Difference between revisions of "HelloWorldCpp"

From ArmadeusWiki
Jump to: navigation, search
m (Source code)
m (Makefile improving)
Line 62: Line 62:
  
 
==Putting it all together in a Makefile==
 
==Putting it all together in a Makefile==
You can put your program compiling and copying in a ''Makefile'' to make things cleaner:
+
You can put your program compilation and installation in a ''Makefile'' to make things cleaner:
 
  CXX=arm-linux-g++
 
  CXX=arm-linux-g++
 
  CXXFLAGS=-W -Wall
 
  CXXFLAGS=-W -Wall
Line 82: Line 82:
 
  clean:
 
  clean:
 
     rm -rf *.o
 
     rm -rf *.o
 +
    rm -f $(EXEC)
 
   
 
   
  install:  
+
  install: all
 
     cp -f $(EXEC) /tftpboot/
 
     cp -f $(EXEC) /tftpboot/
  
'''!! If you do a cut & paste with previous commands, don't forget to check TABS for each target (Makefiles are using TABS and not SPACES) !!'''
+
'''!! If you do a cut & paste with previous commands, don't forget to check TABS for each target (Makefiles are using TABS and not SPACES) !! In that case make will complains about a missing separator line 11'''
  
 
Then, just do:
 
Then, just do:

Revision as of 21:37, 3 April 2007

On this page you will learn how to create your first C++ application for your Armadeus board

Source code

First take your favorite editor/IDE and create the following program:

#include <iostream>

using namespace std;

class Hello
{
    private:
        const char* message;

    public:
       Hello( const char* amessage ) { message = amessage; };
       virtual ~Hello() { cout << "Bye, bye !" << endl; };

       void speaks() { cout << message << endl; };
};

int main()
{
    Hello hello( "APF9328 says Hello everybody !" );
    hello.speaks();
    return 0;
}

Save it as hello.cpp

Compilation

The C++ cross compiler is installed in armadeus/buildroot/build_arm_nofpu/staging_dir/bin/ and is named arm-linux-g++ There are 2 possibilities to use it:

  • either add armadeus/buildroot/build_arm_nofpu/staging_dir/bin/ to your PATH environment variable and then call arm-linux-g++ instead of g++
  • or call directly armadeus/buildroot/build_arm_nofpu/staging_dir/bin/arm-linux-g++

So to compile your small program do (here hello.cpp was saved in armadeus/target/demos/ directory):

[host demos]$ ../../buildroot/build_arm_nofpu/staging_dir/bin/arm-linux-g++ -o hello hello.cpp

Running

Copy your hello executable on your board either through TFTP or NFS

TFTP

Be sure to have TFTP server installed, if not it's explained here
Copy hello to TFTP directory:

[host demos]$ cp hello /tftpboot/

Load your executable on the target (here my host IP is 192.168.0.2):

# tftp -g -r hello -l /usr/bin/hello 192.168.0.2

Give it executable rights, if lost during TFTP transfer:

# chmod a+x /usr/bin/hello

Launch it:

# /usr/bin/hello
APF9328 says: Hello World ! ;-)
#

Now it's up to you ! ;-)

NFS

Be sure to have NFS server installed, if not it's explained here
I assume that your NFS drive is accessible from /mnt/host Launch your prog:

[target]# /mnt/host/hello

Putting it all together in a Makefile

You can put your program compilation and installation in a Makefile to make things cleaner:

CXX=arm-linux-g++
CXXFLAGS=-W -Wall
LDFLAGS=
EXEC=hello
SRC= $(wildcard *.cpp)
OBJ= $(SRC:.c=.o)

all: $(EXEC)

hello: $(OBJ)
    $(CXX) -o $@ $^ $(LDFLAGS)

%.o: %.c
    $(CXX) -o $@ -c $< $(CXXFLAGS)

.PHONY: clean install

clean:
    rm -rf *.o
    rm -f $(EXEC)

install: all
    cp -f $(EXEC) /tftpboot/

!! If you do a cut & paste with previous commands, don't forget to check TABS for each target (Makefiles are using TABS and not SPACES) !! In that case make will complains about a missing separator line 11

Then, just do:

[host demos]$ make

Links