We Keep Coding

Sending a PNG using HTTP in C#

I'm developing my own simple webserver in C#, and I have normal html files and subfolders etc in place and working fine... however my issue lies with sending PNG files.
Once the browser has made the HTTP request and my other code has done its magic, I need to send a response header. Here is my code to do this:
else if (URL.EndsWith(".ico") || URL.EndsWith(".png"))
{
if (File.Exists(Program.Start_Directory + URL))
{
byte[] ImgFile = File.ReadAllBytes(Program.Start_Directory + URL);
sw.Write("HTTP/1.0 200 OK\r\n");
sw.Write("Content-Type: image/png\r\n");
sw.Write("Content-Length: " + ImgFile.Length + "\r\n");
sw.Write("\r\n");
sw.Write(ImgFile);
}
else
{
FileNotFoundPage(sw);
}
}
sw in this instance being the StreamWriter for the socket connection to the browser.
When I run this, the browser screen goes black like it usually does when its about to load an image, but no image loads and the spinning loading wheel remains spinny indefinitely.
How can I get this to work? Thanks.

Things get messy when you use a StreamWriter (which is designed to write strings with a specific Encoding to a stream) with binary data.
It looks like you are invoking this overload of StreamWriter.Write with the belief that it writes the bytes verbatim to the output stream. The docs state that this overload actually...
Writes the text representation of an object to the text string or stream by calling the ToString method on that object
You have two options. Carry on using StreamWriter and Flush, then write the binary data directly to the underlying stream:
byte[] ImgFile = File.ReadAllBytes(Program.Start_Directory + URL);
sw.Write("HTTP/1.0 200 OK\r\n");
//...etc
sw.Flush();
sw.BaseStream.Write(ImgFile,0,ImgFile.Length);
or just do everything in bytes:
var sb=new StringBuilder();
sb.AppendLine("HTTP/1.0 200 OK");
//...etc
var headerBytes = Encoding.ASCII.GetBytes(sb.ToString());
Stream str = iGotTheStreamFromSomewhere;
str.Write(headerBytes,0,headerBytes.Length);
str.Write(ImgFile,0,ImgFile.Length);

Using IISHandler to stream a radio?

I'm trying to stream my own online radio on my site. I'm right now focusing on simply using an http handler and html5 to do so.
How can I keep IISHandler constantly running without ever ending for each user? Like a real online radio.

You have directly redirect to any url inside handler provided that url has audio content, you are are intended to put handler to <audio> src. So you can try this.
public void ProcessRequest(HttpContext context)
{
//context.Response.Redirect("http://api.soundcloud.com/tracks/148976759/stream?client_id=201b55a1a16e7c0a122d112590b32e4a");
// you can use above or else below
byte[] content = null;
string fileName = context.Server.MapPath(#"\mp3\") + context.Request["file_name"];
if (File.Exists(fileName))
{
using (FileStream stream = new FileStream(fileName, FileMode.Open))
{
content = new byte[System.Convert.ToInt32(stream.Length)];
stream.Read(content, 0, System.Convert.ToInt32(stream.Length));
context.Response.ContentType = "audio/mp3";
context.Response.OutputStream.Write(content, 0, content.Length);
}
}
}
<audio src="Handler1.ashx" controls></audio>
Regarding IISHandler constantly running? -- When you host this in IIS, any request to handler ( i.e. you will have a html page having audio's src= handler, so end-user hits that html page, handler will be invoked), it will process the request. The similar concept like aspx . They are hosted in IIS, any request to that will be process. Do you get my point here?

The action can't be completed because the file is open in IIS Worker Process

I am uploading a image for a user & saving it on a specified location. The code is working fine in normal scenario i.e. user selects a image & save it, it works.
Challenge occurs when user selects a image( probably wrong image due to mistake) & don't save it, but again selects a new image & then saves it.
This scenario gives me the error :
"The process cannot access the file because it is being used by another process."
When I try to delete image from the location at the time of error, file can't be deleted with message:
"The action can't be completed because the file is open in IIS Worker Process
Close the file and try again."
Code is like this:
try
{
if (!Directory.Exists(folder))
Directory.CreateDirectory(folder);
msf = new MemoryStream();
bytes=FileUpload1.FileBytes;
msf.Write(bytes, 0, bytes.Length);
using (FileStream stream = new FileStream(folder + "/" + filename, FileMode.Create, FileAccess.ReadWrite, FileShare.ReadWrite))
{
//converting any graphic file type to jpeg format
Image img = Image.FromStream(msf);
img.Save(stream, System.Drawing.Imaging.ImageFormat.Jpeg);
msf.WriteTo(stream);
IsuploadSuccess = true;
img.Dispose();
}
}
catch
{
IsuploadSuccess = false;
}
finally
{
if (msf != null)
{
msf.Close();
msf.Dispose();
}
}
I have tried adding "FileAccess.ReadWrite" & "FileShare.ReadWrite" in file stream, but doesn't work either with all options in File stream.
Please help...

Instead of getting solution of problem, I finally change approach to get rid to source of challenge. I change file name with UserID (unique) appended with CurrentDateTime converted to string & store it to temp folder until user save the change. This forms a different file each time for the challenge scenario. After save each file in temp folder created by the user (preceded by unique userid) is deleted & last changes are saved to respective directory.

I had this problem and took me few hours to figure it out.
I my case the Bitmap new(path) had locked the file. I had to dispose of it before exiting the using block.
However it is the method by which I debugged it is important.
My approach in pinpointing the source of the problem was to open the folder where the image was saved and tracing the code line by line. After each line I deleted the file and did Ctrl+Z to put it back. If it gave an error when I tried to delete it, then that is the line that locked it. So I got to that line within 5 min.
using (MemoryStream memoryStream = new())
{
if (System.IO.File.Exists(path))
{
using(Bitmap bitmap = new(path)) <-- this solved it
{
ImageConverter imageConverter = new();
bitmap.Save(memoryStream, ImageFormat.Bmp);
bytes = memoryStream.ToArray();
}
}
}

Using HttpWebRequest with dynamic URI causes "parameter is not valid" in Image.FromStream

I'm trying to obtain an image to encode to a WordML document. The original version of this function used files, but I needed to change it to get images created on the fly with an aspx page. I've adapted the code to use HttpWebRequest instead of a WebClient. The problem is that I don't think the page request is getting resolved and so the image stream is invalid, generating the error "parameter is not valid" when I invoke Image.FromStream.
public string RenderCitationTableImage(string citation_table_id)
{
string image_content = "";
string _strBaseURL = String.Format("http://{0}",
HttpContext.Current.Request.Url.GetComponents(UriComponents.HostAndPort, UriFormat.Unescaped));
string _strPageURL = String.Format("{0}{1}", _strBaseURL,
ResolveUrl("~/Publication/render_citation_chart.aspx"));
string _staticURL = String.Format("{0}{1}", _strBaseURL,
ResolveUrl("~/Images/table.gif"));
string _fullURL = String.Format("{0}?publication_id={1}&citation_table_layout_id={2}",
_strPageURL, publication_id, citation_table_id);
try
{
HttpWebRequest request = (HttpWebRequest)WebRequest.Create(_fullURL);
HttpWebResponse response = (HttpWebResponse)request.GetResponse();
Stream image_stream = response.GetResponseStream();
// Read the image data
MemoryStream ms = new MemoryStream();
int num_read;
byte[] crlf = System.Text.Encoding.Default.GetBytes("\r\n");
byte[] buffer = new byte[1024];
for (num_read = image_stream.Read(buffer, 0, 1024); num_read > 0; num_read = image_stream.Read(buffer, 0, 1024))
{
ms.Write(buffer, 0, num_read);
}
// Base 64 Encode the image data
byte[] image_bytes = ms.ToArray();
string encodedImage = Convert.ToBase64String(image_bytes);
ms.Position = 0;
System.Drawing.Image image_original = System.Drawing.Image.FromStream(ms); // <---error here: parameter is not valid
image_stream.Close();
image_content = string.Format("<w:p>{4}<w:r><w:pict><w:binData w:name=\"wordml://{0}\">{1}</w:binData>" +
"<v:shape style=\"width:{2}px;height:{3}px\">" +
"<v:imagedata src=\"wordml://{0}\"/>" +
"</v:shape>" +
"</w:pict></w:r></w:p>", _word_image_id, encodedImage, 800, 400, alignment.center);
image_content = "<w:br w:type=\"text-wrapping\"/>" + image_content + "<w:br w:type=\"text-wrapping\"/>";
}
catch (Exception ex)
{
return ex.ToString();
}
return image_content;
Using a static URI it works fine. If I replace "staticURL" with "fullURL" in the WebRequest.Create method I get the error. Any ideas as to why the page request doesn't fully resolve?
And yes, the full URL resolves fine and shows an image if I post it in the address bar.

UPDATE:
Just read your updated question. Since you're running into login issues, try doing this before you execute the request:
request.Credentials = CredentialCache.DefaultCredentials
If this doesn't work, then perhaps the problem is that authentication is not being enforced on static files, but is being enforced on dynamic files. In this case, you'll need to log in first (using your client code) and retain the login cookie (using HttpWebRequest.CookieContainer on the login request as well as on the second request) or turn off authentication on the page you're trying to access.
ORIGINAL:
Since it works with one HTTP URL and doesn't work with another, the place to start diagnosing this is figuring out what's different between the two requests, at the HTTP level, which accounts for the difference in behavior in your code.
To figure out the difference, I'd use Fiddler (http://fiddlertool.com) to compare the two requests. Compare the HTTP headers. Are they the same? In particular, are they the same HTTP content type? If not, that's likely the source of your problem.
If headers are the same, make sure both the static and dynamic image are exactly the same content and file type on the server. (e.g. use File...Save As to save the image in a browser to your disk). Then use Fiddler's Hex View to compare the image content. Can you see any obvious differences?
Finally, I'm sure you've already checked this, but just making sure: /Publication/render_citation_chart.aspx refers to an actual image file, not an HTML wrapper around an IMG element, right? This would account for the behavior you're seeing, where a browser renders the image OK but your code doesn't.

C# - WebRequest Doesn't Return Different Pages

Here's the purpose of my console program: Make a web request > Save results from web request > Use QueryString to get next page from web request > Save those results > Use QueryString to get next page from web request, etc.
So here's some pseudocode for how I set the code up.
for (int i = 0; i < 3; i++)
{
strPageNo = Convert.ToString(i);
//creates the url I want, with incrementing pages
strURL = "http://www.website.com/results.aspx?page=" + strPageNo;
//makes the web request
wrGETURL = WebRequest.Create(strURL);
//gets the web page for me
objStream = wrGETURL.GetResponse().GetResponseStream();
//for reading web page
objReader = new StreamReader(objStream);
//--------
// -snip- code that saves it to file, etc.
//--------
objStream.Close();
objReader.Close();
//so the server doesn't get hammered
System.Threading.Thread.Sleep(1000);
}
Pretty simple, right? The problem is, even though it increments the page number to get a different web page, I'm getting the exact same results page each time the loop runs.
i IS incrementing correctly, and I can cut/paste the url strURL creates into a web browser and it works just fine.
I can manually type in &page=1, &page=2, &page=3, and it'll return the correct pages. Somehow putting the increment in there screws it up.
Does it have anything to do with sessions, or what? I make sure I close both the stream and the reader before it loops again...

Have you tried creating a new WebRequest object for each time during the loop, it could be the Create() method isn't adequately flushing out all of its old data.
Another thing to check is that the ResponseStream is adequately flushed out before the next loop iteration.

This code works fine for me:
var urls = new [] { "http://www.google.com", "http://www.yahoo.com", "http://www.live.com" };
foreach (var url in urls)
{
WebRequest request = WebRequest.Create(url);
using (Stream responseStream = request.GetResponse().GetResponseStream())
using (Stream outputStream = new FileStream("file" + DateTime.Now.Ticks.ToString(), FileMode.Create, FileAccess.Write, FileShare.None))
{
const int chunkSize = 1024;
byte[] buffer = new byte[chunkSize];
int bytesRead;
while ((bytesRead = responseStream.Read(buffer, 0, buffer.Length)) > 0)
{
byte[] actual = new byte[bytesRead];
Buffer.BlockCopy(buffer, 0, actual, 0, bytesRead);
outputStream.Write(actual, 0, actual.Length);
}
}
Thread.Sleep(1000);
}

Just a suggestion, try disposing the Stream, and the Reader. I've seen some weird cases where not disposing objects like these and using them in loops can yield some wacky results....

That URL doesn't quite make sense to me unless you are using MVC or something that can interpret the querystring correctly.
http://www.website.com/results.aspx&page=
should be:
http://www.website.com/results.aspx?page=
Some browsers will accept poorly formed URLs and render them fine. Others may not which may be the problem with your console app.

Here's my terrible, hack-ish, workaround solution:
Make another console app that calls THIS one, in which the first console app passes an argument at the end of strURL. It works, but I feel so dirty.