Open a Text Editor of your choice, copy/paste the code below and save it as “MyHttpServer.dpr”.
program MyHttpServer;
{$APPTYPE CONSOLE}
uses
SysUtils,
rtcTypes,
rtcSystem,
rtcInfo,
rtcConn,
rtcDataSrv,
rtcHttpSrv;
type
// This is our self-contained HTTP Server class ...
TMyServer=class(TObject)
protected
HttpSrv:TRtcHttpServer;
DataProv:TRtcDataProvider;
public
constructor Create(PortNumber:String='80'; IPVer:RtcIPV=rtc_IPVDefault);
destructor Destroy; override;
procedure DataProv_CheckRequest(Sender: TRtcConnection);
procedure DataProv_DataReceived(Sender: TRtcConnection);
procedure HttpSrv_ListenStart(Sender: TRtcConnection);
procedure HttpSrv_ListenError(Sender: TRtcConnection; E:Exception);
procedure HttpSrv_ListenStop(Sender: TRtcConnection);
procedure HttpSrv_RequestNotAccepted(Sender: TRtcConnection);
end;
constructor TMyServer.Create(PortNumber:String='80'; IPVer:RtcIPV=rtc_IPVDefault);
begin
// Create HttpServer and DataProvider components ...
HttpSrv:=TRtcHttpServer.Create(nil);
DataProv:=TRtcDataProvider.Create(nil);
// Assign Server for our Data Provider ...
DataProv.Server:=HttpSrv;
// Assign Data Provider Events (handles Valid Requests) ...
DataProv.OnCheckRequest:=DataProv_CheckRequest;
DataProv.OnDataReceived:=DataProv_DataReceived;
// Assign Server Events (handles the rest) ...
HttpSrv.OnRequestNotAccepted:=HttpSrv_RequestNotAccepted;
HttpSrv.OnListenStart:=HttpSrv_ListenStart;
HttpSrv.OnListenStop:=HttpSrv_ListenStop;
HttpSrv.OnListenError:=HttpSrv_ListenError;
// Configure the Server ...
HttpSrv.ServerPort:=PortNumber;
HttpSrv.ServerIPV:=IPVer;
HttpSrv.MultiThreaded:=True;
// Start the Server listener ...
HttpSrv.Listen();
end;
destructor TMyServer.Destroy;
begin
// Stop the Server listener ...
HttpSrv.StopListenNow();
// Destroy the components ...
HttpSrv.Free;
DataProv.Free;
end;
procedure TMyServer.DataProv_CheckRequest(Sender: TRtcConnection);
begin
// Check Request headers and "Accept" all Requests
// we want to handle with our Data Provider ...
with TRtcDataServer(Sender) do
if (Request.Method='GET') and // we only want "GET" requests
(Request.ContentLength=0) then // ... with no content body
if (Request.URI='/html') or
(Request.URI='/json') or
(Request.URI='/xml') or
(Request.URI='/code') then
Accept; // Accept the Request
end;
procedure TMyServer.DataProv_DataReceived(Sender: TRtcConnection);
var
t:TRtcRecord;
begin
with TRtcDataServer(Sender) do
// We will start processing the request only if
// we have received the complee request content body ...
if Request.Complete then
if Request.URI='/html' then
begin
// We can use multiple "Write" calls
// to prepare our HTML response ...
Response.ContentType:='text/html';
Write('<html><body>');
Write('Your IP: '+PeerAddr+'<br>');
Write('Your Port: '+PeerPort+'<br>');
Write('Date & Time: <b>'+DateTimeToStr(Now)+'</b><br>');
Write('Agent: <i>'+Request['User-Agent']+'</i><br>');
Write('</body></html>');
// All "Write" calls will be buffered,
// RTC will calculate the "Content-Length" for us
// and send the whole content body out as a single
// Response - when we are finished with our event.
end
else
begin
// Using TRtcRecord to prepare our response Object ...
t:=TRtcRecord.Create;
try
t.asText['agent']:=Request['User-Agent'];
t.asText['ip']:=PeerAddr;
t.asText['port']:=PeerPort;
t.asDateTime['now']:=Now;
if Request.URI='/json' then
begin
// Serialize to "JSON" ...
Response.ContentType:='application/json';
Write(t.toJSON);
end
else if Request.URI='/xml' then
begin
// Serialize to "XML-RPC" ...
Response.ContentType:='text/xml';
Write(t.toXMLrpc);
end
else if Request.URI='/code' then
begin
// Serialize to "Code" (RTC format) ...
Response.ContentType:='text/plain';
Write(t.toCode);
end;
finally
t.Free;
end;
end
end;
procedure TMyServer.HttpSrv_RequestNotAccepted(Sender: TRtcConnection);
begin
// Request wasn't accepted ...
with TRtcDataServer(Sender) do
begin
// Send "404" status code back ...
Response.Status(404,'Not Found');
Response.ContentType:='text/plain';
// Something to show in the Web Browser ...
Write('Bad command.');
// And ... Disconnect the Client.
Disconnect;
end;
end;
procedure TMyServer.HttpSrv_ListenError(Sender: TRtcConnection; E: Exception);
begin
Writeln('Server Error: '+E.Message);
end;
procedure TMyServer.HttpSrv_ListenStart(Sender: TRtcConnection);
begin
Writeln('Server started.');
end;
procedure TMyServer.HttpSrv_ListenStop(Sender: TRtcConnection);
begin
Writeln('Server stopped.');
end;
var
MyServer:TMyServer;
begin
try
// Create and start our Server ...
MyServer:=TMyServer.Create('80');
try
// Since this is a console application and our
// Server is Multi-Threaded, we can do whatever
// we want here. For simplicity reasons, we will
// just use "ReadLn" to allow the Server to run
// while we wait for the user to press <Enter>.
Writeln('Press <Enter> to Quit ...');
ReadLn;
// User has pressed <Enter> - time to kill our Server.
finally
MyServer.Free;
end;
except
on E: Exception do
Writeln(E.ClassName, ': ', E.Message);
end;
end.
Now, you only need the RealThinClient SDK and Delphi to compile this Project into a fully functional multi-threaded cross-platform Server. If you compile and run this Project on your local machine, you will see a Console open up, with a message “Server started”. To test the Server, open any Web Browser and try one of these URLs …
http://localhost/html
http://localhost/json
http://localhost/xml
http://localhost/code
If the Server does NOT start, you will see an error message like “Server Error: Address already in use”, which means that another Server is already running on port 80 (which was used in this example), so you will either need to modify your code to use another Port (see line 180), or stop the other Server before you start this Project. Because all the links above are accessing a Server running on the default HTTP port (80), if your Server is running on any other Port, you also need to include the Port number in each URL. For example, if the Server is running on Port 8080 on your local machine, URLs would look like this …
http://localhost:8080/html
http://localhost:8080/json
http://localhost:8080/xml
http://localhost:8080/code
What is happening there? Well … if you take a look at the “DataProv_CheckRequest” method, you will see that it will accept “/html”, “/json”, “/xml” or “/code” as the request URI. And if you then look at the “DataProv_DataReceived” method, you will see how the Server handles each of these requests to prepare a response for the Client.
There are also methods to display a simple message on the Console if there is an error starting the Server, or when the Server is started and stopped. There is also a method which handles all requests not accepted or processed by any Data Provider linked to youor Server, which will (in this case) respond with an 404 page that says “Bad command.”
Ofcourse, the code above is not limited to Console Applications. The exact same code as I’ve used here for the “TMyServer” class would work without any modifications if compiled into a Windows or Android Service, a Windows VCL Application, or cross-platform FireMonkey Application for Windows, MacOSX, iOS or Android.
And … naturally … because RTC SDK is a set of non-visual components for Delphi, all the components install into the Delphi component palette, so you can simply drop them on any Form or DataModule in Delphi, set properties in the Delphi’s Object Inspector and implement all the events you want directly, without the need to implement your own classes.
Now, someone could ask: “What if I want to add more functionality? Do I add more code to DataProv_CheckRequest and DataProv_DataReceived events to handle everything I want my Server to do?”
While you could do that too, it is better to keep your functionality separated by using a different TRtcDataProvider component for each set of requests you want your Server to handle. To make your “data” provided by each TRtcDataProvider component accessible through the same Server, you only have to set the “Server” of every TRtcDataProvider component to the same TRtcHttpServer component – just like I’ve done in the example above for the “DataProv” component.
Since RTC SDK also comes with components for writing Remote Functions, Load Balancers and Gateways (as well as some other stuff), you can always add more functionality to your Server, simply by adding more components from the RTC SDK and linking them to the TRtcHttpServer.
RTC SDK has full support for IPv4 as well as IPv6, but you need to specify which IP version you want your Server to use. By default, RTC will be using IPv4. To have your Server running on IPv6, simply use rtc_IPv6 as the 2nd parameter on line 180.
On Windows, if you want to have a Server listening on the same Port using IPv4 and IPv6, you will need two TRtcHttpServer components. One will have “ServerIPV:=rtc_IPv4” and the other “rtc_IPv6”. For all your Data Providers to be accessible from both HTTP Servers, you will use a TRtcDualDataServerLink component, which has 2 Server properties, so it can be assigned both TRtcHttpServer components, and then you will assign that TRtcDualDataServerLink component to the “Link” property of each TRtcDataProvider component – instead of using the “Server” property there.
I didn’t bother using a TRtcDataServerLink component in my example above to keep it simple and give you an easy start, but keep in mind that you can also use TRtcDataServerLink and TRtcDualDataServerLink components to completely separate your “Data Providers” and all your Application logic from the actual Server component being used.
And if you need to host your Application as an ISAPI extension, for example … if your customers already have IIS or Apache running and you want your Server code to be accessible through their existing Web Server, by separating your code into modules and using a TRtcDataServerLink, you can compile your Server code into ISAPI DLLs by linking your Data Providers and Remote Functions to a TRtcISAPIServer component and compile it into the ISAPI example Project, included in the RTC SDK “Lib” folder.
Best Regards,
Danijel Tkalcec