cba84ed65f243ddac5148f04b913c0745f1d2262
[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 = 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 = L": ";
152 ret.fsect[1].just_numbers = TRUE;
153 ret.fsect[1].number_suffix = 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 = 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 = 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 sfree(conf.fsect);
680 }
681
682 static int xhtml_para_level(paragraph *p)
683 {
684 switch (p->type)
685 {
686 case para_Title:
687 return 0;
688 break;
689 case para_UnnumberedChapter:
690 case para_Chapter:
691 case para_Appendix:
692 return 1;
693 break;
694 /* case para_BiblioCited:
695 return 2;
696 break;*/
697 case para_Heading:
698 case para_Subsect:
699 return p->aux+2;
700 break;
701 default:
702 return -1;
703 break;
704 }
705 }
706
707 static char* xhtml_index_filename = "IndexPage.html";
708
709 /* Output the nav links for the current file.
710 * file == NULL means we're doing the index
711 */
712 static void xhtml_donavlinks(FILE *fp, xhtmlfile *file)
713 {
714 xhtmlfile *xhtml_next_file = NULL;
715 fprintf(fp, "<p");
716 if (conf.nav_attrs!=NULL) {
717 fprintf(fp, " %ls>", conf.nav_attrs);
718 } else {
719 fprintf(fp, ">");
720 }
721 if (xhtml_last_file==NULL) {
722 fprintf(fp, "Previous | ");
723 } else {
724 fprintf(fp, "<a href='%s'>Previous</a> | ", xhtml_last_file->filename);
725 }
726 fprintf(fp, "<a href='Contents.html'>Contents</a> | ");
727 if (file == NULL) {
728 fprintf(fp, "Index | ");
729 } else {
730 fprintf(fp, "<a href='%s'>Index</a> | ", xhtml_index_filename);
731 }
732 if (file != NULL) { /* otherwise we're doing nav links for the index */
733 if (xhtml_next_file==NULL)
734 xhtml_next_file = file->child;
735 if (xhtml_next_file==NULL)
736 xhtml_next_file = file->next;
737 if (xhtml_next_file==NULL)
738 xhtml_next_file = file->parent->next;
739 }
740 if (xhtml_next_file==NULL) {
741 if (file==NULL) { /* index, so no next file */
742 fprintf(fp, "Next ");
743 } else {
744 fprintf(fp, "<a href='%s'>Next</a>", xhtml_index_filename);
745 }
746 } else {
747 fprintf(fp, "<a href='%s'>Next</a>", xhtml_next_file->filename);
748 }
749 fprintf(fp, "</p>\n");
750 }
751
752 /* Write out the index file */
753 static void xhtml_do_index_body(FILE *fp)
754 {
755 indexentry *y;
756 int ti;
757
758 if (count234(idx->entries) == 0)
759 return; /* don't write anything at all */
760
761 fprintf(fp, "<dl>\n");
762 /* iterate over idx->entries using the tree functions and display everything */
763 for (ti = 0; (y = (indexentry *)index234(idx->entries, ti)) != NULL; ti++) {
764 if (y->backend_data) {
765 int i;
766 xhtmlindex *xi;
767
768 fprintf(fp, "<dt>");
769 xhtml_para(fp, y->text, FALSE);
770 fprintf(fp, "</dt>\n<dd>");
771
772 xi = (xhtmlindex*) y->backend_data;
773 for (i=0; i<xi->nsection; i++) {
774 xhtmlsection *sect = xi->sections[i];
775 if (sect) {
776 fprintf(fp, "<a href='%s#%s'>", sect->file->filename, sect->fragment);
777 if (sect->para->kwtext) {
778 xhtml_para(fp, sect->para->kwtext, FALSE);
779 } else if (sect->para->words) {
780 xhtml_para(fp, sect->para->words, FALSE);
781 }
782 fprintf(fp, "</a>");
783 if (i+1<xi->nsection) {
784 fprintf(fp, ", ");
785 }
786 }
787 }
788 fprintf(fp, "</dd>\n");
789 }
790 }
791 fprintf(fp, "</dl>\n");
792 }
793 static void xhtml_do_index()
794 {
795 word temp_word = { NULL, NULL, word_Normal, 0, 0, L"Index", { NULL, 0, 0} };
796 FILE *fp = fopen(xhtml_index_filename, "w");
797
798 if (fp==NULL)
799 fatal(err_cantopenw, xhtml_index_filename);
800 xhtml_doheader(fp, &temp_word);
801 xhtml_donavlinks(fp, NULL);
802
803 xhtml_do_index_body(fp);
804
805 xhtml_donavlinks(fp, NULL);
806 xhtml_dofooter(fp);
807 fclose(fp);
808 }
809
810 /* Output the given file. This includes whatever contents at beginning and end, etc. etc. */
811 static void xhtml_do_file(xhtmlfile *file)
812 {
813 FILE *fp = fopen(file->filename, "w");
814 if (fp==NULL)
815 fatal(err_cantopenw, file->filename);
816
817 if (file->sections->para->words) {
818 xhtml_doheader(fp, file->sections->para->words);
819 } else if (file->sections->para->kwtext) {
820 xhtml_doheader(fp, file->sections->para->kwtext);
821 } else {
822 xhtml_doheader(fp, NULL);
823 }
824
825 xhtml_donavlinks(fp, file);
826
827 if (file->is_leaf && conf.leaf_contains_contents &&
828 xhtml_do_contents(NULL, file)>=conf.leaf_smallest_contents)
829 xhtml_do_contents(fp, file);
830 xhtml_do_sections(fp, file->sections);
831 if (!file->is_leaf)
832 xhtml_do_naked_contents(fp, file);
833
834 xhtml_donavlinks(fp, file);
835
836 xhtml_dofooter(fp);
837 fclose(fp);
838
839 xhtml_last_file = file;
840 }
841
842 /* Output the top-level file. */
843 static void xhtml_do_top_file(xhtmlfile *file, paragraph *sourceform)
844 {
845 paragraph *p;
846 int done=FALSE;
847 FILE *fp = fopen(file->filename, "w");
848 if (fp==NULL)
849 fatal(err_cantopenw, file->filename);
850
851 /* Do the title -- only one allowed */
852 for (p = sourceform; p && !done; p = p->next)
853 {
854 if (p->type == para_Title)
855 {
856 xhtml_doheader(fp, p->words);
857 done=TRUE;
858 }
859 }
860 if (!done)
861 xhtml_doheader(fp, NULL /* Eek! */);
862
863 /*
864 * Display the title.
865 */
866 for (p = sourceform; p; p = p->next)
867 {
868 if (p->type == para_Title) {
869 xhtml_heading(fp, p, FALSE);
870 break;
871 }
872 }
873
874 /* Do the preamble */
875 for (p = sourceform; p; p = p->next)
876 {
877 if (p->type == para_Chapter || p->type == para_Heading ||
878 p->type == para_Subsect || p->type == para_Appendix ||
879 p->type == para_UnnumberedChapter) {
880 /*
881 * We've found the end of the preamble. Do every normal
882 * paragraph up to there.
883 */
884 xhtml_do_paras(fp, sourceform, p, FALSE);
885 break;
886 }
887 }
888
889 xhtml_do_contents(fp, file);
890 xhtml_do_sections(fp, file->sections);
891
892 /*
893 * Put the index in the top file if we're in single-file mode
894 * (leaf-level 0).
895 */
896 if (conf.leaf_level == 0 && count234(idx->entries) > 0) {
897 fprintf(fp, "<a name=\"index\"></a><h1>Index</h1>\n");
898 xhtml_do_index_body(fp);
899 }
900
901 xhtml_dofooter(fp);
902 fclose(fp);
903 }
904
905 /* Convert a Unicode string to an ASCII one. '?' is
906 * used for unmappable characters.
907 */
908 static void xhtml_utostr(wchar_t *in, char **out)
909 {
910 int l = ustrlen(in);
911 int i;
912 *out = smalloc(l+1);
913 for (i=0; i<l; i++)
914 {
915 if (in[i]>=32 && in[i]<=126)
916 (*out)[i]=(char)in[i];
917 else
918 (*out)[i]='?';
919 }
920 (*out)[i]=0;
921 }
922
923 /*
924 * Write contents for the given file, and subfiles, down to
925 * the appropriate contents depth. Returns the number of
926 * entries written.
927 */
928 static int xhtml_do_contents(FILE *fp, xhtmlfile *file)
929 {
930 int level, limit, count = 0;
931 if (!file)
932 return 0;
933
934 level = (file->sections)?(file->sections->level):(0);
935 limit = conf.contents_depth[(level>5)?(5):(level)];
936 start_level = (file->is_leaf) ? (level-1) : (level);
937 last_level = start_level;
938
939 count += xhtml_do_contents_section_limit(fp, file->sections, limit);
940 count += xhtml_do_contents_limit(fp, file->child, limit);
941 if (fp!=NULL) {
942 while (last_level > start_level) {
943 last_level--;
944 fprintf(fp, "</li></ul>\n");
945 }
946 }
947 return count;
948 }
949
950 /* As above, but doesn't do anything in the current file */
951 static int xhtml_do_naked_contents(FILE *fp, xhtmlfile *file)
952 {
953 int level, limit, start_level, count = 0;
954 if (!file)
955 return 0;
956
957 level = (file->sections)?(file->sections->level):(0);
958 limit = conf.contents_depth[(level>5)?(5):(level)];
959 start_level = (file->is_leaf) ? (level-1) : (level);
960 last_level = start_level;
961
962 count = xhtml_do_contents_limit(fp, file->child, limit);
963 if (fp!=NULL) {
964 while (last_level > start_level) {
965 last_level--;
966 fprintf(fp, "</li></ul>\n");
967 }
968 }
969 return count;
970 }
971
972 /*
973 * Write contents for the given file, children, and siblings, down to
974 * given limit contents depth.
975 */
976 static int xhtml_do_contents_limit(FILE *fp, xhtmlfile *file, int limit)
977 {
978 int count = 0;
979 while (file) {
980 count += xhtml_do_contents_section_limit(fp, file->sections, limit);
981 count += xhtml_do_contents_limit(fp, file->child, limit);
982 file = file->next;
983 }
984 return count;
985 }
986
987 /*
988 * Write contents entries for the given section tree, down to the
989 * limit contents depth.
990 */
991 static int xhtml_do_contents_section_deep_limit(FILE *fp, xhtmlsection *section, int limit)
992 {
993 int count = 0;
994 while (section) {
995 if (!xhtml_add_contents_entry(fp, section, limit))
996 return 0;
997 else
998 count++;
999 count += xhtml_do_contents_section_deep_limit(fp, section->child, limit);
1000 section = section->next;
1001 }
1002 return count;
1003 }
1004
1005 /*
1006 * Write contents entries for the given section tree, down to the
1007 * limit contents depth.
1008 */
1009 static int xhtml_do_contents_section_limit(FILE *fp, xhtmlsection *section, int limit)
1010 {
1011 int count = 0;
1012 if (!section)
1013 return 0;
1014 xhtml_add_contents_entry(fp, section, limit);
1015 count=1;
1016 count += xhtml_do_contents_section_deep_limit(fp, section->child, limit);
1017 /* section=section->child;
1018 while (section && xhtml_add_contents_entry(fp, section, limit)) {
1019 section = section->next;
1020 }*/
1021 return count;
1022 }
1023
1024 /*
1025 * Add a section entry, unless we're exceeding the limit, in which
1026 * case return FALSE (otherwise return TRUE).
1027 */
1028 static int xhtml_add_contents_entry(FILE *fp, xhtmlsection *section, int limit)
1029 {
1030 if (!section || section->level > limit)
1031 return FALSE;
1032 if (fp==NULL || section->level < 0)
1033 return TRUE;
1034 if (last_level > section->level) {
1035 while (last_level > section->level) {
1036 last_level--;
1037 fprintf(fp, "</li></ul>\n");
1038 }
1039 fprintf(fp, "</li>\n");
1040 } else if (last_level < section->level) {
1041 assert(last_level == section->level - 1);
1042 last_level++;
1043 fprintf(fp, "<ul>\n");
1044 } else {
1045 fprintf(fp, "</li>\n");
1046 }
1047 fprintf(fp, "<li><a href=\"%s#%s\">", section->file->filename, section->fragment);
1048 if (section->para->kwtext) {
1049 xhtml_para(fp, section->para->kwtext, FALSE);
1050 if (section->para->words) {
1051 fprintf(fp, ": ");
1052 }
1053 }
1054 if (section->para->words) {
1055 xhtml_para(fp, section->para->words, FALSE);
1056 }
1057 fprintf(fp, "</a>\n");
1058 return TRUE;
1059 }
1060
1061 /*
1062 * Write all the sections in this file. Do all paragraphs in this section, then all
1063 * children (recursively), then go on to the next one (tail recursively).
1064 */
1065 static void xhtml_do_sections(FILE *fp, xhtmlsection *sections)
1066 {
1067 while (sections) {
1068 currentsection = sections;
1069 xhtml_do_paras(fp, sections->para, NULL, TRUE);
1070 xhtml_do_sections(fp, sections->child);
1071 sections = sections->next;
1072 }
1073 }
1074
1075 /* Write this list of paragraphs. Close off all lists at the end. */
1076 static void xhtml_do_paras(FILE *fp, paragraph *p, paragraph *end,
1077 int indexable)
1078 {
1079 int last_type = -1, ptype, first=TRUE;
1080 stack lcont_stack = stk_new();
1081 if (!p)
1082 return;
1083
1084 /* for (; p && (xhtml_para_level(p)>limit || xhtml_para_level(p)==-1 || first); p=p->next) {*/
1085 for (; p && p != end && (xhtml_para_level(p)==-1 || first); p=p->next) {
1086 first=FALSE;
1087 switch (ptype = p->type)
1088 {
1089 /*
1090 * Things we ignore because we've already processed them or
1091 * aren't going to touch them in this pass.
1092 */
1093 case para_IM:
1094 case para_BR:
1095 case para_Biblio: /* only touch BiblioCited */
1096 case para_VersionID:
1097 case para_NoCite:
1098 case para_Title:
1099 break;
1100
1101 /*
1102 * Chapter titles.
1103 */
1104 case para_Chapter:
1105 case para_Appendix:
1106 case para_UnnumberedChapter:
1107 xhtml_heading(fp, p, indexable);
1108 break;
1109
1110 case para_Heading:
1111 case para_Subsect:
1112 xhtml_heading(fp, p, indexable);
1113 break;
1114
1115 case para_Rule:
1116 fprintf(fp, "\n<hr />\n");
1117 break;
1118
1119 case para_Normal:
1120 case para_Copyright:
1121 fprintf(fp, "\n<p>");
1122 xhtml_para(fp, p->words, indexable);
1123 fprintf(fp, "</p>\n");
1124 break;
1125
1126 case para_LcontPush:
1127 {
1128 int *p;
1129 p = mknew(int);
1130 *p = last_type;
1131 stk_push(lcont_stack, p);
1132 last_type = para_Normal;
1133 }
1134 break;
1135 case para_LcontPop:
1136 {
1137 int *p = stk_pop(lcont_stack);
1138 assert(p);
1139 ptype = last_type = *p;
1140 sfree(p);
1141 goto closeofflist; /* ick */
1142 }
1143 break;
1144 case para_QuotePush:
1145 fprintf(fp, "<blockquote>\n");
1146 break;
1147 case para_QuotePop:
1148 fprintf(fp, "</blockquote>\n");
1149 break;
1150
1151 case para_Bullet:
1152 case para_NumberedList:
1153 case para_Description:
1154 case para_DescribedThing:
1155 case para_BiblioCited:
1156 if (last_type!=p->type &&
1157 !(last_type==para_DescribedThing && p->type==para_Description) &&
1158 !(last_type==para_Description && p->type==para_DescribedThing)) {
1159 /* start up list if necessary */
1160 if (p->type == para_Bullet) {
1161 fprintf(fp, "<ul>\n");
1162 } else if (p->type == para_NumberedList) {
1163 fprintf(fp, "<ol>\n");
1164 } else if (p->type == para_BiblioCited ||
1165 p->type == para_DescribedThing ||
1166 p->type == para_Description) {
1167 fprintf(fp, "<dl>\n");
1168 }
1169 }
1170 if (p->type == para_Bullet || p->type == para_NumberedList) {
1171 fprintf(fp, "<li>");
1172 } else if (p->type == para_DescribedThing) {
1173 fprintf(fp, "<dt>");
1174 } else if (p->type == para_Description) {
1175 fprintf(fp, "<dd>");
1176 } else if (p->type == para_BiblioCited) {
1177 fprintf(fp, "<dt>");
1178 xhtml_para(fp, p->kwtext, indexable);
1179 fprintf(fp, "</dt>\n<dd>");
1180 }
1181 xhtml_para(fp, p->words, indexable);
1182 {
1183 paragraph *p2 = p->next;
1184 if (p2 && xhtml_para_level(p2)==-1 && p2->type == para_LcontPush)
1185 break;
1186 }
1187
1188 closeofflist:
1189 if (ptype == para_BiblioCited) {
1190 fprintf(fp, "</dd>\n");
1191 } else if (ptype == para_DescribedThing) {
1192 fprintf(fp, "</dt>");
1193 } else if (ptype == para_Description) {
1194 fprintf(fp, "</dd>");
1195 } else if (ptype == para_Bullet || ptype == para_NumberedList) {
1196 fprintf(fp, "</li>");
1197 }
1198 if (ptype == para_Bullet || ptype == para_NumberedList ||
1199 ptype == para_BiblioCited || ptype == para_Description ||
1200 ptype == para_DescribedThing)
1201 /* close off list if necessary */
1202 {
1203 paragraph *p2 = p->next;
1204 int close_off=FALSE;
1205 /* if (p2 && (xhtml_para_level(p2)>limit || xhtml_para_level(p2)==-1)) {*/
1206 if (p2 && xhtml_para_level(p2)==-1) {
1207 if (p2->type != ptype &&
1208 !(p2->type==para_DescribedThing && ptype==para_Description) &&
1209 !(p2->type==para_Description && ptype==para_DescribedThing) &&
1210 p2->type != para_LcontPush)
1211 close_off=TRUE;
1212 } else {
1213 close_off=TRUE;
1214 }
1215 if (close_off) {
1216 if (ptype == para_Bullet) {
1217 fprintf(fp, "</ul>\n");
1218 } else if (ptype == para_NumberedList) {
1219 fprintf(fp, "</ol>\n");
1220 } else if (ptype == para_BiblioCited ||
1221 ptype == para_Description ||
1222 ptype == para_DescribedThing) {
1223 fprintf(fp, "</dl>\n");
1224 }
1225 }
1226 }
1227 break;
1228
1229 case para_Code:
1230 xhtml_codepara(fp, p->words);
1231 break;
1232 }
1233 last_type = ptype;
1234 }
1235
1236 stk_free(lcont_stack);
1237 }
1238
1239 /*
1240 * Output a header for this XHTML file.
1241 */
1242 static void xhtml_doheader(FILE *fp, word *title)
1243 {
1244 fprintf(fp, "<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Transitional//EN\"\n");
1245 fprintf(fp, "\"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd\">\n");
1246 fprintf(fp, "<html xmlns='http://www.w3.org/1999/xhtml'>\n\n<head>\n<title>");
1247 if (title==NULL)
1248 fprintf(fp, "The thing with no name!");
1249 else
1250 xhtml_para(fp, title, FALSE);
1251 fprintf(fp, "</title>\n");
1252 fprintf(fp, "<meta name=\"generator\" content=\"Halibut %s xhtml-backend\" />\n", version);
1253 if (conf.author)
1254 fprintf(fp, "<meta name=\"author\" content=\"%ls\" />\n", conf.author);
1255 if (conf.description)
1256 fprintf(fp, "<meta name=\"description\" content=\"%ls\" />\n", conf.description);
1257 if (conf.head_end)
1258 fprintf(fp, "%ls\n", conf.head_end);
1259 fprintf(fp, "</head>\n\n");
1260 if (conf.body)
1261 fprintf(fp, "%ls\n", conf.body);
1262 else
1263 fprintf(fp, "<body>\n");
1264 if (conf.body_start)
1265 fprintf(fp, "%ls\n", conf.body_start);
1266 }
1267
1268 /*
1269 * Output a footer for this XHTML file.
1270 */
1271 static void xhtml_dofooter(FILE *fp)
1272 {
1273 fprintf(fp, "\n<hr />\n\n");
1274 if (conf.body_end)
1275 fprintf(fp, "%ls\n", conf.body_end);
1276 if (!conf.suppress_address) {
1277 fprintf(fp,"<address>\n");
1278 if (conf.address_start)
1279 fprintf(fp, "%ls\n", conf.address_start);
1280 /* Do the version ID */
1281 if (conf.include_version_id) {
1282 paragraph *p;
1283 int started = 0;
1284 for (p = sourceparas; p; p = p->next)
1285 if (p->type == para_VersionID) {
1286 xhtml_versionid(fp, p->words, started);
1287 started = 1;
1288 }
1289 }
1290 if (conf.address_end)
1291 fprintf(fp, "%ls\n", conf.address_end);
1292 fprintf(fp, "</address>\n");
1293 }
1294 fprintf(fp, "</body>\n\n</html>\n");
1295 }
1296
1297 /*
1298 * Output the versionid paragraph. Typically this is a version control
1299 * ID string (such as $Id...$ in RCS).
1300 */
1301 static void xhtml_versionid(FILE *fp, word *text, int started)
1302 {
1303 rdstringc t = { 0, 0, NULL };
1304
1305 rdaddc(&t, '['); /* FIXME: configurability */
1306 xhtml_rdaddwc(&t, text, NULL, FALSE);
1307 rdaddc(&t, ']'); /* FIXME: configurability */
1308
1309 if (started)
1310 fprintf(fp, "<br />\n");
1311 fprintf(fp, "%s\n", t.text);
1312 sfree(t.text);
1313 }
1314
1315 /* Is this an XHTML reserved character? */
1316 static int xhtml_reservedchar(int c)
1317 {
1318 if (c=='&' || c=='<' || c=='>' || c=='"')
1319 return TRUE;
1320 else
1321 return FALSE;
1322 }
1323
1324 /*
1325 * Convert a wide string into valid XHTML: Anything outside ASCII will
1326 * be fixed up as an entity. Currently we don't worry about constraining the
1327 * encoded character set, which we should probably do at some point (we can
1328 * still fix up and return FALSE - see the last comment here). We also don't
1329 * currently
1330 *
1331 * Because this is only used for words, spaces are HARD spaces (any other
1332 * spaces will be word_Whitespace not word_Normal). So they become &nbsp;
1333 * Unless hard_spaces is FALSE, of course (code paragraphs break the above
1334 * rule).
1335 *
1336 * If `result' is non-NULL, mallocs the resulting string and stores a pointer to
1337 * it in `*result'. If `result' is NULL, merely checks whether all
1338 * characters in the string are feasible.
1339 *
1340 * Return is nonzero if all characters are OK. If not all
1341 * characters are OK but `result' is non-NULL, a result _will_
1342 * still be generated!
1343 */
1344 static int xhtml_convert(wchar_t *s, int maxlen, char **result,
1345 int hard_spaces) {
1346 int doing = (result != 0);
1347 int ok = TRUE;
1348 char *p = NULL;
1349 int plen = 0, psize = 0;
1350
1351 if (maxlen <= 0)
1352 maxlen = -1;
1353
1354 for (; *s && maxlen != 0; s++, maxlen--) {
1355 wchar_t c = *s;
1356
1357 #define ensure_size(i) if (i>=psize) { psize = i+256; p = resize(p, psize); }
1358
1359 if (((c == 32 && !hard_spaces) || (c > 32 && c <= 126 && !xhtml_reservedchar(c)))) {
1360 /* Char is OK. */
1361 if (doing)
1362 {
1363 ensure_size(plen);
1364 p[plen++] = (char)c;
1365 }
1366 } else {
1367 /* Char needs fixing up. */
1368 /* ok = FALSE; -- currently we never return FALSE; we
1369 * might want to when considering a character set for the
1370 * encoded document.
1371 */
1372 if (doing)
1373 {
1374 if (c==32) { /* a space in a word is a hard space */
1375 ensure_size(plen+6); /* includes space for the NUL, which is subsequently stomped on */
1376 sprintf(p+plen, "&nbsp;");
1377 plen+=6;
1378 } else {
1379 /* FIXME: entity names! */
1380 ensure_size(plen+8); /* includes space for the NUL, which is subsequently stomped on */
1381 plen+=sprintf(p+plen, "&#%04i;", (int)c);
1382 }
1383 }
1384 }
1385 }
1386 if (doing) {
1387 p = resize(p, plen+1);
1388 p[plen] = '\0';
1389 *result = p;
1390 }
1391 return ok;
1392 }
1393
1394 /*
1395 * This formats the given words as XHTML.
1396 *
1397 * `indexable', if FALSE, prohibits adding any index references.
1398 * You might use this, for example, if an index reference occurred
1399 * in a section title, to prevent phony index references when the
1400 * section title is processed in strange places such as contents
1401 * sections.
1402 */
1403 static void xhtml_rdaddwc(rdstringc *rs, word *text, word *end, int indexable) {
1404 char *c;
1405 keyword *kwl;
1406 xhtmlsection *sect;
1407 indextag *itag;
1408 int ti;
1409
1410 for (; text && text != end; text = text->next) {
1411 switch (text->type) {
1412 case word_HyperLink:
1413 xhtml_utostr(text->text, &c);
1414 rdaddsc(rs, "<a href=\"");
1415 rdaddsc(rs, c);
1416 rdaddsc(rs, "\">");
1417 sfree(c);
1418 break;
1419
1420 case word_UpperXref:
1421 case word_LowerXref:
1422 kwl = kw_lookup(keywords, text->text);
1423 if (kwl) {
1424 sect=xhtml_find_section(kwl->para);
1425 if (sect) {
1426 rdaddsc(rs, "<a href=\"");
1427 rdaddsc(rs, sect->file->filename);
1428 rdaddc(rs, '#');
1429 rdaddsc(rs, sect->fragment);
1430 rdaddsc(rs, "\">");
1431 } else {
1432 rdaddsc(rs, "<a href=\"Apologies.html\"><!-- probably a bibliography cross reference -->");
1433 error(err_whatever, "Couldn't locate cross-reference! (Probably a bibliography entry.)");
1434 }
1435 } else {
1436 rdaddsc(rs, "<a href=\"Apologies.html\"><!-- unknown cross-reference -->");
1437 error(err_whatever, "Couldn't locate cross-reference! (Wasn't in source file.)");
1438 }
1439 break;
1440
1441 case word_IndexRef: /* in theory we could make an index target here */
1442 /* rdaddsc(rs, "<a name=\"idx-");
1443 xhtml_utostr(text->text, &c);
1444 rdaddsc(rs, c);
1445 sfree(c);
1446 rdaddsc(rs, "\"></a>");*/
1447 /* what we _do_ need to do is to fix up the backend data
1448 * for any indexentry this points to.
1449 */
1450 if (!indexable)
1451 break;
1452
1453 for (ti=0; (itag = (indextag *)index234(idx->tags, ti))!=NULL; ti++) {
1454 /* FIXME: really ustricmp() and not ustrcmp()? */
1455 if (ustricmp(itag->name, text->text)==0) {
1456 break;
1457 }
1458 }
1459 if (itag!=NULL) {
1460 if (itag->refs!=NULL) {
1461 int i;
1462 for (i=0; i<itag->nrefs; i++) {
1463 xhtmlindex *idx_ref;
1464 indexentry *ientry;
1465
1466 ientry = itag->refs[i];
1467 if (ientry->backend_data==NULL) {
1468 idx_ref = (xhtmlindex*) smalloc(sizeof(xhtmlindex));
1469 if (idx_ref==NULL)
1470 fatal(err_nomemory);
1471 idx_ref->nsection = 0;
1472 idx_ref->size = 4;
1473 idx_ref->sections = (xhtmlsection**) smalloc(idx_ref->size * sizeof(xhtmlsection*));
1474 if (idx_ref->sections==NULL)
1475 fatal(err_nomemory);
1476 ientry->backend_data = idx_ref;
1477 } else {
1478 idx_ref = ientry->backend_data;
1479 if (idx_ref->nsection+1 > idx_ref->size) {
1480 int new_size = idx_ref->size * 2;
1481 idx_ref->sections = srealloc(idx_ref->sections, new_size * sizeof(xhtmlsection));
1482 if (idx_ref->sections==NULL) {
1483 fatal(err_nomemory);
1484 }
1485 idx_ref->size = new_size;
1486 }
1487 }
1488 idx_ref->sections[idx_ref->nsection++] = currentsection;
1489 #if 0
1490 #endif
1491 }
1492 } else {
1493 fatal(err_whatever, "Index tag had no entries!");
1494 }
1495 } else {
1496 fprintf(stderr, "Looking for index entry '%ls'\n", text->text);
1497 fatal(err_whatever, "Couldn't locate index entry! (Wasn't in index.)");
1498 }
1499 break;
1500
1501 case word_HyperEnd:
1502 case word_XrefEnd:
1503 rdaddsc(rs, "</a>");
1504 break;
1505
1506 case word_Normal:
1507 case word_Emph:
1508 case word_Code:
1509 case word_WeakCode:
1510 case word_WhiteSpace:
1511 case word_EmphSpace:
1512 case word_CodeSpace:
1513 case word_WkCodeSpace:
1514 case word_Quote:
1515 case word_EmphQuote:
1516 case word_CodeQuote:
1517 case word_WkCodeQuote:
1518 assert(text->type != word_CodeQuote &&
1519 text->type != word_WkCodeQuote);
1520 if (towordstyle(text->type) == word_Emph &&
1521 (attraux(text->aux) == attr_First ||
1522 attraux(text->aux) == attr_Only))
1523 rdaddsc(rs, "<em>");
1524 else if ((towordstyle(text->type) == word_Code || towordstyle(text->type) == word_WeakCode) &&
1525 (attraux(text->aux) == attr_First ||
1526 attraux(text->aux) == attr_Only))
1527 rdaddsc(rs, "<code>");
1528
1529 if (removeattr(text->type) == word_Normal) {
1530 if (xhtml_convert(text->text, 0, &c, TRUE)) /* spaces in the word are hard */
1531 rdaddsc(rs, c);
1532 else
1533 xhtml_rdaddwc(rs, text->alt, NULL, indexable);
1534 sfree(c);
1535 } else if (removeattr(text->type) == word_WhiteSpace) {
1536 rdaddc(rs, ' ');
1537 } else if (removeattr(text->type) == word_Quote) {
1538 rdaddsc(rs, "&quot;");
1539 }
1540
1541 if (towordstyle(text->type) == word_Emph &&
1542 (attraux(text->aux) == attr_Last ||
1543 attraux(text->aux) == attr_Only))
1544 rdaddsc(rs, "</em>");
1545 else if ((towordstyle(text->type) == word_Code || towordstyle(text->type) == word_WeakCode) &&
1546 (attraux(text->aux) == attr_Last ||
1547 attraux(text->aux) == attr_Only))
1548 rdaddsc(rs, "</code>");
1549 break;
1550 }
1551 }
1552 }
1553
1554 /* Output a heading, formatted as XHTML.
1555 */
1556 static void xhtml_heading(FILE *fp, paragraph *p, int indexable)
1557 {
1558 rdstringc t = { 0, 0, NULL };
1559 word *tprefix = p->kwtext;
1560 word *nprefix = p->kwtext2;
1561 word *text = p->words;
1562 int level = xhtml_para_level(p);
1563 xhtmlsection *sect = xhtml_find_section(p);
1564 xhtmlheadfmt *fmt;
1565 char *fragment;
1566 if (sect) {
1567 fragment = sect->fragment;
1568 } else {
1569 if (p->type == para_Title)
1570 fragment = "title";
1571 else {
1572 fragment = ""; /* FIXME: what else can we do? */
1573 error(err_whatever, "Couldn't locate heading cross-reference!");
1574 }
1575 }
1576
1577 if (p->type == para_Title)
1578 fmt = NULL;
1579 else if (level == 1)
1580 fmt = &conf.fchapter;
1581 else if (level-1 < conf.nfsect)
1582 fmt = &conf.fsect[level-1];
1583 else
1584 fmt = &conf.fsect[conf.nfsect-1];
1585
1586 if (fmt && fmt->just_numbers && nprefix) {
1587 xhtml_rdaddwc(&t, nprefix, NULL, indexable);
1588 if (fmt) {
1589 char *c;
1590 if (xhtml_convert(fmt->number_suffix, 0, &c, FALSE)) {
1591 rdaddsc(&t, c);
1592 sfree(c);
1593 }
1594 }
1595 } else if (fmt && !fmt->just_numbers && tprefix) {
1596 xhtml_rdaddwc(&t, tprefix, NULL, indexable);
1597 if (fmt) {
1598 char *c;
1599 if (xhtml_convert(fmt->number_suffix, 0, &c, FALSE)) {
1600 rdaddsc(&t, c);
1601 sfree(c);
1602 }
1603 }
1604 }
1605 xhtml_rdaddwc(&t, text, NULL, indexable);
1606 /*
1607 * If we're outputting in single-file mode, we need to lower
1608 * the level of each heading by one, because the overall
1609 * document title will be sitting right at the top as an <h1>
1610 * and so chapters and sections should start at <h2>.
1611 *
1612 * Even if not, the document title will come back from
1613 * xhtml_para_level() as level zero, so we must increment that
1614 * no matter what leaf_level is set to.
1615 */
1616 if (conf.leaf_level == 0 || level == 0)
1617 level++;
1618 fprintf(fp, "<a name=\"%s\"></a><h%i>%s</h%i>\n", fragment, level, t.text, level);
1619 sfree(t.text);
1620 }
1621
1622 /* Output a paragraph. Styles are handled by xhtml_rdaddwc().
1623 * This looks pretty simple; I may have missed something ...
1624 */
1625 static void xhtml_para(FILE *fp, word *text, int indexable)
1626 {
1627 rdstringc out = { 0, 0, NULL };
1628 xhtml_rdaddwc(&out, text, NULL, indexable);
1629 fprintf(fp, "%s", out.text);
1630 sfree(out.text);
1631 }
1632
1633 /* Output a code paragraph. I'm treating this as preformatted, which
1634 * may not be entirely correct. See xhtml_para() for my worries about
1635 * this being overly-simple; however I think that most of the complexity
1636 * of the text backend came entirely out of word wrapping anyway.
1637 */
1638 static void xhtml_codepara(FILE *fp, word *text)
1639 {
1640 fprintf(fp, "<pre>");
1641 for (; text; text = text->next) if (text->type == word_WeakCode) {
1642 word *here, *next;
1643 char *c;
1644
1645 /*
1646 * See if this WeakCode is followed by an Emph to indicate
1647 * emphasis.
1648 */
1649 here = text;
1650 if (text->next && text->next->type == word_Emph) {
1651 next = text = text->next;
1652 } else
1653 next = NULL;
1654
1655 if (next) {
1656 wchar_t *t, *e;
1657 int n;
1658
1659 t = here->text;
1660 e = next->text;
1661
1662 while (*e) {
1663 int ec = *e;
1664
1665 for (n = 0; t[n] && e[n] && e[n] == ec; n++);
1666 xhtml_convert(t, n, &c, FALSE);
1667 fprintf(fp, "%s%s%s",
1668 (ec == 'i' ? "<em>" : ec == 'b' ? "<b>" : ""),
1669 c,
1670 (ec == 'i' ? "</em>" : ec == 'b' ? "</b>" : ""));
1671 sfree(c);
1672
1673 t += n;
1674 e += n;
1675 }
1676
1677 xhtml_convert(t, 0, &c, FALSE);
1678 fprintf(fp, "%s\n", c);
1679 sfree(c);
1680 } else {
1681 xhtml_convert(here->text, 0, &c, FALSE);
1682 fprintf(fp, "%s\n", c);
1683 sfree(c);
1684 }
1685 }
1686 fprintf(fp, "</pre>\n");
1687 }