Building a Server Application in Delphi

There are two ways to built a server application in Indy: command handlers and through the OnExecute event.

Command handlers are mostly suited to text-based protocols. They enable you to set up your own protocol and implement it. Most protocols are text-based and can be used with the command handlers format, but are not suited to protocols which have a binary command structure or no command structure. Below is an example command handler that implements a command called "gettxt":

procedure TForm1.IdTCPServer1TIdCommandHandler1gettxt(ASender: TIdCommand);

var

receivedtxt: string;

begin

if ASender.Params.Count > 0 then begin

receivedtxt := ASender.Params[0];

end;

ASender.Reply.Text.Text := receivedtxt;

end;

OnExecute Event

For servers that have a binary command structure or no command structure, the OnExecute event should be used. The OnExecute event runs for as long as there is a live connection. Here's an example:

procedure TForm1.IdTCPServer1Execute(AThread: TIdPeerThread);

var

ACmd: string;

begin

with AThread.Connection do begin

ACmd := Trim(ReadLn);

if SameText(ACmd, 'Hi') then begin

WriteLn('200 Hi to you too');

Disconnect;

end else if SameText(ACmd, 'DATE') then begin

WriteLn('200 ' + DateToStr(Date));

end else begin

WriteLn('400 Unknown command');

end;

end;

end;

You can see in the code above that I do not check for a connection nor do I do any looping. Both are done by Indy automatically. Indy repeatedly calls this event until there is no connection. This can be caused by the client disconnecting, or a disconnect call as above.

Conclusion

I've read extensively on both Delphi and indy threading models, but have not covered it all here. So if you want to learn more about indy components and the like, a excellent source is the "Indy in Depth" ebook. It tells you everything about how indy components work and in what situations you can use them.