> Works, though maybe BillW's version IS better
Not if you write mystrncpy(buffer,"Hello",59000).

I've heard C described as "machine-independent assembly language",
which I think says it pretty well.

C-compilers aren't magic, the onus is still on you to ensure that it isn't
going to generate crap.

Trival example: you know and I know that "x/8" is the same as "x>>3". But
does the compiler know? If it doesn't then you'll end up with something
like "call __divideint16" instead of "lsr 3" -- perhaps a 5000% speed
penalty just because of the way you typed it. OK, so this isn't really
relevant anymore, ALL compilers are smart enough to get this one.

But in the case of mystrncpy(), you know and I know that
"while(n++<count)" is functionally identical to "while(count--)". But the
compiler almost certainly will not know. It will do pretty much what you
told it to: create, initialize and increment a local variable and
repeatedly compare it to another local variable.

The problem with C is that it's not obvious what a HUGE impact subtle
syntactical differences can have, you need to look at the .LST (if you
even GET a .LST).

But the rule of thumb is "the smaller the source, the faster the code".
Just comment like crazy.

> some bosses simply cannot DEAL with modifying a function parameter as
> you're doing there
Different problem. I like to say two words to micro-managing bosses (the
second word is "off"). A local variable is a local variable, whether
passed on the stack under the return address or allocated on top of it, or
passed in register or whatever -- the point is, it's LOCAL, use it as you
wish.

> (I'd maybe CONST that puppy?)
You could, although IMHO that's not really what const is for. Take for
example:

const char string_in_rom[]="Hello";

void erase(char *str)
(
  *str = 0;
}

const simply tells the compiler than the string is unwriteable, and, by
exclusion, that erase() can write to str.

Someday, if (when) I headspace and try to erase(string_in_rom), then the
compiler can tell me to, um, cease micro-managing.

-- Rich