Recipe 18.6: Saving and Reusing HTML Output

Problem

To improve the performance of pages that rarely change, you want to capture the output of those pages and save it for reuse when the page is requested.

Solution

Create the page that contains the desired content just as you would any other page, including the server controls you need. At the end of the Page_Load method, use the RenderControl method of the Page control to generate the HTML and then save the HTML to a file.

In the code-behind class for the page, use the .NET language of your choice to:

1. Create an HtmlTextWriter to use for rendering the page.

2. Use the RenderControl method of the Page control to render the output of the page to the HtmlTextWriter.

3. Save the rendered output to a file and redirect to another page.

Example 18-16 and Example 18-17 show the VB and C# code-behind files for our application that demonstrates this solution.

Discussion

Occasionally, it's beneficial to save the HTML output from a generated page. This is commonly done when using the saved HTML can significantly improve web site performance. If the content of a page is static, for example, there is no point in dynamically generating HTML each time the page is requested. Until the advent of ASP.NET, the only way to save the HTML was to use the "Save as Complete Web Page" feature of Internet Explorer or another browser. Although this method does save the HTML, it also copies all of the page images to the local machine and changes the image references to point to the local copies. If you are trying to improve performance by capturing a static copy of the page to use on your web server, this technique does not work very well.

With ASP.NET you can easily capture the HTML exactly as it would be sent to the browser. For our example that illustrates this solution, we have used the page from Recipe 18.3 and added code to the Page_Load method to save the rendered output.

The RenderControl method of the Page control provides the ability to render the output of the page to the HtmlTextWriter passed to the method. Unfortunately, the HtmlTextWriter does not provide any methods for reading the contents, so a little more work is required to access the rendered HTML.

By first creating a StringBuilder and then using it to create a StringWriter, which is then used to create the required HtmlTextWriter, the contents of the HtmlTextWriter are available by way of the original StringBuilder. This works because the underlying storage mechanism for the StringWriter is a StringBuilder; and because the StringWriter (a stream) is used to create the HtmlTextWriter, the RenderControl method is actually writing the rendered output to the StringBuilder. Our example code to accomplish this is shown here:

  renderedOutput = New StringBuilder( )
  strWriter = New StringWriter(renderedOutput)
  tWriter = New HtmlTextWriter(strWriter)


  renderedOutput = new StringBuilder( );
  strWriter = new StringWriter(renderedOutput);
  tWriter = new HtmlTextWriter(strWriter);

After creating the HtmlTextWriter, the RenderControl method of the Page is called to render the HTML for the page:

  Page.RenderControl(tWriter)


  Page.RenderControl(tWriter);

Now that the rendered HTML is available, it needs to be saved to a file on the server. This can be accomplished by creating the file with a FileStream and using a StreamWriter to write the rendered output in the StringBuilder to the file, as shown here:

  filename = Server.MapPath(".") & "\" & OUTPUT_FILENAME
  outputStream = New FileStream(filename, _
                                FileMode.Create)
  sWriter = New StreamWriter(outputStream)
  sWriter.Write(renderedOutput.ToString( ))
  sWriter.Flush( )


  filename = Server.MapPath(".") + "\\" + OUTPUT_FILENAME;
  outputStream = new FileStream(filename,
                                FileMode.Create);
  sWriter = new StreamWriter(outputStream);
  sWriter.Write(renderedOutput.ToString( ));
  sWriter.Flush( );

The last step is to redirect to another page. This is necessary because allowing the page to be displayed would result in an additional rendering and an exception being thrown indicating the page has more than one server-side form element. If you need the page to be displayable anyway, a parameter can be passed in the querystring and checked in the code to determine if the output should be rendered and written to a file or handled normally.

  Response.Redirect([next page])


  Response.Redirect([next page]);

This technique can be used for individual controls in the same manner as for the entire page. For example, if you have a page that contains a DataGrid and you want the rendered HTML for just the DataGrid, you can call the RenderControl method of the DataGrid and then save the output as described earlier.

See Also

Recipe 18.3