Question

I have the following piece of code

try
{
  if (!bDebug)
    smtp.Send(m);
}
catch (Exception e)
{
  wl("Meldingen kunne ikke sendes til en eller flere mottakere.", ConsoleColor.Red);
  wl(e.Message, ConsoleColor.DarkRed);
  using (var errorfile = System.IO.File.CreateText("error-" + DateTime.Now.Ticks + ".txt"))
  {
    errorfile.WriteLine(e.StackTrace);
    if (e.GetType() == typeof(SmtpFailedRecipientException))
    {
      var se = (SmtpFailedRecipientException) e;
      errorfile.WriteLine(se.FailedRecipient);
    }
    errorfile.WriteLine(e.ToString());
  }
}

Where wl is a shortcut for writing to the console with color, and the text in the first line says “The message could not be sent to one or more recipients.

Previously I only caught the SmtpFailedRecipientException, but when it started failing in some other steps I shoved the generic Exception in there. So the part I’m wondering about is where I’m casting the Exception object into a more specific object to get the FailedRecipient property. Could/should this be done in another more proper way? It seems a bit clunky…

Answer 1

You can have multiple catch branches:

catch (SmtpFailedRecipientException se)
{
  using (var errorfile = System.IO.File.CreateText("error-" + DateTime.Now.Ticks + ".txt"))
  {
    errorfile.WriteLine(se.StackTrace);  
    // variable se is already the right type, so no need to cast it      
    errorfile.WriteLine(se.FailedRecipient);       
    errorfile.WriteLine(se.ToString());
  }
}
catch (Exception e)
{
  wl("Meldingen kunne ikke sendes til en eller flere mottakere.", ConsoleColor.Red);
  wl(e.Message, ConsoleColor.DarkRed);   

  // for other error types just write the info without the FailedRecipient
  using (var errorfile = System.IO.File.CreateText("error-" + DateTime.Now.Ticks + ".txt"))
  {
    errorfile.WriteLine(e.StackTrace);        
    errorfile.WriteLine(e.ToString());
  }

}