|
1 /* |
|
2 * libtu/parser.c |
|
3 * |
|
4 * Copyright (c) Tuomo Valkonen 1999-2000. |
|
5 * |
|
6 * This file is distributed under the terms of the "Artistic License". |
|
7 * See the included file LICENSE for details. |
|
8 */ |
|
9 |
|
10 #include <string.h> |
|
11 |
|
12 #include "include/parser.h" |
|
13 #include "include/misc.h" |
|
14 #include "include/output.h" |
|
15 |
|
16 #define MAX_TOKENS 32 |
|
17 #define MAX_NEST 16 |
|
18 |
|
19 |
|
20 enum{ |
|
21 P_EOF, |
|
22 P_OK, |
|
23 P_ERROR, |
|
24 P_BEG_SECT, |
|
25 P_END_SECT |
|
26 }; |
|
27 |
|
28 |
|
29 static bool check_args(const Tokenizer *tokz, Token *tokens, int ntokens, |
|
30 const char *fmt); |
|
31 |
|
32 |
|
33 /* */ |
|
34 |
|
35 |
|
36 static int read_statement(const ConfOpt **opt_ret, Token *tokens, |
|
37 int *ntok_ret, Tokenizer *tokz, |
|
38 const ConfOpt *options) |
|
39 { |
|
40 int ntokens=0; |
|
41 Token *tok=NULL; |
|
42 const ConfOpt *opt=NULL; |
|
43 int had_comma=0; |
|
44 int retval=P_OK; |
|
45 int e=0; |
|
46 int e_line=0; |
|
47 |
|
48 while(1){ |
|
49 if(ntokens==MAX_TOKENS-1){ |
|
50 e=E_TOKZ_TOKEN_LIMIT; |
|
51 goto ret_err; |
|
52 } |
|
53 |
|
54 tok=&tokens[ntokens]; |
|
55 |
|
56 if(!tokz_get_token(tokz, tok)) |
|
57 goto ret_err; |
|
58 |
|
59 ntokens++; |
|
60 |
|
61 if(!TOK_IS_OP(tok)){ |
|
62 if(ntokens==1 && !had_comma){ |
|
63 if(!TOK_IS_IDENT(tok)){ |
|
64 e=E_TOKZ_IDENTIFIER_EXPECTED; |
|
65 goto ret_err; |
|
66 } |
|
67 |
|
68 /* find the option */ |
|
69 for(opt=options; opt->optname; opt++){ |
|
70 if(strcmp(opt->optname, TOK_IDENT_VAL(tok))==0) |
|
71 break; |
|
72 } |
|
73 |
|
74 if(!opt->optname){ |
|
75 e=E_TOKZ_UNKNOWN_OPTION; |
|
76 e_line=tok->line; |
|
77 retval=P_ERROR; |
|
78 } |
|
79 |
|
80 had_comma=2; |
|
81 }else{ |
|
82 if(!had_comma){ |
|
83 e=E_TOKZ_SYNTAX; |
|
84 goto ret_err; |
|
85 } |
|
86 |
|
87 had_comma=0; |
|
88 } |
|
89 continue; |
|
90 } |
|
91 |
|
92 /* It is an operator */ |
|
93 |
|
94 switch(TOK_OP_VAL(tok)){ |
|
95 case OP_SCOLON: |
|
96 if(opt){ |
|
97 if(had_comma || opt->opts){ |
|
98 e=E_TOKZ_SYNTAX; |
|
99 goto ret_err; |
|
100 } |
|
101 goto ret_success; |
|
102 } |
|
103 break; |
|
104 |
|
105 case OP_NEXTLINE: |
|
106 if(had_comma==1){ |
|
107 e=E_TOKZ_SYNTAX; |
|
108 e_line=tok->line-1; |
|
109 goto ret_err2; |
|
110 } |
|
111 |
|
112 if(opt && !opt->opts) |
|
113 goto ret_success; |
|
114 break; |
|
115 |
|
116 case OP_EOF: |
|
117 if(had_comma==1){ |
|
118 e=E_TOKZ_UNEXPECTED_EOF; |
|
119 goto ret_err; |
|
120 } |
|
121 |
|
122 if(opt && opt->opts){ |
|
123 e=E_TOKZ_UNEXPECTED_EOF; |
|
124 goto ret_err; |
|
125 } |
|
126 |
|
127 retval=P_EOF; |
|
128 goto ret_success; |
|
129 |
|
130 case OP_R_BRC: |
|
131 if(had_comma==1){ |
|
132 e=E_TOKZ_SYNTAX; |
|
133 goto ret_err; |
|
134 } |
|
135 |
|
136 if(opt && opt->opts){ |
|
137 e=E_TOKZ_SYNTAX; |
|
138 goto ret_err; |
|
139 } |
|
140 |
|
141 retval=P_END_SECT; |
|
142 goto ret_success; |
|
143 |
|
144 case OP_L_BRC: |
|
145 if(had_comma==1 || !opt || !opt->opts){ |
|
146 e=E_TOKZ_SYNTAX; |
|
147 goto ret_err; |
|
148 } |
|
149 |
|
150 retval=P_BEG_SECT; |
|
151 goto ret_success; |
|
152 |
|
153 case OP_COMMA: |
|
154 if(had_comma){ |
|
155 e=E_TOKZ_SYNTAX; |
|
156 goto ret_err; |
|
157 } |
|
158 had_comma=1; |
|
159 break; |
|
160 |
|
161 default: |
|
162 e=E_TOKZ_SYNTAX; |
|
163 goto ret_err; |
|
164 } |
|
165 |
|
166 ntokens--; |
|
167 } |
|
168 |
|
169 ret_err: |
|
170 e_line=tok->line; |
|
171 ret_err2: |
|
172 retval=P_ERROR; |
|
173 |
|
174 ret_success: |
|
175 if(retval==P_ERROR && e!=0) |
|
176 tokz_warn_error(tokz, e_line, e); |
|
177 |
|
178 *opt_ret=opt; |
|
179 *ntok_ret=ntokens; |
|
180 |
|
181 return retval; |
|
182 } |
|
183 |
|
184 |
|
185 /* */ |
|
186 |
|
187 |
|
188 static bool call_end_sect(Tokenizer *tokz, const ConfOpt *options) |
|
189 { |
|
190 bool retval=TRUE; |
|
191 |
|
192 while(options->optname){ |
|
193 if(strcmp(options->optname, "#end")==0){ |
|
194 retval=options->fn(tokz, 0, NULL); |
|
195 break; |
|
196 } |
|
197 options++; |
|
198 } |
|
199 |
|
200 return retval; |
|
201 } |
|
202 |
|
203 |
|
204 static void call_cancel_sect(Tokenizer *tokz, const ConfOpt *options) |
|
205 { |
|
206 while(options->optname){ |
|
207 if(strcmp(options->optname, "#cancel")==0){ |
|
208 options->fn(tokz, 0, NULL); |
|
209 break; |
|
210 } |
|
211 options++; |
|
212 } |
|
213 } |
|
214 |
|
215 |
|
216 /* */ |
|
217 |
|
218 |
|
219 /* Does the parsing work |
|
220 */ |
|
221 bool parse_config_tokz(Tokenizer *tokz, const ConfOpt *options) |
|
222 { |
|
223 Token tokens[MAX_TOKENS]; |
|
224 bool alloced_optstack=FALSE; |
|
225 int i, t, ntokens; |
|
226 int init_nest_lvl; |
|
227 bool had_error=FALSE; |
|
228 |
|
229 /* Allocate tokz->optstack if it does not yet exist (if it does, |
|
230 * we have been called from an option handler) |
|
231 */ |
|
232 if(!tokz->optstack){ |
|
233 tokz->optstack=ALLOC_N(const ConfOpt*, MAX_NEST); |
|
234 if(!tokz->optstack){ |
|
235 warn_err(); |
|
236 return FALSE; |
|
237 } |
|
238 |
|
239 memset(tokz->optstack, 0, sizeof(ConfOpt*)*MAX_NEST); |
|
240 init_nest_lvl=tokz->nest_lvl=0; |
|
241 alloced_optstack=TRUE; |
|
242 }else{ |
|
243 init_nest_lvl=tokz->nest_lvl; |
|
244 } |
|
245 |
|
246 tokz->optstack[init_nest_lvl]=options; |
|
247 |
|
248 for(i=0;i<MAX_TOKENS;i++) |
|
249 tok_init(&tokens[i]); |
|
250 |
|
251 |
|
252 /* The loop |
|
253 */ |
|
254 while(1){ |
|
255 t=read_statement(&options, tokens, &ntokens, |
|
256 tokz, tokz->optstack[tokz->nest_lvl]); |
|
257 |
|
258 had_error=(t==P_ERROR); |
|
259 |
|
260 /* Check that arguments are ok */ |
|
261 if(!had_error && options) |
|
262 had_error=!check_args(tokz, tokens, ntokens, options->argfmt); |
|
263 |
|
264 if(tokz->flags&TOKZ_PARSER_INDENT_MODE) |
|
265 verbose_indent(tokz->nest_lvl); |
|
266 |
|
267 /* New section? */ |
|
268 if(t==P_BEG_SECT){ |
|
269 if(tokz->nest_lvl==MAX_NEST-1){ |
|
270 tokz_warn_error(tokz, tokz->line, E_TOKZ_MAX_NEST); |
|
271 had_error=TRUE; |
|
272 while(ntokens--) |
|
273 tok_free(&tokens[ntokens]); |
|
274 break; |
|
275 }else{ |
|
276 tokz->optstack[++tokz->nest_lvl]=options->opts; |
|
277 } |
|
278 } |
|
279 |
|
280 /* call the handler */ |
|
281 if(!had_error && options && options->fn) |
|
282 had_error=!options->fn(tokz, ntokens-1, tokens); |
|
283 |
|
284 /* free the tokens */ |
|
285 while(ntokens--) |
|
286 tok_free(&tokens[ntokens]); |
|
287 |
|
288 switch(t){ |
|
289 case P_BEG_SECT: |
|
290 if(!had_error) |
|
291 continue; |
|
292 /* #cancel handler should not be called when |
|
293 * error occured in section begin handler */ |
|
294 tokz->nest_lvl--; |
|
295 break; |
|
296 |
|
297 case P_EOF: |
|
298 if(tokz->nest_lvl>0){ |
|
299 tokz_warn_error(tokz, 0, E_TOKZ_UNEXPECTED_EOF); |
|
300 had_error=TRUE; |
|
301 }else if(!had_error){ |
|
302 had_error=!call_end_sect(tokz, tokz->optstack[0]); |
|
303 } |
|
304 break; |
|
305 |
|
306 case P_END_SECT: |
|
307 if(tokz->nest_lvl==0){ |
|
308 tokz_warn_error(tokz, tokz->line, E_TOKZ_SYNTAX); |
|
309 had_error=TRUE; |
|
310 break; |
|
311 } |
|
312 |
|
313 if(!had_error) |
|
314 had_error=!call_end_sect(tokz, tokz->optstack[tokz->nest_lvl]); |
|
315 |
|
316 tokz->nest_lvl--; |
|
317 |
|
318 if(tokz->nest_lvl<init_nest_lvl) |
|
319 break; |
|
320 |
|
321 /* fall thru */ |
|
322 |
|
323 default: |
|
324 if(!had_error) |
|
325 continue; |
|
326 } |
|
327 break; |
|
328 } |
|
329 |
|
330 /* On error, call all the #cancel-handlers */ |
|
331 while(had_error && tokz->nest_lvl>=init_nest_lvl){ |
|
332 call_cancel_sect(tokz, tokz->optstack[tokz->nest_lvl]); |
|
333 tokz->nest_lvl--; |
|
334 } |
|
335 |
|
336 /* Free optstack if it was alloced by this call */ |
|
337 if(alloced_optstack){ |
|
338 free(tokz->optstack); |
|
339 tokz->optstack=NULL; |
|
340 tokz->nest_lvl=0; |
|
341 } |
|
342 |
|
343 if(tokz->flags&TOKZ_PARSER_INDENT_MODE) |
|
344 verbose_indent(init_nest_lvl); |
|
345 |
|
346 return !had_error; |
|
347 } |
|
348 |
|
349 |
|
350 /* */ |
|
351 |
|
352 |
|
353 bool parse_config(const char *fname, const ConfOpt *options) |
|
354 { |
|
355 Tokenizer *tokz; |
|
356 bool ret; |
|
357 |
|
358 tokz=tokz_open(fname); |
|
359 |
|
360 if(tokz==NULL) |
|
361 return FALSE; |
|
362 |
|
363 ret=parse_config_tokz(tokz, options); |
|
364 |
|
365 tokz_close(tokz); |
|
366 |
|
367 return ret; |
|
368 } |
|
369 |
|
370 |
|
371 bool parse_config_file(FILE *file, const ConfOpt *options) |
|
372 { |
|
373 Tokenizer *tokz; |
|
374 bool ret; |
|
375 |
|
376 tokz=tokz_open_file(file); |
|
377 |
|
378 if(tokz==NULL) |
|
379 return FALSE; |
|
380 |
|
381 ret=parse_config_tokz(tokz, options); |
|
382 |
|
383 tokz_close(tokz); |
|
384 |
|
385 return ret; |
|
386 } |
|
387 |
|
388 |
|
389 /* |
|
390 * Argument validity checking stuff |
|
391 */ |
|
392 |
|
393 |
|
394 static bool arg_match(Token *tok, char c) |
|
395 { |
|
396 static const char chs[]={0, 'l', 'd', 'c', 's', 'i', 0, 0}; |
|
397 char c2; |
|
398 |
|
399 if(c=='.' || c=='*') |
|
400 return TRUE; |
|
401 |
|
402 c2=chs[tok->type]; |
|
403 |
|
404 if(c2==c) |
|
405 return TRUE; |
|
406 |
|
407 if(c2=='c' && c=='l'){ |
|
408 TOK_SET_LONG(tok, TOK_CHAR_VAL(tok)); |
|
409 return TRUE; |
|
410 } |
|
411 |
|
412 if(c2=='l' && c=='c'){ |
|
413 TOK_SET_CHAR(tok, TOK_LONG_VAL(tok)); |
|
414 return TRUE; |
|
415 } |
|
416 |
|
417 if(c2=='l' && c=='d'){ |
|
418 TOK_SET_DOUBLE(tok, TOK_LONG_VAL(tok)); |
|
419 return TRUE; |
|
420 } |
|
421 |
|
422 return FALSE; |
|
423 } |
|
424 |
|
425 |
|
426 static bool check_argument(const char **pret, Token *tok, const char *p) |
|
427 { |
|
428 int mode=0; |
|
429 |
|
430 if(*p=='*'){ |
|
431 *pret=p; |
|
432 return TRUE; |
|
433 }else if(*p=='?'){ |
|
434 mode=1; |
|
435 p++; |
|
436 }else if(*p==':'){ |
|
437 mode=2; |
|
438 p++; |
|
439 }else if(*p=='+'){ |
|
440 *pret=p; |
|
441 return arg_match(tok, *(p-1)); |
|
442 } |
|
443 |
|
444 while(*p!='\0'){ |
|
445 if(arg_match(tok, *p)){ |
|
446 p++; |
|
447 while(mode==2 && *p==':'){ |
|
448 if(*++p=='\0') |
|
449 break; /* invalid argument format string it is... */ |
|
450 p++; |
|
451 } |
|
452 *pret=p; |
|
453 return TRUE; |
|
454 } |
|
455 |
|
456 if(mode==0) |
|
457 break; |
|
458 |
|
459 p++; |
|
460 |
|
461 if(mode==1){ |
|
462 *pret=p; |
|
463 return TRUE; |
|
464 } |
|
465 |
|
466 if(*p!=':') |
|
467 break; |
|
468 p++; |
|
469 } |
|
470 |
|
471 *pret=p; |
|
472 return FALSE; |
|
473 } |
|
474 |
|
475 |
|
476 static bool args_at_end(const char *p) |
|
477 { |
|
478 if(p==NULL) |
|
479 return TRUE; |
|
480 |
|
481 while(*p!='\0'){ |
|
482 if(*p=='*' || *p=='+') |
|
483 p++; |
|
484 else if(*p=='?') |
|
485 p+=2; |
|
486 else |
|
487 return FALSE; |
|
488 } |
|
489 |
|
490 return TRUE; |
|
491 } |
|
492 |
|
493 |
|
494 static bool check_args(const Tokenizer *tokz, Token *tokens, int ntokens, |
|
495 const char *fmt) |
|
496 { |
|
497 int i; |
|
498 |
|
499 if(fmt==NULL) |
|
500 return ntokens==2; |
|
501 |
|
502 for(i=1; i<ntokens-1; i++){ |
|
503 if(!check_argument(&fmt, &tokens[i], fmt)){ |
|
504 tokz_warn_error(tokz, tokens[i].line, |
|
505 *fmt!='\0' ? E_TOKZ_INVALID_ARGUMENT |
|
506 : E_TOKZ_TOO_MANY_ARGS); |
|
507 return FALSE; |
|
508 } |
|
509 } |
|
510 |
|
511 if(!args_at_end(fmt)){ |
|
512 tokz_warn_error(tokz, tokens[i].line, E_TOKZ_TOO_FEW_ARGS); |
|
513 return FALSE; |
|
514 } |
|
515 |
|
516 return TRUE; |
|
517 } |
|
518 |