02e2a328d394383458d5061fe5b45a4bdef836a7
[sgt/halibut] / bk_xhtml.c
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 struct xhtmlsection_Struct {
36 struct xhtmlsection_Struct *next; /* next sibling (NULL if split across files) */
37 struct xhtmlsection_Struct *child; /* NULL if split across files */
38 struct xhtmlsection_Struct *parent; /* NULL if split across files */
39 struct xhtmlsection_Struct *chain; /* single structure independent of weird trees */
40 paragraph *para;
41 struct xhtmlfile_Struct *file; /* which file is this a part of? */
42 char *fragment; /* fragment id within the file */
43 int level;
44 };
45
46 struct xhtmlfile_Struct {
47 struct xhtmlfile_Struct *next;
48 struct xhtmlfile_Struct *child;
49 struct xhtmlfile_Struct *parent;
50 char *filename;
51 struct xhtmlsection_Struct *sections; /* sections within this file (only one for non-leaf) */
52 int is_leaf; /* is this file a leaf file, ie does it not have any children? */
53 };
54
55 typedef struct xhtmlsection_Struct xhtmlsection;
56 typedef struct xhtmlfile_Struct xhtmlfile;
57 typedef struct xhtmlindex_Struct xhtmlindex;
58
59 struct xhtmlindex_Struct {
60 int nsection;
61 int size;
62 xhtmlsection **sections;
63 };
64
65 typedef struct {
66 int just_numbers;
67 wchar_t *number_suffix;
68 } xhtmlheadfmt;
69
70 typedef struct {
71 int contents_depth[6];
72 int leaf_contains_contents;
73 int leaf_level;
74 int leaf_smallest_contents;
75 int include_version_id;
76 wchar_t *author, *description;
77 wchar_t *head_end, *body, *body_start, *body_end, *address_start, *address_end, *nav_attrs;
78 int suppress_address;
79 xhtmlheadfmt fchapter, *fsect;
80 int nfsect;
81 } xhtmlconfig;
82
83 /*static void xhtml_level(paragraph *, int);
84 static void xhtml_level_0(paragraph *);
85 static void xhtml_docontents(FILE *, paragraph *, int);
86 static void xhtml_dosections(FILE *, paragraph *, int);
87 static void xhtml_dobody(FILE *, paragraph *, int);*/
88
89 static void xhtml_doheader(FILE *, word *);
90 static void xhtml_dofooter(FILE *);
91 static void xhtml_versionid(FILE *, word *, int);
92
93 static void xhtml_utostr(wchar_t *, char **);
94 static int xhtml_para_level(paragraph *);
95 static int xhtml_reservedchar(int);
96
97 static int xhtml_convert(wchar_t *, int, char **, int);
98 static void xhtml_rdaddwc(rdstringc *, word *, word *, int);
99 static void xhtml_para(FILE *, word *, int);
100 static void xhtml_codepara(FILE *, word *);
101 static void xhtml_heading(FILE *, paragraph *, int);
102
103 /* File-global variables are much easier than passing these things
104 * all over the place. Evil, but easier. We can replace this with a single
105 * structure at some point.
106 */
107 static xhtmlconfig conf;
108 static keywordlist *keywords;
109 static indexdata *idx;
110 static xhtmlfile *topfile;
111 static xhtmlsection *topsection;
112 static paragraph *sourceparas;
113 static xhtmlfile *lastfile;
114 static xhtmlfile *xhtml_last_file = NULL;
115 static int last_level=-1, start_level;
116 static xhtmlsection *currentsection;
117
118 static xhtmlconfig xhtml_configure(paragraph *source)
119 {
120 xhtmlconfig ret;
121
122 /*
123 * Defaults.
124 */
125 ret.contents_depth[0] = 2;
126 ret.contents_depth[1] = 3;
127 ret.contents_depth[2] = 4;
128 ret.contents_depth[3] = 5;
129 ret.contents_depth[4] = 6;
130 ret.contents_depth[5] = 7;
131 ret.leaf_level = 2;
132 ret.leaf_smallest_contents = 4;
133 ret.leaf_contains_contents = FALSE;
134 ret.include_version_id = TRUE;
135 ret.author = NULL;
136 ret.description = NULL;
137 ret.head_end = NULL;
138 ret.body = NULL;
139 ret.body_start = NULL;
140 ret.body_end = NULL;
141 ret.address_start = NULL;
142 ret.address_end = NULL;
143 ret.nav_attrs = NULL;
144 ret.suppress_address = FALSE;
145
146 ret.fchapter.just_numbers = FALSE;
147 ret.fchapter.number_suffix = ustrdup(L": ");
148 ret.nfsect = 2;
149 ret.fsect = mknewa(xhtmlheadfmt, ret.nfsect);
150 ret.fsect[0].just_numbers = FALSE;
151 ret.fsect[0].number_suffix = ustrdup(L": ");
152 ret.fsect[1].just_numbers = TRUE;
153 ret.fsect[1].number_suffix = ustrdup(L" ");
154
155 for (; source; source = source->next)
156 {
157 if (source->type == para_Config)
158 {
159 if (!ustricmp(source->keyword, L"xhtml-contents-depth-0")) {
160 ret.contents_depth[0] = utoi(uadv(source->keyword));
161 } else if (!ustricmp(source->keyword, L"xhtml-contents-depth-1")) {
162 ret.contents_depth[1] = utoi(uadv(source->keyword));
163 } else if (!ustricmp(source->keyword, L"xhtml-contents-depth-2")) {
164 ret.contents_depth[2] = utoi(uadv(source->keyword));
165 } else if (!ustricmp(source->keyword, L"xhtml-contents-depth-3")) {
166 ret.contents_depth[3] = utoi(uadv(source->keyword));
167 } else if (!ustricmp(source->keyword, L"xhtml-contents-depth-4")) {
168 ret.contents_depth[4] = utoi(uadv(source->keyword));
169 } else if (!ustricmp(source->keyword, L"xhtml-contents-depth-5")) {
170 ret.contents_depth[5] = utoi(uadv(source->keyword));
171 } else if (!ustricmp(source->keyword, L"xhtml-leaf-level")) {
172 ret.leaf_level = utoi(uadv(source->keyword));
173 } else if (!ustricmp(source->keyword, L"xhtml-leaf-smallest-contents")) {
174 ret.leaf_smallest_contents = utoi(uadv(source->keyword));
175 } else if (!ustricmp(source->keyword, L"xhtml-versionid")) {
176 ret.include_version_id = utob(uadv(source->keyword));
177 } else if (!ustricmp(source->keyword, L"xhtml-leaf-contains-contents")) {
178 ret.leaf_contains_contents = utob(uadv(source->keyword));
179 } else if (!ustricmp(source->keyword, L"xhtml-suppress-address")) {
180 ret.suppress_address = utob(uadv(source->keyword));
181 } else if (!ustricmp(source->keyword, L"xhtml-author")) {
182 ret.author = uadv(source->keyword);
183 } else if (!ustricmp(source->keyword, L"xhtml-description")) {
184 ret.description = uadv(source->keyword);
185 } else if (!ustricmp(source->keyword, L"xhtml-head-end")) {
186 ret.head_end = uadv(source->keyword);
187 } else if (!ustricmp(source->keyword, L"xhtml-body-start")) {
188 ret.body_start = uadv(source->keyword);
189 } else if (!ustricmp(source->keyword, L"xhtml-body-tag")) {
190 ret.body = uadv(source->keyword);
191 } else if (!ustricmp(source->keyword, L"xhtml-body-end")) {
192 ret.body_end = uadv(source->keyword);
193 } else if (!ustricmp(source->keyword, L"xhtml-address-start")) {
194 ret.address_start = uadv(source->keyword);
195 } else if (!ustricmp(source->keyword, L"xhtml-address-end")) {
196 ret.address_end = uadv(source->keyword);
197 } else if (!ustricmp(source->keyword, L"xhtml-navigation-attributes")) {
198 ret.nav_attrs = uadv(source->keyword);
199 } else if (!ustricmp(source->keyword, L"xhtml-chapter-numeric")) {
200 ret.fchapter.just_numbers = utob(uadv(source->keyword));
201 } else if (!ustricmp(source->keyword, L"xhtml-chapter-suffix")) {
202 ret.fchapter.number_suffix = ustrdup(uadv(source->keyword));
203 } else if (!ustricmp(source->keyword, L"xhtml-section-numeric")) {
204 wchar_t *p = uadv(source->keyword);
205 int n = 0;
206 if (uisdigit(*p)) {
207 n = utoi(p);
208 p = uadv(p);
209 }
210 if (n >= ret.nfsect) {
211 int i;
212 ret.fsect = resize(ret.fsect, n+1);
213 for (i = ret.nfsect; i <= n; i++)
214 ret.fsect[i] = ret.fsect[ret.nfsect-1];
215 ret.nfsect = n+1;
216 }
217 ret.fsect[n].just_numbers = utob(p);
218 } else if (!ustricmp(source->keyword, L"xhtml-section-suffix")) {
219 wchar_t *p = uadv(source->keyword);
220 int n = 0;
221 if (uisdigit(*p)) {
222 n = utoi(p);
223 p = uadv(p);
224 }
225 if (n >= ret.nfsect) {
226 int i;
227 ret.fsect = resize(ret.fsect, n+1);
228 for (i = ret.nfsect; i <= n; i++)
229 ret.fsect[i] = ret.fsect[ret.nfsect-1];
230 ret.nfsect = n+1;
231 }
232 ret.fsect[n].number_suffix = ustrdup(p);
233 }
234 }
235 }
236
237 /* printf(" !!! leaf_level = %i\n", ret.leaf_level);
238 printf(" !!! contentdepth-0 = %i\n", ret.contents_depth[0]);
239 printf(" !!! contentdepth-1 = %i\n", ret.contents_depth[1]);
240 printf(" !!! contentdepth-2 = %i\n", ret.contents_depth[2]);
241 printf(" !!! contentdepth-3 = %i\n", ret.contents_depth[3]);
242 printf(" !!! contentdepth-4 = %i\n", ret.contents_depth[4]);
243 printf(" !!! contentdepth-5 = %i\n", ret.contents_depth[5]);
244 printf(" !!! leaf_contains_contents = %i\n", ret.leaf_contains_contents);*/
245 return ret;
246 }
247
248 static xhtmlsection *xhtml_new_section(xhtmlsection *last)
249 {
250 xhtmlsection *ret = mknew(xhtmlsection);
251 ret->next=NULL;
252 ret->child=NULL;
253 ret->parent=NULL;
254 ret->chain=last;
255 ret->para=NULL;
256 ret->file=NULL;
257 ret->fragment=NULL;
258 ret->level=-1; /* marker: end of chain */
259 return ret;
260 }
261
262 /* Returns NULL or the section that marks that paragraph */
263 static xhtmlsection *xhtml_find_section(paragraph *p)
264 {
265 xhtmlsection *ret = topsection;
266 if (xhtml_para_level(p)==-1) { /* first, we back-track to a section paragraph */
267 paragraph *p2 = sourceparas;
268 paragraph *p3 = NULL;
269 while (p2 && p2!=p) {
270 if (xhtml_para_level(p2)!=-1) {
271 p3 = p2;
272 }
273 p2=p2->next;
274 }
275 if (p3==NULL) { /* for some reason, we couldn't find a section before this paragraph ... ? */
276 /* Note that this can happen, if you have a cross-reference to before the first chapter starts.
277 * So don't do that, then.
278 */
279 return NULL;
280 }
281 p=p3;
282 }
283 while (ret && ret->para != p) {
284 /* printf(" xhtml_find_section(): checking %s for para @ %p\n", ret->fragment, p);*/
285 ret=ret->chain;
286 }
287 return ret;
288 }
289
290 static xhtmlfile *xhtml_new_file(xhtmlsection *sect)
291 {
292 xhtmlfile *ret = mknew(xhtmlfile);
293
294 ret->next=NULL;
295 ret->child=NULL;
296 ret->parent=NULL;
297 ret->filename=NULL;
298 ret->sections=sect;
299 ret->is_leaf=(sect!=NULL && sect->level==conf.leaf_level);
300 if (sect==NULL) {
301 if (conf.leaf_level==0) { /* currently unused */
302 #define FILENAME_MANUAL "Manual.html"
303 #define FILENAME_CONTENTS "Contents.html"
304 ret->filename = smalloc(strlen(FILENAME_MANUAL)+1);
305 sprintf(ret->filename, FILENAME_MANUAL);
306 } else {
307 ret->filename = smalloc(strlen(FILENAME_CONTENTS)+1);
308 sprintf(ret->filename, FILENAME_CONTENTS);
309 }
310 } else {
311 paragraph *p = sect->para;
312 rdstringc fname_c = { 0, 0, NULL };
313 char *c;
314 word *w;
315 for (w=(p->kwtext)?(p->kwtext):(p->words); w; w=w->next)
316 {
317 switch (removeattr(w->type))
318 {
319 case word_Normal:
320 /*case word_Emph:
321 case word_Code:
322 case word_WeakCode:*/
323 xhtml_utostr(w->text, &c);
324 rdaddsc(&fname_c,c);
325 sfree(c);
326 break;
327 }
328 }
329 rdaddsc(&fname_c, ".html");
330 ret->filename = rdtrimc(&fname_c);
331 }
332 /* printf(" ! new file '%s', is_leaf == %s\n", ret->filename, (ret->is_leaf)?("true"):("false"));*/
333 return ret;
334 }
335
336 /*
337 * Walk the tree fixing up files which are actually leaf (ie
338 * have no children) but aren't at leaf level, so they have the
339 * leaf flag set.
340 */
341 void xhtml_fixup_layout(xhtmlfile* file)
342 {
343 if (file->child==NULL) {
344 file->is_leaf = TRUE;
345 } else {
346 xhtml_fixup_layout(file->child);
347 }
348 if (file->next)
349 xhtml_fixup_layout(file->next);
350 }
351
352 /*
353 * Create the tree structure so we know where everything goes.
354 * Method:
355 *
356 * Ignoring file splitting, we have three choices with each new section:
357 *
358 * +-----------------+-----------------+
359 * | | |
360 * X +----X----+ (1)
361 * | |
362 * Y (2)
363 * |
364 * (3)
365 *
366 * Y is the last section we added (currentsect).
367 * If sect is the section we want to add, then:
368 *
369 * (1) if sect->level < currentsect->level
370 * (2) if sect->level == currentsect->level
371 * (3) if sect->level > currentsect->level
372 *
373 * This requires the constraint that you never skip section numbers
374 * (so you can't have a.b.c.d without all of a, a.b and a.b.c existing).
375 *
376 * Note that you _can_ have 1.1.1.1 followed by 1.2 - you can change
377 * more than one level at a time. Lots of asserts, and probably part of
378 * the algorithm here, rely on this being true. (It currently isn't
379 * enforced by halibut, however.)
380 *
381 * File splitting makes this harder. For instance, say we added at (3)
382 * above and now need to add another section. We are splitting at level
383 * 2, ie the level of Y. Z is the last section we added:
384 *
385 * +-----------------+-----------------+
386 * | | |
387 * X +----X----+ (1)
388 * | |
389 * +----Y----+ (1)
390 * | |
391 * Z (2)
392 * |
393 * (3)
394 *
395 * The (1) case is now split; we need to search upwards to find where
396 * to actually link in. The other two cases remain the same (and will
397 * always be like this).
398 *
399 * File splitting makes this harder, however. The decision of whether
400 * to split to a new file is always on the same condition, however (is
401 * the level of this section higher than the leaf_level configuration
402 * value or not).
403 *
404 * Treating the cases backwards:
405 *
406 * (3) same file if sect->level > conf.leaf_level, otherwise new file
407 *
408 * if in the same file, currentsect->child points to sect
409 * otherwise the linking is done through the file tree (which works
410 * in more or less the same way, ie currentfile->child points to
411 * the new file)
412 *
413 * (2) same file if sect->level > conf.leaf_level, otherwise new file
414 *
415 * if in the same file, currentsect->next points to sect
416 * otherwise file linking and currentfile->next points to the new
417 * file (we know that Z must have caused a new file to be created)
418 *
419 * (1) same file if sect->level > conf.leaf_level, otherwise new file
420 *
421 * this is actually effectively the same case as (2) here,
422 * except that we first have to travel up the sections to figure
423 * out which section this new one will be a sibling of. In doing
424 * so, we may disappear off the top of a file and have to go up
425 * to its parent in the file tree.
426 *
427 */
428 static void xhtml_ponder_layout(paragraph *p)
429 {
430 xhtmlsection *lastsection;
431 xhtmlsection *currentsect;
432 xhtmlfile *currentfile;
433
434 lastfile = NULL;
435 topsection = xhtml_new_section(NULL);
436 topfile = xhtml_new_file(NULL);
437 lastsection = topsection;
438 currentfile = topfile;
439 currentsect = topsection;
440
441 if (conf.leaf_level == 0) {
442 topfile->is_leaf = 1;
443 topfile->sections = topsection;
444 topsection->file = topfile;
445 }
446
447 for (; p; p=p->next)
448 {
449 int level = xhtml_para_level(p);
450 if (level>0) /* actually a section */
451 {
452 xhtmlsection *sect;
453 word *w;
454 char *c;
455 rdstringc fname_c = { 0, 0, NULL };
456
457 sect = xhtml_new_section(lastsection);
458 lastsection = sect;
459 sect->para = p;
460 for (w=(p->kwtext2)?(p->kwtext2):(p->words); w; w=w->next) /* kwtext2 because we want numbers only! */
461 {
462 switch (removeattr(w->type))
463 {
464 case word_Normal:
465 /*case word_Emph:
466 case word_Code:
467 case word_WeakCode:*/
468 xhtml_utostr(w->text, &c);
469 rdaddsc(&fname_c,c);
470 sfree(c);
471 break;
472 }
473 }
474 /* rdaddsc(&fname_c, ".html");*/
475 sect->fragment = rdtrimc(&fname_c);
476 sect->level = level;
477 /* printf(" ! adding para @ %p as sect %s, level %i\n", sect->para, sect->fragment, level);*/
478
479 if (level>currentsect->level) { /* case (3) */
480 if (level>conf.leaf_level) { /* same file */
481 assert(currentfile->is_leaf);
482 currentsect->child = sect;
483 sect->parent=currentsect;
484 sect->file=currentfile;
485 /* printf("connected '%s' to existing file '%s' [I]\n", sect->fragment, currentfile->filename);*/
486 currentsect=sect;
487 } else { /* new file */
488 xhtmlfile *file = xhtml_new_file(sect);
489 assert(!currentfile->is_leaf);
490 currentfile->child=file;
491 sect->file=file;
492 file->parent=currentfile;
493 /* printf("connected '%s' to new file '%s' [I]\n", sect->fragment, file->filename);*/
494 currentfile=file;
495 currentsect=sect;
496 }
497 } else if (level >= currentsect->file->sections->level) {
498 /* Case (1) or (2) *AND* still under the section that starts
499 * the current file.
500 *
501 * I'm not convinced that this couldn't be rolled in with the
502 * final else {} leg further down. It seems a lot of effort
503 * this way.
504 */
505 if (level>conf.leaf_level) { /* stick within the same file */
506 assert(currentfile->is_leaf);
507 sect->file = currentfile;
508 while (currentsect && currentsect->level > level &&
509 currentsect->file==currentsect->parent->file) {
510 currentsect = currentsect->parent;
511 }
512 assert(currentsect);
513 currentsect->next = sect;
514 assert(currentsect->level == sect->level);
515 sect->parent = currentsect->parent;
516 currentsect = sect;
517 /* printf("connected '%s' to existing file '%s' [II]\n", sect->fragment, currentfile->filename);*/
518 } else { /* new file */
519 xhtmlfile *file = xhtml_new_file(sect);
520 sect->file=file;
521 currentfile->next=file;
522 file->parent=currentfile->parent;
523 file->is_leaf=(level==conf.leaf_level);
524 file->sections=sect;
525 /* printf("connected '%s' to new file '%s' [II]\n", sect->fragment, file->filename);*/
526 currentfile=file;
527 currentsect=sect;
528 }
529 } else { /* Case (1) or (2) and we must move up the file tree first */
530 /* this loop is now probably irrelevant - we know we can't connect
531 * to anything in the current file */
532 while (currentsect && level<currentsect->level) {
533 currentsect=currentsect->parent;
534 if (currentsect) {
535 /* printf(" * up one level to '%s'\n", currentsect->fragment);*/
536 } else {
537 /* printf(" * up one level (off top of current file)\n");*/
538 }
539 }
540 if (currentsect) {
541 /* I'm pretty sure this can now never fire */
542 assert(currentfile->is_leaf);
543 /* printf("connected '%s' to existing file '%s' [III]\n", sect->fragment, currentfile->filename);*/
544 sect->file = currentfile;
545 currentsect->next=sect;
546 currentsect=sect;
547 } else { /* find a file we can attach to */
548 while (currentfile && currentfile->sections && level<currentfile->sections->level) {
549 currentfile=currentfile->parent;
550 if (currentfile) {
551 /* printf(" * up one file level to '%s'\n", currentfile->filename);*/
552 } else {
553 /* printf(" * up one file level (off top of tree)\n");*/
554 }
555 }
556 if (currentfile) { /* new file (we had to skip up a file to
557 get here, so we must be dealing with a
558 level no lower than the configured
559 leaf_level */
560 xhtmlfile *file = xhtml_new_file(sect);
561 currentfile->next=file;
562 sect->file=file;
563 file->parent=currentfile->parent;
564 file->is_leaf=(level==conf.leaf_level);
565 file->sections=sect;
566 /* printf("connected '%s' to new file '%s' [III]\n", sect->fragment, file->filename);*/
567 currentfile=file;
568 currentsect=sect;
569 } else {
570 fatal(err_whatever, "Ran off the top trying to connect sibling: strange document.");
571 }
572 }
573 }
574 }
575 }
576 topsection = lastsection; /* get correct end of the chain */
577 xhtml_fixup_layout(topfile); /* leaf files not at leaf level marked as such */
578 }
579
580 static void xhtml_do_index();
581 static void xhtml_do_file(xhtmlfile *file);
582 static void xhtml_do_top_file(xhtmlfile *file, paragraph *sourceform);
583 static void xhtml_do_paras(FILE *fp, paragraph *p, paragraph *end, int indexable);
584 static int xhtml_do_contents_limit(FILE *fp, xhtmlfile *file, int limit);
585 static int xhtml_do_contents_section_limit(FILE *fp, xhtmlsection *section, int limit);
586 static int xhtml_add_contents_entry(FILE *fp, xhtmlsection *section, int limit);
587 static int xhtml_do_contents(FILE *fp, xhtmlfile *file);
588 static int xhtml_do_naked_contents(FILE *fp, xhtmlfile *file);
589 static void xhtml_do_sections(FILE *fp, xhtmlsection *sections);
590
591 /*
592 * Do all the files in this structure.
593 */
594 static void xhtml_do_files(xhtmlfile *file)
595 {
596 xhtml_do_file(file);
597 if (file->child)
598 xhtml_do_files(file->child);
599 if (file->next)
600 xhtml_do_files(file->next);
601 }
602
603 /*
604 * Free up all memory used by the file tree from 'xfile' downwards
605 */
606 static void xhtml_free_file(xhtmlfile* xfile)
607 {
608 if (xfile==NULL) {
609 return;
610 }
611
612 if (xfile->filename) {
613 sfree(xfile->filename);
614 }
615 xhtml_free_file(xfile->child);
616 xhtml_free_file(xfile->next);
617 sfree(xfile);
618 }
619
620 /*
621 * Main function.
622 */
623 void xhtml_backend(paragraph *sourceform, keywordlist *in_keywords,
624 indexdata *in_idx)
625 {
626 /* int i;*/
627 indexentry *ientry;
628 int ti;
629 xhtmlsection *xsect;
630
631 sourceparas = sourceform;
632 conf = xhtml_configure(sourceform);
633 keywords = in_keywords;
634 idx = in_idx;
635
636 /* Clear up the index entries backend data pointers */
637 for (ti=0; (ientry = (indexentry *)index234(idx->entries, ti))!=NULL; ti++) {
638 ientry->backend_data=NULL;
639 }
640
641 xhtml_ponder_layout(sourceform);
642
643 /* old system ... (writes to *.alt, but gets some stuff wrong and is ugly) */
644 /* xhtml_level_0(sourceform);
645 for (i=1; i<=conf.leaf_level; i++)
646 {
647 xhtml_level(sourceform, i);
648 }*/
649
650 /* new system ... (writes to *.html, but isn't fully trusted) */
651 xhtml_do_top_file(topfile, sourceform);
652 assert(!topfile->next); /* shouldn't have a sibling at all */
653 if (topfile->child) {
654 xhtml_do_files(topfile->child);
655 xhtml_do_index();
656 }
657
658 /* release file, section, index data structures */
659 xsect = topsection;
660 while (xsect) {
661 xhtmlsection *tmp = xsect->chain;
662 if (xsect->fragment) {
663 sfree(xsect->fragment);
664 }
665 sfree(xsect);
666 xsect = tmp;
667 }
668 xhtml_free_file(topfile);
669 for (ti = 0; (ientry=(indexentry *)index234(idx->entries, ti))!=NULL; ti++) {
670 if (ientry->backend_data!=NULL) {
671 xhtmlindex *xi = (xhtmlindex*) ientry->backend_data;
672 if (xi->sections!=NULL) {
673 sfree(xi->sections);
674 }
675 sfree(xi);
676 }
677 ientry->backend_data = NULL;
678 }
679 {
680 int i;
681 sfree(conf.fchapter.number_suffix);
682 for (i = 0; i < conf.nfsect; i++)
683 sfree(conf.fsect[i].number_suffix);
684 sfree(conf.fsect);
685 }
686 }
687
688 static int xhtml_para_level(paragraph *p)
689 {
690 switch (p->type)
691 {
692 case para_Title:
693 return 0;
694 break;
695 case para_UnnumberedChapter:
696 case para_Chapter:
697 case para_Appendix:
698 return 1;
699 break;
700 /* case para_BiblioCited:
701 return 2;
702 break;*/
703 case para_Heading:
704 case para_Subsect:
705 return p->aux+2;
706 break;
707 default:
708 return -1;
709 break;
710 }
711 }
712
713 static char* xhtml_index_filename = "IndexPage.html";
714
715 /* Output the nav links for the current file.
716 * file == NULL means we're doing the index
717 */
718 static void xhtml_donavlinks(FILE *fp, xhtmlfile *file)
719 {
720 xhtmlfile *xhtml_next_file = NULL;
721 fprintf(fp, "<p");
722 if (conf.nav_attrs!=NULL) {
723 fprintf(fp, " %ls>", conf.nav_attrs);
724 } else {
725 fprintf(fp, ">");
726 }
727 if (xhtml_last_file==NULL) {
728 fprintf(fp, "Previous | ");
729 } else {
730 fprintf(fp, "<a href='%s'>Previous</a> | ", xhtml_last_file->filename);
731 }
732 fprintf(fp, "<a href='Contents.html'>Contents</a> | ");
733 if (file == NULL) {
734 fprintf(fp, "Index | ");
735 } else {
736 fprintf(fp, "<a href='%s'>Index</a> | ", xhtml_index_filename);
737 }
738 if (file != NULL) { /* otherwise we're doing nav links for the index */
739 if (xhtml_next_file==NULL)
740 xhtml_next_file = file->child;
741 if (xhtml_next_file==NULL)
742 xhtml_next_file = file->next;
743 if (xhtml_next_file==NULL)
744 xhtml_next_file = file->parent->next;
745 }
746 if (xhtml_next_file==NULL) {
747 if (file==NULL) { /* index, so no next file */
748 fprintf(fp, "Next ");
749 } else {
750 fprintf(fp, "<a href='%s'>Next</a>", xhtml_index_filename);
751 }
752 } else {
753 fprintf(fp, "<a href='%s'>Next</a>", xhtml_next_file->filename);
754 }
755 fprintf(fp, "</p>\n");
756 }
757
758 /* Write out the index file */
759 static void xhtml_do_index_body(FILE *fp)
760 {
761 indexentry *y;
762 int ti;
763
764 if (count234(idx->entries) == 0)
765 return; /* don't write anything at all */
766
767 fprintf(fp, "<dl>\n");
768 /* iterate over idx->entries using the tree functions and display everything */
769 for (ti = 0; (y = (indexentry *)index234(idx->entries, ti)) != NULL; ti++) {
770 if (y->backend_data) {
771 int i;
772 xhtmlindex *xi;
773
774 fprintf(fp, "<dt>");
775 xhtml_para(fp, y->text, FALSE);
776 fprintf(fp, "</dt>\n<dd>");
777
778 xi = (xhtmlindex*) y->backend_data;
779 for (i=0; i<xi->nsection; i++) {
780 xhtmlsection *sect = xi->sections[i];
781 if (sect) {
782 fprintf(fp, "<a href='%s#%s'>", sect->file->filename, sect->fragment);
783 if (sect->para->kwtext) {
784 xhtml_para(fp, sect->para->kwtext, FALSE);
785 } else if (sect->para->words) {
786 xhtml_para(fp, sect->para->words, FALSE);
787 }
788 fprintf(fp, "</a>");
789 if (i+1<xi->nsection) {
790 fprintf(fp, ", ");
791 }
792 }
793 }
794 fprintf(fp, "</dd>\n");
795 }
796 }
797 fprintf(fp, "</dl>\n");
798 }
799 static void xhtml_do_index()
800 {
801 word temp_word = { NULL, NULL, word_Normal, 0, 0, L"Index", { NULL, 0, 0} };
802 FILE *fp = fopen(xhtml_index_filename, "w");
803
804 if (fp==NULL)
805 fatal(err_cantopenw, xhtml_index_filename);
806 xhtml_doheader(fp, &temp_word);
807 xhtml_donavlinks(fp, NULL);
808
809 xhtml_do_index_body(fp);
810
811 xhtml_donavlinks(fp, NULL);
812 xhtml_dofooter(fp);
813 fclose(fp);
814 }
815
816 /* Output the given file. This includes whatever contents at beginning and end, etc. etc. */
817 static void xhtml_do_file(xhtmlfile *file)
818 {
819 FILE *fp = fopen(file->filename, "w");
820 if (fp==NULL)
821 fatal(err_cantopenw, file->filename);
822
823 if (file->sections->para->words) {
824 xhtml_doheader(fp, file->sections->para->words);
825 } else if (file->sections->para->kwtext) {
826 xhtml_doheader(fp, file->sections->para->kwtext);
827 } else {
828 xhtml_doheader(fp, NULL);
829 }
830
831 xhtml_donavlinks(fp, file);
832
833 if (file->is_leaf && conf.leaf_contains_contents &&
834 xhtml_do_contents(NULL, file)>=conf.leaf_smallest_contents)
835 xhtml_do_contents(fp, file);
836 xhtml_do_sections(fp, file->sections);
837 if (!file->is_leaf)
838 xhtml_do_naked_contents(fp, file);
839
840 xhtml_donavlinks(fp, file);
841
842 xhtml_dofooter(fp);
843 fclose(fp);
844
845 xhtml_last_file = file;
846 }
847
848 /* Output the top-level file. */
849 static void xhtml_do_top_file(xhtmlfile *file, paragraph *sourceform)
850 {
851 paragraph *p;
852 int done=FALSE;
853 FILE *fp = fopen(file->filename, "w");
854 if (fp==NULL)
855 fatal(err_cantopenw, file->filename);
856
857 /* Do the title -- only one allowed */
858 for (p = sourceform; p && !done; p = p->next)
859 {
860 if (p->type == para_Title)
861 {
862 xhtml_doheader(fp, p->words);
863 done=TRUE;
864 }
865 }
866 if (!done)
867 xhtml_doheader(fp, NULL /* Eek! */);
868
869 /*
870 * Display the title.
871 */
872 for (p = sourceform; p; p = p->next)
873 {
874 if (p->type == para_Title) {
875 xhtml_heading(fp, p, FALSE);
876 break;
877 }
878 }
879
880 /* Do the preamble */
881 for (p = sourceform; p; p = p->next)
882 {
883 if (p->type == para_Chapter || p->type == para_Heading ||
884 p->type == para_Subsect || p->type == para_Appendix ||
885 p->type == para_UnnumberedChapter) {
886 /*
887 * We've found the end of the preamble. Do every normal
888 * paragraph up to there.
889 */
890 xhtml_do_paras(fp, sourceform, p, FALSE);
891 break;
892 }
893 }
894
895 xhtml_do_contents(fp, file);
896 xhtml_do_sections(fp, file->sections);
897
898 /*
899 * Put the index in the top file if we're in single-file mode
900 * (leaf-level 0).
901 */
902 if (conf.leaf_level == 0 && count234(idx->entries) > 0) {
903 fprintf(fp, "<a name=\"index\"></a><h1>Index</h1>\n");
904 xhtml_do_index_body(fp);
905 }
906
907 xhtml_dofooter(fp);
908 fclose(fp);
909 }
910
911 /* Convert a Unicode string to an ASCII one. '?' is
912 * used for unmappable characters.
913 */
914 static void xhtml_utostr(wchar_t *in, char **out)
915 {
916 int l = ustrlen(in);
917 int i;
918 *out = smalloc(l+1);
919 for (i=0; i<l; i++)
920 {
921 if (in[i]>=32 && in[i]<=126)
922 (*out)[i]=(char)in[i];
923 else
924 (*out)[i]='?';
925 }
926 (*out)[i]=0;
927 }
928
929 /*
930 * Write contents for the given file, and subfiles, down to
931 * the appropriate contents depth. Returns the number of
932 * entries written.
933 */
934 static int xhtml_do_contents(FILE *fp, xhtmlfile *file)
935 {
936 int level, limit, count = 0;
937 if (!file)
938 return 0;
939
940 level = (file->sections)?(file->sections->level):(0);
941 limit = conf.contents_depth[(level>5)?(5):(level)];
942 start_level = (file->is_leaf) ? (level-1) : (level);
943 last_level = start_level;
944
945 count += xhtml_do_contents_section_limit(fp, file->sections, limit);
946 count += xhtml_do_contents_limit(fp, file->child, limit);
947 if (fp!=NULL) {
948 while (last_level > start_level) {
949 last_level--;
950 fprintf(fp, "</li></ul>\n");
951 }
952 }
953 return count;
954 }
955
956 /* As above, but doesn't do anything in the current file */
957 static int xhtml_do_naked_contents(FILE *fp, xhtmlfile *file)
958 {
959 int level, limit, start_level, count = 0;
960 if (!file)
961 return 0;
962
963 level = (file->sections)?(file->sections->level):(0);
964 limit = conf.contents_depth[(level>5)?(5):(level)];
965 start_level = (file->is_leaf) ? (level-1) : (level);
966 last_level = start_level;
967
968 count = xhtml_do_contents_limit(fp, file->child, limit);
969 if (fp!=NULL) {
970 while (last_level > start_level) {
971 last_level--;
972 fprintf(fp, "</li></ul>\n");
973 }
974 }
975 return count;
976 }
977
978 /*
979 * Write contents for the given file, children, and siblings, down to
980 * given limit contents depth.
981 */
982 static int xhtml_do_contents_limit(FILE *fp, xhtmlfile *file, int limit)
983 {
984 int count = 0;
985 while (file) {
986 count += xhtml_do_contents_section_limit(fp, file->sections, limit);
987 count += xhtml_do_contents_limit(fp, file->child, limit);
988 file = file->next;
989 }
990 return count;
991 }
992
993 /*
994 * Write contents entries for the given section tree, down to the
995 * limit contents depth.
996 */
997 static int xhtml_do_contents_section_deep_limit(FILE *fp, xhtmlsection *section, int limit)
998 {
999 int count = 0;
1000 while (section) {
1001 if (!xhtml_add_contents_entry(fp, section, limit))
1002 return 0;
1003 else
1004 count++;
1005 count += xhtml_do_contents_section_deep_limit(fp, section->child, limit);
1006 section = section->next;
1007 }
1008 return count;
1009 }
1010
1011 /*
1012 * Write contents entries for the given section tree, down to the
1013 * limit contents depth.
1014 */
1015 static int xhtml_do_contents_section_limit(FILE *fp, xhtmlsection *section, int limit)
1016 {
1017 int count = 0;
1018 if (!section)
1019 return 0;
1020 xhtml_add_contents_entry(fp, section, limit);
1021 count=1;
1022 count += xhtml_do_contents_section_deep_limit(fp, section->child, limit);
1023 /* section=section->child;
1024 while (section && xhtml_add_contents_entry(fp, section, limit)) {
1025 section = section->next;
1026 }*/
1027 return count;
1028 }
1029
1030 /*
1031 * Add a section entry, unless we're exceeding the limit, in which
1032 * case return FALSE (otherwise return TRUE).
1033 */
1034 static int xhtml_add_contents_entry(FILE *fp, xhtmlsection *section, int limit)
1035 {
1036 if (!section || section->level > limit)
1037 return FALSE;
1038 if (fp==NULL || section->level < 0)
1039 return TRUE;
1040 if (last_level > section->level) {
1041 while (last_level > section->level) {
1042 last_level--;
1043 fprintf(fp, "</li></ul>\n");
1044 }
1045 fprintf(fp, "</li>\n");
1046 } else if (last_level < section->level) {
1047 assert(last_level == section->level - 1);
1048 last_level++;
1049 fprintf(fp, "<ul>\n");
1050 } else {
1051 fprintf(fp, "</li>\n");
1052 }
1053 fprintf(fp, "<li><a href=\"%s#%s\">", section->file->filename, section->fragment);
1054 if (section->para->kwtext) {
1055 xhtml_para(fp, section->para->kwtext, FALSE);
1056 if (section->para->words) {
1057 fprintf(fp, ": ");
1058 }
1059 }
1060 if (section->para->words) {
1061 xhtml_para(fp, section->para->words, FALSE);
1062 }
1063 fprintf(fp, "</a>\n");
1064 return TRUE;
1065 }
1066
1067 /*
1068 * Write all the sections in this file. Do all paragraphs in this section, then all
1069 * children (recursively), then go on to the next one (tail recursively).
1070 */
1071 static void xhtml_do_sections(FILE *fp, xhtmlsection *sections)
1072 {
1073 while (sections) {
1074 currentsection = sections;
1075 xhtml_do_paras(fp, sections->para, NULL, TRUE);
1076 xhtml_do_sections(fp, sections->child);
1077 sections = sections->next;
1078 }
1079 }
1080
1081 /* Write this list of paragraphs. Close off all lists at the end. */
1082 static void xhtml_do_paras(FILE *fp, paragraph *p, paragraph *end,
1083 int indexable)
1084 {
1085 int last_type = -1, ptype, first=TRUE;
1086 stack lcont_stack = stk_new();
1087 if (!p)
1088 return;
1089
1090 /* for (; p && (xhtml_para_level(p)>limit || xhtml_para_level(p)==-1 || first); p=p->next) {*/
1091 for (; p && p != end && (xhtml_para_level(p)==-1 || first); p=p->next) {
1092 first=FALSE;
1093 switch (ptype = p->type)
1094 {
1095 /*
1096 * Things we ignore because we've already processed them or
1097 * aren't going to touch them in this pass.
1098 */
1099 case para_IM:
1100 case para_BR:
1101 case para_Biblio: /* only touch BiblioCited */
1102 case para_VersionID:
1103 case para_NoCite:
1104 case para_Title:
1105 break;
1106
1107 /*
1108 * Chapter titles.
1109 */
1110 case para_Chapter:
1111 case para_Appendix:
1112 case para_UnnumberedChapter:
1113 xhtml_heading(fp, p, indexable);
1114 break;
1115
1116 case para_Heading:
1117 case para_Subsect:
1118 xhtml_heading(fp, p, indexable);
1119 break;
1120
1121 case para_Rule:
1122 fprintf(fp, "\n<hr />\n");
1123 break;
1124
1125 case para_Normal:
1126 case para_Copyright:
1127 fprintf(fp, "\n<p>");
1128 xhtml_para(fp, p->words, indexable);
1129 fprintf(fp, "</p>\n");
1130 break;
1131
1132 case para_LcontPush:
1133 {
1134 int *p;
1135 p = mknew(int);
1136 *p = last_type;
1137 stk_push(lcont_stack, p);
1138 last_type = para_Normal;
1139 }
1140 break;
1141 case para_LcontPop:
1142 {
1143 int *p = stk_pop(lcont_stack);
1144 assert(p);
1145 ptype = last_type = *p;
1146 sfree(p);
1147 goto closeofflist; /* ick */
1148 }
1149 break;
1150 case para_QuotePush:
1151 fprintf(fp, "<blockquote>\n");
1152 break;
1153 case para_QuotePop:
1154 fprintf(fp, "</blockquote>\n");
1155 break;
1156
1157 case para_Bullet:
1158 case para_NumberedList:
1159 case para_Description:
1160 case para_DescribedThing:
1161 case para_BiblioCited:
1162 if (last_type!=p->type &&
1163 !(last_type==para_DescribedThing && p->type==para_Description) &&
1164 !(last_type==para_Description && p->type==para_DescribedThing)) {
1165 /* start up list if necessary */
1166 if (p->type == para_Bullet) {
1167 fprintf(fp, "<ul>\n");
1168 } else if (p->type == para_NumberedList) {
1169 fprintf(fp, "<ol>\n");
1170 } else if (p->type == para_BiblioCited ||
1171 p->type == para_DescribedThing ||
1172 p->type == para_Description) {
1173 fprintf(fp, "<dl>\n");
1174 }
1175 }
1176 if (p->type == para_Bullet || p->type == para_NumberedList) {
1177 fprintf(fp, "<li>");
1178 } else if (p->type == para_DescribedThing) {
1179 fprintf(fp, "<dt>");
1180 } else if (p->type == para_Description) {
1181 fprintf(fp, "<dd>");
1182 } else if (p->type == para_BiblioCited) {
1183 fprintf(fp, "<dt>");
1184 xhtml_para(fp, p->kwtext, indexable);
1185 fprintf(fp, "</dt>\n<dd>");
1186 }
1187 xhtml_para(fp, p->words, indexable);
1188 {
1189 paragraph *p2 = p->next;
1190 if (p2 && xhtml_para_level(p2)==-1 && p2->type == para_LcontPush)
1191 break;
1192 }
1193
1194 closeofflist:
1195 if (ptype == para_BiblioCited) {
1196 fprintf(fp, "</dd>\n");
1197 } else if (ptype == para_DescribedThing) {
1198 fprintf(fp, "</dt>");
1199 } else if (ptype == para_Description) {
1200 fprintf(fp, "</dd>");
1201 } else if (ptype == para_Bullet || ptype == para_NumberedList) {
1202 fprintf(fp, "</li>");
1203 }
1204 if (ptype == para_Bullet || ptype == para_NumberedList ||
1205 ptype == para_BiblioCited || ptype == para_Description ||
1206 ptype == para_DescribedThing)
1207 /* close off list if necessary */
1208 {
1209 paragraph *p2 = p->next;
1210 int close_off=FALSE;
1211 /* if (p2 && (xhtml_para_level(p2)>limit || xhtml_para_level(p2)==-1)) {*/
1212 if (p2 && xhtml_para_level(p2)==-1) {
1213 if (p2->type != ptype &&
1214 !(p2->type==para_DescribedThing && ptype==para_Description) &&
1215 !(p2->type==para_Description && ptype==para_DescribedThing) &&
1216 p2->type != para_LcontPush)
1217 close_off=TRUE;
1218 } else {
1219 close_off=TRUE;
1220 }
1221 if (close_off) {
1222 if (ptype == para_Bullet) {
1223 fprintf(fp, "</ul>\n");
1224 } else if (ptype == para_NumberedList) {
1225 fprintf(fp, "</ol>\n");
1226 } else if (ptype == para_BiblioCited ||
1227 ptype == para_Description ||
1228 ptype == para_DescribedThing) {
1229 fprintf(fp, "</dl>\n");
1230 }
1231 }
1232 }
1233 break;
1234
1235 case para_Code:
1236 xhtml_codepara(fp, p->words);
1237 break;
1238 }
1239 last_type = ptype;
1240 }
1241
1242 stk_free(lcont_stack);
1243 }
1244
1245 /*
1246 * Output a header for this XHTML file.
1247 */
1248 static void xhtml_doheader(FILE *fp, word *title)
1249 {
1250 fprintf(fp, "<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Transitional//EN\"\n");
1251 fprintf(fp, "\"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd\">\n");
1252 fprintf(fp, "<html xmlns='http://www.w3.org/1999/xhtml'>\n\n<head>\n<title>");
1253 if (title==NULL)
1254 fprintf(fp, "The thing with no name!");
1255 else
1256 xhtml_para(fp, title, FALSE);
1257 fprintf(fp, "</title>\n");
1258 fprintf(fp, "<meta name=\"generator\" content=\"Halibut %s xhtml-backend\" />\n", version);
1259 if (conf.author)
1260 fprintf(fp, "<meta name=\"author\" content=\"%ls\" />\n", conf.author);
1261 if (conf.description)
1262 fprintf(fp, "<meta name=\"description\" content=\"%ls\" />\n", conf.description);
1263 if (conf.head_end)
1264 fprintf(fp, "%ls\n", conf.head_end);
1265 fprintf(fp, "</head>\n\n");
1266 if (conf.body)
1267 fprintf(fp, "%ls\n", conf.body);
1268 else
1269 fprintf(fp, "<body>\n");
1270 if (conf.body_start)
1271 fprintf(fp, "%ls\n", conf.body_start);
1272 }
1273
1274 /*
1275 * Output a footer for this XHTML file.
1276 */
1277 static void xhtml_dofooter(FILE *fp)
1278 {
1279 fprintf(fp, "\n<hr />\n\n");
1280 if (conf.body_end)
1281 fprintf(fp, "%ls\n", conf.body_end);
1282 if (!conf.suppress_address) {
1283 fprintf(fp,"<address>\n");
1284 if (conf.address_start)
1285 fprintf(fp, "%ls\n", conf.address_start);
1286 /* Do the version ID */
1287 if (conf.include_version_id) {
1288 paragraph *p;
1289 int started = 0;
1290 for (p = sourceparas; p; p = p->next)
1291 if (p->type == para_VersionID) {
1292 xhtml_versionid(fp, p->words, started);
1293 started = 1;
1294 }
1295 }
1296 if (conf.address_end)
1297 fprintf(fp, "%ls\n", conf.address_end);
1298 fprintf(fp, "</address>\n");
1299 }
1300 fprintf(fp, "</body>\n\n</html>\n");
1301 }
1302
1303 /*
1304 * Output the versionid paragraph. Typically this is a version control
1305 * ID string (such as $Id...$ in RCS).
1306 */
1307 static void xhtml_versionid(FILE *fp, word *text, int started)
1308 {
1309 rdstringc t = { 0, 0, NULL };
1310
1311 rdaddc(&t, '['); /* FIXME: configurability */
1312 xhtml_rdaddwc(&t, text, NULL, FALSE);
1313 rdaddc(&t, ']'); /* FIXME: configurability */
1314
1315 if (started)
1316 fprintf(fp, "<br />\n");
1317 fprintf(fp, "%s\n", t.text);
1318 sfree(t.text);
1319 }
1320
1321 /* Is this an XHTML reserved character? */
1322 static int xhtml_reservedchar(int c)
1323 {
1324 if (c=='&' || c=='<' || c=='>' || c=='"')
1325 return TRUE;
1326 else
1327 return FALSE;
1328 }
1329
1330 /*
1331 * Convert a wide string into valid XHTML: Anything outside ASCII will
1332 * be fixed up as an entity. Currently we don't worry about constraining the
1333 * encoded character set, which we should probably do at some point (we can
1334 * still fix up and return FALSE - see the last comment here). We also don't
1335 * currently
1336 *
1337 * Because this is only used for words, spaces are HARD spaces (any other
1338 * spaces will be word_Whitespace not word_Normal). So they become &nbsp;
1339 * Unless hard_spaces is FALSE, of course (code paragraphs break the above
1340 * rule).
1341 *
1342 * If `result' is non-NULL, mallocs the resulting string and stores a pointer to
1343 * it in `*result'. If `result' is NULL, merely checks whether all
1344 * characters in the string are feasible.
1345 *
1346 * Return is nonzero if all characters are OK. If not all
1347 * characters are OK but `result' is non-NULL, a result _will_
1348 * still be generated!
1349 */
1350 static int xhtml_convert(wchar_t *s, int maxlen, char **result,
1351 int hard_spaces) {
1352 int doing = (result != 0);
1353 int ok = TRUE;
1354 char *p = NULL;
1355 int plen = 0, psize = 0;
1356
1357 if (maxlen <= 0)
1358 maxlen = -1;
1359
1360 for (; *s && maxlen != 0; s++, maxlen--) {
1361 wchar_t c = *s;
1362
1363 #define ensure_size(i) if (i>=psize) { psize = i+256; p = resize(p, psize); }
1364
1365 if (((c == 32 && !hard_spaces) || (c > 32 && c <= 126 && !xhtml_reservedchar(c)))) {
1366 /* Char is OK. */
1367 if (doing)
1368 {
1369 ensure_size(plen);
1370 p[plen++] = (char)c;
1371 }
1372 } else {
1373 /* Char needs fixing up. */
1374 /* ok = FALSE; -- currently we never return FALSE; we
1375 * might want to when considering a character set for the
1376 * encoded document.
1377 */
1378 if (doing)
1379 {
1380 if (c==32) { /* a space in a word is a hard space */
1381 ensure_size(plen+6); /* includes space for the NUL, which is subsequently stomped on */
1382 sprintf(p+plen, "&nbsp;");
1383 plen+=6;
1384 } else {
1385 /* FIXME: entity names! */
1386 ensure_size(plen+8); /* includes space for the NUL, which is subsequently stomped on */
1387 plen+=sprintf(p+plen, "&#%04i;", (int)c);
1388 }
1389 }
1390 }
1391 }
1392 if (doing) {
1393 p = resize(p, plen+1);
1394 p[plen] = '\0';
1395 *result = p;
1396 }
1397 return ok;
1398 }
1399
1400 /*
1401 * This formats the given words as XHTML.
1402 *
1403 * `indexable', if FALSE, prohibits adding any index references.
1404 * You might use this, for example, if an index reference occurred
1405 * in a section title, to prevent phony index references when the
1406 * section title is processed in strange places such as contents
1407 * sections.
1408 */
1409 static void xhtml_rdaddwc(rdstringc *rs, word *text, word *end, int indexable) {
1410 char *c;
1411 keyword *kwl;
1412 xhtmlsection *sect;
1413 indextag *itag;
1414 int ti;
1415
1416 for (; text && text != end; text = text->next) {
1417 switch (text->type) {
1418 case word_HyperLink:
1419 xhtml_utostr(text->text, &c);
1420 rdaddsc(rs, "<a href=\"");
1421 rdaddsc(rs, c);
1422 rdaddsc(rs, "\">");
1423 sfree(c);
1424 break;
1425
1426 case word_UpperXref:
1427 case word_LowerXref:
1428 kwl = kw_lookup(keywords, text->text);
1429 if (kwl) {
1430 sect=xhtml_find_section(kwl->para);
1431 if (sect) {
1432 rdaddsc(rs, "<a href=\"");
1433 rdaddsc(rs, sect->file->filename);
1434 rdaddc(rs, '#');
1435 rdaddsc(rs, sect->fragment);
1436 rdaddsc(rs, "\">");
1437 } else {
1438 rdaddsc(rs, "<a href=\"Apologies.html\"><!-- probably a bibliography cross reference -->");
1439 error(err_whatever, "Couldn't locate cross-reference! (Probably a bibliography entry.)");
1440 }
1441 } else {
1442 rdaddsc(rs, "<a href=\"Apologies.html\"><!-- unknown cross-reference -->");
1443 error(err_whatever, "Couldn't locate cross-reference! (Wasn't in source file.)");
1444 }
1445 break;
1446
1447 case word_IndexRef: /* in theory we could make an index target here */
1448 /* rdaddsc(rs, "<a name=\"idx-");
1449 xhtml_utostr(text->text, &c);
1450 rdaddsc(rs, c);
1451 sfree(c);
1452 rdaddsc(rs, "\"></a>");*/
1453 /* what we _do_ need to do is to fix up the backend data
1454 * for any indexentry this points to.
1455 */
1456 if (!indexable)
1457 break;
1458
1459 for (ti=0; (itag = (indextag *)index234(idx->tags, ti))!=NULL; ti++) {
1460 /* FIXME: really ustricmp() and not ustrcmp()? */
1461 if (ustricmp(itag->name, text->text)==0) {
1462 break;
1463 }
1464 }
1465 if (itag!=NULL) {
1466 if (itag->refs!=NULL) {
1467 int i;
1468 for (i=0; i<itag->nrefs; i++) {
1469 xhtmlindex *idx_ref;
1470 indexentry *ientry;
1471
1472 ientry = itag->refs[i];
1473 if (ientry->backend_data==NULL) {
1474 idx_ref = (xhtmlindex*) smalloc(sizeof(xhtmlindex));
1475 if (idx_ref==NULL)
1476 fatal(err_nomemory);
1477 idx_ref->nsection = 0;
1478 idx_ref->size = 4;
1479 idx_ref->sections = (xhtmlsection**) smalloc(idx_ref->size * sizeof(xhtmlsection*));
1480 if (idx_ref->sections==NULL)
1481 fatal(err_nomemory);
1482 ientry->backend_data = idx_ref;
1483 } else {
1484 idx_ref = ientry->backend_data;
1485 if (idx_ref->nsection+1 > idx_ref->size) {
1486 int new_size = idx_ref->size * 2;
1487 idx_ref->sections = srealloc(idx_ref->sections, new_size * sizeof(xhtmlsection));
1488 if (idx_ref->sections==NULL) {
1489 fatal(err_nomemory);
1490 }
1491 idx_ref->size = new_size;
1492 }
1493 }
1494 idx_ref->sections[idx_ref->nsection++] = currentsection;
1495 #if 0
1496 #endif
1497 }
1498 } else {
1499 fatal(err_whatever, "Index tag had no entries!");
1500 }
1501 } else {
1502 fprintf(stderr, "Looking for index entry '%ls'\n", text->text);
1503 fatal(err_whatever, "Couldn't locate index entry! (Wasn't in index.)");
1504 }
1505 break;
1506
1507 case word_HyperEnd:
1508 case word_XrefEnd:
1509 rdaddsc(rs, "</a>");
1510 break;
1511
1512 case word_Normal:
1513 case word_Emph:
1514 case word_Code:
1515 case word_WeakCode:
1516 case word_WhiteSpace:
1517 case word_EmphSpace:
1518 case word_CodeSpace:
1519 case word_WkCodeSpace:
1520 case word_Quote:
1521 case word_EmphQuote:
1522 case word_CodeQuote:
1523 case word_WkCodeQuote:
1524 assert(text->type != word_CodeQuote &&
1525 text->type != word_WkCodeQuote);
1526 if (towordstyle(text->type) == word_Emph &&
1527 (attraux(text->aux) == attr_First ||
1528 attraux(text->aux) == attr_Only))
1529 rdaddsc(rs, "<em>");
1530 else if ((towordstyle(text->type) == word_Code || towordstyle(text->type) == word_WeakCode) &&
1531 (attraux(text->aux) == attr_First ||
1532 attraux(text->aux) == attr_Only))
1533 rdaddsc(rs, "<code>");
1534
1535 if (removeattr(text->type) == word_Normal) {
1536 if (xhtml_convert(text->text, 0, &c, TRUE)) /* spaces in the word are hard */
1537 rdaddsc(rs, c);
1538 else
1539 xhtml_rdaddwc(rs, text->alt, NULL, indexable);
1540 sfree(c);
1541 } else if (removeattr(text->type) == word_WhiteSpace) {
1542 rdaddc(rs, ' ');
1543 } else if (removeattr(text->type) == word_Quote) {
1544 rdaddsc(rs, "&quot;");
1545 }
1546
1547 if (towordstyle(text->type) == word_Emph &&
1548 (attraux(text->aux) == attr_Last ||
1549 attraux(text->aux) == attr_Only))
1550 rdaddsc(rs, "</em>");
1551 else if ((towordstyle(text->type) == word_Code || towordstyle(text->type) == word_WeakCode) &&
1552 (attraux(text->aux) == attr_Last ||
1553 attraux(text->aux) == attr_Only))
1554 rdaddsc(rs, "</code>");
1555 break;
1556 }
1557 }
1558 }
1559
1560 /* Output a heading, formatted as XHTML.
1561 */
1562 static void xhtml_heading(FILE *fp, paragraph *p, int indexable)
1563 {
1564 rdstringc t = { 0, 0, NULL };
1565 word *tprefix = p->kwtext;
1566 word *nprefix = p->kwtext2;
1567 word *text = p->words;
1568 int level = xhtml_para_level(p);
1569 xhtmlsection *sect = xhtml_find_section(p);
1570 xhtmlheadfmt *fmt;
1571 char *fragment;
1572 if (sect) {
1573 fragment = sect->fragment;
1574 } else {
1575 if (p->type == para_Title)
1576 fragment = "title";
1577 else {
1578 fragment = ""; /* FIXME: what else can we do? */
1579 error(err_whatever, "Couldn't locate heading cross-reference!");
1580 }
1581 }
1582
1583 if (p->type == para_Title)
1584 fmt = NULL;
1585 else if (level == 1)
1586 fmt = &conf.fchapter;
1587 else if (level-1 < conf.nfsect)
1588 fmt = &conf.fsect[level-1];
1589 else
1590 fmt = &conf.fsect[conf.nfsect-1];
1591
1592 if (fmt && fmt->just_numbers && nprefix) {
1593 xhtml_rdaddwc(&t, nprefix, NULL, indexable);
1594 if (fmt) {
1595 char *c;
1596 if (xhtml_convert(fmt->number_suffix, 0, &c, FALSE)) {
1597 rdaddsc(&t, c);
1598 sfree(c);
1599 }
1600 }
1601 } else if (fmt && !fmt->just_numbers && tprefix) {
1602 xhtml_rdaddwc(&t, tprefix, NULL, indexable);
1603 if (fmt) {
1604 char *c;
1605 if (xhtml_convert(fmt->number_suffix, 0, &c, FALSE)) {
1606 rdaddsc(&t, c);
1607 sfree(c);
1608 }
1609 }
1610 }
1611 xhtml_rdaddwc(&t, text, NULL, indexable);
1612 /*
1613 * If we're outputting in single-file mode, we need to lower
1614 * the level of each heading by one, because the overall
1615 * document title will be sitting right at the top as an <h1>
1616 * and so chapters and sections should start at <h2>.
1617 *
1618 * Even if not, the document title will come back from
1619 * xhtml_para_level() as level zero, so we must increment that
1620 * no matter what leaf_level is set to.
1621 */
1622 if (conf.leaf_level == 0 || level == 0)
1623 level++;
1624 fprintf(fp, "<a name=\"%s\"></a><h%i>%s</h%i>\n", fragment, level, t.text, level);
1625 sfree(t.text);
1626 }
1627
1628 /* Output a paragraph. Styles are handled by xhtml_rdaddwc().
1629 * This looks pretty simple; I may have missed something ...
1630 */
1631 static void xhtml_para(FILE *fp, word *text, int indexable)
1632 {
1633 rdstringc out = { 0, 0, NULL };
1634 xhtml_rdaddwc(&out, text, NULL, indexable);
1635 fprintf(fp, "%s", out.text);
1636 sfree(out.text);
1637 }
1638
1639 /* Output a code paragraph. I'm treating this as preformatted, which
1640 * may not be entirely correct. See xhtml_para() for my worries about
1641 * this being overly-simple; however I think that most of the complexity
1642 * of the text backend came entirely out of word wrapping anyway.
1643 */
1644 static void xhtml_codepara(FILE *fp, word *text)
1645 {
1646 fprintf(fp, "<pre>");
1647 for (; text; text = text->next) if (text->type == word_WeakCode) {
1648 word *here, *next;
1649 char *c;
1650
1651 /*
1652 * See if this WeakCode is followed by an Emph to indicate
1653 * emphasis.
1654 */
1655 here = text;
1656 if (text->next && text->next->type == word_Emph) {
1657 next = text = text->next;
1658 } else
1659 next = NULL;
1660
1661 if (next) {
1662 wchar_t *t, *e;
1663 int n;
1664
1665 t = here->text;
1666 e = next->text;
1667
1668 while (*e) {
1669 int ec = *e;
1670
1671 for (n = 0; t[n] && e[n] && e[n] == ec; n++);
1672 xhtml_convert(t, n, &c, FALSE);
1673 fprintf(fp, "%s%s%s",
1674 (ec == 'i' ? "<em>" : ec == 'b' ? "<b>" : ""),
1675 c,
1676 (ec == 'i' ? "</em>" : ec == 'b' ? "</b>" : ""));
1677 sfree(c);
1678
1679 t += n;
1680 e += n;
1681 }
1682
1683 xhtml_convert(t, 0, &c, FALSE);
1684 fprintf(fp, "%s\n", c);
1685 sfree(c);
1686 } else {
1687 xhtml_convert(here->text, 0, &c, FALSE);
1688 fprintf(fp, "%s\n", c);
1689 sfree(c);
1690 }
1691 }
1692 fprintf(fp, "</pre>\n");
1693 }