Step Two : Basic configuration handling
First, from that point we will use more and more cim-specific files, like
pictures, map files, music, sound etc. It his tutorial we load the first version
of cim’s configuration file.
The configuration handling is Crystal Space is very easy to use. All plugin has
his own configuration file, you can override the values with your own
configuration file. You can add or override many option from command line. The
manual describes this mechanism in depth, please read link this section for all details.
You can configure many thing from your configuration file.
We can load plugins from the configuration file, this is a very nice feature.
You can change dynamically the plugins, without recompile your application. For
example, the cim.cfg (config file) has this line somewhere:
System.Plugins.iConsoleOutput = crystalspace.console.output.standard
That means, Crystal Space will use the standard output plugin. When you change the line a little bit:
System.Plugins.iConsoleOutput = crystalspace.console.output.fancy
In that case the fancy console output awaits you, when you run the application
(Note, we will change the console source a little bit., now it’s s load the
standard plugin statically. When you edit the walktest.cfg file, in the
/data/config directory , you can see the effect in walktest.).
You can ovveride any settings of the official CS plugins. For example:
Video.ScreenWidth = 800 Video.ScreenHeight = 600 Video.ScreenDepth = 32 Video.FullScreen = yes
When your configuration file has this lines, your app will start at 800*600
resolution, in fullscreen. You can change any system configuration key, by
adding the name and the value in the config file.
You can mount directories at runtime. For example :
VFS.Mount.lev/cim = $@data$/cim$/ </pre will mount your real ''cim.exe/data/cim'' directory to ''lev/cim'' the virtual path. I always use this solution, it’s better, then rewrite the original ''vfs.cfg''. Note, ''vfs.cfg'' is just a normal configuration file, with many mounting keys. You can add your own configuration key. Cim will use some own key, for example: <pre> Cim.Settings.StartLevel = terrain Cim.Settings.EnableConsole = true; </pre We have to handle this config keys from code, of course. The whole ''cim.cfg'' will look like this: <pre> ;mounting cim working folder ;VFS.Mount.lev/cim = $@data$/cim$/ ;cs plugins System.Plugins.iGraphics3D = crystalspace.graphics3d.opengl System.Plugins.iConsoleInput = crystalspace.console.input.standard System.Plugins.iConsoleOutput = crystalspace.console.output.standard System.Plugins.iImageIO = crystalspace.graphic.image.io.multiplexer System.Plugins.iLoader = crystalspace.level.loader System.Plugins.iDocumentSystem = crystalspace.documentsystem.xmlread System.Plugins.iEngine = crystalspace.engine.3d System.ApplicationID = Tutorial.Cim ;video settings Video.ScreenWidth = 1024 Video.ScreenHeight = 768 Video.ScreenDepth = 32 Video.FullScreen = no ;mouse settings MouseDriver.DoubleClickTime = 300 MouseDriver.DoubleClickDist = 2 ;game specific settings Cim.Settings.StartLevel = /lev/terrain Cim.Settings.EnableConsole = false
This file you have to put into your CS/config/ directory.
To load a config file, first we need a pointer to the config manager. Add the
following private member to cim class in cim.h:
/// A pointer to the configuration manager. csRef<iConfigManager> confman;
And in the cim.cpp, on the OnInitialize function:
if (!csInitializer::SetupConfigManager (GetObjectRegistry (),
"/config/cim.cfg"))
return ReportError ("Error reading config file 'cim.cfg'!");
Your config file will be loaded and your settings will be applied. Put cim.cfg
in same directory, that your cim binary, recompile the application and run it.
Try different Video.ScreenWidth Video.ScreenHeight, etc values (i.e 800*600,
1024*768), and you can see, Crystal Space respects your changes. This the the
real power of Crystal Space configuration handling.
Our config file contents the following Cim-specific keys:
Cim.Settings.StartLevel = terrain; Cim.Settings.EnableConsole = true;
The StartLevel defines the name of the starting map, the EnableConsole decides,
the app creates a console or not.
Let’s go to to write the code.
We make many changes. First, we introduce a variable to check, we use console or
not. This will a private member of the cimGame class, called use_console. When
this flag is false, we don’t instance and use the console class. The event
handlers first check, we have console or not, to avoid crashes :
bool cimGame::OnKeyboard(iEvent& ev)
{
// We got a keyboard event.
csKeyEventType eventtype = csKeyEventHelper::GetEventType(&ev);
if (eventtype == csKeyEventTypeDown)
{
// The user pressed a key (as opposed to releasing it).
utf32_char code = csKeyEventHelper::GetCookedCode(&ev);
//The user pressed tab to toggle console, when we have any console
if (use_console)
{
if (code == CSKEY_TAB)
{
console->Toggle();
return false;
}
//more code of course
(This is a very simple solution,. Cs provides us a better way to event handling,
we will deal with this problem later. )
On the cimGame::Setup, we do the following changes:
We have to change the ProcessFrame function from same reason:
void cimGame::ProcessFrame ()
{
// First get elapsed time from the virtual clock.
csTicks elapsed_time = vc->GetElapsedTicks ();
csVector3 obj_move (0);
csVector3 obj_rotate (0);
if (use_console)
if(console->IsVisible() )
return;
// more code
We change the our console class a little bit. First, we change the plugin
loading method, we don’t use concrete plugin name, since we give the used
plugins from the configuration file:
bool cimConsole::Initialize(iObjectRegistry* obj_reg)
{
object_reg = obj_reg;
conout = CS_QUERY_REGISTRY(obj_reg, iConsoleOutput);
if (!conout)
{
csReport (object_reg,
CS_REPORTER_SEVERITY_ERROR, "cim.console",
"Can't load the output console!");
return false;
}
conin = CS_QUERY_REGISTRY(obj_reg,iConsoleInput);
if (!conin)
{
csReport (object_reg,
CS_REPORTER_SEVERITY_ERROR, "cim.console",
"Can't load the input console!");
return false;
}
//more code
In the cimGame class, we introduce a new private function, called LoadConfig.
This function parse the application’s special konfiguration keys, and makes the
desired settings.
The configuration manger has some GetSomeType function to get the key values,
these are GetInt. GetStr, GetBool, GetFloat. This functions look up in the
configuration file, and return with the given key’s value (first parameter). You
can give a default value, when the key was not found, in the second parameter.
bool cimGame::LoadConfig()
{
csRef<iConfigManager> confman (CS_QUERY_REGISTRY (GetObjectRegistry(), iConfigManager));
use_console = confman->GetBool("Cim.Settings.EnableConsole ", true);
startmap = confman->GetStr("Cim.Settings.StartLevel","terrain");
}
Finally, we learn the basics of the command line handling in CS. Cystal Space
uses heavily command lines and provides an easy way to process it by a CS based
application. The iCommandLineParser helps to do it. So we extend the
cimGame::Setup function:
csRef<iCommandLineParser> cmdline(CS_QUERY_REGISTRY (object_reg,
iCommandLineParser));
if (!cmdline)
return ReportError("Failed to set up command line parser!");
Then we look up for the option enable-python-console. When the user runs the
application with this option, we set up the console anyway:
if( cmdline->GetOption("enable-python-console")) use_console=true;
Many more is possible with the command line parser, but probably this function
will you use most often.
And finally, we organize the map loading and collosion detection init code to a
new LoadMap (const char* mapname) function. This function can load any CS map
file by name:
bool cimGame::LoadMap (const char* mapname)
{
//Delete the engine's content.
engine->DeleteAll();
// Set VFS current directory to the level we want to load.
csRef<iVFS> VFS (CS_QUERY_REGISTRY (GetObjectRegistry (), iVFS));
VFS->ChDir (mapname);
// Load the level file which is called 'world'.
if (!loader->LoadMapFile ("world"))
ReportError("Error couldn't load level!");
// Initialize collision objects for all loaded objects.
csColliderHelper::InitializeCollisionWrappers (cdsys, engine);
// Let the engine prepare all lightmaps for use and also free all images
// that were loaded for the texture manager.
engine->Prepare ();
// Find the starting position in this level.
csVector3 pos (0);
if (engine->GetCameraPositions ()->GetCount () > 0)
{
// There is a valid starting position defined in the level file.
iCameraPosition* campos = engine->GetCameraPositions ()->Get (0);
room = engine->GetSectors ()->FindByName (campos->GetSector ());
pos = campos->GetPosition ();
}
else
{
// We didn't find a valid starting position. So we default
// to going to room called 'room' at position (0,0,0).
room = engine->GetSectors ()->FindByName ("room");
pos = csVector3 (0, 0, 0);
}
if (!room)
ReportError("Can't find a valid starting position!");
// Now we need to position the camera in our world.
view->GetCamera ()->SetSector (room);
view->GetCamera ()->GetTransform ().SetOrigin (pos);
// Initialize our collider actor.
collider_actor.SetCollideSystem (cdsys);
collider_actor.SetEngine (engine);
csVector3 legs (.2f, .3f, .2f);
csVector3 body (.2f, 1.2f, .2f);
csVector3 shift (0, -1, 0);
collider_actor.InitializeColliders (view->GetCamera (),
legs, body, shift);
return true;
}
Look at the source for all changes.
Thats it. We can configure CS engine and we can add more cim-specific option
later. In the next step we will create the basement of our window handling solution.
