Hi Bruno,
Here are the implementation for both Pil32 and Pil64:
Pil32:
// (low? 'any) -> sym | NIL
any doLowQ(any x) {
x = cdr(x);
return isSym(x = EVAL(car(x))) && isLowc(symChar(name(x)))? x : Nil;
}
// (upp? 'any) -> sym | NIL
any doUppQ(any x) {
x = cdr(x);
return isSym(x = EVAL(car(x))) && isUppc(symChar(name(x)))? x : Nil;
}
Pil64:
# (low? 'any) -> sym | NIL
(code 'doLowQ 2)
ld E ((E CDR)) # Get arg
eval # Eval it
num E # Number?
jnz retNil # Yes
sym E # Symbol?
jz retNil # No
call firstCharE_A # Get first character
cmp A TOP # Special "top" character?
jeq retNil # Yes
call caseDataA_AC # Get case info
and B (hex "1F") # Character type
cmp B CHAR_LOWERCASE # Lower case?
ldnz E Nil # No
ret
# (upp? 'any) -> sym | NIL
(code 'doUppQ 2)
ld E ((E CDR)) # Get arg
eval # Eval it
num E # Number?
jnz retNil # Yes
sym E # Symbol?
jz retNil # No
call firstCharE_A # Get first character
cmp A TOP # Special "top" character?
jeq retNil # Yes
call caseDataA_AC # Get case info
and B (hex "1F") # Character type
cmp B CHAR_UPPERCASE # Lower case?
ldnz E Nil # No
ret
>From what I understand they are both using same mechanism but Pil32 is using C
>programming while Pil64 is using PilASM.
So to understand better lets just dig with Pil32, here is the ingredients:
static inline int charType(int c) {return Data[Blocks[c>>5]+c & 0xFFFF] & 0x1F;}
static inline bool isLowc(int c) {return charType(c) == CHAR_LOWERCASE;}
static inline bool isUppc(int c) {return charType(c) == CHAR_UPPERCASE;}
So there is an array there so I believe your understanding is correct. And when
looking deeper with Pil64 it mentions about
"Case mappings from the GNU Kaffe Project" so I think this is where the case
info integrated.
BR,
geo
--
UNSUBSCRIBE: mailto:[email protected]?subject=Unsubscribe
