rdiscountwl 1.0.0.1
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/BUILDING +151 -0
- data/COPYING +33 -0
- data/README.markdown +73 -0
- data/Rakefile +224 -0
- data/bin/rdiscount +13 -0
- data/ext/Csio.c +61 -0
- data/ext/VERSION +1 -0
- data/ext/amalloc.c +135 -0
- data/ext/amalloc.h +29 -0
- data/ext/basename.c +43 -0
- data/ext/blocktags +33 -0
- data/ext/config.h +25 -0
- data/ext/css.c +87 -0
- data/ext/cstring.h +77 -0
- data/ext/docheader.c +49 -0
- data/ext/dumptree.c +151 -0
- data/ext/emmatch.c +188 -0
- data/ext/extconf.rb +49 -0
- data/ext/flags.c +91 -0
- data/ext/generate.c +1865 -0
- data/ext/github_flavoured.c +100 -0
- data/ext/html5.c +22 -0
- data/ext/markdown.c +1361 -0
- data/ext/markdown.h +238 -0
- data/ext/mkdio.c +360 -0
- data/ext/mkdio.h +122 -0
- data/ext/mktags.c +89 -0
- data/ext/pgm_options.c +146 -0
- data/ext/pgm_options.h +9 -0
- data/ext/rdiscount.c +150 -0
- data/ext/resource.c +159 -0
- data/ext/setup.c +39 -0
- data/ext/tags.c +94 -0
- data/ext/tags.h +19 -0
- data/ext/toc.c +114 -0
- data/ext/version.c +13 -0
- data/ext/xml.c +82 -0
- data/ext/xmlpage.c +46 -0
- data/lib/markdown.rb +1 -0
- data/lib/rdiscount.rb +108 -0
- data/man/markdown.7 +1020 -0
- data/man/rdiscount.1 +22 -0
- data/man/rdiscount.1.ronn +24 -0
- data/rdiscount.gemspec +71 -0
- data/test/benchmark.rb +56 -0
- data/test/benchmark.txt +306 -0
- data/test/markdown_test.rb +158 -0
- data/test/rdiscount_test.rb +276 -0
- metadata +99 -0
data/ext/markdown.c
ADDED
@@ -0,0 +1,1361 @@
|
|
1
|
+
/* markdown: a C implementation of John Gruber's Markdown markup language.
|
2
|
+
*
|
3
|
+
* Copyright (C) 2007 David L Parsons.
|
4
|
+
* The redistribution terms are provided in the COPYRIGHT file that must
|
5
|
+
* be distributed with this source code.
|
6
|
+
*/
|
7
|
+
#include "config.h"
|
8
|
+
|
9
|
+
#include <stdio.h>
|
10
|
+
#include <string.h>
|
11
|
+
#include <stdarg.h>
|
12
|
+
#include <stdlib.h>
|
13
|
+
#include <time.h>
|
14
|
+
#include <ctype.h>
|
15
|
+
|
16
|
+
#include "cstring.h"
|
17
|
+
#include "markdown.h"
|
18
|
+
#include "amalloc.h"
|
19
|
+
#include "tags.h"
|
20
|
+
|
21
|
+
typedef int (*stfu)(const void*,const void*);
|
22
|
+
|
23
|
+
typedef ANCHOR(Paragraph) ParagraphRoot;
|
24
|
+
|
25
|
+
static Paragraph *Pp(ParagraphRoot *, Line *, int);
|
26
|
+
static Paragraph *compile(Line *, int, MMIOT *);
|
27
|
+
|
28
|
+
/* case insensitive string sort for Footnote tags.
|
29
|
+
*/
|
30
|
+
int
|
31
|
+
__mkd_footsort(Footnote *a, Footnote *b)
|
32
|
+
{
|
33
|
+
int i;
|
34
|
+
char ac, bc;
|
35
|
+
|
36
|
+
if ( S(a->tag) != S(b->tag) )
|
37
|
+
return S(a->tag) - S(b->tag);
|
38
|
+
|
39
|
+
for ( i=0; i < S(a->tag); i++) {
|
40
|
+
ac = tolower(T(a->tag)[i]);
|
41
|
+
bc = tolower(T(b->tag)[i]);
|
42
|
+
|
43
|
+
if ( isspace(ac) && isspace(bc) )
|
44
|
+
continue;
|
45
|
+
if ( ac != bc )
|
46
|
+
return ac - bc;
|
47
|
+
}
|
48
|
+
return 0;
|
49
|
+
}
|
50
|
+
|
51
|
+
|
52
|
+
/* find the first blank character after position <i>
|
53
|
+
*/
|
54
|
+
static int
|
55
|
+
nextblank(Line *t, int i)
|
56
|
+
{
|
57
|
+
while ( (i < S(t->text)) && !isspace(T(t->text)[i]) )
|
58
|
+
++i;
|
59
|
+
return i;
|
60
|
+
}
|
61
|
+
|
62
|
+
|
63
|
+
/* find the next nonblank character after position <i>
|
64
|
+
*/
|
65
|
+
static int
|
66
|
+
nextnonblank(Line *t, int i)
|
67
|
+
{
|
68
|
+
while ( (i < S(t->text)) && isspace(T(t->text)[i]) )
|
69
|
+
++i;
|
70
|
+
return i;
|
71
|
+
}
|
72
|
+
|
73
|
+
|
74
|
+
/* find the first nonblank character on the Line.
|
75
|
+
*/
|
76
|
+
int
|
77
|
+
mkd_firstnonblank(Line *p)
|
78
|
+
{
|
79
|
+
return nextnonblank(p,0);
|
80
|
+
}
|
81
|
+
|
82
|
+
|
83
|
+
static inline int
|
84
|
+
blankline(Line *p)
|
85
|
+
{
|
86
|
+
return ! (p && (S(p->text) > p->dle) );
|
87
|
+
}
|
88
|
+
|
89
|
+
|
90
|
+
static Line *
|
91
|
+
skipempty(Line *p)
|
92
|
+
{
|
93
|
+
while ( p && (p->dle == S(p->text)) )
|
94
|
+
p = p->next;
|
95
|
+
return p;
|
96
|
+
}
|
97
|
+
|
98
|
+
|
99
|
+
void
|
100
|
+
___mkd_tidy(Cstring *t)
|
101
|
+
{
|
102
|
+
while ( S(*t) && isspace(T(*t)[S(*t)-1]) )
|
103
|
+
--S(*t);
|
104
|
+
}
|
105
|
+
|
106
|
+
|
107
|
+
static struct kw comment = { "!--", 3, 0 };
|
108
|
+
|
109
|
+
static struct kw *
|
110
|
+
isopentag(Line *p)
|
111
|
+
{
|
112
|
+
int i=0, len;
|
113
|
+
char *line;
|
114
|
+
|
115
|
+
if ( !p ) return 0;
|
116
|
+
|
117
|
+
line = T(p->text);
|
118
|
+
len = S(p->text);
|
119
|
+
|
120
|
+
if ( len < 3 || line[0] != '<' )
|
121
|
+
return 0;
|
122
|
+
|
123
|
+
if ( line[1] == '!' && line[2] == '-' && line[3] == '-' )
|
124
|
+
/* comments need special case handling, because
|
125
|
+
* the !-- doesn't need to end in a whitespace
|
126
|
+
*/
|
127
|
+
return &comment;
|
128
|
+
|
129
|
+
/* find how long the tag is so we can check to see if
|
130
|
+
* it's a block-level tag
|
131
|
+
*/
|
132
|
+
for ( i=1; i < len && T(p->text)[i] != '>'
|
133
|
+
&& T(p->text)[i] != '/'
|
134
|
+
&& !isspace(T(p->text)[i]); ++i )
|
135
|
+
;
|
136
|
+
|
137
|
+
|
138
|
+
return mkd_search_tags(T(p->text)+1, i-1);
|
139
|
+
}
|
140
|
+
|
141
|
+
|
142
|
+
typedef struct _flo {
|
143
|
+
Line *t;
|
144
|
+
int i;
|
145
|
+
} FLO;
|
146
|
+
|
147
|
+
#define floindex(x) (x.i)
|
148
|
+
|
149
|
+
|
150
|
+
static int
|
151
|
+
flogetc(FLO *f)
|
152
|
+
{
|
153
|
+
if ( f && f->t ) {
|
154
|
+
if ( f->i < S(f->t->text) )
|
155
|
+
return T(f->t->text)[f->i++];
|
156
|
+
f->t = f->t->next;
|
157
|
+
f->i = 0;
|
158
|
+
return flogetc(f);
|
159
|
+
}
|
160
|
+
return EOF;
|
161
|
+
}
|
162
|
+
|
163
|
+
|
164
|
+
static void
|
165
|
+
splitline(Line *t, int cutpoint)
|
166
|
+
{
|
167
|
+
if ( t && (cutpoint < S(t->text)) ) {
|
168
|
+
Line *tmp = calloc(1, sizeof *tmp);
|
169
|
+
|
170
|
+
tmp->next = t->next;
|
171
|
+
t->next = tmp;
|
172
|
+
|
173
|
+
tmp->dle = t->dle;
|
174
|
+
SUFFIX(tmp->text, T(t->text)+cutpoint, S(t->text)-cutpoint);
|
175
|
+
S(t->text) = cutpoint;
|
176
|
+
}
|
177
|
+
}
|
178
|
+
|
179
|
+
#define UNCHECK(l) ((l)->flags &= ~CHECKED)
|
180
|
+
|
181
|
+
#define UNLESS_FENCED(t) if (fenced) { \
|
182
|
+
other = 1; l->count += (c == ' ' ? 0 : -1); \
|
183
|
+
} else { t; }
|
184
|
+
|
185
|
+
/*
|
186
|
+
* walk a line, seeing if it's any of half a dozen interesting regular
|
187
|
+
* types.
|
188
|
+
*/
|
189
|
+
static void
|
190
|
+
checkline(Line *l, DWORD flags)
|
191
|
+
{
|
192
|
+
int eol, i;
|
193
|
+
int dashes = 0, spaces = 0,
|
194
|
+
equals = 0, underscores = 0,
|
195
|
+
stars = 0, tildes = 0, other = 0,
|
196
|
+
backticks = 0, fenced = 0;
|
197
|
+
|
198
|
+
l->flags |= CHECKED;
|
199
|
+
l->kind = chk_text;
|
200
|
+
l->count = 0;
|
201
|
+
|
202
|
+
if (l->dle >= 4) { l->kind=chk_code; return; }
|
203
|
+
|
204
|
+
for ( eol = S(l->text); eol > l->dle && isspace(T(l->text)[eol-1]); --eol )
|
205
|
+
;
|
206
|
+
|
207
|
+
for (i=l->dle; i<eol; i++) {
|
208
|
+
register int c = T(l->text)[i];
|
209
|
+
int is_fence_char = 0;
|
210
|
+
|
211
|
+
if ( c != ' ' ) l->count++;
|
212
|
+
|
213
|
+
switch (c) {
|
214
|
+
case '-': UNLESS_FENCED(dashes = 1); break;
|
215
|
+
case ' ': UNLESS_FENCED(spaces = 1); break;
|
216
|
+
case '=': equals = 1; break;
|
217
|
+
case '_': UNLESS_FENCED(underscores = 1); break;
|
218
|
+
case '*': stars = 1; break;
|
219
|
+
default:
|
220
|
+
if (flags & MKD_FENCEDCODE) {
|
221
|
+
switch (c) {
|
222
|
+
case '~': if (other) return; is_fence_char = 1; tildes = 1; break;
|
223
|
+
case '`': if (other) return; is_fence_char = 1; backticks = 1; break;
|
224
|
+
}
|
225
|
+
if (is_fence_char) {
|
226
|
+
fenced = 1;
|
227
|
+
break;
|
228
|
+
}
|
229
|
+
}
|
230
|
+
other = 1;
|
231
|
+
l->count--;
|
232
|
+
if (!fenced) return;
|
233
|
+
}
|
234
|
+
}
|
235
|
+
|
236
|
+
if ( dashes + equals + underscores + stars + tildes + backticks > 1 )
|
237
|
+
return;
|
238
|
+
|
239
|
+
if ( spaces ) {
|
240
|
+
if ( (underscores || stars || dashes) )
|
241
|
+
l->kind = chk_hr;
|
242
|
+
return;
|
243
|
+
}
|
244
|
+
|
245
|
+
if ( stars || underscores ) { l->kind = chk_hr; }
|
246
|
+
else if ( dashes ) { l->kind = chk_dash; }
|
247
|
+
else if ( equals ) { l->kind = chk_equal; }
|
248
|
+
else if ( tildes ) { l->kind = chk_tilde; }
|
249
|
+
else if ( backticks ) { l->kind = chk_backtick; }
|
250
|
+
}
|
251
|
+
|
252
|
+
|
253
|
+
|
254
|
+
/* markdown only does special handling of comments if the comment end
|
255
|
+
* is at the end of a line
|
256
|
+
*/
|
257
|
+
static Line *
|
258
|
+
commentblock(Paragraph *p, int *unclosed)
|
259
|
+
{
|
260
|
+
Line *t, *ret;
|
261
|
+
char *end;
|
262
|
+
|
263
|
+
for ( t = p->text; t ; t = t->next) {
|
264
|
+
if ( end = strstr(T(t->text), "-->") ) {
|
265
|
+
if ( nextnonblank(t, 3 + (end - T(t->text))) < S(t->text) )
|
266
|
+
continue;
|
267
|
+
/*splitline(t, 3 + (end - T(t->text)) );*/
|
268
|
+
ret = t->next;
|
269
|
+
t->next = 0;
|
270
|
+
return ret;
|
271
|
+
}
|
272
|
+
}
|
273
|
+
|
274
|
+
*unclosed = 1;
|
275
|
+
return t;
|
276
|
+
|
277
|
+
}
|
278
|
+
|
279
|
+
|
280
|
+
static Line *
|
281
|
+
htmlblock(Paragraph *p, struct kw *tag, int *unclosed)
|
282
|
+
{
|
283
|
+
Line *ret;
|
284
|
+
FLO f = { p->text, 0 };
|
285
|
+
int c;
|
286
|
+
int i, closing, depth=0;
|
287
|
+
|
288
|
+
*unclosed = 0;
|
289
|
+
|
290
|
+
if ( tag == &comment )
|
291
|
+
return commentblock(p, unclosed);
|
292
|
+
|
293
|
+
if ( tag->selfclose ) {
|
294
|
+
ret = f.t->next;
|
295
|
+
f.t->next = 0;
|
296
|
+
return ret;
|
297
|
+
}
|
298
|
+
|
299
|
+
while ( (c = flogetc(&f)) != EOF ) {
|
300
|
+
if ( c == '<' ) {
|
301
|
+
/* tag? */
|
302
|
+
c = flogetc(&f);
|
303
|
+
if ( c == '!' ) { /* comment? */
|
304
|
+
if ( flogetc(&f) == '-' && flogetc(&f) == '-' ) {
|
305
|
+
/* yes */
|
306
|
+
while ( (c = flogetc(&f)) != EOF ) {
|
307
|
+
if ( c == '-' && flogetc(&f) == '-'
|
308
|
+
&& flogetc(&f) == '>')
|
309
|
+
/* consumed whole comment */
|
310
|
+
break;
|
311
|
+
}
|
312
|
+
}
|
313
|
+
}
|
314
|
+
else {
|
315
|
+
if ( closing = (c == '/') ) c = flogetc(&f);
|
316
|
+
|
317
|
+
for ( i=0; i < tag->size; i++, c=flogetc(&f) ) {
|
318
|
+
if ( tag->id[i] != toupper(c) )
|
319
|
+
break;
|
320
|
+
}
|
321
|
+
|
322
|
+
if ( (i == tag->size) && !isalnum(c) ) {
|
323
|
+
depth = depth + (closing ? -1 : 1);
|
324
|
+
if ( depth == 0 ) {
|
325
|
+
while ( c != EOF && c != '>' ) {
|
326
|
+
/* consume trailing gunk in close tag */
|
327
|
+
c = flogetc(&f);
|
328
|
+
}
|
329
|
+
if ( c == EOF )
|
330
|
+
break;
|
331
|
+
if ( !f.t )
|
332
|
+
return 0;
|
333
|
+
splitline(f.t, floindex(f));
|
334
|
+
ret = f.t->next;
|
335
|
+
f.t->next = 0;
|
336
|
+
return ret;
|
337
|
+
}
|
338
|
+
}
|
339
|
+
}
|
340
|
+
}
|
341
|
+
}
|
342
|
+
*unclosed = 1;
|
343
|
+
return 0;
|
344
|
+
}
|
345
|
+
|
346
|
+
|
347
|
+
/* footnotes look like ^<whitespace>{0,3}[stuff]: <content>$
|
348
|
+
*/
|
349
|
+
static int
|
350
|
+
isfootnote(Line *t)
|
351
|
+
{
|
352
|
+
int i;
|
353
|
+
|
354
|
+
if ( ( (i = t->dle) > 3) || (T(t->text)[i] != '[') )
|
355
|
+
return 0;
|
356
|
+
|
357
|
+
for ( ++i; i < S(t->text) ; ++i ) {
|
358
|
+
if ( T(t->text)[i] == '[' )
|
359
|
+
return 0;
|
360
|
+
else if ( T(t->text)[i] == ']' )
|
361
|
+
return ( T(t->text)[i+1] == ':' ) ;
|
362
|
+
}
|
363
|
+
return 0;
|
364
|
+
}
|
365
|
+
|
366
|
+
|
367
|
+
static inline int
|
368
|
+
isquote(Line *t)
|
369
|
+
{
|
370
|
+
return (t->dle < 4 && T(t->text)[t->dle] == '>');
|
371
|
+
}
|
372
|
+
|
373
|
+
|
374
|
+
static inline int
|
375
|
+
iscode(Line *t)
|
376
|
+
{
|
377
|
+
return (t->dle >= 4);
|
378
|
+
}
|
379
|
+
|
380
|
+
|
381
|
+
static inline int
|
382
|
+
ishr(Line *t, DWORD flags)
|
383
|
+
{
|
384
|
+
if ( ! (t->flags & CHECKED) )
|
385
|
+
checkline(t, flags);
|
386
|
+
|
387
|
+
if ( t->count > 2 )
|
388
|
+
return t->kind == chk_hr || t->kind == chk_dash || t->kind == chk_equal;
|
389
|
+
return 0;
|
390
|
+
}
|
391
|
+
|
392
|
+
|
393
|
+
static int
|
394
|
+
issetext(Line *t, int *htyp, DWORD flags)
|
395
|
+
{
|
396
|
+
Line *n;
|
397
|
+
|
398
|
+
/* check for setext-style HEADER
|
399
|
+
* ======
|
400
|
+
*/
|
401
|
+
|
402
|
+
if ( (n = t->next) ) {
|
403
|
+
if ( !(n->flags & CHECKED) )
|
404
|
+
checkline(n, flags);
|
405
|
+
|
406
|
+
if ( n->kind == chk_dash || n->kind == chk_equal ) {
|
407
|
+
*htyp = SETEXT;
|
408
|
+
return 1;
|
409
|
+
}
|
410
|
+
}
|
411
|
+
return 0;
|
412
|
+
}
|
413
|
+
|
414
|
+
|
415
|
+
static int
|
416
|
+
ishdr(Line *t, int *htyp, DWORD flags)
|
417
|
+
{
|
418
|
+
/* ANY leading `#`'s make this into an ETX header
|
419
|
+
*/
|
420
|
+
if ( (t->dle == 0) && (S(t->text) > 1) && (T(t->text)[0] == '#') ) {
|
421
|
+
*htyp = ETX;
|
422
|
+
return 1;
|
423
|
+
}
|
424
|
+
|
425
|
+
/* And if not, maybe it's a SETEXT header instead
|
426
|
+
*/
|
427
|
+
return issetext(t, htyp, flags);
|
428
|
+
}
|
429
|
+
|
430
|
+
|
431
|
+
static inline int
|
432
|
+
end_of_block(Line *t, DWORD flags)
|
433
|
+
{
|
434
|
+
int dummy;
|
435
|
+
|
436
|
+
if ( !t )
|
437
|
+
return 0;
|
438
|
+
|
439
|
+
return ( (S(t->text) <= t->dle) || ishr(t, flags) || ishdr(t, &dummy, flags) );
|
440
|
+
}
|
441
|
+
|
442
|
+
|
443
|
+
static Line*
|
444
|
+
is_discount_dt(Line *t, int *clip, DWORD flags)
|
445
|
+
{
|
446
|
+
if ( !(flags & MKD_NODLDISCOUNT)
|
447
|
+
&& t
|
448
|
+
&& t->next
|
449
|
+
&& (S(t->text) > 2)
|
450
|
+
&& (t->dle == 0)
|
451
|
+
&& (T(t->text)[0] == '=')
|
452
|
+
&& (T(t->text)[S(t->text)-1] == '=') ) {
|
453
|
+
if ( t->next->dle >= 4 ) {
|
454
|
+
*clip = 4;
|
455
|
+
return t;
|
456
|
+
}
|
457
|
+
else
|
458
|
+
return is_discount_dt(t->next, clip, flags);
|
459
|
+
}
|
460
|
+
return 0;
|
461
|
+
}
|
462
|
+
|
463
|
+
|
464
|
+
static int
|
465
|
+
is_extra_dd(Line *t)
|
466
|
+
{
|
467
|
+
return (t->dle < 4) && (T(t->text)[t->dle] == ':')
|
468
|
+
&& isspace(T(t->text)[t->dle+1]);
|
469
|
+
}
|
470
|
+
|
471
|
+
|
472
|
+
static Line*
|
473
|
+
is_extra_dt(Line *t, int *clip, DWORD flags)
|
474
|
+
{
|
475
|
+
if ( flags & MKD_DLEXTRA
|
476
|
+
&& t
|
477
|
+
&& t->next && S(t->text) && T(t->text)[0] != '='
|
478
|
+
&& T(t->text)[S(t->text)-1] != '=') {
|
479
|
+
Line *x;
|
480
|
+
|
481
|
+
if ( iscode(t) || end_of_block(t, flags) )
|
482
|
+
return 0;
|
483
|
+
|
484
|
+
if ( (x = skipempty(t->next)) && is_extra_dd(x) ) {
|
485
|
+
*clip = x->dle+2;
|
486
|
+
return t;
|
487
|
+
}
|
488
|
+
|
489
|
+
if ( x=is_extra_dt(t->next, clip, flags) )
|
490
|
+
return x;
|
491
|
+
}
|
492
|
+
return 0;
|
493
|
+
}
|
494
|
+
|
495
|
+
|
496
|
+
static Line*
|
497
|
+
isdefinition(Line *t, int *clip, int *kind, DWORD flags)
|
498
|
+
{
|
499
|
+
Line *ret;
|
500
|
+
|
501
|
+
*kind = 1;
|
502
|
+
if ( ret = is_discount_dt(t,clip,flags) )
|
503
|
+
return ret;
|
504
|
+
|
505
|
+
*kind=2;
|
506
|
+
return is_extra_dt(t,clip,flags);
|
507
|
+
}
|
508
|
+
|
509
|
+
|
510
|
+
static int
|
511
|
+
islist(Line *t, int *clip, DWORD flags, int *list_type)
|
512
|
+
{
|
513
|
+
int i, j;
|
514
|
+
char *q;
|
515
|
+
|
516
|
+
if ( end_of_block(t, flags) )
|
517
|
+
return 0;
|
518
|
+
|
519
|
+
if ( !(flags & (MKD_NODLIST|MKD_STRICT)) && isdefinition(t,clip,list_type,flags) )
|
520
|
+
return DL;
|
521
|
+
|
522
|
+
if ( strchr("*-+", T(t->text)[t->dle]) && isspace(T(t->text)[t->dle+1]) ) {
|
523
|
+
i = nextnonblank(t, t->dle+1);
|
524
|
+
*clip = (i > 4) ? 4 : i;
|
525
|
+
*list_type = UL;
|
526
|
+
return AL;
|
527
|
+
}
|
528
|
+
|
529
|
+
if ( (j = nextblank(t,t->dle)) > t->dle ) {
|
530
|
+
if ( T(t->text)[j-1] == '.' ) {
|
531
|
+
|
532
|
+
if ( !(flags & (MKD_NOALPHALIST|MKD_STRICT))
|
533
|
+
&& (j == t->dle + 2)
|
534
|
+
&& isalpha(T(t->text)[t->dle]) ) {
|
535
|
+
j = nextnonblank(t,j);
|
536
|
+
*clip = (j > 4) ? 4 : j;
|
537
|
+
*list_type = AL;
|
538
|
+
return AL;
|
539
|
+
}
|
540
|
+
|
541
|
+
strtoul(T(t->text)+t->dle, &q, 10);
|
542
|
+
if ( (q > T(t->text)+t->dle) && (q == T(t->text) + (j-1)) ) {
|
543
|
+
j = nextnonblank(t,j);
|
544
|
+
*clip = j;
|
545
|
+
*list_type = OL;
|
546
|
+
return AL;
|
547
|
+
}
|
548
|
+
}
|
549
|
+
}
|
550
|
+
return 0;
|
551
|
+
}
|
552
|
+
|
553
|
+
|
554
|
+
static Line *
|
555
|
+
headerblock(Paragraph *pp, int htyp)
|
556
|
+
{
|
557
|
+
Line *ret = 0;
|
558
|
+
Line *p = pp->text;
|
559
|
+
int i, j;
|
560
|
+
|
561
|
+
switch (htyp) {
|
562
|
+
case SETEXT:
|
563
|
+
/* p->text is header, p->next->text is -'s or ='s
|
564
|
+
*/
|
565
|
+
pp->hnumber = (T(p->next->text)[0] == '=') ? 1 : 2;
|
566
|
+
|
567
|
+
ret = p->next->next;
|
568
|
+
___mkd_freeLine(p->next);
|
569
|
+
p->next = 0;
|
570
|
+
break;
|
571
|
+
|
572
|
+
case ETX:
|
573
|
+
/* p->text is ###header###, so we need to trim off
|
574
|
+
* the leading and trailing `#`'s
|
575
|
+
*/
|
576
|
+
|
577
|
+
for (i=0; (T(p->text)[i] == T(p->text)[0]) && (i < S(p->text)-1)
|
578
|
+
&& (i < 6); i++)
|
579
|
+
;
|
580
|
+
|
581
|
+
pp->hnumber = i;
|
582
|
+
|
583
|
+
while ( (i < S(p->text)) && isspace(T(p->text)[i]) )
|
584
|
+
++i;
|
585
|
+
|
586
|
+
CLIP(p->text, 0, i);
|
587
|
+
UNCHECK(p);
|
588
|
+
|
589
|
+
for (j=S(p->text); (j > 1) && (T(p->text)[j-1] == '#'); --j)
|
590
|
+
;
|
591
|
+
|
592
|
+
while ( j && isspace(T(p->text)[j-1]) )
|
593
|
+
--j;
|
594
|
+
|
595
|
+
S(p->text) = j;
|
596
|
+
|
597
|
+
ret = p->next;
|
598
|
+
p->next = 0;
|
599
|
+
break;
|
600
|
+
}
|
601
|
+
return ret;
|
602
|
+
}
|
603
|
+
|
604
|
+
|
605
|
+
static Line *
|
606
|
+
codeblock(Paragraph *p)
|
607
|
+
{
|
608
|
+
Line *t = p->text, *r;
|
609
|
+
|
610
|
+
for ( ; t; t = r ) {
|
611
|
+
CLIP(t->text,0,4);
|
612
|
+
t->dle = mkd_firstnonblank(t);
|
613
|
+
|
614
|
+
if ( !( (r = skipempty(t->next)) && iscode(r)) ) {
|
615
|
+
___mkd_freeLineRange(t,r);
|
616
|
+
t->next = 0;
|
617
|
+
return r;
|
618
|
+
}
|
619
|
+
}
|
620
|
+
return t;
|
621
|
+
}
|
622
|
+
|
623
|
+
|
624
|
+
static int
|
625
|
+
iscodefence(Line *r, int size, line_type kind, DWORD flags)
|
626
|
+
{
|
627
|
+
if ( !(flags & MKD_FENCEDCODE) )
|
628
|
+
return 0;
|
629
|
+
|
630
|
+
if ( !(r->flags & CHECKED) )
|
631
|
+
checkline(r, flags);
|
632
|
+
|
633
|
+
if ( kind )
|
634
|
+
return (r->kind == kind) && (r->count >= size);
|
635
|
+
else
|
636
|
+
return (r->kind == chk_tilde || r->kind == chk_backtick) && (r->count >= size);
|
637
|
+
}
|
638
|
+
|
639
|
+
static Paragraph *
|
640
|
+
fencedcodeblock(ParagraphRoot *d, Line **ptr, DWORD flags)
|
641
|
+
{
|
642
|
+
Line *first, *r;
|
643
|
+
Paragraph *ret;
|
644
|
+
|
645
|
+
first = (*ptr);
|
646
|
+
|
647
|
+
/* don't allow zero-length code fences
|
648
|
+
*/
|
649
|
+
if ( (first->next == 0) || iscodefence(first->next, first->count, 0, flags) )
|
650
|
+
return 0;
|
651
|
+
|
652
|
+
/* find the closing fence, discard the fences,
|
653
|
+
* return a Paragraph with the contents
|
654
|
+
*/
|
655
|
+
for ( r = first; r && r->next; r = r->next )
|
656
|
+
if ( iscodefence(r->next, first->count, first->kind, flags) ) {
|
657
|
+
(*ptr) = r->next->next;
|
658
|
+
ret = Pp(d, first->next, CODE);
|
659
|
+
if (S(first->text) - first->count > 0) {
|
660
|
+
char *lang_attr = T(first->text) + first->count;
|
661
|
+
while ( *lang_attr != 0 && *lang_attr == ' ' ) lang_attr++;
|
662
|
+
ret->lang = strdup(lang_attr);
|
663
|
+
}
|
664
|
+
else {
|
665
|
+
ret->lang = 0;
|
666
|
+
}
|
667
|
+
___mkd_freeLine(first);
|
668
|
+
___mkd_freeLine(r->next);
|
669
|
+
r->next = 0;
|
670
|
+
return ret;
|
671
|
+
}
|
672
|
+
return 0;
|
673
|
+
}
|
674
|
+
|
675
|
+
|
676
|
+
static int
|
677
|
+
centered(Line *first, Line *last)
|
678
|
+
{
|
679
|
+
|
680
|
+
if ( first&&last ) {
|
681
|
+
int len = S(last->text);
|
682
|
+
|
683
|
+
if ( (len > 2) && (strncmp(T(first->text), "->", 2) == 0)
|
684
|
+
&& (strncmp(T(last->text)+len-2, "<-", 2) == 0) ) {
|
685
|
+
CLIP(first->text, 0, 2);
|
686
|
+
S(last->text) -= 2;
|
687
|
+
return CENTER;
|
688
|
+
}
|
689
|
+
}
|
690
|
+
return 0;
|
691
|
+
}
|
692
|
+
|
693
|
+
|
694
|
+
static int
|
695
|
+
endoftextblock(Line *t, int toplevelblock, DWORD flags)
|
696
|
+
{
|
697
|
+
int z;
|
698
|
+
|
699
|
+
if ( end_of_block(t, flags) || isquote(t) )
|
700
|
+
return 1;
|
701
|
+
|
702
|
+
/* HORRIBLE STANDARDS KLUDGES:
|
703
|
+
* 1. non-toplevel paragraphs absorb adjacent code blocks
|
704
|
+
* 2. Toplevel paragraphs eat absorb adjacent list items,
|
705
|
+
* but sublevel blocks behave properly.
|
706
|
+
* (What this means is that we only need to check for code
|
707
|
+
* blocks at toplevel, and only check for list items at
|
708
|
+
* nested levels.)
|
709
|
+
*/
|
710
|
+
return toplevelblock ? 0 : islist(t,&z,flags,&z);
|
711
|
+
}
|
712
|
+
|
713
|
+
|
714
|
+
static Line *
|
715
|
+
textblock(Paragraph *p, int toplevel, DWORD flags)
|
716
|
+
{
|
717
|
+
Line *t, *next;
|
718
|
+
|
719
|
+
for ( t = p->text; t ; t = next ) {
|
720
|
+
if ( ((next = t->next) == 0) || endoftextblock(next, toplevel, flags) ) {
|
721
|
+
p->align = centered(p->text, t);
|
722
|
+
t->next = 0;
|
723
|
+
return next;
|
724
|
+
}
|
725
|
+
}
|
726
|
+
return t;
|
727
|
+
}
|
728
|
+
|
729
|
+
|
730
|
+
/* length of the id: or class: kind in a special div-not-quote block
|
731
|
+
*/
|
732
|
+
static int
|
733
|
+
szmarkerclass(char *p)
|
734
|
+
{
|
735
|
+
if ( strncasecmp(p, "id:", 3) == 0 )
|
736
|
+
return 3;
|
737
|
+
if ( strncasecmp(p, "class:", 6) == 0 )
|
738
|
+
return 6;
|
739
|
+
return 0;
|
740
|
+
}
|
741
|
+
|
742
|
+
|
743
|
+
/*
|
744
|
+
* check if the first line of a quoted block is the special div-not-quote
|
745
|
+
* marker %[kind:]name%
|
746
|
+
*/
|
747
|
+
#define iscsschar(c) (isalpha(c) || (c == '-') || (c == '_') )
|
748
|
+
|
749
|
+
static int
|
750
|
+
isdivmarker(Line *p, int start, DWORD flags)
|
751
|
+
{
|
752
|
+
char *s;
|
753
|
+
int last, i;
|
754
|
+
|
755
|
+
if ( flags & (MKD_NODIVQUOTE|MKD_STRICT) )
|
756
|
+
return 0;
|
757
|
+
|
758
|
+
start = nextnonblank(p, start);
|
759
|
+
last= S(p->text) - (1 + start);
|
760
|
+
s = T(p->text) + start;
|
761
|
+
|
762
|
+
if ( (last <= 0) || (*s != '%') || (s[last] != '%') )
|
763
|
+
return 0;
|
764
|
+
|
765
|
+
i = szmarkerclass(s+1);
|
766
|
+
|
767
|
+
if ( !iscsschar(s[i+1]) )
|
768
|
+
return 0;
|
769
|
+
while ( ++i < last )
|
770
|
+
if ( !(isdigit(s[i]) || iscsschar(s[i])) )
|
771
|
+
return 0;
|
772
|
+
|
773
|
+
return 1;
|
774
|
+
}
|
775
|
+
|
776
|
+
|
777
|
+
/*
|
778
|
+
* accumulate a blockquote.
|
779
|
+
*
|
780
|
+
* one sick horrible thing about blockquotes is that even though
|
781
|
+
* it just takes ^> to start a quote, following lines, if quoted,
|
782
|
+
* assume that the prefix is ``> ''. This means that code needs
|
783
|
+
* to be indented *5* spaces from the leading '>', but *4* spaces
|
784
|
+
* from the start of the line. This does not appear to be
|
785
|
+
* documented in the reference implementation, but it's the
|
786
|
+
* way the markdown sample web form at Daring Fireball works.
|
787
|
+
*/
|
788
|
+
static Line *
|
789
|
+
quoteblock(Paragraph *p, DWORD flags)
|
790
|
+
{
|
791
|
+
Line *t, *q;
|
792
|
+
int qp;
|
793
|
+
|
794
|
+
for ( t = p->text; t ; t = q ) {
|
795
|
+
if ( isquote(t) ) {
|
796
|
+
/* clip leading spaces */
|
797
|
+
for (qp = 0; T(t->text)[qp] != '>'; qp ++)
|
798
|
+
/* assert: the first nonblank character on this line
|
799
|
+
* will be a >
|
800
|
+
*/;
|
801
|
+
/* clip '>' */
|
802
|
+
qp++;
|
803
|
+
/* clip next space, if any */
|
804
|
+
if ( T(t->text)[qp] == ' ' )
|
805
|
+
qp++;
|
806
|
+
CLIP(t->text, 0, qp);
|
807
|
+
UNCHECK(t);
|
808
|
+
t->dle = mkd_firstnonblank(t);
|
809
|
+
}
|
810
|
+
|
811
|
+
q = skipempty(t->next);
|
812
|
+
|
813
|
+
if ( (q == 0) || ((q != t->next) && (!isquote(q) || isdivmarker(q,1,flags))) ) {
|
814
|
+
___mkd_freeLineRange(t, q);
|
815
|
+
t = q;
|
816
|
+
break;
|
817
|
+
}
|
818
|
+
}
|
819
|
+
if ( isdivmarker(p->text,0,flags) ) {
|
820
|
+
char *prefix = "class";
|
821
|
+
int i;
|
822
|
+
|
823
|
+
q = p->text;
|
824
|
+
p->text = p->text->next;
|
825
|
+
|
826
|
+
if ( (i = szmarkerclass(1+T(q->text))) == 3 )
|
827
|
+
/* and this would be an "%id:" prefix */
|
828
|
+
prefix="id";
|
829
|
+
|
830
|
+
if ( p->ident = malloc(4+strlen(prefix)+S(q->text)) )
|
831
|
+
sprintf(p->ident, "%s=\"%.*s\"", prefix, S(q->text)-(i+2),
|
832
|
+
T(q->text)+(i+1) );
|
833
|
+
|
834
|
+
___mkd_freeLine(q);
|
835
|
+
}
|
836
|
+
return t;
|
837
|
+
}
|
838
|
+
|
839
|
+
|
840
|
+
typedef int (*linefn)(Line *);
|
841
|
+
|
842
|
+
|
843
|
+
/*
|
844
|
+
* pull in a list block. A list block starts with a list marker and
|
845
|
+
* runs until the next list marker, the next non-indented paragraph,
|
846
|
+
* or EOF. You do not have to indent nonblank lines after the list
|
847
|
+
* marker, but multiple paragraphs need to start with a 4-space indent.
|
848
|
+
*/
|
849
|
+
static Line *
|
850
|
+
listitem(Paragraph *p, int indent, DWORD flags, linefn check)
|
851
|
+
{
|
852
|
+
Line *t, *q;
|
853
|
+
int clip = indent;
|
854
|
+
int z;
|
855
|
+
|
856
|
+
for ( t = p->text; t ; t = q) {
|
857
|
+
CLIP(t->text, 0, clip);
|
858
|
+
UNCHECK(t);
|
859
|
+
t->dle = mkd_firstnonblank(t);
|
860
|
+
|
861
|
+
/* even though we had to trim a long leader off this item,
|
862
|
+
* the indent for trailing paragraphs is still 4...
|
863
|
+
*/
|
864
|
+
if (indent > 4) {
|
865
|
+
indent = 4;
|
866
|
+
}
|
867
|
+
if ( (q = skipempty(t->next)) == 0 ) {
|
868
|
+
___mkd_freeLineRange(t,q);
|
869
|
+
return 0;
|
870
|
+
}
|
871
|
+
|
872
|
+
/* after a blank line, the next block needs to start with a line
|
873
|
+
* that's indented 4(? -- reference implementation allows a 1
|
874
|
+
* character indent, but that has unfortunate side effects here)
|
875
|
+
* spaces, but after that the line doesn't need any indentation
|
876
|
+
*/
|
877
|
+
if ( q != t->next ) {
|
878
|
+
if (q->dle < indent) {
|
879
|
+
q = t->next;
|
880
|
+
t->next = 0;
|
881
|
+
return q;
|
882
|
+
}
|
883
|
+
/* indent at least 2, and at most as
|
884
|
+
* as far as the initial line was indented. */
|
885
|
+
indent = clip ? clip : 2;
|
886
|
+
}
|
887
|
+
|
888
|
+
if ( (q->dle < indent) && (ishr(q,flags) || islist(q,&z,flags,&z)
|
889
|
+
|| (check && (*check)(q)))
|
890
|
+
&& !issetext(q,&z,flags) ) {
|
891
|
+
q = t->next;
|
892
|
+
t->next = 0;
|
893
|
+
return q;
|
894
|
+
}
|
895
|
+
|
896
|
+
clip = (q->dle > indent) ? indent : q->dle;
|
897
|
+
}
|
898
|
+
return t;
|
899
|
+
}
|
900
|
+
|
901
|
+
|
902
|
+
static Line *
|
903
|
+
definition_block(Paragraph *top, int clip, MMIOT *f, int kind)
|
904
|
+
{
|
905
|
+
ParagraphRoot d = { 0, 0 };
|
906
|
+
Paragraph *p;
|
907
|
+
Line *q = top->text, *text = 0, *labels;
|
908
|
+
int z, para;
|
909
|
+
|
910
|
+
while (( labels = q )) {
|
911
|
+
|
912
|
+
if ( (q = isdefinition(labels, &z, &kind, f->flags)) == 0 )
|
913
|
+
break;
|
914
|
+
|
915
|
+
if ( (text = skipempty(q->next)) == 0 )
|
916
|
+
break;
|
917
|
+
|
918
|
+
if ( para = (text != q->next) )
|
919
|
+
___mkd_freeLineRange(q, text);
|
920
|
+
|
921
|
+
q->next = 0;
|
922
|
+
if ( kind == 1 /* discount dl */ )
|
923
|
+
for ( q = labels; q; q = q->next ) {
|
924
|
+
CLIP(q->text, 0, 1);
|
925
|
+
UNCHECK(q);
|
926
|
+
S(q->text)--;
|
927
|
+
}
|
928
|
+
|
929
|
+
dd_block:
|
930
|
+
p = Pp(&d, text, LISTITEM);
|
931
|
+
|
932
|
+
text = listitem(p, clip, f->flags, (kind==2) ? is_extra_dd : 0);
|
933
|
+
p->down = compile(p->text, 0, f);
|
934
|
+
p->text = labels; labels = 0;
|
935
|
+
|
936
|
+
if ( para && p->down ) p->down->align = PARA;
|
937
|
+
|
938
|
+
if ( (q = skipempty(text)) == 0 )
|
939
|
+
break;
|
940
|
+
|
941
|
+
if ( para = (q != text) ) {
|
942
|
+
Line anchor;
|
943
|
+
|
944
|
+
anchor.next = text;
|
945
|
+
___mkd_freeLineRange(&anchor,q);
|
946
|
+
text = q;
|
947
|
+
|
948
|
+
}
|
949
|
+
|
950
|
+
if ( kind == 2 && is_extra_dd(q) )
|
951
|
+
goto dd_block;
|
952
|
+
}
|
953
|
+
top->text = 0;
|
954
|
+
top->down = T(d);
|
955
|
+
return text;
|
956
|
+
}
|
957
|
+
|
958
|
+
|
959
|
+
static Line *
|
960
|
+
enumerated_block(Paragraph *top, int clip, MMIOT *f, int list_class)
|
961
|
+
{
|
962
|
+
ParagraphRoot d = { 0, 0 };
|
963
|
+
Paragraph *p;
|
964
|
+
Line *q = top->text, *text;
|
965
|
+
int para = 0, z;
|
966
|
+
|
967
|
+
while (( text = q )) {
|
968
|
+
|
969
|
+
p = Pp(&d, text, LISTITEM);
|
970
|
+
text = listitem(p, clip, f->flags, 0);
|
971
|
+
|
972
|
+
p->down = compile(p->text, 0, f);
|
973
|
+
p->text = 0;
|
974
|
+
|
975
|
+
if ( para && p->down ) p->down->align = PARA;
|
976
|
+
|
977
|
+
if ( (q = skipempty(text)) == 0
|
978
|
+
|| islist(q, &clip, f->flags, &z) != list_class )
|
979
|
+
break;
|
980
|
+
|
981
|
+
if ( para = (q != text) ) {
|
982
|
+
Line anchor;
|
983
|
+
|
984
|
+
anchor.next = text;
|
985
|
+
___mkd_freeLineRange(&anchor, q);
|
986
|
+
|
987
|
+
if ( p->down ) p->down->align = PARA;
|
988
|
+
}
|
989
|
+
}
|
990
|
+
top->text = 0;
|
991
|
+
top->down = T(d);
|
992
|
+
return text;
|
993
|
+
}
|
994
|
+
|
995
|
+
|
996
|
+
static int
|
997
|
+
tgood(char c)
|
998
|
+
{
|
999
|
+
switch (c) {
|
1000
|
+
case '\'':
|
1001
|
+
case '"': return c;
|
1002
|
+
case '(': return ')';
|
1003
|
+
}
|
1004
|
+
return 0;
|
1005
|
+
}
|
1006
|
+
|
1007
|
+
|
1008
|
+
/*
|
1009
|
+
* add a new (image or link) footnote to the footnote table
|
1010
|
+
*/
|
1011
|
+
static Line*
|
1012
|
+
addfootnote(Line *p, MMIOT* f)
|
1013
|
+
{
|
1014
|
+
int j, i;
|
1015
|
+
int c;
|
1016
|
+
Line *np = p->next;
|
1017
|
+
|
1018
|
+
Footnote *foot = &EXPAND(f->footnotes->note);
|
1019
|
+
|
1020
|
+
CREATE(foot->tag);
|
1021
|
+
CREATE(foot->link);
|
1022
|
+
CREATE(foot->title);
|
1023
|
+
foot->flags = foot->height = foot->width = 0;
|
1024
|
+
|
1025
|
+
for (j=i=p->dle+1; T(p->text)[j] != ']'; j++)
|
1026
|
+
EXPAND(foot->tag) = T(p->text)[j];
|
1027
|
+
|
1028
|
+
EXPAND(foot->tag) = 0;
|
1029
|
+
S(foot->tag)--;
|
1030
|
+
j = nextnonblank(p, j+2);
|
1031
|
+
|
1032
|
+
if ( (f->flags & MKD_EXTRA_FOOTNOTE) && (T(foot->tag)[0] == '^') ) {
|
1033
|
+
/* need to consume all lines until non-indented block? */
|
1034
|
+
while ( j < S(p->text) )
|
1035
|
+
EXPAND(foot->title) = T(p->text)[j++];
|
1036
|
+
goto skip_to_end;
|
1037
|
+
}
|
1038
|
+
|
1039
|
+
while ( (j < S(p->text)) && !isspace(T(p->text)[j]) )
|
1040
|
+
EXPAND(foot->link) = T(p->text)[j++];
|
1041
|
+
EXPAND(foot->link) = 0;
|
1042
|
+
S(foot->link)--;
|
1043
|
+
j = nextnonblank(p,j);
|
1044
|
+
|
1045
|
+
if ( T(p->text)[j] == '=' ) {
|
1046
|
+
sscanf(T(p->text)+j, "=%dx%d", &foot->width, &foot->height);
|
1047
|
+
while ( (j < S(p->text)) && !isspace(T(p->text)[j]) )
|
1048
|
+
++j;
|
1049
|
+
j = nextnonblank(p,j);
|
1050
|
+
}
|
1051
|
+
|
1052
|
+
|
1053
|
+
if ( (j >= S(p->text)) && np && np->dle && tgood(T(np->text)[np->dle]) ) {
|
1054
|
+
___mkd_freeLine(p);
|
1055
|
+
p = np;
|
1056
|
+
np = p->next;
|
1057
|
+
j = p->dle;
|
1058
|
+
}
|
1059
|
+
|
1060
|
+
if ( (c = tgood(T(p->text)[j])) ) {
|
1061
|
+
/* Try to take the rest of the line as a comment; read to
|
1062
|
+
* EOL, then shrink the string back to before the final
|
1063
|
+
* quote.
|
1064
|
+
*/
|
1065
|
+
++j; /* skip leading quote */
|
1066
|
+
|
1067
|
+
while ( j < S(p->text) )
|
1068
|
+
EXPAND(foot->title) = T(p->text)[j++];
|
1069
|
+
|
1070
|
+
while ( S(foot->title) && T(foot->title)[S(foot->title)-1] != c )
|
1071
|
+
--S(foot->title);
|
1072
|
+
if ( S(foot->title) ) /* skip trailing quote */
|
1073
|
+
--S(foot->title);
|
1074
|
+
EXPAND(foot->title) = 0;
|
1075
|
+
--S(foot->title);
|
1076
|
+
}
|
1077
|
+
|
1078
|
+
skip_to_end:
|
1079
|
+
___mkd_freeLine(p);
|
1080
|
+
return np;
|
1081
|
+
}
|
1082
|
+
|
1083
|
+
|
1084
|
+
/*
|
1085
|
+
* allocate a paragraph header, link it to the
|
1086
|
+
* tail of the current document
|
1087
|
+
*/
|
1088
|
+
static Paragraph *
|
1089
|
+
Pp(ParagraphRoot *d, Line *ptr, int typ)
|
1090
|
+
{
|
1091
|
+
Paragraph *ret = calloc(sizeof *ret, 1);
|
1092
|
+
|
1093
|
+
ret->text = ptr;
|
1094
|
+
ret->typ = typ;
|
1095
|
+
|
1096
|
+
return ATTACH(*d, ret);
|
1097
|
+
}
|
1098
|
+
|
1099
|
+
|
1100
|
+
|
1101
|
+
static Line*
|
1102
|
+
consume(Line *ptr, int *eaten)
|
1103
|
+
{
|
1104
|
+
Line *next;
|
1105
|
+
int blanks=0;
|
1106
|
+
|
1107
|
+
for (; ptr && blankline(ptr); ptr = next, blanks++ ) {
|
1108
|
+
next = ptr->next;
|
1109
|
+
___mkd_freeLine(ptr);
|
1110
|
+
}
|
1111
|
+
if ( ptr ) *eaten = blanks;
|
1112
|
+
return ptr;
|
1113
|
+
}
|
1114
|
+
|
1115
|
+
|
1116
|
+
/*
|
1117
|
+
* top-level compilation; break the document into
|
1118
|
+
* style, html, and source blocks with footnote links
|
1119
|
+
* weeded out.
|
1120
|
+
*/
|
1121
|
+
static Paragraph *
|
1122
|
+
compile_document(Line *ptr, MMIOT *f)
|
1123
|
+
{
|
1124
|
+
ParagraphRoot d = { 0, 0 };
|
1125
|
+
ANCHOR(Line) source = { 0, 0 };
|
1126
|
+
Paragraph *p = 0;
|
1127
|
+
struct kw *tag;
|
1128
|
+
int eaten, unclosed;
|
1129
|
+
|
1130
|
+
while ( ptr ) {
|
1131
|
+
if ( !(f->flags & MKD_NOHTML) && (tag = isopentag(ptr)) ) {
|
1132
|
+
int blocktype;
|
1133
|
+
/* If we encounter a html/style block, compile and save all
|
1134
|
+
* of the cached source BEFORE processing the html/style.
|
1135
|
+
*/
|
1136
|
+
if ( T(source) ) {
|
1137
|
+
E(source)->next = 0;
|
1138
|
+
p = Pp(&d, 0, SOURCE);
|
1139
|
+
p->down = compile(T(source), 1, f);
|
1140
|
+
T(source) = E(source) = 0;
|
1141
|
+
}
|
1142
|
+
|
1143
|
+
if ( f->flags & MKD_NOSTYLE )
|
1144
|
+
blocktype = HTML;
|
1145
|
+
else
|
1146
|
+
blocktype = strcmp(tag->id, "STYLE") == 0 ? STYLE : HTML;
|
1147
|
+
p = Pp(&d, ptr, blocktype);
|
1148
|
+
ptr = htmlblock(p, tag, &unclosed);
|
1149
|
+
if ( unclosed ) {
|
1150
|
+
p->typ = SOURCE;
|
1151
|
+
p->down = compile(p->text, 1, f);
|
1152
|
+
p->text = 0;
|
1153
|
+
}
|
1154
|
+
}
|
1155
|
+
else if ( isfootnote(ptr) ) {
|
1156
|
+
/* footnotes, like cats, sleep anywhere; pull them
|
1157
|
+
* out of the input stream and file them away for
|
1158
|
+
* later processing
|
1159
|
+
*/
|
1160
|
+
ptr = consume(addfootnote(ptr, f), &eaten);
|
1161
|
+
}
|
1162
|
+
else {
|
1163
|
+
/* source; cache it up to wait for eof or the
|
1164
|
+
* next html/style block
|
1165
|
+
*/
|
1166
|
+
ATTACH(source,ptr);
|
1167
|
+
ptr = ptr->next;
|
1168
|
+
}
|
1169
|
+
}
|
1170
|
+
if ( T(source) ) {
|
1171
|
+
/* if there's any cached source at EOF, compile
|
1172
|
+
* it now.
|
1173
|
+
*/
|
1174
|
+
E(source)->next = 0;
|
1175
|
+
p = Pp(&d, 0, SOURCE);
|
1176
|
+
p->down = compile(T(source), 1, f);
|
1177
|
+
}
|
1178
|
+
return T(d);
|
1179
|
+
}
|
1180
|
+
|
1181
|
+
|
1182
|
+
static int
|
1183
|
+
first_nonblank_before(Line *j, int dle)
|
1184
|
+
{
|
1185
|
+
return (j->dle < dle) ? j->dle : dle;
|
1186
|
+
}
|
1187
|
+
|
1188
|
+
|
1189
|
+
static int
|
1190
|
+
actually_a_table(MMIOT *f, Line *pp)
|
1191
|
+
{
|
1192
|
+
Line *r;
|
1193
|
+
int j;
|
1194
|
+
int c;
|
1195
|
+
|
1196
|
+
/* tables need to be turned on */
|
1197
|
+
if ( f->flags & (MKD_STRICT|MKD_NOTABLES) )
|
1198
|
+
return 0;
|
1199
|
+
|
1200
|
+
/* tables need three lines */
|
1201
|
+
if ( !(pp && pp->next && pp->next->next) ) {
|
1202
|
+
return 0;
|
1203
|
+
}
|
1204
|
+
|
1205
|
+
/* all lines must contain |'s */
|
1206
|
+
for (r = pp; r; r = r->next )
|
1207
|
+
if ( !(r->flags & PIPECHAR) ) {
|
1208
|
+
return 0;
|
1209
|
+
}
|
1210
|
+
|
1211
|
+
/* if the header has a leading |, all lines must have leading |'s */
|
1212
|
+
if ( T(pp->text)[pp->dle] == '|' ) {
|
1213
|
+
for ( r = pp; r; r = r->next )
|
1214
|
+
if ( T(r->text)[first_nonblank_before(r,pp->dle)] != '|' ) {
|
1215
|
+
return 0;
|
1216
|
+
}
|
1217
|
+
}
|
1218
|
+
|
1219
|
+
/* second line must be only whitespace, -, |, or : */
|
1220
|
+
r = pp->next;
|
1221
|
+
|
1222
|
+
for ( j=r->dle; j < S(r->text); ++j ) {
|
1223
|
+
c = T(r->text)[j];
|
1224
|
+
|
1225
|
+
if ( !(isspace(c)||(c=='-')||(c==':')||(c=='|')) ) {
|
1226
|
+
return 0;
|
1227
|
+
}
|
1228
|
+
}
|
1229
|
+
|
1230
|
+
return 1;
|
1231
|
+
}
|
1232
|
+
|
1233
|
+
|
1234
|
+
/*
|
1235
|
+
* break a collection of markdown input into
|
1236
|
+
* blocks of lists, code, html, and text to
|
1237
|
+
* be marked up.
|
1238
|
+
*/
|
1239
|
+
static Paragraph *
|
1240
|
+
compile(Line *ptr, int toplevel, MMIOT *f)
|
1241
|
+
{
|
1242
|
+
ParagraphRoot d = { 0, 0 };
|
1243
|
+
Paragraph *p = 0;
|
1244
|
+
Line *r;
|
1245
|
+
int para = toplevel;
|
1246
|
+
int blocks = 0;
|
1247
|
+
int hdr_type, list_type, list_class, indent;
|
1248
|
+
|
1249
|
+
ptr = consume(ptr, ¶);
|
1250
|
+
|
1251
|
+
while ( ptr ) {
|
1252
|
+
if ( iscode(ptr) ) {
|
1253
|
+
p = Pp(&d, ptr, CODE);
|
1254
|
+
|
1255
|
+
if ( f->flags & MKD_1_COMPAT) {
|
1256
|
+
/* HORRIBLE STANDARDS KLUDGE: the first line of every block
|
1257
|
+
* has trailing whitespace trimmed off.
|
1258
|
+
*/
|
1259
|
+
___mkd_tidy(&p->text->text);
|
1260
|
+
}
|
1261
|
+
|
1262
|
+
ptr = codeblock(p);
|
1263
|
+
}
|
1264
|
+
else if ( iscodefence(ptr,3,0,f->flags) && (p=fencedcodeblock(&d, &ptr, f->flags)) )
|
1265
|
+
/* yay, it's already done */ ;
|
1266
|
+
else if ( ishr(ptr, f->flags) ) {
|
1267
|
+
p = Pp(&d, 0, HR);
|
1268
|
+
r = ptr;
|
1269
|
+
ptr = ptr->next;
|
1270
|
+
___mkd_freeLine(r);
|
1271
|
+
}
|
1272
|
+
else if ( list_class = islist(ptr, &indent, f->flags, &list_type) ) {
|
1273
|
+
if ( list_class == DL ) {
|
1274
|
+
p = Pp(&d, ptr, DL);
|
1275
|
+
ptr = definition_block(p, indent, f, list_type);
|
1276
|
+
}
|
1277
|
+
else {
|
1278
|
+
p = Pp(&d, ptr, list_type);
|
1279
|
+
ptr = enumerated_block(p, indent, f, list_class);
|
1280
|
+
}
|
1281
|
+
}
|
1282
|
+
else if ( isquote(ptr) ) {
|
1283
|
+
p = Pp(&d, ptr, QUOTE);
|
1284
|
+
ptr = quoteblock(p, f->flags);
|
1285
|
+
p->down = compile(p->text, 1, f);
|
1286
|
+
p->text = 0;
|
1287
|
+
}
|
1288
|
+
else if ( ishdr(ptr, &hdr_type, f->flags) ) {
|
1289
|
+
p = Pp(&d, ptr, HDR);
|
1290
|
+
ptr = headerblock(p, hdr_type);
|
1291
|
+
}
|
1292
|
+
else {
|
1293
|
+
p = Pp(&d, ptr, MARKUP);
|
1294
|
+
ptr = textblock(p, toplevel, f->flags);
|
1295
|
+
/* tables are a special kind of paragraph */
|
1296
|
+
if ( actually_a_table(f, p->text) )
|
1297
|
+
p->typ = TABLE;
|
1298
|
+
}
|
1299
|
+
|
1300
|
+
if ( (para||toplevel) && !p->align )
|
1301
|
+
p->align = PARA;
|
1302
|
+
|
1303
|
+
blocks++;
|
1304
|
+
para = toplevel || (blocks > 1);
|
1305
|
+
ptr = consume(ptr, ¶);
|
1306
|
+
|
1307
|
+
if ( para && !p->align )
|
1308
|
+
p->align = PARA;
|
1309
|
+
|
1310
|
+
}
|
1311
|
+
return T(d);
|
1312
|
+
}
|
1313
|
+
|
1314
|
+
|
1315
|
+
/*
|
1316
|
+
* the guts of the markdown() function, ripped out so I can do
|
1317
|
+
* debugging.
|
1318
|
+
*/
|
1319
|
+
|
1320
|
+
/*
|
1321
|
+
* prepare and compile `text`, returning a Paragraph tree.
|
1322
|
+
*/
|
1323
|
+
int
|
1324
|
+
mkd_compile(Document *doc, DWORD flags)
|
1325
|
+
{
|
1326
|
+
if ( !doc )
|
1327
|
+
return 0;
|
1328
|
+
|
1329
|
+
flags &= USER_FLAGS;
|
1330
|
+
|
1331
|
+
if ( doc->compiled ) {
|
1332
|
+
if ( doc->ctx->flags == flags )
|
1333
|
+
return 1;
|
1334
|
+
else {
|
1335
|
+
if ( doc->code)
|
1336
|
+
___mkd_freeParagraph(doc->code);
|
1337
|
+
if ( doc->ctx->footnotes )
|
1338
|
+
___mkd_freefootnotes(doc->ctx);
|
1339
|
+
}
|
1340
|
+
}
|
1341
|
+
|
1342
|
+
doc->compiled = 1;
|
1343
|
+
memset(doc->ctx, 0, sizeof(MMIOT) );
|
1344
|
+
doc->ctx->ref_prefix= doc->ref_prefix;
|
1345
|
+
doc->ctx->cb = &(doc->cb);
|
1346
|
+
doc->ctx->flags = flags;
|
1347
|
+
CREATE(doc->ctx->in);
|
1348
|
+
doc->ctx->footnotes = malloc(sizeof doc->ctx->footnotes[0]);
|
1349
|
+
doc->ctx->footnotes->reference = 0;
|
1350
|
+
CREATE(doc->ctx->footnotes->note);
|
1351
|
+
|
1352
|
+
mkd_initialize();
|
1353
|
+
|
1354
|
+
doc->code = compile_document(T(doc->content), doc->ctx);
|
1355
|
+
qsort(T(doc->ctx->footnotes->note), S(doc->ctx->footnotes->note),
|
1356
|
+
sizeof T(doc->ctx->footnotes->note)[0],
|
1357
|
+
(stfu)__mkd_footsort);
|
1358
|
+
memset(&doc->content, 0, sizeof doc->content);
|
1359
|
+
return 1;
|
1360
|
+
}
|
1361
|
+
|