Mercurial > hg-old > index.cgi
annotate lwasm/parse.c @ 253:c537a3a723fc 2.x
Fixed obscure comment detection bug
author | lost |
---|---|
date | Wed, 09 Dec 2009 03:59:26 +0000 |
parents | 0986834ec7d3 |
children | c7a41b4c89b3 |
rev | line source |
---|---|
36 | 1 /* |
2 parse.c | |
3 Copyright © 2008 William Astle | |
4 | |
5 This file is part of LWASM. | |
6 | |
7 LWASM is free software: you can redistribute it and/or modify it under the | |
8 terms of the GNU General Public License as published by the Free Software | |
9 Foundation, either version 3 of the License, or (at your option) any later | |
10 version. | |
11 | |
12 This program is distributed in the hope that it will be useful, but WITHOUT | |
13 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or | |
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for | |
15 more details. | |
16 | |
17 You should have received a copy of the GNU General Public License along with | |
18 this program. If not, see <http://www.gnu.org/licenses/>. | |
19 */ | |
20 | |
21 /* | |
22 Contains the general parser | |
23 */ | |
24 | |
25 #define __parse_c_seen__ | |
212 | 26 #include <config.h> |
36 | 27 |
28 #include <ctype.h> | |
29 #include <string.h> | |
30 | |
31 #include "lwasm.h" | |
32 #include "instab.h" | |
33 #include "util.h" | |
34 | |
35 // parse a line and despatch to the appropriate handlers for opcodes | |
36 int lwasm_parse_line(asmstate_t *as, lwasm_line_t *l) | |
37 { | |
38 char *p, *p2; | |
39 char *opc; | |
40 int opnum; | |
44 | 41 char *sym = NULL; |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
42 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
43 // if this was a bad op first pass (or otherwise a no-process line) |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
44 // ignore it |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
45 if (l -> badop) |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
46 return; |
36 | 47 |
48 p = l -> text; | |
101
f59c0916753d
Fixed relative branches and PCR addressing to handle constant intra-section references properly
lost
parents:
99
diff
changeset
|
49 l -> sect = as -> csect; |
236
a58f49a77441
Added os9 target, pragma to control whether $ localizes a symbol, and fixed some condition nesting bugs
lost
parents:
230
diff
changeset
|
50 l -> inmod = as -> inmod; |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
51 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
52 // blank lines are a no brainer |
36 | 53 if (!*p) |
37
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
54 { |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
55 as -> context = lwasm_next_context(as); |
36 | 56 return 0; |
37
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
57 } |
36 | 58 |
52 | 59 // for output generation later but only on pass 1 |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
60 // also used by some pseudo ops on pass 2 |
52 | 61 if (as -> passnum == 1) |
62 l -> codeaddr = as -> addr; | |
44 | 63 |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
64 // if it's a comment, return (this doesn't cause a context change) |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
65 if (*p == '*' || *p == ';') |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
66 return; |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
67 |
169 | 68 // if we have C pre-processor directives/output, ignore it |
69 if (*p == '#') | |
70 return; | |
71 | |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
72 // if we start with a non-space character, it's a symbol |
244
c8bcc396ec59
Allow symbols to appear after whitespace when terminated by :
lost
parents:
236
diff
changeset
|
73 symbolagain: |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
74 if (!isspace(*p)) |
36 | 75 { |
253 | 76 // if it starts with a comment character, it's not a symbol |
77 if (*p == ';' || *p == '*') | |
78 return; | |
79 | |
36 | 80 // we have a symbol specified here |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
81 // parse it out and record it for later use |
157 | 82 for (p2 = p; *p2 && !isspace(*p2) && *p2 != ':'; p2++) |
37
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
83 /* do nothing */ ; |
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
84 |
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
85 sym = lwasm_alloc((p2 - p) + 1); |
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
86 sym[p2 - p] = '\0'; |
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
87 memcpy(sym, p, p2 - p); |
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
88 |
44 | 89 p = p2; |
157 | 90 if (!*sym) |
91 { | |
92 register_error(as, l, 1, "Invalid symbol"); | |
93 lwasm_free(sym); | |
94 sym = NULL; | |
95 } | |
96 if (*p == ':') | |
97 p++; | |
37
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
98 } |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
99 l -> sym = sym; |
36 | 100 |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
101 // now skip any whitespace to find the opcode |
36 | 102 while (*p && isspace(*p)) |
103 p++; | |
104 | |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
105 // is the line blank? |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
106 if (!*p && !sym) |
37
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
107 { |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
108 // nothing but white space *is* a context break |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
109 as -> context = lwasm_next_context(as); |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
110 return; |
37
538e15927776
Added symbol handling to expression subsystem; adpated instruction handlers to the new scheme; misc fixes
lost
parents:
36
diff
changeset
|
111 } |
36 | 112 |
113 // parse the opcode | |
114 for (p2 = p; *p2 && !isspace(*p2); p2++) | |
115 /* do nothing */ ; | |
116 | |
117 opc = lwasm_alloc((p2 - p) + 1); | |
118 memcpy(opc, p, p2 - p); | |
119 opc[p2 - p] = '\0'; | |
120 | |
244
c8bcc396ec59
Allow symbols to appear after whitespace when terminated by :
lost
parents:
236
diff
changeset
|
121 if (strchr(opc, ':')) |
c8bcc396ec59
Allow symbols to appear after whitespace when terminated by :
lost
parents:
236
diff
changeset
|
122 { |
c8bcc396ec59
Allow symbols to appear after whitespace when terminated by :
lost
parents:
236
diff
changeset
|
123 lwasm_free(opc); |
c8bcc396ec59
Allow symbols to appear after whitespace when terminated by :
lost
parents:
236
diff
changeset
|
124 goto symbolagain; |
c8bcc396ec59
Allow symbols to appear after whitespace when terminated by :
lost
parents:
236
diff
changeset
|
125 } |
c8bcc396ec59
Allow symbols to appear after whitespace when terminated by :
lost
parents:
236
diff
changeset
|
126 |
190 | 127 l -> forceglobal = 0; |
128 // if the opcode contains an =, treat it as "symbol = expr" | |
129 if (!sym && strchr(opc, '=')) | |
130 { | |
131 for (p2 = opc; *p2 && *p2 != '='; p2++) | |
132 /* do nothing */ ; | |
133 sym = lwasm_alloc((p2 - opc) + 1); | |
134 memcpy(sym, opc, p2 - opc); | |
135 sym[p2 - opc] = '\0'; | |
136 l -> sym = sym; | |
137 | |
138 p2 = p + (p2 - opc) + 1; | |
139 // p2++; | |
140 opc[0] = '='; | |
141 opc[1] = '\0'; | |
142 debug_message(2, "Found opcode = with symbol %s and operand %s", sym, p2); | |
143 l -> forceglobal = 1; | |
144 } | |
145 | |
38 | 146 debug_message(2, "Found operation code: '%s'", opc); |
147 | |
36 | 148 // skip intervening whitespace if present |
149 while (*p2 && isspace(*p2)) | |
150 p2++; | |
151 | |
152 // look up instruction in insn table | |
153 for (opnum = 0; instab[opnum].opcode; opnum++) | |
154 { | |
155 if (!strcasecmp(instab[opnum].opcode, opc)) | |
156 break; | |
157 } | |
158 | |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
159 // if we found no operation, check if we had a comment |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
160 // the reason this check is here is to allow for "private" |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
161 // operation codes like "*pragma" which will be ignored by |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
162 // other assemblers |
68 | 163 // also skip empty ops |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
164 if (!(instab[opnum].opcode)) |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
165 { |
68 | 166 if (*opc == '*' || *opc == ';' || !*opc) |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
167 goto done_line; |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
168 } |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
169 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
170 // now we have the opcode and the symbol, we can decide if we're |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
171 // actually going to do anything with this line |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
172 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
173 // we will NOT call the function if any of the following are true: |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
174 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
175 // - we are skipping a condition and the operation code is not a conditional |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
176 // - we are defining a macro and the operation code is not ENDM |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
177 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
178 // we will call the function in any other circumstance |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
179 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
180 // first condition above |
57 | 181 if (as -> inmacro && instab[opnum].endm == 0) |
182 { | |
183 add_macro_line(as, l -> text); | |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
184 goto done_line; |
57 | 185 } |
186 | |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
187 // second condition above |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
188 if (as -> skipcond && instab[opnum].iscond == 0) |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
189 goto done_line; |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
190 |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
191 // we've registered the symbol as needed |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
192 // now we need to check for a macro call IFF we don't collide with |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
193 // an operation code; otherwise, call the operation function |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
194 if (instab[opnum].opcode) |
36 | 195 { |
230
0df2a39a268c
Added --6809/--6309 switches and some cleanups in error reporting
lost
parents:
223
diff
changeset
|
196 if (instab[opnum].fn && !(as -> no6309 && instab[opnum].is6309)) |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
197 { |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
198 (instab[opnum].fn)(as, l, &p2, opnum); |
223
436b36214e35
Fixed lack of error when there are extraneous characters at the end of the operand.
lost
parents:
212
diff
changeset
|
199 |
436b36214e35
Fixed lack of error when there are extraneous characters at the end of the operand.
lost
parents:
212
diff
changeset
|
200 // if we didn't end on a "space" character or EOL, throw error |
250
0986834ec7d3
Added no-op .bank pseudo-op to support compiling gcc6809
lost
parents:
244
diff
changeset
|
201 if (p2 && *p2 && !isspace(*p2) && !(l -> err) && as -> passnum == 1) |
236
a58f49a77441
Added os9 target, pragma to control whether $ localizes a symbol, and fixed some condition nesting bugs
lost
parents:
230
diff
changeset
|
202 register_error(as, l, 1, "Bad operand: %s (%d)", p2, as -> passnum); |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
203 } |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
204 else |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
205 { |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
206 // carp about unimplemented operation |
230
0df2a39a268c
Added --6809/--6309 switches and some cleanups in error reporting
lost
parents:
223
diff
changeset
|
207 if (instab[opnum].is6309) |
0df2a39a268c
Added --6809/--6309 switches and some cleanups in error reporting
lost
parents:
223
diff
changeset
|
208 register_error(as, l, 1, "Use of 6309 operation code: %s", opc); |
0df2a39a268c
Added --6809/--6309 switches and some cleanups in error reporting
lost
parents:
223
diff
changeset
|
209 else |
0df2a39a268c
Added --6809/--6309 switches and some cleanups in error reporting
lost
parents:
223
diff
changeset
|
210 register_error(as, l, 1, "Unimplemented operation code: %s", opc); |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
211 } |
36 | 212 } |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
213 else |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
214 { |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
215 if (expand_macro(as, l, &p2, opc) == 0) |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
216 goto done_line; |
36 | 217 |
66
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
218 // carp about an unknown operation code and note that fact for |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
219 // pass 2 in case a macro appears later with the same name! |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
220 register_error(as, l, 1, "Uknown operation code: %s", opc); |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
221 l -> badop = 1; |
aa9d9fedfdf4
Redid lwasm_parse_line() to correct overly complex logic flaws
lost
parents:
64
diff
changeset
|
222 } |
57 | 223 |
224 done_line: | |
68 | 225 if (!(as -> skipcond || as -> inmacro)) |
226 { | |
227 // register symbol if the operation didn't | |
228 if (sym && instab[opnum].setsym == 0) | |
229 { | |
230 if (as -> passnum == 1) | |
231 { | |
71
90a5657d5408
Fixed problem with symbols being registered with the address of the NEXT instruction
lost
parents:
68
diff
changeset
|
232 debug_message(1, "Registering symbol '%s' at %04X", sym, l -> codeaddr); |
90a5657d5408
Fixed problem with symbols being registered with the address of the NEXT instruction
lost
parents:
68
diff
changeset
|
233 if (lwasm_register_symbol(as, l, sym, l -> codeaddr, SYMBOL_NORM) < 0) |
68 | 234 l -> sym = NULL; |
235 else | |
236 l -> addrset = 1; | |
237 } | |
238 } | |
239 } | |
99 | 240 |
85 | 241 l -> sect = as -> csect; |
91
718998b673ee
Added incomplete references to object output and added support for section base terms in expression handler
lost
parents:
85
diff
changeset
|
242 l -> context = as -> context; |
85 | 243 |
36 | 244 lwasm_free(opc); |
44 | 245 if (sym) |
246 lwasm_free(sym); | |
36 | 247 } |