Browse Source

Updated Platform support documentation.

Now using generic aliases implicitly everywhere.
pull/34/head
Vladimír Vondruš 13 years ago
parent
commit
a4ab4ad74c
  1. 80
      doc/platform.dox
  2. 16
      doc/portability.dox

80
doc/platform.dox

@ -41,11 +41,18 @@ required methods.
@section platform-windowed Windowed applications
Windowed applications provide a window and keyboard and mouse handling. The
most basic toolkit (and toolkit available on most platforms) is GLUT, which is
is implemented in GlutApplication. As said above, the usage is similar for all
toolkits, you must provide two-argument constructor and implement at least
@ref GlutApplication::viewportEvent() "viewportEvent()" and
@ref GlutApplication::drawEvent() "drawEvent()".
most basic toolkit (and toolkit packaged for most systems) is GLUT, which is
implemented in @ref Platform::GlutApplication. As said above, the usage is
similar for all toolkits, you must provide one-argument constructor and
implement at least @ref GlutApplication::viewportEvent() "viewportEvent()" and
@ref GlutApplication::drawEvent() "drawEvent()". The class can be then used
directly in `main()`, but for convenience and portability it's better to use
@ref MAGNUM_GLUTAPPLICATION_MAIN() macro.
To simplify the porting, the library provides `Platform::Application` typedef
and `MAGNUM_APPLICATION_MAIN()` macro (but only if only one application header
is included, to avoid ambiguity). Changing the code to use different toolkit is
then matter of replacing only the `#include` statement.
Barebone application implementation which will just clear the window to dark
blue color:
@ -56,7 +63,7 @@ blue color:
using namespace Magnum;
class MyApplication: public Platform::GlutApplication {
class MyApplication: public Platform::Application {
public:
MyApplication(const Arguments& arguments);
@ -64,7 +71,7 @@ class MyApplication: public Platform::GlutApplication {
void drawEvent() override;
};
MyApplication::MyApplication(const Arguments& arguments): Platform::GlutApplication(arguments) {
MyApplication::MyApplication(const Arguments& arguments): Platform::Application(arguments) {
// Set clear color to dark blue
Renderer::setClearColor({0.0f, 0.0f, 0.4f});
}
@ -83,7 +90,7 @@ void MyApplication::drawEvent() {
}
// main() function implementation
MAGNUM_GLUTAPPLICATION_MAIN(MyApplication)
MAGNUM_APPLICATION_MAIN(MyApplication)
@endcode
@section platform-windowless Windowless applications
@ -91,8 +98,19 @@ MAGNUM_GLUTAPPLICATION_MAIN(MyApplication)
Windowless applications provide just a context for ofscreen rendering or
performing tasks on GPU. There is not yet any platform-independent toolkit
which could handle this in portable way, thus you have to use platform-specific
ones. As example we use WindowlessGlxApplication, you need to implement just
@ref WindowlessGlxApplication::exec() "exec()" function.
ones. As example we use @ref Platform::WindowlessGlxApplication, you need to
implement just @ref WindowlessGlxApplication::exec() "exec()" function. The
class can be then used directly in `main()`, but again, for convenience and
portability it's better to use @ref MAGNUM_WINDOWLESSGLXAPPLICATION_MAIN()
macro.
Similarly as with windowed applications, to simplify the porting, the library
provides `Platform::WindowlessApplication` typedef and `MAGNUM_WINDOWLESSAPPLICATION_MAIN()`
macro, but only if only one windowless application header is included. Changing
the code to use different toolkit is then matter of replacing only the `#include`
statement. Aliases for windowless applications are separated from aliases for
windowed applications, because projects commonly contain both graphics
application and command-line tools (for data preparation etc.).
Barebone application which will just print out current OpenGL version and
renderer string and exits:
@ -102,14 +120,14 @@ renderer string and exits:
using namespace Magnum;
class MyApplication: public Platform::WindowlessGlxApplication {
class MyApplication: public Platform::WindowlessApplication {
public:
MyApplication(const Arguments& arguments);
int exec() override;
};
MyApplication::MyApplication(const Arguments& arguments): Platform::WindowlessGlxApplication(arguments) {}
MyApplication::MyApplication(const Arguments& arguments): Platform::WindowlessApplication(arguments) {}
int MyApplication::exec() {
Debug() << "OpenGL version:" << Context::current()->versionString();
@ -120,23 +138,31 @@ int MyApplication::exec() {
}
// main() function implementation
MAGNUM_WINDOWLESSGLXAPPLICATION_MAIN(MyApplication)
MAGNUM_WINDOWLESSAPPLICATION_MAIN(MyApplication)
@endcode
@section platform-compilation Compilation with CMake
Barebone compilation consists just of finding %Magnum library with required
`*Application` component, compilation of the executable and linking the
libraries to it:
`*Application` component, adding %Magnum's `${MAGNUM_INCLUDE_DIRS}` and application-specific `${MAGNUM_GLUTAPPLICATION_INCLUDE_DIRS}` to include path, compilation of the
executable and linking `${MAGNUM_LIBRARIES}` and `${MAGNUM_GLUTAPPLICATION_LIBRARIES}`
to it.
Again, to simplify porting, you can also use generic `${MAGNUM_APPLICATION_INCLUDE_DIRS}`
and `${MAGNUM_WAPPLICATION_LIBRARIES}` aliases (or `${MAGNUM_WINDOWLESSAPPLICATION_INCLUDE_DIRS}`, `${MAGNUM_WINDOWLESSAPPLICATION_LIBRARIES}` for windowless applications), but
only if only one application (windowless application) component is requested to
avoid ambiguity. Changing the code to use different toolkit is then matter of
replacing only the requested `*Application` component.
@code
find_package(Magnum REQUIRED GlutApplication)
include_directories(${MAGNUM_INCLUDE_DIRS})
include_directories(${MAGNUM_INCLUDE_DIRS} ${MAGNUM_APPLICATION_INCLUDE_DIRS})
add_executable(myapplication MyApplication.cpp)
target_link_libraries(myapplication
${MAGNUM_LIBRARIES}
${MAGNUM_GLUTAPPLICATION_LIBRARIES})
${MAGNUM_APPLICATION_LIBRARIES})
@endcode
@section platform-configuration Specifying configuration
@ -147,20 +173,20 @@ window size 800x600 pixels). If you want something else, you can pass
constructor. Using method chaining it can be done conveniently like this:
@code
MyApplication::MyApplication(int& argc, char** argv):
Platform::GlutApplication(argc, argv, Configuration()
Platform::Application(argc, argv, Configuration()
.setTitle("My Application")
.setSize({800, 600})
{
{
// ...
}
@endcode
However, sometimes you would need to configure the application based on some
configuration file or system introspection. In that case you can pass `nullptr`
instead of Configuration instance and then specify it later with
@ref GlutApplication::createContext() "createContext()":
instead of @ref GlutApplication::Configuration "Configuration" instance and
then specify it later with @ref GlutApplication::createContext() "createContext()":
@code
MyApplication::MyApplication(int& argc, char** argv): Platform::GlutApplication(argc, argv, nullptr) {
MyApplication::MyApplication(int& argc, char** argv): Platform::Application(argc, argv, nullptr) {
// ...
createContext(Configuration()
@ -173,12 +199,12 @@ MyApplication::MyApplication(int& argc, char** argv): Platform::GlutApplication(
If the context creation in constructor or @ref GlutApplication::createContext() "createContext()"
fails, the application exits. However, it is also possible to negotiate the
context using @ref GlutApplication::tryCreateContext() "tryCreateContext()". The
only difference is that this function returns `false` instead of exiting. You
can for example try enabling MSAA and if the context creation fails, fall back
to no-AA rendering:
context using @ref GlutApplication::tryCreateContext() "tryCreateContext()".
The only difference is that this function returns `false` instead of exiting.
You can for example try enabling MSAA and if the context creation fails, fall
back to no-AA rendering:
@code
MyApplication::MyApplication(int& argc, char** argv): Platform::GlutApplication(argc, argv, nullptr) {
MyApplication::MyApplication(int& argc, char** argv): Platform::Application(argc, argv, nullptr) {
// ...
Configuration conf;

16
doc/portability.dox

@ -175,19 +175,9 @@ can just switch to another base class and in many cases you won't need to
change any other code. It has its limitations, though - some toolkits don't
support all keys, mouse movement events etc.
In most cases the entry point is classic `main()` function, but some platforms
(e.g. Native Client) have different requirements. To make things easier, entry
points are handled using macros, which take care of the rest.
If exactly one `*Application` or `*Windowless*Application` header is included,
the application class is aliased to `Platform::Application` or
`Platform::WindowlessApplication` and the macro is aliased to
`MAGNUM_APPLICATION_MAIN()` or `MAGNUM_WINDOWLESSAPPLICATION_MAIN()` to
simplify porting. The same is with CMake code, if exactly one `*Application` or
`Windowless*Application` component , the libraries and include dirs are
available in `MAGNUM_APPLICATION_LIBRARIES` / `MAGNUM_WINDOWLESSAPPLICATION_LIBRARIES`
and `MAGNUM_APPLICATION_INCLUDE_DIRS` / `MAGNUM_WINDOWLESSAPPLICATION_INCLUDE_DIRS`
variables.
As mentioned in @ref platform, all the classes, macros and CMake variables have
generic aliases, thus using different toolkit is in most cases only matter of
replacing two lines of code.
Example application, which targets both embedded Linux (using plain X and EGL)
and desktop (using SDL2 toolkit). Thanks to static polymorphism most of the

Loading…
Cancel
Save