BoostC Explaining Dot H Files
Draft Not quite Ready
This is my russ_hensel explanation of *.h. Others are welcome to contribute. It is a work in progress, and may or may not be correct.
Should almost always be included in all your files. It has limited functionality, largely a bridge to other include files. For the boostC compiler these files are:
- BoostCPic18.h defines for the 18 series chips
- BoostCPic16.h defines for the 16 series chips
- Boostc.h defines and other utility stuff for the compiler.
Apparently the compiler reads the target chip and effectively #defines a value for _PIC16F722 or whatever for the chip that you have designated the target chip.
Links for the macro processor and related
http://publications.gbdirect.co.uk/c_book/chapter2/textual_program_structure.html 2.3. The Textual Structure of Programs
http://publications.gbdirect.co.uk/c_book/chapter7/ 2.3 The Textual Structure of Programs
BoostC.h is almost always included, but via System.h, so you may not “see” the include. The following refers to section within BoostC.h, pull up a copy while you are reading this.
#define clear_bit( reg, bitNumb ) ((reg) &= ~(1 << (bitNumb)))
These are inline functions using #defines with arguments. ( see [] ) many C programmers use these expressions directly in there code, but using clear_bit instead of equivalent is bothe easier to read and write. It make the intent of the program cleared, and may make the code easier to port.
#define MAKESHORT( dst, lobyte, hibyte ) asm movf _##lobyte, W \ asm movwf _##dst \ asm movf _##hibyte, W \ asm movwf _##dst##+1
Again an inline functions using #define with arguments. There are few other things going on here:
- using the asm keyword to directly code assembler code
- using \ to make the multiline macro be read as a single line
- using ## for Token pasting.
with token pasting when
asm movf _##lobyte, W
is processed in MAKESORHT( A, B, C ) becomes
asm movf _B, W
_B in asm represents the addres of the variable B in the C program.
Additionally from reading this macro we can see that BoostC stores the low byte in low memory and the high byte in higher memory. This is called bigendien or the reverse, I need to look it up. Useful to know if you want to do byte manipulations on multibyte data.
Not quite sure how the linker generates these, but I do understand that the linker needs to know the clock speed to do it. This makes it hard to use precompiled libary functions unless a different one was used for clock speed. Including this allows the compiler to compile what ever call is needed in your C code.
These macros offer a way of using assembler without using the assembler key words _asm or asm. Using asm makes sure that the compiler will not optimize out the code of the macro.
Functions used internally
look like multibyte arithmitic, not sure how or why, working on it.
see BoostC Inline Functions for more information on inline functions and #define.
Other project Files
- .cof not a text file, purpose unknown
- .stat text file, purpose: Linker Function Resource Usage Dump
- obj not a text file, purpose unknown
- .lst text assembler code ( no ref to source ) output of xx input to yyy
- .csam source and assembler code output of xx input to yyy
- brws text, symbols etc and their source code reference.
- tree call tree in text format, display in ide by....