15 * Standard scanner routine used by parser.c and some output
16 * formats. It keeps a succession of temporary-storage strings in
17 * stdscan_tempstorage, which can be cleared using stdscan_reset.
19 static char **stdscan_tempstorage
= NULL
;
20 static int stdscan_tempsize
= 0, stdscan_templen
= 0;
21 #define STDSCAN_TEMP_DELTA 256
23 static void stdscan_pop(void)
25 nasm_free(stdscan_tempstorage
[--stdscan_templen
]);
28 void stdscan_reset(void)
30 while (stdscan_templen
> 0)
35 * Unimportant cleanup is done to avoid confusing people who are trying
36 * to debug real memory leaks
38 void stdscan_cleanup(void)
41 nasm_free(stdscan_tempstorage
);
44 static char *stdscan_copy(char *p
, int len
)
48 text
= nasm_malloc(len
+ 1);
49 strncpy(text
, p
, len
);
52 if (stdscan_templen
>= stdscan_tempsize
) {
53 stdscan_tempsize
+= STDSCAN_TEMP_DELTA
;
54 stdscan_tempstorage
= nasm_realloc(stdscan_tempstorage
,
58 stdscan_tempstorage
[stdscan_templen
++] = text
;
63 char *stdscan_bufptr
= NULL
;
64 int stdscan(void *private_data
, struct tokenval
*tv
)
66 char ourcopy
[MAX_KEYWORD
+ 1], *r
, *s
;
68 (void)private_data
; /* Don't warn that this parameter is unused */
70 while (isspace(*stdscan_bufptr
))
73 return tv
->t_type
= 0;
75 /* we have a token; either an id, a number or a char */
76 if (isidstart(*stdscan_bufptr
) ||
77 (*stdscan_bufptr
== '$' && isidstart(stdscan_bufptr
[1]))) {
78 /* now we've got an identifier */
81 if (*stdscan_bufptr
== '$') {
87 /* read the entire buffer to advance the buffer pointer but... */
88 while (isidchar(*stdscan_bufptr
))
91 /* ... copy only up to IDLEN_MAX-1 characters */
92 tv
->t_charptr
= stdscan_copy(r
, stdscan_bufptr
- r
< IDLEN_MAX
?
93 stdscan_bufptr
- r
: IDLEN_MAX
- 1);
95 if (is_sym
|| stdscan_bufptr
- r
> MAX_KEYWORD
)
96 return tv
->t_type
= TOKEN_ID
; /* bypass all other checks */
98 for (s
= tv
->t_charptr
, r
= ourcopy
; *s
; s
++)
101 /* right, so we have an identifier sitting in temp storage. now,
102 * is it actually a register or instruction name, or what? */
103 return nasm_token_hash(ourcopy
, tv
);
104 } else if (*stdscan_bufptr
== '$' && !isnumchar(stdscan_bufptr
[1])) {
106 * It's a $ sign with no following hex number; this must
107 * mean it's a Here token ($), evaluating to the current
108 * assembly location, or a Base token ($$), evaluating to
109 * the base of the current segment.
112 if (*stdscan_bufptr
== '$') {
114 return tv
->t_type
= TOKEN_BASE
;
116 return tv
->t_type
= TOKEN_HERE
;
117 } else if (isnumstart(*stdscan_bufptr
)) { /* now we've got a number */
120 bool is_float
= false;
126 if (*stdscan_bufptr
== '$') {
132 c
= *stdscan_bufptr
++;
134 if (!is_hex
&& (c
== 'e' || c
== 'E')) {
136 if (*stdscan_bufptr
== '+' || *stdscan_bufptr
== '-') {
137 /* e can only be followed by +/- if it is either a
138 prefixed hex number or a floating-point number */
142 } else if (c
== 'H' || c
== 'h' || c
== 'X' || c
== 'x') {
144 } else if (c
== 'P' || c
== 'p') {
146 if (*stdscan_bufptr
== '+' || *stdscan_bufptr
== '-')
148 } else if (isnumchar(c
) || c
== '_')
155 stdscan_bufptr
--; /* Point to first character beyond number */
157 if (has_e
&& !is_hex
) {
158 /* 1e13 is floating-point, but 1e13h is not */
163 tv
->t_charptr
= stdscan_copy(r
, stdscan_bufptr
- r
);
164 return tv
->t_type
= TOKEN_FLOAT
;
166 r
= stdscan_copy(r
, stdscan_bufptr
- r
);
167 tv
->t_integer
= readnum(r
, &rn_error
);
170 return tv
->t_type
= TOKEN_ERRNUM
; /* some malformation occurred */
171 tv
->t_charptr
= NULL
;
172 return tv
->t_type
= TOKEN_NUM
;
174 } else if (*stdscan_bufptr
== '\'' || *stdscan_bufptr
== '"') {
175 /* a char constant */
176 char quote
= *stdscan_bufptr
++, *r
;
178 r
= tv
->t_charptr
= stdscan_bufptr
;
179 while (*stdscan_bufptr
&& *stdscan_bufptr
!= quote
)
181 tv
->t_inttwo
= stdscan_bufptr
- r
; /* store full version */
182 if (!*stdscan_bufptr
)
183 return tv
->t_type
= TOKEN_ERRNUM
; /* unmatched quotes */
184 stdscan_bufptr
++; /* skip over final quote */
185 tv
->t_integer
= readstrnum(r
, tv
->t_inttwo
, &rn_warn
);
186 /* FIXME: rn_warn is not checked! */
187 return tv
->t_type
= TOKEN_NUM
;
188 } else if (*stdscan_bufptr
== ';') { /* a comment has happened - stay */
189 return tv
->t_type
= 0;
190 } else if (stdscan_bufptr
[0] == '>' && stdscan_bufptr
[1] == '>') {
192 return tv
->t_type
= TOKEN_SHR
;
193 } else if (stdscan_bufptr
[0] == '<' && stdscan_bufptr
[1] == '<') {
195 return tv
->t_type
= TOKEN_SHL
;
196 } else if (stdscan_bufptr
[0] == '/' && stdscan_bufptr
[1] == '/') {
198 return tv
->t_type
= TOKEN_SDIV
;
199 } else if (stdscan_bufptr
[0] == '%' && stdscan_bufptr
[1] == '%') {
201 return tv
->t_type
= TOKEN_SMOD
;
202 } else if (stdscan_bufptr
[0] == '=' && stdscan_bufptr
[1] == '=') {
204 return tv
->t_type
= TOKEN_EQ
;
205 } else if (stdscan_bufptr
[0] == '<' && stdscan_bufptr
[1] == '>') {
207 return tv
->t_type
= TOKEN_NE
;
208 } else if (stdscan_bufptr
[0] == '!' && stdscan_bufptr
[1] == '=') {
210 return tv
->t_type
= TOKEN_NE
;
211 } else if (stdscan_bufptr
[0] == '<' && stdscan_bufptr
[1] == '=') {
213 return tv
->t_type
= TOKEN_LE
;
214 } else if (stdscan_bufptr
[0] == '>' && stdscan_bufptr
[1] == '=') {
216 return tv
->t_type
= TOKEN_GE
;
217 } else if (stdscan_bufptr
[0] == '&' && stdscan_bufptr
[1] == '&') {
219 return tv
->t_type
= TOKEN_DBL_AND
;
220 } else if (stdscan_bufptr
[0] == '^' && stdscan_bufptr
[1] == '^') {
222 return tv
->t_type
= TOKEN_DBL_XOR
;
223 } else if (stdscan_bufptr
[0] == '|' && stdscan_bufptr
[1] == '|') {
225 return tv
->t_type
= TOKEN_DBL_OR
;
226 } else /* just an ordinary char */
227 return tv
->t_type
= (uint8_t)(*stdscan_bufptr
++);