The Crossware ARM C/C++ compiler generates code for the microprocessors and microcontrollers based on the ARM7, ARM9, Cortex-M3 and Cortex-M4 cores. It comes as part of the Crossware ARM Development Suite and runs under Windows 7 and above.
The C compiler conforms to the 1989 ANSI C specification and in addition provides a number of general enhancements including:
It also supports a number of features from the 1999 ANSI C Standard including:
In addition, the compiler supports the C++ feature that allows variables to be defined within the conditional expressions of the for loop, while loop, if statement and switch statement
The support libraries are a subset of the ANSI standard libraries. The supported functions are listed below.
The C++ compiler supports the Embedded C++ language. Embedded C++ is a subset of C++ specifically formulated for embedded systems. It is defined at http://www.caravan.net/ec2plus.
Embedded C++ excludes templates, exceptions, namespaces, run-time type information, localization, file operations and some other features.
Also Embedded C++ does not support multiple-inheritance and virtual base classes. However the Crossware compiler does support multiple-inheritance but does not support virtual base classes.
Since Embedded C++ does not support templates, basic_string is not supported. As an alternative, Embedded C++ provides the string class. The Crossware C++ library includes the string class.
The Crossware C++ library also includes the operators new, delete, new[], delete[], placement new and placement new[].
Other Embedded C++ library features such as streams and complex are not yet available in the Crossware library.
The compiler uses the following sizes for the various C data types:
| char and unsigned char | 1 byte | |
| short int and unsigned short int | 2 bytes | |
| int and unsigned int | 4 bytes | |
| long and unsigned long | 4 bytes | |
| long long and unsigned long long | 8 bytes | |
| float | 4bytes (32 bits) | |
| double | 8 bytes (64 bits) | |
| long double | 8 bytes (64 bits) | |
| enum | 4 bytes | |
| bit fields | up to 32 bits |
Optimizations include:
The compiler supports three calling conventions:
Assembler code can be embedded in your C source code using two methods.
Strings inserted using the _asm keyword are scanned by the C preprocessor and so assembler statements can be generated using C macros with full macro token replacement. Additionally, the compiler can replace C variable names with the appropriate sub-string allowing easy access to global, static and local variables and parameters.
On the other hand, assembler macros are best defined outside of functions and the #asm/#endasm directive allows this.
Compiler generated code and data are automatically located in appropriate memory sections. Linker options allow these sections to be located at user defined memory locations. String constants and objects declared as const are located in code space. Initialised and uninitialised data are located in separate sections and initialisation of these sections is automatically carried out at run time.
Directives are also available allowing you to tell the compiler to place code and data in your own named sections. This gives you further control on where code and data are placed and how data space is initialised.
The optional use of linker scripts adds more flexibilty to the location of sections and in particular allows the storage and run-time addresses to be different.
A unique drag-and-drop linker script editor simplifies the change to a linker script approach and eliminates the need for an understanding of the linker script language.
Optional section name translation enables the use of linker scripts written for other tool chains.
The compiler optionally generates information on all of the definitions of and references to the identifiers used in your program. This includes functions, function parameters, local variables, global and static variables, enum identifiers, typedefs, goto labels and the tag names of structures, unions and enums. The Embedded Development Studio will then use this information to allow you to quickly navigate through your source code.
Comprehensive debug records are generated by the compiler and embedded in the object files for each module. The linker combines these to produce program wide debugging records.
| abs( ) | coshl( ) | gets( ) | longjmp() | sin( ) | strrchr( ) |
| acos( ) | exit( ) | getvect( ) | ltoa() | sinl( ) | strspn( ) |
| acosl( ) | exp( ) | gmtime() | malloc( ) | sinh( ) | strstr( ) |
| alloca() | expl( ) | isalnum( ) | memchr( ) | sinhl( ) | tan( ) |
| asctime() | fabs( ) | isalpha( ) | memcmp( ) | sprintf( ) | tanl( ) |
| asin( ) | fabsl( ) | isascii( ) | memcpy( ) | sqrt( ) | tanh( ) |
| asinl( ) | _fcvt( ) | iscntrl( ) | memmove( ) | sqrtl( ) | tanhl( ) |
| atan( ) | ferroe( ) | isdigit( ) | memset( ) | srand( ) | time( ) |
| atanl( ) | fgetc( ) | isgraph( ) | mktime() | sscanf( ) | toascii( ) |
| atoi( ) | fgets( ) | islower( ) | pow( ) | strcat( ) | tolower( ) |
| atol( ) | fileno( ) | isprint( ) | powl( ) | sscanf( ) | toupper( ) |
| atoff( ) | floor( ) | ispuntc( ) | printf( ) | strchr( ) | ultoa() |
| atolf( ) | floorl( ) | isspace( ) | putc( ) | strcmp( ) | ungetc( ) |
| calloc() | fprintf( ) | isupper( ) | putchar( ) | strcpy( ) | ungetch( ) |
| clearerr( ) | fputc() | isxdigit( ) | puts( ) | strcspn( ) | vfprintf() |
| ceil( ) | fputs() | labs( ) | qsort() | stricmp() | vprintf() |
| ceill( ) | free( ) | ldexp() | rand( ) | strlen( ) | vsprintf() |
| clock() | fscanf( ) | log( ) | sbrk( ) | strncat( ) | |
| cos( ) | getc( ) | logl( ) | scanf( ) | strncmp( ) | |
| cosl( ) | getchar( ) | log10( ) | setjmp() | strncpy( ) | |
| cosh( ) | getche( ) | log10l( ) | setvect( ) | strpbrk( ) |