Factor out size tokens and annotate with the corresponding size

There is space in the token table to explicitly encode the size
corresponding to a size token. We might as well do so...

Signed-off-by: H. Peter Anvin <hpa@zytor.com>

3 files changed, 42 insertions, 21 deletions

diff --git a/asm/parser.c b/asm/parser.c
index c76696c2..9349e25e 100644
--- a/asm/parser.c
+++ b/asm/parser.c
@@ -767,7 +767,8 @@ is_expression:
         }
         first = false;
         op->type = 0; /* so far, no override */
-        while (i == TOKEN_SPECIAL) {    /* size specifiers */
+        /* size specifiers */
+        while (i == TOKEN_SPECIAL || i == TOKEN_SIZE) {
             switch (tokval.t_integer) {
             case S_BYTE:
                 if (!setsize)   /* we want to use only the first */
@@ -835,7 +836,8 @@ is_expression:
             mref = true;
             bracket = (i == '[');
             i = stdscan(NULL, &tokval); /* then skip the colon */
-            while (i == TOKEN_SPECIAL || i == TOKEN_PREFIX) {
+            while (i == TOKEN_SPECIAL || i == TOKEN_SIZE ||
+                   i == TOKEN_PREFIX) {
                 process_size_override(result, op);
                 i = stdscan(NULL, &tokval);
             }
@@ -880,7 +882,8 @@ is_expression:
             }
 
             i = stdscan(NULL, &tokval); /* then skip the colon */
-            while (i == TOKEN_SPECIAL || i == TOKEN_PREFIX) {
+            while (i == TOKEN_SPECIAL || i == TOKEN_SIZE ||
+                   i == TOKEN_PREFIX) {
                 process_size_override(result, op);
                 i = stdscan(NULL, &tokval);
             }
diff --git a/asm/tokens.dat b/asm/tokens.dat
index d5345e19..bd89d0a0 100644
--- a/asm/tokens.dat
+++ b/asm/tokens.dat
@@ -60,24 +60,26 @@ xrelease
 bnd
 nobnd
 
-% TOKEN_SPECIAL, 0, 0, S_*
-abs
+% TOKEN_SIZE, 0, SIZE_*, S_*
 byte
+word
 dword
+qword
+tword
+oword
+yword
+zword
+
+% TOKEN_SPECIAL, 0, 0, S_*
+abs
 far
 long
 near
 nosplit
-oword
-qword
 rel
 short
 strict
 to
-tword
-word
-yword
-zword
 
 % TOKEN_ID, 0, TFLAG_WARN, 0
 ptr
diff --git a/include/nasm.h b/include/nasm.h
index eb376b16..c2a80562 100644
--- a/include/nasm.h
+++ b/include/nasm.h
@@ -167,7 +167,8 @@ enum token_type { /* token types, other than chars */
     TOKEN_INSN,         /* instruction name */
     TOKEN_HERE,         /* $ */
     TOKEN_BASE,         /* $$ */
-    TOKEN_SPECIAL,      /* BYTE, WORD, DWORD, QWORD, FAR, NEAR, etc */
+    TOKEN_SIZE,		/* BYTE, WORD, DWORD, QWORD, etc */
+    TOKEN_SPECIAL,      /* REL, FAR, NEAR, STRICT, NOSPLIT, etc */
     TOKEN_PREFIX,       /* A32, O16, LOCK, REPNZ, TIMES, etc */
     TOKEN_SHL,          /* << or <<< */
     TOKEN_SHR,          /* >> */
@@ -1081,25 +1082,40 @@ extern const struct dfmt *dfmt;
 
 #define TYS_ELEMENTS(x) ((x) << 8)
 
+/* Sizes corresponding to various tokens */
+enum byte_sizes {
+    SIZE_BYTE	=  1,
+    SIZE_WORD	=  2,
+    SIZE_DWORD	=  4,
+    SIZE_QWORD	=  8,
+    SIZE_TWORD  = 10,
+    SIZE_OWORD  = 16,
+    SIZE_YWORD  = 32,
+    SIZE_ZWORD  = 64
+};
+
 enum special_tokens {
-    SPECIAL_ENUM_START  = PREFIX_ENUM_LIMIT,
-    S_ABS               = SPECIAL_ENUM_START,
-    S_BYTE,
+    SIZE_ENUM_START     = PREFIX_ENUM_LIMIT,
+    S_BYTE              = SIZE_ENUM_START,
+    S_WORD,
     S_DWORD,
+    S_QWORD,
+    S_TWORD,
+    S_OWORD,
+    S_YWORD,
+    S_ZWORD,
+    SIZE_ENUM_LIMIT,
+
+    SPECIAL_ENUM_START  = SIZE_ENUM_LIMIT,
+    S_ABS		= SPECIAL_ENUM_START,
     S_FAR,
     S_LONG,
     S_NEAR,
     S_NOSPLIT,
-    S_OWORD,
-    S_QWORD,
     S_REL,
     S_SHORT,
     S_STRICT,
     S_TO,
-    S_TWORD,
-    S_WORD,
-    S_YWORD,
-    S_ZWORD,
     SPECIAL_ENUM_LIMIT
 };