| 1 | /* |
| 2 | * man page backend for Halibut |
| 3 | */ |
| 4 | |
| 5 | #include <stdio.h> |
| 6 | #include <stdlib.h> |
| 7 | #include <assert.h> |
| 8 | #include "halibut.h" |
| 9 | |
| 10 | static void man_text(FILE *, word *, int newline, int quote_props); |
| 11 | static void man_codepara(FILE *, word *); |
| 12 | static int man_convert(wchar_t *s, int maxlen, |
| 13 | char **result, int quote_props); |
| 14 | |
| 15 | typedef struct { |
| 16 | wchar_t *th; |
| 17 | int headnumbers; |
| 18 | int mindepth; |
| 19 | } manconfig; |
| 20 | |
| 21 | static manconfig man_configure(paragraph *source) { |
| 22 | manconfig ret; |
| 23 | |
| 24 | /* |
| 25 | * Defaults. |
| 26 | */ |
| 27 | ret.th = NULL; |
| 28 | ret.headnumbers = FALSE; |
| 29 | ret.mindepth = 0; |
| 30 | |
| 31 | for (; source; source = source->next) { |
| 32 | if (source->type == para_Config) { |
| 33 | if (!ustricmp(source->keyword, L"man-identity")) { |
| 34 | wchar_t *wp, *ep; |
| 35 | |
| 36 | wp = uadv(source->keyword); |
| 37 | ep = wp; |
| 38 | while (*ep) |
| 39 | ep = uadv(ep); |
| 40 | ret.th = mknewa(wchar_t, ep - wp + 1); |
| 41 | memcpy(ret.th, wp, (ep - wp + 1) * sizeof(wchar_t)); |
| 42 | } else if (!ustricmp(source->keyword, L"man-headnumbers")) { |
| 43 | ret.headnumbers = utob(uadv(source->keyword)); |
| 44 | } else if (!ustricmp(source->keyword, L"man-mindepth")) { |
| 45 | ret.mindepth = utoi(uadv(source->keyword)); |
| 46 | } |
| 47 | } |
| 48 | } |
| 49 | |
| 50 | return ret; |
| 51 | } |
| 52 | |
| 53 | static void man_conf_cleanup(manconfig cf) |
| 54 | { |
| 55 | sfree(cf.th); |
| 56 | } |
| 57 | |
| 58 | #define QUOTE_INITCTRL 1 /* quote initial . and ' on a line */ |
| 59 | #define QUOTE_QUOTES 2 /* quote double quotes by doubling them */ |
| 60 | |
| 61 | void man_backend(paragraph *sourceform, keywordlist *keywords, |
| 62 | indexdata *idx) { |
| 63 | paragraph *p; |
| 64 | FILE *fp; |
| 65 | char const *sep; |
| 66 | manconfig conf; |
| 67 | |
| 68 | IGNORE(keywords); /* we don't happen to need this */ |
| 69 | IGNORE(idx); /* or this */ |
| 70 | |
| 71 | conf = man_configure(sourceform); |
| 72 | |
| 73 | /* |
| 74 | * Determine the output file name, and open the output file |
| 75 | * |
| 76 | * FIXME: want configurable output file names here. For the |
| 77 | * moment, we'll just call it `output.1'. |
| 78 | */ |
| 79 | fp = fopen("output.1", "w"); |
| 80 | if (!fp) { |
| 81 | error(err_cantopenw, "output.1"); |
| 82 | return; |
| 83 | } |
| 84 | |
| 85 | /* Do the version ID */ |
| 86 | for (p = sourceform; p; p = p->next) |
| 87 | if (p->type == para_VersionID) { |
| 88 | fprintf(fp, ".\\\" "); |
| 89 | man_text(fp, p->words, TRUE, 0); |
| 90 | } |
| 91 | |
| 92 | /* .TH name-of-program manual-section */ |
| 93 | fprintf(fp, ".TH"); |
| 94 | if (conf.th && *conf.th) { |
| 95 | char *c; |
| 96 | wchar_t *wp; |
| 97 | |
| 98 | for (wp = conf.th; *wp; wp = uadv(wp)) { |
| 99 | fputs(" \"", fp); |
| 100 | man_convert(wp, 0, &c, QUOTE_QUOTES); |
| 101 | fputs(c, fp); |
| 102 | sfree(c); |
| 103 | fputc('"', fp); |
| 104 | } |
| 105 | } |
| 106 | fputc('\n', fp); |
| 107 | |
| 108 | fprintf(fp, ".UC\n"); |
| 109 | |
| 110 | /* Do the preamble and copyright */ |
| 111 | sep = ""; |
| 112 | for (p = sourceform; p; p = p->next) |
| 113 | if (p->type == para_Preamble) { |
| 114 | fprintf(fp, "%s", sep); |
| 115 | man_text(fp, p->words, TRUE, 0); |
| 116 | sep = "\n"; |
| 117 | } |
| 118 | for (p = sourceform; p; p = p->next) |
| 119 | if (p->type == para_Copyright) { |
| 120 | fprintf(fp, "%s", sep); |
| 121 | man_text(fp, p->words, TRUE, 0); |
| 122 | sep = "\n"; |
| 123 | } |
| 124 | |
| 125 | for (p = sourceform; p; p = p->next) switch (p->type) { |
| 126 | /* |
| 127 | * Things we ignore because we've already processed them or |
| 128 | * aren't going to touch them in this pass. |
| 129 | */ |
| 130 | case para_IM: |
| 131 | case para_BR: |
| 132 | case para_Biblio: /* only touch BiblioCited */ |
| 133 | case para_VersionID: |
| 134 | case para_Copyright: |
| 135 | case para_Preamble: |
| 136 | case para_NoCite: |
| 137 | case para_Title: |
| 138 | break; |
| 139 | |
| 140 | /* |
| 141 | * Headings. |
| 142 | */ |
| 143 | case para_Chapter: |
| 144 | case para_Appendix: |
| 145 | case para_UnnumberedChapter: |
| 146 | case para_Heading: |
| 147 | case para_Subsect: |
| 148 | { |
| 149 | int depth; |
| 150 | if (p->type == para_Subsect) |
| 151 | depth = p->aux + 2; |
| 152 | else if (p->type == para_Heading) |
| 153 | depth = 1; |
| 154 | else |
| 155 | depth = 0; |
| 156 | if (depth >= conf.mindepth) { |
| 157 | fprintf(fp, ".SH \""); |
| 158 | if (conf.headnumbers && p->kwtext) { |
| 159 | man_text(fp, p->kwtext, FALSE, QUOTE_QUOTES); |
| 160 | fprintf(fp, " "); |
| 161 | } |
| 162 | man_text(fp, p->words, FALSE, QUOTE_QUOTES); |
| 163 | fprintf(fp, "\"\n"); |
| 164 | } |
| 165 | break; |
| 166 | } |
| 167 | |
| 168 | /* |
| 169 | * Code paragraphs. |
| 170 | */ |
| 171 | case para_Code: |
| 172 | fprintf(fp, ".PP\n"); |
| 173 | man_codepara(fp, p->words); |
| 174 | break; |
| 175 | |
| 176 | /* |
| 177 | * Normal paragraphs. |
| 178 | */ |
| 179 | case para_Normal: |
| 180 | fprintf(fp, ".PP\n"); |
| 181 | man_text(fp, p->words, TRUE, 0); |
| 182 | break; |
| 183 | |
| 184 | /* |
| 185 | * List paragraphs. |
| 186 | */ |
| 187 | case para_Description: |
| 188 | case para_BiblioCited: |
| 189 | case para_Bullet: |
| 190 | case para_NumberedList: |
| 191 | if (p->type == para_Bullet) { |
| 192 | fprintf(fp, ".IP \"\\fBo\\fP\"\n"); /* FIXME: configurable? */ |
| 193 | } else if (p->type == para_NumberedList) { |
| 194 | fprintf(fp, ".IP \""); |
| 195 | man_text(fp, p->kwtext, FALSE, QUOTE_QUOTES); |
| 196 | fprintf(fp, "\"\n"); |
| 197 | } else if (p->type == para_Description) { |
| 198 | /* |
| 199 | * Do nothing; the .xP for this paragraph is the .IP |
| 200 | * which has come before it in the DescribedThing. |
| 201 | */ |
| 202 | } else if (p->type == para_BiblioCited) { |
| 203 | fprintf(fp, ".IP \""); |
| 204 | man_text(fp, p->kwtext, FALSE, QUOTE_QUOTES); |
| 205 | fprintf(fp, "\"\n"); |
| 206 | } |
| 207 | man_text(fp, p->words, TRUE, 0); |
| 208 | break; |
| 209 | |
| 210 | case para_DescribedThing: |
| 211 | fprintf(fp, ".IP \""); |
| 212 | man_text(fp, p->words, FALSE, QUOTE_QUOTES); |
| 213 | fprintf(fp, "\"\n"); |
| 214 | break; |
| 215 | |
| 216 | case para_Rule: |
| 217 | /* |
| 218 | * This isn't terribly good. Anyone who wants to do better |
| 219 | * should feel free! |
| 220 | */ |
| 221 | fprintf(fp, ".PP\n----------------------------------------\n"); |
| 222 | break; |
| 223 | |
| 224 | case para_LcontPush: |
| 225 | case para_QuotePush: |
| 226 | fprintf(fp, ".RS\n"); |
| 227 | break; |
| 228 | case para_LcontPop: |
| 229 | case para_QuotePop: |
| 230 | fprintf(fp, ".RE\n"); |
| 231 | break; |
| 232 | } |
| 233 | |
| 234 | /* |
| 235 | * Tidy up. |
| 236 | */ |
| 237 | fclose(fp); |
| 238 | man_conf_cleanup(conf); |
| 239 | } |
| 240 | |
| 241 | /* |
| 242 | * Convert a wide string into a string of chars. If `result' is |
| 243 | * non-NULL, mallocs the resulting string and stores a pointer to |
| 244 | * it in `*result'. If `result' is NULL, merely checks whether all |
| 245 | * characters in the string are feasible for the output character |
| 246 | * set. |
| 247 | * |
| 248 | * Return is nonzero if all characters are OK. If not all |
| 249 | * characters are OK but `result' is non-NULL, a result _will_ |
| 250 | * still be generated! |
| 251 | * |
| 252 | * FIXME: Here is probably also a good place to do escaping sorts |
| 253 | * of things. I know I at least need to escape backslash, and full |
| 254 | * stops at the starts of words are probably trouble as well. |
| 255 | */ |
| 256 | static int man_convert(wchar_t *s, int maxlen, |
| 257 | char **result, int quote_props) { |
| 258 | /* |
| 259 | * FIXME. Currently this is ISO8859-1 only. |
| 260 | */ |
| 261 | int doing = (result != 0); |
| 262 | int ok = TRUE; |
| 263 | char *p = NULL; |
| 264 | int plen = 0, psize = 0; |
| 265 | |
| 266 | if (maxlen <= 0) |
| 267 | maxlen = -1; |
| 268 | |
| 269 | for (; *s && maxlen != 0; s++, maxlen--) { |
| 270 | wchar_t c = *s; |
| 271 | char outc; |
| 272 | |
| 273 | if ((c >= 32 && c <= 126) || |
| 274 | (c >= 160 && c <= 255)) { |
| 275 | /* Char is OK. */ |
| 276 | outc = (char)c; |
| 277 | } else { |
| 278 | /* Char is not OK. */ |
| 279 | ok = FALSE; |
| 280 | outc = 0xBF; /* approximate the good old DEC `uh?' */ |
| 281 | } |
| 282 | if (doing) { |
| 283 | if (plen+3 >= psize) { |
| 284 | psize = plen + 256; |
| 285 | p = resize(p, psize); |
| 286 | } |
| 287 | if (plen == 0 && (outc == '.' || outc == '\'') && |
| 288 | (quote_props & QUOTE_INITCTRL)) { |
| 289 | /* |
| 290 | * Control character (. or ') at the start of a |
| 291 | * line. Quote it by putting \& (troff zero-width |
| 292 | * space) before it. |
| 293 | */ |
| 294 | p[plen++] = '\\'; |
| 295 | p[plen++] = '&'; |
| 296 | } else if (outc == '\\') { |
| 297 | /* |
| 298 | * Quote backslashes by doubling them, always. |
| 299 | */ |
| 300 | p[plen++] = '\\'; |
| 301 | } else if (outc == '"' && (quote_props & QUOTE_QUOTES)) { |
| 302 | /* |
| 303 | * Double quote within double quotes. Quote it by |
| 304 | * doubling. |
| 305 | */ |
| 306 | p[plen++] = '"'; |
| 307 | } |
| 308 | p[plen++] = outc; |
| 309 | } |
| 310 | } |
| 311 | if (doing) { |
| 312 | p = resize(p, plen+1); |
| 313 | p[plen] = '\0'; |
| 314 | *result = p; |
| 315 | } |
| 316 | return ok; |
| 317 | } |
| 318 | |
| 319 | static void man_rdaddwc(rdstringc *rs, word *text, word *end, |
| 320 | int quote_props) { |
| 321 | char *c; |
| 322 | |
| 323 | for (; text && text != end; text = text->next) switch (text->type) { |
| 324 | case word_HyperLink: |
| 325 | case word_HyperEnd: |
| 326 | case word_UpperXref: |
| 327 | case word_LowerXref: |
| 328 | case word_XrefEnd: |
| 329 | case word_IndexRef: |
| 330 | break; |
| 331 | |
| 332 | case word_Normal: |
| 333 | case word_Emph: |
| 334 | case word_Code: |
| 335 | case word_WeakCode: |
| 336 | case word_WhiteSpace: |
| 337 | case word_EmphSpace: |
| 338 | case word_CodeSpace: |
| 339 | case word_WkCodeSpace: |
| 340 | case word_Quote: |
| 341 | case word_EmphQuote: |
| 342 | case word_CodeQuote: |
| 343 | case word_WkCodeQuote: |
| 344 | assert(text->type != word_CodeQuote && |
| 345 | text->type != word_WkCodeQuote); |
| 346 | if (towordstyle(text->type) == word_Emph && |
| 347 | (attraux(text->aux) == attr_First || |
| 348 | attraux(text->aux) == attr_Only)) |
| 349 | rdaddsc(rs, "\\fI"); |
| 350 | else if ((towordstyle(text->type) == word_Code || |
| 351 | towordstyle(text->type) == word_WeakCode) && |
| 352 | (attraux(text->aux) == attr_First || |
| 353 | attraux(text->aux) == attr_Only)) |
| 354 | rdaddsc(rs, "\\fB"); |
| 355 | if (removeattr(text->type) == word_Normal) { |
| 356 | if (rs->pos > 0) |
| 357 | quote_props &= ~QUOTE_INITCTRL; /* not at start any more */ |
| 358 | if (man_convert(text->text, 0, &c, quote_props)) |
| 359 | rdaddsc(rs, c); |
| 360 | else |
| 361 | man_rdaddwc(rs, text->alt, NULL, quote_props); |
| 362 | sfree(c); |
| 363 | } else if (removeattr(text->type) == word_WhiteSpace) { |
| 364 | rdaddc(rs, ' '); |
| 365 | } else if (removeattr(text->type) == word_Quote) { |
| 366 | rdaddc(rs, '"'); |
| 367 | if (quote_props & QUOTE_QUOTES) |
| 368 | rdaddc(rs, '"'); |
| 369 | } |
| 370 | if (towordstyle(text->type) == word_Emph && |
| 371 | (attraux(text->aux) == attr_Last || |
| 372 | attraux(text->aux) == attr_Only)) |
| 373 | rdaddsc(rs, "\\fP"); |
| 374 | else if ((towordstyle(text->type) == word_Code || |
| 375 | towordstyle(text->type) == word_WeakCode) && |
| 376 | (attraux(text->aux) == attr_Last || |
| 377 | attraux(text->aux) == attr_Only)) |
| 378 | rdaddsc(rs, "\\fP"); |
| 379 | break; |
| 380 | } |
| 381 | } |
| 382 | |
| 383 | static void man_text(FILE *fp, word *text, int newline, int quote_props) { |
| 384 | rdstringc t = { 0, 0, NULL }; |
| 385 | |
| 386 | man_rdaddwc(&t, text, NULL, quote_props | QUOTE_INITCTRL); |
| 387 | fprintf(fp, "%s", t.text); |
| 388 | sfree(t.text); |
| 389 | if (newline) |
| 390 | fputc('\n', fp); |
| 391 | } |
| 392 | |
| 393 | static void man_codepara(FILE *fp, word *text) { |
| 394 | fprintf(fp, ".nf\n"); |
| 395 | for (; text; text = text->next) if (text->type == word_WeakCode) { |
| 396 | char *c; |
| 397 | wchar_t *t, *e; |
| 398 | int quote_props = QUOTE_INITCTRL; |
| 399 | |
| 400 | t = text->text; |
| 401 | if (text->next && text->next->type == word_Emph) { |
| 402 | e = text->next->text; |
| 403 | text = text->next; |
| 404 | } else |
| 405 | e = NULL; |
| 406 | |
| 407 | while (e && *e && *t) { |
| 408 | int n; |
| 409 | int ec = *e; |
| 410 | |
| 411 | for (n = 0; t[n] && e[n] && e[n] == ec; n++); |
| 412 | if (ec == 'i') |
| 413 | fprintf(fp, "\\fI"); |
| 414 | else if (ec == 'b') |
| 415 | fprintf(fp, "\\fB"); |
| 416 | man_convert(t, n, &c, quote_props); |
| 417 | quote_props &= ~QUOTE_INITCTRL; |
| 418 | fprintf(fp, "%s", c); |
| 419 | sfree(c); |
| 420 | if (ec == 'i' || ec == 'b') |
| 421 | fprintf(fp, "\\fP"); |
| 422 | t += n; |
| 423 | e += n; |
| 424 | } |
| 425 | man_convert(t, 0, &c, quote_props); |
| 426 | fprintf(fp, "%s\n", c); |
| 427 | sfree(c); |
| 428 | } |
| 429 | fprintf(fp, ".fi\n"); |
| 430 | } |