Mercurial > hg > index.cgi
annotate lwbasic/parser.c @ 34:bfea77812e64
Start of assignment code
author | Lost Wizard (lost@starbug3) |
---|---|
date | Fri, 04 Feb 2011 21:27:03 -0700 |
parents | 890a8f688889 |
children | cdb0175e1063 |
rev | line source |
---|---|
25 | 1 /* |
2 compiler.c | |
3 | |
4 Copyright © 2011 William Astle | |
5 | |
6 This file is part of LWTOOLS. | |
7 | |
8 LWTOOLS is free software: you can redistribute it and/or modify it under the | |
9 terms of the GNU General Public License as published by the Free Software | |
10 Foundation, either version 3 of the License, or (at your option) any later | |
11 version. | |
12 | |
13 This program is distributed in the hope that it will be useful, but WITHOUT | |
14 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or | |
15 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for | |
16 more details. | |
17 | |
18 You should have received a copy of the GNU General Public License along with | |
19 this program. If not, see <http://www.gnu.org/licenses/>. | |
20 */ | |
21 | |
22 /* | |
23 This is the actual compiler bit; it drives the parser and code generation | |
24 */ | |
25 | |
26 #include <stdio.h> | |
27 | |
26
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
28 #include <lw_alloc.h> |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
29 #include <lw_string.h> |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
30 |
25 | 31 #include "lwbasic.h" |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
32 #include "symtab.h" |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
33 |
34 | 34 static void expect(cstate *state, int tt) |
35 { | |
36 if (state -> lexer_token != tt) | |
37 lwb_error("Expecting %s, got %s\n", lexer_token_name(tt), lexer_return_token(state)); | |
38 lexer(state); | |
39 } | |
40 | |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
41 |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
42 /* size of a type */ |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
43 static int sizeof_type(int type) |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
44 { |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
45 /* everything is an "int" right now; 2 bytes */ |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
46 return 2; |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
47 } |
25 | 48 |
49 /* parse a type; the next token will be acquired as a result */ | |
50 /* the token advancement is to provide consistency */ | |
51 static int parse_type(cstate *state) | |
52 { | |
53 int pt = -1; | |
54 | |
55 switch (state -> lexer_token) | |
56 { | |
57 case token_kw_integer: | |
58 pt = 1; | |
59 break; | |
60 | |
61 default: | |
62 lwb_error("Invalid type specification"); | |
63 } | |
64 lexer(state); | |
65 /* look for "unsigned" modifier for integer types */ | |
66 return pt; | |
67 } | |
68 | |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
69 static void parse_decls(cstate *state) |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
70 { |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
71 /* declarations */ |
33 | 72 /* the first thing that doesn't look like a declaration is assumed */ |
73 /* to be a statement and will trigger a bailout */ | |
74 int vt; | |
75 char *vn; | |
76 symtab_entry_t *se; | |
77 | |
78 for (;;) | |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
79 { |
33 | 80 switch (state -> lexer_token) |
81 { | |
82 /* DIM keyword */ | |
83 case token_kw_dim: | |
84 lexer(state); | |
85 if (state -> lexer_token != token_identifier) | |
86 { | |
87 lwb_error("Expecting identifier, got %s\n", lexer_return_token(state)); | |
88 } | |
89 vn = lw_strdup(state -> lexer_token_string); | |
90 lexer(state); | |
91 if (state -> lexer_token != token_kw_as) | |
92 { | |
93 lwb_error("Expecting AS, got %s\n", lexer_return_token(state)); | |
94 } | |
95 lexer(state); | |
96 vt = parse_type(state); | |
97 | |
98 se = symtab_find(state -> local_syms, vn); | |
99 if (se) | |
100 { | |
101 lwb_error("Multiply defined local variable %s", vn); | |
102 } | |
103 state -> framesize += sizeof_type(vt); | |
104 symtab_register(state -> local_syms, vn, -(state -> framesize), symtype_var, NULL); | |
105 | |
106 lw_free(vn); | |
107 break; | |
34 | 108 |
109 /* blank lines allowed */ | |
110 case token_eol: | |
111 break; | |
112 | |
33 | 113 default: |
114 return; | |
115 } | |
116 if (state -> lexer_token != token_eol) | |
34 | 117 lwb_error("Expecting end of line; got %s\n", lexer_return_token(state)); |
118 lexer(state); | |
119 } | |
120 } | |
121 | |
122 static void parse_statements(cstate *state) | |
123 { | |
124 symtab_entry_t *se; | |
125 | |
126 for (;;) | |
127 { | |
128 switch (state -> lexer_token) | |
129 { | |
130 /* blank lines allowed */ | |
131 case token_eol: | |
132 break; | |
133 | |
134 /* variable assignment */ | |
135 case token_identifier: | |
136 se = symtab_find(state -> local_syms, state -> lexer_token_string); | |
137 if (!se) | |
138 { | |
139 se = symtab_find(state -> global_syms, state -> lexer_token_string); | |
140 } | |
141 if (!se) | |
142 lwb_error("Unknown variable %s\n", state -> lexer_token_string); | |
143 lexer(state); | |
144 expect(state, token_op_assignment); | |
145 | |
146 /* parse the expression */ | |
147 /* parse_expression(state); */ | |
148 | |
149 /* actually do the assignment */ | |
150 | |
151 break; | |
152 | |
153 /* anything we don't recognize as a statement token breaks out */ | |
154 default: | |
155 return; | |
156 } | |
157 if (state -> lexer_token != token_eol) | |
158 lwb_error("Expecting end of line; got %s\n", lexer_return_token(state)); | |
33 | 159 lexer(state); |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
160 } |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
161 } |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
162 |
25 | 163 |
164 /* issub means RETURNS is not allowed; !issub means RETURNS is required */ | |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
165 |
25 | 166 static void parse_subfunc(cstate *state, int issub) |
167 { | |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
168 int pt, rt; |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
169 char *subname, *pn; |
26
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
170 int vis = 0; |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
171 symtab_entry_t *se; |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
172 int paramsize = 0; |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
173 |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
174 state -> local_syms = symtab_init(); |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
175 state -> framesize = 0; |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
176 |
25 | 177 lexer(state); |
178 if (state -> lexer_token != token_identifier) | |
179 { | |
180 lwb_error("Invalid sub name '%s'", state -> lexer_token_string); | |
181 } | |
182 | |
26
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
183 subname = lw_strdup(state -> lexer_token_string); |
25 | 184 |
185 lexer(state); | |
186 if (state -> lexer_token == token_kw_public || state -> lexer_token == token_kw_private) | |
187 { | |
26
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
188 if (state -> lexer_token == token_kw_public) |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
189 vis = 1; |
25 | 190 lexer(state); |
191 } | |
192 | |
193 /* ignore the "PARAMS" keyword if present */ | |
194 if (state -> lexer_token == token_kw_params) | |
195 lexer(state); | |
196 | |
26
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
197 if (state -> lexer_token == token_eol || state -> lexer_token == token_kw_returns) |
25 | 198 goto noparms; |
199 | |
200 paramagain: | |
201 if (state -> lexer_token != token_identifier) | |
202 { | |
31
574931d87abd
Created a function to prettyprint the current lexer token
lost@l-w.ca
parents:
30
diff
changeset
|
203 lwb_error("Parameter name expected, got %s\n", lexer_return_token(state)); |
25 | 204 } |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
205 pn = lw_strdup(state -> lexer_token_string); |
25 | 206 lexer(state); |
207 | |
208 if (state -> lexer_token != token_kw_as) | |
209 lwb_error("Expecting AS\n"); | |
210 lexer(state); | |
211 | |
212 pt = parse_type(state); | |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
213 |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
214 se = symtab_find(state -> local_syms, pn); |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
215 if (se) |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
216 { |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
217 lwb_error("Duplicate parameter name %s\n", pn); |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
218 } |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
219 symtab_register(state -> local_syms, pn, paramsize, symtype_param, NULL); |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
220 paramsize += sizeof_type(pt); |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
221 lw_free(pn); |
25 | 222 |
223 if (state -> lexer_token == token_char && state -> lexer_token_string[0] == ',') | |
224 { | |
225 lexer(state); | |
226 goto paramagain; | |
227 } | |
228 | |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
229 noparms: |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
230 rt = -1; |
25 | 231 if (!issub) |
232 { | |
233 if (state -> lexer_token != token_kw_returns) | |
234 { | |
235 lwb_error("FUNCTION must have RETURNS\n"); | |
236 } | |
237 lexer(state); | |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
238 /* if (state -> lexer_token == token_identifier) |
25 | 239 { |
240 printf("Return value named: %s\n", state -> lexer_token_string); | |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
241 |
25 | 242 lexer(state); |
243 if (state -> lexer_token != token_kw_as) | |
244 lwb_error("Execting AS after RETURNS"); | |
245 lexer(state); | |
246 } | |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
247 */ |
25 | 248 rt = parse_type(state); |
249 } | |
250 else | |
251 { | |
252 if (state -> lexer_token == token_kw_returns) | |
253 { | |
254 lwb_error("SUB cannot specify RETURNS\n"); | |
255 } | |
256 } | |
257 | |
258 | |
259 if (state -> lexer_token != token_eol) | |
260 { | |
31
574931d87abd
Created a function to prettyprint the current lexer token
lost@l-w.ca
parents:
30
diff
changeset
|
261 lwb_error("EOL expected; found %s\n", lexer_return_token(state)); |
25 | 262 } |
26
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
263 |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
264 |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
265 se = symtab_find(state -> global_syms, subname); |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
266 if (se) |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
267 { |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
268 lwb_error("Multiply defined symbol %s\n", subname); |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
269 } |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
270 |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
271 symtab_register(state -> global_syms, subname, -1, issub ? symtype_sub : symtype_func, NULL); |
26
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
272 |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
273 state -> currentsub = subname; |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
274 state -> returntype = rt; |
26
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
275 /* consume EOL */ |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
276 lexer(state); |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
277 |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
278 /* variable declarations */ |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
279 parse_decls(state); |
26
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
280 |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
281 /* output function/sub prolog */ |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
282 emit_prolog(state, vis); |
26
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
283 |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
284 /* parse statement block */ |
34 | 285 parse_statements(state); |
26
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
286 |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
287 if (issub) |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
288 { |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
289 if (state -> lexer_token != token_kw_endsub) |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
290 { |
31
574931d87abd
Created a function to prettyprint the current lexer token
lost@l-w.ca
parents:
30
diff
changeset
|
291 lwb_error("Expecting ENDSUB, got %s\n", lexer_return_token(state)); |
26
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
292 } |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
293 } |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
294 else |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
295 { |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
296 if (state -> lexer_token != token_kw_endfunction) |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
297 { |
31
574931d87abd
Created a function to prettyprint the current lexer token
lost@l-w.ca
parents:
30
diff
changeset
|
298 lwb_error("Expecting ENDFUNCTION, got %s\n", lexer_return_token(state)); |
26
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
299 } |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
300 } |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
301 /* output function/sub epilog */ |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
302 emit_epilog(state); |
26
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
303 |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
304 lw_free(state -> currentsub); |
26aa76da75ad
Additional parsing in function/sub; emission of prolog/epilog code
lost@l-w.ca
parents:
25
diff
changeset
|
305 state -> currentsub = NULL; |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
306 symtab_destroy(state -> local_syms); |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
307 state -> local_syms = NULL; |
25 | 308 } |
309 | |
30
bcd532a90e53
Renamed "compiler" to "parser" for more consistent terminology
lost@l-w.ca
parents:
27
diff
changeset
|
310 void parser(cstate *state) |
25 | 311 { |
312 state -> lexer_curchar = -1; | |
32
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
313 state -> global_syms = symtab_init(); |
49d608aecc4d
Framework for handling local stack frame and/or variables
lost@l-w.ca
parents:
31
diff
changeset
|
314 |
25 | 315 /* now look for a global declaration */ |
316 for (;;) | |
317 { | |
318 state -> parser_state = parser_state_global; | |
319 lexer(state); | |
320 switch (state -> lexer_token) | |
321 { | |
322 case token_kw_function: | |
323 printf("Function\n"); | |
324 parse_subfunc(state, 0); | |
325 break; | |
326 | |
327 case token_kw_sub: | |
328 printf("Sub\n"); | |
329 parse_subfunc(state, 1); | |
330 break; | |
331 | |
332 /* blank lines are allowed */ | |
333 case token_eol: | |
334 continue; | |
335 | |
336 /* EOF is allowed - end of parsing */ | |
337 case token_eof: | |
338 return; | |
339 | |
340 default: | |
31
574931d87abd
Created a function to prettyprint the current lexer token
lost@l-w.ca
parents:
30
diff
changeset
|
341 lwb_error("Invalid token '%s' in global state\n", lexer_return_token(state)); |
25 | 342 } |
343 } | |
344 } |