| 1 | /* |
| 2 | * xhtml backend for Halibut |
| 3 | * (initial implementation by James Aylett) |
| 4 | * |
| 5 | * Still to do: |
| 6 | * |
| 7 | * +++ doesn't handle non-breaking hyphens. Not sure how to yet. |
| 8 | * +++ entity names (from a file -- ideally supply normal SGML files) |
| 9 | * +++ configuration directive to file split where the current layout |
| 10 | * code wouldn't. Needs changes to _ponder_layout() and _do_paras(), |
| 11 | * perhaps others. |
| 12 | * |
| 13 | * Limitations: |
| 14 | * |
| 15 | * +++ biblio/index references target the nearest section marker, rather |
| 16 | * than having a dedicated target themselves. In large bibliographies |
| 17 | * this will cause problems. (The solution is to fake up a response |
| 18 | * from xhtml_find_section(), probably linking it into the sections |
| 19 | * chain just in case we need it again, and to make freeing it up |
| 20 | * easier.) docsrc.pl used to work as we do, however, and SGT agrees that |
| 21 | * this is acceptable for now. |
| 22 | * +++ can't cope with leaf-level == 0. It's all to do with the |
| 23 | * top-level file not being normal, probably not even having a valid |
| 24 | * section level, and stuff like that. I question whether this is an |
| 25 | * issue, frankly; small manuals that fit on one page should probably |
| 26 | * not be written in halibut at all. |
| 27 | */ |
| 28 | |
| 29 | #include <stdio.h> |
| 30 | #include <stdlib.h> |
| 31 | #include <string.h> |
| 32 | #include <assert.h> |
| 33 | #include "halibut.h" |
| 34 | |
| 35 | /* |
| 36 | * FILENAME_TEMPLATE (overridable in config of course) allows you |
| 37 | * to choose the general form for your HTML file names. It is |
| 38 | * slightly printf-styled (% followed by a single character is a |
| 39 | * formatting directive, %% is a literal %). Formatting directives |
| 40 | * are: |
| 41 | * |
| 42 | * - %n is the section type-plus-number, minus whitespace (`Chapter1.2'). |
| 43 | * - %b is the section number on its own (`1.2'). |
| 44 | * - %k is the section's _internal_ keyword. |
| 45 | * - %N is the section's visible title in the output, again minus |
| 46 | * whitespace. |
| 47 | * |
| 48 | * %n, %b and %k will all default to %N if the section is |
| 49 | * unnumbered (`Bibliography' is often a good example). |
| 50 | * |
| 51 | * FRAGMENT_TEMPLATE is the same, but defines the <a name="foo"> |
| 52 | * markers used to cross-reference to particular subsections of a |
| 53 | * file. |
| 54 | */ |
| 55 | |
| 56 | #define FILENAME_SINGLE "Manual.html" |
| 57 | #define FILENAME_CONTENTS "Contents.html" |
| 58 | #define FILENAME_INDEX "IndexPage.html" |
| 59 | #define FILENAME_TEMPLATE "%n.html" |
| 60 | #define FRAGMENT_TEMPLATE "%b" |
| 61 | |
| 62 | struct xhtmlsection_Struct { |
| 63 | struct xhtmlsection_Struct *next; /* next sibling (NULL if split across files) */ |
| 64 | struct xhtmlsection_Struct *child; /* NULL if split across files */ |
| 65 | struct xhtmlsection_Struct *parent; /* NULL if split across files */ |
| 66 | struct xhtmlsection_Struct *chain; /* single structure independent of weird trees */ |
| 67 | paragraph *para; |
| 68 | struct xhtmlfile_Struct *file; /* which file is this a part of? */ |
| 69 | char *fragment; /* fragment id within the file */ |
| 70 | int level; |
| 71 | }; |
| 72 | |
| 73 | struct xhtmlfile_Struct { |
| 74 | struct xhtmlfile_Struct *next; |
| 75 | struct xhtmlfile_Struct *child; |
| 76 | struct xhtmlfile_Struct *parent; |
| 77 | char *filename; |
| 78 | struct xhtmlsection_Struct *sections; /* sections within this file (only one for non-leaf) */ |
| 79 | int is_leaf; /* is this file a leaf file, ie does it not have any children? */ |
| 80 | }; |
| 81 | |
| 82 | typedef struct xhtmlsection_Struct xhtmlsection; |
| 83 | typedef struct xhtmlfile_Struct xhtmlfile; |
| 84 | typedef struct xhtmlindex_Struct xhtmlindex; |
| 85 | |
| 86 | struct xhtmlindex_Struct { |
| 87 | int nsection; |
| 88 | int size; |
| 89 | xhtmlsection **sections; |
| 90 | }; |
| 91 | |
| 92 | typedef struct { |
| 93 | int just_numbers; |
| 94 | wchar_t *number_suffix; |
| 95 | } xhtmlheadfmt; |
| 96 | |
| 97 | typedef struct { |
| 98 | int contents_depth[6]; |
| 99 | int leaf_contains_contents; |
| 100 | int leaf_level; |
| 101 | int leaf_smallest_contents; |
| 102 | int include_version_id; |
| 103 | wchar_t *author, *description; |
| 104 | wchar_t *head_end, *body, *body_start, *body_end, *address_start, *address_end, *nav_attrs; |
| 105 | int suppress_address; |
| 106 | xhtmlheadfmt fchapter, *fsect; |
| 107 | int nfsect; |
| 108 | char *contents_filename, *index_filename; |
| 109 | char *single_filename, *template_filename, *template_fragment; |
| 110 | } xhtmlconfig; |
| 111 | |
| 112 | /*static void xhtml_level(paragraph *, int); |
| 113 | static void xhtml_level_0(paragraph *); |
| 114 | static void xhtml_docontents(FILE *, paragraph *, int); |
| 115 | static void xhtml_dosections(FILE *, paragraph *, int); |
| 116 | static void xhtml_dobody(FILE *, paragraph *, int);*/ |
| 117 | |
| 118 | static void xhtml_doheader(FILE *, word *); |
| 119 | static void xhtml_dofooter(FILE *); |
| 120 | static void xhtml_versionid(FILE *, word *, int); |
| 121 | |
| 122 | static void xhtml_utostr(wchar_t *, char **); |
| 123 | static int xhtml_para_level(paragraph *); |
| 124 | static int xhtml_reservedchar(int); |
| 125 | |
| 126 | static int xhtml_convert(wchar_t *, int, char **, int); |
| 127 | static void xhtml_rdaddwc(rdstringc *, word *, word *, int); |
| 128 | static void xhtml_para(FILE *, word *, int); |
| 129 | static void xhtml_codepara(FILE *, word *); |
| 130 | static void xhtml_heading(FILE *, paragraph *, int); |
| 131 | |
| 132 | /* File-global variables are much easier than passing these things |
| 133 | * all over the place. Evil, but easier. We can replace this with a single |
| 134 | * structure at some point. |
| 135 | */ |
| 136 | static xhtmlconfig conf; |
| 137 | static keywordlist *keywords; |
| 138 | static indexdata *idx; |
| 139 | static xhtmlfile *topfile; |
| 140 | static xhtmlsection *topsection; |
| 141 | static paragraph *sourceparas; |
| 142 | static xhtmlfile *lastfile; |
| 143 | static xhtmlfile *xhtml_last_file = NULL; |
| 144 | static int last_level=-1, start_level; |
| 145 | static xhtmlsection *currentsection; |
| 146 | |
| 147 | static xhtmlconfig xhtml_configure(paragraph *source) |
| 148 | { |
| 149 | xhtmlconfig ret; |
| 150 | |
| 151 | /* |
| 152 | * Defaults. |
| 153 | */ |
| 154 | ret.contents_depth[0] = 2; |
| 155 | ret.contents_depth[1] = 3; |
| 156 | ret.contents_depth[2] = 4; |
| 157 | ret.contents_depth[3] = 5; |
| 158 | ret.contents_depth[4] = 6; |
| 159 | ret.contents_depth[5] = 7; |
| 160 | ret.leaf_level = 2; |
| 161 | ret.leaf_smallest_contents = 4; |
| 162 | ret.leaf_contains_contents = FALSE; |
| 163 | ret.include_version_id = TRUE; |
| 164 | ret.author = NULL; |
| 165 | ret.description = NULL; |
| 166 | ret.head_end = NULL; |
| 167 | ret.body = NULL; |
| 168 | ret.body_start = NULL; |
| 169 | ret.body_end = NULL; |
| 170 | ret.address_start = NULL; |
| 171 | ret.address_end = NULL; |
| 172 | ret.nav_attrs = NULL; |
| 173 | ret.suppress_address = FALSE; |
| 174 | |
| 175 | ret.fchapter.just_numbers = FALSE; |
| 176 | ret.fchapter.number_suffix = L": "; |
| 177 | ret.nfsect = 2; |
| 178 | ret.fsect = mknewa(xhtmlheadfmt, ret.nfsect); |
| 179 | ret.fsect[0].just_numbers = FALSE; |
| 180 | ret.fsect[0].number_suffix = L": "; |
| 181 | ret.fsect[1].just_numbers = TRUE; |
| 182 | ret.fsect[1].number_suffix = L" "; |
| 183 | ret.contents_filename = strdup(FILENAME_CONTENTS); |
| 184 | ret.single_filename = strdup(FILENAME_SINGLE); |
| 185 | ret.index_filename = strdup(FILENAME_INDEX); |
| 186 | ret.template_filename = strdup(FILENAME_TEMPLATE); |
| 187 | ret.template_fragment = strdup(FRAGMENT_TEMPLATE); |
| 188 | |
| 189 | for (; source; source = source->next) |
| 190 | { |
| 191 | if (source->type == para_Config) |
| 192 | { |
| 193 | if (!ustricmp(source->keyword, L"xhtml-contents-filename")) { |
| 194 | sfree(ret.contents_filename); |
| 195 | ret.contents_filename = utoa_dup(uadv(source->keyword)); |
| 196 | } else if (!ustricmp(source->keyword, L"xhtml-single-filename")) { |
| 197 | sfree(ret.single_filename); |
| 198 | ret.single_filename = utoa_dup(uadv(source->keyword)); |
| 199 | } else if (!ustricmp(source->keyword, L"xhtml-index-filename")) { |
| 200 | sfree(ret.index_filename); |
| 201 | ret.index_filename = utoa_dup(uadv(source->keyword)); |
| 202 | } else if (!ustricmp(source->keyword, L"xhtml-template-filename")) { |
| 203 | sfree(ret.template_filename); |
| 204 | ret.template_filename = utoa_dup(uadv(source->keyword)); |
| 205 | } else if (!ustricmp(source->keyword, L"xhtml-template-fragment")) { |
| 206 | sfree(ret.template_fragment); |
| 207 | ret.template_fragment = utoa_dup(uadv(source->keyword)); |
| 208 | } else if (!ustricmp(source->keyword, L"xhtml-contents-depth-0")) { |
| 209 | ret.contents_depth[0] = utoi(uadv(source->keyword)); |
| 210 | } else if (!ustricmp(source->keyword, L"xhtml-contents-depth-1")) { |
| 211 | ret.contents_depth[1] = utoi(uadv(source->keyword)); |
| 212 | } else if (!ustricmp(source->keyword, L"xhtml-contents-depth-2")) { |
| 213 | ret.contents_depth[2] = utoi(uadv(source->keyword)); |
| 214 | } else if (!ustricmp(source->keyword, L"xhtml-contents-depth-3")) { |
| 215 | ret.contents_depth[3] = utoi(uadv(source->keyword)); |
| 216 | } else if (!ustricmp(source->keyword, L"xhtml-contents-depth-4")) { |
| 217 | ret.contents_depth[4] = utoi(uadv(source->keyword)); |
| 218 | } else if (!ustricmp(source->keyword, L"xhtml-contents-depth-5")) { |
| 219 | ret.contents_depth[5] = utoi(uadv(source->keyword)); |
| 220 | } else if (!ustricmp(source->keyword, L"xhtml-leaf-level")) { |
| 221 | ret.leaf_level = utoi(uadv(source->keyword)); |
| 222 | } else if (!ustricmp(source->keyword, L"xhtml-leaf-smallest-contents")) { |
| 223 | ret.leaf_smallest_contents = utoi(uadv(source->keyword)); |
| 224 | } else if (!ustricmp(source->keyword, L"xhtml-versionid")) { |
| 225 | ret.include_version_id = utob(uadv(source->keyword)); |
| 226 | } else if (!ustricmp(source->keyword, L"xhtml-leaf-contains-contents")) { |
| 227 | ret.leaf_contains_contents = utob(uadv(source->keyword)); |
| 228 | } else if (!ustricmp(source->keyword, L"xhtml-suppress-address")) { |
| 229 | ret.suppress_address = utob(uadv(source->keyword)); |
| 230 | } else if (!ustricmp(source->keyword, L"xhtml-author")) { |
| 231 | ret.author = uadv(source->keyword); |
| 232 | } else if (!ustricmp(source->keyword, L"xhtml-description")) { |
| 233 | ret.description = uadv(source->keyword); |
| 234 | } else if (!ustricmp(source->keyword, L"xhtml-head-end")) { |
| 235 | ret.head_end = uadv(source->keyword); |
| 236 | } else if (!ustricmp(source->keyword, L"xhtml-body-start")) { |
| 237 | ret.body_start = uadv(source->keyword); |
| 238 | } else if (!ustricmp(source->keyword, L"xhtml-body-tag")) { |
| 239 | ret.body = uadv(source->keyword); |
| 240 | } else if (!ustricmp(source->keyword, L"xhtml-body-end")) { |
| 241 | ret.body_end = uadv(source->keyword); |
| 242 | } else if (!ustricmp(source->keyword, L"xhtml-address-start")) { |
| 243 | ret.address_start = uadv(source->keyword); |
| 244 | } else if (!ustricmp(source->keyword, L"xhtml-address-end")) { |
| 245 | ret.address_end = uadv(source->keyword); |
| 246 | } else if (!ustricmp(source->keyword, L"xhtml-navigation-attributes")) { |
| 247 | ret.nav_attrs = uadv(source->keyword); |
| 248 | } else if (!ustricmp(source->keyword, L"xhtml-chapter-numeric")) { |
| 249 | ret.fchapter.just_numbers = utob(uadv(source->keyword)); |
| 250 | } else if (!ustricmp(source->keyword, L"xhtml-chapter-suffix")) { |
| 251 | ret.fchapter.number_suffix = uadv(source->keyword); |
| 252 | } else if (!ustricmp(source->keyword, L"xhtml-section-numeric")) { |
| 253 | wchar_t *p = uadv(source->keyword); |
| 254 | int n = 0; |
| 255 | if (uisdigit(*p)) { |
| 256 | n = utoi(p); |
| 257 | p = uadv(p); |
| 258 | } |
| 259 | if (n >= ret.nfsect) { |
| 260 | int i; |
| 261 | ret.fsect = resize(ret.fsect, n+1); |
| 262 | for (i = ret.nfsect; i <= n; i++) |
| 263 | ret.fsect[i] = ret.fsect[ret.nfsect-1]; |
| 264 | ret.nfsect = n+1; |
| 265 | } |
| 266 | ret.fsect[n].just_numbers = utob(p); |
| 267 | } else if (!ustricmp(source->keyword, L"xhtml-section-suffix")) { |
| 268 | wchar_t *p = uadv(source->keyword); |
| 269 | int n = 0; |
| 270 | if (uisdigit(*p)) { |
| 271 | n = utoi(p); |
| 272 | p = uadv(p); |
| 273 | } |
| 274 | if (n >= ret.nfsect) { |
| 275 | int i; |
| 276 | ret.fsect = resize(ret.fsect, n+1); |
| 277 | for (i = ret.nfsect; i <= n; i++) |
| 278 | ret.fsect[i] = ret.fsect[ret.nfsect-1]; |
| 279 | ret.nfsect = n+1; |
| 280 | } |
| 281 | ret.fsect[n].number_suffix = p; |
| 282 | } |
| 283 | } |
| 284 | } |
| 285 | |
| 286 | /* printf(" !!! leaf_level = %i\n", ret.leaf_level); |
| 287 | printf(" !!! contentdepth-0 = %i\n", ret.contents_depth[0]); |
| 288 | printf(" !!! contentdepth-1 = %i\n", ret.contents_depth[1]); |
| 289 | printf(" !!! contentdepth-2 = %i\n", ret.contents_depth[2]); |
| 290 | printf(" !!! contentdepth-3 = %i\n", ret.contents_depth[3]); |
| 291 | printf(" !!! contentdepth-4 = %i\n", ret.contents_depth[4]); |
| 292 | printf(" !!! contentdepth-5 = %i\n", ret.contents_depth[5]); |
| 293 | printf(" !!! leaf_contains_contents = %i\n", ret.leaf_contains_contents);*/ |
| 294 | return ret; |
| 295 | } |
| 296 | |
| 297 | paragraph *xhtml_config_filename(char *filename) |
| 298 | { |
| 299 | /* |
| 300 | * If the user passes in a single filename as a parameter to |
| 301 | * the `--html' command-line option, then we should assume it |
| 302 | * to imply _two_ config directives: |
| 303 | * \cfg{xhtml-single-filename}{whatever} and |
| 304 | * \cfg{xhtml-leaf-level}{0}; the rationale being that the user |
| 305 | * wants their output _in that file_. |
| 306 | */ |
| 307 | |
| 308 | paragraph *p[2]; |
| 309 | int i, len; |
| 310 | wchar_t *ufilename, *up; |
| 311 | |
| 312 | for (i = 0; i < 2; i++) { |
| 313 | p[i] = mknew(paragraph); |
| 314 | memset(p[i], 0, sizeof(*p[i])); |
| 315 | p[i]->type = para_Config; |
| 316 | p[i]->next = NULL; |
| 317 | p[i]->fpos.filename = "<command line>"; |
| 318 | p[i]->fpos.line = p[i]->fpos.col = -1; |
| 319 | } |
| 320 | |
| 321 | ufilename = ufroma_dup(filename); |
| 322 | len = ustrlen(ufilename) + 2 + lenof(L"xhtml-single-filename"); |
| 323 | p[0]->keyword = mknewa(wchar_t, len); |
| 324 | up = p[0]->keyword; |
| 325 | ustrcpy(up, L"xhtml-single-filename"); |
| 326 | up = uadv(up); |
| 327 | ustrcpy(up, ufilename); |
| 328 | up = uadv(up); |
| 329 | *up = L'\0'; |
| 330 | assert(up - p[0]->keyword < len); |
| 331 | sfree(ufilename); |
| 332 | |
| 333 | len = lenof(L"xhtml-leaf-level") + lenof(L"0") + 1; |
| 334 | p[1]->keyword = mknewa(wchar_t, len); |
| 335 | up = p[1]->keyword; |
| 336 | ustrcpy(up, L"xhtml-leaf-level"); |
| 337 | up = uadv(up); |
| 338 | ustrcpy(up, L"0"); |
| 339 | up = uadv(up); |
| 340 | *up = L'\0'; |
| 341 | assert(up - p[1]->keyword < len); |
| 342 | |
| 343 | p[0]->next = p[1]; |
| 344 | |
| 345 | return p[0]; |
| 346 | } |
| 347 | |
| 348 | static xhtmlsection *xhtml_new_section(xhtmlsection *last) |
| 349 | { |
| 350 | xhtmlsection *ret = mknew(xhtmlsection); |
| 351 | ret->next=NULL; |
| 352 | ret->child=NULL; |
| 353 | ret->parent=NULL; |
| 354 | ret->chain=last; |
| 355 | ret->para=NULL; |
| 356 | ret->file=NULL; |
| 357 | ret->fragment=NULL; |
| 358 | ret->level=-1; /* marker: end of chain */ |
| 359 | return ret; |
| 360 | } |
| 361 | |
| 362 | /* Returns NULL or the section that marks that paragraph */ |
| 363 | static xhtmlsection *xhtml_find_section(paragraph *p) |
| 364 | { |
| 365 | xhtmlsection *ret = topsection; |
| 366 | if (xhtml_para_level(p)==-1) { /* first, we back-track to a section paragraph */ |
| 367 | paragraph *p2 = sourceparas; |
| 368 | paragraph *p3 = NULL; |
| 369 | while (p2 && p2!=p) { |
| 370 | if (xhtml_para_level(p2)!=-1) { |
| 371 | p3 = p2; |
| 372 | } |
| 373 | p2=p2->next; |
| 374 | } |
| 375 | if (p3==NULL) { /* for some reason, we couldn't find a section before this paragraph ... ? */ |
| 376 | /* Note that this can happen, if you have a cross-reference to before the first chapter starts. |
| 377 | * So don't do that, then. |
| 378 | */ |
| 379 | return NULL; |
| 380 | } |
| 381 | p=p3; |
| 382 | } |
| 383 | while (ret && ret->para != p) { |
| 384 | /* printf(" xhtml_find_section(): checking %s for para @ %p\n", ret->fragment, p);*/ |
| 385 | ret=ret->chain; |
| 386 | } |
| 387 | return ret; |
| 388 | } |
| 389 | |
| 390 | static void xhtml_format(paragraph *p, char *template_string, rdstringc *r) |
| 391 | { |
| 392 | char *c, *t; |
| 393 | word *w; |
| 394 | wchar_t *ws; |
| 395 | |
| 396 | t = template_string; |
| 397 | while (*t) { |
| 398 | if (*t == '%' && t[1]) { |
| 399 | int fmt; |
| 400 | |
| 401 | t++; |
| 402 | fmt = *t++; |
| 403 | |
| 404 | if (fmt == '%') { |
| 405 | rdaddc(r, fmt); |
| 406 | continue; |
| 407 | } |
| 408 | |
| 409 | w = NULL; |
| 410 | ws = NULL; |
| 411 | |
| 412 | if (p->kwtext && fmt == 'n') |
| 413 | w = p->kwtext; |
| 414 | else if (p->kwtext2 && fmt == 'b') |
| 415 | w = p->kwtext2; |
| 416 | else if (p->keyword && *p->keyword && fmt == 'k') |
| 417 | ws = p->keyword; |
| 418 | else |
| 419 | w = p->words; |
| 420 | |
| 421 | while (w) { |
| 422 | switch (removeattr(w->type)) |
| 423 | { |
| 424 | case word_Normal: |
| 425 | /*case word_Emph: |
| 426 | case word_Code: |
| 427 | case word_WeakCode:*/ |
| 428 | xhtml_utostr(w->text, &c); |
| 429 | rdaddsc(r,c); |
| 430 | sfree(c); |
| 431 | break; |
| 432 | } |
| 433 | w = w->next; |
| 434 | } |
| 435 | if (ws) { |
| 436 | xhtml_utostr(ws, &c); |
| 437 | rdaddsc(r,c); |
| 438 | sfree(c); |
| 439 | } |
| 440 | } else { |
| 441 | rdaddc(r, *t++); |
| 442 | } |
| 443 | } |
| 444 | } |
| 445 | |
| 446 | static xhtmlfile *xhtml_new_file(xhtmlsection *sect) |
| 447 | { |
| 448 | xhtmlfile *ret = mknew(xhtmlfile); |
| 449 | |
| 450 | ret->next=NULL; |
| 451 | ret->child=NULL; |
| 452 | ret->parent=NULL; |
| 453 | ret->filename=NULL; |
| 454 | ret->sections=sect; |
| 455 | ret->is_leaf=(sect!=NULL && sect->level==conf.leaf_level); |
| 456 | if (sect==NULL) { |
| 457 | if (conf.leaf_level==0) { /* currently unused */ |
| 458 | ret->filename = smalloc(strlen(conf.single_filename)+1); |
| 459 | sprintf(ret->filename, conf.single_filename); |
| 460 | } else { |
| 461 | ret->filename = smalloc(strlen(conf.contents_filename)+1); |
| 462 | sprintf(ret->filename, conf.contents_filename); |
| 463 | } |
| 464 | } else { |
| 465 | paragraph *p = sect->para; |
| 466 | rdstringc fname_c = { 0, 0, NULL }; |
| 467 | xhtml_format(p, conf.template_filename, &fname_c); |
| 468 | ret->filename = rdtrimc(&fname_c); |
| 469 | } |
| 470 | /* printf(" ! new file '%s', is_leaf == %s\n", ret->filename, (ret->is_leaf)?("true"):("false"));*/ |
| 471 | return ret; |
| 472 | } |
| 473 | |
| 474 | /* |
| 475 | * Walk the tree fixing up files which are actually leaf (ie |
| 476 | * have no children) but aren't at leaf level, so they have the |
| 477 | * leaf flag set. |
| 478 | */ |
| 479 | void xhtml_fixup_layout(xhtmlfile* file) |
| 480 | { |
| 481 | if (file->child==NULL) { |
| 482 | file->is_leaf = TRUE; |
| 483 | } else { |
| 484 | xhtml_fixup_layout(file->child); |
| 485 | } |
| 486 | if (file->next) |
| 487 | xhtml_fixup_layout(file->next); |
| 488 | } |
| 489 | |
| 490 | /* |
| 491 | * Create the tree structure so we know where everything goes. |
| 492 | * Method: |
| 493 | * |
| 494 | * Ignoring file splitting, we have three choices with each new section: |
| 495 | * |
| 496 | * +-----------------+-----------------+ |
| 497 | * | | | |
| 498 | * X +----X----+ (1) |
| 499 | * | | |
| 500 | * Y (2) |
| 501 | * | |
| 502 | * (3) |
| 503 | * |
| 504 | * Y is the last section we added (currentsect). |
| 505 | * If sect is the section we want to add, then: |
| 506 | * |
| 507 | * (1) if sect->level < currentsect->level |
| 508 | * (2) if sect->level == currentsect->level |
| 509 | * (3) if sect->level > currentsect->level |
| 510 | * |
| 511 | * This requires the constraint that you never skip section numbers |
| 512 | * (so you can't have a.b.c.d without all of a, a.b and a.b.c existing). |
| 513 | * |
| 514 | * Note that you _can_ have 1.1.1.1 followed by 1.2 - you can change |
| 515 | * more than one level at a time. Lots of asserts, and probably part of |
| 516 | * the algorithm here, rely on this being true. (It currently isn't |
| 517 | * enforced by halibut, however.) |
| 518 | * |
| 519 | * File splitting makes this harder. For instance, say we added at (3) |
| 520 | * above and now need to add another section. We are splitting at level |
| 521 | * 2, ie the level of Y. Z is the last section we added: |
| 522 | * |
| 523 | * +-----------------+-----------------+ |
| 524 | * | | | |
| 525 | * X +----X----+ (1) |
| 526 | * | | |
| 527 | * +----Y----+ (1) |
| 528 | * | | |
| 529 | * Z (2) |
| 530 | * | |
| 531 | * (3) |
| 532 | * |
| 533 | * The (1) case is now split; we need to search upwards to find where |
| 534 | * to actually link in. The other two cases remain the same (and will |
| 535 | * always be like this). |
| 536 | * |
| 537 | * File splitting makes this harder, however. The decision of whether |
| 538 | * to split to a new file is always on the same condition, however (is |
| 539 | * the level of this section higher than the leaf_level configuration |
| 540 | * value or not). |
| 541 | * |
| 542 | * Treating the cases backwards: |
| 543 | * |
| 544 | * (3) same file if sect->level > conf.leaf_level, otherwise new file |
| 545 | * |
| 546 | * if in the same file, currentsect->child points to sect |
| 547 | * otherwise the linking is done through the file tree (which works |
| 548 | * in more or less the same way, ie currentfile->child points to |
| 549 | * the new file) |
| 550 | * |
| 551 | * (2) same file if sect->level > conf.leaf_level, otherwise new file |
| 552 | * |
| 553 | * if in the same file, currentsect->next points to sect |
| 554 | * otherwise file linking and currentfile->next points to the new |
| 555 | * file (we know that Z must have caused a new file to be created) |
| 556 | * |
| 557 | * (1) same file if sect->level > conf.leaf_level, otherwise new file |
| 558 | * |
| 559 | * this is actually effectively the same case as (2) here, |
| 560 | * except that we first have to travel up the sections to figure |
| 561 | * out which section this new one will be a sibling of. In doing |
| 562 | * so, we may disappear off the top of a file and have to go up |
| 563 | * to its parent in the file tree. |
| 564 | * |
| 565 | */ |
| 566 | static void xhtml_ponder_layout(paragraph *p) |
| 567 | { |
| 568 | xhtmlsection *lastsection; |
| 569 | xhtmlsection *currentsect; |
| 570 | xhtmlfile *currentfile; |
| 571 | |
| 572 | lastfile = NULL; |
| 573 | topsection = xhtml_new_section(NULL); |
| 574 | topfile = xhtml_new_file(NULL); |
| 575 | lastsection = topsection; |
| 576 | currentfile = topfile; |
| 577 | currentsect = topsection; |
| 578 | |
| 579 | if (conf.leaf_level == 0) { |
| 580 | topfile->is_leaf = 1; |
| 581 | topfile->sections = topsection; |
| 582 | topsection->file = topfile; |
| 583 | } |
| 584 | |
| 585 | for (; p; p=p->next) |
| 586 | { |
| 587 | int level = xhtml_para_level(p); |
| 588 | if (level>0) /* actually a section */ |
| 589 | { |
| 590 | xhtmlsection *sect; |
| 591 | rdstringc frag_c = { 0, 0, NULL }; |
| 592 | |
| 593 | sect = xhtml_new_section(lastsection); |
| 594 | lastsection = sect; |
| 595 | sect->para = p; |
| 596 | |
| 597 | xhtml_format(p, conf.template_fragment, &frag_c); |
| 598 | sect->fragment = rdtrimc(&frag_c); |
| 599 | sect->level = level; |
| 600 | /* printf(" ! adding para @ %p as sect %s, level %i\n", sect->para, sect->fragment, level);*/ |
| 601 | |
| 602 | if (level>currentsect->level) { /* case (3) */ |
| 603 | if (level>conf.leaf_level) { /* same file */ |
| 604 | assert(currentfile->is_leaf); |
| 605 | currentsect->child = sect; |
| 606 | sect->parent=currentsect; |
| 607 | sect->file=currentfile; |
| 608 | /* printf("connected '%s' to existing file '%s' [I]\n", sect->fragment, currentfile->filename);*/ |
| 609 | currentsect=sect; |
| 610 | } else { /* new file */ |
| 611 | xhtmlfile *file = xhtml_new_file(sect); |
| 612 | assert(!currentfile->is_leaf); |
| 613 | currentfile->child=file; |
| 614 | sect->file=file; |
| 615 | file->parent=currentfile; |
| 616 | /* printf("connected '%s' to new file '%s' [I]\n", sect->fragment, file->filename);*/ |
| 617 | currentfile=file; |
| 618 | currentsect=sect; |
| 619 | } |
| 620 | } else if (level >= currentsect->file->sections->level) { |
| 621 | /* Case (1) or (2) *AND* still under the section that starts |
| 622 | * the current file. |
| 623 | * |
| 624 | * I'm not convinced that this couldn't be rolled in with the |
| 625 | * final else {} leg further down. It seems a lot of effort |
| 626 | * this way. |
| 627 | */ |
| 628 | if (level>conf.leaf_level) { /* stick within the same file */ |
| 629 | assert(currentfile->is_leaf); |
| 630 | sect->file = currentfile; |
| 631 | while (currentsect && currentsect->level > level && |
| 632 | currentsect->file==currentsect->parent->file) { |
| 633 | currentsect = currentsect->parent; |
| 634 | } |
| 635 | assert(currentsect); |
| 636 | currentsect->next = sect; |
| 637 | assert(currentsect->level == sect->level); |
| 638 | sect->parent = currentsect->parent; |
| 639 | currentsect = sect; |
| 640 | /* printf("connected '%s' to existing file '%s' [II]\n", sect->fragment, currentfile->filename);*/ |
| 641 | } else { /* new file */ |
| 642 | xhtmlfile *file = xhtml_new_file(sect); |
| 643 | sect->file=file; |
| 644 | currentfile->next=file; |
| 645 | file->parent=currentfile->parent; |
| 646 | file->is_leaf=(level==conf.leaf_level); |
| 647 | file->sections=sect; |
| 648 | /* printf("connected '%s' to new file '%s' [II]\n", sect->fragment, file->filename);*/ |
| 649 | currentfile=file; |
| 650 | currentsect=sect; |
| 651 | } |
| 652 | } else { /* Case (1) or (2) and we must move up the file tree first */ |
| 653 | /* this loop is now probably irrelevant - we know we can't connect |
| 654 | * to anything in the current file */ |
| 655 | while (currentsect && level<currentsect->level) { |
| 656 | currentsect=currentsect->parent; |
| 657 | if (currentsect) { |
| 658 | /* printf(" * up one level to '%s'\n", currentsect->fragment);*/ |
| 659 | } else { |
| 660 | /* printf(" * up one level (off top of current file)\n");*/ |
| 661 | } |
| 662 | } |
| 663 | if (currentsect) { |
| 664 | /* I'm pretty sure this can now never fire */ |
| 665 | assert(currentfile->is_leaf); |
| 666 | /* printf("connected '%s' to existing file '%s' [III]\n", sect->fragment, currentfile->filename);*/ |
| 667 | sect->file = currentfile; |
| 668 | currentsect->next=sect; |
| 669 | currentsect=sect; |
| 670 | } else { /* find a file we can attach to */ |
| 671 | while (currentfile && currentfile->sections && level<currentfile->sections->level) { |
| 672 | currentfile=currentfile->parent; |
| 673 | if (currentfile) { |
| 674 | /* printf(" * up one file level to '%s'\n", currentfile->filename);*/ |
| 675 | } else { |
| 676 | /* printf(" * up one file level (off top of tree)\n");*/ |
| 677 | } |
| 678 | } |
| 679 | if (currentfile) { /* new file (we had to skip up a file to |
| 680 | get here, so we must be dealing with a |
| 681 | level no lower than the configured |
| 682 | leaf_level */ |
| 683 | xhtmlfile *file = xhtml_new_file(sect); |
| 684 | currentfile->next=file; |
| 685 | sect->file=file; |
| 686 | file->parent=currentfile->parent; |
| 687 | file->is_leaf=(level==conf.leaf_level); |
| 688 | file->sections=sect; |
| 689 | /* printf("connected '%s' to new file '%s' [III]\n", sect->fragment, file->filename);*/ |
| 690 | currentfile=file; |
| 691 | currentsect=sect; |
| 692 | } else { |
| 693 | fatal(err_whatever, "Ran off the top trying to connect sibling: strange document."); |
| 694 | } |
| 695 | } |
| 696 | } |
| 697 | } |
| 698 | } |
| 699 | topsection = lastsection; /* get correct end of the chain */ |
| 700 | xhtml_fixup_layout(topfile); /* leaf files not at leaf level marked as such */ |
| 701 | } |
| 702 | |
| 703 | static void xhtml_do_index(); |
| 704 | static void xhtml_do_file(xhtmlfile *file); |
| 705 | static void xhtml_do_top_file(xhtmlfile *file, paragraph *sourceform); |
| 706 | static void xhtml_do_paras(FILE *fp, paragraph *p, paragraph *end, int indexable); |
| 707 | static int xhtml_do_contents_limit(FILE *fp, xhtmlfile *file, int limit); |
| 708 | static int xhtml_do_contents_section_limit(FILE *fp, xhtmlsection *section, int limit); |
| 709 | static int xhtml_add_contents_entry(FILE *fp, xhtmlsection *section, int limit); |
| 710 | static int xhtml_do_contents(FILE *fp, xhtmlfile *file); |
| 711 | static int xhtml_do_naked_contents(FILE *fp, xhtmlfile *file); |
| 712 | static void xhtml_do_sections(FILE *fp, xhtmlsection *sections); |
| 713 | |
| 714 | /* |
| 715 | * Do all the files in this structure. |
| 716 | */ |
| 717 | static void xhtml_do_files(xhtmlfile *file) |
| 718 | { |
| 719 | xhtml_do_file(file); |
| 720 | if (file->child) |
| 721 | xhtml_do_files(file->child); |
| 722 | if (file->next) |
| 723 | xhtml_do_files(file->next); |
| 724 | } |
| 725 | |
| 726 | /* |
| 727 | * Free up all memory used by the file tree from 'xfile' downwards |
| 728 | */ |
| 729 | static void xhtml_free_file(xhtmlfile* xfile) |
| 730 | { |
| 731 | if (xfile==NULL) { |
| 732 | return; |
| 733 | } |
| 734 | |
| 735 | if (xfile->filename) { |
| 736 | sfree(xfile->filename); |
| 737 | } |
| 738 | xhtml_free_file(xfile->child); |
| 739 | xhtml_free_file(xfile->next); |
| 740 | sfree(xfile); |
| 741 | } |
| 742 | |
| 743 | /* |
| 744 | * Main function. |
| 745 | */ |
| 746 | void xhtml_backend(paragraph *sourceform, keywordlist *in_keywords, |
| 747 | indexdata *in_idx) |
| 748 | { |
| 749 | /* int i;*/ |
| 750 | indexentry *ientry; |
| 751 | int ti; |
| 752 | xhtmlsection *xsect; |
| 753 | |
| 754 | sourceparas = sourceform; |
| 755 | conf = xhtml_configure(sourceform); |
| 756 | keywords = in_keywords; |
| 757 | idx = in_idx; |
| 758 | |
| 759 | /* Clear up the index entries backend data pointers */ |
| 760 | for (ti=0; (ientry = (indexentry *)index234(idx->entries, ti))!=NULL; ti++) { |
| 761 | ientry->backend_data=NULL; |
| 762 | } |
| 763 | |
| 764 | xhtml_ponder_layout(sourceform); |
| 765 | |
| 766 | /* old system ... (writes to *.alt, but gets some stuff wrong and is ugly) */ |
| 767 | /* xhtml_level_0(sourceform); |
| 768 | for (i=1; i<=conf.leaf_level; i++) |
| 769 | { |
| 770 | xhtml_level(sourceform, i); |
| 771 | }*/ |
| 772 | |
| 773 | /* new system ... (writes to *.html, but isn't fully trusted) */ |
| 774 | xhtml_do_top_file(topfile, sourceform); |
| 775 | assert(!topfile->next); /* shouldn't have a sibling at all */ |
| 776 | if (topfile->child) { |
| 777 | xhtml_do_files(topfile->child); |
| 778 | xhtml_do_index(); |
| 779 | } |
| 780 | |
| 781 | /* release file, section, index data structures */ |
| 782 | xsect = topsection; |
| 783 | while (xsect) { |
| 784 | xhtmlsection *tmp = xsect->chain; |
| 785 | if (xsect->fragment) { |
| 786 | sfree(xsect->fragment); |
| 787 | } |
| 788 | sfree(xsect); |
| 789 | xsect = tmp; |
| 790 | } |
| 791 | xhtml_free_file(topfile); |
| 792 | for (ti = 0; (ientry=(indexentry *)index234(idx->entries, ti))!=NULL; ti++) { |
| 793 | if (ientry->backend_data!=NULL) { |
| 794 | xhtmlindex *xi = (xhtmlindex*) ientry->backend_data; |
| 795 | if (xi->sections!=NULL) { |
| 796 | sfree(xi->sections); |
| 797 | } |
| 798 | sfree(xi); |
| 799 | } |
| 800 | ientry->backend_data = NULL; |
| 801 | } |
| 802 | sfree(conf.fsect); |
| 803 | } |
| 804 | |
| 805 | static int xhtml_para_level(paragraph *p) |
| 806 | { |
| 807 | switch (p->type) |
| 808 | { |
| 809 | case para_Title: |
| 810 | return 0; |
| 811 | break; |
| 812 | case para_UnnumberedChapter: |
| 813 | case para_Chapter: |
| 814 | case para_Appendix: |
| 815 | return 1; |
| 816 | break; |
| 817 | /* case para_BiblioCited: |
| 818 | return 2; |
| 819 | break;*/ |
| 820 | case para_Heading: |
| 821 | case para_Subsect: |
| 822 | return p->aux+2; |
| 823 | break; |
| 824 | default: |
| 825 | return -1; |
| 826 | break; |
| 827 | } |
| 828 | } |
| 829 | |
| 830 | /* Output the nav links for the current file. |
| 831 | * file == NULL means we're doing the index |
| 832 | */ |
| 833 | static void xhtml_donavlinks(FILE *fp, xhtmlfile *file) |
| 834 | { |
| 835 | xhtmlfile *xhtml_next_file = NULL; |
| 836 | fprintf(fp, "<p"); |
| 837 | if (conf.nav_attrs!=NULL) { |
| 838 | fprintf(fp, " %ls>", conf.nav_attrs); |
| 839 | } else { |
| 840 | fprintf(fp, ">"); |
| 841 | } |
| 842 | if (xhtml_last_file==NULL) { |
| 843 | fprintf(fp, "Previous | "); |
| 844 | } else { |
| 845 | fprintf(fp, "<a href='%s'>Previous</a> | ", xhtml_last_file->filename); |
| 846 | } |
| 847 | fprintf(fp, "<a href='%s'>Contents</a> | ", conf.contents_filename); |
| 848 | if (file == NULL) { |
| 849 | fprintf(fp, "Index | "); |
| 850 | } else { |
| 851 | fprintf(fp, "<a href='%s'>Index</a> | ", conf.index_filename); |
| 852 | } |
| 853 | if (file != NULL) { /* otherwise we're doing nav links for the index */ |
| 854 | if (xhtml_next_file==NULL) |
| 855 | xhtml_next_file = file->child; |
| 856 | if (xhtml_next_file==NULL) |
| 857 | xhtml_next_file = file->next; |
| 858 | if (xhtml_next_file==NULL) |
| 859 | xhtml_next_file = file->parent->next; |
| 860 | } |
| 861 | if (xhtml_next_file==NULL) { |
| 862 | if (file==NULL) { /* index, so no next file */ |
| 863 | fprintf(fp, "Next "); |
| 864 | } else { |
| 865 | fprintf(fp, "<a href='%s'>Next</a>", conf.index_filename); |
| 866 | } |
| 867 | } else { |
| 868 | fprintf(fp, "<a href='%s'>Next</a>", xhtml_next_file->filename); |
| 869 | } |
| 870 | fprintf(fp, "</p>\n"); |
| 871 | } |
| 872 | |
| 873 | /* Write out the index file */ |
| 874 | static void xhtml_do_index_body(FILE *fp) |
| 875 | { |
| 876 | indexentry *y; |
| 877 | int ti; |
| 878 | |
| 879 | if (count234(idx->entries) == 0) |
| 880 | return; /* don't write anything at all */ |
| 881 | |
| 882 | fprintf(fp, "<dl>\n"); |
| 883 | /* iterate over idx->entries using the tree functions and display everything */ |
| 884 | for (ti = 0; (y = (indexentry *)index234(idx->entries, ti)) != NULL; ti++) { |
| 885 | if (y->backend_data) { |
| 886 | int i; |
| 887 | xhtmlindex *xi; |
| 888 | |
| 889 | fprintf(fp, "<dt>"); |
| 890 | xhtml_para(fp, y->text, FALSE); |
| 891 | fprintf(fp, "</dt>\n<dd>"); |
| 892 | |
| 893 | xi = (xhtmlindex*) y->backend_data; |
| 894 | for (i=0; i<xi->nsection; i++) { |
| 895 | xhtmlsection *sect = xi->sections[i]; |
| 896 | if (sect) { |
| 897 | fprintf(fp, "<a href='%s#%s'>", sect->file->filename, sect->fragment); |
| 898 | if (sect->para->kwtext) { |
| 899 | xhtml_para(fp, sect->para->kwtext, FALSE); |
| 900 | } else if (sect->para->words) { |
| 901 | xhtml_para(fp, sect->para->words, FALSE); |
| 902 | } |
| 903 | fprintf(fp, "</a>"); |
| 904 | if (i+1<xi->nsection) { |
| 905 | fprintf(fp, ", "); |
| 906 | } |
| 907 | } |
| 908 | } |
| 909 | fprintf(fp, "</dd>\n"); |
| 910 | } |
| 911 | } |
| 912 | fprintf(fp, "</dl>\n"); |
| 913 | } |
| 914 | static void xhtml_do_index() |
| 915 | { |
| 916 | word temp_word = { NULL, NULL, word_Normal, 0, 0, L"Index", { NULL, 0, 0} }; |
| 917 | FILE *fp = fopen(conf.index_filename, "w"); |
| 918 | |
| 919 | if (fp==NULL) |
| 920 | fatal(err_cantopenw, conf.index_filename); |
| 921 | xhtml_doheader(fp, &temp_word); |
| 922 | xhtml_donavlinks(fp, NULL); |
| 923 | |
| 924 | xhtml_do_index_body(fp); |
| 925 | |
| 926 | xhtml_donavlinks(fp, NULL); |
| 927 | xhtml_dofooter(fp); |
| 928 | fclose(fp); |
| 929 | } |
| 930 | |
| 931 | /* Output the given file. This includes whatever contents at beginning and end, etc. etc. */ |
| 932 | static void xhtml_do_file(xhtmlfile *file) |
| 933 | { |
| 934 | FILE *fp = fopen(file->filename, "w"); |
| 935 | if (fp==NULL) |
| 936 | fatal(err_cantopenw, file->filename); |
| 937 | |
| 938 | if (file->sections->para->words) { |
| 939 | xhtml_doheader(fp, file->sections->para->words); |
| 940 | } else if (file->sections->para->kwtext) { |
| 941 | xhtml_doheader(fp, file->sections->para->kwtext); |
| 942 | } else { |
| 943 | xhtml_doheader(fp, NULL); |
| 944 | } |
| 945 | |
| 946 | xhtml_donavlinks(fp, file); |
| 947 | |
| 948 | if (file->is_leaf && conf.leaf_contains_contents && |
| 949 | xhtml_do_contents(NULL, file)>=conf.leaf_smallest_contents) |
| 950 | xhtml_do_contents(fp, file); |
| 951 | xhtml_do_sections(fp, file->sections); |
| 952 | if (!file->is_leaf) |
| 953 | xhtml_do_naked_contents(fp, file); |
| 954 | |
| 955 | xhtml_donavlinks(fp, file); |
| 956 | |
| 957 | xhtml_dofooter(fp); |
| 958 | fclose(fp); |
| 959 | |
| 960 | xhtml_last_file = file; |
| 961 | } |
| 962 | |
| 963 | /* Output the top-level file. */ |
| 964 | static void xhtml_do_top_file(xhtmlfile *file, paragraph *sourceform) |
| 965 | { |
| 966 | paragraph *p; |
| 967 | int done=FALSE; |
| 968 | FILE *fp = fopen(file->filename, "w"); |
| 969 | if (fp==NULL) |
| 970 | fatal(err_cantopenw, file->filename); |
| 971 | |
| 972 | /* Do the title -- only one allowed */ |
| 973 | for (p = sourceform; p && !done; p = p->next) |
| 974 | { |
| 975 | if (p->type == para_Title) |
| 976 | { |
| 977 | xhtml_doheader(fp, p->words); |
| 978 | done=TRUE; |
| 979 | } |
| 980 | } |
| 981 | if (!done) |
| 982 | xhtml_doheader(fp, NULL /* Eek! */); |
| 983 | |
| 984 | /* |
| 985 | * Display the title. |
| 986 | */ |
| 987 | for (p = sourceform; p; p = p->next) |
| 988 | { |
| 989 | if (p->type == para_Title) { |
| 990 | xhtml_heading(fp, p, FALSE); |
| 991 | break; |
| 992 | } |
| 993 | } |
| 994 | |
| 995 | /* Do the preamble */ |
| 996 | for (p = sourceform; p; p = p->next) |
| 997 | { |
| 998 | if (p->type == para_Chapter || p->type == para_Heading || |
| 999 | p->type == para_Subsect || p->type == para_Appendix || |
| 1000 | p->type == para_UnnumberedChapter) { |
| 1001 | /* |
| 1002 | * We've found the end of the preamble. Do every normal |
| 1003 | * paragraph up to there. |
| 1004 | */ |
| 1005 | xhtml_do_paras(fp, sourceform, p, FALSE); |
| 1006 | break; |
| 1007 | } |
| 1008 | } |
| 1009 | |
| 1010 | xhtml_do_contents(fp, file); |
| 1011 | xhtml_do_sections(fp, file->sections); |
| 1012 | |
| 1013 | /* |
| 1014 | * Put the index in the top file if we're in single-file mode |
| 1015 | * (leaf-level 0). |
| 1016 | */ |
| 1017 | if (conf.leaf_level == 0 && count234(idx->entries) > 0) { |
| 1018 | fprintf(fp, "<a name=\"index\"></a><h1>Index</h1>\n"); |
| 1019 | xhtml_do_index_body(fp); |
| 1020 | } |
| 1021 | |
| 1022 | xhtml_dofooter(fp); |
| 1023 | fclose(fp); |
| 1024 | } |
| 1025 | |
| 1026 | /* Convert a Unicode string to an ASCII one. '?' is |
| 1027 | * used for unmappable characters. |
| 1028 | */ |
| 1029 | static void xhtml_utostr(wchar_t *in, char **out) |
| 1030 | { |
| 1031 | int l = ustrlen(in); |
| 1032 | int i; |
| 1033 | *out = smalloc(l+1); |
| 1034 | for (i=0; i<l; i++) |
| 1035 | { |
| 1036 | if (in[i]>=32 && in[i]<=126) |
| 1037 | (*out)[i]=(char)in[i]; |
| 1038 | else |
| 1039 | (*out)[i]='?'; |
| 1040 | } |
| 1041 | (*out)[i]=0; |
| 1042 | } |
| 1043 | |
| 1044 | /* |
| 1045 | * Write contents for the given file, and subfiles, down to |
| 1046 | * the appropriate contents depth. Returns the number of |
| 1047 | * entries written. |
| 1048 | */ |
| 1049 | static int xhtml_do_contents(FILE *fp, xhtmlfile *file) |
| 1050 | { |
| 1051 | int level, limit, count = 0; |
| 1052 | if (!file) |
| 1053 | return 0; |
| 1054 | |
| 1055 | level = (file->sections)?(file->sections->level):(0); |
| 1056 | limit = conf.contents_depth[(level>5)?(5):(level)]; |
| 1057 | start_level = (file->is_leaf) ? (level-1) : (level); |
| 1058 | last_level = start_level; |
| 1059 | |
| 1060 | count += xhtml_do_contents_section_limit(fp, file->sections, limit); |
| 1061 | count += xhtml_do_contents_limit(fp, file->child, limit); |
| 1062 | if (fp!=NULL) { |
| 1063 | while (last_level > start_level) { |
| 1064 | last_level--; |
| 1065 | fprintf(fp, "</li></ul>\n"); |
| 1066 | } |
| 1067 | } |
| 1068 | return count; |
| 1069 | } |
| 1070 | |
| 1071 | /* As above, but doesn't do anything in the current file */ |
| 1072 | static int xhtml_do_naked_contents(FILE *fp, xhtmlfile *file) |
| 1073 | { |
| 1074 | int level, limit, start_level, count = 0; |
| 1075 | if (!file) |
| 1076 | return 0; |
| 1077 | |
| 1078 | level = (file->sections)?(file->sections->level):(0); |
| 1079 | limit = conf.contents_depth[(level>5)?(5):(level)]; |
| 1080 | start_level = (file->is_leaf) ? (level-1) : (level); |
| 1081 | last_level = start_level; |
| 1082 | |
| 1083 | count = xhtml_do_contents_limit(fp, file->child, limit); |
| 1084 | if (fp!=NULL) { |
| 1085 | while (last_level > start_level) { |
| 1086 | last_level--; |
| 1087 | fprintf(fp, "</li></ul>\n"); |
| 1088 | } |
| 1089 | } |
| 1090 | return count; |
| 1091 | } |
| 1092 | |
| 1093 | /* |
| 1094 | * Write contents for the given file, children, and siblings, down to |
| 1095 | * given limit contents depth. |
| 1096 | */ |
| 1097 | static int xhtml_do_contents_limit(FILE *fp, xhtmlfile *file, int limit) |
| 1098 | { |
| 1099 | int count = 0; |
| 1100 | while (file) { |
| 1101 | count += xhtml_do_contents_section_limit(fp, file->sections, limit); |
| 1102 | count += xhtml_do_contents_limit(fp, file->child, limit); |
| 1103 | file = file->next; |
| 1104 | } |
| 1105 | return count; |
| 1106 | } |
| 1107 | |
| 1108 | /* |
| 1109 | * Write contents entries for the given section tree, down to the |
| 1110 | * limit contents depth. |
| 1111 | */ |
| 1112 | static int xhtml_do_contents_section_deep_limit(FILE *fp, xhtmlsection *section, int limit) |
| 1113 | { |
| 1114 | int count = 0; |
| 1115 | while (section) { |
| 1116 | if (!xhtml_add_contents_entry(fp, section, limit)) |
| 1117 | return 0; |
| 1118 | else |
| 1119 | count++; |
| 1120 | count += xhtml_do_contents_section_deep_limit(fp, section->child, limit); |
| 1121 | section = section->next; |
| 1122 | } |
| 1123 | return count; |
| 1124 | } |
| 1125 | |
| 1126 | /* |
| 1127 | * Write contents entries for the given section tree, down to the |
| 1128 | * limit contents depth. |
| 1129 | */ |
| 1130 | static int xhtml_do_contents_section_limit(FILE *fp, xhtmlsection *section, int limit) |
| 1131 | { |
| 1132 | int count = 0; |
| 1133 | if (!section) |
| 1134 | return 0; |
| 1135 | xhtml_add_contents_entry(fp, section, limit); |
| 1136 | count=1; |
| 1137 | count += xhtml_do_contents_section_deep_limit(fp, section->child, limit); |
| 1138 | /* section=section->child; |
| 1139 | while (section && xhtml_add_contents_entry(fp, section, limit)) { |
| 1140 | section = section->next; |
| 1141 | }*/ |
| 1142 | return count; |
| 1143 | } |
| 1144 | |
| 1145 | /* |
| 1146 | * Add a section entry, unless we're exceeding the limit, in which |
| 1147 | * case return FALSE (otherwise return TRUE). |
| 1148 | */ |
| 1149 | static int xhtml_add_contents_entry(FILE *fp, xhtmlsection *section, int limit) |
| 1150 | { |
| 1151 | if (!section || section->level > limit) |
| 1152 | return FALSE; |
| 1153 | if (fp==NULL || section->level < 0) |
| 1154 | return TRUE; |
| 1155 | if (last_level > section->level) { |
| 1156 | while (last_level > section->level) { |
| 1157 | last_level--; |
| 1158 | fprintf(fp, "</li></ul>\n"); |
| 1159 | } |
| 1160 | fprintf(fp, "</li>\n"); |
| 1161 | } else if (last_level < section->level) { |
| 1162 | assert(last_level == section->level - 1); |
| 1163 | last_level++; |
| 1164 | fprintf(fp, "<ul>\n"); |
| 1165 | } else { |
| 1166 | fprintf(fp, "</li>\n"); |
| 1167 | } |
| 1168 | fprintf(fp, "<li><a href=\"%s#%s\">", section->file->filename, section->fragment); |
| 1169 | if (section->para->kwtext) { |
| 1170 | xhtml_para(fp, section->para->kwtext, FALSE); |
| 1171 | if (section->para->words) { |
| 1172 | fprintf(fp, ": "); |
| 1173 | } |
| 1174 | } |
| 1175 | if (section->para->words) { |
| 1176 | xhtml_para(fp, section->para->words, FALSE); |
| 1177 | } |
| 1178 | fprintf(fp, "</a>\n"); |
| 1179 | return TRUE; |
| 1180 | } |
| 1181 | |
| 1182 | /* |
| 1183 | * Write all the sections in this file. Do all paragraphs in this section, then all |
| 1184 | * children (recursively), then go on to the next one (tail recursively). |
| 1185 | */ |
| 1186 | static void xhtml_do_sections(FILE *fp, xhtmlsection *sections) |
| 1187 | { |
| 1188 | while (sections) { |
| 1189 | currentsection = sections; |
| 1190 | xhtml_do_paras(fp, sections->para, NULL, TRUE); |
| 1191 | xhtml_do_sections(fp, sections->child); |
| 1192 | sections = sections->next; |
| 1193 | } |
| 1194 | } |
| 1195 | |
| 1196 | /* Write this list of paragraphs. Close off all lists at the end. */ |
| 1197 | static void xhtml_do_paras(FILE *fp, paragraph *p, paragraph *end, |
| 1198 | int indexable) |
| 1199 | { |
| 1200 | int last_type = -1, ptype, first=TRUE; |
| 1201 | stack lcont_stack = stk_new(); |
| 1202 | if (!p) |
| 1203 | return; |
| 1204 | |
| 1205 | /* for (; p && (xhtml_para_level(p)>limit || xhtml_para_level(p)==-1 || first); p=p->next) {*/ |
| 1206 | for (; p && p != end && (xhtml_para_level(p)==-1 || first); p=p->next) { |
| 1207 | first=FALSE; |
| 1208 | switch (ptype = p->type) |
| 1209 | { |
| 1210 | /* |
| 1211 | * Things we ignore because we've already processed them or |
| 1212 | * aren't going to touch them in this pass. |
| 1213 | */ |
| 1214 | case para_IM: |
| 1215 | case para_BR: |
| 1216 | case para_Biblio: /* only touch BiblioCited */ |
| 1217 | case para_VersionID: |
| 1218 | case para_NoCite: |
| 1219 | case para_Title: |
| 1220 | break; |
| 1221 | |
| 1222 | /* |
| 1223 | * Chapter titles. |
| 1224 | */ |
| 1225 | case para_Chapter: |
| 1226 | case para_Appendix: |
| 1227 | case para_UnnumberedChapter: |
| 1228 | xhtml_heading(fp, p, indexable); |
| 1229 | break; |
| 1230 | |
| 1231 | case para_Heading: |
| 1232 | case para_Subsect: |
| 1233 | xhtml_heading(fp, p, indexable); |
| 1234 | break; |
| 1235 | |
| 1236 | case para_Rule: |
| 1237 | fprintf(fp, "\n<hr />\n"); |
| 1238 | break; |
| 1239 | |
| 1240 | case para_Normal: |
| 1241 | case para_Copyright: |
| 1242 | fprintf(fp, "\n<p>"); |
| 1243 | xhtml_para(fp, p->words, indexable); |
| 1244 | fprintf(fp, "</p>\n"); |
| 1245 | break; |
| 1246 | |
| 1247 | case para_LcontPush: |
| 1248 | { |
| 1249 | int *p; |
| 1250 | p = mknew(int); |
| 1251 | *p = last_type; |
| 1252 | stk_push(lcont_stack, p); |
| 1253 | last_type = para_Normal; |
| 1254 | } |
| 1255 | break; |
| 1256 | case para_LcontPop: |
| 1257 | { |
| 1258 | int *p = stk_pop(lcont_stack); |
| 1259 | assert(p); |
| 1260 | ptype = last_type = *p; |
| 1261 | sfree(p); |
| 1262 | goto closeofflist; /* ick */ |
| 1263 | } |
| 1264 | break; |
| 1265 | case para_QuotePush: |
| 1266 | fprintf(fp, "<blockquote>\n"); |
| 1267 | break; |
| 1268 | case para_QuotePop: |
| 1269 | fprintf(fp, "</blockquote>\n"); |
| 1270 | break; |
| 1271 | |
| 1272 | case para_Bullet: |
| 1273 | case para_NumberedList: |
| 1274 | case para_Description: |
| 1275 | case para_DescribedThing: |
| 1276 | case para_BiblioCited: |
| 1277 | if (last_type!=p->type && |
| 1278 | !(last_type==para_DescribedThing && p->type==para_Description) && |
| 1279 | !(last_type==para_Description && p->type==para_DescribedThing)) { |
| 1280 | /* start up list if necessary */ |
| 1281 | if (p->type == para_Bullet) { |
| 1282 | fprintf(fp, "<ul>\n"); |
| 1283 | } else if (p->type == para_NumberedList) { |
| 1284 | fprintf(fp, "<ol>\n"); |
| 1285 | } else if (p->type == para_BiblioCited || |
| 1286 | p->type == para_DescribedThing || |
| 1287 | p->type == para_Description) { |
| 1288 | fprintf(fp, "<dl>\n"); |
| 1289 | } |
| 1290 | } |
| 1291 | if (p->type == para_Bullet || p->type == para_NumberedList) { |
| 1292 | fprintf(fp, "<li>"); |
| 1293 | } else if (p->type == para_DescribedThing) { |
| 1294 | fprintf(fp, "<dt>"); |
| 1295 | } else if (p->type == para_Description) { |
| 1296 | fprintf(fp, "<dd>"); |
| 1297 | } else if (p->type == para_BiblioCited) { |
| 1298 | fprintf(fp, "<dt>"); |
| 1299 | xhtml_para(fp, p->kwtext, indexable); |
| 1300 | fprintf(fp, "</dt>\n<dd>"); |
| 1301 | } |
| 1302 | xhtml_para(fp, p->words, indexable); |
| 1303 | { |
| 1304 | paragraph *p2 = p->next; |
| 1305 | if (p2 && xhtml_para_level(p2)==-1 && p2->type == para_LcontPush) |
| 1306 | break; |
| 1307 | } |
| 1308 | |
| 1309 | closeofflist: |
| 1310 | if (ptype == para_BiblioCited) { |
| 1311 | fprintf(fp, "</dd>\n"); |
| 1312 | } else if (ptype == para_DescribedThing) { |
| 1313 | fprintf(fp, "</dt>"); |
| 1314 | } else if (ptype == para_Description) { |
| 1315 | fprintf(fp, "</dd>"); |
| 1316 | } else if (ptype == para_Bullet || ptype == para_NumberedList) { |
| 1317 | fprintf(fp, "</li>"); |
| 1318 | } |
| 1319 | if (ptype == para_Bullet || ptype == para_NumberedList || |
| 1320 | ptype == para_BiblioCited || ptype == para_Description || |
| 1321 | ptype == para_DescribedThing) |
| 1322 | /* close off list if necessary */ |
| 1323 | { |
| 1324 | paragraph *p2 = p->next; |
| 1325 | int close_off=FALSE; |
| 1326 | /* if (p2 && (xhtml_para_level(p2)>limit || xhtml_para_level(p2)==-1)) {*/ |
| 1327 | if (p2 && xhtml_para_level(p2)==-1) { |
| 1328 | if (p2->type != ptype && |
| 1329 | !(p2->type==para_DescribedThing && ptype==para_Description) && |
| 1330 | !(p2->type==para_Description && ptype==para_DescribedThing) && |
| 1331 | p2->type != para_LcontPush) |
| 1332 | close_off=TRUE; |
| 1333 | } else { |
| 1334 | close_off=TRUE; |
| 1335 | } |
| 1336 | if (close_off) { |
| 1337 | if (ptype == para_Bullet) { |
| 1338 | fprintf(fp, "</ul>\n"); |
| 1339 | } else if (ptype == para_NumberedList) { |
| 1340 | fprintf(fp, "</ol>\n"); |
| 1341 | } else if (ptype == para_BiblioCited || |
| 1342 | ptype == para_Description || |
| 1343 | ptype == para_DescribedThing) { |
| 1344 | fprintf(fp, "</dl>\n"); |
| 1345 | } |
| 1346 | } |
| 1347 | } |
| 1348 | break; |
| 1349 | |
| 1350 | case para_Code: |
| 1351 | xhtml_codepara(fp, p->words); |
| 1352 | break; |
| 1353 | } |
| 1354 | last_type = ptype; |
| 1355 | } |
| 1356 | |
| 1357 | stk_free(lcont_stack); |
| 1358 | } |
| 1359 | |
| 1360 | /* |
| 1361 | * Output a header for this XHTML file. |
| 1362 | */ |
| 1363 | static void xhtml_doheader(FILE *fp, word *title) |
| 1364 | { |
| 1365 | fprintf(fp, "<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Transitional//EN\"\n"); |
| 1366 | fprintf(fp, "\"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd\">\n"); |
| 1367 | fprintf(fp, "<html xmlns='http://www.w3.org/1999/xhtml'>\n\n<head>\n<title>"); |
| 1368 | if (title==NULL) |
| 1369 | fprintf(fp, "The thing with no name!"); |
| 1370 | else |
| 1371 | xhtml_para(fp, title, FALSE); |
| 1372 | fprintf(fp, "</title>\n"); |
| 1373 | fprintf(fp, "<meta name=\"generator\" content=\"Halibut %s xhtml-backend\" />\n", version); |
| 1374 | if (conf.author) |
| 1375 | fprintf(fp, "<meta name=\"author\" content=\"%ls\" />\n", conf.author); |
| 1376 | if (conf.description) |
| 1377 | fprintf(fp, "<meta name=\"description\" content=\"%ls\" />\n", conf.description); |
| 1378 | if (conf.head_end) |
| 1379 | fprintf(fp, "%ls\n", conf.head_end); |
| 1380 | fprintf(fp, "</head>\n\n"); |
| 1381 | if (conf.body) |
| 1382 | fprintf(fp, "%ls\n", conf.body); |
| 1383 | else |
| 1384 | fprintf(fp, "<body>\n"); |
| 1385 | if (conf.body_start) |
| 1386 | fprintf(fp, "%ls\n", conf.body_start); |
| 1387 | } |
| 1388 | |
| 1389 | /* |
| 1390 | * Output a footer for this XHTML file. |
| 1391 | */ |
| 1392 | static void xhtml_dofooter(FILE *fp) |
| 1393 | { |
| 1394 | fprintf(fp, "\n<hr />\n\n"); |
| 1395 | if (conf.body_end) |
| 1396 | fprintf(fp, "%ls\n", conf.body_end); |
| 1397 | if (!conf.suppress_address) { |
| 1398 | fprintf(fp,"<address>\n"); |
| 1399 | if (conf.address_start) |
| 1400 | fprintf(fp, "%ls\n", conf.address_start); |
| 1401 | /* Do the version ID */ |
| 1402 | if (conf.include_version_id) { |
| 1403 | paragraph *p; |
| 1404 | int started = 0; |
| 1405 | for (p = sourceparas; p; p = p->next) |
| 1406 | if (p->type == para_VersionID) { |
| 1407 | xhtml_versionid(fp, p->words, started); |
| 1408 | started = 1; |
| 1409 | } |
| 1410 | } |
| 1411 | if (conf.address_end) |
| 1412 | fprintf(fp, "%ls\n", conf.address_end); |
| 1413 | fprintf(fp, "</address>\n"); |
| 1414 | } |
| 1415 | fprintf(fp, "</body>\n\n</html>\n"); |
| 1416 | } |
| 1417 | |
| 1418 | /* |
| 1419 | * Output the versionid paragraph. Typically this is a version control |
| 1420 | * ID string (such as $Id...$ in RCS). |
| 1421 | */ |
| 1422 | static void xhtml_versionid(FILE *fp, word *text, int started) |
| 1423 | { |
| 1424 | rdstringc t = { 0, 0, NULL }; |
| 1425 | |
| 1426 | rdaddc(&t, '['); /* FIXME: configurability */ |
| 1427 | xhtml_rdaddwc(&t, text, NULL, FALSE); |
| 1428 | rdaddc(&t, ']'); /* FIXME: configurability */ |
| 1429 | |
| 1430 | if (started) |
| 1431 | fprintf(fp, "<br />\n"); |
| 1432 | fprintf(fp, "%s\n", t.text); |
| 1433 | sfree(t.text); |
| 1434 | } |
| 1435 | |
| 1436 | /* Is this an XHTML reserved character? */ |
| 1437 | static int xhtml_reservedchar(int c) |
| 1438 | { |
| 1439 | if (c=='&' || c=='<' || c=='>' || c=='"') |
| 1440 | return TRUE; |
| 1441 | else |
| 1442 | return FALSE; |
| 1443 | } |
| 1444 | |
| 1445 | /* |
| 1446 | * Convert a wide string into valid XHTML: Anything outside ASCII will |
| 1447 | * be fixed up as an entity. Currently we don't worry about constraining the |
| 1448 | * encoded character set, which we should probably do at some point (we can |
| 1449 | * still fix up and return FALSE - see the last comment here). We also don't |
| 1450 | * currently |
| 1451 | * |
| 1452 | * Because this is only used for words, spaces are HARD spaces (any other |
| 1453 | * spaces will be word_Whitespace not word_Normal). So they become |
| 1454 | * Unless hard_spaces is FALSE, of course (code paragraphs break the above |
| 1455 | * rule). |
| 1456 | * |
| 1457 | * If `result' is non-NULL, mallocs the resulting string and stores a pointer to |
| 1458 | * it in `*result'. If `result' is NULL, merely checks whether all |
| 1459 | * characters in the string are feasible. |
| 1460 | * |
| 1461 | * Return is nonzero if all characters are OK. If not all |
| 1462 | * characters are OK but `result' is non-NULL, a result _will_ |
| 1463 | * still be generated! |
| 1464 | */ |
| 1465 | static int xhtml_convert(wchar_t *s, int maxlen, char **result, |
| 1466 | int hard_spaces) { |
| 1467 | int doing = (result != 0); |
| 1468 | int ok = TRUE; |
| 1469 | char *p = NULL; |
| 1470 | int plen = 0, psize = 0; |
| 1471 | |
| 1472 | if (maxlen <= 0) |
| 1473 | maxlen = -1; |
| 1474 | |
| 1475 | for (; *s && maxlen != 0; s++, maxlen--) { |
| 1476 | wchar_t c = *s; |
| 1477 | |
| 1478 | #define ensure_size(i) if (i>=psize) { psize = i+256; p = resize(p, psize); } |
| 1479 | |
| 1480 | if (((c == 32 && !hard_spaces) || (c > 32 && c <= 126 && !xhtml_reservedchar(c)))) { |
| 1481 | /* Char is OK. */ |
| 1482 | if (doing) |
| 1483 | { |
| 1484 | ensure_size(plen); |
| 1485 | p[plen++] = (char)c; |
| 1486 | } |
| 1487 | } else { |
| 1488 | /* Char needs fixing up. */ |
| 1489 | /* ok = FALSE; -- currently we never return FALSE; we |
| 1490 | * might want to when considering a character set for the |
| 1491 | * encoded document. |
| 1492 | */ |
| 1493 | if (doing) |
| 1494 | { |
| 1495 | if (c==32) { /* a space in a word is a hard space */ |
| 1496 | ensure_size(plen+6); /* includes space for the NUL, which is subsequently stomped on */ |
| 1497 | sprintf(p+plen, " "); |
| 1498 | plen+=6; |
| 1499 | } else { |
| 1500 | /* FIXME: entity names! */ |
| 1501 | ensure_size(plen+8); /* includes space for the NUL, which is subsequently stomped on */ |
| 1502 | plen+=sprintf(p+plen, "&#%04i;", (int)c); |
| 1503 | } |
| 1504 | } |
| 1505 | } |
| 1506 | } |
| 1507 | if (doing) { |
| 1508 | p = resize(p, plen+1); |
| 1509 | p[plen] = '\0'; |
| 1510 | *result = p; |
| 1511 | } |
| 1512 | return ok; |
| 1513 | } |
| 1514 | |
| 1515 | /* |
| 1516 | * This formats the given words as XHTML. |
| 1517 | * |
| 1518 | * `indexable', if FALSE, prohibits adding any index references. |
| 1519 | * You might use this, for example, if an index reference occurred |
| 1520 | * in a section title, to prevent phony index references when the |
| 1521 | * section title is processed in strange places such as contents |
| 1522 | * sections. |
| 1523 | */ |
| 1524 | static void xhtml_rdaddwc(rdstringc *rs, word *text, word *end, int indexable) { |
| 1525 | char *c; |
| 1526 | keyword *kwl; |
| 1527 | xhtmlsection *sect; |
| 1528 | indextag *itag; |
| 1529 | int ti; |
| 1530 | |
| 1531 | for (; text && text != end; text = text->next) { |
| 1532 | switch (text->type) { |
| 1533 | case word_HyperLink: |
| 1534 | xhtml_utostr(text->text, &c); |
| 1535 | rdaddsc(rs, "<a href=\""); |
| 1536 | rdaddsc(rs, c); |
| 1537 | rdaddsc(rs, "\">"); |
| 1538 | sfree(c); |
| 1539 | break; |
| 1540 | |
| 1541 | case word_UpperXref: |
| 1542 | case word_LowerXref: |
| 1543 | kwl = kw_lookup(keywords, text->text); |
| 1544 | if (kwl) { |
| 1545 | sect=xhtml_find_section(kwl->para); |
| 1546 | if (sect) { |
| 1547 | rdaddsc(rs, "<a href=\""); |
| 1548 | rdaddsc(rs, sect->file->filename); |
| 1549 | rdaddc(rs, '#'); |
| 1550 | rdaddsc(rs, sect->fragment); |
| 1551 | rdaddsc(rs, "\">"); |
| 1552 | } else { |
| 1553 | rdaddsc(rs, "<a href=\"Apologies.html\"><!-- probably a bibliography cross reference -->"); |
| 1554 | error(err_whatever, "Couldn't locate cross-reference! (Probably a bibliography entry.)"); |
| 1555 | } |
| 1556 | } else { |
| 1557 | rdaddsc(rs, "<a href=\"Apologies.html\"><!-- unknown cross-reference -->"); |
| 1558 | error(err_whatever, "Couldn't locate cross-reference! (Wasn't in source file.)"); |
| 1559 | } |
| 1560 | break; |
| 1561 | |
| 1562 | case word_IndexRef: /* in theory we could make an index target here */ |
| 1563 | /* rdaddsc(rs, "<a name=\"idx-"); |
| 1564 | xhtml_utostr(text->text, &c); |
| 1565 | rdaddsc(rs, c); |
| 1566 | sfree(c); |
| 1567 | rdaddsc(rs, "\"></a>");*/ |
| 1568 | /* what we _do_ need to do is to fix up the backend data |
| 1569 | * for any indexentry this points to. |
| 1570 | */ |
| 1571 | if (!indexable) |
| 1572 | break; |
| 1573 | |
| 1574 | for (ti=0; (itag = (indextag *)index234(idx->tags, ti))!=NULL; ti++) { |
| 1575 | /* FIXME: really ustricmp() and not ustrcmp()? */ |
| 1576 | if (ustricmp(itag->name, text->text)==0) { |
| 1577 | break; |
| 1578 | } |
| 1579 | } |
| 1580 | if (itag!=NULL) { |
| 1581 | if (itag->refs!=NULL) { |
| 1582 | int i; |
| 1583 | for (i=0; i<itag->nrefs; i++) { |
| 1584 | xhtmlindex *idx_ref; |
| 1585 | indexentry *ientry; |
| 1586 | |
| 1587 | ientry = itag->refs[i]; |
| 1588 | if (ientry->backend_data==NULL) { |
| 1589 | idx_ref = (xhtmlindex*) smalloc(sizeof(xhtmlindex)); |
| 1590 | if (idx_ref==NULL) |
| 1591 | fatal(err_nomemory); |
| 1592 | idx_ref->nsection = 0; |
| 1593 | idx_ref->size = 4; |
| 1594 | idx_ref->sections = (xhtmlsection**) smalloc(idx_ref->size * sizeof(xhtmlsection*)); |
| 1595 | if (idx_ref->sections==NULL) |
| 1596 | fatal(err_nomemory); |
| 1597 | ientry->backend_data = idx_ref; |
| 1598 | } else { |
| 1599 | idx_ref = ientry->backend_data; |
| 1600 | if (idx_ref->nsection+1 > idx_ref->size) { |
| 1601 | int new_size = idx_ref->size * 2; |
| 1602 | idx_ref->sections = srealloc(idx_ref->sections, new_size * sizeof(xhtmlsection)); |
| 1603 | if (idx_ref->sections==NULL) { |
| 1604 | fatal(err_nomemory); |
| 1605 | } |
| 1606 | idx_ref->size = new_size; |
| 1607 | } |
| 1608 | } |
| 1609 | idx_ref->sections[idx_ref->nsection++] = currentsection; |
| 1610 | #if 0 |
| 1611 | #endif |
| 1612 | } |
| 1613 | } else { |
| 1614 | fatal(err_whatever, "Index tag had no entries!"); |
| 1615 | } |
| 1616 | } else { |
| 1617 | fprintf(stderr, "Looking for index entry '%ls'\n", text->text); |
| 1618 | fatal(err_whatever, "Couldn't locate index entry! (Wasn't in index.)"); |
| 1619 | } |
| 1620 | break; |
| 1621 | |
| 1622 | case word_HyperEnd: |
| 1623 | case word_XrefEnd: |
| 1624 | rdaddsc(rs, "</a>"); |
| 1625 | break; |
| 1626 | |
| 1627 | case word_Normal: |
| 1628 | case word_Emph: |
| 1629 | case word_Code: |
| 1630 | case word_WeakCode: |
| 1631 | case word_WhiteSpace: |
| 1632 | case word_EmphSpace: |
| 1633 | case word_CodeSpace: |
| 1634 | case word_WkCodeSpace: |
| 1635 | case word_Quote: |
| 1636 | case word_EmphQuote: |
| 1637 | case word_CodeQuote: |
| 1638 | case word_WkCodeQuote: |
| 1639 | assert(text->type != word_CodeQuote && |
| 1640 | text->type != word_WkCodeQuote); |
| 1641 | if (towordstyle(text->type) == word_Emph && |
| 1642 | (attraux(text->aux) == attr_First || |
| 1643 | attraux(text->aux) == attr_Only)) |
| 1644 | rdaddsc(rs, "<em>"); |
| 1645 | else if ((towordstyle(text->type) == word_Code || towordstyle(text->type) == word_WeakCode) && |
| 1646 | (attraux(text->aux) == attr_First || |
| 1647 | attraux(text->aux) == attr_Only)) |
| 1648 | rdaddsc(rs, "<code>"); |
| 1649 | |
| 1650 | if (removeattr(text->type) == word_Normal) { |
| 1651 | if (xhtml_convert(text->text, 0, &c, TRUE)) /* spaces in the word are hard */ |
| 1652 | rdaddsc(rs, c); |
| 1653 | else |
| 1654 | xhtml_rdaddwc(rs, text->alt, NULL, indexable); |
| 1655 | sfree(c); |
| 1656 | } else if (removeattr(text->type) == word_WhiteSpace) { |
| 1657 | rdaddc(rs, ' '); |
| 1658 | } else if (removeattr(text->type) == word_Quote) { |
| 1659 | rdaddsc(rs, """); |
| 1660 | } |
| 1661 | |
| 1662 | if (towordstyle(text->type) == word_Emph && |
| 1663 | (attraux(text->aux) == attr_Last || |
| 1664 | attraux(text->aux) == attr_Only)) |
| 1665 | rdaddsc(rs, "</em>"); |
| 1666 | else if ((towordstyle(text->type) == word_Code || towordstyle(text->type) == word_WeakCode) && |
| 1667 | (attraux(text->aux) == attr_Last || |
| 1668 | attraux(text->aux) == attr_Only)) |
| 1669 | rdaddsc(rs, "</code>"); |
| 1670 | break; |
| 1671 | } |
| 1672 | } |
| 1673 | } |
| 1674 | |
| 1675 | /* Output a heading, formatted as XHTML. |
| 1676 | */ |
| 1677 | static void xhtml_heading(FILE *fp, paragraph *p, int indexable) |
| 1678 | { |
| 1679 | rdstringc t = { 0, 0, NULL }; |
| 1680 | word *tprefix = p->kwtext; |
| 1681 | word *nprefix = p->kwtext2; |
| 1682 | word *text = p->words; |
| 1683 | int level = xhtml_para_level(p); |
| 1684 | xhtmlsection *sect = xhtml_find_section(p); |
| 1685 | xhtmlheadfmt *fmt; |
| 1686 | char *fragment; |
| 1687 | if (sect) { |
| 1688 | fragment = sect->fragment; |
| 1689 | } else { |
| 1690 | if (p->type == para_Title) |
| 1691 | fragment = "title"; |
| 1692 | else { |
| 1693 | fragment = ""; /* FIXME: what else can we do? */ |
| 1694 | error(err_whatever, "Couldn't locate heading cross-reference!"); |
| 1695 | } |
| 1696 | } |
| 1697 | |
| 1698 | if (p->type == para_Title) |
| 1699 | fmt = NULL; |
| 1700 | else if (level == 1) |
| 1701 | fmt = &conf.fchapter; |
| 1702 | else if (level-1 < conf.nfsect) |
| 1703 | fmt = &conf.fsect[level-1]; |
| 1704 | else |
| 1705 | fmt = &conf.fsect[conf.nfsect-1]; |
| 1706 | |
| 1707 | if (fmt && fmt->just_numbers && nprefix) { |
| 1708 | xhtml_rdaddwc(&t, nprefix, NULL, indexable); |
| 1709 | if (fmt) { |
| 1710 | char *c; |
| 1711 | if (xhtml_convert(fmt->number_suffix, 0, &c, FALSE)) { |
| 1712 | rdaddsc(&t, c); |
| 1713 | sfree(c); |
| 1714 | } |
| 1715 | } |
| 1716 | } else if (fmt && !fmt->just_numbers && tprefix) { |
| 1717 | xhtml_rdaddwc(&t, tprefix, NULL, indexable); |
| 1718 | if (fmt) { |
| 1719 | char *c; |
| 1720 | if (xhtml_convert(fmt->number_suffix, 0, &c, FALSE)) { |
| 1721 | rdaddsc(&t, c); |
| 1722 | sfree(c); |
| 1723 | } |
| 1724 | } |
| 1725 | } |
| 1726 | xhtml_rdaddwc(&t, text, NULL, indexable); |
| 1727 | /* |
| 1728 | * If we're outputting in single-file mode, we need to lower |
| 1729 | * the level of each heading by one, because the overall |
| 1730 | * document title will be sitting right at the top as an <h1> |
| 1731 | * and so chapters and sections should start at <h2>. |
| 1732 | * |
| 1733 | * Even if not, the document title will come back from |
| 1734 | * xhtml_para_level() as level zero, so we must increment that |
| 1735 | * no matter what leaf_level is set to. |
| 1736 | */ |
| 1737 | if (conf.leaf_level == 0 || level == 0) |
| 1738 | level++; |
| 1739 | fprintf(fp, "<a name=\"%s\"></a><h%i>%s</h%i>\n", fragment, level, t.text, level); |
| 1740 | sfree(t.text); |
| 1741 | } |
| 1742 | |
| 1743 | /* Output a paragraph. Styles are handled by xhtml_rdaddwc(). |
| 1744 | * This looks pretty simple; I may have missed something ... |
| 1745 | */ |
| 1746 | static void xhtml_para(FILE *fp, word *text, int indexable) |
| 1747 | { |
| 1748 | rdstringc out = { 0, 0, NULL }; |
| 1749 | xhtml_rdaddwc(&out, text, NULL, indexable); |
| 1750 | fprintf(fp, "%s", out.text); |
| 1751 | sfree(out.text); |
| 1752 | } |
| 1753 | |
| 1754 | /* Output a code paragraph. I'm treating this as preformatted, which |
| 1755 | * may not be entirely correct. See xhtml_para() for my worries about |
| 1756 | * this being overly-simple; however I think that most of the complexity |
| 1757 | * of the text backend came entirely out of word wrapping anyway. |
| 1758 | */ |
| 1759 | static void xhtml_codepara(FILE *fp, word *text) |
| 1760 | { |
| 1761 | fprintf(fp, "<pre>"); |
| 1762 | for (; text; text = text->next) if (text->type == word_WeakCode) { |
| 1763 | word *here, *next; |
| 1764 | char *c; |
| 1765 | |
| 1766 | /* |
| 1767 | * See if this WeakCode is followed by an Emph to indicate |
| 1768 | * emphasis. |
| 1769 | */ |
| 1770 | here = text; |
| 1771 | if (text->next && text->next->type == word_Emph) { |
| 1772 | next = text = text->next; |
| 1773 | } else |
| 1774 | next = NULL; |
| 1775 | |
| 1776 | if (next) { |
| 1777 | wchar_t *t, *e; |
| 1778 | int n; |
| 1779 | |
| 1780 | t = here->text; |
| 1781 | e = next->text; |
| 1782 | |
| 1783 | while (*e) { |
| 1784 | int ec = *e; |
| 1785 | |
| 1786 | for (n = 0; t[n] && e[n] && e[n] == ec; n++); |
| 1787 | xhtml_convert(t, n, &c, FALSE); |
| 1788 | fprintf(fp, "%s%s%s", |
| 1789 | (ec == 'i' ? "<em>" : ec == 'b' ? "<b>" : ""), |
| 1790 | c, |
| 1791 | (ec == 'i' ? "</em>" : ec == 'b' ? "</b>" : "")); |
| 1792 | sfree(c); |
| 1793 | |
| 1794 | t += n; |
| 1795 | e += n; |
| 1796 | } |
| 1797 | |
| 1798 | xhtml_convert(t, 0, &c, FALSE); |
| 1799 | fprintf(fp, "%s\n", c); |
| 1800 | sfree(c); |
| 1801 | } else { |
| 1802 | xhtml_convert(here->text, 0, &c, FALSE); |
| 1803 | fprintf(fp, "%s\n", c); |
| 1804 | sfree(c); |
| 1805 | } |
| 1806 | } |
| 1807 | fprintf(fp, "</pre>\n"); |
| 1808 | } |