Guava Files 源码分析(一)

Files中的工厂

Files类中对InputStream, OutputStream以及Reader,Writer的操作封装了抽象工厂模式,抽象工厂是InputSupplier与OutputSupplier,具体工厂是Files中的newInputStreamSupplier(), newOutputStreamSupplier()等方法

而InputStream, OutputStream以及Reader,Writer则是抽象产品, 他们的各种实现和装饰器包装则为具体产品

 

Input与Output工厂

Files中将Input与Output(包括InputStream,OutputStream和Reader,Writer两大IO派系)的生成使用抽象工厂来实现,其中代表抽象工厂的两个类为

public interface InputSupplier<T> {

  T getInput() throws IOException;
}

public interface InputSupplier<T> {

  T getInput() throws IOException;
}

在某些类的对Input和Output操作的工具方法则会使用这个Supplier作为形参,例如

CharStreams.newReaderSupplier(

com.google.common.io.InputSupplier<? extends java.io.InputStream> in,

java.nio.charset.Charset charset

)

这样调用CharStreams.newReaderSupplier()便可以通过传入不同的Supplier具体实现来达到多态和解耦

Files类中的具体工厂实现

Files中工厂的具体实现抽象为了一个方法,如下:

public static InputSupplier<FileInputStream> newInputStreamSupplier(
            final File file) {
        Preconditions.checkNotNull(file);
        return new InputSupplier<FileInputStream>() {
            @Override
            public FileInputStream getInput() throws IOException {
                return new FileInputStream(file);
            }
        };
    }


    public static OutputSupplier<FileOutputStream> newOutputStreamSupplier(
            File file) {
        return newOutputStreamSupplier(file, false);
    }


    public static OutputSupplier<FileOutputStream> newOutputStreamSupplier(
            final File file, final boolean append) {
        Preconditions.checkNotNull(file);
        return new OutputSupplier<FileOutputStream>() {
            @Override
            public FileOutputStream getOutput() throws IOException {
                return new FileOutputStream(file, append);
            }
        };
    }

    public static InputSupplier<InputStreamReader> newReaderSupplier(File file,
                                                                     Charset charset) {
        return CharStreams.newReaderSupplier(newInputStreamSupplier(file), charset);
    }


    public static OutputSupplier<OutputStreamWriter> newWriterSupplier(File file,
                                                                       Charset charset) {
        return newWriterSupplier(file, charset, false);
    }


    public static OutputSupplier<OutputStreamWriter> newWriterSupplier(File file,
                                                                       Charset charset) {
        return newWriterSupplier(file, charset, false);
    }

这些工厂返回的产品大致包括了FileInputStream, FileOutputStream, Reader, Writer

文件读取与写入函数

    /**
     * 这里有意思的有些特殊文件是file.length() == 0,但是文件却有实际内容
     * 所以不能直接开辟一个file.length()大小的byte[] buff来读取文件内容
     * Guava的解决方法是通过一个buff[0x1000]大小的buff来逐步读取这个特殊的文件
     * 将其写入到一个OutputStream以后再一次性将它out.toByteArray()并返回
     *
     * 而对于file.length() != 0 的情况,则是直接开辟一个byte[] buff[file.length()]大小的buff
     * 一次性将file里的内容读取到buff中并返回,从而避免额外像读取上面说的特殊文件那样频繁的开辟小的byte[] buff
     */
public static byte[] toByteArray(File file) throws IOException {
        Preconditions.checkArgument(file.length() <= Integer.MAX_VALUE);
        if (file.length() == 0) {
            // Some special files are length 0 but have content nonetheless.
            return ByteStreams.toByteArray(newInputStreamSupplier(file));
        } else {
            // Avoid an extra allocation and copy.
            byte[] b = new byte[(int) file.length()];
            boolean threw = true;
            InputStream in = new FileInputStream(file);
            try {
                ByteStreams.readFully(in, b);
                threw = false;
            } finally {
                Closeables.close(in, threw);
            }
            return b;
        }
    }


    /**
     * 通过toByteArray()方法将文件内容包装成字符串
     */
    public static String toString(File file, Charset charset) throws IOException {
        return new String(toByteArray(file), charset.name());
    }


    /**
     * 使用小buff byte[0x1000] 的方法来copy文件
     * 因为直接使用InputStream不像File.length()那样可以直接获得长度
     */
    public static void copy(InputSupplier<? extends InputStream> from, File to)
            throws IOException {
        ByteStreams.copy(from, newOutputStreamSupplier(to));
    }

另外还有很多关于文件读取和写入函数的重载方法,实现方式大同小异,只是参数变化了一下

文件比较方法

/**
     * 依旧是那个问题,有些特殊文件显示的文件长度为0,所以必须通过读取文件的byte内容去比较是否相等
     */
    public static boolean equal(File file1, File file2) throws IOException {
        if (file1 == file2 || file1.equals(file2)) {
            return true;
        }

    /*
     * Some operating systems may return zero as the length for files
     * denoting system-dependent entities such as devices or pipes, in
     * which case we must fall back on comparing the bytes directly.
     */
        long len1 = file1.length();
        long len2 = file2.length();
        if (len1 != 0 && len2 != 0 && len1 != len2) {
            return false;
        }
        return ByteStreams.equal(newInputStreamSupplier(file1),
                newInputStreamSupplier(file2));
    }

临时目录的创建

    /**
     * Atomically creates a new directory somewhere beneath the system's
     * temporary directory (as defined by the {@code java.io.tmpdir} system
     * property), and returns its name.
     *
     * <p>Use this method instead of {@link File#createTempFile(String, String)}
     * when you wish to create a directory, not a regular file.  A common pitfall
     * is to call {@code createTempFile}, delete the file and create a
     * directory in its place, but this leads a race condition which can be
     * exploited to create security vulnerabilities, especially when executable
     * files are to be written into the directory.
     *
     * <p>This method assumes that the temporary volume is writable, has free
     * inodes and free blocks, and that it will not be called thousands of times
     * per second.
     *
     * @return the newly-created directory
     * @throws IllegalStateException if the directory could not be created
     */
    public static File createTempDir() {
        File baseDir = new File(System.getProperty("java.io.tmpdir"));
        String baseName = System.currentTimeMillis() + "-";

        for (int counter = 0; counter < TEMP_DIR_ATTEMPTS; counter++) {
            File tempDir = new File(baseDir, baseName + counter);
            if (tempDir.mkdir()) {
                return tempDir;
            }
        }
        throw new IllegalStateException("Failed to create directory within "
                + TEMP_DIR_ATTEMPTS + " attempts (tried "
                + baseName + "0 to " + baseName + (TEMP_DIR_ATTEMPTS - 1) + ')');
    }

Guava Doc中说到,如果直接使用File.createTempFile()会有安全问题,先介绍一下createTempFile()

public static java.io.File createTempFile(java.lang.String prefix,
                                          java.lang.String suffix,
                                          java.io.File directory)
                                  throws java.io.IOException

它的功能是创建临时文件,提供prefix和suffix以及创建的dir,文件名会使用prefix+random+suffix的形式构成,而中间的random随机数则是使LazyInitialization.random.nextLong()生成,最后使用递归的方式生成这个临时文件

createTempFile()方法的返回值是生成文件以后的抽象路径File对象

根据doc的提示,应该使用 deleteOnExit()方法来删除临时文件,deleteOnExit()的意思是在JVM停止时才删除这个文件.相当于缓存了删除命令,如果有多个文件有deleteOnExit(),他们在JVM停止时会逆序开始删除(最后调用deleteOnExit()方法的文件最先删除,类似于栈)

所谓安全问题, StackOverflow上有人提到了,加入使用createTempFile()来创建临时目录,会这么写

public static File createTempDirectory()
    throws IOException
{
    final File temp;

    temp = File.createTempFile("temp", Long.toString(System.nanoTime()));

    if(!(temp.delete()))
    {
        throw new IOException("Could not delete temp file: " + temp.getAbsolutePath());
    }

    if(!(temp.mkdir()))
    {
        throw new IOException("Could not create temp directory: " + temp.getAbsolutePath());
    }

    return (temp);
}

然后参见几位网友的讨论

 按照Sarel Botha的说法,使用了 file.mkdir() 的返回值做判断后抛出异常是不会有问题的

而如果没有使用file.mkdir()的返回值做判断, 潜在的问题是在于Linux的tmp目录使用了sticky bit(只能删除用户自己创建的文件), 在多线程情况下会存在file.delete()和file.mkdir()操作的线程竞争问题.具体是什么我也没搞清楚

原帖地址: http://stackoverflow.com/questions/617414/create-a-temporary-directory-in-java

随意总而言之, 就是使用Guava提供的createTmpDir()更安全

未完待续 ...