1 | /* GNU SED, a batch stream editor.
|
---|
2 | Copyright (C) 2018-2022 Free Software Foundation, Inc.
|
---|
3 |
|
---|
4 | This program is free software; you can redistribute it and/or modify
|
---|
5 | it under the terms of the GNU General Public License as published by
|
---|
6 | the Free Software Foundation; either version 3, or (at your option)
|
---|
7 | any later version.
|
---|
8 |
|
---|
9 | This program is distributed in the hope that it will be useful,
|
---|
10 | but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
11 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
12 | GNU General Public License for more details.
|
---|
13 |
|
---|
14 | You should have received a copy of the GNU General Public License
|
---|
15 | along with this program; If not, see <https://www.gnu.org/licenses/>. */
|
---|
16 |
|
---|
17 | /* Written by Assaf Gordon. */
|
---|
18 |
|
---|
19 | /* debug.c: debugging functions */
|
---|
20 |
|
---|
21 | #include "sed.h"
|
---|
22 | #include "basicdefs.h"
|
---|
23 | #include <stdio.h>
|
---|
24 | #include <ctype.h>
|
---|
25 | #include <string.h>
|
---|
26 | #include <stdlib.h>
|
---|
27 | #include <sys/types.h>
|
---|
28 | #include <assert.h>
|
---|
29 |
|
---|
30 | /* indentation level when printing the program */
|
---|
31 | static int block_level = 0;
|
---|
32 |
|
---|
33 |
|
---|
34 | void
|
---|
35 | debug_print_char (char c)
|
---|
36 | {
|
---|
37 | if (ISPRINT (c) && c != '\\')
|
---|
38 | {
|
---|
39 | putchar (c);
|
---|
40 | return;
|
---|
41 | }
|
---|
42 |
|
---|
43 | putchar ('\\');
|
---|
44 | switch (c)
|
---|
45 | {
|
---|
46 | case '\a':
|
---|
47 | putchar ('a');
|
---|
48 | break;
|
---|
49 | case '\f':
|
---|
50 | putchar ('f');
|
---|
51 | break;
|
---|
52 | case '\r':
|
---|
53 | putchar ('r');
|
---|
54 | break;
|
---|
55 | case '\t':
|
---|
56 | putchar ('t');
|
---|
57 | break;
|
---|
58 | case '\v':
|
---|
59 | putchar ('v');
|
---|
60 | break;
|
---|
61 | case '\n':
|
---|
62 | putchar ('n');
|
---|
63 | break;
|
---|
64 | case '\\':
|
---|
65 | putchar ('\\');
|
---|
66 | break;
|
---|
67 |
|
---|
68 | default:
|
---|
69 | printf ("o%03o", (unsigned int) c);
|
---|
70 | }
|
---|
71 | }
|
---|
72 |
|
---|
73 | static void
|
---|
74 | debug_print_regex_pattern (const char *pat, size_t len)
|
---|
75 | {
|
---|
76 | const char *p = pat;
|
---|
77 | while (len--)
|
---|
78 | {
|
---|
79 | if (*p == '/')
|
---|
80 | fputs ("\\/", stdout);
|
---|
81 | else
|
---|
82 | debug_print_char (*p);
|
---|
83 | ++p;
|
---|
84 | }
|
---|
85 | }
|
---|
86 |
|
---|
87 | static void
|
---|
88 | debug_print_regex_flags (const struct regex *r, bool addr)
|
---|
89 | {
|
---|
90 | if (!r)
|
---|
91 | return;
|
---|
92 |
|
---|
93 | #ifdef REG_PERL
|
---|
94 | if (r->flags & REG_DOTALL) /* REG_PERL */
|
---|
95 | putchar ('s');
|
---|
96 | if (r->flags & REG_EXTENDED) /* REG_PERL */
|
---|
97 | putchar ('x');
|
---|
98 | #endif
|
---|
99 |
|
---|
100 | if (r->flags & REG_ICASE)
|
---|
101 | putchar (addr ? 'I' : 'i');
|
---|
102 | if (r->flags & REG_NEWLINE)
|
---|
103 | putchar (addr ? 'M' : 'm');
|
---|
104 | }
|
---|
105 |
|
---|
106 | static void
|
---|
107 | debug_print_regex (const struct regex *r)
|
---|
108 | {
|
---|
109 | if (!r)
|
---|
110 | {
|
---|
111 | /* Previous Regex */
|
---|
112 | fputs ("//", stdout);
|
---|
113 | return;
|
---|
114 | }
|
---|
115 |
|
---|
116 | putchar ('/');
|
---|
117 | debug_print_regex_pattern (r->re, r->sz);
|
---|
118 | putchar ('/');
|
---|
119 | }
|
---|
120 |
|
---|
121 | static void
|
---|
122 | debug_print_addr (const struct addr *a)
|
---|
123 | {
|
---|
124 | if (!a)
|
---|
125 | return;
|
---|
126 | switch (a->addr_type)
|
---|
127 | {
|
---|
128 | case ADDR_IS_NULL:
|
---|
129 | fputs ("[ADDR-NULL]", stdout);
|
---|
130 | break;
|
---|
131 | case ADDR_IS_REGEX:
|
---|
132 | debug_print_regex (a->addr_regex);
|
---|
133 | debug_print_regex_flags (a->addr_regex, true);
|
---|
134 | break;
|
---|
135 | case ADDR_IS_NUM:
|
---|
136 | printf ("%lu", a->addr_number);
|
---|
137 | break;
|
---|
138 | case ADDR_IS_NUM_MOD:
|
---|
139 | printf ("%lu~%lu", a->addr_number, a->addr_step);
|
---|
140 | break;
|
---|
141 | case ADDR_IS_STEP:
|
---|
142 | printf ("+%lu", a->addr_step);
|
---|
143 | break;
|
---|
144 | case ADDR_IS_STEP_MOD:
|
---|
145 | printf ("~%lu", a->addr_step);
|
---|
146 | break;
|
---|
147 | case ADDR_IS_LAST:
|
---|
148 | putchar ('$');
|
---|
149 | break;
|
---|
150 | }
|
---|
151 | }
|
---|
152 |
|
---|
153 | static void
|
---|
154 | debug_print_subst_replacement (const struct replacement *r)
|
---|
155 | {
|
---|
156 | enum replacement_types last_repl_type = REPL_ASIS;
|
---|
157 |
|
---|
158 | if (!r)
|
---|
159 | return;
|
---|
160 |
|
---|
161 | const struct replacement *p = r;
|
---|
162 | while (p)
|
---|
163 | {
|
---|
164 | if (p->repl_type != last_repl_type)
|
---|
165 | {
|
---|
166 | /* Special GNU replacements \E\U\u\L\l should be printed
|
---|
167 | BEFORE the 'prefix' .... the 'prefix' refers to being
|
---|
168 | before the backreference. */
|
---|
169 | putchar ('\\');
|
---|
170 | if (p->repl_type == 0)
|
---|
171 | putchar ('E');
|
---|
172 | else if (p->repl_type == REPL_UPPERCASE)
|
---|
173 | putchar ('U');
|
---|
174 | else if (p->repl_type == REPL_LOWERCASE)
|
---|
175 | putchar ('L');
|
---|
176 | else if ((p->repl_type & REPL_MODIFIERS) == REPL_UPPERCASE_FIRST)
|
---|
177 | putchar ('u');
|
---|
178 | else if ((p->repl_type & REPL_MODIFIERS) == REPL_LOWERCASE_FIRST)
|
---|
179 | putchar ('l');
|
---|
180 |
|
---|
181 | last_repl_type = p->repl_type;
|
---|
182 | }
|
---|
183 |
|
---|
184 | if (p->prefix_length)
|
---|
185 | fwrite (p->prefix, 1, p->prefix_length, stdout);
|
---|
186 |
|
---|
187 | if (p->subst_id != -1)
|
---|
188 | {
|
---|
189 | if (p->subst_id == 0)
|
---|
190 | putchar ('&');
|
---|
191 | else
|
---|
192 | printf ("\\%d", p->subst_id);
|
---|
193 | }
|
---|
194 |
|
---|
195 | p = p->next;
|
---|
196 | }
|
---|
197 | }
|
---|
198 |
|
---|
199 | static void
|
---|
200 | debug_print_output_file (const struct output *o)
|
---|
201 | {
|
---|
202 | if (!o)
|
---|
203 | return;
|
---|
204 |
|
---|
205 | fputs (o->name, stdout);
|
---|
206 | }
|
---|
207 |
|
---|
208 | static void
|
---|
209 | debug_print_subst (const struct subst *s)
|
---|
210 | {
|
---|
211 | if (!s)
|
---|
212 | return;
|
---|
213 |
|
---|
214 | debug_print_regex (s->regx);
|
---|
215 | debug_print_subst_replacement (s->replacement);
|
---|
216 | putchar ('/');
|
---|
217 |
|
---|
218 | debug_print_regex_flags (s->regx, false);
|
---|
219 |
|
---|
220 | if (s->global)
|
---|
221 | putchar ('g');
|
---|
222 | if (s->eval)
|
---|
223 | putchar ('e');
|
---|
224 | if (s->print)
|
---|
225 | putchar ('p');
|
---|
226 | if (s->numb)
|
---|
227 | printf ("%lu", s->numb);
|
---|
228 | if (s->outf)
|
---|
229 | {
|
---|
230 | putchar ('w');
|
---|
231 | debug_print_output_file (s->outf);
|
---|
232 | }
|
---|
233 | }
|
---|
234 |
|
---|
235 | static void
|
---|
236 | debug_print_translation (const struct sed_cmd *sc)
|
---|
237 | {
|
---|
238 | unsigned int i;
|
---|
239 |
|
---|
240 | if (mb_cur_max > 1)
|
---|
241 | {
|
---|
242 | /* multibyte translation */
|
---|
243 | putchar ('/');
|
---|
244 | for (i = 0; sc->x.translatemb[2 * i] != NULL; i++)
|
---|
245 | fputs (sc->x.translatemb[2 * i], stdout);
|
---|
246 | putchar ('/');
|
---|
247 | for (i = 0; sc->x.translatemb[2 * i] != NULL; i++)
|
---|
248 | fputs (sc->x.translatemb[2 * i + 1], stdout);
|
---|
249 | putchar ('/');
|
---|
250 | }
|
---|
251 | else
|
---|
252 | {
|
---|
253 | /* unibyte translation */
|
---|
254 | putchar ('/');
|
---|
255 | for (i = 0; i < 256; ++i)
|
---|
256 | if (sc->x.translate[i] != (unsigned char) i)
|
---|
257 | putchar ((unsigned char) i);
|
---|
258 | putchar ('/');
|
---|
259 | for (i = 0; i < 256; ++i)
|
---|
260 | if (sc->x.translate[i] != (unsigned char) i)
|
---|
261 | putchar (sc->x.translate[i]);
|
---|
262 | putchar ('/');
|
---|
263 | }
|
---|
264 | }
|
---|
265 |
|
---|
266 | static void
|
---|
267 | debug_print_function (const struct vector *program, const struct sed_cmd *sc)
|
---|
268 | {
|
---|
269 | if (!sc)
|
---|
270 | return;
|
---|
271 |
|
---|
272 | putchar (sc->cmd);
|
---|
273 |
|
---|
274 | switch (sc->cmd) /* LCOV_EXCL_BR */
|
---|
275 | {
|
---|
276 | case '=':
|
---|
277 | break;
|
---|
278 |
|
---|
279 | case ':':
|
---|
280 | printf ("%s", sc->x.label_name);
|
---|
281 | break;
|
---|
282 |
|
---|
283 | case '{':
|
---|
284 | break;
|
---|
285 |
|
---|
286 | case '}':
|
---|
287 | break;
|
---|
288 |
|
---|
289 | case '#': /* LCOV_EXCL_LINE */
|
---|
290 | /* should not happen - discarded during compilation. */
|
---|
291 | assert (0); /* LCOV_EXCL_LINE */
|
---|
292 |
|
---|
293 | case 'a':
|
---|
294 | case 'c':
|
---|
295 | case 'i':
|
---|
296 | fputs ("\\", stdout);
|
---|
297 | if (sc->x.cmd_txt.text_length)
|
---|
298 | fwrite (sc->x.cmd_txt.text, 1, sc->x.cmd_txt.text_length, stdout);
|
---|
299 | break;
|
---|
300 |
|
---|
301 | case 'b':
|
---|
302 | case 't':
|
---|
303 | case 'T':
|
---|
304 | {
|
---|
305 | if (sc->x.jump_index < program->v_length)
|
---|
306 | {
|
---|
307 | const char *label_name = program->v[sc->x.jump_index].x.label_name;
|
---|
308 | if (label_name)
|
---|
309 | printf (" %s", label_name);
|
---|
310 | }
|
---|
311 | }
|
---|
312 | break;
|
---|
313 |
|
---|
314 | case 'D':
|
---|
315 | break;
|
---|
316 |
|
---|
317 | case 'd':
|
---|
318 | break;
|
---|
319 |
|
---|
320 | case 'e':
|
---|
321 | putchar (' ');
|
---|
322 | fwrite (sc->x.cmd_txt.text, 1, sc->x.cmd_txt.text_length, stdout);
|
---|
323 | break;
|
---|
324 |
|
---|
325 | case 'F':
|
---|
326 | break;
|
---|
327 |
|
---|
328 | case 'g':
|
---|
329 | break;
|
---|
330 |
|
---|
331 | case 'G':
|
---|
332 | break;
|
---|
333 |
|
---|
334 | case 'h':
|
---|
335 | break;
|
---|
336 |
|
---|
337 | case 'H':
|
---|
338 | break;
|
---|
339 |
|
---|
340 | /* 'i' is lumped above with 'a' and 'c' */
|
---|
341 |
|
---|
342 | case 'L':
|
---|
343 | case 'l':
|
---|
344 | case 'q':
|
---|
345 | case 'Q':
|
---|
346 | if (sc->x.int_arg != -1)
|
---|
347 | printf (" %d", sc->x.int_arg);
|
---|
348 | break;
|
---|
349 |
|
---|
350 | case 'n':
|
---|
351 | break;
|
---|
352 |
|
---|
353 | case 'N':
|
---|
354 | break;
|
---|
355 |
|
---|
356 | case 'P':
|
---|
357 | break;
|
---|
358 |
|
---|
359 | case 'p':
|
---|
360 | break;
|
---|
361 |
|
---|
362 | /* 'q','Q' are lumped above with 'L' and 'l' */
|
---|
363 |
|
---|
364 | case 'r':
|
---|
365 | putchar (' ');
|
---|
366 | fputs (sc->x.readcmd.fname, stdout);
|
---|
367 | break;
|
---|
368 |
|
---|
369 | case 'R':
|
---|
370 | putchar (' ');
|
---|
371 | fputs (sc->x.inf->name, stdout);
|
---|
372 | break;
|
---|
373 |
|
---|
374 | case 's':
|
---|
375 | debug_print_subst (sc->x.cmd_subst);
|
---|
376 | break;
|
---|
377 |
|
---|
378 | /* 't','T' are lumped above with 'b' */
|
---|
379 |
|
---|
380 | case 'v': /* LCOV_EXCL_LINE */
|
---|
381 | /* should not happen - handled during compilation then discarded. */
|
---|
382 | assert (0); /* LCOV_EXCL_LINE */
|
---|
383 |
|
---|
384 | case 'W':
|
---|
385 | debug_print_output_file (sc->x.outf);
|
---|
386 | break;
|
---|
387 |
|
---|
388 | case 'w':
|
---|
389 | debug_print_output_file (sc->x.outf);
|
---|
390 | break;
|
---|
391 |
|
---|
392 | case 'x':
|
---|
393 | break;
|
---|
394 |
|
---|
395 | case 'y':
|
---|
396 | debug_print_translation (sc);
|
---|
397 | break;
|
---|
398 |
|
---|
399 | case 'z':
|
---|
400 | break;
|
---|
401 |
|
---|
402 | default: /* LCOV_EXCL_LINE */
|
---|
403 | /* should not happen - unless missed a sed command. */
|
---|
404 | assert (0); /* LCOV_EXCL_LINE */
|
---|
405 | }
|
---|
406 | }
|
---|
407 |
|
---|
408 | void
|
---|
409 | debug_print_command (const struct vector *program, const struct sed_cmd *sc)
|
---|
410 | {
|
---|
411 | bool addr_bang;
|
---|
412 | if (!program)
|
---|
413 | return;
|
---|
414 |
|
---|
415 | if (sc->cmd == '}')
|
---|
416 | --block_level;
|
---|
417 |
|
---|
418 | for (int j = 0; j < block_level; ++j)
|
---|
419 | fputs (" ", stdout);
|
---|
420 |
|
---|
421 | debug_print_addr (sc->a1);
|
---|
422 | if (sc->a2)
|
---|
423 | putchar (',');
|
---|
424 | debug_print_addr (sc->a2);
|
---|
425 |
|
---|
426 | addr_bang = sc->addr_bang;
|
---|
427 | /* Implmentation detail: GNU Sed implements beginning of block
|
---|
428 | by negating the matched address and jumping if there's no match. */
|
---|
429 | if (sc->cmd == '{')
|
---|
430 | addr_bang = !addr_bang;
|
---|
431 | if (addr_bang)
|
---|
432 | putchar ('!');
|
---|
433 |
|
---|
434 | if (sc->a1 || sc->a2)
|
---|
435 | putchar (' ');
|
---|
436 |
|
---|
437 | debug_print_function (program, sc);
|
---|
438 |
|
---|
439 | putchar ('\n');
|
---|
440 |
|
---|
441 | if (sc->cmd == '{')
|
---|
442 | ++block_level;
|
---|
443 | }
|
---|
444 |
|
---|
445 | void
|
---|
446 | debug_print_program (const struct vector *program)
|
---|
447 | {
|
---|
448 | if (!program)
|
---|
449 | return;
|
---|
450 |
|
---|
451 | block_level = 1;
|
---|
452 | puts ("SED PROGRAM:");
|
---|
453 | for (size_t i = 0; i < program->v_length; ++i)
|
---|
454 | debug_print_command (program, &program->v[i]);
|
---|
455 | block_level = 0;
|
---|
456 | }
|
---|