2 * xhtml backend for Halibut
3 * (initial implementation by James Aylett)
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(),
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.
36 * FILENAME_TEMPLATE (overridable in config of course) allows you
37 * to choose the general form for your HTML file names. It is
38 * slightly printf-styled (% followed by a single character is a
39 * formatting directive, %% is a literal %). Formatting directives
42 * - %n is the section type-plus-number, minus whitespace (`Chapter1.2').
43 * - %b is the section number on its own (`1.2').
44 * - %k is the section's _internal_ keyword.
45 * - %N is the section's visible title in the output, again minus
48 * %n, %b and %k will all default to %N if the section is
49 * unnumbered (`Bibliography' is often a good example).
52 #define FILENAME_SINGLE "Manual.html"
53 #define FILENAME_CONTENTS "Contents.html"
54 #define FILENAME_INDEX "IndexPage.html"
55 #define FILENAME_TEMPLATE "%n.html"
57 struct xhtmlsection_Struct
{
58 struct xhtmlsection_Struct
*next
; /* next sibling (NULL if split across files) */
59 struct xhtmlsection_Struct
*child
; /* NULL if split across files */
60 struct xhtmlsection_Struct
*parent
; /* NULL if split across files */
61 struct xhtmlsection_Struct
*chain
; /* single structure independent of weird trees */
63 struct xhtmlfile_Struct
*file
; /* which file is this a part of? */
64 char *fragment
; /* fragment id within the file */
68 struct xhtmlfile_Struct
{
69 struct xhtmlfile_Struct
*next
;
70 struct xhtmlfile_Struct
*child
;
71 struct xhtmlfile_Struct
*parent
;
73 struct xhtmlsection_Struct
*sections
; /* sections within this file (only one for non-leaf) */
74 int is_leaf
; /* is this file a leaf file, ie does it not have any children? */
77 typedef struct xhtmlsection_Struct xhtmlsection
;
78 typedef struct xhtmlfile_Struct xhtmlfile
;
79 typedef struct xhtmlindex_Struct xhtmlindex
;
81 struct xhtmlindex_Struct
{
84 xhtmlsection
**sections
;
89 wchar_t *number_suffix
;
93 int contents_depth
[6];
94 int leaf_contains_contents
;
96 int leaf_smallest_contents
;
97 int include_version_id
;
98 wchar_t *author
, *description
;
99 wchar_t *head_end
, *body
, *body_start
, *body_end
, *address_start
, *address_end
, *nav_attrs
;
100 int suppress_address
;
101 xhtmlheadfmt fchapter
, *fsect
;
103 char *contents_filename
, *index_filename
;
104 char *single_filename
, *template_filename
;
107 /*static void xhtml_level(paragraph *, int);
108 static void xhtml_level_0(paragraph *);
109 static void xhtml_docontents(FILE *, paragraph *, int);
110 static void xhtml_dosections(FILE *, paragraph *, int);
111 static void xhtml_dobody(FILE *, paragraph *, int);*/
113 static void xhtml_doheader(FILE *, word
*);
114 static void xhtml_dofooter(FILE *);
115 static void xhtml_versionid(FILE *, word
*, int);
117 static void xhtml_utostr(wchar_t *, char **);
118 static int xhtml_para_level(paragraph
*);
119 static int xhtml_reservedchar(int);
121 static int xhtml_convert(wchar_t *, int, char **, int);
122 static void xhtml_rdaddwc(rdstringc
*, word
*, word
*, int);
123 static void xhtml_para(FILE *, word
*, int);
124 static void xhtml_codepara(FILE *, word
*);
125 static void xhtml_heading(FILE *, paragraph
*, int);
127 /* File-global variables are much easier than passing these things
128 * all over the place. Evil, but easier. We can replace this with a single
129 * structure at some point.
131 static xhtmlconfig conf
;
132 static keywordlist
*keywords
;
133 static indexdata
*idx
;
134 static xhtmlfile
*topfile
;
135 static xhtmlsection
*topsection
;
136 static paragraph
*sourceparas
;
137 static xhtmlfile
*lastfile
;
138 static xhtmlfile
*xhtml_last_file
= NULL
;
139 static int last_level
=-1, start_level
;
140 static xhtmlsection
*currentsection
;
142 static xhtmlconfig
xhtml_configure(paragraph
*source
)
149 ret
.contents_depth
[0] = 2;
150 ret
.contents_depth
[1] = 3;
151 ret
.contents_depth
[2] = 4;
152 ret
.contents_depth
[3] = 5;
153 ret
.contents_depth
[4] = 6;
154 ret
.contents_depth
[5] = 7;
156 ret
.leaf_smallest_contents
= 4;
157 ret
.leaf_contains_contents
= FALSE
;
158 ret
.include_version_id
= TRUE
;
160 ret
.description
= NULL
;
163 ret
.body_start
= NULL
;
165 ret
.address_start
= NULL
;
166 ret
.address_end
= NULL
;
167 ret
.nav_attrs
= NULL
;
168 ret
.suppress_address
= FALSE
;
170 ret
.fchapter
.just_numbers
= FALSE
;
171 ret
.fchapter
.number_suffix
= L
": ";
173 ret
.fsect
= mknewa(xhtmlheadfmt
, ret
.nfsect
);
174 ret
.fsect
[0].just_numbers
= FALSE
;
175 ret
.fsect
[0].number_suffix
= L
": ";
176 ret
.fsect
[1].just_numbers
= TRUE
;
177 ret
.fsect
[1].number_suffix
= L
" ";
178 ret
.contents_filename
= strdup(FILENAME_CONTENTS
);
179 ret
.single_filename
= strdup(FILENAME_SINGLE
);
180 ret
.index_filename
= strdup(FILENAME_INDEX
);
181 ret
.template_filename
= strdup(FILENAME_TEMPLATE
);
183 for (; source
; source
= source
->next
)
185 if (source
->type
== para_Config
)
187 if (!ustricmp(source
->keyword
, L
"xhtml-contents-filename")) {
188 sfree(ret
.contents_filename
);
189 ret
.contents_filename
= utoa_dup(uadv(source
->keyword
));
190 } else if (!ustricmp(source
->keyword
, L
"xhtml-single-filename")) {
191 sfree(ret
.single_filename
);
192 ret
.single_filename
= utoa_dup(uadv(source
->keyword
));
193 } else if (!ustricmp(source
->keyword
, L
"xhtml-index-filename")) {
194 sfree(ret
.index_filename
);
195 ret
.index_filename
= utoa_dup(uadv(source
->keyword
));
196 } else if (!ustricmp(source
->keyword
, L
"xhtml-template-filename")) {
197 sfree(ret
.template_filename
);
198 ret
.template_filename
= utoa_dup(uadv(source
->keyword
));
199 } else if (!ustricmp(source
->keyword
, L
"xhtml-contents-depth-0")) {
200 ret
.contents_depth
[0] = utoi(uadv(source
->keyword
));
201 } else if (!ustricmp(source
->keyword
, L
"xhtml-contents-depth-1")) {
202 ret
.contents_depth
[1] = utoi(uadv(source
->keyword
));
203 } else if (!ustricmp(source
->keyword
, L
"xhtml-contents-depth-2")) {
204 ret
.contents_depth
[2] = utoi(uadv(source
->keyword
));
205 } else if (!ustricmp(source
->keyword
, L
"xhtml-contents-depth-3")) {
206 ret
.contents_depth
[3] = utoi(uadv(source
->keyword
));
207 } else if (!ustricmp(source
->keyword
, L
"xhtml-contents-depth-4")) {
208 ret
.contents_depth
[4] = utoi(uadv(source
->keyword
));
209 } else if (!ustricmp(source
->keyword
, L
"xhtml-contents-depth-5")) {
210 ret
.contents_depth
[5] = utoi(uadv(source
->keyword
));
211 } else if (!ustricmp(source
->keyword
, L
"xhtml-leaf-level")) {
212 ret
.leaf_level
= utoi(uadv(source
->keyword
));
213 } else if (!ustricmp(source
->keyword
, L
"xhtml-leaf-smallest-contents")) {
214 ret
.leaf_smallest_contents
= utoi(uadv(source
->keyword
));
215 } else if (!ustricmp(source
->keyword
, L
"xhtml-versionid")) {
216 ret
.include_version_id
= utob(uadv(source
->keyword
));
217 } else if (!ustricmp(source
->keyword
, L
"xhtml-leaf-contains-contents")) {
218 ret
.leaf_contains_contents
= utob(uadv(source
->keyword
));
219 } else if (!ustricmp(source
->keyword
, L
"xhtml-suppress-address")) {
220 ret
.suppress_address
= utob(uadv(source
->keyword
));
221 } else if (!ustricmp(source
->keyword
, L
"xhtml-author")) {
222 ret
.author
= uadv(source
->keyword
);
223 } else if (!ustricmp(source
->keyword
, L
"xhtml-description")) {
224 ret
.description
= uadv(source
->keyword
);
225 } else if (!ustricmp(source
->keyword
, L
"xhtml-head-end")) {
226 ret
.head_end
= uadv(source
->keyword
);
227 } else if (!ustricmp(source
->keyword
, L
"xhtml-body-start")) {
228 ret
.body_start
= uadv(source
->keyword
);
229 } else if (!ustricmp(source
->keyword
, L
"xhtml-body-tag")) {
230 ret
.body
= uadv(source
->keyword
);
231 } else if (!ustricmp(source
->keyword
, L
"xhtml-body-end")) {
232 ret
.body_end
= uadv(source
->keyword
);
233 } else if (!ustricmp(source
->keyword
, L
"xhtml-address-start")) {
234 ret
.address_start
= uadv(source
->keyword
);
235 } else if (!ustricmp(source
->keyword
, L
"xhtml-address-end")) {
236 ret
.address_end
= uadv(source
->keyword
);
237 } else if (!ustricmp(source
->keyword
, L
"xhtml-navigation-attributes")) {
238 ret
.nav_attrs
= uadv(source
->keyword
);
239 } else if (!ustricmp(source
->keyword
, L
"xhtml-chapter-numeric")) {
240 ret
.fchapter
.just_numbers
= utob(uadv(source
->keyword
));
241 } else if (!ustricmp(source
->keyword
, L
"xhtml-chapter-suffix")) {
242 ret
.fchapter
.number_suffix
= uadv(source
->keyword
);
243 } else if (!ustricmp(source
->keyword
, L
"xhtml-section-numeric")) {
244 wchar_t *p
= uadv(source
->keyword
);
250 if (n
>= ret
.nfsect
) {
252 ret
.fsect
= resize(ret
.fsect
, n
+1);
253 for (i
= ret
.nfsect
; i
<= n
; i
++)
254 ret
.fsect
[i
] = ret
.fsect
[ret
.nfsect
-1];
257 ret
.fsect
[n
].just_numbers
= utob(p
);
258 } else if (!ustricmp(source
->keyword
, L
"xhtml-section-suffix")) {
259 wchar_t *p
= uadv(source
->keyword
);
265 if (n
>= ret
.nfsect
) {
267 ret
.fsect
= resize(ret
.fsect
, n
+1);
268 for (i
= ret
.nfsect
; i
<= n
; i
++)
269 ret
.fsect
[i
] = ret
.fsect
[ret
.nfsect
-1];
272 ret
.fsect
[n
].number_suffix
= p
;
277 /* printf(" !!! leaf_level = %i\n", ret.leaf_level);
278 printf(" !!! contentdepth-0 = %i\n", ret.contents_depth[0]);
279 printf(" !!! contentdepth-1 = %i\n", ret.contents_depth[1]);
280 printf(" !!! contentdepth-2 = %i\n", ret.contents_depth[2]);
281 printf(" !!! contentdepth-3 = %i\n", ret.contents_depth[3]);
282 printf(" !!! contentdepth-4 = %i\n", ret.contents_depth[4]);
283 printf(" !!! contentdepth-5 = %i\n", ret.contents_depth[5]);
284 printf(" !!! leaf_contains_contents = %i\n", ret.leaf_contains_contents);*/
288 paragraph
*xhtml_config_filename(char *filename
)
291 * If the user passes in a single filename as a parameter to
292 * the `--html' command-line option, then we should assume it
293 * to imply _two_ config directives:
294 * \cfg{xhtml-single-filename}{whatever} and
295 * \cfg{xhtml-leaf-level}{0}; the rationale being that the user
296 * wants their output _in that file_.
301 wchar_t *ufilename
, *up
;
303 for (i
= 0; i
< 2; i
++) {
304 p
[i
] = mknew(paragraph
);
305 memset(p
[i
], 0, sizeof(*p
[i
]));
306 p
[i
]->type
= para_Config
;
308 p
[i
]->fpos
.filename
= "<command line>";
309 p
[i
]->fpos
.line
= p
[i
]->fpos
.col
= -1;
312 ufilename
= ufroma_dup(filename
);
313 len
= ustrlen(ufilename
) + 2 + lenof(L
"xhtml-single-filename");
314 p
[0]->keyword
= mknewa(wchar_t, len
);
316 ustrcpy(up
, L
"xhtml-single-filename");
318 ustrcpy(up
, ufilename
);
321 assert(up
- p
[0]->keyword
< len
);
324 len
= lenof(L
"xhtml-leaf-level") + lenof(L
"0") + 1;
325 p
[1]->keyword
= mknewa(wchar_t, len
);
327 ustrcpy(up
, L
"xhtml-leaf-level");
332 assert(up
- p
[1]->keyword
< len
);
339 static xhtmlsection
*xhtml_new_section(xhtmlsection
*last
)
341 xhtmlsection
*ret
= mknew(xhtmlsection
);
349 ret
->level
=-1; /* marker: end of chain */
353 /* Returns NULL or the section that marks that paragraph */
354 static xhtmlsection
*xhtml_find_section(paragraph
*p
)
356 xhtmlsection
*ret
= topsection
;
357 if (xhtml_para_level(p
)==-1) { /* first, we back-track to a section paragraph */
358 paragraph
*p2
= sourceparas
;
359 paragraph
*p3
= NULL
;
360 while (p2
&& p2
!=p
) {
361 if (xhtml_para_level(p2
)!=-1) {
366 if (p3
==NULL
) { /* for some reason, we couldn't find a section before this paragraph ... ? */
367 /* Note that this can happen, if you have a cross-reference to before the first chapter starts.
368 * So don't do that, then.
374 while (ret
&& ret
->para
!= p
) {
375 /* printf(" xhtml_find_section(): checking %s for para @ %p\n", ret->fragment, p);*/
381 static xhtmlfile
*xhtml_new_file(xhtmlsection
*sect
)
383 xhtmlfile
*ret
= mknew(xhtmlfile
);
390 ret
->is_leaf
=(sect
!=NULL
&& sect
->level
==conf
.leaf_level
);
392 if (conf
.leaf_level
==0) { /* currently unused */
393 ret
->filename
= smalloc(strlen(conf
.single_filename
)+1);
394 sprintf(ret
->filename
, conf
.single_filename
);
396 ret
->filename
= smalloc(strlen(conf
.contents_filename
)+1);
397 sprintf(ret
->filename
, conf
.contents_filename
);
400 paragraph
*p
= sect
->para
;
401 rdstringc fname_c
= { 0, 0, NULL
};
406 t
= conf
.template_filename
;
408 if (*t
== '%' && t
[1]) {
415 rdaddc(&fname_c
, fmt
);
422 if (p
->kwtext
&& fmt
== 'n')
424 else if (p
->kwtext2
&& fmt
== 'b')
426 else if (p
->keyword
&& *p
->keyword
&& fmt
== 'k')
432 switch (removeattr(w
->type
))
437 case word_WeakCode:*/
438 xhtml_utostr(w
->text
, &c
);
446 xhtml_utostr(ws
, &c
);
451 rdaddc(&fname_c
, *t
++);
455 ret
->filename
= rdtrimc(&fname_c
);
457 /* printf(" ! new file '%s', is_leaf == %s\n", ret->filename, (ret->is_leaf)?("true"):("false"));*/
462 * Walk the tree fixing up files which are actually leaf (ie
463 * have no children) but aren't at leaf level, so they have the
466 void xhtml_fixup_layout(xhtmlfile
* file
)
468 if (file
->child
==NULL
) {
469 file
->is_leaf
= TRUE
;
471 xhtml_fixup_layout(file
->child
);
474 xhtml_fixup_layout(file
->next
);
478 * Create the tree structure so we know where everything goes.
481 * Ignoring file splitting, we have three choices with each new section:
483 * +-----------------+-----------------+
491 * Y is the last section we added (currentsect).
492 * If sect is the section we want to add, then:
494 * (1) if sect->level < currentsect->level
495 * (2) if sect->level == currentsect->level
496 * (3) if sect->level > currentsect->level
498 * This requires the constraint that you never skip section numbers
499 * (so you can't have a.b.c.d without all of a, a.b and a.b.c existing).
501 * Note that you _can_ have 1.1.1.1 followed by 1.2 - you can change
502 * more than one level at a time. Lots of asserts, and probably part of
503 * the algorithm here, rely on this being true. (It currently isn't
504 * enforced by halibut, however.)
506 * File splitting makes this harder. For instance, say we added at (3)
507 * above and now need to add another section. We are splitting at level
508 * 2, ie the level of Y. Z is the last section we added:
510 * +-----------------+-----------------+
520 * The (1) case is now split; we need to search upwards to find where
521 * to actually link in. The other two cases remain the same (and will
522 * always be like this).
524 * File splitting makes this harder, however. The decision of whether
525 * to split to a new file is always on the same condition, however (is
526 * the level of this section higher than the leaf_level configuration
529 * Treating the cases backwards:
531 * (3) same file if sect->level > conf.leaf_level, otherwise new file
533 * if in the same file, currentsect->child points to sect
534 * otherwise the linking is done through the file tree (which works
535 * in more or less the same way, ie currentfile->child points to
538 * (2) same file if sect->level > conf.leaf_level, otherwise new file
540 * if in the same file, currentsect->next points to sect
541 * otherwise file linking and currentfile->next points to the new
542 * file (we know that Z must have caused a new file to be created)
544 * (1) same file if sect->level > conf.leaf_level, otherwise new file
546 * this is actually effectively the same case as (2) here,
547 * except that we first have to travel up the sections to figure
548 * out which section this new one will be a sibling of. In doing
549 * so, we may disappear off the top of a file and have to go up
550 * to its parent in the file tree.
553 static void xhtml_ponder_layout(paragraph
*p
)
555 xhtmlsection
*lastsection
;
556 xhtmlsection
*currentsect
;
557 xhtmlfile
*currentfile
;
560 topsection
= xhtml_new_section(NULL
);
561 topfile
= xhtml_new_file(NULL
);
562 lastsection
= topsection
;
563 currentfile
= topfile
;
564 currentsect
= topsection
;
566 if (conf
.leaf_level
== 0) {
567 topfile
->is_leaf
= 1;
568 topfile
->sections
= topsection
;
569 topsection
->file
= topfile
;
574 int level
= xhtml_para_level(p
);
575 if (level
>0) /* actually a section */
580 rdstringc fname_c
= { 0, 0, NULL
};
582 sect
= xhtml_new_section(lastsection
);
585 for (w
=(p
->kwtext2
)?
(p
->kwtext2
):(p
->words
); w
; w
=w
->next
) /* kwtext2 because we want numbers only! */
587 switch (removeattr(w
->type
))
592 case word_WeakCode:*/
593 xhtml_utostr(w
->text
, &c
);
599 /* rdaddsc(&fname_c, ".html");*/
600 sect
->fragment
= rdtrimc(&fname_c
);
602 /* printf(" ! adding para @ %p as sect %s, level %i\n", sect->para, sect->fragment, level);*/
604 if (level
>currentsect
->level
) { /* case (3) */
605 if (level
>conf
.leaf_level
) { /* same file */
606 assert(currentfile
->is_leaf
);
607 currentsect
->child
= sect
;
608 sect
->parent
=currentsect
;
609 sect
->file
=currentfile
;
610 /* printf("connected '%s' to existing file '%s' [I]\n", sect->fragment, currentfile->filename);*/
612 } else { /* new file */
613 xhtmlfile
*file
= xhtml_new_file(sect
);
614 assert(!currentfile
->is_leaf
);
615 currentfile
->child
=file
;
617 file
->parent
=currentfile
;
618 /* printf("connected '%s' to new file '%s' [I]\n", sect->fragment, file->filename);*/
622 } else if (level
>= currentsect
->file
->sections
->level
) {
623 /* Case (1) or (2) *AND* still under the section that starts
626 * I'm not convinced that this couldn't be rolled in with the
627 * final else {} leg further down. It seems a lot of effort
630 if (level
>conf
.leaf_level
) { /* stick within the same file */
631 assert(currentfile
->is_leaf
);
632 sect
->file
= currentfile
;
633 while (currentsect
&& currentsect
->level
> level
&&
634 currentsect
->file
==currentsect
->parent
->file
) {
635 currentsect
= currentsect
->parent
;
638 currentsect
->next
= sect
;
639 assert(currentsect
->level
== sect
->level
);
640 sect
->parent
= currentsect
->parent
;
642 /* printf("connected '%s' to existing file '%s' [II]\n", sect->fragment, currentfile->filename);*/
643 } else { /* new file */
644 xhtmlfile
*file
= xhtml_new_file(sect
);
646 currentfile
->next
=file
;
647 file
->parent
=currentfile
->parent
;
648 file
->is_leaf
=(level
==conf
.leaf_level
);
650 /* printf("connected '%s' to new file '%s' [II]\n", sect->fragment, file->filename);*/
654 } else { /* Case (1) or (2) and we must move up the file tree first */
655 /* this loop is now probably irrelevant - we know we can't connect
656 * to anything in the current file */
657 while (currentsect
&& level
<currentsect
->level
) {
658 currentsect
=currentsect
->parent
;
660 /* printf(" * up one level to '%s'\n", currentsect->fragment);*/
662 /* printf(" * up one level (off top of current file)\n");*/
666 /* I'm pretty sure this can now never fire */
667 assert(currentfile
->is_leaf
);
668 /* printf("connected '%s' to existing file '%s' [III]\n", sect->fragment, currentfile->filename);*/
669 sect
->file
= currentfile
;
670 currentsect
->next
=sect
;
672 } else { /* find a file we can attach to */
673 while (currentfile
&& currentfile
->sections
&& level
<currentfile
->sections
->level
) {
674 currentfile
=currentfile
->parent
;
676 /* printf(" * up one file level to '%s'\n", currentfile->filename);*/
678 /* printf(" * up one file level (off top of tree)\n");*/
681 if (currentfile
) { /* new file (we had to skip up a file to
682 get here, so we must be dealing with a
683 level no lower than the configured
685 xhtmlfile
*file
= xhtml_new_file(sect
);
686 currentfile
->next
=file
;
688 file
->parent
=currentfile
->parent
;
689 file
->is_leaf
=(level
==conf
.leaf_level
);
691 /* printf("connected '%s' to new file '%s' [III]\n", sect->fragment, file->filename);*/
695 fatal(err_whatever
, "Ran off the top trying to connect sibling: strange document.");
701 topsection
= lastsection
; /* get correct end of the chain */
702 xhtml_fixup_layout(topfile
); /* leaf files not at leaf level marked as such */
705 static void xhtml_do_index();
706 static void xhtml_do_file(xhtmlfile
*file
);
707 static void xhtml_do_top_file(xhtmlfile
*file
, paragraph
*sourceform
);
708 static void xhtml_do_paras(FILE *fp
, paragraph
*p
, paragraph
*end
, int indexable
);
709 static int xhtml_do_contents_limit(FILE *fp
, xhtmlfile
*file
, int limit
);
710 static int xhtml_do_contents_section_limit(FILE *fp
, xhtmlsection
*section
, int limit
);
711 static int xhtml_add_contents_entry(FILE *fp
, xhtmlsection
*section
, int limit
);
712 static int xhtml_do_contents(FILE *fp
, xhtmlfile
*file
);
713 static int xhtml_do_naked_contents(FILE *fp
, xhtmlfile
*file
);
714 static void xhtml_do_sections(FILE *fp
, xhtmlsection
*sections
);
717 * Do all the files in this structure.
719 static void xhtml_do_files(xhtmlfile
*file
)
723 xhtml_do_files(file
->child
);
725 xhtml_do_files(file
->next
);
729 * Free up all memory used by the file tree from 'xfile' downwards
731 static void xhtml_free_file(xhtmlfile
* xfile
)
737 if (xfile
->filename
) {
738 sfree(xfile
->filename
);
740 xhtml_free_file(xfile
->child
);
741 xhtml_free_file(xfile
->next
);
748 void xhtml_backend(paragraph
*sourceform
, keywordlist
*in_keywords
,
756 sourceparas
= sourceform
;
757 conf
= xhtml_configure(sourceform
);
758 keywords
= in_keywords
;
761 /* Clear up the index entries backend data pointers */
762 for (ti
=0; (ientry
= (indexentry
*)index234(idx
->entries
, ti
))!=NULL
; ti
++) {
763 ientry
->backend_data
=NULL
;
766 xhtml_ponder_layout(sourceform
);
768 /* old system ... (writes to *.alt, but gets some stuff wrong and is ugly) */
769 /* xhtml_level_0(sourceform);
770 for (i=1; i<=conf.leaf_level; i++)
772 xhtml_level(sourceform, i);
775 /* new system ... (writes to *.html, but isn't fully trusted) */
776 xhtml_do_top_file(topfile
, sourceform
);
777 assert(!topfile
->next
); /* shouldn't have a sibling at all */
778 if (topfile
->child
) {
779 xhtml_do_files(topfile
->child
);
783 /* release file, section, index data structures */
786 xhtmlsection
*tmp
= xsect
->chain
;
787 if (xsect
->fragment
) {
788 sfree(xsect
->fragment
);
793 xhtml_free_file(topfile
);
794 for (ti
= 0; (ientry
=(indexentry
*)index234(idx
->entries
, ti
))!=NULL
; ti
++) {
795 if (ientry
->backend_data
!=NULL
) {
796 xhtmlindex
*xi
= (xhtmlindex
*) ientry
->backend_data
;
797 if (xi
->sections
!=NULL
) {
802 ientry
->backend_data
= NULL
;
807 static int xhtml_para_level(paragraph
*p
)
814 case para_UnnumberedChapter
:
819 /* case para_BiblioCited:
832 /* Output the nav links for the current file.
833 * file == NULL means we're doing the index
835 static void xhtml_donavlinks(FILE *fp
, xhtmlfile
*file
)
837 xhtmlfile
*xhtml_next_file
= NULL
;
839 if (conf
.nav_attrs
!=NULL
) {
840 fprintf(fp
, " %ls>", conf
.nav_attrs
);
844 if (xhtml_last_file
==NULL
) {
845 fprintf(fp
, "Previous | ");
847 fprintf(fp
, "<a href='%s'>Previous</a> | ", xhtml_last_file
->filename
);
849 fprintf(fp
, "<a href='%s'>Contents</a> | ", conf
.contents_filename
);
851 fprintf(fp
, "Index | ");
853 fprintf(fp
, "<a href='%s'>Index</a> | ", conf
.index_filename
);
855 if (file
!= NULL
) { /* otherwise we're doing nav links for the index */
856 if (xhtml_next_file
==NULL
)
857 xhtml_next_file
= file
->child
;
858 if (xhtml_next_file
==NULL
)
859 xhtml_next_file
= file
->next
;
860 if (xhtml_next_file
==NULL
)
861 xhtml_next_file
= file
->parent
->next
;
863 if (xhtml_next_file
==NULL
) {
864 if (file
==NULL
) { /* index, so no next file */
865 fprintf(fp
, "Next ");
867 fprintf(fp
, "<a href='%s'>Next</a>", conf
.index_filename
);
870 fprintf(fp
, "<a href='%s'>Next</a>", xhtml_next_file
->filename
);
872 fprintf(fp
, "</p>\n");
875 /* Write out the index file */
876 static void xhtml_do_index_body(FILE *fp
)
881 if (count234(idx
->entries
) == 0)
882 return; /* don't write anything at all */
884 fprintf(fp
, "<dl>\n");
885 /* iterate over idx->entries using the tree functions and display everything */
886 for (ti
= 0; (y
= (indexentry
*)index234(idx
->entries
, ti
)) != NULL
; ti
++) {
887 if (y
->backend_data
) {
892 xhtml_para(fp
, y
->text
, FALSE
);
893 fprintf(fp
, "</dt>\n<dd>");
895 xi
= (xhtmlindex
*) y
->backend_data
;
896 for (i
=0; i
<xi
->nsection
; i
++) {
897 xhtmlsection
*sect
= xi
->sections
[i
];
899 fprintf(fp
, "<a href='%s#%s'>", sect
->file
->filename
, sect
->fragment
);
900 if (sect
->para
->kwtext
) {
901 xhtml_para(fp
, sect
->para
->kwtext
, FALSE
);
902 } else if (sect
->para
->words
) {
903 xhtml_para(fp
, sect
->para
->words
, FALSE
);
906 if (i
+1<xi
->nsection
) {
911 fprintf(fp
, "</dd>\n");
914 fprintf(fp
, "</dl>\n");
916 static void xhtml_do_index()
918 word temp_word
= { NULL
, NULL
, word_Normal
, 0, 0, L
"Index", { NULL
, 0, 0} };
919 FILE *fp
= fopen(conf
.index_filename
, "w");
922 fatal(err_cantopenw
, conf
.index_filename
);
923 xhtml_doheader(fp
, &temp_word
);
924 xhtml_donavlinks(fp
, NULL
);
926 xhtml_do_index_body(fp
);
928 xhtml_donavlinks(fp
, NULL
);
933 /* Output the given file. This includes whatever contents at beginning and end, etc. etc. */
934 static void xhtml_do_file(xhtmlfile
*file
)
936 FILE *fp
= fopen(file
->filename
, "w");
938 fatal(err_cantopenw
, file
->filename
);
940 if (file
->sections
->para
->words
) {
941 xhtml_doheader(fp
, file
->sections
->para
->words
);
942 } else if (file
->sections
->para
->kwtext
) {
943 xhtml_doheader(fp
, file
->sections
->para
->kwtext
);
945 xhtml_doheader(fp
, NULL
);
948 xhtml_donavlinks(fp
, file
);
950 if (file
->is_leaf
&& conf
.leaf_contains_contents
&&
951 xhtml_do_contents(NULL
, file
)>=conf
.leaf_smallest_contents
)
952 xhtml_do_contents(fp
, file
);
953 xhtml_do_sections(fp
, file
->sections
);
955 xhtml_do_naked_contents(fp
, file
);
957 xhtml_donavlinks(fp
, file
);
962 xhtml_last_file
= file
;
965 /* Output the top-level file. */
966 static void xhtml_do_top_file(xhtmlfile
*file
, paragraph
*sourceform
)
970 FILE *fp
= fopen(file
->filename
, "w");
972 fatal(err_cantopenw
, file
->filename
);
974 /* Do the title -- only one allowed */
975 for (p
= sourceform
; p
&& !done
; p
= p
->next
)
977 if (p
->type
== para_Title
)
979 xhtml_doheader(fp
, p
->words
);
984 xhtml_doheader(fp
, NULL
/* Eek! */);
989 for (p
= sourceform
; p
; p
= p
->next
)
991 if (p
->type
== para_Title
) {
992 xhtml_heading(fp
, p
, FALSE
);
997 /* Do the preamble */
998 for (p
= sourceform
; p
; p
= p
->next
)
1000 if (p
->type
== para_Chapter
|| p
->type
== para_Heading
||
1001 p
->type
== para_Subsect
|| p
->type
== para_Appendix
||
1002 p
->type
== para_UnnumberedChapter
) {
1004 * We've found the end of the preamble. Do every normal
1005 * paragraph up to there.
1007 xhtml_do_paras(fp
, sourceform
, p
, FALSE
);
1012 xhtml_do_contents(fp
, file
);
1013 xhtml_do_sections(fp
, file
->sections
);
1016 * Put the index in the top file if we're in single-file mode
1019 if (conf
.leaf_level
== 0 && count234(idx
->entries
) > 0) {
1020 fprintf(fp
, "<a name=\"index\"></a><h1>Index</h1>\n");
1021 xhtml_do_index_body(fp
);
1028 /* Convert a Unicode string to an ASCII one. '?' is
1029 * used for unmappable characters.
1031 static void xhtml_utostr(wchar_t *in
, char **out
)
1033 int l
= ustrlen(in
);
1035 *out
= smalloc(l
+1);
1038 if (in
[i
]>=32 && in
[i
]<=126)
1039 (*out
)[i
]=(char)in
[i
];
1047 * Write contents for the given file, and subfiles, down to
1048 * the appropriate contents depth. Returns the number of
1051 static int xhtml_do_contents(FILE *fp
, xhtmlfile
*file
)
1053 int level
, limit
, count
= 0;
1057 level
= (file
->sections
)?
(file
->sections
->level
):(0);
1058 limit
= conf
.contents_depth
[(level
>5)?
(5):(level
)];
1059 start_level
= (file
->is_leaf
) ?
(level
-1) : (level
);
1060 last_level
= start_level
;
1062 count
+= xhtml_do_contents_section_limit(fp
, file
->sections
, limit
);
1063 count
+= xhtml_do_contents_limit(fp
, file
->child
, limit
);
1065 while (last_level
> start_level
) {
1067 fprintf(fp
, "</li></ul>\n");
1073 /* As above, but doesn't do anything in the current file */
1074 static int xhtml_do_naked_contents(FILE *fp
, xhtmlfile
*file
)
1076 int level
, limit
, start_level
, count
= 0;
1080 level
= (file
->sections
)?
(file
->sections
->level
):(0);
1081 limit
= conf
.contents_depth
[(level
>5)?
(5):(level
)];
1082 start_level
= (file
->is_leaf
) ?
(level
-1) : (level
);
1083 last_level
= start_level
;
1085 count
= xhtml_do_contents_limit(fp
, file
->child
, limit
);
1087 while (last_level
> start_level
) {
1089 fprintf(fp
, "</li></ul>\n");
1096 * Write contents for the given file, children, and siblings, down to
1097 * given limit contents depth.
1099 static int xhtml_do_contents_limit(FILE *fp
, xhtmlfile
*file
, int limit
)
1103 count
+= xhtml_do_contents_section_limit(fp
, file
->sections
, limit
);
1104 count
+= xhtml_do_contents_limit(fp
, file
->child
, limit
);
1111 * Write contents entries for the given section tree, down to the
1112 * limit contents depth.
1114 static int xhtml_do_contents_section_deep_limit(FILE *fp
, xhtmlsection
*section
, int limit
)
1118 if (!xhtml_add_contents_entry(fp
, section
, limit
))
1122 count
+= xhtml_do_contents_section_deep_limit(fp
, section
->child
, limit
);
1123 section
= section
->next
;
1129 * Write contents entries for the given section tree, down to the
1130 * limit contents depth.
1132 static int xhtml_do_contents_section_limit(FILE *fp
, xhtmlsection
*section
, int limit
)
1137 xhtml_add_contents_entry(fp
, section
, limit
);
1139 count
+= xhtml_do_contents_section_deep_limit(fp
, section
->child
, limit
);
1140 /* section=section->child;
1141 while (section && xhtml_add_contents_entry(fp, section, limit)) {
1142 section = section->next;
1148 * Add a section entry, unless we're exceeding the limit, in which
1149 * case return FALSE (otherwise return TRUE).
1151 static int xhtml_add_contents_entry(FILE *fp
, xhtmlsection
*section
, int limit
)
1153 if (!section
|| section
->level
> limit
)
1155 if (fp
==NULL
|| section
->level
< 0)
1157 if (last_level
> section
->level
) {
1158 while (last_level
> section
->level
) {
1160 fprintf(fp
, "</li></ul>\n");
1162 fprintf(fp
, "</li>\n");
1163 } else if (last_level
< section
->level
) {
1164 assert(last_level
== section
->level
- 1);
1166 fprintf(fp
, "<ul>\n");
1168 fprintf(fp
, "</li>\n");
1170 fprintf(fp
, "<li><a href=\"%s#%s\">", section
->file
->filename
, section
->fragment
);
1171 if (section
->para
->kwtext
) {
1172 xhtml_para(fp
, section
->para
->kwtext
, FALSE
);
1173 if (section
->para
->words
) {
1177 if (section
->para
->words
) {
1178 xhtml_para(fp
, section
->para
->words
, FALSE
);
1180 fprintf(fp
, "</a>\n");
1185 * Write all the sections in this file. Do all paragraphs in this section, then all
1186 * children (recursively), then go on to the next one (tail recursively).
1188 static void xhtml_do_sections(FILE *fp
, xhtmlsection
*sections
)
1191 currentsection
= sections
;
1192 xhtml_do_paras(fp
, sections
->para
, NULL
, TRUE
);
1193 xhtml_do_sections(fp
, sections
->child
);
1194 sections
= sections
->next
;
1198 /* Write this list of paragraphs. Close off all lists at the end. */
1199 static void xhtml_do_paras(FILE *fp
, paragraph
*p
, paragraph
*end
,
1202 int last_type
= -1, ptype
, first
=TRUE
;
1203 stack lcont_stack
= stk_new();
1207 /* for (; p && (xhtml_para_level(p)>limit || xhtml_para_level(p)==-1 || first); p=p->next) {*/
1208 for (; p
&& p
!= end
&& (xhtml_para_level(p
)==-1 || first
); p
=p
->next
) {
1210 switch (ptype
= p
->type
)
1213 * Things we ignore because we've already processed them or
1214 * aren't going to touch them in this pass.
1218 case para_Biblio
: /* only touch BiblioCited */
1219 case para_VersionID
:
1229 case para_UnnumberedChapter
:
1230 xhtml_heading(fp
, p
, indexable
);
1235 xhtml_heading(fp
, p
, indexable
);
1239 fprintf(fp
, "\n<hr />\n");
1243 case para_Copyright
:
1244 fprintf(fp
, "\n<p>");
1245 xhtml_para(fp
, p
->words
, indexable
);
1246 fprintf(fp
, "</p>\n");
1249 case para_LcontPush
:
1254 stk_push(lcont_stack
, p
);
1255 last_type
= para_Normal
;
1260 int *p
= stk_pop(lcont_stack
);
1262 ptype
= last_type
= *p
;
1264 goto closeofflist
; /* ick */
1267 case para_QuotePush
:
1268 fprintf(fp
, "<blockquote>\n");
1271 fprintf(fp
, "</blockquote>\n");
1275 case para_NumberedList
:
1276 case para_Description
:
1277 case para_DescribedThing
:
1278 case para_BiblioCited
:
1279 if (last_type
!=p
->type
&&
1280 !(last_type
==para_DescribedThing
&& p
->type
==para_Description
) &&
1281 !(last_type
==para_Description
&& p
->type
==para_DescribedThing
)) {
1282 /* start up list if necessary */
1283 if (p
->type
== para_Bullet
) {
1284 fprintf(fp
, "<ul>\n");
1285 } else if (p
->type
== para_NumberedList
) {
1286 fprintf(fp
, "<ol>\n");
1287 } else if (p
->type
== para_BiblioCited
||
1288 p
->type
== para_DescribedThing
||
1289 p
->type
== para_Description
) {
1290 fprintf(fp
, "<dl>\n");
1293 if (p
->type
== para_Bullet
|| p
->type
== para_NumberedList
) {
1294 fprintf(fp
, "<li>");
1295 } else if (p
->type
== para_DescribedThing
) {
1296 fprintf(fp
, "<dt>");
1297 } else if (p
->type
== para_Description
) {
1298 fprintf(fp
, "<dd>");
1299 } else if (p
->type
== para_BiblioCited
) {
1300 fprintf(fp
, "<dt>");
1301 xhtml_para(fp
, p
->kwtext
, indexable
);
1302 fprintf(fp
, "</dt>\n<dd>");
1304 xhtml_para(fp
, p
->words
, indexable
);
1306 paragraph
*p2
= p
->next
;
1307 if (p2
&& xhtml_para_level(p2
)==-1 && p2
->type
== para_LcontPush
)
1312 if (ptype
== para_BiblioCited
) {
1313 fprintf(fp
, "</dd>\n");
1314 } else if (ptype
== para_DescribedThing
) {
1315 fprintf(fp
, "</dt>");
1316 } else if (ptype
== para_Description
) {
1317 fprintf(fp
, "</dd>");
1318 } else if (ptype
== para_Bullet
|| ptype
== para_NumberedList
) {
1319 fprintf(fp
, "</li>");
1321 if (ptype
== para_Bullet
|| ptype
== para_NumberedList
||
1322 ptype
== para_BiblioCited
|| ptype
== para_Description
||
1323 ptype
== para_DescribedThing
)
1324 /* close off list if necessary */
1326 paragraph
*p2
= p
->next
;
1327 int close_off
=FALSE
;
1328 /* if (p2 && (xhtml_para_level(p2)>limit || xhtml_para_level(p2)==-1)) {*/
1329 if (p2
&& xhtml_para_level(p2
)==-1) {
1330 if (p2
->type
!= ptype
&&
1331 !(p2
->type
==para_DescribedThing
&& ptype
==para_Description
) &&
1332 !(p2
->type
==para_Description
&& ptype
==para_DescribedThing
) &&
1333 p2
->type
!= para_LcontPush
)
1339 if (ptype
== para_Bullet
) {
1340 fprintf(fp
, "</ul>\n");
1341 } else if (ptype
== para_NumberedList
) {
1342 fprintf(fp
, "</ol>\n");
1343 } else if (ptype
== para_BiblioCited
||
1344 ptype
== para_Description
||
1345 ptype
== para_DescribedThing
) {
1346 fprintf(fp
, "</dl>\n");
1353 xhtml_codepara(fp
, p
->words
);
1359 stk_free(lcont_stack
);
1363 * Output a header for this XHTML file.
1365 static void xhtml_doheader(FILE *fp
, word
*title
)
1367 fprintf(fp
, "<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Transitional//EN\"\n");
1368 fprintf(fp
, "\"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd\">\n");
1369 fprintf(fp
, "<html xmlns='http://www.w3.org/1999/xhtml'>\n\n<head>\n<title>");
1371 fprintf(fp
, "The thing with no name!");
1373 xhtml_para(fp
, title
, FALSE
);
1374 fprintf(fp
, "</title>\n");
1375 fprintf(fp
, "<meta name=\"generator\" content=\"Halibut %s xhtml-backend\" />\n", version
);
1377 fprintf(fp
, "<meta name=\"author\" content=\"%ls\" />\n", conf
.author
);
1378 if (conf
.description
)
1379 fprintf(fp
, "<meta name=\"description\" content=\"%ls\" />\n", conf
.description
);
1381 fprintf(fp
, "%ls\n", conf
.head_end
);
1382 fprintf(fp
, "</head>\n\n");
1384 fprintf(fp
, "%ls\n", conf
.body
);
1386 fprintf(fp
, "<body>\n");
1387 if (conf
.body_start
)
1388 fprintf(fp
, "%ls\n", conf
.body_start
);
1392 * Output a footer for this XHTML file.
1394 static void xhtml_dofooter(FILE *fp
)
1396 fprintf(fp
, "\n<hr />\n\n");
1398 fprintf(fp
, "%ls\n", conf
.body_end
);
1399 if (!conf
.suppress_address
) {
1400 fprintf(fp
,"<address>\n");
1401 if (conf
.address_start
)
1402 fprintf(fp
, "%ls\n", conf
.address_start
);
1403 /* Do the version ID */
1404 if (conf
.include_version_id
) {
1407 for (p
= sourceparas
; p
; p
= p
->next
)
1408 if (p
->type
== para_VersionID
) {
1409 xhtml_versionid(fp
, p
->words
, started
);
1413 if (conf
.address_end
)
1414 fprintf(fp
, "%ls\n", conf
.address_end
);
1415 fprintf(fp
, "</address>\n");
1417 fprintf(fp
, "</body>\n\n</html>\n");
1421 * Output the versionid paragraph. Typically this is a version control
1422 * ID string (such as $Id...$ in RCS).
1424 static void xhtml_versionid(FILE *fp
, word
*text
, int started
)
1426 rdstringc t
= { 0, 0, NULL
};
1428 rdaddc(&t
, '['); /* FIXME: configurability */
1429 xhtml_rdaddwc(&t
, text
, NULL
, FALSE
);
1430 rdaddc(&t
, ']'); /* FIXME: configurability */
1433 fprintf(fp
, "<br />\n");
1434 fprintf(fp
, "%s\n", t
.text
);
1438 /* Is this an XHTML reserved character? */
1439 static int xhtml_reservedchar(int c
)
1441 if (c
=='&' || c
=='<' || c
=='>' || c
=='"')
1448 * Convert a wide string into valid XHTML: Anything outside ASCII will
1449 * be fixed up as an entity. Currently we don't worry about constraining the
1450 * encoded character set, which we should probably do at some point (we can
1451 * still fix up and return FALSE - see the last comment here). We also don't
1454 * Because this is only used for words, spaces are HARD spaces (any other
1455 * spaces will be word_Whitespace not word_Normal). So they become
1456 * Unless hard_spaces is FALSE, of course (code paragraphs break the above
1459 * If `result' is non-NULL, mallocs the resulting string and stores a pointer to
1460 * it in `*result'. If `result' is NULL, merely checks whether all
1461 * characters in the string are feasible.
1463 * Return is nonzero if all characters are OK. If not all
1464 * characters are OK but `result' is non-NULL, a result _will_
1465 * still be generated!
1467 static int xhtml_convert(wchar_t *s
, int maxlen
, char **result
,
1469 int doing
= (result
!= 0);
1472 int plen
= 0, psize
= 0;
1477 for (; *s
&& maxlen
!= 0; s
++, maxlen
--) {
1480 #define ensure_size(i) if (i>=psize) { psize = i+256; p = resize(p, psize); }
1482 if (((c
== 32 && !hard_spaces
) || (c
> 32 && c
<= 126 && !xhtml_reservedchar(c
)))) {
1487 p
[plen
++] = (char)c
;
1490 /* Char needs fixing up. */
1491 /* ok = FALSE; -- currently we never return FALSE; we
1492 * might want to when considering a character set for the
1497 if (c
==32) { /* a space in a word is a hard space */
1498 ensure_size(plen
+6); /* includes space for the NUL, which is subsequently stomped on */
1499 sprintf(p
+plen
, " ");
1502 /* FIXME: entity names! */
1503 ensure_size(plen
+8); /* includes space for the NUL, which is subsequently stomped on */
1504 plen
+=sprintf(p
+plen
, "&#%04i;", (int)c
);
1510 p
= resize(p
, plen
+1);
1518 * This formats the given words as XHTML.
1520 * `indexable', if FALSE, prohibits adding any index references.
1521 * You might use this, for example, if an index reference occurred
1522 * in a section title, to prevent phony index references when the
1523 * section title is processed in strange places such as contents
1526 static void xhtml_rdaddwc(rdstringc
*rs
, word
*text
, word
*end
, int indexable
) {
1533 for (; text
&& text
!= end
; text
= text
->next
) {
1534 switch (text
->type
) {
1535 case word_HyperLink
:
1536 xhtml_utostr(text
->text
, &c
);
1537 rdaddsc(rs
, "<a href=\"");
1543 case word_UpperXref
:
1544 case word_LowerXref
:
1545 kwl
= kw_lookup(keywords
, text
->text
);
1547 sect
=xhtml_find_section(kwl
->para
);
1549 rdaddsc(rs
, "<a href=\"");
1550 rdaddsc(rs
, sect
->file
->filename
);
1552 rdaddsc(rs
, sect
->fragment
);
1555 rdaddsc(rs
, "<a href=\"Apologies.html\"><!-- probably a bibliography cross reference -->");
1556 error(err_whatever
, "Couldn't locate cross-reference! (Probably a bibliography entry.)");
1559 rdaddsc(rs
, "<a href=\"Apologies.html\"><!-- unknown cross-reference -->");
1560 error(err_whatever
, "Couldn't locate cross-reference! (Wasn't in source file.)");
1564 case word_IndexRef
: /* in theory we could make an index target here */
1565 /* rdaddsc(rs, "<a name=\"idx-");
1566 xhtml_utostr(text->text, &c);
1569 rdaddsc(rs, "\"></a>");*/
1570 /* what we _do_ need to do is to fix up the backend data
1571 * for any indexentry this points to.
1576 for (ti
=0; (itag
= (indextag
*)index234(idx
->tags
, ti
))!=NULL
; ti
++) {
1577 /* FIXME: really ustricmp() and not ustrcmp()? */
1578 if (ustricmp(itag
->name
, text
->text
)==0) {
1583 if (itag
->refs
!=NULL
) {
1585 for (i
=0; i
<itag
->nrefs
; i
++) {
1586 xhtmlindex
*idx_ref
;
1589 ientry
= itag
->refs
[i
];
1590 if (ientry
->backend_data
==NULL
) {
1591 idx_ref
= (xhtmlindex
*) smalloc(sizeof(xhtmlindex
));
1593 fatal(err_nomemory
);
1594 idx_ref
->nsection
= 0;
1596 idx_ref
->sections
= (xhtmlsection
**) smalloc(idx_ref
->size
* sizeof(xhtmlsection
*));
1597 if (idx_ref
->sections
==NULL
)
1598 fatal(err_nomemory
);
1599 ientry
->backend_data
= idx_ref
;
1601 idx_ref
= ientry
->backend_data
;
1602 if (idx_ref
->nsection
+1 > idx_ref
->size
) {
1603 int new_size
= idx_ref
->size
* 2;
1604 idx_ref
->sections
= srealloc(idx_ref
->sections
, new_size
* sizeof(xhtmlsection
));
1605 if (idx_ref
->sections
==NULL
) {
1606 fatal(err_nomemory
);
1608 idx_ref
->size
= new_size
;
1611 idx_ref
->sections
[idx_ref
->nsection
++] = currentsection
;
1616 fatal(err_whatever
, "Index tag had no entries!");
1619 fprintf(stderr
, "Looking for index entry '%ls'\n", text
->text
);
1620 fatal(err_whatever
, "Couldn't locate index entry! (Wasn't in index.)");
1626 rdaddsc(rs
, "</a>");
1633 case word_WhiteSpace
:
1634 case word_EmphSpace
:
1635 case word_CodeSpace
:
1636 case word_WkCodeSpace
:
1638 case word_EmphQuote
:
1639 case word_CodeQuote
:
1640 case word_WkCodeQuote
:
1641 assert(text
->type
!= word_CodeQuote
&&
1642 text
->type
!= word_WkCodeQuote
);
1643 if (towordstyle(text
->type
) == word_Emph
&&
1644 (attraux(text
->aux
) == attr_First
||
1645 attraux(text
->aux
) == attr_Only
))
1646 rdaddsc(rs
, "<em>");
1647 else if ((towordstyle(text
->type
) == word_Code
|| towordstyle(text
->type
) == word_WeakCode
) &&
1648 (attraux(text
->aux
) == attr_First
||
1649 attraux(text
->aux
) == attr_Only
))
1650 rdaddsc(rs
, "<code>");
1652 if (removeattr(text
->type
) == word_Normal
) {
1653 if (xhtml_convert(text
->text
, 0, &c
, TRUE
)) /* spaces in the word are hard */
1656 xhtml_rdaddwc(rs
, text
->alt
, NULL
, indexable
);
1658 } else if (removeattr(text
->type
) == word_WhiteSpace
) {
1660 } else if (removeattr(text
->type
) == word_Quote
) {
1661 rdaddsc(rs
, """);
1664 if (towordstyle(text
->type
) == word_Emph
&&
1665 (attraux(text
->aux
) == attr_Last
||
1666 attraux(text
->aux
) == attr_Only
))
1667 rdaddsc(rs
, "</em>");
1668 else if ((towordstyle(text
->type
) == word_Code
|| towordstyle(text
->type
) == word_WeakCode
) &&
1669 (attraux(text
->aux
) == attr_Last
||
1670 attraux(text
->aux
) == attr_Only
))
1671 rdaddsc(rs
, "</code>");
1677 /* Output a heading, formatted as XHTML.
1679 static void xhtml_heading(FILE *fp
, paragraph
*p
, int indexable
)
1681 rdstringc t
= { 0, 0, NULL
};
1682 word
*tprefix
= p
->kwtext
;
1683 word
*nprefix
= p
->kwtext2
;
1684 word
*text
= p
->words
;
1685 int level
= xhtml_para_level(p
);
1686 xhtmlsection
*sect
= xhtml_find_section(p
);
1690 fragment
= sect
->fragment
;
1692 if (p
->type
== para_Title
)
1695 fragment
= ""; /* FIXME: what else can we do? */
1696 error(err_whatever
, "Couldn't locate heading cross-reference!");
1700 if (p
->type
== para_Title
)
1702 else if (level
== 1)
1703 fmt
= &conf
.fchapter
;
1704 else if (level
-1 < conf
.nfsect
)
1705 fmt
= &conf
.fsect
[level
-1];
1707 fmt
= &conf
.fsect
[conf
.nfsect
-1];
1709 if (fmt
&& fmt
->just_numbers
&& nprefix
) {
1710 xhtml_rdaddwc(&t
, nprefix
, NULL
, indexable
);
1713 if (xhtml_convert(fmt
->number_suffix
, 0, &c
, FALSE
)) {
1718 } else if (fmt
&& !fmt
->just_numbers
&& tprefix
) {
1719 xhtml_rdaddwc(&t
, tprefix
, NULL
, indexable
);
1722 if (xhtml_convert(fmt
->number_suffix
, 0, &c
, FALSE
)) {
1728 xhtml_rdaddwc(&t
, text
, NULL
, indexable
);
1730 * If we're outputting in single-file mode, we need to lower
1731 * the level of each heading by one, because the overall
1732 * document title will be sitting right at the top as an <h1>
1733 * and so chapters and sections should start at <h2>.
1735 * Even if not, the document title will come back from
1736 * xhtml_para_level() as level zero, so we must increment that
1737 * no matter what leaf_level is set to.
1739 if (conf
.leaf_level
== 0 || level
== 0)
1741 fprintf(fp
, "<a name=\"%s\"></a><h%i>%s</h%i>\n", fragment
, level
, t
.text
, level
);
1745 /* Output a paragraph. Styles are handled by xhtml_rdaddwc().
1746 * This looks pretty simple; I may have missed something ...
1748 static void xhtml_para(FILE *fp
, word
*text
, int indexable
)
1750 rdstringc out
= { 0, 0, NULL
};
1751 xhtml_rdaddwc(&out
, text
, NULL
, indexable
);
1752 fprintf(fp
, "%s", out
.text
);
1756 /* Output a code paragraph. I'm treating this as preformatted, which
1757 * may not be entirely correct. See xhtml_para() for my worries about
1758 * this being overly-simple; however I think that most of the complexity
1759 * of the text backend came entirely out of word wrapping anyway.
1761 static void xhtml_codepara(FILE *fp
, word
*text
)
1763 fprintf(fp
, "<pre>");
1764 for (; text
; text
= text
->next
) if (text
->type
== word_WeakCode
) {
1769 * See if this WeakCode is followed by an Emph to indicate
1773 if (text
->next
&& text
->next
->type
== word_Emph
) {
1774 next
= text
= text
->next
;
1788 for (n
= 0; t
[n
] && e
[n
] && e
[n
] == ec
; n
++);
1789 xhtml_convert(t
, n
, &c
, FALSE
);
1790 fprintf(fp
, "%s%s%s",
1791 (ec
== 'i' ?
"<em>" : ec
== 'b' ?
"<b>" : ""),
1793 (ec
== 'i' ?
"</em>" : ec
== 'b' ?
"</b>" : ""));
1800 xhtml_convert(t
, 0, &c
, FALSE
);
1801 fprintf(fp
, "%s\n", c
);
1804 xhtml_convert(here
->text
, 0, &c
, FALSE
);
1805 fprintf(fp
, "%s\n", c
);
1809 fprintf(fp
, "</pre>\n");