]>
Commit | Line | Data |
---|---|---|
104b92f8 PB |
1 | /* Align/Truncate a string in a given screen width |
2 | Copyright (C) 2009-2010 Free Software Foundation, Inc. | |
3 | ||
4 | This program is free software: you can redistribute it and/or modify | |
36c7f785 PB |
5 | it under the terms of the GNU Lesser General Public License as published by |
6 | the Free Software Foundation, either version 2.1 of the License, or | |
104b92f8 PB |
7 | (at your option) 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 <http://www.gnu.org/licenses/>. */ | |
16 | ||
17 | /* Written by Pádraig Brady. */ | |
18 | ||
19 | #include <config.h> | |
20 | ||
21 | #include <stdlib.h> | |
22 | #include <string.h> | |
23 | #include <stdio.h> | |
24 | #include <stdbool.h> | |
25 | #include <limits.h> | |
1b1f66e4 | 26 | #include <ctype.h> |
104b92f8 PB |
27 | |
28 | #include "c.h" | |
29 | #include "mbsalign.h" | |
30 | #include "widechar.h" | |
31 | ||
104b92f8 PB |
32 | #ifdef HAVE_WIDECHAR |
33 | /* Replace non printable chars. | |
34 | Note \t and \n etc. are non printable. | |
35 | Return 1 if replacement made, 0 otherwise. */ | |
36 | ||
1b1f66e4 KZ |
37 | /* |
38 | * Counts number of cells in multibyte string. For all control and | |
39 | * non-printable chars is the result width enlarged to store \x?? hex | |
40 | * sequence. See mbs_safe_encode(). | |
41 | */ | |
42 | size_t mbs_safe_width(const char *s) | |
43 | { | |
44 | mbstate_t st; | |
45 | const char *p = s; | |
46 | size_t width = 0; | |
47 | ||
48 | memset(&st, 0, sizeof(st)); | |
49 | ||
50 | while (p && *p) { | |
51 | if (iscntrl((unsigned char) *p)) { | |
52 | width += 4; /* *p encoded to \x?? */ | |
53 | p++; | |
54 | } | |
55 | #ifdef HAVE_WIDECHAR | |
56 | else { | |
57 | wchar_t wc; | |
58 | size_t len = mbrtowc(&wc, p, MB_CUR_MAX, &st); | |
59 | ||
60 | if (len == 0) | |
61 | break; | |
62 | ||
63 | if (len == (size_t) -1 || len == (size_t) -2) { | |
64 | len = 1; | |
65 | width += (isprint((unsigned char) *p) ? 1 : 4); | |
66 | ||
67 | } if (!iswprint(wc)) | |
68 | width += len * 4; /* hex encode whole sequence */ | |
69 | else | |
70 | width += wcwidth(wc); /* number of cells */ | |
71 | p += len; | |
72 | } | |
73 | #else | |
74 | else if (!isprint((unsigned char) *p)) { | |
75 | width += 4; /* *p encoded to \x?? */ | |
76 | p++; | |
77 | } else { | |
78 | width++; | |
79 | p++; | |
80 | } | |
81 | #endif | |
82 | } | |
83 | ||
84 | return width; | |
85 | } | |
86 | ||
87 | /* | |
88 | * Returns allocated string where all control and non-printable chars are | |
89 | * replaced with \x?? hex sequence. | |
90 | */ | |
91 | char *mbs_safe_encode(const char *s, size_t *width) | |
92 | { | |
93 | mbstate_t st; | |
94 | const char *p = s; | |
95 | char *res, *r; | |
96 | size_t sz = s ? strlen(s) : 0; | |
97 | ||
98 | ||
99 | if (!sz) | |
100 | return NULL; | |
101 | ||
102 | memset(&st, 0, sizeof(st)); | |
103 | ||
104 | res = malloc((sz * 4) + 1); | |
105 | if (!res) | |
106 | return NULL; | |
107 | ||
108 | r = res; | |
109 | *width = 0; | |
110 | ||
111 | while (p && *p) { | |
112 | if (iscntrl((unsigned char) *p)) { | |
113 | sprintf(r, "\\x%02x", (unsigned char) *p); | |
114 | r += 4; | |
115 | *width += 4; | |
116 | p++; | |
117 | } | |
118 | #ifdef HAVE_WIDECHAR | |
119 | else { | |
120 | wchar_t wc; | |
121 | size_t len = mbrtowc(&wc, p, MB_CUR_MAX, &st); | |
122 | ||
123 | if (len == 0) | |
124 | break; /* end of string */ | |
125 | ||
126 | if (len == (size_t) -1 || len == (size_t) -2) { | |
127 | len = 1; | |
128 | /* | |
129 | * Not valid multibyte sequence -- maybe it's | |
130 | * printable char according to the current locales. | |
131 | */ | |
132 | if (!isprint((unsigned char) *p)) { | |
133 | sprintf(r, "\\x%02x", (unsigned char) *p); | |
134 | r += 4; | |
135 | *width += 4; | |
136 | } else { | |
137 | width++; | |
138 | *r++ = *p; | |
139 | } | |
140 | } else if (!iswprint(wc)) { | |
141 | size_t i; | |
142 | for (i = 0; i < len; i++) { | |
143 | sprintf(r, "\\x%02x", (unsigned char) *p); | |
144 | r += 4; | |
145 | *width += 4; | |
146 | } | |
147 | } else { | |
148 | memcpy(r, p, len); | |
149 | r += len; | |
150 | *width += wcwidth(wc); | |
151 | } | |
152 | p += len; | |
153 | } | |
154 | #else | |
155 | else if (!isprint((unsigned char) *p)) { | |
156 | sprintf(r, "\\x%02x", (unsigned char) *p); | |
157 | p++; | |
158 | r += 4; | |
159 | *width += 4; | |
160 | } else { | |
161 | *r++ = *p++; | |
162 | *width++; | |
163 | } | |
164 | #endif | |
165 | } | |
166 | ||
167 | *r = '\0'; | |
168 | ||
169 | return res; | |
170 | } | |
171 | ||
104b92f8 PB |
172 | static bool |
173 | wc_ensure_printable (wchar_t *wchars) | |
174 | { | |
175 | bool replaced = false; | |
176 | wchar_t *wc = wchars; | |
177 | while (*wc) | |
178 | { | |
179 | if (!iswprint ((wint_t) *wc)) | |
180 | { | |
181 | *wc = 0xFFFD; /* L'\uFFFD' (replacement char) */ | |
182 | replaced = true; | |
183 | } | |
184 | wc++; | |
185 | } | |
186 | return replaced; | |
187 | } | |
188 | ||
189 | /* Truncate wchar string to width cells. | |
190 | * Returns number of cells used. */ | |
191 | ||
192 | static size_t | |
193 | wc_truncate (wchar_t *wc, size_t width) | |
194 | { | |
195 | size_t cells = 0; | |
196 | int next_cells = 0; | |
197 | ||
198 | while (*wc) | |
199 | { | |
200 | next_cells = wcwidth (*wc); | |
201 | if (next_cells == -1) /* non printable */ | |
202 | { | |
203 | *wc = 0xFFFD; /* L'\uFFFD' (replacement char) */ | |
204 | next_cells = 1; | |
205 | } | |
206 | if (cells + next_cells > width) | |
207 | break; | |
208 | cells += next_cells; | |
209 | wc++; | |
210 | } | |
211 | *wc = L'\0'; | |
212 | return cells; | |
213 | } | |
214 | ||
215 | /* FIXME: move this function to gnulib as it's missing on: | |
216 | OpenBSD 3.8, IRIX 5.3, Solaris 2.5.1, mingw, BeOS */ | |
217 | ||
218 | static int | |
219 | rpl_wcswidth (const wchar_t *s, size_t n) | |
220 | { | |
221 | int ret = 0; | |
222 | ||
223 | while (n-- > 0 && *s != L'\0') | |
224 | { | |
225 | int nwidth = wcwidth (*s++); | |
226 | if (nwidth == -1) /* non printable */ | |
227 | return -1; | |
228 | if (ret > (INT_MAX - nwidth)) /* overflow */ | |
229 | return -1; | |
230 | ret += nwidth; | |
231 | } | |
232 | ||
233 | return ret; | |
234 | } | |
235 | #endif | |
236 | ||
5f94ca33 KZ |
237 | /* Truncate multi-byte string to @width and returns number of |
238 | * bytes of the new string @str, and in @width returns number | |
239 | * of cells. | |
240 | */ | |
241 | size_t | |
242 | mbs_truncate(char *str, size_t *width) | |
243 | { | |
2897f29a | 244 | ssize_t bytes = strlen(str); |
5f94ca33 | 245 | #ifdef HAVE_WIDECHAR |
2897f29a | 246 | ssize_t sz = mbstowcs(NULL, str, 0); |
5f94ca33 KZ |
247 | wchar_t *wcs = NULL; |
248 | ||
2897f29a | 249 | if (sz == (ssize_t) -1) |
5f94ca33 KZ |
250 | goto done; |
251 | ||
252 | wcs = malloc((sz + 1) * sizeof(wchar_t)); | |
253 | if (!wcs) | |
254 | goto done; | |
255 | ||
256 | if (!mbstowcs(wcs, str, sz)) | |
257 | goto done; | |
258 | *width = wc_truncate(wcs, *width); | |
259 | bytes = wcstombs(str, wcs, bytes); | |
260 | done: | |
261 | free(wcs); | |
262 | #else | |
263 | if (*width < bytes) | |
264 | bytes = *width; | |
265 | #endif | |
266 | if (bytes >= 0) | |
267 | str[bytes] = '\0'; | |
268 | return bytes; | |
269 | } | |
270 | ||
104b92f8 PB |
271 | /* Write N_SPACES space characters to DEST while ensuring |
272 | nothing is written beyond DEST_END. A terminating NUL | |
273 | is always added to DEST. | |
274 | A pointer to the terminating NUL is returned. */ | |
275 | ||
276 | static char* | |
277 | mbs_align_pad (char *dest, const char* dest_end, size_t n_spaces) | |
278 | { | |
279 | /* FIXME: Should we pad with "figure space" (\u2007) | |
280 | if non ascii data present? */ | |
281 | while (n_spaces-- && (dest < dest_end)) | |
282 | *dest++ = ' '; | |
283 | *dest = '\0'; | |
284 | return dest; | |
285 | } | |
286 | ||
287 | /* Align a string, SRC, in a field of *WIDTH columns, handling multi-byte | |
288 | characters; write the result into the DEST_SIZE-byte buffer, DEST. | |
289 | ALIGNMENT specifies whether to left- or right-justify or to center. | |
290 | If SRC requires more than *WIDTH columns, truncate it to fit. | |
291 | When centering, the number of trailing spaces may be one less than the | |
292 | number of leading spaces. The FLAGS parameter is unused at present. | |
293 | Return the length in bytes required for the final result, not counting | |
294 | the trailing NUL. A return value of DEST_SIZE or larger means there | |
295 | wasn't enough space. DEST will be NUL terminated in any case. | |
296 | Return (size_t) -1 upon error (invalid multi-byte sequence in SRC, | |
297 | or malloc failure), unless MBA_UNIBYTE_FALLBACK is specified. | |
298 | Update *WIDTH to indicate how many columns were used before padding. */ | |
299 | ||
300 | size_t | |
301 | mbsalign (const char *src, char *dest, size_t dest_size, | |
302 | size_t *width, mbs_align_t align, int flags) | |
303 | { | |
304 | size_t ret = -1; | |
305 | size_t src_size = strlen (src) + 1; | |
306 | char *newstr = NULL; | |
307 | wchar_t *str_wc = NULL; | |
308 | const char *str_to_print = src; | |
309 | size_t n_cols = src_size - 1; | |
310 | size_t n_used_bytes = n_cols; /* Not including NUL */ | |
3acc206d | 311 | size_t n_spaces = 0, space_left; |
104b92f8 PB |
312 | bool conversion = false; |
313 | bool wc_enabled = false; | |
314 | ||
315 | #ifdef HAVE_WIDECHAR | |
316 | /* In multi-byte locales convert to wide characters | |
317 | to allow easy truncation. Also determine number | |
318 | of screen columns used. */ | |
319 | if (MB_CUR_MAX > 1) | |
320 | { | |
321 | size_t src_chars = mbstowcs (NULL, src, 0); | |
322 | if (src_chars == (size_t) -1) | |
323 | { | |
324 | if (flags & MBA_UNIBYTE_FALLBACK) | |
325 | goto mbsalign_unibyte; | |
326 | else | |
327 | goto mbsalign_cleanup; | |
328 | } | |
329 | src_chars += 1; /* make space for NUL */ | |
330 | str_wc = malloc (src_chars * sizeof (wchar_t)); | |
331 | if (str_wc == NULL) | |
332 | { | |
333 | if (flags & MBA_UNIBYTE_FALLBACK) | |
334 | goto mbsalign_unibyte; | |
335 | else | |
336 | goto mbsalign_cleanup; | |
337 | } | |
338 | if (mbstowcs (str_wc, src, src_chars) != 0) | |
339 | { | |
340 | str_wc[src_chars - 1] = L'\0'; | |
341 | wc_enabled = true; | |
342 | conversion = wc_ensure_printable (str_wc); | |
343 | n_cols = rpl_wcswidth (str_wc, src_chars); | |
344 | } | |
345 | } | |
346 | ||
347 | /* If we transformed or need to truncate the source string | |
348 | then create a modified copy of it. */ | |
349 | if (wc_enabled && (conversion || (n_cols > *width))) | |
350 | { | |
351 | if (conversion) | |
352 | { | |
353 | /* May have increased the size by converting | |
354 | \t to \uFFFD for example. */ | |
355 | src_size = wcstombs(NULL, str_wc, 0) + 1; | |
356 | } | |
357 | newstr = malloc (src_size); | |
358 | if (newstr == NULL) | |
359 | { | |
360 | if (flags & MBA_UNIBYTE_FALLBACK) | |
361 | goto mbsalign_unibyte; | |
362 | else | |
363 | goto mbsalign_cleanup; | |
364 | } | |
365 | str_to_print = newstr; | |
366 | n_cols = wc_truncate (str_wc, *width); | |
367 | n_used_bytes = wcstombs (newstr, str_wc, src_size); | |
368 | } | |
369 | #endif | |
370 | ||
371 | mbsalign_unibyte: | |
372 | ||
373 | if (n_cols > *width) /* Unibyte truncation required. */ | |
374 | { | |
375 | n_cols = *width; | |
376 | n_used_bytes = n_cols; | |
377 | } | |
378 | ||
379 | if (*width > n_cols) /* Padding required. */ | |
380 | n_spaces = *width - n_cols; | |
381 | ||
382 | /* indicate to caller how many cells needed (not including padding). */ | |
383 | *width = n_cols; | |
384 | ||
385 | /* indicate to caller how many bytes needed (not including NUL). */ | |
386 | ret = n_used_bytes + (n_spaces * 1); | |
387 | ||
388 | /* Write as much NUL terminated output to DEST as possible. */ | |
389 | if (dest_size != 0) | |
390 | { | |
391 | char *dest_end = dest + dest_size - 1; | |
f7ff0414 SK |
392 | size_t start_spaces; |
393 | size_t end_spaces; | |
104b92f8 PB |
394 | |
395 | switch (align) | |
396 | { | |
397 | case MBS_ALIGN_CENTER: | |
398 | start_spaces = n_spaces / 2 + n_spaces % 2; | |
399 | end_spaces = n_spaces / 2; | |
400 | break; | |
401 | case MBS_ALIGN_LEFT: | |
402 | start_spaces = 0; | |
403 | end_spaces = n_spaces; | |
404 | break; | |
405 | case MBS_ALIGN_RIGHT: | |
406 | start_spaces = n_spaces; | |
407 | end_spaces = 0; | |
408 | break; | |
ce7b11a9 SK |
409 | default: |
410 | abort(); | |
104b92f8 PB |
411 | } |
412 | ||
413 | dest = mbs_align_pad (dest, dest_end, start_spaces); | |
3acc206d | 414 | space_left = dest_end - dest; |
104b92f8 PB |
415 | dest = mempcpy (dest, str_to_print, min (n_used_bytes, space_left)); |
416 | mbs_align_pad (dest, dest_end, end_spaces); | |
417 | } | |
418 | ||
419 | mbsalign_cleanup: | |
420 | ||
421 | free (str_wc); | |
422 | free (newstr); | |
423 | ||
424 | return ret; | |
425 | } |