Annotation of mandoc/term_ascii.c, Revision 1.62
1.62 ! schwarze 1: /* $Id: term_ascii.c,v 1.61 2018/05/20 21:37:34 schwarze Exp $ */
1.1 kristaps 2: /*
1.18 schwarze 3: * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.60 schwarze 4: * Copyright (c) 2014, 2015, 2017, 2018 Ingo Schwarze <schwarze@openbsd.org>
1.1 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
9: *
1.44 schwarze 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.1 kristaps 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.44 schwarze 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.1 kristaps 13: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
17: */
18: #include "config.h"
19:
20: #include <sys/types.h>
21:
1.39 schwarze 22: #include <assert.h>
1.31 schwarze 23: #if HAVE_WCHAR
1.60 schwarze 24: #include <langinfo.h>
1.28 schwarze 25: #include <locale.h>
1.15 kristaps 26: #endif
1.1 kristaps 27: #include <stdint.h>
28: #include <stdio.h>
29: #include <stdlib.h>
1.60 schwarze 30: #include <string.h>
1.2 kristaps 31: #include <unistd.h>
1.31 schwarze 32: #if HAVE_WCHAR
1.28 schwarze 33: #include <wchar.h>
1.15 kristaps 34: #endif
1.1 kristaps 35:
1.11 kristaps 36: #include "mandoc.h"
1.22 schwarze 37: #include "mandoc_aux.h"
1.1 kristaps 38: #include "out.h"
39: #include "term.h"
1.44 schwarze 40: #include "manconf.h"
1.1 kristaps 41: #include "main.h"
42:
1.51 schwarze 43: static struct termp *ascii_init(enum termenc, const struct manoutput *);
1.45 schwarze 44: static int ascii_hspan(const struct termp *,
1.7 kristaps 45: const struct roffsu *);
1.13 kristaps 46: static size_t ascii_width(const struct termp *, int);
1.7 kristaps 47: static void ascii_advance(struct termp *, size_t);
48: static void ascii_begin(struct termp *);
49: static void ascii_end(struct termp *);
1.1 kristaps 50: static void ascii_endline(struct termp *);
1.13 kristaps 51: static void ascii_letter(struct termp *, int);
1.45 schwarze 52: static void ascii_setwidth(struct termp *, int, int);
1.1 kristaps 53:
1.31 schwarze 54: #if HAVE_WCHAR
1.15 kristaps 55: static void locale_advance(struct termp *, size_t);
56: static void locale_endline(struct termp *);
57: static void locale_letter(struct termp *, int);
58: static size_t locale_width(const struct termp *, int);
59: #endif
60:
1.25 schwarze 61:
1.14 kristaps 62: static struct termp *
1.51 schwarze 63: ascii_init(enum termenc enc, const struct manoutput *outopts)
1.1 kristaps 64: {
1.46 schwarze 65: #if HAVE_WCHAR
1.1 kristaps 66: char *v;
1.46 schwarze 67: #endif
1.14 kristaps 68: struct termp *p;
1.1 kristaps 69:
1.57 schwarze 70: p = mandoc_calloc(1, sizeof(*p));
71: p->tcol = p->tcols = mandoc_calloc(1, sizeof(*p->tcol));
72: p->maxtcol = 1;
1.1 kristaps 73:
1.47 schwarze 74: p->line = 1;
1.23 schwarze 75: p->defrmargin = p->lastrmargin = 78;
1.41 schwarze 76: p->fontq = mandoc_reallocarray(NULL,
1.57 schwarze 77: (p->fontsz = 8), sizeof(*p->fontq));
1.41 schwarze 78: p->fontq[0] = p->fontl = TERMFONT_NONE;
1.5 kristaps 79:
1.1 kristaps 80: p->begin = ascii_begin;
81: p->end = ascii_end;
1.15 kristaps 82: p->hspan = ascii_hspan;
83: p->type = TERMTYPE_CHAR;
84:
85: p->enc = TERMENC_ASCII;
86: p->advance = ascii_advance;
1.1 kristaps 87: p->endline = ascii_endline;
1.7 kristaps 88: p->letter = ascii_letter;
1.23 schwarze 89: p->setwidth = ascii_setwidth;
1.5 kristaps 90: p->width = ascii_width;
1.1 kristaps 91:
1.31 schwarze 92: #if HAVE_WCHAR
1.17 kristaps 93: if (TERMENC_ASCII != enc) {
1.52 schwarze 94:
95: /*
96: * Do not change any of this to LC_ALL. It might break
97: * the formatting by subtly changing the behaviour of
98: * various functions, for example strftime(3). As a
99: * worst case, it might even cause buffer overflows.
100: */
101:
1.17 kristaps 102: v = TERMENC_LOCALE == enc ?
1.52 schwarze 103: setlocale(LC_CTYPE, "") :
1.54 schwarze 104: setlocale(LC_CTYPE, UTF8_LOCALE);
1.60 schwarze 105:
106: /*
107: * We only support UTF-8,
108: * so revert to ASCII for anything else.
109: */
110:
111: if (v != NULL &&
112: strcmp(nl_langinfo(CODESET), "UTF-8") != 0)
113: v = setlocale(LC_CTYPE, "C");
114:
115: if (v != NULL && MB_CUR_MAX > 1) {
1.15 kristaps 116: p->enc = enc;
117: p->advance = locale_advance;
118: p->endline = locale_endline;
119: p->letter = locale_letter;
120: p->width = locale_width;
121: }
1.17 kristaps 122: }
1.15 kristaps 123: #endif
124:
1.44 schwarze 125: if (outopts->mdoc) {
126: p->mdocstyle = 1;
127: p->defindent = 5;
128: }
129: if (outopts->indent)
130: p->defindent = outopts->indent;
131: if (outopts->width)
132: p->defrmargin = outopts->width;
133: if (outopts->synopsisonly)
134: p->synopsisonly = 1;
1.1 kristaps 135:
1.61 schwarze 136: assert(p->defindent < UINT16_MAX);
137: assert(p->defrmargin < UINT16_MAX);
1.49 schwarze 138: return p;
1.5 kristaps 139: }
140:
1.14 kristaps 141: void *
1.51 schwarze 142: ascii_alloc(const struct manoutput *outopts)
1.14 kristaps 143: {
144:
1.51 schwarze 145: return ascii_init(TERMENC_ASCII, outopts);
1.14 kristaps 146: }
1.17 kristaps 147:
148: void *
1.51 schwarze 149: utf8_alloc(const struct manoutput *outopts)
1.17 kristaps 150: {
151:
1.51 schwarze 152: return ascii_init(TERMENC_UTF8, outopts);
1.17 kristaps 153: }
154:
1.14 kristaps 155: void *
1.51 schwarze 156: locale_alloc(const struct manoutput *outopts)
1.14 kristaps 157: {
158:
1.51 schwarze 159: return ascii_init(TERMENC_LOCALE, outopts);
1.23 schwarze 160: }
161:
162: static void
1.45 schwarze 163: ascii_setwidth(struct termp *p, int iop, int width)
1.23 schwarze 164: {
165:
1.45 schwarze 166: width /= 24;
1.57 schwarze 167: p->tcol->rmargin = p->defrmargin;
1.40 schwarze 168: if (iop > 0)
1.24 schwarze 169: p->defrmargin += width;
1.40 schwarze 170: else if (iop == 0)
1.45 schwarze 171: p->defrmargin = width ? (size_t)width : p->lastrmargin;
172: else if (p->defrmargin > (size_t)width)
1.24 schwarze 173: p->defrmargin -= width;
174: else
1.40 schwarze 175: p->defrmargin = 0;
1.61 schwarze 176: if (p->defrmargin > 1000)
177: p->defrmargin = 1000;
1.57 schwarze 178: p->lastrmargin = p->tcol->rmargin;
179: p->tcol->rmargin = p->maxrmargin = p->defrmargin;
1.42 schwarze 180: }
181:
182: void
1.53 schwarze 183: terminal_sepline(void *arg)
1.42 schwarze 184: {
185: struct termp *p;
186: size_t i;
187:
188: p = (struct termp *)arg;
1.53 schwarze 189: (*p->endline)(p);
1.42 schwarze 190: for (i = 0; i < p->defrmargin; i++)
1.53 schwarze 191: (*p->letter)(p, '-');
192: (*p->endline)(p);
193: (*p->endline)(p);
1.34 schwarze 194: }
195:
1.5 kristaps 196: static size_t
1.13 kristaps 197: ascii_width(const struct termp *p, int c)
1.5 kristaps 198: {
1.62 ! schwarze 199: return c != ASCII_BREAK;
1.1 kristaps 200: }
201:
202: void
203: ascii_free(void *arg)
204: {
205:
206: term_free((struct termp *)arg);
207: }
208:
209: static void
1.13 kristaps 210: ascii_letter(struct termp *p, int c)
1.1 kristaps 211: {
1.25 schwarze 212:
1.1 kristaps 213: putchar(c);
214: }
215:
216: static void
217: ascii_begin(struct termp *p)
218: {
219:
220: (*p->headf)(p, p->argf);
221: }
222:
223: static void
224: ascii_end(struct termp *p)
225: {
226:
227: (*p->footf)(p, p->argf);
228: }
229:
230: static void
231: ascii_endline(struct termp *p)
232: {
233:
1.47 schwarze 234: p->line++;
1.57 schwarze 235: p->tcol->offset -= p->ti;
1.56 schwarze 236: p->ti = 0;
1.1 kristaps 237: putchar('\n');
238: }
239:
240: static void
241: ascii_advance(struct termp *p, size_t len)
242: {
1.25 schwarze 243: size_t i;
1.1 kristaps 244:
1.61 schwarze 245: assert(len < UINT16_MAX);
1.1 kristaps 246: for (i = 0; i < len; i++)
247: putchar(' ');
248: }
1.7 kristaps 249:
1.45 schwarze 250: static int
1.7 kristaps 251: ascii_hspan(const struct termp *p, const struct roffsu *su)
252: {
253: double r;
254:
255: switch (su->unit) {
1.29 kristaps 256: case SCALE_BU:
1.45 schwarze 257: r = su->scale;
1.29 kristaps 258: break;
1.25 schwarze 259: case SCALE_CM:
1.45 schwarze 260: r = su->scale * 240.0 / 2.54;
1.29 kristaps 261: break;
262: case SCALE_FS:
1.45 schwarze 263: r = su->scale * 65536.0;
1.7 kristaps 264: break;
1.25 schwarze 265: case SCALE_IN:
1.45 schwarze 266: r = su->scale * 240.0;
1.7 kristaps 267: break;
1.29 kristaps 268: case SCALE_MM:
1.45 schwarze 269: r = su->scale * 0.24;
1.29 kristaps 270: break;
1.45 schwarze 271: case SCALE_VS:
1.25 schwarze 272: case SCALE_PC:
1.45 schwarze 273: r = su->scale * 40.0;
1.7 kristaps 274: break;
1.25 schwarze 275: case SCALE_PT:
1.45 schwarze 276: r = su->scale * 10.0 / 3.0;
1.7 kristaps 277: break;
1.29 kristaps 278: case SCALE_EN:
279: case SCALE_EM:
1.45 schwarze 280: r = su->scale * 24.0;
1.29 kristaps 281: break;
1.32 schwarze 282: default:
1.29 kristaps 283: abort();
1.7 kristaps 284: }
1.49 schwarze 285: return r > 0.0 ? r + 0.01 : r - 0.01;
1.36 schwarze 286: }
287:
288: const char *
289: ascii_uc2str(int uc)
290: {
291: static const char nbrsp[2] = { ASCII_NBRSP, '\0' };
292: static const char *tab[] = {
293: "<NUL>","<SOH>","<STX>","<ETX>","<EOT>","<ENQ>","<ACK>","<BEL>",
294: "<BS>", "\t", "<LF>", "<VT>", "<FF>", "<CR>", "<SO>", "<SI>",
295: "<DLE>","<DC1>","<DC2>","<DC3>","<DC4>","<NAK>","<SYN>","<ETB>",
296: "<CAN>","<EM>", "<SUB>","<ESC>","<FS>", "<GS>", "<RS>", "<US>",
297: " ", "!", "\"", "#", "$", "%", "&", "'",
298: "(", ")", "*", "+", ",", "-", ".", "/",
299: "0", "1", "2", "3", "4", "5", "6", "7",
300: "8", "9", ":", ";", "<", "=", ">", "?",
301: "@", "A", "B", "C", "D", "E", "F", "G",
302: "H", "I", "J", "K", "L", "M", "N", "O",
303: "P", "Q", "R", "S", "T", "U", "V", "W",
304: "X", "Y", "Z", "[", "\\", "]", "^", "_",
305: "`", "a", "b", "c", "d", "e", "f", "g",
306: "h", "i", "j", "k", "l", "m", "n", "o",
307: "p", "q", "r", "s", "t", "u", "v", "w",
308: "x", "y", "z", "{", "|", "}", "~", "<DEL>",
309: "<80>", "<81>", "<82>", "<83>", "<84>", "<85>", "<86>", "<87>",
310: "<88>", "<89>", "<8A>", "<8B>", "<8C>", "<8D>", "<8E>", "<8F>",
311: "<90>", "<91>", "<92>", "<93>", "<94>", "<95>", "<96>", "<97>",
1.58 schwarze 312: "<98>", "<99>", "<9A>", "<9B>", "<9C>", "<9D>", "<9E>", "<9F>",
1.59 schwarze 313: nbrsp, "!", "/\bc", "GBP", "o\bx", "=\bY", "|", "<section>",
1.37 schwarze 314: "\"", "(C)", "_\ba", "<<", "~", "", "(R)", "-",
1.59 schwarze 315: "<degree>","+-","^2", "^3", "'","<micro>","<paragraph>",".",
316: ",", "^1", "_\bo", ">>", "1/4", "1/2", "3/4", "?",
1.37 schwarze 317: "`\bA", "'\bA", "^\bA", "~\bA", "\"\bA","o\bA", "AE", ",\bC",
318: "`\bE", "'\bE", "^\bE", "\"\bE","`\bI", "'\bI", "^\bI", "\"\bI",
1.59 schwarze 319: "Dh", "~\bN", "`\bO", "'\bO", "^\bO", "~\bO", "\"\bO","x",
1.37 schwarze 320: "/\bO", "`\bU", "'\bU", "^\bU", "\"\bU","'\bY", "Th", "ss",
321: "`\ba", "'\ba", "^\ba", "~\ba", "\"\ba","o\ba", "ae", ",\bc",
322: "`\be", "'\be", "^\be", "\"\be","`\bi", "'\bi", "^\bi", "\"\bi",
1.59 schwarze 323: "dh", "~\bn", "`\bo", "'\bo", "^\bo", "~\bo", "\"\bo","/",
1.37 schwarze 324: "/\bo", "`\bu", "'\bu", "^\bu", "\"\bu","'\by", "th", "\"\by",
325: "A", "a", "A", "a", "A", "a", "'\bC", "'\bc",
326: "^\bC", "^\bc", "C", "c", "C", "c", "D", "d",
327: "/\bD", "/\bd", "E", "e", "E", "e", "E", "e",
328: "E", "e", "E", "e", "^\bG", "^\bg", "G", "g",
329: "G", "g", ",\bG", ",\bg", "^\bH", "^\bh", "/\bH", "/\bh",
330: "~\bI", "~\bi", "I", "i", "I", "i", "I", "i",
331: "I", "i", "IJ", "ij", "^\bJ", "^\bj", ",\bK", ",\bk",
332: "q", "'\bL", "'\bl", ",\bL", ",\bl", "L", "l", "L",
333: "l", "/\bL", "/\bl", "'\bN", "'\bn", ",\bN", ",\bn", "N",
1.36 schwarze 334: "n", "'n", "Ng", "ng", "O", "o", "O", "o",
1.37 schwarze 335: "O", "o", "OE", "oe", "'\bR", "'\br", ",\bR", ",\br",
336: "R", "r", "'\bS", "'\bs", "^\bS", "^\bs", ",\bS", ",\bs",
337: "S", "s", ",\bT", ",\bt", "T", "t", "/\bT", "/\bt",
338: "~\bU", "~\bu", "U", "u", "U", "u", "U", "u",
339: "U", "u", "U", "u", "^\bW", "^\bw", "^\bY", "^\by",
340: "\"\bY","'\bZ", "'\bz", "Z", "z", "Z", "z", "s",
1.36 schwarze 341: "b", "B", "B", "b", "6", "6", "O", "C",
342: "c", "D", "D", "D", "d", "d", "3", "@",
1.37 schwarze 343: "E", "F", ",\bf", "G", "G", "hv", "I", "/\bI",
344: "K", "k", "/\bl", "l", "W", "N", "n", "~\bO",
1.36 schwarze 345: "O", "o", "OI", "oi", "P", "p", "YR", "2",
346: "2", "SH", "sh", "t", "T", "t", "T", "U",
1.37 schwarze 347: "u", "Y", "V", "Y", "y", "/\bZ", "/\bz", "ZH",
348: "ZH", "zh", "zh", "/\b2", "5", "5", "ts", "w",
1.36 schwarze 349: "|", "||", "|=", "!", "DZ", "Dz", "dz", "LJ",
350: "Lj", "lj", "NJ", "Nj", "nj", "A", "a", "I",
351: "i", "O", "o", "U", "u", "U", "u", "U",
352: "u", "U", "u", "U", "u", "@", "A", "a",
1.37 schwarze 353: "A", "a", "AE", "ae", "/\bG", "/\bg", "G", "g",
1.36 schwarze 354: "K", "k", "O", "o", "O", "o", "ZH", "zh",
1.37 schwarze 355: "j", "DZ", "Dz", "dz", "'\bG", "'\bg", "HV", "W",
356: "`\bN", "`\bn", "A", "a", "'\bAE","'\bae","O", "o"};
1.36 schwarze 357:
1.39 schwarze 358: assert(uc >= 0);
1.36 schwarze 359: if ((size_t)uc < sizeof(tab)/sizeof(tab[0]))
1.49 schwarze 360: return tab[uc];
361: return mchars_uc2str(uc);
1.7 kristaps 362: }
363:
1.31 schwarze 364: #if HAVE_WCHAR
1.15 kristaps 365: static size_t
366: locale_width(const struct termp *p, int c)
367: {
368: int rc;
369:
1.26 schwarze 370: if (c == ASCII_NBRSP)
371: c = ' ';
372: rc = wcwidth(c);
373: if (rc < 0)
374: rc = 0;
1.49 schwarze 375: return rc;
1.15 kristaps 376: }
377:
378: static void
379: locale_advance(struct termp *p, size_t len)
380: {
1.25 schwarze 381: size_t i;
1.15 kristaps 382:
1.61 schwarze 383: assert(len < UINT16_MAX);
1.15 kristaps 384: for (i = 0; i < len; i++)
385: putwchar(L' ');
386: }
387:
388: static void
389: locale_endline(struct termp *p)
390: {
391:
1.47 schwarze 392: p->line++;
1.57 schwarze 393: p->tcol->offset -= p->ti;
1.56 schwarze 394: p->ti = 0;
1.15 kristaps 395: putwchar(L'\n');
396: }
397:
398: static void
399: locale_letter(struct termp *p, int c)
400: {
1.25 schwarze 401:
1.15 kristaps 402: putwchar(c);
403: }
404: #endif
CVSweb