aboutsummaryrefslogtreecommitdiff
path: root/src/Text/Pandoc/Readers/Markdown.hs
blob: 3b363263d0ba6e131e9aba09daa599ca0c608bc0 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
{-# LANGUAGE RelaxedPolyRec      #-}
{-# LANGUAGE ScopedTypeVariables #-}
{-# LANGUAGE TupleSections       #-}
{-# LANGUAGE OverloadedStrings   #-}
{-# LANGUAGE ViewPatterns        #-}
{- |
   Module      : Text.Pandoc.Readers.Markdown
   Copyright   : Copyright (C) 2006-2020 John MacFarlane
   License     : GNU GPL, version 2 or above

   Maintainer  : John MacFarlane <jgm@berkeley.edu>
   Stability   : alpha
   Portability : portable

Conversion of markdown-formatted plain text to 'Pandoc' document.
-}
module Text.Pandoc.Readers.Markdown ( readMarkdown, yamlToMeta ) where

import Control.Monad
import Control.Monad.Except (throwError)
import Data.Char (isAlphaNum, isPunctuation, isSpace)
import Data.List (transpose, elemIndex, sortOn)
import qualified Data.Map as M
import Data.Maybe
import qualified Data.Set as Set
import Data.Text (Text)
import qualified Data.Text as T
import qualified Data.Text.Lazy as TL
import qualified Data.ByteString.Lazy as BL
import System.FilePath (addExtension, takeExtension)
import Text.HTML.TagSoup hiding (Row)
import Text.Pandoc.Builder (Blocks, Inlines)
import qualified Text.Pandoc.Builder as B
import Text.Pandoc.Class.PandocMonad (PandocMonad (..), report)
import Text.Pandoc.Definition as Pandoc
import Text.Pandoc.Emoji (emojiToInline)
import Text.Pandoc.Error
import Text.Pandoc.Logging
import Text.Pandoc.Options
import Text.Pandoc.Parsing hiding (tableWith)
import Text.Pandoc.Readers.HTML (htmlInBalanced, htmlTag, isBlockTag,
                                 isCommentTag, isInlineTag, isTextTag)
import Text.Pandoc.Readers.LaTeX (applyMacros, rawLaTeXBlock, rawLaTeXInline)
import Text.Pandoc.Shared
import qualified Text.Pandoc.UTF8 as UTF8
import Text.Pandoc.XML (fromEntities)
import Text.Pandoc.Readers.Metadata (yamlBsToMeta)

type MarkdownParser m = ParserT Text ParserState m

-- | Read markdown from an input string and return a Pandoc document.
readMarkdown :: PandocMonad m
             => ReaderOptions -- ^ Reader options
             -> Text      -- ^ String to parse (assuming @'\n'@ line endings)
             -> m Pandoc
readMarkdown opts s = do
  parsed <- readWithM parseMarkdown def{ stateOptions = opts }
               (crFilter s <> "\n\n")
  case parsed of
    Right result -> return result
    Left e       -> throwError e

-- | Read a YAML string and convert it to pandoc metadata.
-- String scalars in the YAML are parsed as Markdown.
yamlToMeta :: PandocMonad m
           => ReaderOptions
           -> BL.ByteString
           -> m Meta
yamlToMeta opts bstr = do
  let parser = do
        meta <- yamlBsToMeta parseBlocks bstr
        return $ runF meta defaultParserState
  parsed <- readWithM parser def{ stateOptions = opts } ""
  case parsed of
    Right result -> return result
    Left e       -> throwError e



--
-- Constants and data structure definitions
--

isBulletListMarker :: Char -> Bool
isBulletListMarker '*' = True
isBulletListMarker '+' = True
isBulletListMarker '-' = True
isBulletListMarker _   = False

isHruleChar :: Char -> Bool
isHruleChar '*' = True
isHruleChar '-' = True
isHruleChar '_' = True
isHruleChar _   = False

setextHChars :: [Char]
setextHChars = "=-"

isBlank :: Char -> Bool
isBlank ' '  = True
isBlank '\t' = True
isBlank '\n' = True
isBlank _    = False

--
-- auxiliary functions
--

-- | Succeeds when we're in list context.
inList :: PandocMonad m => MarkdownParser m ()
inList = do
  ctx <- stateParserContext <$> getState
  guard (ctx == ListItemState)

spnl :: PandocMonad m => ParserT Text st m ()
spnl = try $ do
  skipSpaces
  optional newline
  skipSpaces
  notFollowedBy (char '\n')

spnl' :: PandocMonad m => ParserT Text st m Text
spnl' = try $ do
  xs <- many spaceChar
  ys <- option "" $ try $ (:) <$> newline
                              <*> (many spaceChar <* notFollowedBy (char '\n'))
  return $ T.pack $ xs ++ ys

indentSpaces :: PandocMonad m => MarkdownParser m Text
indentSpaces = try $ do
  tabStop <- getOption readerTabStop
  countChar tabStop (char ' ') <|>
    textStr "\t" <?> "indentation"

nonindentSpaces :: PandocMonad m => MarkdownParser m Text
nonindentSpaces = do
  n <- skipNonindentSpaces
  return $ T.replicate n " "

-- returns number of spaces parsed
skipNonindentSpaces :: PandocMonad m => MarkdownParser m Int
skipNonindentSpaces = do
  tabStop <- getOption readerTabStop
  gobbleAtMostSpaces (tabStop - 1) <* notFollowedBy spaceChar

litChar :: PandocMonad m => MarkdownParser m Char
litChar = escapedChar'
       <|> characterReference
       <|> noneOf "\n"
       <|> try (newline >> notFollowedBy blankline >> return ' ')

-- | Parse a sequence of inline elements between square brackets,
-- including inlines between balanced pairs of square brackets.
inlinesInBalancedBrackets :: PandocMonad m => MarkdownParser m (F Inlines)
inlinesInBalancedBrackets =
  try $ char '[' >> withRaw (go 1) >>=
          parseFromString inlines . stripBracket . snd
  where stripBracket t = case T.unsnoc t of
          Just (t', ']') -> t'
          _              -> t
        go :: PandocMonad m => Int -> MarkdownParser m ()
        go 0 = return ()
        go openBrackets =
          (() <$ (escapedChar <|>
                code <|>
                rawHtmlInline <|>
                rawLaTeXInline') >> go openBrackets)
          <|>
          (do char ']'
              Control.Monad.when (openBrackets > 1) $ go (openBrackets - 1))
          <|>
          (char '[' >> go (openBrackets + 1))
          <|>
          (anyChar >> go openBrackets)

--
-- document structure
--

rawTitleBlockLine :: PandocMonad m => MarkdownParser m Text
rawTitleBlockLine = do
  char '%'
  skipSpaces
  first <- anyLine
  rest <- many $ try $ do spaceChar
                          notFollowedBy blankline
                          skipSpaces
                          anyLine
  return $ trim $ T.unlines (first:rest)

titleLine :: PandocMonad m => MarkdownParser m (F Inlines)
titleLine = try $ do
  raw <- rawTitleBlockLine
  res <- parseFromString' inlines raw
  return $ trimInlinesF res

authorsLine :: PandocMonad m => MarkdownParser m (F [Inlines])
authorsLine = try $ do
  raw <- rawTitleBlockLine
  let sep = (char ';' <* spaces) <|> newline
  let pAuthors = sepEndBy
            (trimInlinesF . mconcat <$> many
                 (try $ notFollowedBy sep >> inline))
            sep
  sequence <$> parseFromString' pAuthors raw

dateLine :: PandocMonad m => MarkdownParser m (F Inlines)
dateLine = try $ do
  raw <- rawTitleBlockLine
  res <- parseFromString' inlines raw
  return $ trimInlinesF res

titleBlock :: PandocMonad m => MarkdownParser m ()
titleBlock = pandocTitleBlock <|> mmdTitleBlock

pandocTitleBlock :: PandocMonad m => MarkdownParser m ()
pandocTitleBlock = try $ do
  guardEnabled Ext_pandoc_title_block
  lookAhead (char '%')
  title <- option mempty titleLine
  author <- option (return []) authorsLine
  date <- option mempty dateLine
  optional blanklines
  let meta' = do title' <- title
                 author' <- author
                 date' <- date
                 return $
                     (if B.isNull title' then id else B.setMeta "title" title')
                   . (if null author' then id else B.setMeta "author" author')
                   . (if B.isNull date' then id else B.setMeta "date" date')
                   $ nullMeta
  updateState $ \st -> st{ stateMeta' = stateMeta' st <> meta' }

yamlMetaBlock :: PandocMonad m => MarkdownParser m (F Blocks)
yamlMetaBlock = try $ do
  guardEnabled Ext_yaml_metadata_block
  string "---"
  blankline
  notFollowedBy blankline  -- if --- is followed by a blank it's an HRULE
  rawYamlLines <- manyTill anyLine stopLine
  -- by including --- and ..., we allow yaml blocks with just comments:
  let rawYaml = T.unlines ("---" : (rawYamlLines ++ ["..."]))
  optional blanklines
  newMetaF <- yamlBsToMeta parseBlocks
              $ UTF8.fromTextLazy $ TL.fromStrict rawYaml
  -- Since `<>` is left-biased, existing values are not touched:
  updateState $ \st -> st{ stateMeta' = stateMeta' st <> newMetaF }
  return mempty

stopLine :: PandocMonad m => MarkdownParser m ()
stopLine = try $ (string "---" <|> string "...") >> blankline >> return ()

mmdTitleBlock :: PandocMonad m => MarkdownParser m ()
mmdTitleBlock = try $ do
  guardEnabled Ext_mmd_title_block
  firstPair <- kvPair False
  restPairs <- many (kvPair True)
  let kvPairs = firstPair : restPairs
  blanklines
  updateState $ \st -> st{ stateMeta' = stateMeta' st <>
                             return (Meta $ M.fromList kvPairs) }

kvPair :: PandocMonad m => Bool -> MarkdownParser m (Text, MetaValue)
kvPair allowEmpty = try $ do
  key <- many1TillChar (alphaNum <|> oneOf "_- ") (char ':')
  val <- trim <$> manyTillChar anyChar
          (try $ newline >> lookAhead (blankline <|> nonspaceChar))
  guard $ allowEmpty || not (T.null val)
  let key' = T.concat $ T.words $ T.toLower key
  let val' = MetaBlocks $ B.toList $ B.plain $ B.text val
  return (key',val')

parseMarkdown :: PandocMonad m => MarkdownParser m Pandoc
parseMarkdown = do
  optional titleBlock
  blocks <- parseBlocks
  st <- getState
  -- check for notes with no corresponding note references
  let notesUsed = stateNoteRefs st
  let notesDefined = M.keys (stateNotes' st)
  mapM_ (\n -> unless (n `Set.member` notesUsed) $
                case M.lookup n (stateNotes' st) of
                   Just (pos, _) -> report (NoteDefinedButNotUsed n pos)
                   Nothing -> throwError $
                     PandocShouldNeverHappenError "note not found")
         notesDefined
  let doc = runF (do Pandoc _ bs <- B.doc <$> blocks
                     meta <- stateMeta' st
                     return $ Pandoc meta bs) st
  reportLogMessages
  return doc

referenceKey :: PandocMonad m => MarkdownParser m (F Blocks)
referenceKey = try $ do
  pos <- getPosition
  skipNonindentSpaces
  (_,raw) <- reference
  char ':'
  skipSpaces >> optional newline >> skipSpaces >> notFollowedBy (char '[')
  let sourceURL = fmap T.unwords $ many $ try $ do
                    skipMany spaceChar
                    notFollowedBy' referenceTitle
                    notFollowedBy' $ guardEnabled Ext_link_attributes >> attributes
                    notFollowedBy' (() <$ reference)
                    many1Char $ notFollowedBy space >> litChar
  let betweenAngles = try $ char '<' >> manyTillChar litChar (char '>')
  src <- try betweenAngles <|> sourceURL
  tit <- option "" referenceTitle
  attr   <- option nullAttr $ try $
              do guardEnabled Ext_link_attributes
                 skipSpaces >> optional newline >> skipSpaces
                 attributes
  addKvs <- option [] $ guardEnabled Ext_mmd_link_attributes
                          >> many (try $ spnl >> keyValAttr)
  blanklines
  let attr'  = extractIdClass $ foldl (\x f -> f x) attr addKvs
      target = (escapeURI $ trimr src, tit)
  st <- getState
  let oldkeys = stateKeys st
  let key = toKey raw
  case M.lookup key oldkeys of
    Just (t,a) | not (t == target && a == attr') ->
      -- We don't warn on two duplicate keys if the targets are also
      -- the same. This can happen naturally with --reference-location=block
      -- or section. See #3701.
      logMessage $ DuplicateLinkReference raw pos
    _ -> return ()
  updateState $ \s -> s { stateKeys = M.insert key (target, attr') oldkeys }
  return $ return mempty

referenceTitle :: PandocMonad m => MarkdownParser m Text
referenceTitle = try $ do
  skipSpaces >> optional newline >> skipSpaces
  quotedTitle '"' <|> quotedTitle '\'' <|> charsInBalanced '(' ')' litChar

-- A link title in quotes
quotedTitle :: PandocMonad m => Char -> MarkdownParser m Text
quotedTitle c = try $ do
  char c
  notFollowedBy spaces
  let pEnder = try $ char c >> notFollowedBy (satisfy isAlphaNum)
  let regChunk = many1Char (noneOf ['\\','\n','&',c]) <|> countChar 1 litChar
  let nestedChunk = (\x -> T.singleton c <> x <> T.singleton c) <$> quotedTitle c
  T.unwords . T.words . T.concat <$> manyTill (nestedChunk <|> regChunk) pEnder

-- | PHP Markdown Extra style abbreviation key.  Currently
-- we just skip them, since Pandoc doesn't have an element for
-- an abbreviation.
abbrevKey :: PandocMonad m => MarkdownParser m (F Blocks)
abbrevKey = do
  guardEnabled Ext_abbreviations
  try $ do
    char '*'
    reference
    char ':'
    skipMany (satisfy (/= '\n'))
    blanklines
    return $ return mempty

noteMarker :: PandocMonad m => MarkdownParser m Text
noteMarker = string "[^" >> many1TillChar (satisfy $ not . isBlank) (char ']')

rawLine :: PandocMonad m => MarkdownParser m Text
rawLine = try $ do
  notFollowedBy blankline
  notFollowedBy' $ try $ skipNonindentSpaces >> noteMarker
  optional indentSpaces
  anyLine

rawLines :: PandocMonad m => MarkdownParser m Text
rawLines = do
  first <- anyLine
  rest <- many rawLine
  return $ T.unlines (first:rest)

noteBlock :: PandocMonad m => MarkdownParser m (F Blocks)
noteBlock = do
  guardEnabled Ext_footnotes
  try $ do
     pos <- getPosition
     skipNonindentSpaces
     ref <- noteMarker
     char ':'
     optional blankline
     optional indentSpaces
     first <- rawLines
     rest <- many $ try $ blanklines >> indentSpaces >> rawLines
     let raw = T.unlines (first:rest) <> "\n"
     optional blanklines
     parsed <- parseFromString' parseBlocks raw
     oldnotes <- stateNotes' <$> getState
     case M.lookup ref oldnotes of
       Just _  -> logMessage $ DuplicateNoteReference ref pos
       Nothing -> return ()
     updateState $ \s -> s { stateNotes' =
       M.insert ref (pos, parsed) oldnotes }
     return mempty

--
-- parsing blocks
--

parseBlocks :: PandocMonad m => MarkdownParser m (F Blocks)
parseBlocks = mconcat <$> manyTill block eof

block :: PandocMonad m => MarkdownParser m (F Blocks)
block = do
  res <- choice [ mempty <$ blanklines
               , codeBlockFenced
               , yamlMetaBlock
               -- note: bulletList needs to be before header because of
               -- the possibility of empty list items: -
               , bulletList
               , divHtml
               , divFenced
               , header
               , lhsCodeBlock
               , htmlBlock
               , table
               , codeBlockIndented
               , rawTeXBlock
               , lineBlock
               , blockQuote
               , hrule
               , orderedList
               , definitionList
               , noteBlock
               , referenceKey
               , abbrevKey
               , para
               , plain
               ] <?> "block"
  trace (T.take 60 $ tshow $ B.toList $ runF res defaultParserState)
  return res

--
-- header blocks
--

header :: PandocMonad m => MarkdownParser m (F Blocks)
header = setextHeader <|> atxHeader <?> "header"

atxChar :: PandocMonad m => MarkdownParser m Char
atxChar = do
  exts <- getOption readerExtensions
  return $ if extensionEnabled Ext_literate_haskell exts
              then '='
              else '#'

atxHeader :: PandocMonad m => MarkdownParser m (F Blocks)
atxHeader = try $ do
  level <- fmap length (atxChar >>= many1 . char)
  notFollowedBy $ guardEnabled Ext_fancy_lists >>
                  (char '.' <|> char ')') -- this would be a list
  guardDisabled Ext_space_in_atx_header <|> notFollowedBy nonspaceChar
  skipSpaces
  (text, raw) <- withRaw $ do
    oldAllowLineBreaks <- stateAllowLineBreaks <$> getState
    updateState $ \st -> st{ stateAllowLineBreaks = False }
    res <- trimInlinesF . mconcat <$>
               many (notFollowedBy atxClosing >> inline)
    updateState $ \st -> st{ stateAllowLineBreaks = oldAllowLineBreaks }
    return res
  attr <- atxClosing
  attr' <- registerHeader attr (runF text defaultParserState)
  guardDisabled Ext_implicit_header_references
    <|> registerImplicitHeader raw attr'
  return $ B.headerWith attr' level <$> text

atxClosing :: PandocMonad m => MarkdownParser m Attr
atxClosing = try $ do
  attr' <- option nullAttr
             (guardEnabled Ext_mmd_header_identifiers >> mmdHeaderIdentifier)
  skipMany . char =<< atxChar
  skipSpaces
  attr <- option attr'
             (guardEnabled Ext_header_attributes >> attributes)
  blanklines
  return attr

setextHeaderEnd :: PandocMonad m => MarkdownParser m Attr
setextHeaderEnd = try $ do
  attr <- option nullAttr
          $ (guardEnabled Ext_mmd_header_identifiers >> mmdHeaderIdentifier)
           <|> (guardEnabled Ext_header_attributes >> attributes)
  blanklines
  return attr

mmdHeaderIdentifier :: PandocMonad m => MarkdownParser m Attr
mmdHeaderIdentifier = do
  (_, raw) <- reference
  let raw' = trim $ stripFirstAndLast raw
  let ident = T.concat $ T.words $ T.toLower raw'
  let attr = (ident, [], [])
  guardDisabled Ext_implicit_header_references
    <|> registerImplicitHeader raw' attr
  skipSpaces
  return attr

setextHeader :: PandocMonad m => MarkdownParser m (F Blocks)
setextHeader = try $ do
  -- This lookahead prevents us from wasting time parsing Inlines
  -- unless necessary -- it gives a significant performance boost.
  lookAhead $ anyLine >> many1 (oneOf setextHChars) >> blankline
  skipSpaces
  (text, raw) <- withRaw $ do
    oldAllowLineBreaks <- stateAllowLineBreaks <$> getState
    updateState $ \st -> st{ stateAllowLineBreaks = False }
    res <- trimInlinesF . mconcat <$>
               many (notFollowedBy setextHeaderEnd >> inline)
    updateState $ \st -> st{ stateAllowLineBreaks = oldAllowLineBreaks }
    return res
  attr <- setextHeaderEnd
  underlineChar <- oneOf setextHChars
  many (char underlineChar)
  blanklines
  let level = fromMaybe 0 (elemIndex underlineChar setextHChars) + 1
  attr' <- registerHeader attr (runF text defaultParserState)
  guardDisabled Ext_implicit_header_references
    <|> registerImplicitHeader raw attr'
  return $ B.headerWith attr' level <$> text

registerImplicitHeader :: PandocMonad m => Text -> Attr -> MarkdownParser m ()
registerImplicitHeader raw attr@(ident, _, _)
  | T.null raw = return ()
  | otherwise = do
      let key = toKey $ "[" <> raw <> "]"
      updateState $ \s ->
        s { stateHeaderKeys = M.insert key (("#" <> ident,""), attr)
                                     (stateHeaderKeys s) }

--
-- hrule block
--

hrule :: PandocMonad m => ParserT Text st m (F Blocks)
hrule = try $ do
  skipSpaces
  start <- satisfy isHruleChar
  count 2 (skipSpaces >> char start)
  skipMany (spaceChar <|> char start)
  newline
  optional blanklines
  return $ return B.horizontalRule

--
-- code blocks
--

indentedLine :: PandocMonad m => MarkdownParser m Text
indentedLine = indentSpaces >> anyLineNewline

blockDelimiter :: PandocMonad m
               => (Char -> Bool)
               -> Maybe Int
               -> ParserT Text ParserState m Int
blockDelimiter f len = try $ do
  skipNonindentSpaces
  c <- lookAhead (satisfy f)
  case len of
      Just l  -> count l (char c) >> many (char c) >> return l
      Nothing -> fmap ((+ 3) . length) (count 3 (char c) >> many (char c))

attributes :: PandocMonad m => MarkdownParser m Attr
attributes = try $ do
  char '{'
  spnl
  attrs <- many (attribute <* spnl)
  char '}'
  return $ foldl (\x f -> f x) nullAttr attrs

attribute :: PandocMonad m => MarkdownParser m (Attr -> Attr)
attribute = identifierAttr <|> classAttr <|> keyValAttr <|> specialAttr

identifier :: PandocMonad m => MarkdownParser m Text
identifier = do
  first <- letter
  rest <- many $ alphaNum <|> oneOf "-_:."
  return $ T.pack (first:rest)

identifierAttr :: PandocMonad m => MarkdownParser m (Attr -> Attr)
identifierAttr = try $ do
  char '#'
  result <- identifier
  return $ \(_,cs,kvs) -> (result,cs,kvs)

classAttr :: PandocMonad m => MarkdownParser m (Attr -> Attr)
classAttr = try $ do
  char '.'
  result <- identifier
  return $ \(id',cs,kvs) -> (id',cs ++ [result],kvs)

keyValAttr :: PandocMonad m => MarkdownParser m (Attr -> Attr)
keyValAttr = try $ do
  key <- identifier
  char '='
  val <- T.pack <$> enclosed (char '"') (char '"') litChar
     <|> T.pack <$> enclosed (char '\'') (char '\'') litChar
     <|> ("" <$ try (string "\"\""))
     <|> ("" <$ try (string "''"))
     <|> manyChar (escapedChar' <|> noneOf " \t\n\r}")
  return $ \(id',cs,kvs) ->
    case key of
         "id"    -> (val,cs,kvs)
         "class" -> (id',cs ++ T.words val,kvs)
         _       -> (id',cs,kvs ++ [(key,val)])

specialAttr :: PandocMonad m => MarkdownParser m (Attr -> Attr)
specialAttr = do
  char '-'
  return $ \(id',cs,kvs) -> (id',cs ++ ["unnumbered"],kvs)

rawAttribute :: PandocMonad m => MarkdownParser m Text
rawAttribute = do
  char '{'
  skipMany spaceChar
  char '='
  format <- many1Char $ satisfy (\c -> isAlphaNum c || c `elem` ['-', '_'])
  skipMany spaceChar
  char '}'
  return format

codeBlockFenced :: PandocMonad m => MarkdownParser m (F Blocks)
codeBlockFenced = try $ do
  indentchars <- nonindentSpaces
  let indentLevel = T.length indentchars
  c <- try (guardEnabled Ext_fenced_code_blocks >> lookAhead (char '~'))
     <|> (guardEnabled Ext_backtick_code_blocks >> lookAhead (char '`'))
  size <- blockDelimiter (== c) Nothing
  skipMany spaceChar
  rawattr <-
     (Left <$> try (guardEnabled Ext_raw_attribute >> rawAttribute))
    <|>
     (Right <$> option ("",[],[])
         (try (guardEnabled Ext_fenced_code_attributes >> attributes)
          <|> ((\x -> ("",[toLanguageId x],[])) <$> many1Char nonspaceChar)))
  blankline
  contents <- T.intercalate "\n" <$>
                 manyTill (gobbleAtMostSpaces indentLevel >> anyLine)
                          (try $ do
                            blockDelimiter (== c) (Just size)
                            blanklines)
  return $ return $
    case rawattr of
          Left syn   -> B.rawBlock syn contents
          Right attr -> B.codeBlockWith attr contents

-- correctly handle github language identifiers
toLanguageId :: Text -> Text
toLanguageId = T.toLower . go
  where go "c++"         = "cpp"
        go "objective-c" = "objectivec"
        go x             = x

codeBlockIndented :: PandocMonad m => MarkdownParser m (F Blocks)
codeBlockIndented = do
  contents <- many1 (indentedLine <|>
                     try (do b <- blanklines
                             l <- indentedLine
                             return $ b <> l))
  optional blanklines
  classes <- getOption readerIndentedCodeClasses
  return $ return $ B.codeBlockWith ("", classes, []) $
           stripTrailingNewlines $ T.concat contents

lhsCodeBlock :: PandocMonad m => MarkdownParser m (F Blocks)
lhsCodeBlock = do
  guardEnabled Ext_literate_haskell
  (return . B.codeBlockWith ("",["haskell","literate"],[]) <$>
          (lhsCodeBlockBird <|> lhsCodeBlockLaTeX))
    <|> (return . B.codeBlockWith ("",["haskell"],[]) <$>
          lhsCodeBlockInverseBird)

lhsCodeBlockLaTeX :: PandocMonad m => MarkdownParser m Text
lhsCodeBlockLaTeX = try $ do
  string "\\begin{code}"
  manyTill spaceChar newline
  contents <- many1TillChar anyChar (try $ string "\\end{code}")
  blanklines
  return $ stripTrailingNewlines contents

lhsCodeBlockBird :: PandocMonad m => MarkdownParser m Text
lhsCodeBlockBird = lhsCodeBlockBirdWith '>'

lhsCodeBlockInverseBird :: PandocMonad m => MarkdownParser m Text
lhsCodeBlockInverseBird = lhsCodeBlockBirdWith '<'

lhsCodeBlockBirdWith :: PandocMonad m => Char -> MarkdownParser m Text
lhsCodeBlockBirdWith c = try $ do
  pos <- getPosition
  when (sourceColumn pos /= 1) $ Prelude.fail "Not in first column"
  lns <- many1 $ birdTrackLine c
  -- if (as is normal) there is always a space after >, drop it
  let lns' = if all (\ln -> T.null ln || T.take 1 ln == " ") lns
                then map (T.drop 1) lns
                else lns
  blanklines
  return $ T.intercalate "\n" lns'

birdTrackLine :: PandocMonad m => Char -> ParserT Text st m Text
birdTrackLine c = try $ do
  char c
  -- allow html tags on left margin:
  when (c == '<') $ notFollowedBy letter
  anyLine

--
-- block quotes
--

emailBlockQuoteStart :: PandocMonad m => MarkdownParser m Char
emailBlockQuoteStart = try $ skipNonindentSpaces >> char '>' <* optional (char ' ')

emailBlockQuote :: PandocMonad m => MarkdownParser m [Text]
emailBlockQuote = try $ do
  emailBlockQuoteStart
  let emailLine = manyChar $ nonEndline <|> try
                              (endline >> notFollowedBy emailBlockQuoteStart >>
                               return '\n')
  let emailSep = try (newline >> emailBlockQuoteStart)
  first <- emailLine
  rest <- many $ try $ emailSep >> emailLine
  let raw = first:rest
  newline <|> (eof >> return '\n')
  optional blanklines
  return raw

blockQuote :: PandocMonad m => MarkdownParser m (F Blocks)
blockQuote = do
  raw <- emailBlockQuote
  -- parse the extracted block, which may contain various block elements:
  contents <- parseFromString' parseBlocks $ T.intercalate "\n" raw <> "\n\n"
  return $ B.blockQuote <$> contents

--
-- list blocks
--

bulletListStart :: PandocMonad m => MarkdownParser m ()
bulletListStart = try $ do
  optional newline -- if preceded by a Plain block in a list context
  skipNonindentSpaces
  notFollowedBy' (() <$ hrule)     -- because hrules start out just like lists
  satisfy isBulletListMarker
  gobbleSpaces 1 <|> () <$ lookAhead newline
  try (gobbleAtMostSpaces 3 >> notFollowedBy spaceChar) <|> return ()

orderedListStart :: PandocMonad m
                 => Maybe (ListNumberStyle, ListNumberDelim)
                 -> MarkdownParser m (Int, ListNumberStyle, ListNumberDelim)
orderedListStart mbstydelim = try $ do
  optional newline -- if preceded by a Plain block in a list context
  skipNonindentSpaces
  notFollowedBy $ string "p." >> spaceChar >> digit  -- page number
  (do guardDisabled Ext_fancy_lists
      start <- many1Char digit >>= safeRead
      char '.'
      gobbleSpaces 1 <|> () <$ lookAhead newline
      optional $ try (gobbleAtMostSpaces 3 >> notFollowedBy spaceChar)
      return (start, DefaultStyle, DefaultDelim))
   <|>
   (do (num, style, delim) <- maybe
          anyOrderedListMarker
          (\(sty,delim) -> (\start -> (start,sty,delim)) <$>
               orderedListMarker sty delim)
          mbstydelim
       gobbleSpaces 1 <|> () <$ lookAhead newline
       -- if it could be an abbreviated first name,
       -- insist on more than one space
       when (delim == Period && (style == UpperAlpha ||
            (style == UpperRoman &&
             num `elem` [1, 5, 10, 50, 100, 500, 1000]))) $
              () <$ lookAhead (newline <|> spaceChar)
       optional $ try (gobbleAtMostSpaces 3 >> notFollowedBy spaceChar)
       return (num, style, delim))

listStart :: PandocMonad m => MarkdownParser m ()
listStart = bulletListStart <|> Control.Monad.void (orderedListStart Nothing)

listLine :: PandocMonad m => Int -> MarkdownParser m Text
listLine continuationIndent = try $ do
  notFollowedBy' (do gobbleSpaces continuationIndent
                     skipMany spaceChar
                     listStart)
  notFollowedByHtmlCloser
  notFollowedByDivCloser
  optional (() <$ gobbleSpaces continuationIndent)
  listLineCommon

listLineCommon :: PandocMonad m => MarkdownParser m Text
listLineCommon = T.concat <$> manyTill
              (  many1Char (satisfy $ \c -> c `notElem` ['\n', '<', '`'])
             <|> fmap snd (withRaw code)
             <|> fmap snd (htmlTag isCommentTag)
             <|> countChar 1 anyChar
              ) newline

-- parse raw text for one list item, excluding start marker and continuations
rawListItem :: PandocMonad m
            => Bool -- four space rule
            -> MarkdownParser m a
            -> MarkdownParser m (Text, Int)
rawListItem fourSpaceRule start = try $ do
  pos1 <- getPosition
  start
  pos2 <- getPosition
  let continuationIndent = if fourSpaceRule
                              then 4
                              else sourceColumn pos2 - sourceColumn pos1
  first <- listLineCommon
  rest <- many (do notFollowedBy listStart
                   notFollowedBy (() <$ codeBlockFenced)
                   notFollowedBy blankline
                   listLine continuationIndent)
  blanks <- manyChar blankline
  let result = T.unlines (first:rest) <> blanks
  return (result, continuationIndent)

-- continuation of a list item - indented and separated by blankline
-- or (in compact lists) endline.
-- note: nested lists are parsed as continuations
listContinuation :: PandocMonad m => Int -> MarkdownParser m Text
listContinuation continuationIndent = try $ do
  x <- try $ do
         notFollowedBy blankline
         notFollowedByHtmlCloser
         notFollowedByDivCloser
         gobbleSpaces continuationIndent
         anyLineNewline
  xs <- many $ try $ do
         notFollowedBy blankline
         notFollowedByHtmlCloser
         notFollowedByDivCloser
         gobbleSpaces continuationIndent <|> notFollowedBy' listStart
         anyLineNewline
  blanks <- manyChar blankline
  return $ T.concat (x:xs) <> blanks

-- Variant of blanklines that doesn't require blank lines
-- before a fence or eof.
blanklines' :: PandocMonad m => MarkdownParser m Text
blanklines' = blanklines <|> try checkDivCloser
  where checkDivCloser = do
          guardEnabled Ext_fenced_divs
          divLevel <- stateFencedDivLevel <$> getState
          guard (divLevel >= 1)
          lookAhead divFenceEnd
          return ""

notFollowedByDivCloser :: PandocMonad m => MarkdownParser m ()
notFollowedByDivCloser =
  guardDisabled Ext_fenced_divs <|>
  do divLevel <- stateFencedDivLevel <$> getState
     guard (divLevel < 1) <|> notFollowedBy divFenceEnd

notFollowedByHtmlCloser :: PandocMonad m => MarkdownParser m ()
notFollowedByHtmlCloser = do
  inHtmlBlock <- stateInHtmlBlock <$> getState
  case inHtmlBlock of
        Just t  -> notFollowedBy' $ htmlTag (~== TagClose t)
        Nothing -> return ()

listItem :: PandocMonad m
         => Bool -- four-space rule
         -> MarkdownParser m a
         -> MarkdownParser m (F Blocks)
listItem fourSpaceRule start = try $ do
  -- parsing with ListItemState forces markers at beginning of lines to
  -- count as list item markers, even if not separated by blank space.
  -- see definition of "endline"
  state <- getState
  let oldContext = stateParserContext state
  setState $ state {stateParserContext = ListItemState}
  (first, continuationIndent) <- rawListItem fourSpaceRule start
  continuations <- many (listContinuation continuationIndent)
  -- parse the extracted block, which may contain various block elements:
  let raw = T.concat (first:continuations)
  contents <- parseFromString' parseBlocks raw
  updateState (\st -> st {stateParserContext = oldContext})
  exts <- getOption readerExtensions
  return $ B.fromList . taskListItemFromAscii exts . B.toList <$> contents

orderedList :: PandocMonad m => MarkdownParser m (F Blocks)
orderedList = try $ do
  (start, style, delim) <- lookAhead (orderedListStart Nothing)
  unless (style `elem` [DefaultStyle, Decimal, Example] &&
          delim `elem` [DefaultDelim, Period]) $
    guardEnabled Ext_fancy_lists
  when (style == Example) $ guardEnabled Ext_example_lists
  fourSpaceRule <- (True <$ guardEnabled Ext_four_space_rule)
               <|> return (style == Example)
  items <- fmap sequence $ many1 $ listItem fourSpaceRule
                 (orderedListStart (Just (style, delim)))
  start' <- if style == Example
               then return start
               else (start <$ guardEnabled Ext_startnum) <|> return 1
  return $ B.orderedListWith (start', style, delim) <$> fmap compactify items

bulletList :: PandocMonad m => MarkdownParser m (F Blocks)
bulletList = do
  fourSpaceRule <- (True <$ guardEnabled Ext_four_space_rule)
               <|> return False
  items <- fmap sequence $ many1 $ listItem fourSpaceRule bulletListStart
  return $ B.bulletList <$> fmap compactify items

-- definition lists

defListMarker :: PandocMonad m => MarkdownParser m ()
defListMarker = do
  sps <- nonindentSpaces
  char ':' <|> char '~'
  tabStop <- getOption readerTabStop
  let remaining = tabStop - (T.length sps + 1)
  if remaining > 0
     then try (count remaining (char ' ')) <|> string "\t" <|> many1 spaceChar
     else mzero
  return ()

definitionListItem :: PandocMonad m => Bool -> MarkdownParser m (F (Inlines, [Blocks]))
definitionListItem compact = try $ do
  rawLine' <- anyLine
  raw <- many1 $ defRawBlock compact
  term <- parseFromString' (trimInlinesF <$> inlines) rawLine'
  contents <- mapM (parseFromString' parseBlocks . (<> "\n")) raw
  optional blanklines
  return $ liftM2 (,) term (sequence contents)

defRawBlock :: PandocMonad m => Bool -> MarkdownParser m Text
defRawBlock compact = try $ do
  hasBlank <- option False $ blankline >> return True
  defListMarker
  firstline <- anyLineNewline
  let dline = try
               ( do notFollowedBy blankline
                    notFollowedByHtmlCloser
                    notFollowedByDivCloser
                    if compact -- laziness not compatible with compact
                       then () <$ indentSpaces
                       else (() <$ indentSpaces)
                             <|> notFollowedBy defListMarker
                    anyLine )
  rawlines <- many dline
  cont <- fmap T.concat $ many $ try $ do
            trailing <- option "" blanklines
            ln <- indentSpaces >> notFollowedBy blankline >> anyLine
            lns <- many dline
            return $ trailing <> T.unlines (ln:lns)
  return $ trimr (firstline <> T.unlines rawlines <> cont) <>
            if hasBlank || not (T.null cont) then "\n\n" else ""

definitionList :: PandocMonad m => MarkdownParser m (F Blocks)
definitionList = try $ do
  lookAhead (anyLine >>
             optional (blankline >> notFollowedBy (Control.Monad.void table)) >>
             -- don't capture table caption as def list!
             defListMarker)
  compactDefinitionList <|> normalDefinitionList

compactDefinitionList :: PandocMonad m => MarkdownParser m (F Blocks)
compactDefinitionList = do
  guardEnabled Ext_compact_definition_lists
  items <- fmap sequence $ many1 $ definitionListItem True
  return $ B.definitionList <$> fmap compactifyDL items

normalDefinitionList :: PandocMonad m => MarkdownParser m (F Blocks)
normalDefinitionList = do
  guardEnabled Ext_definition_lists
  items <- fmap sequence $ many1 $ definitionListItem False
  return $ B.definitionList <$> items

--
-- paragraph block
--

para :: PandocMonad m => MarkdownParser m (F Blocks)
para = try $ do
  exts <- getOption readerExtensions
  let implicitFigures x
       | extensionEnabled Ext_implicit_figures exts = do
         x' <- x
         case B.toList x' of
               [Image attr alt (src,tit)]
                 | not (null alt) ->
                    -- the fig: at beginning of title indicates a figure
                    return $ B.singleton
                           $ Image attr alt (src, "fig:" <> tit)
               _ -> return x'
       | otherwise = x
  result <- implicitFigures . trimInlinesF <$> inlines1
  option (B.plain <$> result)
    $ try $ do
            newline
            (blanklines >> return mempty)
              <|> (guardDisabled Ext_blank_before_blockquote >> () <$ lookAhead blockQuote)
              <|> (guardEnabled Ext_backtick_code_blocks >> () <$ lookAhead codeBlockFenced)
              <|> (guardDisabled Ext_blank_before_header >> () <$ lookAhead header)
              <|> (guardEnabled Ext_lists_without_preceding_blankline >>
                       -- Avoid creating a paragraph in a nested list.
                       notFollowedBy' inList >>
                       () <$ lookAhead listStart)
              <|> do guardEnabled Ext_native_divs
                     inHtmlBlock <- stateInHtmlBlock <$> getState
                     case inHtmlBlock of
                          Just "div" -> () <$
                                       lookAhead (htmlTag (~== TagClose ("div" :: Text)))
                          _          -> mzero
              <|> do guardEnabled Ext_fenced_divs
                     divLevel <- stateFencedDivLevel <$> getState
                     if divLevel > 0
                        then lookAhead divFenceEnd
                        else mzero
            return $ B.para <$> result

plain :: PandocMonad m => MarkdownParser m (F Blocks)
plain = fmap B.plain . trimInlinesF <$> inlines1

--
-- raw html
--

htmlElement :: PandocMonad m => MarkdownParser m Text
htmlElement = rawVerbatimBlock
          <|> strictHtmlBlock
          <|> fmap snd (htmlTag isBlockTag)

htmlBlock :: PandocMonad m => MarkdownParser m (F Blocks)
htmlBlock = do
  guardEnabled Ext_raw_html
  try (do
      (TagOpen _ attrs) <- lookAhead $ fst <$> htmlTag isBlockTag
      (return . B.rawBlock "html") <$> rawVerbatimBlock
        <|> (do guardEnabled Ext_markdown_attribute
                oldMarkdownAttribute <- stateMarkdownAttribute <$> getState
                markdownAttribute <-
                   case lookup "markdown" attrs of
                        Just "0" -> False <$ updateState (\st -> st{
                                       stateMarkdownAttribute = False })
                        Just _   -> True <$ updateState (\st -> st{
                                       stateMarkdownAttribute = True })
                        Nothing  -> return oldMarkdownAttribute
                res <- if markdownAttribute
                          then rawHtmlBlocks
                          else htmlBlock'
                updateState $ \st -> st{ stateMarkdownAttribute =
                                         oldMarkdownAttribute }
                return res)
        <|> (guardEnabled Ext_markdown_in_html_blocks >> rawHtmlBlocks))
    <|> htmlBlock'

htmlBlock' :: PandocMonad m => MarkdownParser m (F Blocks)
htmlBlock' = try $ do
    first <- htmlElement
    skipMany spaceChar
    optional blanklines
    return $ if T.null first
                then mempty
                else return $ B.rawBlock "html" first

strictHtmlBlock :: PandocMonad m => MarkdownParser m Text
strictHtmlBlock = htmlInBalanced (not . isInlineTag)

rawVerbatimBlock :: PandocMonad m => MarkdownParser m Text
rawVerbatimBlock = htmlInBalanced isVerbTag
  where isVerbTag (TagOpen "pre" _)      = True
        isVerbTag (TagOpen "style" _)    = True
        isVerbTag (TagOpen "script" _)   = True
        isVerbTag (TagOpen "textarea" _) = True
        isVerbTag _                      = False

rawTeXBlock :: PandocMonad m => MarkdownParser m (F Blocks)
rawTeXBlock = do
  guardEnabled Ext_raw_tex
  result <- (B.rawBlock "tex" . trim . T.concat <$>
                many1 ((<>) <$> rawConTeXtEnvironment <*> spnl'))
          <|> (B.rawBlock "tex" . trim . T.concat <$>
                many1 ((<>) <$> rawLaTeXBlock <*> spnl'))
  return $ case B.toList result of
                [RawBlock _ cs]
                  | T.all (`elem` [' ','\t','\n']) cs -> return mempty
                -- don't create a raw block for suppressed macro defs
                _ -> return result

rawHtmlBlocks :: PandocMonad m => MarkdownParser m (F Blocks)
rawHtmlBlocks = do
  (TagOpen tagtype _, raw) <- htmlTag isBlockTag
  -- we don't want '<td>    text' to be a code block:
  skipMany spaceChar
  indentlevel <- (blankline >> length <$> many (char ' ')) <|> return 0
  -- try to find closing tag
  -- we set stateInHtmlBlock so that closing tags that can be either block or
  -- inline will not be parsed as inline tags
  oldInHtmlBlock <- stateInHtmlBlock <$> getState
  updateState $ \st -> st{ stateInHtmlBlock = Just tagtype }
  let closer = htmlTag (~== TagClose tagtype)
  let block' = try $ do
                 gobbleAtMostSpaces indentlevel
                 notFollowedBy' closer
                 block
  contents <- mconcat <$> many block'
  result <-
    try
    (do gobbleAtMostSpaces indentlevel
        (_, rawcloser) <- closer
        return (return (B.rawBlock "html" $ stripMarkdownAttribute raw) <>
                contents <>
                return (B.rawBlock "html" rawcloser)))
      <|> return (return (B.rawBlock "html" raw) <> contents)
  updateState $ \st -> st{ stateInHtmlBlock = oldInHtmlBlock }
  return result

-- remove markdown="1" attribute
stripMarkdownAttribute :: Text -> Text
stripMarkdownAttribute s = renderTags' $ map filterAttrib $ parseTags s
  where filterAttrib (TagOpen t as) = TagOpen t
                                        [(k,v) | (k,v) <- as, k /= "markdown"]
        filterAttrib              x = x

--
-- line block
--

lineBlock :: PandocMonad m => MarkdownParser m (F Blocks)
lineBlock = try $ do
  guardEnabled Ext_line_blocks
  lines' <- lineBlockLines >>=
            mapM (parseFromString' (trimInlinesF <$> inlines))
  return $ B.lineBlock <$> sequence lines'

--
-- Tables
--

-- Parse a dashed line with optional trailing spaces; return its length
-- and the length including trailing space.
dashedLine :: PandocMonad m
           => Char
           -> ParserT Text st m (Int, Int)
dashedLine ch = do
  dashes <- many1 (char ch)
  sp     <- many spaceChar
  let lengthDashes = length dashes
      lengthSp     = length sp
  return (lengthDashes, lengthDashes + lengthSp)

-- Parse a table header with dashed lines of '-' preceded by
-- one (or zero) line of text.
simpleTableHeader :: PandocMonad m
                  => Bool  -- ^ Headerless table
                  -> MarkdownParser m (F [Blocks], [Alignment], [Int])
simpleTableHeader headless = try $ do
  rawContent  <- if headless
                    then return ""
                    else anyLine
  initSp      <- nonindentSpaces
  dashes      <- many1 (dashedLine '-')
  newline
  let (lengths, lines') = unzip dashes
  let indices  = scanl (+) (T.length initSp) lines'
  -- If no header, calculate alignment on basis of first row of text
  rawHeads <- fmap (tail . splitTextByIndices (init indices)) $
              if headless
                 then lookAhead anyLine
                 else return rawContent
  let aligns   = zipWith alignType (map (: []) rawHeads) lengths
  let rawHeads' = if headless
                     then []
                     else rawHeads
  heads <- fmap sequence
           $
            mapM (parseFromString' (mconcat <$> many plain).trim) rawHeads'
  return (heads, aligns, indices)

-- Returns an alignment type for a table, based on a list of strings
-- (the rows of the column header) and a number (the length of the
-- dashed line under the rows.
alignType :: [Text]
          -> Int
          -> Alignment
alignType [] _ = AlignDefault
alignType strLst len =
  let nonempties = filter (not . T.null) $ map trimr strLst
      (leftSpace, rightSpace) =
           case sortOn T.length nonempties of
                 (x:_) -> (T.head x `elem` [' ', '\t'], T.length x < len)
                 []    -> (False, False)
  in  case (leftSpace, rightSpace) of
        (True,  False) -> AlignRight
        (False, True)  -> AlignLeft
        (True,  True)  -> AlignCenter
        (False, False) -> AlignDefault

-- Parse a table footer - dashed lines followed by blank line.
tableFooter :: PandocMonad m => MarkdownParser m Text
tableFooter = try $ skipNonindentSpaces >> many1 (dashedLine '-') >> blanklines'

-- Parse a table separator - dashed line.
tableSep :: PandocMonad m => MarkdownParser m Char
tableSep = try $ skipNonindentSpaces >> many1 (dashedLine '-') >> char '\n'

-- Parse a raw line and split it into chunks by indices.
rawTableLine :: PandocMonad m
             => [Int]
             -> MarkdownParser m [Text]
rawTableLine indices = do
  notFollowedBy' (blanklines' <|> tableFooter)
  line <- take1WhileP (/='\n') <* newline
  return $ map trim $ tail $
           splitTextByIndices (init indices) line

-- Parse a table line and return a list of lists of blocks (columns).
tableLine :: PandocMonad m
          => [Int]
          -> MarkdownParser m (F [Blocks])
tableLine indices = rawTableLine indices >>=
  fmap sequence . mapM (parseFromString' (mconcat <$> many plain))

-- Parse a multiline table row and return a list of blocks (columns).
multilineRow :: PandocMonad m
             => [Int]
             -> MarkdownParser m (F [Blocks])
multilineRow indices = do
  colLines <- many1 (rawTableLine indices)
  let cols = map T.unlines $ transpose colLines
  fmap sequence $ mapM (parseFromString' (mconcat <$> many plain)) cols

-- Parses a table caption:  inlines beginning with 'Table:'
-- and followed by blank lines.
tableCaption :: PandocMonad m => MarkdownParser m (F Inlines)
tableCaption = try $ do
  guardEnabled Ext_table_captions
  skipNonindentSpaces
  (string ":" <* notFollowedBy (satisfy isPunctuation)) <|> string "Table:"
  trimInlinesF <$> inlines1 <* blanklines

-- Parse a simple table with '---' header and one line per row.
simpleTable :: PandocMonad m
            => Bool  -- ^ Headerless table
            -> MarkdownParser m ([Alignment], [Double], F [Row], F [Row])
simpleTable headless = do
  (aligns, _widths, heads', lines') <-
       tableWith (simpleTableHeader headless) tableLine
              (return ())
              (if headless then tableFooter else tableFooter <|> blanklines')
  -- Simple tables get 0s for relative column widths (i.e., use default)
  return (aligns, replicate (length aligns) 0, heads', lines')

-- Parse a multiline table:  starts with row of '-' on top, then header
-- (which may be multiline), then the rows,
-- which may be multiline, separated by blank lines, and
-- ending with a footer (dashed line followed by blank line).
multilineTable :: PandocMonad m
               => Bool -- ^ Headerless table
               -> MarkdownParser m ([Alignment], [Double], F [Row], F [Row])
multilineTable headless =
  tableWith (multilineTableHeader headless) multilineRow blanklines tableFooter

multilineTableHeader :: PandocMonad m
                     => Bool -- ^ Headerless table
                     -> MarkdownParser m (F [Blocks], [Alignment], [Int])
multilineTableHeader headless = try $ do
  unless headless $
     tableSep >> notFollowedBy blankline
  rawContent  <- if headless
                    then return $ repeat ""
                    else many1 $ notFollowedBy tableSep >> anyLine
  initSp      <- nonindentSpaces
  dashes      <- many1 (dashedLine '-')
  newline
  let (lengths, lines') = unzip dashes
  let indices  = scanl (+) (T.length initSp) lines'
  -- compensate for the fact that intercolumn spaces are
  -- not included in the last index:
  let indices' = case reverse indices of
                      []     -> []
                      (x:xs) -> reverse (x+1:xs)
  rawHeadsList <- if headless
                     then fmap (map (:[]) . tail .
                              splitTextByIndices (init indices')) $ lookAhead anyLine
                     else return $ transpose $ map
                           (tail . splitTextByIndices (init indices'))
                           rawContent
  let aligns   = zipWith alignType rawHeadsList lengths
  let rawHeads = if headless
                    then []
                    else map (T.unlines . map trim) rawHeadsList
  heads <- fmap sequence $
            mapM (parseFromString' (mconcat <$> many plain).trim) rawHeads
  return (heads, aligns, indices')

-- Parse a grid table:  starts with row of '-' on top, then header
-- (which may be grid), then the rows,
-- which may be grid, separated by blank lines, and
-- ending with a footer (dashed line followed by blank line).
gridTable :: PandocMonad m => Bool -- ^ Headerless table
          -> MarkdownParser m ([Alignment], [Double], F [Row], F [Row])
gridTable headless = gridTableWith' parseBlocks headless

pipeBreak :: PandocMonad m => MarkdownParser m ([Alignment], [Int])
pipeBreak = try $ do
  nonindentSpaces
  openPipe <- (True <$ char '|') <|> return False
  first <- pipeTableHeaderPart
  rest <- many $ sepPipe *> pipeTableHeaderPart
  -- surrounding pipes needed for a one-column table:
  guard $ not (null rest && not openPipe)
  optional (char '|')
  blankline
  return $ unzip (first:rest)

pipeTable :: PandocMonad m => MarkdownParser m ([Alignment], [Double], F [Row], F [Row])
pipeTable = try $ do
  nonindentSpaces
  lookAhead nonspaceChar
  (heads,(aligns, seplengths)) <- (,) <$> pipeTableRow <*> pipeBreak
  let heads' = take (length aligns) <$> heads
  lines' <- many pipeTableRow
  let lines'' = map (take (length aligns) <$>) lines'
  let maxlength = maximum $
       map (\x -> T.length . stringify $ runF x def) (heads' : lines'')
  numColumns <- getOption readerColumns
  let widths = if maxlength > numColumns
                  then map (\len ->
                         fromIntegral len / fromIntegral (sum seplengths))
                         seplengths
                  else replicate (length aligns) 0.0
  return (aligns, widths, toHeaderRow <$> heads', map toRow <$> sequence lines'')

sepPipe :: PandocMonad m => MarkdownParser m ()
sepPipe = try $ do
  char '|' <|> char '+'
  notFollowedBy blankline

-- parse a row, also returning probable alignments for org-table cells
pipeTableRow :: PandocMonad m => MarkdownParser m (F [Blocks])
pipeTableRow = try $ do
  scanForPipe
  skipMany spaceChar
  openPipe <- (True <$ char '|') <|> return False
  -- split into cells
  let chunk = void (code <|> math <|> rawHtmlInline <|> escapedChar <|> rawLaTeXInline')
       <|> void (noneOf "|\n\r")
  let cellContents = withRaw (many chunk) >>=
        parseFromString' pipeTableCell . trim . snd
  cells <- cellContents `sepEndBy1` char '|'
  -- surrounding pipes needed for a one-column table:
  guard $ not (length cells == 1 && not openPipe)
  blankline
  return $ sequence cells

pipeTableCell :: PandocMonad m => MarkdownParser m (F Blocks)
pipeTableCell =
  (do result <- inlines1
      return $ B.plain <$> result)
    <|> return mempty

pipeTableHeaderPart :: PandocMonad m => ParserT Text st m (Alignment, Int)
pipeTableHeaderPart = try $ do
  skipMany spaceChar
  left <- optionMaybe (char ':')
  pipe <- many1 (char '-')
  right <- optionMaybe (char ':')
  skipMany spaceChar
  let len = length pipe + maybe 0 (const 1) left + maybe 0 (const 1) right
  return
    (case (left,right) of
      (Nothing,Nothing) -> AlignDefault
      (Just _,Nothing)  -> AlignLeft
      (Nothing,Just _)  -> AlignRight
      (Just _,Just _)   -> AlignCenter, len)

-- Succeed only if current line contains a pipe.
scanForPipe :: PandocMonad m => ParserT Text st m ()
scanForPipe = do
  inp <- getInput
  case T.break (\c -> c == '\n' || c == '|') inp of
       (_, T.uncons -> Just ('|', _)) -> return ()
       _                              -> mzero

-- | Parse a table using 'headerParser', 'rowParser',
-- 'lineParser', and 'footerParser'.  Variant of the version in
-- Text.Pandoc.Parsing.
tableWith :: PandocMonad m
          => MarkdownParser m (F [Blocks], [Alignment], [Int])
          -> ([Int] -> MarkdownParser m (F [Blocks]))
          -> MarkdownParser m sep
          -> MarkdownParser m end
          -> MarkdownParser m ([Alignment], [Double], F [Row], F [Row])
tableWith headerParser rowParser lineParser footerParser = try $ do
    (heads, aligns, indices) <- headerParser
    lines' <- fmap sequence $ rowParser indices `sepEndBy1` lineParser
    footerParser
    numColumns <- getOption readerColumns
    let widths = if null indices
                    then replicate (length aligns) 0.0
                    else widthsFromIndices numColumns indices
    return (aligns, widths, toHeaderRow <$> heads, map toRow <$> lines')

table :: PandocMonad m => MarkdownParser m (F Blocks)
table = try $ do
  frontCaption <- option Nothing (Just <$> tableCaption)
  (aligns, widths, heads, lns) <-
         try (guardEnabled Ext_pipe_tables >> scanForPipe >> pipeTable) <|>
         try (guardEnabled Ext_multiline_tables >>
                multilineTable False) <|>
         try (guardEnabled Ext_simple_tables >>
                (simpleTable True <|> simpleTable False)) <|>
         try (guardEnabled Ext_multiline_tables >>
                multilineTable True) <|>
         try (guardEnabled Ext_grid_tables >>
                (gridTable False <|> gridTable True)) <?> "table"
  optional blanklines
  caption <- case frontCaption of
                  Nothing -> option (return mempty) tableCaption
                  Just c  -> return c
  -- renormalize widths if greater than 100%:
  let totalWidth = sum widths
  let widths' = if totalWidth < 1
                   then widths
                   else map (/ totalWidth) widths
  let strictPos w
        | w > 0     = ColWidth w
        | otherwise = ColWidthDefault
  return $ do
    caption' <- caption
    heads' <- heads
    lns' <- lns
    return $ B.table (B.simpleCaption $ B.plain caption')
                     (zip aligns (strictPos <$> widths'))
                     (TableHead nullAttr heads')
                     [TableBody nullAttr 0 [] lns']
                     (TableFoot nullAttr [])

--
-- inline
--

inlines :: PandocMonad m => MarkdownParser m (F Inlines)
inlines = mconcat <$> many inline

inlines1 :: PandocMonad m => MarkdownParser m (F Inlines)
inlines1 = mconcat <$> many1 inline

inline :: PandocMonad m => MarkdownParser m (F Inlines)
inline = choice [ whitespace
                , bareURL
                , str
                , endline
                , code
                , strongOrEmph
                , note
                , cite
                , bracketedSpan
                , link
                , image
                , math
                , strikeout
                , subscript
                , superscript
                , inlineNote  -- after superscript because of ^[link](/foo)^
                , autoLink
                , spanHtml
                , rawHtmlInline
                , escapedNewline
                , escapedChar
                , rawLaTeXInline'
                , exampleRef
                , smart
                , return . B.singleton <$> charRef
                , emoji
                , symbol
                , ltSign
                ] <?> "inline"

escapedChar' :: PandocMonad m => MarkdownParser m Char
escapedChar' = try $ do
  char '\\'
  (guardEnabled Ext_all_symbols_escapable >> satisfy (not . isAlphaNum))
     <|> (guardEnabled Ext_angle_brackets_escapable >>
            oneOf "\\`*_{}[]()>#+-.!~\"<>")
     <|> oneOf "\\`*_{}[]()>#+-.!~\""

escapedNewline :: PandocMonad m => MarkdownParser m (F Inlines)
escapedNewline = try $ do
  guardEnabled Ext_escaped_line_breaks
  char '\\'
  lookAhead (char '\n') -- don't consume the newline (see #3730)
  return $ return B.linebreak

escapedChar :: PandocMonad m => MarkdownParser m (F Inlines)
escapedChar = do
  result <- escapedChar'
  case result of
       ' ' -> return $ return $ B.str "\160" -- "\ " is a nonbreaking space
       _   -> return $ return $ B.str $ T.singleton result

ltSign :: PandocMonad m => MarkdownParser m (F Inlines)
ltSign = do
  guardDisabled Ext_raw_html
    <|> (notFollowedByHtmlCloser >> notFollowedBy' (htmlTag isBlockTag))
  char '<'
  return $ return $ B.str "<"

exampleRef :: PandocMonad m => MarkdownParser m (F Inlines)
exampleRef = try $ do
  guardEnabled Ext_example_lists
  char '@'
  lab <- many1Char (alphaNum <|> oneOf "-_")
  return $ do
    st <- askF
    return $ case M.lookup lab (stateExamples st) of
                  Just n  -> B.str $ tshow n
                  Nothing -> B.str $ "@" <> lab

symbol :: PandocMonad m => MarkdownParser m (F Inlines)
symbol = do
  result <- noneOf "<\\\n\t "
         <|> try (do lookAhead $ char '\\'
                     notFollowedBy' (() <$ rawTeXBlock)
                     char '\\')
  return $ return $ B.str $ T.singleton result

-- parses inline code, between n `s and n `s
code :: PandocMonad m => MarkdownParser m (F Inlines)
code = try $ do
  starts <- many1 (char '`')
  skipSpaces
  result <- trim . T.concat
        <$> manyTill
              (   many1Char (noneOf "`\n")
              <|> many1Char (char '`')
              <|> (char '\n'
                    >> notFollowedBy (inList >> listStart)
                    >> notFollowedBy' blankline
                    >> return " "))
              (try $ skipSpaces
                  >> count (length starts) (char '`')
                  >> notFollowedBy (char '`'))
  rawattr <-
     (Left <$> try (guardEnabled Ext_raw_attribute >> rawAttribute))
    <|>
     (Right <$> option ("",[],[])
         (try (guardEnabled Ext_inline_code_attributes >> attributes)))
  return $ return $
    case rawattr of
         Left syn   -> B.rawInline syn result
         Right attr -> B.codeWith attr result

math :: PandocMonad m => MarkdownParser m (F Inlines)
math =  (return . B.displayMath <$> (mathDisplay >>= applyMacros))
     <|> (return . B.math <$> (mathInline >>= applyMacros)) <+?>
               (guardEnabled Ext_smart *> (return <$> apostrophe)
                <* notFollowedBy (space <|> satisfy isPunctuation))

-- Parses material enclosed in *s, **s, _s, or __s.
-- Designed to avoid backtracking.
enclosure :: PandocMonad m
          => Char
          -> MarkdownParser m (F Inlines)
enclosure c = do
  -- we can't start an enclosure with _ if after a string and
  -- the intraword_underscores extension is enabled:
  guardDisabled Ext_intraword_underscores
    <|> guard (c == '*')
    <|> (guard =<< notAfterString)
  cs <- many1Char (char c)
  (return (B.str cs) <>) <$> whitespace
    <|>
        case T.length cs of
             3 -> three c
             2 -> two   c mempty
             1 -> one   c mempty
             _ -> return (return $ B.str cs)

ender :: PandocMonad m => Char -> Int -> MarkdownParser m ()
ender c n = try $ do
  count n (char c)
  guard (c == '*')
    <|> guardDisabled Ext_intraword_underscores
    <|> notFollowedBy alphaNum

-- Parse inlines til you hit one c or a sequence of two cs.
-- If one c, emit emph and then parse two.
-- If two cs, emit strong and then parse one.
-- Otherwise, emit ccc then the results.
three :: PandocMonad m => Char -> MarkdownParser m (F Inlines)
three c = do
  contents <- mconcat <$> many (notFollowedBy (ender c 1) >> inline)
  (ender c 3 >> updateLastStrPos >> return ((B.strong . B.emph) <$> contents))
    <|> (ender c 2 >> updateLastStrPos >> one c (B.strong <$> contents))
    <|> (ender c 1 >> updateLastStrPos >> two c (B.emph <$> contents))
    <|> return (return (B.str $ T.pack [c,c,c]) <> contents)

-- Parse inlines til you hit two c's, and emit strong.
-- If you never do hit two cs, emit ** plus inlines parsed.
two :: PandocMonad m => Char -> F Inlines -> MarkdownParser m (F Inlines)
two c prefix' = do
  contents <- mconcat <$> many (try $ notFollowedBy (ender c 2) >> inline)
  (ender c 2 >> updateLastStrPos >>
                return (B.strong <$> (prefix' <> contents)))
    <|> return (return (B.str $ T.pack [c,c]) <> (prefix' <> contents))

-- Parse inlines til you hit a c, and emit emph.
-- If you never hit a c, emit * plus inlines parsed.
one :: PandocMonad m => Char -> F Inlines -> MarkdownParser m (F Inlines)
one c prefix' = do
  contents <- mconcat <$> many (  (notFollowedBy (ender c 1) >> inline)
                           <|> try (string [c,c] >>
                                    notFollowedBy (ender c 1) >>
                                    two c mempty) )
  (ender c 1 >> updateLastStrPos >> return (B.emph <$> (prefix' <> contents)))
    <|> return (return (B.str $ T.singleton c) <> (prefix' <> contents))

strongOrEmph :: PandocMonad m => MarkdownParser m (F Inlines)
strongOrEmph =  enclosure '*' <|> enclosure '_'

-- | Parses a list of inlines between start and end delimiters.
inlinesBetween :: PandocMonad m
               => (Show b)
               => MarkdownParser m a
               -> MarkdownParser m b
               -> MarkdownParser m (F Inlines)
inlinesBetween start end =
  (trimInlinesF . mconcat) <$> try (start >> many1Till inner end)
    where inner      = innerSpace <|> (notFollowedBy' (() <$ whitespace) >> inline)
          innerSpace = try $ whitespace <* notFollowedBy' end

strikeout :: PandocMonad m => MarkdownParser m (F Inlines)
strikeout = fmap B.strikeout <$>
 (guardEnabled Ext_strikeout >> inlinesBetween strikeStart strikeEnd)
    where strikeStart = string "~~" >> lookAhead nonspaceChar
                        >> notFollowedBy (char '~')
          strikeEnd   = try $ string "~~"

superscript :: PandocMonad m => MarkdownParser m (F Inlines)
superscript = fmap B.superscript <$> try (do
  guardEnabled Ext_superscript
  char '^'
  mconcat <$> many1Till (do notFollowedBy spaceChar
                            notFollowedBy newline
                            inline) (char '^'))

subscript :: PandocMonad m => MarkdownParser m (F Inlines)
subscript = fmap B.subscript <$> try (do
  guardEnabled Ext_subscript
  char '~'
  mconcat <$> many1Till (do notFollowedBy spaceChar
                            notFollowedBy newline
                            inline) (char '~'))

whitespace :: PandocMonad m => MarkdownParser m (F Inlines)
whitespace = spaceChar >> return <$> (lb <|> regsp) <?> "whitespace"
  where lb = spaceChar >> skipMany spaceChar >> option B.space (endline >> return B.linebreak)
        regsp = skipMany spaceChar >> return B.space

nonEndline :: PandocMonad m => ParserT Text st m Char
nonEndline = satisfy (/='\n')

str :: PandocMonad m => MarkdownParser m (F Inlines)
str = do
  result <- mconcat <$> many1
             ( take1WhileP isAlphaNum
              <|> "." <$ try (char '.' <* notFollowedBy (char '.')) )
  updateLastStrPos
  (do guardEnabled Ext_smart
      abbrevs <- getOption readerAbbreviations
      if not (T.null result) && T.last result == '.' && result `Set.member` abbrevs
         then try (do ils <- whitespace
                      -- ?? lookAhead alphaNum
                      -- replace space after with nonbreaking space
                      -- if softbreak, move before abbrev if possible (#4635)
                      return $ do
                        ils' <- ils
                        case B.toList ils' of
                             [Space] ->
                                 return (B.str result <> B.str "\160")
                             _ -> return (B.str result <> ils'))
                <|> return (return (B.str result))
         else return (return (B.str result)))
     <|> return (return (B.str result))

-- an endline character that can be treated as a space, not a structural break
endline :: PandocMonad m => MarkdownParser m (F Inlines)
endline = try $ do
  newline
  notFollowedBy blankline
  getState >>= guard . stateAllowLineBreaks
  -- parse potential list-starts differently if in a list:
  notFollowedBy (inList >> listStart)
  guardDisabled Ext_lists_without_preceding_blankline <|> notFollowedBy listStart
  guardEnabled Ext_blank_before_blockquote <|> notFollowedBy emailBlockQuoteStart
  guardEnabled Ext_blank_before_header <|> (notFollowedBy . char =<< atxChar) -- atx header
  guardDisabled Ext_backtick_code_blocks <|>
     notFollowedBy (() <$ (lookAhead (char '`') >> codeBlockFenced))
  notFollowedByHtmlCloser
  notFollowedByDivCloser
  (eof >> return mempty)
    <|> (guardEnabled Ext_hard_line_breaks >> return (return B.linebreak))
    <|> (guardEnabled Ext_ignore_line_breaks >> return mempty)
    <|> (skipMany spaceChar >> return (return B.softbreak))

--
-- links
--

-- a reference label for a link
reference :: PandocMonad m => MarkdownParser m (F Inlines, Text)
reference = do
  guardDisabled Ext_footnotes <|> notFollowedBy' (string "[^")
  guardDisabled Ext_citations <|> notFollowedBy' (string "[@")
  withRaw $ trimInlinesF <$> inlinesInBalancedBrackets

parenthesizedChars :: PandocMonad m => MarkdownParser m Text
parenthesizedChars = do
  result <- charsInBalanced '(' ')' litChar
  return $ "(" <> result <> ")"

-- source for a link, with optional title
source :: PandocMonad m => MarkdownParser m (Text, Text)
source = do
  char '('
  skipSpaces
  let urlChunk =
            try parenthesizedChars
        <|> (notFollowedBy (oneOf " )") >> countChar 1 litChar)
        <|> try (many1Char spaceChar <* notFollowedBy (oneOf "\"')"))
  let sourceURL = (T.unwords . T.words . T.concat) <$> many urlChunk
  let betweenAngles = try $
         char '<' >> manyTillChar litChar (char '>')
  src <- try betweenAngles <|> sourceURL
  tit <- option "" $ try $ spnl >> linkTitle
  skipSpaces
  char ')'
  return (escapeURI $ trimr src, tit)

linkTitle :: PandocMonad m => MarkdownParser m Text
linkTitle = quotedTitle '"' <|> quotedTitle '\''

link :: PandocMonad m => MarkdownParser m (F Inlines)
link = try $ do
  st <- getState
  guard $ stateAllowLinks st
  setState $ st{ stateAllowLinks = False }
  (lab,raw) <- reference
  setState $ st{ stateAllowLinks = True }
  regLink B.linkWith lab <|> referenceLink B.linkWith (lab,raw)

bracketedSpan :: PandocMonad m => MarkdownParser m (F Inlines)
bracketedSpan = try $ do
  guardEnabled Ext_bracketed_spans
  (lab,_) <- reference
  attr <- attributes
  return $ if isSmallCaps attr
              then B.smallcaps <$> lab
              else if isUnderline attr
                      then B.underline <$> lab
                      else B.spanWith attr <$> lab

-- | We treat a span as SmallCaps if class is "smallcaps" (and
-- no other attributes are set or if style is "font-variant:small-caps"
-- (and no other attributes are set).
isSmallCaps :: Attr -> Bool
isSmallCaps ("",["smallcaps"],[]) = True
isSmallCaps ("",[],kvs) =
  case lookup "style" kvs of
       Just s -> T.toLower (T.filter (`notElem` [' ', '\t', ';']) s) ==
                    "font-variant:small-caps"
       Nothing -> False
isSmallCaps _ = False

-- | We treat a span as Underline if class is "ul" or
-- "underline" (and no other attributes are set).
isUnderline :: Attr -> Bool
isUnderline ("",["ul"],[]) = True
isUnderline ("",["underline"],[]) = True
isUnderline _ = False

regLink :: PandocMonad m
        => (Attr -> Text -> Text -> Inlines -> Inlines)
        -> F Inlines
        -> MarkdownParser m (F Inlines)
regLink constructor lab = try $ do
  (src, tit) <- source
  attr <- option nullAttr $
          guardEnabled Ext_link_attributes >> attributes
  return $ constructor attr src tit <$> lab

-- a link like [this][ref] or [this][] or [this]
referenceLink :: PandocMonad m
              => (Attr -> Text -> Text -> Inlines -> Inlines)
              -> (F Inlines, Text)
              -> MarkdownParser m (F Inlines)
referenceLink constructor (lab, raw) = do
  sp <- (True <$ lookAhead (char ' ')) <|> return False
  (_,raw') <- option (mempty, "") $
      lookAhead (try (do guardEnabled Ext_citations
                         guardDisabled Ext_spaced_reference_links <|> spnl
                         normalCite
                         return (mempty, "")))
      <|>
      try ((guardDisabled Ext_spaced_reference_links <|> spnl) >> reference)
  when (raw' == "") $ guardEnabled Ext_shortcut_reference_links
  let labIsRef = raw' == "" || raw' == "[]"
  let key = toKey $ if labIsRef then raw else raw'
  parsedRaw <- parseFromString' inlines raw'
  fallback  <- parseFromString' inlines $ dropBrackets raw
  implicitHeaderRefs <- option False $
                         True <$ guardEnabled Ext_implicit_header_references
  let makeFallback = do
       parsedRaw' <- parsedRaw
       fallback' <- fallback
       return $ B.str "[" <> fallback' <> B.str "]" <>
                (if sp && not (T.null raw) then B.space else mempty) <>
                parsedRaw'
  return $ do
    keys <- asksF stateKeys
    case M.lookup key keys of
       Nothing        ->
         if implicitHeaderRefs
            then do
              headerKeys <- asksF stateHeaderKeys
              case M.lookup key headerKeys of
                   Just ((src, tit), _) -> constructor nullAttr src tit <$> lab
                   Nothing              -> makeFallback
            else makeFallback
       Just ((src,tit), attr) -> constructor attr src tit <$> lab

dropBrackets :: Text -> Text
dropBrackets = dropRB . dropLB
  where dropRB (T.unsnoc -> Just (xs,']')) = xs
        dropRB xs                          = xs
        dropLB (T.uncons -> Just ('[',xs)) = xs
        dropLB xs                          = xs

bareURL :: PandocMonad m => MarkdownParser m (F Inlines)
bareURL = try $ do
  guardEnabled Ext_autolink_bare_uris
  getState >>= guard . stateAllowLinks
  (cls, (orig, src)) <- (("uri",) <$> uri) <|> (("email",) <$> emailAddress)
  notFollowedBy $ try $ spaces >> htmlTag (~== TagClose ("a" :: Text))
  return $ return $ B.linkWith ("",[cls],[]) src "" (B.str orig)

autoLink :: PandocMonad m => MarkdownParser m (F Inlines)
autoLink = try $ do
  getState >>= guard . stateAllowLinks
  char '<'
  (cls, (orig, src)) <- (("uri",) <$> uri) <|> (("email",) <$> emailAddress)
  -- in rare cases, something may remain after the uri parser
  -- is finished, because the uri parser tries to avoid parsing
  -- final punctuation.  for example:  in `<http://hi---there>`,
  -- the URI parser will stop before the dashes.
  extra <- fromEntities <$> manyTillChar nonspaceChar (char '>')
  attr  <- option ("", [cls], []) $ try $
            guardEnabled Ext_link_attributes >> attributes
  return $ return $ B.linkWith attr (src <> escapeURI extra) ""
                     (B.str $ orig <> extra)

image :: PandocMonad m => MarkdownParser m (F Inlines)
image = try $ do
  char '!'
  (lab,raw) <- reference
  defaultExt <- getOption readerDefaultImageExtension
  let constructor attr' src = case takeExtension (T.unpack src) of
                          "" -> B.imageWith attr' (T.pack $ addExtension (T.unpack src)
                                                          $ T.unpack defaultExt)
                          _  -> B.imageWith attr' src
  regLink constructor lab <|> referenceLink constructor (lab,raw)

note :: PandocMonad m => MarkdownParser m (F Inlines)
note = try $ do
  guardEnabled Ext_footnotes
  ref <- noteMarker
  updateState $ \st -> st{ stateNoteRefs = Set.insert ref (stateNoteRefs st) }
  return $ do
    notes <- asksF stateNotes'
    case M.lookup ref notes of
        Nothing       -> return $ B.str $ "[^" <> ref <> "]"
        Just (_pos, contents) -> do
          st <- askF
          -- process the note in a context that doesn't resolve
          -- notes, to avoid infinite looping with notes inside
          -- notes:
          let contents' = runF contents st{ stateNotes' = M.empty }
          return $ B.note contents'

inlineNote :: PandocMonad m => MarkdownParser m (F Inlines)
inlineNote = try $ do
  guardEnabled Ext_inline_notes
  char '^'
  contents <- inlinesInBalancedBrackets
  return $ B.note . B.para <$> contents

rawLaTeXInline' :: PandocMonad m => MarkdownParser m (F Inlines)
rawLaTeXInline' = try $ do
  guardEnabled Ext_raw_tex
  notFollowedBy' rawConTeXtEnvironment
  s <- rawLaTeXInline
  return $ return $ B.rawInline "tex" s -- "tex" because it might be context

rawConTeXtEnvironment :: PandocMonad m => ParserT Text st m Text
rawConTeXtEnvironment = try $ do
  string "\\start"
  completion <- inBrackets (letter <|> digit <|> spaceChar)
               <|> many1Char letter
  contents <- manyTill (rawConTeXtEnvironment <|> countChar 1 anyChar)
                       (try $ string "\\stop" >> textStr completion)
  return $ "\\start" <> completion <> T.concat contents <> "\\stop" <> completion

inBrackets :: PandocMonad m => ParserT Text st m Char -> ParserT Text st m Text
inBrackets parser = do
  char '['
  contents <- manyChar parser
  char ']'
  return $ "[" <> contents <> "]"

spanHtml :: PandocMonad m => MarkdownParser m (F Inlines)
spanHtml = try $ do
  guardEnabled Ext_native_spans
  (TagOpen _ attrs, _) <- htmlTag (~== TagOpen ("span" :: Text) [])
  contents <- mconcat <$> manyTill inline (htmlTag (~== TagClose ("span" :: Text)))
  let ident = fromMaybe "" $ lookup "id" attrs
  let classes = maybe [] T.words $ lookup "class" attrs
  let keyvals = [(k,v) | (k,v) <- attrs, k /= "id" && k /= "class"]
  return $ if isSmallCaps (ident, classes, keyvals)
              then B.smallcaps <$> contents
              else if isUnderline (ident, classes, keyvals)
                      then B.underline <$> contents
                      else B.spanWith (ident, classes, keyvals) <$> contents

divHtml :: PandocMonad m => MarkdownParser m (F Blocks)
divHtml = try $ do
  guardEnabled Ext_native_divs
  (TagOpen _ attrs, rawtag) <- htmlTag (~== TagOpen ("div" :: Text) [])
  -- we set stateInHtmlBlock so that closing tags that can be either block or
  -- inline will not be parsed as inline tags
  oldInHtmlBlock <- stateInHtmlBlock <$> getState
  updateState $ \st -> st{ stateInHtmlBlock = Just "div" }
  bls <- option "" (blankline >> option "" blanklines)
  contents <- mconcat <$>
              many (notFollowedBy' (htmlTag (~== TagClose ("div" :: Text))) >> block)
  closed <- option False (True <$ htmlTag (~== TagClose ("div" :: Text)))
  if closed
     then do
       updateState $ \st -> st{ stateInHtmlBlock = oldInHtmlBlock }
       let ident = fromMaybe "" $ lookup "id" attrs
       let classes = maybe [] T.words $ lookup "class" attrs
       let keyvals = [(k,v) | (k,v) <- attrs, k /= "id" && k /= "class"]
       return $ B.divWith (ident, classes, keyvals) <$> contents
     else -- avoid backtracing
       return $ return (B.rawBlock "html" (rawtag <> bls)) <> contents

divFenced :: PandocMonad m => MarkdownParser m (F Blocks)
divFenced = try $ do
  guardEnabled Ext_fenced_divs
  string ":::"
  skipMany (char ':')
  skipMany spaceChar
  attribs <- attributes <|> ((\x -> ("",[x],[])) <$> many1Char nonspaceChar)
  skipMany spaceChar
  skipMany (char ':')
  blankline
  updateState $ \st -> st{ stateFencedDivLevel = stateFencedDivLevel st + 1 }
  bs <- mconcat <$> manyTill block divFenceEnd
  updateState $ \st -> st{ stateFencedDivLevel = stateFencedDivLevel st - 1 }
  return $ B.divWith attribs <$> bs

divFenceEnd :: PandocMonad m => MarkdownParser m ()
divFenceEnd = try $ do
  string ":::"
  skipMany (char ':')
  blanklines
  return ()

rawHtmlInline :: PandocMonad m => MarkdownParser m (F Inlines)
rawHtmlInline = do
  guardEnabled Ext_raw_html
  inHtmlBlock <- stateInHtmlBlock <$> getState
  let isCloseBlockTag t = case inHtmlBlock of
                               Just t' -> t ~== TagClose t'
                               Nothing -> False
  mdInHtml <- option False $
                (    guardEnabled Ext_markdown_in_html_blocks
                 <|> guardEnabled Ext_markdown_attribute
                ) >> return True
  (_,result) <- htmlTag $ if mdInHtml
                             then (\x -> isInlineTag x &&
                                         not (isCloseBlockTag x))
                             else not . isTextTag
  return $ return $ B.rawInline "html" result

-- Emoji

emojiChars :: [Char]
emojiChars = ['a'..'z'] ++ ['0'..'9'] ++ ['_','+','-']

emoji :: PandocMonad m => MarkdownParser m (F Inlines)
emoji = try $ do
  guardEnabled Ext_emoji
  char ':'
  emojikey <- many1Char (oneOf emojiChars)
  char ':'
  case emojiToInline emojikey of
    Just i -> return (return $ B.singleton i)
    Nothing -> mzero

-- Citations

cite :: PandocMonad m => MarkdownParser m (F Inlines)
cite = do
  guardEnabled Ext_citations
  textualCite
            <|> do (cs, raw) <- withRaw normalCite
                   return $ flip B.cite (B.text raw) <$> cs

textualCite :: PandocMonad m => MarkdownParser m (F Inlines)
textualCite = try $ do
  (suppressAuthor, key) <- citeKey
  let first = Citation{ citationId      = key
                      , citationPrefix  = []
                      , citationSuffix  = []
                      , citationMode    = if suppressAuthor
                                             then SuppressAuthor
                                             else AuthorInText
                      , citationNoteNum = 0
                      , citationHash    = 0
                      }
  mbrest <- option Nothing $ try $ spnl >> Just <$> withRaw normalCite
  case mbrest of
       Just (rest, raw) ->
         return $ (flip B.cite (B.text $ "@" <> key <> " " <> raw) . (first:))
               <$> rest
       Nothing   ->
         (do
          (cs, raw) <- withRaw $ bareloc first
          let (spaces',raw') = T.span isSpace raw
              spc | T.null spaces' = mempty
                  | otherwise      = B.space
          lab <- parseFromString' inlines $ dropBrackets raw'
          fallback <- referenceLink B.linkWith (lab,raw')
          return $ do
            fallback' <- fallback
            cs' <- cs
            return $
              case B.toList fallback' of
                Link{}:_ -> B.cite [first] (B.str $ "@" <> key) <> spc <> fallback'
                _        -> B.cite cs' (B.text $ "@" <> key <> " " <> raw))
         <|> return (do st <- askF
                        return $ case M.lookup key (stateExamples st) of
                                 Just n -> B.str $ tshow n
                                 _      -> B.cite [first] $ B.str $ "@" <> key)

bareloc :: PandocMonad m => Citation -> MarkdownParser m (F [Citation])
bareloc c = try $ do
  spnl
  char '['
  notFollowedBy $ char '^'
  suff <- suffix
  rest <- option (return []) $ try $ char ';' >> citeList
  spnl
  char ']'
  notFollowedBy $ oneOf "[("
  return $ do
    suff' <- suff
    rest' <- rest
    return $ c{ citationSuffix = B.toList suff' } : rest'

normalCite :: PandocMonad m => MarkdownParser m (F [Citation])
normalCite = try $ do
  char '['
  spnl
  citations <- citeList
  spnl
  char ']'
  return citations

suffix :: PandocMonad m => MarkdownParser m (F Inlines)
suffix = try $ do
  hasSpace <- option False (notFollowedBy nonspaceChar >> return True)
  spnl
  rest <- trimInlinesF . mconcat <$> many (notFollowedBy (oneOf ";]") >> inline)
  return $ if hasSpace
              then (B.space <>) <$> rest
              else rest

prefix :: PandocMonad m => MarkdownParser m (F Inlines)
prefix = trimInlinesF . mconcat <$>
  manyTill inline (char ']' <|> fmap (const ']') (lookAhead citeKey))

citeList :: PandocMonad m => MarkdownParser m (F [Citation])
citeList = fmap sequence $ sepBy1 citation (try $ char ';' >> spnl)

citation :: PandocMonad m => MarkdownParser m (F Citation)
citation = try $ do
  pref <- prefix
  (suppress_author, key) <- citeKey
  suff <- suffix
  return $ do
    x <- pref
    y <- suff
    return Citation{ citationId      = key
                   , citationPrefix  = B.toList x
                   , citationSuffix  = B.toList y
                   , citationMode    = if suppress_author
                                          then SuppressAuthor
                                          else NormalCitation
                   , citationNoteNum = 0
                   , citationHash    = 0
                   }

smart :: PandocMonad m => MarkdownParser m (F Inlines)
smart = do
  guardEnabled Ext_smart
  doubleQuoted <|> singleQuoted <|>
    choice (map (return <$>) [apostrophe, dash, ellipses])

singleQuoted :: PandocMonad m => MarkdownParser m (F Inlines)
singleQuoted = try $ do
  singleQuoteStart
  withQuoteContext InSingleQuote $
    fmap B.singleQuoted . trimInlinesF . mconcat <$>
      many1Till inline singleQuoteEnd

-- doubleQuoted will handle regular double-quoted sections, as well
-- as dialogues with an open double-quote without a close double-quote
-- in the same paragraph.
doubleQuoted :: PandocMonad m => MarkdownParser m (F Inlines)
doubleQuoted = try $ do
  doubleQuoteStart
  withQuoteContext InDoubleQuote $
    fmap B.doubleQuoted . trimInlinesF . mconcat <$>
      many1Till inline doubleQuoteEnd

toRow :: [Blocks] -> Row
toRow = Row nullAttr . map B.simpleCell

toHeaderRow :: [Blocks] -> [Row]
toHeaderRow l = if null l then [] else [toRow l]