implementation of {@link InputStream}
 * @throws IOException in any case is stream cannot be opened
 */
public static InputStream getInputStreamFromPath(String path) throws IOException {
    InputStream is;
    String protocol = path.replaceFirst("^(\\w+):.+$", "$1").toLowerCase();
    switch (protocol) {
        case "http":
        case "https":
            HttpURLConnection connection = (HttpURLConnection) new URL(path).openConnection();
            int code = connection.getResponseCode();
            if (code >= 400) throw new IOException("Server returned error code #" + code);
            is = connection.getInputStream();
            String contentEncoding = connection.getContentEncoding();
            if (contentEncoding != null && contentEncoding.equalsIgnoreCase("gzip"))
                is = new GZIPInputStream(is);
            break;
        case "file":
            is = new URL(path).openStream();
            break;
        case "classpath":
            is = Thread.currentThread().getContextClassLoader().getResourceAsStream(path.replaceFirst("^\\w+:", ""));
            break;
        default:
            throw new IOException("Missed or unsupported protocol in path '" + path + "'");
    }
    return is;
}