1 /* $OpenBSD: term_ascii.c,v 1.27 2014/11/20 13:55:23 schwarze Exp $ */ 2 /* 3 * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> 4 * Copyright (c) 2014 Ingo Schwarze <schwarze@openbsd.org> 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 * 10 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES 11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF 12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR 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 <sys/types.h> 19 20 #include <assert.h> 21 #include <locale.h> 22 #include <stdint.h> 23 #include <stdio.h> 24 #include <stdlib.h> 25 #include <unistd.h> 26 #include <wchar.h> 27 28 #include "mandoc.h" 29 #include "mandoc_aux.h" 30 #include "out.h" 31 #include "term.h" 32 #include "main.h" 33 34 static struct termp *ascii_init(enum termenc, 35 const struct mchars *, char *); 36 static double ascii_hspan(const struct termp *, 37 const struct roffsu *); 38 static size_t ascii_width(const struct termp *, int); 39 static void ascii_advance(struct termp *, size_t); 40 static void ascii_begin(struct termp *); 41 static void ascii_end(struct termp *); 42 static void ascii_endline(struct termp *); 43 static void ascii_letter(struct termp *, int); 44 static void ascii_setwidth(struct termp *, int, size_t); 45 46 static void locale_advance(struct termp *, size_t); 47 static void locale_endline(struct termp *); 48 static void locale_letter(struct termp *, int); 49 static size_t locale_width(const struct termp *, int); 50 51 52 static struct termp * 53 ascii_init(enum termenc enc, const struct mchars *mchars, char *outopts) 54 { 55 const char *toks[5]; 56 char *v; 57 struct termp *p; 58 59 p = mandoc_calloc(1, sizeof(struct termp)); 60 61 p->symtab = mchars; 62 p->tabwidth = 5; 63 p->defrmargin = p->lastrmargin = 78; 64 65 p->begin = ascii_begin; 66 p->end = ascii_end; 67 p->hspan = ascii_hspan; 68 p->type = TERMTYPE_CHAR; 69 70 p->enc = TERMENC_ASCII; 71 p->advance = ascii_advance; 72 p->endline = ascii_endline; 73 p->letter = ascii_letter; 74 p->setwidth = ascii_setwidth; 75 p->width = ascii_width; 76 77 if (TERMENC_ASCII != enc) { 78 v = TERMENC_LOCALE == enc ? 79 setlocale(LC_ALL, "") : 80 setlocale(LC_CTYPE, "en_US.UTF-8"); 81 if (NULL != v && MB_CUR_MAX > 1) { 82 p->enc = enc; 83 p->advance = locale_advance; 84 p->endline = locale_endline; 85 p->letter = locale_letter; 86 p->width = locale_width; 87 } 88 } 89 90 toks[0] = "indent"; 91 toks[1] = "width"; 92 toks[2] = "mdoc"; 93 toks[3] = "synopsis"; 94 toks[4] = NULL; 95 96 while (outopts && *outopts) 97 switch (getsubopt(&outopts, UNCONST(toks), &v)) { 98 case 0: 99 p->defindent = (size_t)atoi(v); 100 break; 101 case 1: 102 p->defrmargin = (size_t)atoi(v); 103 break; 104 case 2: 105 /* 106 * Temporary, undocumented mode 107 * to imitate mdoc(7) output style. 108 */ 109 p->mdocstyle = 1; 110 p->defindent = 5; 111 break; 112 case 3: 113 p->synopsisonly = 1; 114 break; 115 default: 116 break; 117 } 118 119 /* Enforce a lower boundary. */ 120 if (p->defrmargin < 58) 121 p->defrmargin = 58; 122 123 return(p); 124 } 125 126 void * 127 ascii_alloc(const struct mchars *mchars, char *outopts) 128 { 129 130 return(ascii_init(TERMENC_ASCII, mchars, outopts)); 131 } 132 133 void * 134 utf8_alloc(const struct mchars *mchars, char *outopts) 135 { 136 137 return(ascii_init(TERMENC_UTF8, mchars, outopts)); 138 } 139 140 void * 141 locale_alloc(const struct mchars *mchars, char *outopts) 142 { 143 144 return(ascii_init(TERMENC_LOCALE, mchars, outopts)); 145 } 146 147 static void 148 ascii_setwidth(struct termp *p, int iop, size_t width) 149 { 150 151 p->rmargin = p->defrmargin; 152 if (iop > 0) 153 p->defrmargin += width; 154 else if (iop == 0) 155 p->defrmargin = width ? width : p->lastrmargin; 156 else if (p->defrmargin > width) 157 p->defrmargin -= width; 158 else 159 p->defrmargin = 0; 160 p->lastrmargin = p->rmargin; 161 p->rmargin = p->maxrmargin = p->defrmargin; 162 } 163 164 static size_t 165 ascii_width(const struct termp *p, int c) 166 { 167 168 return(1); 169 } 170 171 void 172 ascii_free(void *arg) 173 { 174 175 term_free((struct termp *)arg); 176 } 177 178 static void 179 ascii_letter(struct termp *p, int c) 180 { 181 182 putchar(c); 183 } 184 185 static void 186 ascii_begin(struct termp *p) 187 { 188 189 (*p->headf)(p, p->argf); 190 } 191 192 static void 193 ascii_end(struct termp *p) 194 { 195 196 (*p->footf)(p, p->argf); 197 } 198 199 static void 200 ascii_endline(struct termp *p) 201 { 202 203 putchar('\n'); 204 } 205 206 static void 207 ascii_advance(struct termp *p, size_t len) 208 { 209 size_t i; 210 211 for (i = 0; i < len; i++) 212 putchar(' '); 213 } 214 215 static double 216 ascii_hspan(const struct termp *p, const struct roffsu *su) 217 { 218 double r; 219 220 /* 221 * Approximate based on character width. 222 * None of these will be actually correct given that an inch on 223 * the screen depends on character size, terminal, etc., etc. 224 */ 225 switch (su->unit) { 226 case SCALE_BU: 227 r = su->scale * 10.0 / 240.0; 228 break; 229 case SCALE_CM: 230 r = su->scale * 10.0 / 2.54; 231 break; 232 case SCALE_FS: 233 r = su->scale * 2730.666; 234 break; 235 case SCALE_IN: 236 r = su->scale * 10.0; 237 break; 238 case SCALE_MM: 239 r = su->scale / 100.0; 240 break; 241 case SCALE_PC: 242 r = su->scale * 10.0 / 6.0; 243 break; 244 case SCALE_PT: 245 r = su->scale * 10.0 / 72.0; 246 break; 247 case SCALE_VS: 248 r = su->scale * 2.0 - 1.0; 249 break; 250 case SCALE_EN: 251 /* FALLTHROUGH */ 252 case SCALE_EM: 253 r = su->scale; 254 break; 255 default: 256 abort(); 257 /* NOTREACHED */ 258 } 259 260 return(r); 261 } 262 263 const char * 264 ascii_uc2str(int uc) 265 { 266 static const char nbrsp[2] = { ASCII_NBRSP, '\0' }; 267 static const char *tab[] = { 268 "<NUL>","<SOH>","<STX>","<ETX>","<EOT>","<ENQ>","<ACK>","<BEL>", 269 "<BS>", "\t", "<LF>", "<VT>", "<FF>", "<CR>", "<SO>", "<SI>", 270 "<DLE>","<DC1>","<DC2>","<DC3>","<DC4>","<NAK>","<SYN>","<ETB>", 271 "<CAN>","<EM>", "<SUB>","<ESC>","<FS>", "<GS>", "<RS>", "<US>", 272 " ", "!", "\"", "#", "$", "%", "&", "'", 273 "(", ")", "*", "+", ",", "-", ".", "/", 274 "0", "1", "2", "3", "4", "5", "6", "7", 275 "8", "9", ":", ";", "<", "=", ">", "?", 276 "@", "A", "B", "C", "D", "E", "F", "G", 277 "H", "I", "J", "K", "L", "M", "N", "O", 278 "P", "Q", "R", "S", "T", "U", "V", "W", 279 "X", "Y", "Z", "[", "\\", "]", "^", "_", 280 "`", "a", "b", "c", "d", "e", "f", "g", 281 "h", "i", "j", "k", "l", "m", "n", "o", 282 "p", "q", "r", "s", "t", "u", "v", "w", 283 "x", "y", "z", "{", "|", "}", "~", "<DEL>", 284 "<80>", "<81>", "<82>", "<83>", "<84>", "<85>", "<86>", "<87>", 285 "<88>", "<89>", "<8A>", "<8B>", "<8C>", "<8D>", "<8E>", "<8F>", 286 "<90>", "<91>", "<92>", "<93>", "<94>", "<95>", "<96>", "<97>", 287 "<99>", "<99>", "<9A>", "<9B>", "<9C>", "<9D>", "<9E>", "<9F>", 288 nbrsp, "!", "/\bc", "GBP", "o\bx", "=\bY", "|", "<sec>", 289 "\"", "(C)", "_\ba", "<<", "~", "", "(R)", "-", 290 "<deg>","+-", "2", "3", "'", ",\bu", "<par>",".", 291 ",", "1", "_\bo", ">>", "1/4", "1/2", "3/4", "?", 292 "`\bA", "'\bA", "^\bA", "~\bA", "\"\bA","o\bA", "AE", ",\bC", 293 "`\bE", "'\bE", "^\bE", "\"\bE","`\bI", "'\bI", "^\bI", "\"\bI", 294 "-\bD", "~\bN", "`\bO", "'\bO", "^\bO", "~\bO", "\"\bO","x", 295 "/\bO", "`\bU", "'\bU", "^\bU", "\"\bU","'\bY", "Th", "ss", 296 "`\ba", "'\ba", "^\ba", "~\ba", "\"\ba","o\ba", "ae", ",\bc", 297 "`\be", "'\be", "^\be", "\"\be","`\bi", "'\bi", "^\bi", "\"\bi", 298 "d", "~\bn", "`\bo", "'\bo", "^\bo", "~\bo", "\"\bo","-:-", 299 "/\bo", "`\bu", "'\bu", "^\bu", "\"\bu","'\by", "th", "\"\by", 300 "A", "a", "A", "a", "A", "a", "'\bC", "'\bc", 301 "^\bC", "^\bc", "C", "c", "C", "c", "D", "d", 302 "/\bD", "/\bd", "E", "e", "E", "e", "E", "e", 303 "E", "e", "E", "e", "^\bG", "^\bg", "G", "g", 304 "G", "g", ",\bG", ",\bg", "^\bH", "^\bh", "/\bH", "/\bh", 305 "~\bI", "~\bi", "I", "i", "I", "i", "I", "i", 306 "I", "i", "IJ", "ij", "^\bJ", "^\bj", ",\bK", ",\bk", 307 "q", "'\bL", "'\bl", ",\bL", ",\bl", "L", "l", "L", 308 "l", "/\bL", "/\bl", "'\bN", "'\bn", ",\bN", ",\bn", "N", 309 "n", "'n", "Ng", "ng", "O", "o", "O", "o", 310 "O", "o", "OE", "oe", "'\bR", "'\br", ",\bR", ",\br", 311 "R", "r", "'\bS", "'\bs", "^\bS", "^\bs", ",\bS", ",\bs", 312 "S", "s", ",\bT", ",\bt", "T", "t", "/\bT", "/\bt", 313 "~\bU", "~\bu", "U", "u", "U", "u", "U", "u", 314 "U", "u", "U", "u", "^\bW", "^\bw", "^\bY", "^\by", 315 "\"\bY","'\bZ", "'\bz", "Z", "z", "Z", "z", "s", 316 "b", "B", "B", "b", "6", "6", "O", "C", 317 "c", "D", "D", "D", "d", "d", "3", "@", 318 "E", "F", ",\bf", "G", "G", "hv", "I", "/\bI", 319 "K", "k", "/\bl", "l", "W", "N", "n", "~\bO", 320 "O", "o", "OI", "oi", "P", "p", "YR", "2", 321 "2", "SH", "sh", "t", "T", "t", "T", "U", 322 "u", "Y", "V", "Y", "y", "/\bZ", "/\bz", "ZH", 323 "ZH", "zh", "zh", "/\b2", "5", "5", "ts", "w", 324 "|", "||", "|=", "!", "DZ", "Dz", "dz", "LJ", 325 "Lj", "lj", "NJ", "Nj", "nj", "A", "a", "I", 326 "i", "O", "o", "U", "u", "U", "u", "U", 327 "u", "U", "u", "U", "u", "@", "A", "a", 328 "A", "a", "AE", "ae", "/\bG", "/\bg", "G", "g", 329 "K", "k", "O", "o", "O", "o", "ZH", "zh", 330 "j", "DZ", "Dz", "dz", "'\bG", "'\bg", "HV", "W", 331 "`\bN", "`\bn", "A", "a", "'\bAE","'\bae","O", "o"}; 332 333 assert(uc >= 0); 334 if ((size_t)uc < sizeof(tab)/sizeof(tab[0])) 335 return(tab[uc]); 336 return(mchars_uc2str(uc)); 337 } 338 339 static size_t 340 locale_width(const struct termp *p, int c) 341 { 342 int rc; 343 344 if (c == ASCII_NBRSP) 345 c = ' '; 346 rc = wcwidth(c); 347 if (rc < 0) 348 rc = 0; 349 return(rc); 350 } 351 352 static void 353 locale_advance(struct termp *p, size_t len) 354 { 355 size_t i; 356 357 for (i = 0; i < len; i++) 358 putwchar(L' '); 359 } 360 361 static void 362 locale_endline(struct termp *p) 363 { 364 365 putwchar(L'\n'); 366 } 367 368 static void 369 locale_letter(struct termp *p, int c) 370 { 371 372 putwchar(c); 373 } 374