5dd44dce |
1 | /* |
2 | * info backend for Halibut |
3 | * |
b921687e |
4 | * Possible future work: |
5dd44dce |
5 | * |
b921687e |
6 | * - configurable indentation, bullets, emphasis, quotes etc? |
5dd44dce |
7 | * |
b921687e |
8 | * - configurable choice of how to allocate node names? |
9 | * + possibly a template-like approach, choosing node names to |
10 | * be the full section title or perhaps the internal keyword? |
11 | * + neither of those seems quite right. Perhaps instead a |
12 | * Windows Help-like mechanism, where a magic config |
13 | * directive allows user choice of name for every node. |
14 | * + Only trouble with that is, now what happens to the section |
15 | * numbers? Do they become completely vestigial and just sit |
16 | * in the title text of each node? Or do we keep them in the |
17 | * menus somehow? I think people might occasionally want to |
18 | * go to a section by number, if only because all the _other_ |
19 | * formats of the same document will reference the numbers |
20 | * all the time. So our menu lines could look like one of |
21 | * these: |
22 | * * Nodename: Section 1.2. Title of section. |
23 | * * Section 1.2: Nodename. Title of section. |
5dd44dce |
24 | * |
b921687e |
25 | * - might be helpful to diagnose duplicate node names! |
5dd44dce |
26 | */ |
27 | |
91f93b94 |
28 | /* |
29 | * FIXME: |
30 | * |
31 | * - alignment in the index is broken when a non-representable |
32 | * character appears with no alternative. More generally, I |
33 | * fear, this is the fault of the info_rdadd* functions failing |
34 | * to return correct width figures in this circumstance (so it |
35 | * will affect list paragraph prefixes and paragraph wrapping as |
36 | * well). |
37 | */ |
38 | |
5dd44dce |
39 | #include <stdio.h> |
40 | #include <stdlib.h> |
41 | #include <assert.h> |
42 | #include "halibut.h" |
43 | |
44 | typedef struct { |
45 | char *filename; |
46 | int maxfilesize; |
91f93b94 |
47 | int charset; |
5dd44dce |
48 | } infoconfig; |
49 | |
91f93b94 |
50 | typedef struct { |
51 | rdstringc output; |
52 | int charset; |
53 | charset_state state; |
54 | int wcmode; |
55 | } info_data; |
56 | #define EMPTY_INFO_DATA { { 0, 0, NULL }, 0, CHARSET_INIT_STATE, FALSE } |
57 | static const info_data empty_info_data = EMPTY_INFO_DATA; |
58 | |
5dd44dce |
59 | typedef struct node_tag node; |
60 | struct node_tag { |
61 | node *listnext; |
62 | node *up, *prev, *next, *lastchild; |
63 | int pos, started_menu, filenum; |
64 | char *name; |
91f93b94 |
65 | info_data text; |
5dd44dce |
66 | }; |
67 | |
68 | typedef struct { |
69 | char *text; |
91f93b94 |
70 | int length; |
5dd44dce |
71 | int nnodes, nodesize; |
72 | node **nodes; |
73 | } info_idx; |
74 | |
91f93b94 |
75 | static int info_rdadd(info_data *, wchar_t); |
76 | static int info_rdadds(info_data *, wchar_t const *); |
77 | static int info_rdaddc(info_data *, char); |
78 | static int info_rdaddsc(info_data *, char const *); |
5dd44dce |
79 | |
91f93b94 |
80 | static void info_heading(info_data *, word *, word *, int); |
81 | static void info_rule(info_data *, int, int); |
82 | static void info_para(info_data *, word *, wchar_t *, word *, keywordlist *, |
5dd44dce |
83 | int, int, int); |
91f93b94 |
84 | static void info_codepara(info_data *, word *, int, int); |
85 | static void info_versionid(info_data *, word *); |
86 | static void info_menu_item(info_data *, node *, paragraph *); |
5dd44dce |
87 | static word *info_transform_wordlist(word *, keywordlist *); |
88 | static int info_check_index(word *, node *, indexdata *); |
89 | |
91f93b94 |
90 | static int info_rdaddwc(info_data *, word *, word *, int); |
5dd44dce |
91 | |
91f93b94 |
92 | static node *info_node_new(char *name, int charset); |
93 | static char *info_node_name(paragraph *p, int charset); |
5dd44dce |
94 | |
95 | static infoconfig info_configure(paragraph *source) { |
96 | infoconfig ret; |
97 | |
98 | /* |
99 | * Defaults. |
100 | */ |
101 | ret.filename = dupstr("output.info"); |
102 | ret.maxfilesize = 64 << 10; |
91f93b94 |
103 | ret.charset = CS_ASCII; |
5dd44dce |
104 | |
105 | for (; source; source = source->next) { |
106 | if (source->type == para_Config) { |
107 | if (!ustricmp(source->keyword, L"info-filename")) { |
108 | sfree(ret.filename); |
e4ea58f8 |
109 | ret.filename = dupstr(adv(source->origkeyword)); |
91f93b94 |
110 | } else if (!ustricmp(source->keyword, L"info-charset")) { |
111 | char *csname = utoa_dup(uadv(source->keyword), CS_ASCII); |
112 | ret.charset = charset_from_localenc(csname); |
113 | sfree(csname); |
5dd44dce |
114 | } else if (!ustricmp(source->keyword, L"info-max-file-size")) { |
115 | ret.maxfilesize = utoi(uadv(source->keyword)); |
116 | } |
117 | } |
118 | } |
119 | |
120 | return ret; |
121 | } |
122 | |
123 | paragraph *info_config_filename(char *filename) |
124 | { |
e4ea58f8 |
125 | return cmdline_cfg_simple("info-filename", filename, NULL); |
5dd44dce |
126 | } |
127 | |
128 | void info_backend(paragraph *sourceform, keywordlist *keywords, |
43341922 |
129 | indexdata *idx, void *unused) { |
5dd44dce |
130 | paragraph *p; |
131 | infoconfig conf; |
132 | word *prefix, *body, *wp; |
133 | word spaceword; |
91f93b94 |
134 | wchar_t *prefixextra; |
5dd44dce |
135 | int nesting, nestindent; |
136 | int indentb, indenta; |
137 | int filepos; |
138 | int has_index; |
91f93b94 |
139 | info_data intro_text = EMPTY_INFO_DATA; |
5dd44dce |
140 | node *topnode, *currnode; |
141 | word bullet; |
142 | FILE *fp; |
143 | |
144 | /* |
91f93b94 |
145 | * FIXME: possibly configurability? |
5dd44dce |
146 | */ |
147 | int width = 70, listindentbefore = 1, listindentafter = 3; |
148 | int indent_code = 2, index_width = 40; |
149 | |
43341922 |
150 | IGNORE(unused); |
5dd44dce |
151 | |
152 | conf = info_configure(sourceform); |
153 | |
154 | /* |
155 | * Go through and create a node for each section. |
156 | */ |
91f93b94 |
157 | topnode = info_node_new("Top", conf.charset); |
5dd44dce |
158 | currnode = topnode; |
159 | for (p = sourceform; p; p = p->next) switch (p->type) { |
160 | /* |
161 | * Chapter titles. |
162 | */ |
163 | case para_Chapter: |
164 | case para_Appendix: |
165 | case para_UnnumberedChapter: |
166 | case para_Heading: |
167 | case para_Subsect: |
168 | { |
169 | node *newnode, *upnode; |
170 | char *nodename; |
171 | |
91f93b94 |
172 | nodename = info_node_name(p, conf.charset); |
173 | newnode = info_node_new(nodename, conf.charset); |
5dd44dce |
174 | sfree(nodename); |
175 | |
176 | p->private_data = newnode; |
177 | |
178 | if (p->parent) |
179 | upnode = (node *)p->parent->private_data; |
180 | else |
181 | upnode = topnode; |
182 | assert(upnode); |
183 | newnode->up = upnode; |
184 | |
185 | currnode->next = newnode; |
186 | newnode->prev = currnode; |
187 | |
188 | currnode->listnext = newnode; |
189 | currnode = newnode; |
190 | } |
191 | break; |
192 | } |
193 | |
194 | /* |
195 | * Set up the display form of each index entry. |
196 | */ |
197 | { |
198 | int i; |
199 | indexentry *entry; |
200 | |
201 | for (i = 0; (entry = index234(idx->entries, i)) != NULL; i++) { |
202 | info_idx *ii = mknew(info_idx); |
91f93b94 |
203 | info_data id = EMPTY_INFO_DATA; |
204 | |
205 | id.charset = conf.charset; |
5dd44dce |
206 | |
207 | ii->nnodes = ii->nodesize = 0; |
208 | ii->nodes = NULL; |
209 | |
91f93b94 |
210 | ii->length = info_rdaddwc(&id, entry->text, NULL, FALSE); |
f4551933 |
211 | |
91f93b94 |
212 | ii->text = id.output.text; |
5dd44dce |
213 | |
214 | entry->backend_data = ii; |
215 | } |
216 | } |
217 | |
218 | /* |
219 | * An Info file begins with a piece of introductory text which |
220 | * is apparently never shown anywhere. This seems to me to be a |
d4c7e130 |
221 | * good place to put the copyright notice and the version IDs. |
222 | * Also, Info directory entries are expected to go here. |
5dd44dce |
223 | */ |
91f93b94 |
224 | intro_text.charset = conf.charset; |
5dd44dce |
225 | |
91f93b94 |
226 | info_rdaddsc(&intro_text, |
5dd44dce |
227 | "This Info file generated by Halibut, "); |
91f93b94 |
228 | info_rdaddsc(&intro_text, version); |
229 | info_rdaddsc(&intro_text, "\n\n"); |
5dd44dce |
230 | |
231 | for (p = sourceform; p; p = p->next) |
d4c7e130 |
232 | if (p->type == para_Config && |
233 | !ustricmp(p->keyword, L"info-dir-entry")) { |
234 | wchar_t *section, *shortname, *longname, *kw; |
235 | char *s; |
236 | |
237 | section = uadv(p->keyword); |
238 | shortname = *section ? uadv(section) : NULL; |
239 | longname = *shortname ? uadv(shortname) : NULL; |
240 | kw = *longname ? uadv(longname) : NULL; |
241 | |
242 | if (!*longname) { |
243 | error(err_infodirentry, &p->fpos); |
244 | continue; |
245 | } |
246 | |
91f93b94 |
247 | info_rdaddsc(&intro_text, "INFO-DIR-SECTION "); |
248 | info_rdadds(&intro_text, section); |
249 | info_rdaddsc(&intro_text, "\nSTART-INFO-DIR-ENTRY\n* "); |
250 | info_rdadds(&intro_text, shortname); |
251 | info_rdaddsc(&intro_text, ": ("); |
d4c7e130 |
252 | s = dupstr(conf.filename); |
253 | if (strlen(s) > 5 && !strcmp(s+strlen(s)-5, ".info")) |
254 | s[strlen(s)-5] = '\0'; |
91f93b94 |
255 | info_rdaddsc(&intro_text, s); |
d4c7e130 |
256 | sfree(s); |
91f93b94 |
257 | info_rdaddsc(&intro_text, ")"); |
d4c7e130 |
258 | if (*kw) { |
259 | keyword *kwl = kw_lookup(keywords, kw); |
260 | if (kwl && kwl->para->private_data) { |
261 | node *n = (node *)kwl->para->private_data; |
91f93b94 |
262 | info_rdaddsc(&intro_text, n->name); |
d4c7e130 |
263 | } |
264 | } |
91f93b94 |
265 | info_rdaddsc(&intro_text, ". "); |
266 | info_rdadds(&intro_text, longname); |
267 | info_rdaddsc(&intro_text, "\nEND-INFO-DIR-ENTRY\n\n"); |
d4c7e130 |
268 | } |
269 | |
270 | for (p = sourceform; p; p = p->next) |
5dd44dce |
271 | if (p->type == para_Copyright) |
272 | info_para(&intro_text, NULL, NULL, p->words, keywords, |
273 | 0, 0, width); |
274 | |
275 | for (p = sourceform; p; p = p->next) |
276 | if (p->type == para_VersionID) |
277 | info_versionid(&intro_text, p->words); |
278 | |
91f93b94 |
279 | if (intro_text.output.text[intro_text.output.pos-1] != '\n') |
280 | info_rdaddc(&intro_text, '\n'); |
5dd44dce |
281 | |
282 | /* Do the title */ |
283 | for (p = sourceform; p; p = p->next) |
284 | if (p->type == para_Title) |
285 | info_heading(&topnode->text, NULL, p->words, width); |
286 | |
287 | nestindent = listindentbefore + listindentafter; |
288 | nesting = 0; |
289 | |
290 | currnode = topnode; |
291 | |
292 | /* Do the main document */ |
293 | for (p = sourceform; p; p = p->next) switch (p->type) { |
294 | |
295 | case para_QuotePush: |
296 | nesting += 2; |
297 | break; |
298 | case para_QuotePop: |
299 | nesting -= 2; |
300 | assert(nesting >= 0); |
301 | break; |
302 | |
303 | case para_LcontPush: |
304 | nesting += nestindent; |
305 | break; |
306 | case para_LcontPop: |
307 | nesting -= nestindent; |
308 | assert(nesting >= 0); |
309 | break; |
310 | |
311 | /* |
312 | * Things we ignore because we've already processed them or |
313 | * aren't going to touch them in this pass. |
314 | */ |
315 | case para_IM: |
316 | case para_BR: |
317 | case para_Biblio: /* only touch BiblioCited */ |
318 | case para_VersionID: |
319 | case para_NoCite: |
320 | case para_Title: |
321 | break; |
322 | |
323 | /* |
324 | * Chapter titles. |
325 | */ |
326 | case para_Chapter: |
327 | case para_Appendix: |
328 | case para_UnnumberedChapter: |
329 | case para_Heading: |
330 | case para_Subsect: |
331 | currnode = p->private_data; |
332 | assert(currnode); |
333 | assert(currnode->up); |
334 | |
335 | if (!currnode->up->started_menu) { |
91f93b94 |
336 | info_rdaddsc(&currnode->up->text, "* Menu:\n\n"); |
5dd44dce |
337 | currnode->up->started_menu = TRUE; |
338 | } |
339 | info_menu_item(&currnode->up->text, currnode, p); |
340 | |
341 | has_index |= info_check_index(p->words, currnode, idx); |
342 | info_heading(&currnode->text, p->kwtext, p->words, width); |
343 | nesting = 0; |
344 | break; |
345 | |
346 | case para_Rule: |
347 | info_rule(&currnode->text, nesting, width - nesting); |
348 | break; |
349 | |
350 | case para_Normal: |
351 | case para_Copyright: |
352 | case para_DescribedThing: |
353 | case para_Description: |
354 | case para_BiblioCited: |
355 | case para_Bullet: |
356 | case para_NumberedList: |
357 | has_index |= info_check_index(p->words, currnode, idx); |
358 | if (p->type == para_Bullet) { |
359 | bullet.next = NULL; |
360 | bullet.alt = NULL; |
361 | bullet.type = word_Normal; |
362 | bullet.text = L"-"; /* FIXME: configurability */ |
363 | prefix = • |
364 | prefixextra = NULL; |
365 | indentb = listindentbefore; |
366 | indenta = listindentafter; |
367 | } else if (p->type == para_NumberedList) { |
368 | prefix = p->kwtext; |
91f93b94 |
369 | prefixextra = L"."; /* FIXME: configurability */ |
5dd44dce |
370 | indentb = listindentbefore; |
371 | indenta = listindentafter; |
372 | } else if (p->type == para_Description) { |
373 | prefix = NULL; |
374 | prefixextra = NULL; |
375 | indentb = listindentbefore; |
376 | indenta = listindentafter; |
377 | } else { |
378 | prefix = NULL; |
379 | prefixextra = NULL; |
380 | indentb = indenta = 0; |
381 | } |
382 | if (p->type == para_BiblioCited) { |
383 | body = dup_word_list(p->kwtext); |
384 | for (wp = body; wp->next; wp = wp->next); |
385 | wp->next = &spaceword; |
386 | spaceword.next = p->words; |
387 | spaceword.alt = NULL; |
388 | spaceword.type = word_WhiteSpace; |
389 | spaceword.text = NULL; |
390 | } else { |
391 | wp = NULL; |
392 | body = p->words; |
393 | } |
394 | info_para(&currnode->text, prefix, prefixextra, body, keywords, |
395 | nesting + indentb, indenta, |
396 | width - nesting - indentb - indenta); |
397 | if (wp) { |
398 | wp->next = NULL; |
399 | free_word_list(body); |
400 | } |
401 | break; |
402 | |
403 | case para_Code: |
404 | info_codepara(&currnode->text, p->words, |
405 | nesting + indent_code, |
406 | width - nesting - 2 * indent_code); |
407 | break; |
408 | } |
409 | |
410 | /* |
411 | * Create an index node if required. |
412 | */ |
413 | if (has_index) { |
414 | node *newnode; |
415 | int i, j, k; |
416 | indexentry *entry; |
417 | |
91f93b94 |
418 | newnode = info_node_new("Index", conf.charset); |
5dd44dce |
419 | newnode->up = topnode; |
420 | |
421 | currnode->next = newnode; |
422 | newnode->prev = currnode; |
423 | currnode->listnext = newnode; |
424 | |
91f93b94 |
425 | info_rdaddsc(&newnode->text, "Index\n-----\n\n"); |
5dd44dce |
426 | |
427 | info_menu_item(&topnode->text, newnode, NULL); |
428 | |
429 | for (i = 0; (entry = index234(idx->entries, i)) != NULL; i++) { |
430 | info_idx *ii = (info_idx *)entry->backend_data; |
431 | |
432 | for (j = 0; j < ii->nnodes; j++) { |
5dd44dce |
433 | /* |
434 | * When we have multiple references for a single |
435 | * index term, we only display the actual term on |
436 | * the first line, to make it clear that the terms |
437 | * really are the same. |
438 | */ |
439 | if (j == 0) |
91f93b94 |
440 | info_rdaddsc(&newnode->text, ii->text); |
441 | for (k = (j ? 0 : ii->length); k < index_width; k++) |
442 | info_rdaddc(&newnode->text, ' '); |
443 | info_rdaddsc(&newnode->text, " *Note "); |
444 | info_rdaddsc(&newnode->text, ii->nodes[j]->name); |
445 | info_rdaddsc(&newnode->text, "::\n"); |
5dd44dce |
446 | } |
447 | } |
448 | } |
449 | |
450 | /* |
451 | * Finalise the text of each node, by adding the ^_ delimiter |
452 | * and the node line at the top. |
453 | */ |
454 | for (currnode = topnode; currnode; currnode = currnode->listnext) { |
91f93b94 |
455 | char *origtext = currnode->text.output.text; |
456 | currnode->text = empty_info_data; |
457 | currnode->text.charset = conf.charset; |
458 | info_rdaddsc(&currnode->text, "\037\nFile: "); |
459 | info_rdaddsc(&currnode->text, conf.filename); |
460 | info_rdaddsc(&currnode->text, ", Node: "); |
461 | info_rdaddsc(&currnode->text, currnode->name); |
5dd44dce |
462 | if (currnode->prev) { |
91f93b94 |
463 | info_rdaddsc(&currnode->text, ", Prev: "); |
464 | info_rdaddsc(&currnode->text, currnode->prev->name); |
5dd44dce |
465 | } |
91f93b94 |
466 | info_rdaddsc(&currnode->text, ", Up: "); |
467 | info_rdaddsc(&currnode->text, (currnode->up ? |
468 | currnode->up->name : "(dir)")); |
5dd44dce |
469 | if (currnode->next) { |
91f93b94 |
470 | info_rdaddsc(&currnode->text, ", Next: "); |
471 | info_rdaddsc(&currnode->text, currnode->next->name); |
5dd44dce |
472 | } |
91f93b94 |
473 | info_rdaddsc(&currnode->text, "\n\n"); |
474 | info_rdaddsc(&currnode->text, origtext); |
5dd44dce |
475 | /* |
476 | * Just make _absolutely_ sure we end with a newline. |
477 | */ |
91f93b94 |
478 | if (currnode->text.output.text[currnode->text.output.pos-1] != '\n') |
479 | info_rdaddc(&currnode->text, '\n'); |
5dd44dce |
480 | |
481 | sfree(origtext); |
482 | } |
483 | |
484 | /* |
485 | * Compute the offsets for the tag table. |
486 | */ |
91f93b94 |
487 | filepos = intro_text.output.pos; |
5dd44dce |
488 | for (currnode = topnode; currnode; currnode = currnode->listnext) { |
489 | currnode->pos = filepos; |
91f93b94 |
490 | filepos += currnode->text.output.pos; |
5dd44dce |
491 | } |
492 | |
493 | /* |
494 | * Split into sub-files. |
495 | */ |
496 | if (conf.maxfilesize > 0) { |
91f93b94 |
497 | int currfilesize = intro_text.output.pos, currfilenum = 1; |
5dd44dce |
498 | for (currnode = topnode; currnode; currnode = currnode->listnext) { |
91f93b94 |
499 | if (currfilesize > intro_text.output.pos && |
500 | currfilesize + currnode->text.output.pos > conf.maxfilesize) { |
5dd44dce |
501 | currfilenum++; |
91f93b94 |
502 | currfilesize = intro_text.output.pos; |
5dd44dce |
503 | } |
504 | currnode->filenum = currfilenum; |
91f93b94 |
505 | currfilesize += currnode->text.output.pos; |
5dd44dce |
506 | } |
507 | } |
508 | |
509 | /* |
510 | * Write the primary output file. |
511 | */ |
512 | fp = fopen(conf.filename, "w"); |
513 | if (!fp) { |
514 | error(err_cantopenw, conf.filename); |
515 | return; |
516 | } |
91f93b94 |
517 | fputs(intro_text.output.text, fp); |
5dd44dce |
518 | if (conf.maxfilesize == 0) { |
519 | for (currnode = topnode; currnode; currnode = currnode->listnext) |
91f93b94 |
520 | fputs(currnode->text.output.text, fp); |
5dd44dce |
521 | } else { |
522 | int filenum = 0; |
523 | fprintf(fp, "\037\nIndirect:\n"); |
524 | for (currnode = topnode; currnode; currnode = currnode->listnext) |
525 | if (filenum != currnode->filenum) { |
526 | filenum = currnode->filenum; |
527 | fprintf(fp, "%s-%d: %d\n", conf.filename, filenum, |
528 | currnode->pos); |
529 | } |
530 | } |
531 | fprintf(fp, "\037\nTag Table:\n"); |
532 | if (conf.maxfilesize > 0) |
533 | fprintf(fp, "(Indirect)\n"); |
534 | for (currnode = topnode; currnode; currnode = currnode->listnext) |
535 | fprintf(fp, "Node: %s\177%d\n", currnode->name, currnode->pos); |
536 | fprintf(fp, "\037\nEnd Tag Table\n"); |
537 | fclose(fp); |
538 | |
539 | /* |
540 | * Write the subfiles. |
541 | */ |
542 | if (conf.maxfilesize > 0) { |
543 | int filenum = 0; |
544 | fp = NULL; |
545 | |
546 | for (currnode = topnode; currnode; currnode = currnode->listnext) { |
547 | if (filenum != currnode->filenum) { |
548 | char *fname; |
549 | |
550 | filenum = currnode->filenum; |
551 | |
552 | if (fp) |
553 | fclose(fp); |
554 | fname = mknewa(char, strlen(conf.filename) + 40); |
555 | sprintf(fname, "%s-%d", conf.filename, filenum); |
556 | fp = fopen(fname, "w"); |
557 | if (!fp) { |
558 | error(err_cantopenw, fname); |
559 | return; |
560 | } |
561 | sfree(fname); |
91f93b94 |
562 | fputs(intro_text.output.text, fp); |
5dd44dce |
563 | } |
91f93b94 |
564 | fputs(currnode->text.output.text, fp); |
5dd44dce |
565 | } |
566 | |
567 | if (fp) |
568 | fclose(fp); |
569 | } |
570 | } |
571 | |
572 | static int info_check_index(word *w, node *n, indexdata *idx) |
573 | { |
574 | int ret = 0; |
575 | |
576 | for (; w; w = w->next) { |
577 | if (w->type == word_IndexRef) { |
578 | indextag *tag; |
579 | int i; |
580 | |
581 | tag = index_findtag(idx, w->text); |
582 | if (!tag) |
583 | break; |
584 | |
585 | for (i = 0; i < tag->nrefs; i++) { |
586 | indexentry *entry = tag->refs[i]; |
587 | info_idx *ii = (info_idx *)entry->backend_data; |
588 | |
589 | if (ii->nnodes > 0 && ii->nodes[ii->nnodes-1] == n) { |
590 | /* |
591 | * If the same index term is indexed twice |
592 | * within the same section, we only want to |
593 | * mention it once in the index. So do nothing |
594 | * here. |
595 | */ |
596 | continue; |
597 | } |
598 | |
599 | if (ii->nnodes >= ii->nodesize) { |
600 | ii->nodesize += 32; |
601 | ii->nodes = resize(ii->nodes, ii->nodesize); |
602 | } |
603 | |
604 | ii->nodes[ii->nnodes++] = n; |
605 | |
606 | ret = 1; |
607 | } |
608 | } |
609 | } |
610 | |
611 | return ret; |
612 | } |
613 | |
5dd44dce |
614 | static word *info_transform_wordlist(word *words, keywordlist *keywords) |
615 | { |
616 | word *ret = dup_word_list(words); |
617 | word *w; |
618 | keyword *kwl; |
619 | |
620 | for (w = ret; w; w = w->next) { |
621 | w->private_data = NULL; |
622 | if (w->type == word_UpperXref || w->type == word_LowerXref) { |
623 | kwl = kw_lookup(keywords, w->text); |
624 | if (kwl) { |
625 | if (kwl->para->type == para_NumberedList || |
626 | kwl->para->type == para_BiblioCited) { |
627 | /* |
628 | * In Info, we do nothing special for xrefs to |
629 | * numbered list items or bibliography entries. |
630 | */ |
631 | break; |
632 | } else { |
633 | /* |
634 | * An xref to a different section has its text |
635 | * completely replaced. |
636 | */ |
637 | word *w2, *w3, *w4; |
638 | w2 = w3 = w->next; |
639 | w4 = NULL; |
640 | while (w2) { |
641 | if (w2->type == word_XrefEnd) { |
642 | w4 = w2->next; |
643 | w2->next = NULL; |
644 | break; |
645 | } |
646 | w2 = w2->next; |
647 | } |
648 | free_word_list(w3); |
649 | |
650 | /* |
651 | * Now w is the UpperXref / LowerXref we |
652 | * started with, and w4 is the next word after |
653 | * the corresponding XrefEnd (if any). The |
654 | * simplest thing is just to stick a pointer to |
655 | * the target node structure in the private |
656 | * data field of the xref word, and let |
657 | * info_rdaddwc and friends read the node name |
658 | * out from there. |
659 | */ |
660 | w->next = w4; |
661 | w->private_data = kwl->para->private_data; |
662 | assert(w->private_data); |
663 | } |
664 | } |
665 | } |
666 | } |
667 | |
668 | return ret; |
669 | } |
670 | |
91f93b94 |
671 | static int info_rdaddwc(info_data *id, word *words, word *end, int xrefs) { |
672 | int ret = 0; |
5dd44dce |
673 | |
674 | for (; words && words != end; words = words->next) switch (words->type) { |
675 | case word_HyperLink: |
676 | case word_HyperEnd: |
677 | case word_XrefEnd: |
678 | case word_IndexRef: |
679 | break; |
680 | |
681 | case word_Normal: |
682 | case word_Emph: |
683 | case word_Code: |
684 | case word_WeakCode: |
685 | case word_WhiteSpace: |
686 | case word_EmphSpace: |
687 | case word_CodeSpace: |
688 | case word_WkCodeSpace: |
689 | case word_Quote: |
690 | case word_EmphQuote: |
691 | case word_CodeQuote: |
692 | case word_WkCodeQuote: |
693 | assert(words->type != word_CodeQuote && |
694 | words->type != word_WkCodeQuote); |
695 | if (towordstyle(words->type) == word_Emph && |
696 | (attraux(words->aux) == attr_First || |
697 | attraux(words->aux) == attr_Only)) |
91f93b94 |
698 | ret += info_rdadd(id, L'_'); /* FIXME: configurability */ |
5dd44dce |
699 | else if (towordstyle(words->type) == word_Code && |
700 | (attraux(words->aux) == attr_First || |
701 | attraux(words->aux) == attr_Only)) |
91f93b94 |
702 | ret += info_rdadd(id, L'`'); /* FIXME: configurability */ |
5dd44dce |
703 | if (removeattr(words->type) == word_Normal) { |
91f93b94 |
704 | if (cvt_ok(id->charset, words->text) || !words->alt) |
705 | ret += info_rdadds(id, words->text); |
5dd44dce |
706 | else |
91f93b94 |
707 | ret += info_rdaddwc(id, words->alt, NULL, FALSE); |
5dd44dce |
708 | } else if (removeattr(words->type) == word_WhiteSpace) { |
91f93b94 |
709 | ret += info_rdadd(id, L' '); |
5dd44dce |
710 | } else if (removeattr(words->type) == word_Quote) { |
91f93b94 |
711 | ret += info_rdadd(id, quoteaux(words->aux) == quote_Open ? L'`' : L'\''); |
5dd44dce |
712 | /* FIXME: configurability */ |
713 | } |
714 | if (towordstyle(words->type) == word_Emph && |
715 | (attraux(words->aux) == attr_Last || |
716 | attraux(words->aux) == attr_Only)) |
91f93b94 |
717 | ret += info_rdadd(id, L'_'); /* FIXME: configurability */ |
5dd44dce |
718 | else if (towordstyle(words->type) == word_Code && |
719 | (attraux(words->aux) == attr_Last || |
720 | attraux(words->aux) == attr_Only)) |
91f93b94 |
721 | ret += info_rdadd(id, L'\''); /* FIXME: configurability */ |
5dd44dce |
722 | break; |
723 | |
724 | case word_UpperXref: |
725 | case word_LowerXref: |
726 | if (xrefs && words->private_data) { |
91f93b94 |
727 | /* |
728 | * This bit is structural and so must be done in char |
729 | * rather than wchar_t. |
730 | */ |
731 | ret += info_rdaddsc(id, "*Note "); |
732 | ret += info_rdaddsc(id, ((node *)words->private_data)->name); |
733 | ret += info_rdaddsc(id, "::"); |
5dd44dce |
734 | } |
735 | break; |
736 | } |
91f93b94 |
737 | |
738 | return ret; |
5dd44dce |
739 | } |
740 | |
91f93b94 |
741 | static int info_width_internal(word *words, int xrefs, int charset); |
5dd44dce |
742 | |
91f93b94 |
743 | static int info_width_internal_list(word *words, int xrefs, int charset) { |
5dd44dce |
744 | int w = 0; |
745 | while (words) { |
91f93b94 |
746 | w += info_width_internal(words, xrefs, charset); |
5dd44dce |
747 | words = words->next; |
748 | } |
749 | return w; |
750 | } |
751 | |
91f93b94 |
752 | static int info_width_internal(word *words, int xrefs, int charset) { |
5dd44dce |
753 | switch (words->type) { |
754 | case word_HyperLink: |
755 | case word_HyperEnd: |
756 | case word_XrefEnd: |
757 | case word_IndexRef: |
758 | return 0; |
759 | |
760 | case word_Normal: |
761 | case word_Emph: |
762 | case word_Code: |
763 | case word_WeakCode: |
764 | return (((words->type == word_Emph || |
765 | words->type == word_Code) |
766 | ? (attraux(words->aux) == attr_Only ? 2 : |
767 | attraux(words->aux) == attr_Always ? 0 : 1) |
768 | : 0) + |
91f93b94 |
769 | (cvt_ok(charset, words->text) || !words->alt ? |
5dd44dce |
770 | ustrlen(words->text) : |
91f93b94 |
771 | info_width_internal_list(words->alt, xrefs, charset))); |
5dd44dce |
772 | |
773 | case word_WhiteSpace: |
774 | case word_EmphSpace: |
775 | case word_CodeSpace: |
776 | case word_WkCodeSpace: |
777 | case word_Quote: |
778 | case word_EmphQuote: |
779 | case word_CodeQuote: |
780 | case word_WkCodeQuote: |
781 | assert(words->type != word_CodeQuote && |
782 | words->type != word_WkCodeQuote); |
783 | return (((towordstyle(words->type) == word_Emph || |
784 | towordstyle(words->type) == word_Code) |
785 | ? (attraux(words->aux) == attr_Only ? 2 : |
786 | attraux(words->aux) == attr_Always ? 0 : 1) |
787 | : 0) + 1); |
788 | |
789 | case word_UpperXref: |
790 | case word_LowerXref: |
791 | if (xrefs && words->private_data) { |
792 | /* "*Note " plus "::" comes to 8 characters */ |
793 | return 8 + strlen(((node *)words->private_data)->name); |
794 | } |
795 | break; |
796 | } |
797 | return 0; /* should never happen */ |
798 | } |
799 | |
43341922 |
800 | static int info_width_noxrefs(void *ctx, word *words) |
5dd44dce |
801 | { |
91f93b94 |
802 | return info_width_internal(words, FALSE, *(int *)ctx); |
5dd44dce |
803 | } |
43341922 |
804 | static int info_width_xrefs(void *ctx, word *words) |
5dd44dce |
805 | { |
91f93b94 |
806 | return info_width_internal(words, TRUE, *(int *)ctx); |
5dd44dce |
807 | } |
808 | |
91f93b94 |
809 | static void info_heading(info_data *text, word *tprefix, |
5dd44dce |
810 | word *words, int width) { |
91f93b94 |
811 | int length; |
5dd44dce |
812 | int firstlinewidth, wrapwidth; |
5dd44dce |
813 | wrappedline *wrapping, *p; |
814 | |
91f93b94 |
815 | length = 0; |
5dd44dce |
816 | if (tprefix) { |
91f93b94 |
817 | length += info_rdaddwc(text, tprefix, NULL, FALSE); |
818 | length += info_rdadds(text, L": ");/* FIXME: configurability */ |
5dd44dce |
819 | } |
5dd44dce |
820 | |
5dd44dce |
821 | wrapwidth = width; |
91f93b94 |
822 | firstlinewidth = width - length; |
5dd44dce |
823 | |
43341922 |
824 | wrapping = wrap_para(words, firstlinewidth, wrapwidth, |
91f93b94 |
825 | info_width_noxrefs, &text->charset, 0); |
5dd44dce |
826 | for (p = wrapping; p; p = p->next) { |
91f93b94 |
827 | length += info_rdaddwc(text, p->begin, p->end, FALSE); |
828 | info_rdadd(text, L'\n'); |
5dd44dce |
829 | while (length--) |
91f93b94 |
830 | info_rdadd(text, L'-'); /* FIXME: configurability */ |
831 | info_rdadd(text, L'\n'); |
832 | length = 0; |
5dd44dce |
833 | } |
834 | wrap_free(wrapping); |
91f93b94 |
835 | info_rdadd(text, L'\n'); |
5dd44dce |
836 | } |
837 | |
91f93b94 |
838 | static void info_rule(info_data *text, int indent, int width) { |
839 | while (indent--) info_rdadd(text, L' '); |
840 | while (width--) info_rdadd(text, L'-'); |
841 | info_rdadd(text, L'\n'); |
842 | info_rdadd(text, L'\n'); |
5dd44dce |
843 | } |
844 | |
91f93b94 |
845 | static void info_para(info_data *text, word *prefix, wchar_t *prefixextra, |
5dd44dce |
846 | word *input, keywordlist *keywords, |
847 | int indent, int extraindent, int width) { |
848 | wrappedline *wrapping, *p; |
849 | word *words; |
5dd44dce |
850 | int e; |
851 | int i; |
852 | int firstlinewidth = width; |
853 | |
854 | words = info_transform_wordlist(input, keywords); |
855 | |
856 | if (prefix) { |
5dd44dce |
857 | for (i = 0; i < indent; i++) |
91f93b94 |
858 | info_rdadd(text, L' '); |
859 | e = info_rdaddwc(text, prefix, NULL, FALSE); |
860 | if (prefixextra) |
861 | e += info_rdadds(text, prefixextra); |
5dd44dce |
862 | /* If the prefix is too long, shorten the first line to fit. */ |
91f93b94 |
863 | e = extraindent - e; |
5dd44dce |
864 | if (e < 0) { |
865 | firstlinewidth += e; /* this decreases it, since e < 0 */ |
866 | if (firstlinewidth < 0) { |
867 | e = indent + extraindent; |
868 | firstlinewidth = width; |
91f93b94 |
869 | info_rdadd(text, L'\n'); |
5dd44dce |
870 | } else |
871 | e = 0; |
872 | } |
5dd44dce |
873 | } else |
874 | e = indent + extraindent; |
875 | |
43341922 |
876 | wrapping = wrap_para(words, firstlinewidth, width, info_width_xrefs, |
91f93b94 |
877 | &text->charset, 0); |
5dd44dce |
878 | for (p = wrapping; p; p = p->next) { |
879 | for (i = 0; i < e; i++) |
91f93b94 |
880 | info_rdadd(text, L' '); |
5dd44dce |
881 | info_rdaddwc(text, p->begin, p->end, TRUE); |
91f93b94 |
882 | info_rdadd(text, L'\n'); |
5dd44dce |
883 | e = indent + extraindent; |
884 | } |
885 | wrap_free(wrapping); |
91f93b94 |
886 | info_rdadd(text, L'\n'); |
5dd44dce |
887 | |
888 | free_word_list(words); |
889 | } |
890 | |
91f93b94 |
891 | static void info_codepara(info_data *text, word *words, |
5dd44dce |
892 | int indent, int width) { |
893 | int i; |
894 | |
895 | for (; words; words = words->next) if (words->type == word_WeakCode) { |
91f93b94 |
896 | for (i = 0; i < indent; i++) |
897 | info_rdadd(text, L' '); |
898 | if (info_rdadds(text, words->text) > width) { |
5dd44dce |
899 | /* FIXME: warn */ |
900 | } |
91f93b94 |
901 | info_rdadd(text, L'\n'); |
5dd44dce |
902 | } |
903 | |
91f93b94 |
904 | info_rdadd(text, L'\n'); |
5dd44dce |
905 | } |
906 | |
91f93b94 |
907 | static void info_versionid(info_data *text, word *words) { |
908 | info_rdadd(text, L'['); /* FIXME: configurability */ |
5dd44dce |
909 | info_rdaddwc(text, words, NULL, FALSE); |
91f93b94 |
910 | info_rdadds(text, L"]\n"); |
5dd44dce |
911 | } |
912 | |
91f93b94 |
913 | static node *info_node_new(char *name, int charset) |
5dd44dce |
914 | { |
915 | node *n; |
916 | |
917 | n = mknew(node); |
91f93b94 |
918 | n->text = empty_info_data; |
919 | n->text.charset = charset; |
5dd44dce |
920 | n->up = n->next = n->prev = n->lastchild = n->listnext = NULL; |
921 | n->name = dupstr(name); |
922 | n->started_menu = FALSE; |
923 | |
924 | return n; |
925 | } |
926 | |
91f93b94 |
927 | static char *info_node_name(paragraph *par, int charset) |
5dd44dce |
928 | { |
91f93b94 |
929 | info_data id = EMPTY_INFO_DATA; |
f4551933 |
930 | char *p, *q; |
91f93b94 |
931 | |
932 | id.charset = charset; |
933 | info_rdaddwc(&id, par->kwtext ? par->kwtext : par->words, NULL, FALSE); |
934 | info_rdaddsc(&id, NULL); |
f4551933 |
935 | |
936 | /* |
937 | * We cannot have commas or colons in a node name. Remove any |
938 | * that we find, with a warning. |
939 | */ |
91f93b94 |
940 | p = q = id.output.text; |
f4551933 |
941 | while (*p) { |
942 | if (*p == ':' || *p == ',') { |
943 | error(err_infonodechar, &par->fpos, *p); |
944 | } else { |
945 | *q++ = *p; |
946 | } |
947 | p++; |
948 | } |
949 | *p = '\0'; |
950 | |
91f93b94 |
951 | return id.output.text; |
5dd44dce |
952 | } |
953 | |
91f93b94 |
954 | static void info_menu_item(info_data *text, node *n, paragraph *p) |
5dd44dce |
955 | { |
956 | /* |
957 | * FIXME: Depending on how we're doing node names in this info |
958 | * file, we might want to do |
959 | * |
960 | * * Node name:: Chapter title |
961 | * |
962 | * _or_ |
963 | * |
964 | * * Chapter number: Node name. |
965 | * |
91f93b94 |
966 | * This function mostly works in char rather than wchar_t, |
967 | * because a menu item is a structural component. |
5dd44dce |
968 | */ |
91f93b94 |
969 | info_rdaddsc(text, "* "); |
970 | info_rdaddsc(text, n->name); |
971 | info_rdaddsc(text, "::"); |
5dd44dce |
972 | if (p) { |
91f93b94 |
973 | info_rdaddc(text, ' '); |
5dd44dce |
974 | info_rdaddwc(text, p->words, NULL, FALSE); |
975 | } |
91f93b94 |
976 | info_rdaddc(text, '\n'); |
977 | } |
978 | |
979 | /* |
980 | * These functions implement my wrapper on the rdadd* calls which |
981 | * allows me to switch arbitrarily between literal octet-string |
982 | * text and charset-translated Unicode. (Because no matter what |
983 | * character set I write the actual text in, I expect info readers |
984 | * to treat node names and file names literally and to expect |
985 | * keywords like `*Note' in their canonical form, so I have to take |
986 | * steps to ensure that those structural elements of the file |
987 | * aren't messed with.) |
988 | */ |
989 | static int info_rdadds(info_data *d, wchar_t const *wcs) |
990 | { |
991 | if (!d->wcmode) { |
992 | d->state = charset_init_state; |
993 | d->wcmode = TRUE; |
994 | } |
995 | |
996 | if (wcs) { |
997 | char buf[256]; |
998 | int len, origlen, ret; |
999 | |
1000 | origlen = len = ustrlen(wcs); |
1001 | while (len > 0) { |
1002 | int prevlen = len; |
1003 | |
1004 | ret = charset_from_unicode(&wcs, &len, buf, lenof(buf), |
1005 | d->charset, &d->state, NULL); |
1006 | |
1007 | assert(len < prevlen); |
1008 | |
1009 | if (ret > 0) { |
1010 | buf[ret] = '\0'; |
1011 | rdaddsc(&d->output, buf); |
1012 | } |
1013 | } |
1014 | |
1015 | return origlen; |
1016 | } else |
1017 | return 0; |
1018 | } |
1019 | |
1020 | static int info_rdaddsc(info_data *d, char const *cs) |
1021 | { |
1022 | if (d->wcmode) { |
1023 | char buf[256]; |
1024 | int ret; |
1025 | |
1026 | ret = charset_from_unicode(NULL, 0, buf, lenof(buf), |
1027 | d->charset, &d->state, NULL); |
1028 | if (ret > 0) { |
1029 | buf[ret] = '\0'; |
1030 | rdaddsc(&d->output, buf); |
1031 | } |
1032 | |
1033 | d->wcmode = FALSE; |
1034 | } |
1035 | |
1036 | if (cs) { |
1037 | rdaddsc(&d->output, cs); |
1038 | return strlen(cs); |
1039 | } else |
1040 | return 0; |
1041 | } |
1042 | |
1043 | static int info_rdadd(info_data *d, wchar_t wc) |
1044 | { |
1045 | wchar_t wcs[2]; |
1046 | wcs[0] = wc; |
1047 | wcs[1] = L'\0'; |
1048 | return info_rdadds(d, wcs); |
1049 | } |
1050 | |
1051 | static int info_rdaddc(info_data *d, char c) |
1052 | { |
1053 | char cs[2]; |
1054 | cs[0] = c; |
1055 | cs[1] = '\0'; |
1056 | return info_rdaddsc(d, cs); |
5dd44dce |
1057 | } |