5 * Permission is hereby granted, free of charge, to any person obtaining a copy of
6 * this software and associated documentation files (the "Software"), to deal in
7 * the Software without restriction, including without limitation the rights to
8 * use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies
9 * of the Software, and to permit persons to whom the Software is furnished to do
10 * so, subject to the following conditions:
12 * The above copyright notice and this permission notice shall be included in all
13 * copies or substantial portions of the Software.
15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
29 * These are not lexical tokens: These are parse tree types. Most people
30 * perform tokenizing on language punctuation which is wrong. That stuff
31 * is technically already tokenized, it just needs to be parsed into a tree
33 #define PARSE_TYPE_DO 0
34 #define PARSE_TYPE_ELSE 1
35 #define PARSE_TYPE_IF 2
36 #define PARSE_TYPE_WHILE 3
37 #define PARSE_TYPE_BREAK 4
38 #define PARSE_TYPE_CONTINUE 5
39 #define PARSE_TYPE_RETURN 6
40 #define PARSE_TYPE_GOTO 7
41 #define PARSE_TYPE_FOR 8
42 #define PARSE_TYPE_VOID 9
43 #define PARSE_TYPE_STRING 10
44 #define PARSE_TYPE_FLOAT 11
45 #define PARSE_TYPE_VECTOR 12
46 #define PARSE_TYPE_ENTITY 13
47 #define PARSE_TYPE_LAND 14
48 #define PARSE_TYPE_LOR 15
49 #define PARSE_TYPE_LTEQ 16
50 #define PARSE_TYPE_GTEQ 17
51 #define PARSE_TYPE_EQEQ 18
52 #define PARSE_TYPE_LNEQ 19
53 #define PARSE_TYPE_COMMA 20
54 #define PARSE_TYPE_LNOT 21
55 #define PARSE_TYPE_STAR 22
56 #define PARSE_TYPE_DIVIDE 23
57 #define PARSE_TYPE_LPARTH 24
58 #define PARSE_TYPE_RPARTH 25
59 #define PARSE_TYPE_MINUS 26
60 #define PARSE_TYPE_ADD 27
61 #define PARSE_TYPE_EQUAL 28
62 #define PARSE_TYPE_LBS 29
63 #define PARSE_TYPE_RBS 30
64 #define PARSE_TYPE_ELIP 31
65 #define PARSE_TYPE_DOT 32
66 #define PARSE_TYPE_LT 33
67 #define PARSE_TYPE_GT 34
68 #define PARSE_TYPE_BAND 35
69 #define PARSE_TYPE_BOR 36
70 #define PARSE_TYPE_DONE 37
71 #define PARSE_TYPE_IDENT 38
74 * Adds a parse type to the parse tree, this is where all the hard
75 * work actually begins.
77 #define PARSE_TREE_ADD(X) \
79 parsetree->next = mem_a(sizeof(struct parsenode)); \
80 parsetree->next->next = NULL; \
81 parsetree->next->type = (X); \
82 parsetree = parsetree->next; \
86 * This is all the punctuation handled in the parser, these don't
87 * need tokens, they're already tokens.
90 "&&", "||", "<=", ">=", "==", "!=", ";", ",", "!", "*",
91 "/" , "(" , ")" , "-" , "+" , "=" , "[" , "]", "{", "}", "...",
92 "." , "<" , ">" , "&" , "|" ,
100 void parse_debug(struct parsenode *tree) {
102 switch (tree->type) {
103 case PARSE_TYPE_ADD: STORE("OPERATOR: ADD \n");
104 case PARSE_TYPE_BAND: STORE("OPERATOR: BITAND \n");
105 case PARSE_TYPE_BOR: STORE("OPERATOR: BITOR \n");
106 case PARSE_TYPE_COMMA: STORE("OPERATOR: SEPERATOR\n");
107 case PARSE_TYPE_DOT: STORE("OPERATOR: DOT\n");
108 case PARSE_TYPE_DIVIDE: STORE("OPERATOR: DIVIDE\n");
109 case PARSE_TYPE_EQUAL: STORE("OPERATOR: ASSIGNMENT\n");
111 case PARSE_TYPE_BREAK: STORE("STATEMENT: BREAK \n");
112 case PARSE_TYPE_CONTINUE: STORE("STATEMENT: CONTINUE\n");
113 case PARSE_TYPE_GOTO: STORE("STATEMENT: GOTO\n");
114 case PARSE_TYPE_RETURN: STORE("STATEMENT: RETURN\n");
115 case PARSE_TYPE_DONE: STORE("STATEMENT: DONE\n");
117 case PARSE_TYPE_VOID: STORE("DECLTYPE: VOID\n");
118 case PARSE_TYPE_STRING: STORE("DECLTYPE: STRING\n");
119 case PARSE_TYPE_ELIP: STORE("DECLTYPE: VALIST\n");
120 case PARSE_TYPE_ENTITY: STORE("DECLTYPE: ENTITY\n");
121 case PARSE_TYPE_FLOAT: STORE("DECLTYPE: FLOAT\n");
122 case PARSE_TYPE_VECTOR: STORE("DECLTYPE: VECTOR\n");
124 case PARSE_TYPE_GT: STORE("TEST: GREATER THAN\n");
125 case PARSE_TYPE_LT: STORE("TEST: LESS THAN\n");
126 case PARSE_TYPE_GTEQ: STORE("TEST: GREATER THAN OR EQUAL\n");
127 case PARSE_TYPE_LTEQ: STORE("TEST: LESS THAN OR EQUAL\n");
128 case PARSE_TYPE_LNEQ: STORE("TEST: NOT EQUAL\n");
129 case PARSE_TYPE_EQEQ: STORE("TEST: EQUAL-EQUAL\n");
131 case PARSE_TYPE_LBS: STORE("BLOCK: BEG\n");
132 case PARSE_TYPE_RBS: STORE("BLOCK: END\n");
133 case PARSE_TYPE_ELSE: STORE("BLOCK: ELSE\n");
134 case PARSE_TYPE_IF: STORE("BLOCK: IF\n");
136 case PARSE_TYPE_LAND: STORE("LOGICAL: AND\n");
137 case PARSE_TYPE_LNOT: STORE("LOGICAL: NOT\n");
138 case PARSE_TYPE_LOR: STORE("LOGICAL: OR\n");
140 case PARSE_TYPE_LPARTH: STORE("PARTH: BEG\n");
141 case PARSE_TYPE_RPARTH: STORE("PARTH: END\n");
143 case PARSE_TYPE_WHILE: STORE("LOOP: WHILE\n");
144 case PARSE_TYPE_FOR: STORE("LOOP: FOR\n");
145 case PARSE_TYPE_DO: STORE("LOOP: DO\n");
147 //case PARSE_TYPE_IDENT: STORE("IDENT: ???\n");
154 * Performs a parse operation: This is a macro to prevent bugs, if the
155 * calls to lex_token are'nt exactly enough to feed to the end of the
156 * actual lexees for the current thing that is being parsed, the state
157 * of the next iteration in the creation of the parse tree will be wrong
158 * and everything will fail.
160 #define PARSE_PERFORM(X,C) { \
161 token = lex_token(file); \
163 while (token != '\n') { \
164 token = lex_token(file); \
170 void parse_clear(struct parsenode *tree) {
172 struct parsenode *temp = NULL;
173 while (tree != NULL) {
179 /* free any potential typedefs */
184 * Generates a parse tree out of the lexees generated by the lexer. This
185 * is where the tree is built. This is where valid check is performed.
187 int parse_tree(struct lex_file *file) {
188 struct parsenode *parsetree = NULL;
189 struct parsenode *parseroot = NULL;
192 * Allocate memory for our parse tree:
193 * the parse tree is just a singly linked list which will contain
194 * all the data for code generation.
197 parseroot = mem_a(sizeof(struct parsenode));
199 return error(ERROR_INTERNAL, "Ran out of memory", " ");
200 parsetree = parseroot;
201 parsetree->type = -1; /* not a valid type -- root element */
205 while ((token = lex_token(file)) != ERROR_LEX && \
206 token != ERROR_COMPILER && \
207 token != ERROR_INTERNAL && \
208 token != ERROR_PARSE && \
209 token != ERROR_PREPRO && file->length >= 0) {
212 while ((token == ' ' || token == '\n') && file->length >= 0)
213 token = lex_token(file);
214 PARSE_TREE_ADD(PARSE_TYPE_IF);
217 token = lex_token(file);
218 PARSE_TREE_ADD(PARSE_TYPE_ELSE);
221 //token = lex_token(file);
222 while ((token == ' ' || token == '\n') && file->length >= 0)
223 token = lex_token(file);
224 PARSE_TREE_ADD(PARSE_TYPE_FOR);
228 * This is a quick and easy way to do typedefs at parse time
229 * all power is in typedef_add(), in typedef.c. We handle
230 * the tokens accordingly here.
232 case TOKEN_TYPEDEF: {
235 token = lex_token(file);
236 token = lex_token(file); f = util_strdup(file->lastok);
237 token = lex_token(file);
238 token = lex_token(file); t = util_strdup(file->lastok);
245 while (token != '\n')
246 token = lex_token(file);
251 * Returns are addable as-is, statement checking is during
252 * the actual parse tree check.
255 PARSE_TREE_ADD(PARSE_TYPE_RETURN);
258 case TOKEN_DO: PARSE_PERFORM(PARSE_TYPE_DO, {});
259 case TOKEN_WHILE: PARSE_PERFORM(PARSE_TYPE_WHILE, {});
260 case TOKEN_BREAK: PARSE_PERFORM(PARSE_TYPE_BREAK, {});
261 case TOKEN_CONTINUE: PARSE_PERFORM(PARSE_TYPE_CONTINUE,{});
262 case TOKEN_GOTO: PARSE_PERFORM(PARSE_TYPE_GOTO, {});
263 case TOKEN_VOID: PARSE_PERFORM(PARSE_TYPE_VOID, {});
264 case TOKEN_STRING: PARSE_PERFORM(PARSE_TYPE_STRING, {});
265 case TOKEN_FLOAT: PARSE_PERFORM(PARSE_TYPE_FLOAT, {});
266 case TOKEN_VECTOR: PARSE_PERFORM(PARSE_TYPE_VECTOR, {});
267 case TOKEN_ENTITY: PARSE_PERFORM(PARSE_TYPE_ENTITY, {});
270 * From here down is all language punctuation: There is no
271 * need to actual create tokens from these because they're already
272 * tokenized as these individual tokens (which are in a special area
273 * of the ascii table which doesn't conflict with our other tokens
274 * which are higer than the ascii table.)
278 * Skip the preprocessor for now: We'll implement our own
279 * eventually. For now we need to make sure directives are
280 * not accidently tokenized.
282 token = lex_token(file);
283 token = lex_token(file);
285 /* skip all tokens to end of directive */
286 while (token != '\n')
287 token = lex_token(file);
291 token = lex_token(file);
292 PARSE_TREE_ADD(PARSE_TYPE_DOT);
295 token = lex_token(file);
296 PARSE_TREE_ADD(PARSE_TYPE_LPARTH);
299 token = lex_token(file);
300 PARSE_TREE_ADD(PARSE_TYPE_RPARTH);
304 token = lex_token(file);
305 if (token == '&') { /* && */
306 token = lex_token(file);
307 PARSE_TREE_ADD(PARSE_TYPE_LAND);
310 PARSE_TREE_ADD(PARSE_TYPE_BAND);
313 token = lex_token(file);
314 if (token == '|') { /* || */
315 token = lex_token(file);
316 PARSE_TREE_ADD(PARSE_TYPE_LOR);
319 PARSE_TREE_ADD(PARSE_TYPE_BOR);
322 token = lex_token(file);
323 if (token == '=') { /* != */
324 token = lex_token(file);
325 PARSE_TREE_ADD(PARSE_TYPE_LNEQ);
328 PARSE_TREE_ADD(PARSE_TYPE_LNOT);
331 token = lex_token(file);
332 if (token == '=') { /* <= */
333 token = lex_token(file);
334 PARSE_TREE_ADD(PARSE_TYPE_LTEQ);
337 PARSE_TREE_ADD(PARSE_TYPE_LT);
340 token = lex_token(file);
341 if (token == '=') { /* >= */
342 token = lex_token(file);
343 PARSE_TREE_ADD(PARSE_TYPE_GTEQ);
346 PARSE_TREE_ADD(PARSE_TYPE_GT);
349 token = lex_token(file);
350 if (token == '=') { /* == */
351 token = lex_token(file);
352 PARSE_TREE_ADD(PARSE_TYPE_EQEQ);
355 PARSE_TREE_ADD(PARSE_TYPE_EQUAL);
358 token = lex_token(file);
359 PARSE_TREE_ADD(PARSE_TYPE_DONE);
362 token = lex_token(file);
363 PARSE_TREE_ADD(PARSE_TYPE_MINUS);
366 token = lex_token(file);
367 PARSE_TREE_ADD(PARSE_TYPE_ADD);
370 token = lex_token(file);
371 PARSE_TREE_ADD(PARSE_TYPE_LBS);
374 token = lex_token(file);
375 PARSE_TREE_ADD(PARSE_TYPE_RBS);
379 * TODO: Fix lexer to spit out ( ) as tokens, it seems the
380 * using '(' or ')' in parser doesn't work properly unless
381 * there are spaces before them to allow the lexer to properly
382 * seperate identifiers. -- otherwise it eats all of it.
385 token = lex_token(file);
386 PARSE_TREE_ADD(PARSE_TYPE_IDENT);
390 parse_debug(parseroot);
392 parse_clear(parseroot);