ref: 4c4c232a9f307952470065c30c82048decb8c6fd
parent: 4deece51bfadf521d8509a1bb62531f9937719ab
author: Roberto E. Vargas Caballero <k0ga@shike2.com>
date: Tue Apr 22 12:24:32 EDT 2014
Separate char and signed/unsigned char They are different types, and this small difference can be important if you make a pointer from them, because in this case the pointers are not equivalent between them.
--- a/types.c
+++ b/types.c
@@ -23,6 +23,12 @@
.defined = 1,
.u.rank = RANK_BOOL
},
+ *schartype = &(Type) {+ .op = INT,
+ .letter = 'C',
+ .defined = 1,
+ .u.rank = RANK_SCHAR
+ },
*uchartype = &(Type) {.op = INT,
.letter = 'M',
@@ -32,9 +38,10 @@
},
*chartype = &(Type) {.op = INT,
- .letter = 'C',
+ .letter = 'M',
+ .sign = 1,
.defined = 1,
- .u.rank = RANK_SCHAR
+ .u.rank = RANK_UCHAR
},
*ushortype = &(Type) {.op = INT,
@@ -109,27 +116,25 @@
Type *
ctype(int8_t type, int8_t sign, int8_t size)
{- if (type == CHAR && !sign)
- sign = options.charsign;
- if (sign == SIGNED)
- sign = 0;
if (type == DOUBLE)
type = FLOAT, size += LONG;
switch (type) {- case VOID: return voidtype;
- case BOOL: return booltype;
- case CHAR: return (sign) ? uchartype : chartype;
+ case CHAR: if (sign == 0)
+ return chartype;
+ return (sign == UNSIGNED) ? uchartype : schartype;
+ case VOID: return voidtype;
+ case BOOL: return booltype;
case INT: switch (size) {- case 0: return (sign) ? uinttype : inttype;
- case SHORT: return (sign) ? ushortype : shortype;
- case LONG: return (sign) ? ulongtype : longtype;
- case LONG+LONG: return (sign) ? ullongtype : llongtype;
+ case 0: return (sign == UNSIGNED) ? uinttype : inttype;
+ case SHORT: return (sign == UNSIGNED) ? ushortype : shortype;
+ case LONG: return (sign == UNSIGNED) ? ulongtype : longtype;
+ case LONG+LONG: return (sign == UNSIGNED) ? ullongtype : llongtype;
}
case FLOAT: switch (size) {- case 0: return floattype;
- case LONG: return doubletype;
- case LONG+LONG: return ldoubletype;
+ case 0: return floattype;
+ case LONG: return doubletype;
+ case LONG+LONG: return ldoubletype;
}
}
}
--
⑨