Re: [avr-gcc-list] unexpected compiler behaviour

[Brian Dean]

On Tue, Sep 03, 2002 at 01:19:10PM +0200, Joerg Wunsch wrote:

> Perhaps it would make some sense to
> write a second lcd_putstring_p() for this purpose.

I've found that I end up doing this type of thing a lot, so instead of
writing sister functions for the PROGMEM case, I wrote an accessor
function in which you pass in the memory type and the address, and it
returns the data byte at that address, i.e.:

        unsigned char memio_r(unsigned char memtype, uint16_t addr)
        {
          unsigned char c;
        
          if (memtype == MT_RAM) {
            c = *((unsigned char *)addr);
          }
          else if (memtype == MT_FLASH) {
            c = __lpm_inline(addr);
          }
          else if (memtype == MT_EEPROM) {
            c = eeprom_rb(addr);
          }
          else {
            c = 0;
          }
        
          return c;
        }

Then, in your various 'puts()' routines, do something like this:

        void puts(uint8_t memtype, const char *s)
        {
          uint16_t a = (uint16_t)s;
          char c;
        
          for (c = memio_r(memtype, a); c; c = memio_r(memtype, ++a)) {
            putc(c);
          }
        }

Thus, you can use the same function(s) whether your output data is
coming from RAM, FLASH, or EEPROM.

I find this handy ... just thought I'd share it.

Cheers,
-Brian
-- 
Brian Dean                                      address@hidden
avr-gcc-list at http://avr1.org