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 | |
28 | #include <stdio.h> |
29 | #include <stdlib.h> |
30 | #include <assert.h> |
31 | #include "halibut.h" |
32 | |
33 | typedef struct { |
34 | char *filename; |
35 | int maxfilesize; |
91f93b94 |
36 | int charset; |
5dd44dce |
37 | } infoconfig; |
38 | |
91f93b94 |
39 | typedef struct { |
40 | rdstringc output; |
41 | int charset; |
42 | charset_state state; |
43 | int wcmode; |
44 | } info_data; |
45 | #define EMPTY_INFO_DATA { { 0, 0, NULL }, 0, CHARSET_INIT_STATE, FALSE } |
46 | static const info_data empty_info_data = EMPTY_INFO_DATA; |
47 | |
5dd44dce |
48 | typedef struct node_tag node; |
49 | struct node_tag { |
50 | node *listnext; |
51 | node *up, *prev, *next, *lastchild; |
52 | int pos, started_menu, filenum; |
53 | char *name; |
91f93b94 |
54 | info_data text; |
5dd44dce |
55 | }; |
56 | |
57 | typedef struct { |
58 | char *text; |
91f93b94 |
59 | int length; |
5dd44dce |
60 | int nnodes, nodesize; |
61 | node **nodes; |
62 | } info_idx; |
63 | |
91f93b94 |
64 | static int info_rdadd(info_data *, wchar_t); |
65 | static int info_rdadds(info_data *, wchar_t const *); |
66 | static int info_rdaddc(info_data *, char); |
67 | static int info_rdaddsc(info_data *, char const *); |
5dd44dce |
68 | |
91f93b94 |
69 | static void info_heading(info_data *, word *, word *, int); |
70 | static void info_rule(info_data *, int, int); |
71 | static void info_para(info_data *, word *, wchar_t *, word *, keywordlist *, |
5dd44dce |
72 | int, int, int); |
91f93b94 |
73 | static void info_codepara(info_data *, word *, int, int); |
74 | static void info_versionid(info_data *, word *); |
75 | static void info_menu_item(info_data *, node *, paragraph *); |
5dd44dce |
76 | static word *info_transform_wordlist(word *, keywordlist *); |
77 | static int info_check_index(word *, node *, indexdata *); |
78 | |
91f93b94 |
79 | static int info_rdaddwc(info_data *, word *, word *, int); |
5dd44dce |
80 | |
91f93b94 |
81 | static node *info_node_new(char *name, int charset); |
82 | static char *info_node_name(paragraph *p, int charset); |
5dd44dce |
83 | |
84 | static infoconfig info_configure(paragraph *source) { |
85 | infoconfig ret; |
86 | |
87 | /* |
88 | * Defaults. |
89 | */ |
90 | ret.filename = dupstr("output.info"); |
91 | ret.maxfilesize = 64 << 10; |
91f93b94 |
92 | ret.charset = CS_ASCII; |
5dd44dce |
93 | |
94 | for (; source; source = source->next) { |
95 | if (source->type == para_Config) { |
96 | if (!ustricmp(source->keyword, L"info-filename")) { |
97 | sfree(ret.filename); |
e4ea58f8 |
98 | ret.filename = dupstr(adv(source->origkeyword)); |
91f93b94 |
99 | } else if (!ustricmp(source->keyword, L"info-charset")) { |
100 | char *csname = utoa_dup(uadv(source->keyword), CS_ASCII); |
101 | ret.charset = charset_from_localenc(csname); |
102 | sfree(csname); |
5dd44dce |
103 | } else if (!ustricmp(source->keyword, L"info-max-file-size")) { |
104 | ret.maxfilesize = utoi(uadv(source->keyword)); |
105 | } |
106 | } |
107 | } |
108 | |
109 | return ret; |
110 | } |
111 | |
112 | paragraph *info_config_filename(char *filename) |
113 | { |
e4ea58f8 |
114 | return cmdline_cfg_simple("info-filename", filename, NULL); |
5dd44dce |
115 | } |
116 | |
117 | void info_backend(paragraph *sourceform, keywordlist *keywords, |
43341922 |
118 | indexdata *idx, void *unused) { |
5dd44dce |
119 | paragraph *p; |
120 | infoconfig conf; |
121 | word *prefix, *body, *wp; |
122 | word spaceword; |
91f93b94 |
123 | wchar_t *prefixextra; |
5dd44dce |
124 | int nesting, nestindent; |
125 | int indentb, indenta; |
126 | int filepos; |
127 | int has_index; |
91f93b94 |
128 | info_data intro_text = EMPTY_INFO_DATA; |
5dd44dce |
129 | node *topnode, *currnode; |
130 | word bullet; |
131 | FILE *fp; |
132 | |
133 | /* |
91f93b94 |
134 | * FIXME: possibly configurability? |
5dd44dce |
135 | */ |
136 | int width = 70, listindentbefore = 1, listindentafter = 3; |
137 | int indent_code = 2, index_width = 40; |
138 | |
43341922 |
139 | IGNORE(unused); |
5dd44dce |
140 | |
141 | conf = info_configure(sourceform); |
142 | |
143 | /* |
144 | * Go through and create a node for each section. |
145 | */ |
91f93b94 |
146 | topnode = info_node_new("Top", conf.charset); |
5dd44dce |
147 | currnode = topnode; |
148 | for (p = sourceform; p; p = p->next) switch (p->type) { |
149 | /* |
150 | * Chapter titles. |
151 | */ |
152 | case para_Chapter: |
153 | case para_Appendix: |
154 | case para_UnnumberedChapter: |
155 | case para_Heading: |
156 | case para_Subsect: |
157 | { |
158 | node *newnode, *upnode; |
159 | char *nodename; |
160 | |
91f93b94 |
161 | nodename = info_node_name(p, conf.charset); |
162 | newnode = info_node_new(nodename, conf.charset); |
5dd44dce |
163 | sfree(nodename); |
164 | |
165 | p->private_data = newnode; |
166 | |
167 | if (p->parent) |
168 | upnode = (node *)p->parent->private_data; |
169 | else |
170 | upnode = topnode; |
171 | assert(upnode); |
172 | newnode->up = upnode; |
173 | |
174 | currnode->next = newnode; |
175 | newnode->prev = currnode; |
176 | |
177 | currnode->listnext = newnode; |
178 | currnode = newnode; |
179 | } |
180 | break; |
181 | } |
182 | |
183 | /* |
184 | * Set up the display form of each index entry. |
185 | */ |
186 | { |
187 | int i; |
188 | indexentry *entry; |
189 | |
190 | for (i = 0; (entry = index234(idx->entries, i)) != NULL; i++) { |
191 | info_idx *ii = mknew(info_idx); |
91f93b94 |
192 | info_data id = EMPTY_INFO_DATA; |
193 | |
194 | id.charset = conf.charset; |
5dd44dce |
195 | |
196 | ii->nnodes = ii->nodesize = 0; |
197 | ii->nodes = NULL; |
198 | |
91f93b94 |
199 | ii->length = info_rdaddwc(&id, entry->text, NULL, FALSE); |
f4551933 |
200 | |
91f93b94 |
201 | ii->text = id.output.text; |
5dd44dce |
202 | |
203 | entry->backend_data = ii; |
204 | } |
205 | } |
206 | |
207 | /* |
208 | * An Info file begins with a piece of introductory text which |
209 | * is apparently never shown anywhere. This seems to me to be a |
d4c7e130 |
210 | * good place to put the copyright notice and the version IDs. |
211 | * Also, Info directory entries are expected to go here. |
5dd44dce |
212 | */ |
91f93b94 |
213 | intro_text.charset = conf.charset; |
5dd44dce |
214 | |
91f93b94 |
215 | info_rdaddsc(&intro_text, |
5dd44dce |
216 | "This Info file generated by Halibut, "); |
91f93b94 |
217 | info_rdaddsc(&intro_text, version); |
218 | info_rdaddsc(&intro_text, "\n\n"); |
5dd44dce |
219 | |
220 | for (p = sourceform; p; p = p->next) |
d4c7e130 |
221 | if (p->type == para_Config && |
222 | !ustricmp(p->keyword, L"info-dir-entry")) { |
223 | wchar_t *section, *shortname, *longname, *kw; |
224 | char *s; |
225 | |
226 | section = uadv(p->keyword); |
227 | shortname = *section ? uadv(section) : NULL; |
228 | longname = *shortname ? uadv(shortname) : NULL; |
229 | kw = *longname ? uadv(longname) : NULL; |
230 | |
231 | if (!*longname) { |
232 | error(err_infodirentry, &p->fpos); |
233 | continue; |
234 | } |
235 | |
91f93b94 |
236 | info_rdaddsc(&intro_text, "INFO-DIR-SECTION "); |
237 | info_rdadds(&intro_text, section); |
238 | info_rdaddsc(&intro_text, "\nSTART-INFO-DIR-ENTRY\n* "); |
239 | info_rdadds(&intro_text, shortname); |
240 | info_rdaddsc(&intro_text, ": ("); |
d4c7e130 |
241 | s = dupstr(conf.filename); |
242 | if (strlen(s) > 5 && !strcmp(s+strlen(s)-5, ".info")) |
243 | s[strlen(s)-5] = '\0'; |
91f93b94 |
244 | info_rdaddsc(&intro_text, s); |
d4c7e130 |
245 | sfree(s); |
91f93b94 |
246 | info_rdaddsc(&intro_text, ")"); |
d4c7e130 |
247 | if (*kw) { |
248 | keyword *kwl = kw_lookup(keywords, kw); |
249 | if (kwl && kwl->para->private_data) { |
250 | node *n = (node *)kwl->para->private_data; |
91f93b94 |
251 | info_rdaddsc(&intro_text, n->name); |
d4c7e130 |
252 | } |
253 | } |
91f93b94 |
254 | info_rdaddsc(&intro_text, ". "); |
255 | info_rdadds(&intro_text, longname); |
256 | info_rdaddsc(&intro_text, "\nEND-INFO-DIR-ENTRY\n\n"); |
d4c7e130 |
257 | } |
258 | |
259 | for (p = sourceform; p; p = p->next) |
5dd44dce |
260 | if (p->type == para_Copyright) |
261 | info_para(&intro_text, NULL, NULL, p->words, keywords, |
262 | 0, 0, width); |
263 | |
264 | for (p = sourceform; p; p = p->next) |
265 | if (p->type == para_VersionID) |
266 | info_versionid(&intro_text, p->words); |
267 | |
91f93b94 |
268 | if (intro_text.output.text[intro_text.output.pos-1] != '\n') |
269 | info_rdaddc(&intro_text, '\n'); |
5dd44dce |
270 | |
271 | /* Do the title */ |
272 | for (p = sourceform; p; p = p->next) |
273 | if (p->type == para_Title) |
274 | info_heading(&topnode->text, NULL, p->words, width); |
275 | |
276 | nestindent = listindentbefore + listindentafter; |
277 | nesting = 0; |
278 | |
279 | currnode = topnode; |
280 | |
281 | /* Do the main document */ |
282 | for (p = sourceform; p; p = p->next) switch (p->type) { |
283 | |
284 | case para_QuotePush: |
285 | nesting += 2; |
286 | break; |
287 | case para_QuotePop: |
288 | nesting -= 2; |
289 | assert(nesting >= 0); |
290 | break; |
291 | |
292 | case para_LcontPush: |
293 | nesting += nestindent; |
294 | break; |
295 | case para_LcontPop: |
296 | nesting -= nestindent; |
297 | assert(nesting >= 0); |
298 | break; |
299 | |
300 | /* |
301 | * Things we ignore because we've already processed them or |
302 | * aren't going to touch them in this pass. |
303 | */ |
304 | case para_IM: |
305 | case para_BR: |
306 | case para_Biblio: /* only touch BiblioCited */ |
307 | case para_VersionID: |
308 | case para_NoCite: |
309 | case para_Title: |
310 | break; |
311 | |
312 | /* |
313 | * Chapter titles. |
314 | */ |
315 | case para_Chapter: |
316 | case para_Appendix: |
317 | case para_UnnumberedChapter: |
318 | case para_Heading: |
319 | case para_Subsect: |
320 | currnode = p->private_data; |
321 | assert(currnode); |
322 | assert(currnode->up); |
323 | |
324 | if (!currnode->up->started_menu) { |
91f93b94 |
325 | info_rdaddsc(&currnode->up->text, "* Menu:\n\n"); |
5dd44dce |
326 | currnode->up->started_menu = TRUE; |
327 | } |
328 | info_menu_item(&currnode->up->text, currnode, p); |
329 | |
330 | has_index |= info_check_index(p->words, currnode, idx); |
331 | info_heading(&currnode->text, p->kwtext, p->words, width); |
332 | nesting = 0; |
333 | break; |
334 | |
335 | case para_Rule: |
336 | info_rule(&currnode->text, nesting, width - nesting); |
337 | break; |
338 | |
339 | case para_Normal: |
340 | case para_Copyright: |
341 | case para_DescribedThing: |
342 | case para_Description: |
343 | case para_BiblioCited: |
344 | case para_Bullet: |
345 | case para_NumberedList: |
346 | has_index |= info_check_index(p->words, currnode, idx); |
347 | if (p->type == para_Bullet) { |
348 | bullet.next = NULL; |
349 | bullet.alt = NULL; |
350 | bullet.type = word_Normal; |
351 | bullet.text = L"-"; /* FIXME: configurability */ |
352 | prefix = • |
353 | prefixextra = NULL; |
354 | indentb = listindentbefore; |
355 | indenta = listindentafter; |
356 | } else if (p->type == para_NumberedList) { |
357 | prefix = p->kwtext; |
91f93b94 |
358 | prefixextra = L"."; /* FIXME: configurability */ |
5dd44dce |
359 | indentb = listindentbefore; |
360 | indenta = listindentafter; |
361 | } else if (p->type == para_Description) { |
362 | prefix = NULL; |
363 | prefixextra = NULL; |
364 | indentb = listindentbefore; |
365 | indenta = listindentafter; |
366 | } else { |
367 | prefix = NULL; |
368 | prefixextra = NULL; |
369 | indentb = indenta = 0; |
370 | } |
371 | if (p->type == para_BiblioCited) { |
372 | body = dup_word_list(p->kwtext); |
373 | for (wp = body; wp->next; wp = wp->next); |
374 | wp->next = &spaceword; |
375 | spaceword.next = p->words; |
376 | spaceword.alt = NULL; |
377 | spaceword.type = word_WhiteSpace; |
378 | spaceword.text = NULL; |
379 | } else { |
380 | wp = NULL; |
381 | body = p->words; |
382 | } |
383 | info_para(&currnode->text, prefix, prefixextra, body, keywords, |
384 | nesting + indentb, indenta, |
385 | width - nesting - indentb - indenta); |
386 | if (wp) { |
387 | wp->next = NULL; |
388 | free_word_list(body); |
389 | } |
390 | break; |
391 | |
392 | case para_Code: |
393 | info_codepara(&currnode->text, p->words, |
394 | nesting + indent_code, |
395 | width - nesting - 2 * indent_code); |
396 | break; |
397 | } |
398 | |
399 | /* |
400 | * Create an index node if required. |
401 | */ |
402 | if (has_index) { |
403 | node *newnode; |
404 | int i, j, k; |
405 | indexentry *entry; |
406 | |
91f93b94 |
407 | newnode = info_node_new("Index", conf.charset); |
5dd44dce |
408 | newnode->up = topnode; |
409 | |
410 | currnode->next = newnode; |
411 | newnode->prev = currnode; |
412 | currnode->listnext = newnode; |
413 | |
91f93b94 |
414 | info_rdaddsc(&newnode->text, "Index\n-----\n\n"); |
5dd44dce |
415 | |
416 | info_menu_item(&topnode->text, newnode, NULL); |
417 | |
418 | for (i = 0; (entry = index234(idx->entries, i)) != NULL; i++) { |
419 | info_idx *ii = (info_idx *)entry->backend_data; |
420 | |
421 | for (j = 0; j < ii->nnodes; j++) { |
5dd44dce |
422 | /* |
423 | * When we have multiple references for a single |
424 | * index term, we only display the actual term on |
425 | * the first line, to make it clear that the terms |
426 | * really are the same. |
427 | */ |
428 | if (j == 0) |
91f93b94 |
429 | info_rdaddsc(&newnode->text, ii->text); |
430 | for (k = (j ? 0 : ii->length); k < index_width; k++) |
431 | info_rdaddc(&newnode->text, ' '); |
432 | info_rdaddsc(&newnode->text, " *Note "); |
433 | info_rdaddsc(&newnode->text, ii->nodes[j]->name); |
434 | info_rdaddsc(&newnode->text, "::\n"); |
5dd44dce |
435 | } |
436 | } |
437 | } |
438 | |
439 | /* |
440 | * Finalise the text of each node, by adding the ^_ delimiter |
441 | * and the node line at the top. |
442 | */ |
443 | for (currnode = topnode; currnode; currnode = currnode->listnext) { |
91f93b94 |
444 | char *origtext = currnode->text.output.text; |
445 | currnode->text = empty_info_data; |
446 | currnode->text.charset = conf.charset; |
447 | info_rdaddsc(&currnode->text, "\037\nFile: "); |
448 | info_rdaddsc(&currnode->text, conf.filename); |
449 | info_rdaddsc(&currnode->text, ", Node: "); |
450 | info_rdaddsc(&currnode->text, currnode->name); |
5dd44dce |
451 | if (currnode->prev) { |
91f93b94 |
452 | info_rdaddsc(&currnode->text, ", Prev: "); |
453 | info_rdaddsc(&currnode->text, currnode->prev->name); |
5dd44dce |
454 | } |
91f93b94 |
455 | info_rdaddsc(&currnode->text, ", Up: "); |
456 | info_rdaddsc(&currnode->text, (currnode->up ? |
457 | currnode->up->name : "(dir)")); |
5dd44dce |
458 | if (currnode->next) { |
91f93b94 |
459 | info_rdaddsc(&currnode->text, ", Next: "); |
460 | info_rdaddsc(&currnode->text, currnode->next->name); |
5dd44dce |
461 | } |
91f93b94 |
462 | info_rdaddsc(&currnode->text, "\n\n"); |
463 | info_rdaddsc(&currnode->text, origtext); |
5dd44dce |
464 | /* |
465 | * Just make _absolutely_ sure we end with a newline. |
466 | */ |
91f93b94 |
467 | if (currnode->text.output.text[currnode->text.output.pos-1] != '\n') |
468 | info_rdaddc(&currnode->text, '\n'); |
5dd44dce |
469 | |
470 | sfree(origtext); |
471 | } |
472 | |
473 | /* |
474 | * Compute the offsets for the tag table. |
475 | */ |
91f93b94 |
476 | filepos = intro_text.output.pos; |
5dd44dce |
477 | for (currnode = topnode; currnode; currnode = currnode->listnext) { |
478 | currnode->pos = filepos; |
91f93b94 |
479 | filepos += currnode->text.output.pos; |
5dd44dce |
480 | } |
481 | |
482 | /* |
483 | * Split into sub-files. |
484 | */ |
485 | if (conf.maxfilesize > 0) { |
91f93b94 |
486 | int currfilesize = intro_text.output.pos, currfilenum = 1; |
5dd44dce |
487 | for (currnode = topnode; currnode; currnode = currnode->listnext) { |
91f93b94 |
488 | if (currfilesize > intro_text.output.pos && |
489 | currfilesize + currnode->text.output.pos > conf.maxfilesize) { |
5dd44dce |
490 | currfilenum++; |
91f93b94 |
491 | currfilesize = intro_text.output.pos; |
5dd44dce |
492 | } |
493 | currnode->filenum = currfilenum; |
91f93b94 |
494 | currfilesize += currnode->text.output.pos; |
5dd44dce |
495 | } |
496 | } |
497 | |
498 | /* |
499 | * Write the primary output file. |
500 | */ |
501 | fp = fopen(conf.filename, "w"); |
502 | if (!fp) { |
503 | error(err_cantopenw, conf.filename); |
504 | return; |
505 | } |
91f93b94 |
506 | fputs(intro_text.output.text, fp); |
5dd44dce |
507 | if (conf.maxfilesize == 0) { |
508 | for (currnode = topnode; currnode; currnode = currnode->listnext) |
91f93b94 |
509 | fputs(currnode->text.output.text, fp); |
5dd44dce |
510 | } else { |
511 | int filenum = 0; |
512 | fprintf(fp, "\037\nIndirect:\n"); |
513 | for (currnode = topnode; currnode; currnode = currnode->listnext) |
514 | if (filenum != currnode->filenum) { |
515 | filenum = currnode->filenum; |
516 | fprintf(fp, "%s-%d: %d\n", conf.filename, filenum, |
517 | currnode->pos); |
518 | } |
519 | } |
520 | fprintf(fp, "\037\nTag Table:\n"); |
521 | if (conf.maxfilesize > 0) |
522 | fprintf(fp, "(Indirect)\n"); |
523 | for (currnode = topnode; currnode; currnode = currnode->listnext) |
524 | fprintf(fp, "Node: %s\177%d\n", currnode->name, currnode->pos); |
525 | fprintf(fp, "\037\nEnd Tag Table\n"); |
526 | fclose(fp); |
527 | |
528 | /* |
529 | * Write the subfiles. |
530 | */ |
531 | if (conf.maxfilesize > 0) { |
532 | int filenum = 0; |
533 | fp = NULL; |
534 | |
535 | for (currnode = topnode; currnode; currnode = currnode->listnext) { |
536 | if (filenum != currnode->filenum) { |
537 | char *fname; |
538 | |
539 | filenum = currnode->filenum; |
540 | |
541 | if (fp) |
542 | fclose(fp); |
543 | fname = mknewa(char, strlen(conf.filename) + 40); |
544 | sprintf(fname, "%s-%d", conf.filename, filenum); |
545 | fp = fopen(fname, "w"); |
546 | if (!fp) { |
547 | error(err_cantopenw, fname); |
548 | return; |
549 | } |
550 | sfree(fname); |
91f93b94 |
551 | fputs(intro_text.output.text, fp); |
5dd44dce |
552 | } |
91f93b94 |
553 | fputs(currnode->text.output.text, fp); |
5dd44dce |
554 | } |
555 | |
556 | if (fp) |
557 | fclose(fp); |
558 | } |
559 | } |
560 | |
561 | static int info_check_index(word *w, node *n, indexdata *idx) |
562 | { |
563 | int ret = 0; |
564 | |
565 | for (; w; w = w->next) { |
566 | if (w->type == word_IndexRef) { |
567 | indextag *tag; |
568 | int i; |
569 | |
570 | tag = index_findtag(idx, w->text); |
571 | if (!tag) |
572 | break; |
573 | |
574 | for (i = 0; i < tag->nrefs; i++) { |
575 | indexentry *entry = tag->refs[i]; |
576 | info_idx *ii = (info_idx *)entry->backend_data; |
577 | |
578 | if (ii->nnodes > 0 && ii->nodes[ii->nnodes-1] == n) { |
579 | /* |
580 | * If the same index term is indexed twice |
581 | * within the same section, we only want to |
582 | * mention it once in the index. So do nothing |
583 | * here. |
584 | */ |
585 | continue; |
586 | } |
587 | |
588 | if (ii->nnodes >= ii->nodesize) { |
589 | ii->nodesize += 32; |
590 | ii->nodes = resize(ii->nodes, ii->nodesize); |
591 | } |
592 | |
593 | ii->nodes[ii->nnodes++] = n; |
594 | |
595 | ret = 1; |
596 | } |
597 | } |
598 | } |
599 | |
600 | return ret; |
601 | } |
602 | |
5dd44dce |
603 | static word *info_transform_wordlist(word *words, keywordlist *keywords) |
604 | { |
605 | word *ret = dup_word_list(words); |
606 | word *w; |
607 | keyword *kwl; |
608 | |
609 | for (w = ret; w; w = w->next) { |
610 | w->private_data = NULL; |
611 | if (w->type == word_UpperXref || w->type == word_LowerXref) { |
612 | kwl = kw_lookup(keywords, w->text); |
613 | if (kwl) { |
614 | if (kwl->para->type == para_NumberedList || |
615 | kwl->para->type == para_BiblioCited) { |
616 | /* |
617 | * In Info, we do nothing special for xrefs to |
618 | * numbered list items or bibliography entries. |
619 | */ |
620 | break; |
621 | } else { |
622 | /* |
623 | * An xref to a different section has its text |
624 | * completely replaced. |
625 | */ |
626 | word *w2, *w3, *w4; |
627 | w2 = w3 = w->next; |
628 | w4 = NULL; |
629 | while (w2) { |
630 | if (w2->type == word_XrefEnd) { |
631 | w4 = w2->next; |
632 | w2->next = NULL; |
633 | break; |
634 | } |
635 | w2 = w2->next; |
636 | } |
637 | free_word_list(w3); |
638 | |
639 | /* |
640 | * Now w is the UpperXref / LowerXref we |
641 | * started with, and w4 is the next word after |
642 | * the corresponding XrefEnd (if any). The |
643 | * simplest thing is just to stick a pointer to |
644 | * the target node structure in the private |
645 | * data field of the xref word, and let |
646 | * info_rdaddwc and friends read the node name |
647 | * out from there. |
648 | */ |
649 | w->next = w4; |
650 | w->private_data = kwl->para->private_data; |
651 | assert(w->private_data); |
652 | } |
653 | } |
654 | } |
655 | } |
656 | |
657 | return ret; |
658 | } |
659 | |
91f93b94 |
660 | static int info_rdaddwc(info_data *id, word *words, word *end, int xrefs) { |
661 | int ret = 0; |
5dd44dce |
662 | |
663 | for (; words && words != end; words = words->next) switch (words->type) { |
664 | case word_HyperLink: |
665 | case word_HyperEnd: |
666 | case word_XrefEnd: |
667 | case word_IndexRef: |
668 | break; |
669 | |
670 | case word_Normal: |
671 | case word_Emph: |
672 | case word_Code: |
673 | case word_WeakCode: |
674 | case word_WhiteSpace: |
675 | case word_EmphSpace: |
676 | case word_CodeSpace: |
677 | case word_WkCodeSpace: |
678 | case word_Quote: |
679 | case word_EmphQuote: |
680 | case word_CodeQuote: |
681 | case word_WkCodeQuote: |
682 | assert(words->type != word_CodeQuote && |
683 | words->type != word_WkCodeQuote); |
684 | if (towordstyle(words->type) == word_Emph && |
685 | (attraux(words->aux) == attr_First || |
686 | attraux(words->aux) == attr_Only)) |
91f93b94 |
687 | ret += info_rdadd(id, L'_'); /* FIXME: configurability */ |
5dd44dce |
688 | else if (towordstyle(words->type) == word_Code && |
689 | (attraux(words->aux) == attr_First || |
690 | attraux(words->aux) == attr_Only)) |
91f93b94 |
691 | ret += info_rdadd(id, L'`'); /* FIXME: configurability */ |
5dd44dce |
692 | if (removeattr(words->type) == word_Normal) { |
91f93b94 |
693 | if (cvt_ok(id->charset, words->text) || !words->alt) |
694 | ret += info_rdadds(id, words->text); |
5dd44dce |
695 | else |
91f93b94 |
696 | ret += info_rdaddwc(id, words->alt, NULL, FALSE); |
5dd44dce |
697 | } else if (removeattr(words->type) == word_WhiteSpace) { |
91f93b94 |
698 | ret += info_rdadd(id, L' '); |
5dd44dce |
699 | } else if (removeattr(words->type) == word_Quote) { |
91f93b94 |
700 | ret += info_rdadd(id, quoteaux(words->aux) == quote_Open ? L'`' : L'\''); |
5dd44dce |
701 | /* FIXME: configurability */ |
702 | } |
703 | if (towordstyle(words->type) == word_Emph && |
704 | (attraux(words->aux) == attr_Last || |
705 | attraux(words->aux) == attr_Only)) |
91f93b94 |
706 | ret += info_rdadd(id, L'_'); /* FIXME: configurability */ |
5dd44dce |
707 | else if (towordstyle(words->type) == word_Code && |
708 | (attraux(words->aux) == attr_Last || |
709 | attraux(words->aux) == attr_Only)) |
91f93b94 |
710 | ret += info_rdadd(id, L'\''); /* FIXME: configurability */ |
5dd44dce |
711 | break; |
712 | |
713 | case word_UpperXref: |
714 | case word_LowerXref: |
715 | if (xrefs && words->private_data) { |
91f93b94 |
716 | /* |
717 | * This bit is structural and so must be done in char |
718 | * rather than wchar_t. |
719 | */ |
720 | ret += info_rdaddsc(id, "*Note "); |
721 | ret += info_rdaddsc(id, ((node *)words->private_data)->name); |
722 | ret += info_rdaddsc(id, "::"); |
5dd44dce |
723 | } |
724 | break; |
725 | } |
91f93b94 |
726 | |
727 | return ret; |
5dd44dce |
728 | } |
729 | |
91f93b94 |
730 | static int info_width_internal(word *words, int xrefs, int charset); |
5dd44dce |
731 | |
91f93b94 |
732 | static int info_width_internal_list(word *words, int xrefs, int charset) { |
5dd44dce |
733 | int w = 0; |
734 | while (words) { |
91f93b94 |
735 | w += info_width_internal(words, xrefs, charset); |
5dd44dce |
736 | words = words->next; |
737 | } |
738 | return w; |
739 | } |
740 | |
91f93b94 |
741 | static int info_width_internal(word *words, int xrefs, int charset) { |
5dd44dce |
742 | switch (words->type) { |
743 | case word_HyperLink: |
744 | case word_HyperEnd: |
745 | case word_XrefEnd: |
746 | case word_IndexRef: |
747 | return 0; |
748 | |
749 | case word_Normal: |
750 | case word_Emph: |
751 | case word_Code: |
752 | case word_WeakCode: |
753 | return (((words->type == word_Emph || |
754 | words->type == word_Code) |
755 | ? (attraux(words->aux) == attr_Only ? 2 : |
756 | attraux(words->aux) == attr_Always ? 0 : 1) |
757 | : 0) + |
91f93b94 |
758 | (cvt_ok(charset, words->text) || !words->alt ? |
e5cd393f |
759 | ustrwid(words->text, charset) : |
91f93b94 |
760 | info_width_internal_list(words->alt, xrefs, charset))); |
5dd44dce |
761 | |
762 | case word_WhiteSpace: |
763 | case word_EmphSpace: |
764 | case word_CodeSpace: |
765 | case word_WkCodeSpace: |
766 | case word_Quote: |
767 | case word_EmphQuote: |
768 | case word_CodeQuote: |
769 | case word_WkCodeQuote: |
770 | assert(words->type != word_CodeQuote && |
771 | words->type != word_WkCodeQuote); |
772 | return (((towordstyle(words->type) == word_Emph || |
773 | towordstyle(words->type) == word_Code) |
774 | ? (attraux(words->aux) == attr_Only ? 2 : |
775 | attraux(words->aux) == attr_Always ? 0 : 1) |
776 | : 0) + 1); |
777 | |
778 | case word_UpperXref: |
779 | case word_LowerXref: |
780 | if (xrefs && words->private_data) { |
781 | /* "*Note " plus "::" comes to 8 characters */ |
e5cd393f |
782 | return 8 + strwid(((node *)words->private_data)->name, charset); |
5dd44dce |
783 | } |
784 | break; |
785 | } |
786 | return 0; /* should never happen */ |
787 | } |
788 | |
43341922 |
789 | static int info_width_noxrefs(void *ctx, word *words) |
5dd44dce |
790 | { |
91f93b94 |
791 | return info_width_internal(words, FALSE, *(int *)ctx); |
5dd44dce |
792 | } |
43341922 |
793 | static int info_width_xrefs(void *ctx, word *words) |
5dd44dce |
794 | { |
91f93b94 |
795 | return info_width_internal(words, TRUE, *(int *)ctx); |
5dd44dce |
796 | } |
797 | |
91f93b94 |
798 | static void info_heading(info_data *text, word *tprefix, |
5dd44dce |
799 | word *words, int width) { |
91f93b94 |
800 | int length; |
5dd44dce |
801 | int firstlinewidth, wrapwidth; |
5dd44dce |
802 | wrappedline *wrapping, *p; |
803 | |
91f93b94 |
804 | length = 0; |
5dd44dce |
805 | if (tprefix) { |
91f93b94 |
806 | length += info_rdaddwc(text, tprefix, NULL, FALSE); |
807 | length += info_rdadds(text, L": ");/* FIXME: configurability */ |
5dd44dce |
808 | } |
5dd44dce |
809 | |
5dd44dce |
810 | wrapwidth = width; |
91f93b94 |
811 | firstlinewidth = width - length; |
5dd44dce |
812 | |
43341922 |
813 | wrapping = wrap_para(words, firstlinewidth, wrapwidth, |
91f93b94 |
814 | info_width_noxrefs, &text->charset, 0); |
5dd44dce |
815 | for (p = wrapping; p; p = p->next) { |
91f93b94 |
816 | length += info_rdaddwc(text, p->begin, p->end, FALSE); |
817 | info_rdadd(text, L'\n'); |
5dd44dce |
818 | while (length--) |
91f93b94 |
819 | info_rdadd(text, L'-'); /* FIXME: configurability */ |
820 | info_rdadd(text, L'\n'); |
821 | length = 0; |
5dd44dce |
822 | } |
823 | wrap_free(wrapping); |
91f93b94 |
824 | info_rdadd(text, L'\n'); |
5dd44dce |
825 | } |
826 | |
91f93b94 |
827 | static void info_rule(info_data *text, int indent, int width) { |
828 | while (indent--) info_rdadd(text, L' '); |
829 | while (width--) info_rdadd(text, L'-'); |
830 | info_rdadd(text, L'\n'); |
831 | info_rdadd(text, L'\n'); |
5dd44dce |
832 | } |
833 | |
91f93b94 |
834 | static void info_para(info_data *text, word *prefix, wchar_t *prefixextra, |
5dd44dce |
835 | word *input, keywordlist *keywords, |
836 | int indent, int extraindent, int width) { |
837 | wrappedline *wrapping, *p; |
838 | word *words; |
5dd44dce |
839 | int e; |
840 | int i; |
841 | int firstlinewidth = width; |
842 | |
843 | words = info_transform_wordlist(input, keywords); |
844 | |
845 | if (prefix) { |
5dd44dce |
846 | for (i = 0; i < indent; i++) |
91f93b94 |
847 | info_rdadd(text, L' '); |
848 | e = info_rdaddwc(text, prefix, NULL, FALSE); |
849 | if (prefixextra) |
850 | e += info_rdadds(text, prefixextra); |
5dd44dce |
851 | /* If the prefix is too long, shorten the first line to fit. */ |
91f93b94 |
852 | e = extraindent - e; |
5dd44dce |
853 | if (e < 0) { |
854 | firstlinewidth += e; /* this decreases it, since e < 0 */ |
855 | if (firstlinewidth < 0) { |
856 | e = indent + extraindent; |
857 | firstlinewidth = width; |
91f93b94 |
858 | info_rdadd(text, L'\n'); |
5dd44dce |
859 | } else |
860 | e = 0; |
861 | } |
5dd44dce |
862 | } else |
863 | e = indent + extraindent; |
864 | |
43341922 |
865 | wrapping = wrap_para(words, firstlinewidth, width, info_width_xrefs, |
91f93b94 |
866 | &text->charset, 0); |
5dd44dce |
867 | for (p = wrapping; p; p = p->next) { |
868 | for (i = 0; i < e; i++) |
91f93b94 |
869 | info_rdadd(text, L' '); |
5dd44dce |
870 | info_rdaddwc(text, p->begin, p->end, TRUE); |
91f93b94 |
871 | info_rdadd(text, L'\n'); |
5dd44dce |
872 | e = indent + extraindent; |
873 | } |
874 | wrap_free(wrapping); |
91f93b94 |
875 | info_rdadd(text, L'\n'); |
5dd44dce |
876 | |
877 | free_word_list(words); |
878 | } |
879 | |
91f93b94 |
880 | static void info_codepara(info_data *text, word *words, |
5dd44dce |
881 | int indent, int width) { |
882 | int i; |
883 | |
884 | for (; words; words = words->next) if (words->type == word_WeakCode) { |
91f93b94 |
885 | for (i = 0; i < indent; i++) |
886 | info_rdadd(text, L' '); |
887 | if (info_rdadds(text, words->text) > width) { |
5dd44dce |
888 | /* FIXME: warn */ |
889 | } |
91f93b94 |
890 | info_rdadd(text, L'\n'); |
5dd44dce |
891 | } |
892 | |
91f93b94 |
893 | info_rdadd(text, L'\n'); |
5dd44dce |
894 | } |
895 | |
91f93b94 |
896 | static void info_versionid(info_data *text, word *words) { |
897 | info_rdadd(text, L'['); /* FIXME: configurability */ |
5dd44dce |
898 | info_rdaddwc(text, words, NULL, FALSE); |
91f93b94 |
899 | info_rdadds(text, L"]\n"); |
5dd44dce |
900 | } |
901 | |
91f93b94 |
902 | static node *info_node_new(char *name, int charset) |
5dd44dce |
903 | { |
904 | node *n; |
905 | |
906 | n = mknew(node); |
91f93b94 |
907 | n->text = empty_info_data; |
908 | n->text.charset = charset; |
5dd44dce |
909 | n->up = n->next = n->prev = n->lastchild = n->listnext = NULL; |
910 | n->name = dupstr(name); |
911 | n->started_menu = FALSE; |
912 | |
913 | return n; |
914 | } |
915 | |
91f93b94 |
916 | static char *info_node_name(paragraph *par, int charset) |
5dd44dce |
917 | { |
91f93b94 |
918 | info_data id = EMPTY_INFO_DATA; |
f4551933 |
919 | char *p, *q; |
91f93b94 |
920 | |
921 | id.charset = charset; |
922 | info_rdaddwc(&id, par->kwtext ? par->kwtext : par->words, NULL, FALSE); |
923 | info_rdaddsc(&id, NULL); |
f4551933 |
924 | |
925 | /* |
926 | * We cannot have commas or colons in a node name. Remove any |
927 | * that we find, with a warning. |
928 | */ |
91f93b94 |
929 | p = q = id.output.text; |
f4551933 |
930 | while (*p) { |
931 | if (*p == ':' || *p == ',') { |
932 | error(err_infonodechar, &par->fpos, *p); |
933 | } else { |
934 | *q++ = *p; |
935 | } |
936 | p++; |
937 | } |
938 | *p = '\0'; |
939 | |
91f93b94 |
940 | return id.output.text; |
5dd44dce |
941 | } |
942 | |
91f93b94 |
943 | static void info_menu_item(info_data *text, node *n, paragraph *p) |
5dd44dce |
944 | { |
945 | /* |
946 | * FIXME: Depending on how we're doing node names in this info |
947 | * file, we might want to do |
948 | * |
949 | * * Node name:: Chapter title |
950 | * |
951 | * _or_ |
952 | * |
953 | * * Chapter number: Node name. |
954 | * |
91f93b94 |
955 | * This function mostly works in char rather than wchar_t, |
956 | * because a menu item is a structural component. |
5dd44dce |
957 | */ |
91f93b94 |
958 | info_rdaddsc(text, "* "); |
959 | info_rdaddsc(text, n->name); |
960 | info_rdaddsc(text, "::"); |
5dd44dce |
961 | if (p) { |
91f93b94 |
962 | info_rdaddc(text, ' '); |
5dd44dce |
963 | info_rdaddwc(text, p->words, NULL, FALSE); |
964 | } |
91f93b94 |
965 | info_rdaddc(text, '\n'); |
966 | } |
967 | |
968 | /* |
969 | * These functions implement my wrapper on the rdadd* calls which |
970 | * allows me to switch arbitrarily between literal octet-string |
971 | * text and charset-translated Unicode. (Because no matter what |
972 | * character set I write the actual text in, I expect info readers |
973 | * to treat node names and file names literally and to expect |
974 | * keywords like `*Note' in their canonical form, so I have to take |
975 | * steps to ensure that those structural elements of the file |
976 | * aren't messed with.) |
977 | */ |
978 | static int info_rdadds(info_data *d, wchar_t const *wcs) |
979 | { |
980 | if (!d->wcmode) { |
981 | d->state = charset_init_state; |
982 | d->wcmode = TRUE; |
983 | } |
984 | |
985 | if (wcs) { |
986 | char buf[256]; |
e5cd393f |
987 | int len, width, ret; |
988 | |
989 | width = ustrwid(wcs, d->charset); |
91f93b94 |
990 | |
e5cd393f |
991 | len = ustrlen(wcs); |
91f93b94 |
992 | while (len > 0) { |
993 | int prevlen = len; |
994 | |
995 | ret = charset_from_unicode(&wcs, &len, buf, lenof(buf), |
996 | d->charset, &d->state, NULL); |
997 | |
998 | assert(len < prevlen); |
999 | |
1000 | if (ret > 0) { |
1001 | buf[ret] = '\0'; |
1002 | rdaddsc(&d->output, buf); |
1003 | } |
1004 | } |
1005 | |
e5cd393f |
1006 | return width; |
91f93b94 |
1007 | } else |
1008 | return 0; |
1009 | } |
1010 | |
1011 | static int info_rdaddsc(info_data *d, char const *cs) |
1012 | { |
1013 | if (d->wcmode) { |
1014 | char buf[256]; |
1015 | int ret; |
1016 | |
1017 | ret = charset_from_unicode(NULL, 0, buf, lenof(buf), |
1018 | d->charset, &d->state, NULL); |
1019 | if (ret > 0) { |
1020 | buf[ret] = '\0'; |
1021 | rdaddsc(&d->output, buf); |
1022 | } |
1023 | |
1024 | d->wcmode = FALSE; |
1025 | } |
1026 | |
1027 | if (cs) { |
1028 | rdaddsc(&d->output, cs); |
1029 | return strlen(cs); |
1030 | } else |
1031 | return 0; |
1032 | } |
1033 | |
1034 | static int info_rdadd(info_data *d, wchar_t wc) |
1035 | { |
1036 | wchar_t wcs[2]; |
1037 | wcs[0] = wc; |
1038 | wcs[1] = L'\0'; |
1039 | return info_rdadds(d, wcs); |
1040 | } |
1041 | |
1042 | static int info_rdaddc(info_data *d, char c) |
1043 | { |
1044 | char cs[2]; |
1045 | cs[0] = c; |
1046 | cs[1] = '\0'; |
1047 | return info_rdaddsc(d, cs); |
5dd44dce |
1048 | } |