1 | // Coded by Ralf Westram (coder@reallysoft.de) in March 2011 // |
---|
2 | // Institute of Microbiology (Technical University Munich) // |
---|
3 | // http://www.arb-home.de/ // |
---|
4 | |
---|
5 | #include "aisc_parser.h" |
---|
6 | |
---|
7 | const char *Parser::currentLocation(const char *io) { |
---|
8 | const int LOCBUFFERSIZE = 1024; |
---|
9 | static char loc_buf[LOCBUFFERSIZE+1]; |
---|
10 | int printed; |
---|
11 | |
---|
12 | if (last_line_start) { |
---|
13 | int column = io - last_line_start + 1; |
---|
14 | |
---|
15 | aisc_assert(column >= 0 && column<10000); |
---|
16 | printed = sprintf(loc_buf, "%s:%i:%i", loc.get_path(), loc.get_linenr(), column); |
---|
17 | } |
---|
18 | else { |
---|
19 | printed = sprintf(loc_buf, "%s:%i", loc.get_path(), loc.get_linenr()); |
---|
20 | } |
---|
21 | |
---|
22 | if (printed>LOCBUFFERSIZE) { |
---|
23 | fprintf(stderr, "AISC: Internal buffer overflow detected -- terminating [loc_buf]\n"); |
---|
24 | error_flag = 1; |
---|
25 | } |
---|
26 | |
---|
27 | return loc_buf; |
---|
28 | } |
---|
29 | |
---|
30 | void Parser::p_err(const char *io, const char *error) { |
---|
31 | fprintf(stderr, "%s: Error: %s\n", currentLocation(io), error); |
---|
32 | error_flag = 1; |
---|
33 | } |
---|
34 | |
---|
35 | #ifdef LINUX |
---|
36 | # define HAVE_VSNPRINTF |
---|
37 | #endif |
---|
38 | #ifdef HAVE_VSNPRINTF |
---|
39 | # define PRINT2BUFFER(buffer, bufsize, templat, parg) vsnprintf(buffer, bufsize, templat, parg); |
---|
40 | #else |
---|
41 | # define PRINT2BUFFER(buffer, bufsize, templat, parg) vsprintf(buffer, templat, parg); |
---|
42 | #endif |
---|
43 | |
---|
44 | #define ERRBUFFERSIZE 1024 |
---|
45 | void Parser::p_errf(const char *io, const char *formatString, ...) { |
---|
46 | static char buf[ERRBUFFERSIZE+1]; |
---|
47 | int printed; |
---|
48 | va_list varArgs; |
---|
49 | |
---|
50 | va_start(varArgs, formatString); |
---|
51 | printed = PRINT2BUFFER(buf, ERRBUFFERSIZE, formatString, varArgs); |
---|
52 | va_end(varArgs); |
---|
53 | |
---|
54 | if (printed>ERRBUFFERSIZE) { |
---|
55 | fprintf(stderr, "AISC: Internal buffer overflow detected -- terminating [err]\n"); |
---|
56 | error_flag = 1; |
---|
57 | } |
---|
58 | p_err(io, buf); |
---|
59 | } |
---|
60 | |
---|
61 | void Parser::copyTillQuotesTo(const char*& in, char*& out) { |
---|
62 | // closing quotes are END_STR1 plus END_STR2 |
---|
63 | bool quotes_closed = false; |
---|
64 | while (!quotes_closed) { |
---|
65 | while ((lastchar != EOSTR) && (lastchar != END_STR1)) { |
---|
66 | *(out++) = lastchar; |
---|
67 | get_byte(in); |
---|
68 | } |
---|
69 | |
---|
70 | if (lastchar == END_STR1) { |
---|
71 | get_byte(in); |
---|
72 | if (lastchar == END_STR2) { |
---|
73 | get_byte(in); |
---|
74 | quotes_closed = true; |
---|
75 | } |
---|
76 | else { |
---|
77 | *(out++) = END_STR1; |
---|
78 | } |
---|
79 | } |
---|
80 | } |
---|
81 | } |
---|
82 | |
---|
83 | char *Parser::readWord(const char *& in) { |
---|
84 | char buf[1024]; |
---|
85 | char *cp = buf; |
---|
86 | |
---|
87 | if (lastchar == BEG_STR1) { |
---|
88 | get_byte(in); |
---|
89 | if (lastchar == BEG_STR2) { |
---|
90 | get_byte(in); |
---|
91 | copyTillQuotesTo(in, cp); |
---|
92 | } |
---|
93 | else { |
---|
94 | *(cp++) = BEG_STR1; |
---|
95 | copyWordTo(in, cp); |
---|
96 | } |
---|
97 | } |
---|
98 | else copyWordTo(in, cp); |
---|
99 | |
---|
100 | if (lastchar != EOSTR) skip_over_spaces_and_comments(in); |
---|
101 | |
---|
102 | char *result = NULp; |
---|
103 | if (lastchar == EOSTR) { |
---|
104 | p_err_exp_but_saw_EOF(in, "terminator after string"); |
---|
105 | } |
---|
106 | else if (cp != buf) { // do not return empty string, return NULp instead |
---|
107 | *cp = 0; |
---|
108 | result = strdup(buf); |
---|
109 | } |
---|
110 | |
---|
111 | return result; |
---|
112 | } |
---|
113 | |
---|
114 | char *Parser::SETSOURCE(const char *& in, enum TOKEN& foundTokenType) { |
---|
115 | char *result = NULp; |
---|
116 | const char *space = in+9; |
---|
117 | |
---|
118 | if (*space != ' ') p_err(in, "space expected after '@SETSOURCE' (injected code)"); |
---|
119 | else { |
---|
120 | in = space; |
---|
121 | get_byte(in); |
---|
122 | skip_over_spaces(in); |
---|
123 | |
---|
124 | const char *file = in-1; |
---|
125 | const char *comma = strchr(file, ','); |
---|
126 | |
---|
127 | if (!comma) p_err(in, "comma expected after '@SETSOURCE filename' (injected code)"); |
---|
128 | else { |
---|
129 | const char *end = strchr(comma, '@'); |
---|
130 | |
---|
131 | if (!end) p_err(in, "'@' expected after '@SETSOURCE filename,line' (injected code)"); |
---|
132 | else { |
---|
133 | char *filename = copy_string_part(file, comma-1); |
---|
134 | set_source(filename, atoi(comma+1)); |
---|
135 | free(filename); |
---|
136 | |
---|
137 | in = end+1; |
---|
138 | get_byte(in); |
---|
139 | |
---|
140 | result = parse_token(in, foundTokenType); |
---|
141 | } |
---|
142 | } |
---|
143 | } |
---|
144 | return result; |
---|
145 | } |
---|
146 | |
---|
147 | char *Parser::parse_token(const char *& in, enum TOKEN& foundTokenType) { |
---|
148 | skip_over_spaces_and_comments_multiple_lines(in); |
---|
149 | |
---|
150 | char *result = NULp; |
---|
151 | foundTokenType = TOK_INVALID; |
---|
152 | |
---|
153 | switch (lastchar) { |
---|
154 | case EOSTR: foundTokenType = TOK_EOS; break; |
---|
155 | case '{': foundTokenType = TOK_BRACE_OPEN; break; |
---|
156 | case '}': foundTokenType = TOK_BRACE_CLOSE; break; |
---|
157 | case ',': foundTokenType = TOK_COMMA; break; |
---|
158 | case ';': foundTokenType = TOK_SEMI; break; |
---|
159 | |
---|
160 | case '@': |
---|
161 | if (strncmp(in, "SETSOURCE", 9) == 0) { |
---|
162 | result = SETSOURCE(in, foundTokenType); |
---|
163 | } |
---|
164 | else { |
---|
165 | get_byte(in); |
---|
166 | skip_over_spaces(in); |
---|
167 | if (lastchar == EOSTR) { |
---|
168 | p_err_exp_but_saw_EOF(in, "ID behind '@'"); |
---|
169 | } |
---|
170 | else { |
---|
171 | result = readWord(in); |
---|
172 | foundTokenType = TOK_AT_WORD; |
---|
173 | } |
---|
174 | } |
---|
175 | break; |
---|
176 | |
---|
177 | default: |
---|
178 | result = readWord(in); |
---|
179 | foundTokenType = TOK_WORD; |
---|
180 | break; |
---|
181 | } |
---|
182 | |
---|
183 | return result; |
---|
184 | } |
---|
185 | |
---|
186 | class Header : virtual Noncopyable { |
---|
187 | char *key; |
---|
188 | Header *next; |
---|
189 | public: |
---|
190 | Header(const char *key_) { key = strdup(key_); next = NULp; } |
---|
191 | ~Header() { free(key); delete next; } |
---|
192 | |
---|
193 | void set_next(Header *header) { aisc_assert(!next); next = header; } |
---|
194 | |
---|
195 | const char *get_key() const { return key; } |
---|
196 | const Header *next_header() const { return next; } |
---|
197 | }; |
---|
198 | |
---|
199 | class HeaderList : virtual Noncopyable { |
---|
200 | Header *head; |
---|
201 | Header *tail; |
---|
202 | |
---|
203 | char *loc_defined_at; |
---|
204 | |
---|
205 | public: |
---|
206 | HeaderList() { |
---|
207 | head = tail = NULp; |
---|
208 | loc_defined_at = NULp; |
---|
209 | } |
---|
210 | void reset() { |
---|
211 | delete head; |
---|
212 | head = tail = NULp; |
---|
213 | free(loc_defined_at); |
---|
214 | loc_defined_at = NULp; |
---|
215 | } |
---|
216 | ~HeaderList() { reset(); } |
---|
217 | |
---|
218 | void set_first_header(Header *header, const char *location) { |
---|
219 | aisc_assert(!head); |
---|
220 | head = tail = header; |
---|
221 | loc_defined_at = strdup(location); |
---|
222 | } |
---|
223 | void append(Header *header) { |
---|
224 | aisc_assert(head); // use set_first_header() |
---|
225 | tail->set_next(header); |
---|
226 | tail = header; |
---|
227 | } |
---|
228 | |
---|
229 | const Header *first_header() const { return head; } |
---|
230 | const char *defined_at() const { return loc_defined_at; } |
---|
231 | }; |
---|
232 | |
---|
233 | Token *Parser::parseBrace(const char *& in, const char *key) { |
---|
234 | Token *res = NULp; |
---|
235 | if (!error_flag) { |
---|
236 | char *openLoc = strdup(currentLocation(in)); |
---|
237 | TokenListBlock *block = parseTokenListBlock(in); |
---|
238 | |
---|
239 | if (block) { |
---|
240 | res = new Token(key, block); |
---|
241 | block = NULp; |
---|
242 | expect_and_skip_closing_brace(in, openLoc); |
---|
243 | } |
---|
244 | else { |
---|
245 | p_err_empty_braces(in); |
---|
246 | } |
---|
247 | delete block; |
---|
248 | free(openLoc); |
---|
249 | } |
---|
250 | return res; |
---|
251 | } |
---|
252 | |
---|
253 | TokenList *Parser::parseTokenList(const char *& in, HeaderList& headerList) { |
---|
254 | TokenList *items = new TokenList; |
---|
255 | const Header *header = headerList.first_header(); |
---|
256 | |
---|
257 | get_byte(in); |
---|
258 | |
---|
259 | bool reached_end_of_list = false; |
---|
260 | while (!error_flag && !reached_end_of_list) { |
---|
261 | TOKEN foundTokenType; |
---|
262 | char *str = parse_token(in, foundTokenType); |
---|
263 | |
---|
264 | if (!error_flag) { |
---|
265 | switch (foundTokenType) { |
---|
266 | case TOK_SEMI: |
---|
267 | case TOK_BRACE_CLOSE: |
---|
268 | reached_end_of_list = true; |
---|
269 | break; |
---|
270 | |
---|
271 | case TOK_BRACE_OPEN: { |
---|
272 | Token *sub = parseBrace(in, header ? header->get_key() : "{"); |
---|
273 | if (sub) items->append(sub); |
---|
274 | break; |
---|
275 | } |
---|
276 | case TOK_COMMA: |
---|
277 | if (!header) p_err_exp_but_saw(in, "string", "','"); |
---|
278 | else get_byte(in); |
---|
279 | break; |
---|
280 | |
---|
281 | case TOK_AT_WORD: |
---|
282 | if (header != headerList.first_header()) { |
---|
283 | p_err_ill_atWord(in); |
---|
284 | } |
---|
285 | else { |
---|
286 | if (!str) { |
---|
287 | p_err_expected(in, "ID behind '@'"); |
---|
288 | } |
---|
289 | else { |
---|
290 | headerList.reset(); |
---|
291 | headerList.set_first_header(new Header(str), currentLocation(in)); |
---|
292 | header = headerList.first_header(); |
---|
293 | } |
---|
294 | while (lastchar == ',' && !error_flag) { |
---|
295 | get_byte(in); |
---|
296 | char *str2 = parse_token(in, foundTokenType); |
---|
297 | if (foundTokenType != TOK_AT_WORD) p_err_exp_atWord(in); |
---|
298 | else headerList.append(new Header(str2)); |
---|
299 | free(str2); |
---|
300 | } |
---|
301 | if (!error_flag) expect_and_skip(in, ';'); |
---|
302 | if (!error_flag) { |
---|
303 | aisc_assert(headerList.first_header()->get_key()); |
---|
304 | reached_end_of_list = true; |
---|
305 | } |
---|
306 | } |
---|
307 | break; |
---|
308 | |
---|
309 | case TOK_WORD: { |
---|
310 | Token *new_token = NULp; |
---|
311 | if (header) { |
---|
312 | new_token = new Token(header->get_key(), str); |
---|
313 | expect_line_terminator(in); |
---|
314 | } |
---|
315 | else { |
---|
316 | char *str2 = parse_token(in, foundTokenType); |
---|
317 | switch (foundTokenType) { |
---|
318 | case TOK_BRACE_OPEN: { |
---|
319 | new_token = parseBrace(in, str); |
---|
320 | break; |
---|
321 | } |
---|
322 | case TOK_WORD: |
---|
323 | new_token = new Token(str, str2); |
---|
324 | expect_line_terminator(in); |
---|
325 | break; |
---|
326 | |
---|
327 | case TOK_COMMA: |
---|
328 | case TOK_SEMI: |
---|
329 | new_token = new Token(str, ""); |
---|
330 | break; |
---|
331 | |
---|
332 | case TOK_AT_WORD: p_err_exp_string_but_saw(in, "'@'"); break; |
---|
333 | case TOK_BRACE_CLOSE: p_err_exp_string_but_saw(in, "'}'"); break; |
---|
334 | |
---|
335 | case TOK_INVALID: aisc_assert(0); break; |
---|
336 | } |
---|
337 | free(str2); |
---|
338 | } |
---|
339 | |
---|
340 | aisc_assert(new_token || error_flag); |
---|
341 | |
---|
342 | if (new_token) items->append(new_token); |
---|
343 | |
---|
344 | if (!error_flag) { |
---|
345 | if (lastchar == ';') { |
---|
346 | const Header *missingVal = header ? header->next_header() : NULp; |
---|
347 | if (missingVal) { |
---|
348 | char buf[1000]; |
---|
349 | sprintf(buf, "value for @%s", missingVal->get_key()); |
---|
350 | p_err_exp_but_saw(in, buf, "';'"); |
---|
351 | } |
---|
352 | else reached_end_of_list = true; |
---|
353 | get_byte(in); |
---|
354 | } |
---|
355 | else get_byte(in); |
---|
356 | } |
---|
357 | break; |
---|
358 | } |
---|
359 | |
---|
360 | case TOK_INVALID: |
---|
361 | p_err(in, "Invalid token (internal error)"); |
---|
362 | break; |
---|
363 | } |
---|
364 | } |
---|
365 | |
---|
366 | if (!error_flag && header) header = header->next_header(); |
---|
367 | |
---|
368 | free(str); |
---|
369 | } |
---|
370 | |
---|
371 | if (error_flag || items->empty()) { |
---|
372 | delete items; |
---|
373 | items = NULp; |
---|
374 | } |
---|
375 | |
---|
376 | return items; |
---|
377 | } |
---|
378 | |
---|
379 | |
---|
380 | TokenListBlock *Parser::parseTokenListBlock(const char *& in) { |
---|
381 | TokenListBlock *block = new TokenListBlock; |
---|
382 | HeaderList headerList; |
---|
383 | |
---|
384 | while ((EOSTR != lastchar) && (lastchar != '}')) { |
---|
385 | TokenList *list = parseTokenList(in, headerList); |
---|
386 | if (!error_flag && list) { |
---|
387 | block->append(list); |
---|
388 | } |
---|
389 | } |
---|
390 | |
---|
391 | if (block->empty() || error_flag) { |
---|
392 | delete block; |
---|
393 | block = NULp; |
---|
394 | } |
---|
395 | |
---|
396 | return block; |
---|
397 | } |
---|
398 | |
---|
399 | Code *Parser::parse_program(const char *in, const char *filename) { |
---|
400 | Code *first_cl = NULp; |
---|
401 | Code *cl = NULp; |
---|
402 | |
---|
403 | set_source(filename, 0); |
---|
404 | |
---|
405 | while (lastchar != EOSTR) { |
---|
406 | skip_over_spaces_and_comments_multiple_lines(in); |
---|
407 | |
---|
408 | if (lastchar == EOSTR) break; |
---|
409 | |
---|
410 | const char *p = in-1; |
---|
411 | while ((lastchar != EOSTR) && (lastchar != '\n')) get_byte(in); |
---|
412 | |
---|
413 | { |
---|
414 | Code *hcl = new Code; |
---|
415 | Code *& next = cl ? cl->next : first_cl; |
---|
416 | cl = next = hcl; |
---|
417 | } |
---|
418 | cl->str = copy_string_part(p, in-2); |
---|
419 | cl->source = Location(loc.get_linenr(), filename); |
---|
420 | } |
---|
421 | |
---|
422 | return first_cl; |
---|
423 | } |
---|
424 | |
---|