Commit | Line | Data |
---|---|---|
c0bc4ef7 DF |
1 | /* |
2 | * Copyright (c) 1980 Regents of the University of California. | |
3 | * All rights reserved. The Berkeley software License Agreement | |
4 | * specifies the terms and conditions for redistribution. | |
5 | */ | |
6 | ||
7 | #ifndef lint | |
8 | static char sccsid[] = "@(#)lexi.c 5.1 (Berkeley) %G%"; | |
9 | #endif not lint | |
4b365fcd KM |
10 | |
11 | /* | |
12 | ||
13 | Copyright (C) 1976 | |
14 | by the | |
15 | Board of Trustees | |
16 | of the | |
17 | University of Illinois | |
18 | ||
19 | All rights reserved | |
20 | ||
21 | ||
22 | NAME: | |
23 | lexi | |
24 | ||
25 | FUNCTION: | |
26 | This is the token scanner for indent | |
27 | ||
28 | ALGORITHM: | |
29 | 1) Strip off intervening blanks and/or tabs. | |
30 | 2) If it is an alphanumeric token, move it to the token buffer "token". | |
31 | Check if it is a special reserved word that indent will want to | |
32 | know about. | |
33 | 3) Non-alphanumeric tokens are handled with a big switch statement. A | |
34 | flag is kept to remember if the last token was a "unary delimiter", | |
35 | which forces a following operator to be unary as opposed to binary. | |
36 | ||
37 | PARAMETERS: | |
38 | None | |
39 | ||
40 | RETURNS: | |
41 | An integer code indicating the type of token scanned. | |
42 | ||
43 | GLOBALS: | |
44 | buf_ptr = | |
45 | had_eof | |
46 | last_u_d = Set to true iff this token is a "unary delimiter" | |
47 | ||
48 | CALLS: | |
49 | fill_buffer | |
50 | printf (lib) | |
51 | ||
52 | CALLED BY: | |
53 | main | |
54 | ||
55 | NOTES: | |
56 | Start of comment is passed back so that the comment can be scanned by | |
57 | pr_comment. | |
58 | ||
59 | Strings and character literals are returned just like identifiers. | |
60 | ||
61 | HISTORY: | |
62 | initial coding November 1976 D A Willcox of CAC | |
63 | 1/7/77 D A Willcox of CAC Fix to provide proper handling | |
64 | of "int a -1;" | |
65 | ||
66 | */\f | |
67 | ||
68 | /* Here we have the token scanner for indent. It scans off one token and | |
69 | puts it in the global variable "token". It returns a code, indicating the | |
70 | type of token scanned. */ | |
71 | ||
72 | #include "indent_globs.h"; | |
73 | #include "indent_codes.h"; | |
74 | ||
75 | ||
76 | ||
77 | #define alphanum 1 | |
78 | #define opchar 3 | |
79 | ||
80 | struct templ { | |
81 | char *rwd; | |
82 | int rwcode; | |
83 | }; | |
84 | ||
85 | struct templ specials[] = | |
86 | { | |
87 | "switch", 1, | |
88 | "case", 2, | |
89 | "struct", 3, | |
90 | "default", 2, | |
91 | "int", 4, | |
92 | "char", 4, | |
93 | "float", 4, | |
94 | "double", 4, | |
95 | "long", 4, | |
96 | "short", 4, | |
97 | "typdef", 4, | |
98 | "unsigned", 4, | |
99 | "register", 4, | |
100 | "static", 4, | |
101 | "global", 4, | |
102 | "extern", 4, | |
103 | "if", 5, | |
104 | "while", 5, | |
105 | "for", 5, | |
106 | "else", 6, | |
107 | "do", 6, | |
108 | "sizeof", 0, | |
109 | 0, 0 | |
110 | }; | |
111 | ||
112 | char chartype[128] = | |
113 | { /* this is used to facilitate the decision of what type | |
114 | (alphanumeric, operator) each character is */ | |
115 | 0, 0, 0, 0, 0, 0, 0, 0, | |
116 | 0, 0, 0, 0, 0, 0, 0, 0, | |
117 | 0, 0, 0, 0, 0, 0, 0, 0, | |
118 | 0, 0, 0, 0, 0, 0, 0, 0, | |
119 | 0, 3, 0, 0, 0, 3, 3, 0, | |
120 | 0, 0, 3, 3, 0, 3, 3, 3, | |
121 | 1, 1, 1, 1, 1, 1, 1, 1, | |
122 | 1, 1, 0, 0, 3, 3, 3, 3, | |
123 | 0, 1, 1, 1, 1, 1, 1, 1, | |
124 | 1, 1, 1, 1, 1, 1, 1, 1, | |
125 | 1, 1, 1, 1, 1, 1, 1, 1, | |
126 | 1, 1, 1, 0, 0, 0, 3, 1, | |
127 | 0, 1, 1, 1, 1, 1, 1, 1, | |
128 | 1, 1, 1, 1, 1, 1, 1, 1, | |
129 | 1, 1, 1, 1, 1, 1, 1, 1, | |
130 | 1, 1, 1, 0, 3, 0, 3, 0 | |
131 | }; | |
132 | ||
133 | int last_nl = true; | |
134 | /* this is true if the last thing scanned was a newline */ | |
135 | ||
136 | ||
137 | ||
138 | int lexi () { | |
139 | register char *tok; | |
140 | /* local pointer to next char in token */ | |
141 | register int i; | |
142 | /* local loop counter */ | |
143 | register char *j; | |
144 | /* used for searching thru list of reserved words */ | |
145 | int unary_delim; | |
146 | /* this is set to 1 if the current token forces a following operator to be | |
147 | unary */ | |
148 | static int last_code; | |
149 | /* the last token type returned */ | |
150 | static int l_struct; | |
151 | /* set to 1 if the last token was 'struct' */ | |
152 | int found_it; | |
153 | int code; /* internal code to be returned */ | |
154 | char qchar; /* the delimiter character for a string */ | |
155 | ||
156 | tok = token; /* point to start of place to save token */ | |
157 | unary_delim = false; | |
158 | col_1 = last_nl; /* tell world that this token started in column | |
159 | 1 iff the last thing scanned was nl */ | |
160 | last_nl = false; | |
161 | ||
162 | while (*buf_ptr == ' ' || *buf_ptr == '\t') { | |
163 | /* get rid of blanks */ | |
164 | col_1 = false; /* leading blanks imply token is not in column 1 | |
165 | */ | |
166 | if (++buf_ptr >= buf_end) | |
167 | fill_buffer (); | |
168 | } | |
169 | ||
170 | /*----------------------------------------------------------*\ | |
171 | | Scan an alphanumeric token | |
172 | \*----------------------------------------------------------*/ | |
173 | ||
174 | if (chartype[*buf_ptr & 0177] == alphanum) { | |
175 | /* we have a character or number */ | |
176 | while (chartype[*buf_ptr & 0177] == alphanum) { | |
177 | /* copy it over */ | |
178 | *tok++ = *buf_ptr++; | |
179 | if (buf_ptr >= buf_end) | |
180 | fill_buffer (); | |
181 | } | |
182 | ||
183 | *tok++ = '\0'; | |
184 | ||
185 | if (l_struct) { /* if last token was 'struct', then this token | |
186 | should be treated as a declaration */ | |
187 | l_struct = false; | |
188 | last_code = ident; | |
189 | last_u_d = true; | |
190 | return (decl); | |
191 | } | |
192 | ||
193 | last_u_d = false; /* operator after indentifier is binary */ | |
194 | ||
195 | for (i = 0; specials[i].rwd != 0; ++i) { | |
196 | /* this loop will check if the token is a keyword. if so, a following | |
197 | operator is unary */ | |
198 | last_code = ident; /* remember that this is the code we will return | |
199 | */ | |
200 | j = specials[i].rwd; | |
201 | /* point at ith reserved word */ | |
202 | tok = token; /* point at scanned toekn */ | |
203 | found_it = true; /* set to false if not found */ | |
204 | do { | |
205 | if (*tok++ != *j) { | |
206 | found_it = false; | |
207 | break; | |
208 | } | |
209 | } while (*j++); | |
210 | ||
211 | if (found_it) { /* we have a keyword */ | |
212 | last_u_d = true; | |
213 | switch (specials[i].rwcode) { | |
214 | case 1: /* it is a switch */ | |
215 | return (swstmt); | |
216 | case 2: /* a case or default */ | |
217 | return (casestmt); | |
218 | ||
219 | case 3: /* a "struct" */ | |
220 | l_struct = true; | |
221 | /* Next time around, we will want to know that we have had | |
222 | a 'struct' */ | |
223 | case 4: /* one of the declaration keywords */ | |
224 | if(p_l_follow) break; /* inside parens: cast */ | |
225 | last_code = decl; | |
226 | return (decl); | |
227 | ||
228 | case 5: /* if, while, for */ | |
229 | return (sp_paren); | |
230 | ||
231 | case 6: /* do, else */ | |
232 | return (sp_nparen); | |
233 | ||
234 | default: /* all others are treated like any other | |
235 | identifier */ | |
236 | return (ident); | |
237 | } /* end of switch */ | |
238 | } /* end of if (found_it) */ | |
239 | ||
240 | } | |
241 | ||
242 | if (last_code == decl) /* if this is a declared variable, then | |
243 | following sign is unary */ | |
244 | last_u_d = true; /* will make "int a -1" work */ | |
245 | last_code = ident; | |
246 | return (ident); /* the ident is not in the list */ | |
247 | } /* end of procesing for alpanum character */ | |
248 | ||
249 | ||
250 | ||
251 | /*----------------------------------------------------------*\ | |
252 | | Scan a non-alphanumeric token | |
253 | \*----------------------------------------------------------*/ | |
254 | ||
255 | *tok++ = *buf_ptr; /* if it is only a one-character token, it is | |
256 | moved here */ | |
257 | *tok = '\0'; | |
258 | if (++buf_ptr >= buf_end) | |
259 | fill_buffer (); | |
260 | ||
261 | switch (*token) { | |
262 | case '\n': | |
263 | unary_delim = last_u_d; | |
264 | last_nl = true; /* remember that we just had a newline */ | |
265 | code = (had_eof ? 0 : newline); | |
266 | /* if data has been exausted, the newline is a dummy, and we should | |
267 | return code to stop */ | |
268 | break; | |
269 | ||
270 | case '\'': /* start of quoted character */ | |
271 | qchar = '\''; /* remember final delimiter */ | |
272 | goto copy_lit; /* and go to common literal code */ | |
273 | ||
274 | case '"': /* start of string */ | |
275 | qchar = '"'; | |
276 | ||
277 | copy_lit: | |
278 | do { /* copy the string */ | |
279 | while (1) { /* move one character or [/<char>]<char> */ | |
280 | if (*buf_ptr == '\n') { | |
281 | /* check for unterminated literal */ | |
282 | printf ("%d: Unterminated literal\n", line_no); | |
283 | goto stop_lit; | |
284 | /* Don't copy any more */ | |
285 | } | |
286 | ||
287 | *tok = *buf_ptr++; | |
288 | if (buf_ptr >= buf_end) | |
289 | fill_buffer (); | |
290 | if (had_eof || ((tok - token) > (bufsize - 2))) { | |
291 | printf ("Unterminated literal\n"); | |
292 | ++tok; | |
293 | goto stop_lit; | |
294 | /* get outof literal copying loop */ | |
295 | } | |
296 | ||
297 | if (*tok == '\\') { | |
298 | /* if escape, copy extra char */ | |
299 | if (*buf_ptr == '\n') | |
300 | /* check for escaped newline */ | |
301 | ++line_no; | |
302 | *(++tok) = *buf_ptr++; | |
303 | ++tok; /* we must increment this again because we | |
304 | copied two chars */ | |
305 | if (buf_ptr >= buf_end) | |
306 | fill_buffer (); | |
307 | } | |
308 | else | |
309 | break; /* we copied one character */ | |
310 | } /* end of while (1) */ | |
311 | } while (*tok++ != qchar); | |
312 | ||
313 | stop_lit: | |
314 | code = ident; | |
315 | break; | |
316 | ||
317 | case ('('): | |
318 | case ('['): | |
319 | unary_delim = true; | |
320 | code = lparen; | |
321 | break; | |
322 | ||
323 | case (')'): | |
324 | case (']'): | |
325 | code = rparen; | |
326 | break; | |
327 | ||
328 | case '#': | |
329 | unary_delim = last_u_d; | |
330 | code = preesc; | |
331 | break; | |
332 | ||
333 | case '?': | |
334 | unary_delim = true; | |
335 | code = question; | |
336 | break; | |
337 | ||
338 | case (':'): | |
339 | code = colon; | |
340 | unary_delim = true; | |
341 | break; | |
342 | ||
343 | case (';'): | |
344 | unary_delim = true; | |
345 | code = semicolon; | |
346 | break; | |
347 | ||
348 | case ('{'): | |
349 | unary_delim = true; | |
350 | code = lbrace; | |
351 | break; | |
352 | ||
353 | case ('}'): | |
354 | unary_delim = true; | |
355 | code = rbrace; | |
356 | break; | |
357 | ||
358 | case 014: /* a form feed */ | |
359 | unary_delim = last_u_d; | |
360 | last_nl = true; /* remember this so we can set 'col_1' right */ | |
361 | code = form_feed; | |
362 | break; | |
363 | ||
364 | case (','): | |
365 | unary_delim = true; | |
366 | code = comma; | |
367 | break; | |
368 | ||
369 | case '.': | |
370 | unary_delim = false; | |
371 | code = period; | |
372 | break; | |
373 | ||
374 | case '-': | |
375 | case '+': /* check for -, +, --, ++ */ | |
376 | code = (last_u_d ? unary_op : binary_op); | |
377 | unary_delim = true; | |
378 | ||
379 | if (*buf_ptr == token[0]) { | |
380 | /* check for doubled character */ | |
381 | *tok++ = *buf_ptr++; | |
382 | /* buffer overflow will be checked at end of loop */ | |
383 | if (last_code == ident || last_code == rparen) { | |
384 | code = (last_u_d ? unary_op : postop); | |
385 | /* check for following ++ or -- */ | |
386 | unary_delim = false; | |
387 | } | |
388 | } | |
389 | else | |
390 | if (*buf_ptr == '>' || *buf_ptr == '=') | |
391 | /* check for operator -> or += */ | |
392 | *tok++ = *buf_ptr++; | |
393 | /* buffer overflow will be checked at end of switch */ | |
394 | ||
395 | break; | |
396 | ||
397 | case '=': | |
398 | if (chartype[*buf_ptr] == opchar) { | |
399 | /* we have two char assignment */ | |
400 | *tok++ = *buf_ptr; | |
401 | /* move second character */ | |
402 | if (++buf_ptr >= buf_end) | |
403 | fill_buffer (); | |
404 | } | |
405 | ||
406 | code = binary_op; | |
407 | unary_delim = true; | |
408 | if (token[1] != '<' && token[1] != '>') | |
409 | /* check for possible 3 char operator */ | |
410 | break; | |
411 | /* can drop thru!!! */ | |
412 | ||
413 | case '>': | |
414 | case '<': | |
415 | case '!': /* ops like <, <<, <=, !=, etc */ | |
416 | if (*buf_ptr == '>' || *buf_ptr == '<' || *buf_ptr == '=') { | |
417 | *tok++ = *buf_ptr; | |
418 | if (++buf_ptr >= buf_end) | |
419 | fill_buffer (); | |
420 | } | |
421 | ||
422 | if (*buf_ptr == '=') | |
423 | *tok++ = *buf_ptr++; | |
424 | code = (last_u_d ? unary_op : binary_op); | |
425 | unary_delim = true; | |
426 | break; | |
427 | ||
428 | default: | |
429 | if (token[0] == '/' && *buf_ptr == '*') { | |
430 | /* it is start of comment */ | |
431 | *tok++ = '*'; | |
432 | ||
433 | if (++buf_ptr >= buf_end) | |
434 | fill_buffer (); | |
435 | ||
436 | code = comment; | |
437 | unary_delim = last_u_d; | |
438 | break; | |
439 | } | |
440 | ||
441 | while (*(tok - 1) == *buf_ptr || *buf_ptr=='=') { | |
442 | /* handle ||, &&, etc, and also things as in int *****i */ | |
443 | *tok++ = *buf_ptr; | |
444 | if (++buf_ptr >= buf_end) | |
445 | fill_buffer (); | |
446 | } | |
447 | ||
448 | ||
449 | code = (last_u_d ? unary_op : binary_op); | |
450 | unary_delim = true; | |
451 | ||
452 | ||
453 | } /* end of switch */ | |
454 | ||
455 | if (code != newline) { | |
456 | l_struct = false; | |
457 | last_code = code; | |
458 | } | |
459 | ||
460 | if (buf_ptr >= buf_end) /* check for input buffer empty */ | |
461 | fill_buffer (); | |
462 | last_u_d = unary_delim; | |
463 | *tok = '\0'; /* null terminate the token */ | |
464 | return (code); | |
465 | }; |