<?xml version="1.0"?>
<!--
 Licensed to the Apache Software Foundation (ASF) under one or more
 contributor license agreements.  See the NOTICE file distributed with
 this work for additional information regarding copyright ownership.
 The ASF licenses this file to You under the Apache License, Version 2.0
 (the "License"); you may not use this file except in compliance with
 the License.  You may obtain a copy of the License at

      http://www.apache.org/licenses/LICENSE-2.0

 Unless required by applicable law or agreed to in writing, software
 distributed under the License is distributed on an "AS IS" BASIS,
 WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License.
-->

<document xmlns="http://maven.apache.org/XDOC/2.0"
  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  xsi:schemaLocation="http://maven.apache.org/XDOC/2.0 http://maven.apache.org/xsd/xdoc-2.0.xsd">

  <properties>
    <title>Using FileUpload</title>
    <author email="martinc@apache.org">Martin Cooper</author>
  </properties>

  <body>
    <section name="Using FileUpload">
      <p>
        FileUpload can be used in a number of different ways, depending upon the
        requirements of your application. In the simplest case, you will call a
        single method to parse the servlet request, and then process the list of
        items as they apply to your application. At the other end of the scale,
        you might decide to customize FileUpload to take full control of the way
        in which individual items are stored; for example, you might decide to
        stream the content into a database.
      </p>
      <p>
        Here, we will describe the basic principles of FileUpload, and illustrate
        some of the simpler - and most common - usage patterns. Customization of
        FileUpload is described <a href="customizing.html">elsewhere</a>.
      </p>
      <p>
         FileUpload depends on Commons IO, so make sure you have the version 
         mentioned on the <a href="dependencies.html">dependencies page</a> in 
         your classpath before continuing.
      </p>
    </section>

    <section name="How it works">
      <p>
        A file upload request comprises an ordered list of <em>items</em> that
        are encoded according to
        <a href="http://www.ietf.org/rfc/rfc1867.txt">RFC 1867</a>,
        "Form-based File Upload in HTML". FileUpload can parse such a request
        and provide your application with a list of the individual uploaded
        items. Each such item implements the <code>FileItem</code> interface,
        regardless of its underlying implementation.
      </p>
      <p>
        This page describes the traditional API of the commons fileupload
        library. The traditional API is a convenient approach. However, for
        ultimate performance, you might prefer the faster
        <a href="streaming.html">Streaming API</a>.
      </p>
      <p>
        Each file item has a number of properties that might be of interest for
        your application. For example, every item has a name and a content type,
        and can provide an <code>InputStream</code> to access its data. On the
        other hand, you may need to process items differently, depending upon
        whether the item is a regular form field - that is, the data came from
        an ordinary text box or similar HTML field - or an uploaded file. The
        <code>FileItem</code> interface provides the methods to make such a
        determination, and to access the data in the most appropriate manner.
      </p>
      <p>
        FileUpload creates new file items using a <code>FileItemFactory</code>.
        This is what gives FileUpload most of its flexibility. The factory has
        ultimate control over how each item is created. The factory implementation
        that currently ships with FileUpload stores the item's data in memory or
        on disk, depending on the size of the item (i.e. bytes of data). However,
        this behavior can be customized to suit your application.
      </p>
    </section>

    <section name="Servlets and Portlets">
      <p>
        Starting with version 1.1, FileUpload supports file upload requests in
        both servlet and portlet environments. The usage is almost identical in
        the two environments, so the remainder of this document refers only to
        the servlet environment.
      </p>
      <p>
        If you are building a portlet application, the following are the two
        distinctions you should make as you read this document:</p>
        <ul>
          <li>
            Where you see references to the <code>ServletFileUpload</code> class,
            substitute the <code>PortletFileUpload</code> class.
          </li>
          <li>
            Where you see references to the <code>HttpServletRequest</code> class,
            substitute the <code>ActionRequest</code> class.
          </li>
        </ul>
    </section>

    <section name="Parsing the request">
      <p>
        Before you can work with the uploaded items, of course, you need to parse
        the request itself. Ensuring that the request is actually a file upload
        request is straightforward, but FileUpload makes it simplicity itself, by
        providing a static method to do just that.
      </p>
  <source><![CDATA[// Check that we have a file upload request
boolean isMultipart = ServletFileUpload.isMultipartContent(request);]]></source>
        <p>
          Now we are ready to parse the request into its constituent items.
        </p>
      <subsection name="The simplest case">
        <p>The simplest usage scenario is the following:</p>
        <ul>
          <li>
            Uploaded items should be retained in memory as long as they are
            reasonably small.
          </li>
          <li>
            Larger items should be written to a temporary file on disk.
          </li>
          <li>
            Very large upload requests should not be permitted.
          </li>
          <li>
            The built-in defaults for the maximum size of an item to
            be retained in memory, the maximum permitted size of an upload
            request, and the location of temporary files are acceptable.
          </li>
        </ul>
        <p>
          Handling a request in this scenario couldn't be much simpler:
        </p>
  <source><![CDATA[// Create a factory for disk-based file items
DiskFileItemFactory factory = new DiskFileItemFactory();

// Configure a repository (to ensure a secure temp location is used)
ServletContext servletContext = this.getServletConfig().getServletContext();
File repository = (File) servletContext.getAttribute("javax.servlet.context.tempdir");
factory.setRepository(repository);

// Create a new file upload handler
ServletFileUpload upload = new ServletFileUpload(factory);

// Parse the request
List<FileItem> items = upload.parseRequest(request);]]></source>
        <p>
          That's all that's needed. Really!
        </p>
        <p>
          The result of the parse is a <code>List</code> of file items, each of
          which implements the <code>FileItem</code> interface. Processing these
          items is discussed below.
        </p>
      </subsection>

      <subsection name="Exercising more control"> 
        <p>
          If your usage scenario is close to the simplest case, described above,
          but you need a little more control, you can easily customize the
          behavior of the upload handler or the file item factory or both. The
          following example shows several configuration options:
        </p>
<source><![CDATA[// Create a factory for disk-based file items
DiskFileItemFactory factory = new DiskFileItemFactory();

// Set factory constraints
factory.setSizeThreshold(yourMaxMemorySize);
factory.setRepository(yourTempDirectory);

// Create a new file upload handler
ServletFileUpload upload = new ServletFileUpload(factory);

// Set overall request size constraint
upload.setSizeMax(yourMaxRequestSize);

// Parse the request
List<FileItem> items = upload.parseRequest(request);]]></source>
        <p>
          Of course, each of the configuration methods is independent of the
          others, but if you want to configure the factory all at once, you can
          do that with an alternative constructor, like this:
        </p>
  <source><![CDATA[// Create a factory for disk-based file items
DiskFileItemFactory factory = new DiskFileItemFactory(yourMaxMemorySize, yourTempDirectory);]]></source>
        <p>
          Should you need further control over the parsing of the request, such
          as storing the items elsewhere - for example, in a database - you will
          need to look into <a href="customizing.html">customizing</a> FileUpload.
        </p>
      </subsection>
    </section>

    <section name="Processing the uploaded items">
      <p>
        Once the parse has completed, you will have a <code>List</code> of file
        items that you need to process. In most cases, you will want to handle
        file uploads differently from regular form fields, so you might process
        the list like this:
      </p>
  <source><![CDATA[// Process the uploaded items
Iterator<FileItem> iter = items.iterator();
while (iter.hasNext()) {
    FileItem item = iter.next();

    if (item.isFormField()) {
        processFormField(item);
    } else {
        processUploadedFile(item);
    }
}]]></source>
      <p>
        For a regular form field, you will most likely be interested only in the
        name of the item, and its <code>String</code> value. As you might expect,
        accessing these is very simple.
      </p>
  <source><![CDATA[// Process a regular form field
if (item.isFormField()) {
    String name = item.getFieldName();
    String value = item.getString();
    ...
}]]></source>
      <p>
        For a file upload, there are several different things you might want to
        know before you process the content. Here is an example of some of the
        methods you might be interested in.
      </p>
  <source><![CDATA[// Process a file upload
if (!item.isFormField()) {
    String fieldName = item.getFieldName();
    String fileName = item.getName();
    String contentType = item.getContentType();
    boolean isInMemory = item.isInMemory();
    long sizeInBytes = item.getSize();
    ...
}]]></source>
      <p>
        With uploaded files, you generally will not want to access them via
        memory, unless they are small, or unless you have no other alternative.
        Rather, you will want to process the content as a stream, or write the
        entire file to its ultimate location. FileUpload provides simple means of
        accomplishing both of these.
      </p>
  <source><![CDATA[// Process a file upload
if (writeToFile) {
    File uploadedFile = new File(...);
    item.write(uploadedFile);
} else {
    InputStream uploadedStream = item.getInputStream();
    ...
    uploadedStream.close();
}]]></source>
      <p>
        Note that, in the default implementation of FileUpload, <code>write()</code>
        will attempt to rename the file to the specified destination, if the data
        is already in a temporary file. Actually copying the data is only done if
        the the rename fails, for some reason, or if the data was in memory.
      </p>
      <p>
        If you do need to access the uploaded data in memory, you need simply
        call the <code>get()</code> method to obtain the data as an array of
        bytes.
      </p>
  <source><![CDATA[// Process a file upload in memory
byte[] data = item.get();
...]]></source>
    </section>

    <section name="Resource cleanup">
      <p>
        This section applies only, if you are using the
        <a href="./apidocs/org/apache/commons/fileupload/disk/DiskFileItem.html">DiskFileItem</a>.
        In other words, it applies, if your uploaded files are written to
        temporary files before processing them.
      </p>
      <p>
        Such temporary files are deleted automatically, if they are no longer
        used (more precisely, if the corresponding instance of <code>DiskFileItem</code>
        is garbage collected. This is done silently by the <code>org.apache.commons.io.FileCleanerTracker</code>
        class, which starts a reaper thread.
      </p>
      <p>
        This reaper thread should be stopped, if it is no longer needed. In
        a servlet environment, this is done by using a special servlet
        context listener, called
        <a href="./apidocs/org/apache/commons/fileupload/servlet/FileCleanerCleanup.html">FileCleanerCleanup</a>.
        To do so, add a section like the following to your <code>web.xml</code>:
      </p>
  <source><![CDATA[<web-app>
  ...
  <listener>
    <listener-class>
      org.apache.commons.fileupload.servlet.FileCleanerCleanup
    </listener-class>
  </listener>
  ...
</web-app>]]></source>

      <subsection name="Creating a DiskFileItemFactory">
        <p>
          The FileCleanerCleanup provides an instance of
          <code>org.apache.commons.io.FileCleaningTracker</code>. This
          instance must be used when creating a
          <code>org.apache.commons.fileupload.disk.DiskFileItemFactory</code>.
          This should be done by calling a method like the following:
        </p>
  <source><![CDATA[public static DiskFileItemFactory newDiskFileItemFactory(ServletContext context,
                                                         File repository) {
    FileCleaningTracker fileCleaningTracker
        = FileCleanerCleanup.getFileCleaningTracker(context);
    DiskFileItemFactory factory
        = new DiskFileItemFactory(DiskFileItemFactory.DEFAULT_SIZE_THRESHOLD,
                                  repository);
    factory.setFileCleaningTracker(fileCleaningTracker);
    return factory;
}]]></source>
      </subsection>

      <subsection name="Disabling cleanup of temporary files">
        <p>
          To disable tracking of temporary files, you may set the
          <code>FileCleaningTracker</code> to null. Consequently,
          created files will no longer be tracked. In particular,
          they will no longer be deleted automatically.</p>
      </subsection>
    </section>

    <section name="Interaction with virus scanners">
      <p>
        Virus scanners running on the same system as the web container can cause
        some unexpected behaviours for applications using FileUpload. This section
        describes some of the behaviours that you might encounter, and provides
        some ideas for how to handle them.
      </p>
      <p>
        The default implementation of FileUpload will cause uploaded items above
        a certain size threshold to be written to disk. As soon as such a file is
        closed, any virus scanner on the system will wake up and inspect it, and
        potentially quarantine the file - that is, move it to a special location
        where it will not cause problems. This, of course, will be a surprise to
        the application developer, since the uploaded file item will no longer be
        available for processing. On the other hand, uploaded items below that
        same threshold will be held in memory, and therefore will not be seen by
        virus scanners. This allows for the possibility of a virus being retained
        in some form (although if it is ever written to disk, the virus scanner
        would locate and inspect it).
      </p>
      <p>
        One commonly used solution is to set aside one directory on the system
        into which all uploaded files will be placed, and to configure the virus
        scanner to ignore that directory. This ensures that files will not be
        ripped out from under the application, but then leaves responsibility for
        virus scanning up to the application developer. Scanning the uploaded
        files for viruses can then be performed by an external process, which
        might move clean or cleaned files to an "approved" location, or by
        integrating a virus scanner within the application itself. The details of
        configuring an external process or integrating virus scanning into an
        application are outside the scope of this document.
      </p>
    </section>

    <section name="Watching progress">
      <p>
        If you expect really large file uploads, then it would be nice to report
        to your users, how much is already received. Even HTML pages allow to
        implement a progress bar by returning a multipart/replace response,
        or something like that.
      </p>
      <p>
        Watching the upload progress may be done by supplying a progress listener:
      </p>
  <source><![CDATA[//Create a progress listener
ProgressListener progressListener = new ProgressListener(){
   public void update(long pBytesRead, long pContentLength, int pItems) {
       System.out.println("We are currently reading item " + pItems);
       if (pContentLength == -1) {
           System.out.println("So far, " + pBytesRead + " bytes have been read.");
       } else {
           System.out.println("So far, " + pBytesRead + " of " + pContentLength
                              + " bytes have been read.");
       }
   }
};
upload.setProgressListener(progressListener);]]></source>
      <p>
        Do yourself a favour and implement your first progress listener just
        like the above, because it shows you a pitfall: The progress listener
        is called quite frequently. Depending on the servlet engine and other
        environment factory, it may be called for any network packet! In
        other words, your progress listener may become a performance problem!
        A typical solution might be, to reduce the progress listeners activity.
        For example, you might emit a message only, if the number of megabytes
        has changed:
      </p>
  <source><![CDATA[//Create a progress listener
ProgressListener progressListener = new ProgressListener(){
   private long megaBytes = -1;
   public void update(long pBytesRead, long pContentLength, int pItems) {
       long mBytes = pBytesRead / 1000000;
       if (megaBytes == mBytes) {
           return;
       }
       megaBytes = mBytes;
       System.out.println("We are currently reading item " + pItems);
       if (pContentLength == -1) {
           System.out.println("So far, " + pBytesRead + " bytes have been read.");
       } else {
           System.out.println("So far, " + pBytesRead + " of " + pContentLength
                              + " bytes have been read.");
       }
   }
};]]></source>
    </section>

    <section name="What's next">
      <p>
        Hopefully this page has provided you with a good idea of how to use
        FileUpload in your own applications. For more detail on the methods
        introduced here, as well as other available methods, you should refer
        to the <a href="./apidocs/index.html">Javadocs</a>.
      </p>
      <p>
        The usage described here should satisfy a large majority of file upload
        needs. However, should you have more complex requirements, FileUpload
        should still be able to help you, with it's flexible
        <a href="./customizing.html">customization</a> capabilities.
      </p>
    </section>
  </body>

</document>