1 # This file is part of NIT ( http://www.nitlanguage.org ).
3 # Licensed under the Apache License, Version 2.0 (the "License");
4 # you may not use this file except in compliance with the License.
5 # You may obtain a copy of the License at
7 # http://www.apache.org/licenses/LICENSE-2.0
9 # Unless required by applicable law or agreed to in writing, software
10 # distributed under the License is distributed on an "AS IS" BASIS,
11 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12 # See the License for the specific language governing permissions and
13 # limitations under the License.
20 # Parse a markdown string and split it in blocks.
22 # Blocks are then outputed by an `MarkdownEmitter`.
26 # var proc = new MarkdownProcessor
27 # var html = proc.process("**Hello World!**")
28 # assert html == "<p><strong>Hello World!</strong></p>\n"
30 # SEE: `String::md_to_html` for a shortcut.
31 class MarkdownProcessor
33 # `MarkdownEmitter` used for ouput.
34 var emitter
: MarkdownEmitter is noinit
, protected writable
36 # Work in extended mode (default).
38 # Behavior changes when using extended mode:
40 # * Lists and code blocks end a paragraph
42 # In normal markdown the following:
46 # * and this is not a list
52 # <p>This is a paragraph
53 # * and this is not a list</p>
56 # When using extended mode this changes to:
59 # <p>This is a paragraph</p>
61 # <li>and this is not a list</li>
65 # * Fences code blocks
67 # If you don't want to indent your all your code with 4 spaces,
68 # you can wrap your code in ``` ``` ``` or `~~~`.
74 # print "Hello World!"
80 # If you want to use syntax highlighting tools, most of them need to know what kind
81 # of language they are highlighting.
82 # You can add an optional language identifier after the fence declaration to output
83 # it in the HTML render.
88 # print "# Hello World!".md_to_html
94 # <pre class="nit"><code>import markdown
96 # print "Hello World!".md_to_html
100 # * Underscores (Emphasis)
102 # Underscores in the middle of a word like:
108 # normally produces this:
111 # <p>Con<em>cat</em>this</p>
114 # With extended mode they don't result in emphasis.
117 # <p>Con_cat_this</p>
122 # Like in [GFM](https://help.github.com/articles/github-flavored-markdown),
123 # strikethrought span is marked with `~~`.
132 # <del>Mistaken text.</del>
136 init do self.emitter
= new MarkdownEmitter(self)
138 # Process the mardown `input` string and return the processed output.
139 fun process
(input
: String): Writable do
146 var parent
= read_lines
(input
)
147 parent
.remove_surrounding_empty_lines
148 recurse
(parent
, false)
149 # output processed text
150 return emitter
.emit
(parent
.kind
)
153 # Split `input` string into `MDLines` and create a parent `MDBlock` with it.
154 private fun read_lines
(input
: String): MDBlock do
155 var block
= new MDBlock(new MDLocation(1, 1, 1, 1))
156 var value
= new FlatBuffer
162 while i
< input
.length
do
166 while not eol
and i
< input
.length
do
171 else if c
== '\t' then
172 var np
= pos
+ (4 - (pos
& 3))
185 var loc
= new MDLocation(line_pos
, 1, line_pos
, col_pos
)
186 var line
= new MDLine(loc
, value
.write_to_string
)
187 var is_link_ref
= check_link_ref
(line
)
189 if not is_link_ref
then block
.add_line line
195 # Check if line is a block link definition.
196 # Return `true` if line contains a valid link ref and save it into `link_refs`.
197 private fun check_link_ref
(line
: MDLine): Bool do
199 var is_link_ref
= false
200 var id
= new FlatBuffer
201 var link
= new FlatBuffer
202 var comment
= new FlatBuffer
204 if not line
.is_empty
and line
.leading
< 4 and line
.value
[line
.leading
] == '[' then
205 pos
= line
.leading
+ 1
206 pos
= md
.read_until
(id
, pos
, ']')
207 if not id
.is_empty
and pos
+ 2 < line
.value
.length
then
208 if line
.value
[pos
+ 1] == ':' then
210 pos
= md
.skip_spaces
(pos
)
211 if line
.value
[pos
] == '<' then
213 pos
= md
.read_until
(link
, pos
, '>')
216 pos
= md
.read_until
(link
, pos
, ' ', '\n')
218 if not link
.is_empty
then
219 pos
= md
.skip_spaces
(pos
)
220 if pos
> 0 and pos
< line
.value
.length
then
221 var c
= line
.value
[pos
]
222 if c
== '\"' or c
== '\'' or c == '(' then
225 pos = md.read_until(comment, pos, ')')
227 pos = md.read_until(comment, pos, c)
229 if pos > 0 then is_link_ref = true
238 if is_link_ref and not id.is_empty and not link.is_empty then
239 var lr = new LinkRef.with_title(link.write_to_string, comment.write_to_string)
240 add_link_ref(id.write_to_string, lr)
241 if comment.is_empty then last_link_ref = lr
244 comment = new FlatBuffer
245 if not line.is_empty and last_link_ref != null then
247 var c = line.value[pos]
248 if c == '\
"' or c == '\'' or c == '(' then
251 pos = md.read_until(comment, pos, ')')
253 pos = md.read_until(comment, pos, c)
256 if not comment.is_empty then last_link_ref.title = comment.write_to_string
258 if comment.is_empty then return false
264 # This list will be needed during output to expand links.
265 var link_refs: Map[String, LinkRef] = new HashMap[String, LinkRef]
267 # Last encountered link ref (for multiline definitions)
269 # Markdown allows link refs to be defined over two lines:
272 # [id]: http://example.com/longish/path/to/resource/here
273 # "Optional Title Here"
276 private var last_link_ref: nullable LinkRef = null
278 # Add a link ref to the list
279 fun add_link_ref(key: String, ref: LinkRef) do link_refs[key.to_lower] = ref
281 # Recursively split a `block`.
283 # The block is splitted according to the type of lines it contains.
284 # Some blocks can be splited again recursively like lists.
285 # The `in_list` mode is used to recurse on list and build
286 # nested paragraphs or code blocks.
287 fun recurse(root: MDBlock, in_list: Bool) do
288 var old_mode = self.in_list
289 var old_root = self.current_block
290 self.in_list = in_list
292 var line = root.first_line
293 while line != null and line.is_empty do
295 if line == null then return
300 while current_line != null do
301 line_kind(current_line.as(not null)).process(self)
303 self.in_list = old_mode
304 self.current_block = old_root
307 # Currently processed line.
308 # Used when visiting blocks with `recurse`.
309 var current_line: nullable MDLine = null is writable
311 # Currently processed block.
312 # Used when visiting blocks with `recurse`.
313 var current_block: nullable MDBlock = null is writable
315 # Is the current recursion in list mode?
316 # Used when visiting blocks with `recurse`
317 private var in_list = false
321 fun line_kind(md: MDLine): Line do
323 var leading = md.leading
324 var trailing = md.trailing
325 if md.is_empty then return new LineEmpty
326 if md.leading > 3 then return new LineCode
327 if value[leading] == '#' then return new LineHeadline
328 if value[leading] == '>' then return new LineBlockquote
331 if value.length - leading - trailing > 2 then
332 if value[leading] == '`' and md.count_chars_start('`') >= 3 then
335 if value[leading] == '~' and md.count_chars_start('~') >= 3 then
341 if value.length - leading - trailing > 2 and
342 (value[leading] == '*' or value[leading] == '-' or value[leading] == '_') then
343 if md.count_chars(value[leading]) >= 3 then
348 if value.length - leading >= 2 and value[leading + 1] == ' ' then
349 var c = value[leading]
350 if c == '*' or c == '-' or c == '+' then return new LineUList
353 if value.length - leading >= 3 and value[leading].is_digit then
355 while i < value.length and value[i].is_digit do i += 1
356 if i + 1 < value.length and value[i] == '.' and value[i + 1] == ' ' then
361 if value[leading] == '<' and md.check_html then return new LineXML
364 if next != null and not next.is_empty then
365 if next.count_chars('=') > 0 then
366 return new LineHeadline1
368 if next.count_chars('-') > 0 then
369 return new LineHeadline2
375 # Get the token kind at `pos`.
376 fun token_at(text: Text, pos: Int): Token do
388 if pos + 1 < text.length then
393 if pos + 2 < text.length then
399 var loc = new MDLocation(
400 current_loc.line_start,
401 current_loc.column_start + pos,
402 current_loc.line_start,
403 current_loc.column_start + pos)
407 if c0 != ' ' or c2 != ' ' then
408 return new TokenStrongStar(loc, pos, c)
410 return new TokenEmStar(loc, pos, c)
413 if c0 != ' ' or c1 != ' ' then
414 return new TokenEmStar(loc, pos, c)
416 return new TokenNone(loc, pos, c)
418 else if c == '_' then
420 if c0 != ' ' or c2 != ' 'then
421 return new TokenStrongUnderscore(loc, pos, c)
423 return new TokenEmUnderscore(loc, pos, c)
427 if (c0.is_letter or c0.is_digit) and c0 != '_' and
428 (c1.is_letter or c1.is_digit) then
429 return new TokenNone(loc, pos, c)
431 return new TokenEmUnderscore(loc, pos, c)
434 if c0 != ' ' or c1 != ' ' then
435 return new TokenEmUnderscore(loc, pos, c)
437 return new TokenNone(loc, pos, c)
439 else if c == '!' then
440 if c1 == '[' then return new TokenImage(loc, pos, c)
441 return new TokenNone(loc, pos, c)
442 else if c == '[' then
443 return new TokenLink(loc, pos, c)
444 else if c == ']' then
445 return new TokenNone(loc, pos, c)
446 else if c == '`' then
448 return new TokenCodeDouble(loc, pos, c)
450 return new TokenCodeSingle(loc, pos, c)
452 else if c == '\\' then
453 if c1 == '\\' or c1 == '[' or c1 == ']' or c1 == '(' or c1 == ')' or c1 == '{' or c1 == '}' or c1 == '#' or c1 == '"' or c1 == '\
'' or c1
== '.' or c1
== '<' or c1
== '>' or c1
== '*' or c1
== '+' or c1
== '-' or c1
== '_' or c1
== '!' or c1
== '`' or c1
== '~' or c1
== '^' then
454 return new TokenEscape(loc
, pos
, c
)
456 return new TokenNone(loc
, pos
, c
)
458 else if c
== '<' then
459 return new TokenHTML(loc
, pos
, c
)
460 else if c
== '&' then
461 return new TokenEntity(loc
, pos
, c
)
464 if c
== '~' and c1
== '~' then
465 return new TokenStrike(loc
, pos
, c
)
468 return new TokenNone(loc
, pos
, c
)
472 # Find the position of a `token` in `self`.
473 fun find_token
(text
: Text, start
: Int, token
: Token): Int do
475 while pos
< text
.length
do
476 if token_at
(text
, pos
).is_same_type
(token
) then
484 # Location used for next parsed token.
486 # This location can be changed by the emitter to adjust with `\n` found
488 private fun current_loc
: MDLocation do return emitter
.current_loc
491 # Emit output corresponding to blocks content.
493 # Blocks are created by a previous pass in `MarkdownProcessor`.
494 # The emitter use a `Decorator` to select the output format.
495 class MarkdownEmitter
497 # Kind of processor used for parsing.
498 type PROCESSOR: MarkdownProcessor
500 # Processor containing link refs.
501 var processor
: PROCESSOR
503 # Kind of decorator used for decoration.
504 type DECORATOR: Decorator
506 # Decorator used for output.
507 # Default is `HTMLDecorator`
508 var decorator
: DECORATOR is writable, lazy
do
509 return new HTMLDecorator
512 # Create a new `MarkdownEmitter` using a custom `decorator`.
513 init with_decorator
(processor
: PROCESSOR, decorator
: DECORATOR) do
515 self.decorator
= decorator
518 # Output `block` using `decorator` in the current buffer.
519 fun emit
(block
: Block): Text do
520 var buffer
= push_buffer
526 # Output the content of `block`.
527 fun emit_in
(block
: Block) do block
.emit_in
(self)
529 # Transform and emit mardown text
530 fun emit_text
(text
: Text) do emit_text_until
(text
, 0, null)
532 # Transform and emit mardown text starting at `start` and
533 # until a token with the same type as `token` is found.
534 # Go until the end of `text` if `token` is null.
535 fun emit_text_until
(text
: Text, start
: Int, token
: nullable Token): Int do
536 var old_text
= current_text
537 var old_pos
= current_pos
540 while current_pos
< text
.length
do
541 if text
[current_pos
] == '\n' then
542 current_loc
.line_start
+= 1
543 current_loc
.column_start
= -current_pos
545 var mt
= processor
.token_at
(text
, current_pos
)
546 if (token
!= null and not token
isa TokenNone) and
547 (mt
.is_same_type
(token
) or
548 (token
isa TokenEmStar and mt
isa TokenStrongStar) or
549 (token
isa TokenEmUnderscore and mt
isa TokenStrongUnderscore)) then
555 current_text
= old_text
556 current_pos
= old_pos
560 # Currently processed position in `current_text`.
561 # Used when visiting inline production with `emit_text_until`.
562 private var current_pos
: Int = -1
564 # Currently processed text.
565 # Used when visiting inline production with `emit_text_until`.
566 private var current_text
: nullable Text = null
569 private var buffer_stack
= new List[FlatBuffer]
571 # Push a new buffer on the stack.
572 private fun push_buffer
: FlatBuffer do
573 var buffer
= new FlatBuffer
574 buffer_stack
.add buffer
578 # Pop the last buffer.
579 private fun pop_buffer
do buffer_stack
.pop
581 # Current output buffer.
582 private fun current_buffer
: FlatBuffer do
583 assert not buffer_stack
.is_empty
584 return buffer_stack
.last
588 private var loc_stack
= new List[MDLocation]
590 # Push a new MDLocation on the stack.
591 private fun push_loc
(location
: MDLocation) do loc_stack
.add location
593 # Pop the last buffer.
594 private fun pop_loc
: MDLocation do return loc_stack
.pop
596 # Current output buffer.
597 private fun current_loc
: MDLocation do
598 assert not loc_stack
.is_empty
599 return loc_stack
.last
602 # Append `e` to current buffer.
603 fun add
(e
: Writable) do
605 current_buffer
.append e
607 current_buffer
.append e
.write_to_string
611 # Append `c` to current buffer.
612 fun addc
(c
: Char) do add c
.to_s
614 # Append a "\n" line break.
619 # Links that are specified somewhere in the mardown document to be reused as shortcuts.
622 # [1]: http://example.com/ "Optional title"
629 # Optional link title
630 var title
: nullable String = null
632 # Is the link an abreviation?
633 var is_abbrev
= false
635 # Create a link with a title.
636 init with_title
(link
: String, title
: nullable String) do
642 # A `Decorator` is used to emit mardown into a specific format.
643 # Default decorator used is `HTMLDecorator`.
646 # Kind of emitter used for decoration.
647 type EMITTER: MarkdownEmitter
649 # Render a single plain char.
651 # Redefine this method to add special escaping for plain text.
652 fun add_char
(v
: EMITTER, c
: Char) do v
.addc c
654 # Render a ruler block.
655 fun add_ruler
(v
: EMITTER, block
: BlockRuler) is abstract
657 # Render a headline block with corresponding level.
658 fun add_headline
(v
: EMITTER, block
: BlockHeadline) is abstract
660 # Render a paragraph block.
661 fun add_paragraph
(v
: EMITTER, block
: BlockParagraph) is abstract
663 # Render a code or fence block.
664 fun add_code
(v
: EMITTER, block
: BlockCode) is abstract
666 # Render a blockquote.
667 fun add_blockquote
(v
: EMITTER, block
: BlockQuote) is abstract
669 # Render an unordered list.
670 fun add_unorderedlist
(v
: EMITTER, block
: BlockUnorderedList) is abstract
672 # Render an ordered list.
673 fun add_orderedlist
(v
: EMITTER, block
: BlockOrderedList) is abstract
675 # Render a list item.
676 fun add_listitem
(v
: EMITTER, block
: BlockListItem) is abstract
678 # Render an emphasis text.
679 fun add_em
(v
: EMITTER, text
: Text) is abstract
681 # Render a strong text.
682 fun add_strong
(v
: EMITTER, text
: Text) is abstract
684 # Render a strike text.
686 # Extended mode only (see `MarkdownProcessor::ext_mode`)
687 fun add_strike
(v
: EMITTER, text
: Text) is abstract
690 fun add_link
(v
: EMITTER, link
: Text, name
: Text, comment
: nullable Text) is abstract
693 fun add_image
(v
: EMITTER, link
: Text, name
: Text, comment
: nullable Text) is abstract
695 # Render an abbreviation.
696 fun add_abbr
(v
: EMITTER, name
: Text, comment
: Text) is abstract
698 # Render a code span reading from a buffer.
699 fun add_span_code
(v
: EMITTER, buffer
: Text, from
, to
: Int) is abstract
701 # Render a text and escape it.
702 fun append_value
(v
: EMITTER, value
: Text) is abstract
704 # Render code text from buffer and escape it.
705 fun append_code
(v
: EMITTER, buffer
: Text, from
, to
: Int) is abstract
707 # Render a character escape.
708 fun escape_char
(v
: EMITTER, char
: Char) is abstract
710 # Render a line break
711 fun add_line_break
(v
: EMITTER) is abstract
713 # Generate a new html valid id from a `String`.
714 fun strip_id
(txt
: String): String is abstract
716 # Found headlines during the processing labeled by their ids.
717 fun headlines
: ArrayMap[String, HeadLine] is abstract
720 # Class representing a markdown headline.
722 # Unique identifier of this headline.
725 # Text of the headline.
728 # Level of this headline.
730 # According toe the markdown specification, level must be in `[1..6]`.
734 # `Decorator` that outputs HTML.
738 redef var headlines
= new ArrayMap[String, HeadLine]
740 redef fun add_ruler
(v
, block
) do v
.add
"<hr/>\n"
742 redef fun add_headline
(v
, block
) do
744 var txt
= block
.block
.first_line
.value
745 var id
= strip_id
(txt
)
746 var lvl
= block
.depth
747 headlines
[id
] = new HeadLine(id
, txt
, lvl
)
749 v
.add
"<h{lvl} id=\"{id}\
">"
754 redef fun add_paragraph
(v
, block
) do
760 redef fun add_code
(v
, block
) do
761 if block
isa BlockFence and block
.meta
!= null then
762 v
.add
"<pre class=\"{block.meta.to_s}\
"><code>"
767 v
.add
"</code></pre>\n"
770 redef fun add_blockquote
(v
, block
) do
771 v
.add
"<blockquote>\n"
773 v
.add
"</blockquote>\n"
776 redef fun add_unorderedlist
(v
, block
) do
782 redef fun add_orderedlist
(v
, block
) do
788 redef fun add_listitem
(v
, block
) do
794 redef fun add_em
(v
, text
) do
800 redef fun add_strong
(v
, text
) do
806 redef fun add_strike
(v
, text
) do
812 redef fun add_image
(v
, link
, name
, comment
) do
814 append_value(v, link)
816 append_value(v, name)
818 if comment
!= null and not comment
.is_empty
then
820 append_value(v, comment)
826 redef fun add_link
(v
, link
, name
, comment
) do
828 append_value(v, link)
830 if comment
!= null and not comment
.is_empty
then
832 append_value(v, comment)
840 redef fun add_abbr
(v
, name
, comment
) do
841 v
.add
"<abbr title=\""
842 append_value(v, comment)
848 redef fun add_span_code
(v
, text
, from
, to
) do
850 append_code
(v
, text
, from
, to
)
854 redef fun add_line_break
(v
) do
858 redef fun append_value
(v
, text
) do for c
in text
do escape_char
(v
, c
)
860 redef fun escape_char
(v
, c
) do
863 else if c
== '<' then
865 else if c
== '>' then
867 else if c
== '"' then
869 else if c
== '\'' then
876 redef fun append_code(v, buffer, from, to) do
877 for i in [from..to[ do
881 else if c == '<' then
883 else if c == '>' then
891 redef fun strip_id(txt) do
893 var b = new FlatBuffer
898 if not c.is_letter and
900 not allowed_id_chars.has(c) then continue
906 # check for multiple id definitions
907 if headlines.has_key(key) then
910 while headlines.has_key(key) do
918 private var allowed_id_chars: Array[Char] = ['-', '_
', ':', '.']
921 # Location in a Markdown input.
924 # Starting line number (starting from 1).
927 # Starting column number (starting from 1).
928 var column_start: Int
930 # Stopping line number (starting from 1).
933 # Stopping column number (starting from 1).
936 redef fun to_s do return "{line_start},{column_start}--{line_end},{column_end}"
938 # Return a copy of `self`.
939 fun copy: MDLocation do
940 return new MDLocation(line_start, column_start, line_end, column_end)
944 # A block of markdown lines.
945 # A `MDBlock` can contains lines and/or sub-blocks.
948 # Position of `self` in the input.
949 var location: MDLocation
953 var kind: Block = new BlockNone(self) is writable
956 var first_line: nullable MDLine = null is writable
959 var last_line: nullable MDLine = null is writable
961 # First sub-block if any.
962 var first_block: nullable MDBlock = null is writable
964 # Last sub-block if any.
965 var last_block: nullable MDBlock = null is writable
967 # Previous block if any.
968 var prev: nullable MDBlock = null is writable
971 var next: nullable MDBlock = null is writable
973 # Does this block contain subblocks?
974 fun has_blocks: Bool do return first_block != null
977 fun count_blocks: Int do
979 var block = first_block
980 while block != null do
987 # Does this block contain lines?
988 fun has_lines: Bool do return first_line != null
991 fun count_lines: Int do
993 var line = first_line
994 while line != null do
1001 # Split `self` creating a new sub-block having `line` has `last_line`.
1002 fun split(line: MDLine): MDBlock do
1003 # location for new block
1004 var new_loc = new MDLocation(
1005 first_line.location.line_start,
1006 first_line.location.column_start,
1007 line.location.line_end,
1008 line.location.column_end)
1010 var block = new MDBlock(new_loc)
1011 block.first_line = first_line
1012 block.last_line = line
1013 first_line = line.next
1015 if first_line == null then
1018 first_line.prev = null
1019 # update current block loc
1020 location.line_start = first_line.location.line_start
1021 location.column_start = first_line.location.column_start
1023 if first_block == null then
1027 last_block.next = block
1033 # Add a `line` to this block.
1034 fun add_line(line: MDLine) do
1035 if last_line == null then
1039 last_line.next_empty = line.is_empty
1040 line.prev_empty = last_line.is_empty
1041 line.prev = last_line
1042 last_line.next = line
1047 # Remove `line` from this block.
1048 fun remove_line(line: MDLine) do
1049 if line.prev == null then
1050 first_line = line.next
1052 line.prev.next = line.next
1054 if line.next == null then
1055 last_line = line.prev
1057 line.next.prev = line.prev
1063 # Remove leading empty lines.
1064 fun remove_leading_empty_lines: Bool do
1065 var was_empty = false
1066 var line = first_line
1067 while line != null and line.is_empty do
1075 # Remove trailing empty lines.
1076 fun remove_trailing_empty_lines: Bool do
1077 var was_empty = false
1078 var line = last_line
1079 while line != null and line.is_empty do
1087 # Remove leading and trailing empty lines.
1088 fun remove_surrounding_empty_lines: Bool do
1089 var was_empty = false
1090 if remove_leading_empty_lines then was_empty = true
1091 if remove_trailing_empty_lines then was_empty = true
1095 # Remove list markers and up to 4 leading spaces.
1096 # Used to clean nested lists.
1097 fun remove_list_indent(v: MarkdownProcessor) do
1098 var line = first_line
1099 while line != null do
1100 if not line.is_empty then
1101 var kind = v.line_kind(line)
1102 if kind isa LineList then
1103 line.value = kind.extract_value(line)
1105 line.value = line.value.substring_from(line.leading.min(4))
1107 line.leading = line.process_leading
1113 # Collect block line text.
1115 var text = new FlatBuffer
1116 var line = first_line
1117 while line != null do
1118 if not line.is_empty then
1119 text.append line.text
1124 return text.write_to_string
1128 # Representation of a markdown block in the AST.
1129 # Each `Block` is linked to a `MDBlock` that contains mardown code.
1130 abstract class Block
1132 # The markdown block `self` is related to.
1135 # Output `self` using `v.decorator`.
1136 fun emit(v: MarkdownEmitter) do v.emit_in(self)
1138 # Emit the containts of `self`, lines or blocks.
1139 fun emit_in(v: MarkdownEmitter) do
1140 block.remove_surrounding_empty_lines
1141 if block.has_lines then
1148 # Emit lines contained in `block`.
1149 fun emit_lines(v: MarkdownEmitter) do
1150 var tpl = v.push_buffer
1151 var line = block.first_line
1152 while line != null do
1153 if not line.is_empty then
1154 v.add line.value.substring(line.leading, line.value.length - line.trailing)
1155 if line.trailing >= 2 then v.decorator.add_line_break(v)
1157 if line.next != null then
1166 # Emit sub-blocks contained in `block`.
1167 fun emit_blocks(v: MarkdownEmitter) do
1168 var block = self.block.first_block
1169 while block != null do
1170 v.push_loc(block.location)
1177 # The raw content of the block as a multi-line string.
1178 fun raw_content: String do
1179 var infence = self isa BlockFence
1180 var text = new FlatBuffer
1181 var line = self.block.first_line
1182 while line != null do
1183 if not line.is_empty then
1184 var str = line.value
1185 if not infence and str.has_prefix(" ") then
1186 text.append str.substring(4, str.length - line.trailing)
1194 return text.write_to_string
1198 # A block without any markdown specificities.
1200 # Actually use the same implementation than `BlockCode`,
1201 # this class is only used for typing purposes.
1206 # A markdown blockquote.
1210 redef fun emit(v) do v.decorator.add_blockquote(v, self)
1212 # Remove blockquote markers.
1213 private fun remove_block_quote_prefix(block: MDBlock) do
1214 var line = block.first_line
1215 while line != null do
1216 if not line.is_empty then
1217 if line.value[line.leading] == '>' then
1218 var rem = line.leading + 1
1219 if line.leading + 1 < line.value.length and
1220 line.value[line.leading + 1] == ' ' then
1223 line.value = line.value.substring_from(rem)
1224 line.leading = line.process_leading
1232 # A markdown code block.
1236 # Number of char to skip at the beginning of the line.
1238 # Block code lines start at 4 spaces.
1239 protected var line_start = 4
1241 redef fun emit(v) do v.decorator.add_code(v, self)
1243 redef fun emit_lines(v) do
1244 var line = block.first_line
1245 while line != null do
1246 if not line.is_empty then
1247 v.decorator.append_code(v, line.value, line_start, line.value.length)
1255 # A markdown code-fence block.
1257 # Actually use the same implementation than `BlockCode`,
1258 # this class is only used for typing purposes.
1262 # Any string found after fence token.
1263 var meta: nullable Text
1265 # Fence code lines start at 0 spaces.
1266 redef var line_start = 0
1269 # A markdown headline.
1273 redef fun emit(v) do
1274 var loc = block.location.copy
1275 loc.column_start += start
1277 v.decorator.add_headline(v, self)
1281 private var start = 0
1283 # Depth of the headline used to determine the headline level.
1286 # Remove healine marks from lines contained in `self`.
1287 private fun transform_headline(block: MDBlock) do
1288 if depth > 0 then return
1290 var line = block.first_line
1291 if line.is_empty then return
1292 var start = line.leading
1293 while start < line.value.length and line.value[start] == '#' do
1297 while start
< line
.value
.length
and line
.value
[start
] == ' ' do
1300 if start
>= line
.value
.length
then
1301 line
.is_empty
= true
1303 var nend
= line
.value
.length
- line
.trailing
- 1
1304 while line
.value
[nend
] == '#' do nend
-= 1
1305 while line
.value
[nend
] == ' ' do nend
-= 1
1306 line
.value
= line
.value
.substring
(start
, nend
- start
+ 1)
1311 depth
= level
.min
(6)
1315 # A markdown list item block.
1319 redef fun emit
(v
) do v
.decorator
.add_listitem
(v
, self)
1322 # A markdown list block.
1323 # Can be either an ordered or unordered list, this class is mainly used to factorize code.
1324 abstract class BlockList
1327 # Split list block into list items sub-blocks.
1328 private fun init_block
(v
: MarkdownProcessor) do
1329 var line
= block
.first_line
1331 while line
!= null do
1332 var t
= v
.line_kind
(line
)
1333 if t
isa LineList or
1334 (not line
.is_empty
and (line
.prev_empty
and line
.leading
== 0 and
1335 not (t
isa LineList))) then
1336 var sblock
= block
.split
(line
.prev
.as(not null))
1337 sblock
.kind
= new BlockListItem(sblock
)
1341 var sblock
= block
.split
(block
.last_line
.as(not null))
1342 sblock
.kind
= new BlockListItem(sblock
)
1345 # Expand list items as paragraphs if needed.
1346 private fun expand_paragraphs
(block
: MDBlock) do
1347 var outer
= block
.first_block
1348 var inner
: nullable MDBlock
1349 var has_paragraph
= false
1350 while outer
!= null and not has_paragraph
do
1351 if outer
.kind
isa BlockListItem then
1352 inner
= outer
.first_block
1353 while inner
!= null and not has_paragraph
do
1354 if inner
.kind
isa BlockParagraph then
1355 has_paragraph
= true
1362 if has_paragraph
then
1363 outer
= block
.first_block
1364 while outer
!= null do
1365 if outer
.kind
isa BlockListItem then
1366 inner
= outer
.first_block
1367 while inner
!= null do
1368 if inner
.kind
isa BlockNone then
1369 inner
.kind
= new BlockParagraph(inner
)
1380 # A markdown ordered list.
1381 class BlockOrderedList
1384 redef fun emit
(v
) do v
.decorator
.add_orderedlist
(v
, self)
1387 # A markdown unordred list.
1388 class BlockUnorderedList
1391 redef fun emit
(v
) do v
.decorator
.add_unorderedlist
(v
, self)
1394 # A markdown paragraph block.
1395 class BlockParagraph
1398 redef fun emit
(v
) do v
.decorator
.add_paragraph
(v
, self)
1405 redef fun emit
(v
) do v
.decorator
.add_ruler
(v
, self)
1408 # Xml blocks that can be found in markdown markup.
1412 redef fun emit_lines
(v
) do
1413 var line
= block
.first_line
1414 while line
!= null do
1415 if not line
.is_empty
then v
.add line
.value
1425 # Location of `self` in the original input.
1426 var location
: MDLocation
1428 # Text contained in this line.
1429 var value
: String is writable
1431 # Is this line empty?
1432 # Lines containing only spaces are considered empty.
1433 var is_empty
: Bool = true is writable
1435 # Previous line in `MDBlock` or null if first line.
1436 var prev
: nullable MDLine = null is writable
1438 # Next line in `MDBlock` or null if last line.
1439 var next
: nullable MDLine = null is writable
1441 # Is the previous line empty?
1442 var prev_empty
: Bool = false is writable
1444 # Is the next line empty?
1445 var next_empty
: Bool = false is writable
1447 # Initialize a new MDLine from its string value
1449 self.leading
= process_leading
1450 if leading
!= value
.length
then
1451 self.is_empty
= false
1452 self.trailing
= process_trailing
1456 # Set `value` as an empty String and update `leading`, `trailing` and is_`empty`.
1462 if prev
!= null then prev
.next_empty
= true
1463 if next
!= null then next
.prev_empty
= true
1466 # Number or leading spaces on this line.
1467 var leading
: Int = 0 is writable
1469 # Compute `leading` depending on `value`.
1470 fun process_leading
: Int do
1472 var value
= self.value
1473 while count
< value
.length
and value
[count
] == ' ' do count
+= 1
1474 if leading
== value
.length
then clear
1478 # Number of trailing spaces on this line.
1479 var trailing
: Int = 0 is writable
1481 # Compute `trailing` depending on `value`.
1482 fun process_trailing
: Int do
1484 var value
= self.value
1485 while value
[value
.length
- count
- 1] == ' ' do
1491 # Count the amount of `ch` in this line.
1492 # Return A value > 0 if this line only consists of `ch` end spaces.
1493 fun count_chars
(ch
: Char): Int do
1509 # Count the amount of `ch` at the start of this line ignoring spaces.
1510 fun count_chars_start
(ch
: Char): Int do
1525 # Last XML line if any.
1526 private var xml_end_line
: nullable MDLine = null
1528 # Does `value` contains valid XML markup?
1529 private fun check_html
: Bool do
1530 var tags
= new Array[String]
1531 var tmp
= new FlatBuffer
1533 if pos
+ 1 < value
.length
and value
[pos
+ 1] == '!' then
1534 if read_xml_comment
(self, pos
) > 0 then return true
1536 pos
= value
.read_xml
(tmp
, pos
, false)
1540 if not tag
.is_html_block
then
1548 var line
: nullable MDLine = self
1549 while line
!= null do
1550 while pos
< line
.value
.length
and line
.value
[pos
] != '<' do
1553 if pos
>= line
.value
.length
then
1554 if pos
- 2 >= 0 and line
.value
[pos
- 2] == '/' then
1556 if tags
.is_empty
then
1564 tmp
= new FlatBuffer
1565 var new_pos
= line
.value
.read_xml
(tmp
, pos
, false)
1568 if tag
.is_html_block
and not tag
== "hr" then
1569 if tmp
[1] == '/' then
1570 if tags
.last
!= tag
then
1578 if tags
.is_empty
then
1588 return tags
.is_empty
1593 # Read a XML comment.
1594 # Used by `check_html`.
1595 private fun read_xml_comment
(first_line
: MDLine, start
: Int): Int do
1596 var line
: nullable MDLine = first_line
1597 if start
+ 3 < line
.value
.length
then
1598 if line
.value
[2] == '-' and line
.value
[3] == '-' then
1600 while line
!= null do
1601 while pos
< line
.value
.length
and line
.value
[pos
] != '-' do
1604 if pos
== line
.value
.length
then
1608 if pos
+ 2 < line
.value
.length
then
1609 if line
.value
[pos
+ 1] == '-' and line
.value
[pos
+ 2] == '>' then
1610 first_line
.xml_end_line
= line
1622 # Extract the text of `self` without leading and trailing.
1623 fun text
: String do return value
.substring
(leading
, value
.length
- trailing
)
1630 # See `MarkdownProcessor::recurse`.
1631 fun process
(v
: MarkdownProcessor) is abstract
1634 # An empty markdown line.
1638 redef fun process
(v
) do
1639 v
.current_line
= v
.current_line
.next
1643 # A non-specific markdown construction.
1644 # Mainly used as part of another line construct such as paragraphs or lists.
1648 redef fun process
(v
) do
1649 var line
= v
.current_line
1651 var was_empty
= line
.prev_empty
1652 while line
!= null and not line
.is_empty
do
1653 var t
= v
.line_kind
(line
)
1654 if (v
.in_list
or v
.ext_mode
) and t
isa LineList then
1657 if v
.ext_mode
and (t
isa LineCode or t
isa LineFence) then
1660 if t
isa LineHeadline or t
isa LineHeadline1 or t
isa LineHeadline2 or
1661 t
isa LineHR or t
isa LineBlockquote or t
isa LineXML then
1667 if line
!= null and not line
.is_empty
then
1668 var block
= v
.current_block
.split
(line
.prev
.as(not null))
1669 if v
.in_list
and not was_empty
then
1670 block
.kind
= new BlockNone(block
)
1672 block
.kind
= new BlockParagraph(block
)
1674 v
.current_block
.remove_leading_empty_lines
1677 if line
!= null then
1678 block
= v
.current_block
.split
(line
)
1680 block
= v
.current_block
.split
(v
.current_block
.last_line
.as(not null))
1682 if v
.in_list
and (line
== null or not line
.is_empty
) and not was_empty
then
1683 block
.kind
= new BlockNone(block
)
1685 block
.kind
= new BlockParagraph(block
)
1687 v
.current_block
.remove_leading_empty_lines
1689 v
.current_line
= v
.current_block
.first_line
1693 # A line of markdown code.
1697 redef fun process
(v
) do
1698 var line
= v
.current_line
1700 while line
!= null and (line
.is_empty
or v
.line_kind
(line
) isa LineCode) do
1703 # split at block end line
1705 if line
!= null then
1706 block
= v
.current_block
.split
(line
.prev
.as(not null))
1708 block
= v
.current_block
.split
(v
.current_block
.last_line
.as(not null))
1710 block
.kind
= new BlockCode(block
)
1711 block
.remove_surrounding_empty_lines
1712 v
.current_line
= v
.current_block
.first_line
1716 # A line of raw XML.
1720 redef fun process
(v
) do
1721 var line
= v
.current_line
1722 var prev
= line
.prev
1723 if prev
!= null then v
.current_block
.split
(prev
)
1724 var block
= v
.current_block
.split
(line
.xml_end_line
.as(not null))
1725 block
.kind
= new BlockXML(block
)
1726 v
.current_block
.remove_leading_empty_lines
1727 v
.current_line
= v
.current_block
.first_line
1731 # A markdown blockquote line.
1732 class LineBlockquote
1735 redef fun process
(v
) do
1736 var line
= v
.current_line
1738 while line
!= null do
1739 if not line
.is_empty
and (line
.prev_empty
and
1740 line
.leading
== 0 and
1741 not v
.line_kind
(line
) isa LineBlockquote) then break
1746 if line
!= null then
1747 block
= v
.current_block
.split
(line
.prev
.as(not null))
1749 block
= v
.current_block
.split
(v
.current_block
.last_line
.as(not null))
1751 var kind
= new BlockQuote(block
)
1753 block
.remove_surrounding_empty_lines
1754 kind
.remove_block_quote_prefix
(block
)
1755 v
.current_line
= line
1756 v
.recurse
(block
, false)
1757 v
.current_line
= v
.current_block
.first_line
1761 # A markdown ruler line.
1765 redef fun process
(v
) do
1766 var line
= v
.current_line
1767 if line
.prev
!= null then v
.current_block
.split
(line
.prev
.as(not null))
1768 var block
= v
.current_block
.split
(line
.as(not null))
1769 block
.kind
= new BlockRuler(block
)
1770 v
.current_block
.remove_leading_empty_lines
1771 v
.current_line
= v
.current_block
.first_line
1775 # A markdown fence code line.
1779 redef fun process
(v
) do
1781 var line
= v
.current_line
.next
1782 while line
!= null do
1783 if v
.line_kind
(line
) isa LineFence then break
1786 if line
!= null then
1791 if line
!= null then
1792 block
= v
.current_block
.split
(line
.prev
.as(not null))
1794 block
= v
.current_block
.split
(v
.current_block
.last_line
.as(not null))
1796 block
.remove_surrounding_empty_lines
1797 var meta
= block
.first_line
.value
.meta_from_fence
1798 block
.kind
= new BlockFence(block
, meta
)
1799 block
.first_line
.clear
1800 var last
= block
.last_line
1801 if last
!= null and v
.line_kind
(last
) isa LineFence then
1802 block
.last_line
.clear
1804 block
.remove_surrounding_empty_lines
1805 v
.current_line
= line
1809 # A markdown headline.
1813 redef fun process
(v
) do
1814 var line
= v
.current_line
1815 var lprev
= line
.prev
1816 if lprev
!= null then v
.current_block
.split
(lprev
)
1817 var block
= v
.current_block
.split
(line
.as(not null))
1818 var kind
= new BlockHeadline(block
)
1820 kind
.transform_headline
(block
)
1821 v
.current_block
.remove_leading_empty_lines
1822 v
.current_line
= v
.current_block
.first_line
1826 # A markdown headline of level 1.
1830 redef fun process
(v
) do
1831 var line
= v
.current_line
1832 var lprev
= line
.prev
1833 if lprev
!= null then v
.current_block
.split
(lprev
)
1835 var block
= v
.current_block
.split
(line
.as(not null))
1836 var kind
= new BlockHeadline(block
)
1838 kind
.transform_headline
(block
)
1840 v
.current_block
.remove_leading_empty_lines
1841 v
.current_line
= v
.current_block
.first_line
1845 # A markdown headline of level 2.
1849 redef fun process
(v
) do
1850 var line
= v
.current_line
1851 var lprev
= line
.prev
1852 if lprev
!= null then v
.current_block
.split
(lprev
)
1854 var block
= v
.current_block
.split
(line
.as(not null))
1855 var kind
= new BlockHeadline(block
)
1857 kind
.transform_headline
(block
)
1859 v
.current_block
.remove_leading_empty_lines
1860 v
.current_line
= v
.current_block
.first_line
1864 # A markdown list line.
1865 # Mainly used to factorize code between ordered and unordered lists.
1866 abstract class LineList
1869 redef fun process
(v
) do
1870 var line
= v
.current_line
1872 while line
!= null do
1873 var t
= v
.line_kind
(line
)
1874 if not line
.is_empty
and (line
.prev_empty
and line
.leading
== 0 and
1875 not t
isa LineList) then break
1880 if line
!= null then
1881 list
= v
.current_block
.split
(line
.prev
.as(not null))
1883 list
= v
.current_block
.split
(v
.current_block
.last_line
.as(not null))
1885 var kind
= block_kind
(list
)
1887 list
.first_line
.prev_empty
= false
1888 list
.last_line
.next_empty
= false
1889 list
.remove_surrounding_empty_lines
1890 list
.first_line
.prev_empty
= false
1891 list
.last_line
.next_empty
= false
1893 var block
= list
.first_block
1894 while block
!= null do
1895 block
.remove_list_indent
(v
)
1896 v
.recurse
(block
, true)
1899 kind
.expand_paragraphs
(list
)
1900 v
.current_line
= line
1903 # Create a new block kind based on this line.
1904 protected fun block_kind
(block
: MDBlock): BlockList is abstract
1906 # Extract string value from `MDLine`.
1907 protected fun extract_value
(line
: MDLine): String is abstract
1910 # An ordered list line.
1914 redef fun block_kind
(block
) do return new BlockOrderedList(block
)
1916 redef fun extract_value
(line
) do
1917 return line
.value
.substring_from
(line
.value
.index_of
('.') + 2)
1921 # An unordered list line.
1925 redef fun block_kind
(block
) do return new BlockUnorderedList(block
)
1927 redef fun extract_value
(line
) do
1928 return line
.value
.substring_from
(line
.leading
+ 2)
1932 # A token represent a character in the markdown input.
1933 # Some tokens have a specific markup behaviour that is handled here.
1934 abstract class Token
1936 # Location of `self` in the original input.
1937 var location
: MDLocation
1939 # Position of `self` in input independant from lines.
1942 # Character found at `pos` in the markdown input.
1945 # Output that token using `MarkdownEmitter::decorator`.
1946 fun emit
(v
: MarkdownEmitter) do v
.decorator
.add_char
(v
, char
)
1949 # A token without a specific meaning.
1954 # An emphasis token.
1955 abstract class TokenEm
1958 redef fun emit
(v
) do
1959 var tmp
= v
.push_buffer
1960 var b
= v
.emit_text_until
(v
.current_text
.as(not null), pos
+ 1, self)
1963 v
.decorator
.add_em
(v
, tmp
)
1971 # An emphasis star token.
1976 # An emphasis underscore token.
1977 class TokenEmUnderscore
1982 abstract class TokenStrong
1985 redef fun emit
(v
) do
1986 var tmp
= v
.push_buffer
1987 var b
= v
.emit_text_until
(v
.current_text
.as(not null), pos
+ 2, self)
1990 v
.decorator
.add_strong
(v
, tmp
)
1991 v
.current_pos
= b
+ 1
1998 # A strong star token.
1999 class TokenStrongStar
2003 # A strong underscore token.
2004 class TokenStrongUnderscore
2009 # This class is mainly used to factorize work between single and double quoted span codes.
2010 abstract class TokenCode
2013 redef fun emit
(v
) do
2014 var a
= pos
+ next_pos
+ 1
2015 var b
= v
.processor
.find_token
(v
.current_text
.as(not null), a
, self)
2017 v
.current_pos
= b
+ next_pos
2018 while a
< b
and v
.current_text
[a
] == ' ' do a
+= 1
2020 while v
.current_text
[b
- 1] == ' ' do b
-= 1
2021 v
.decorator
.add_span_code
(v
, v
.current_text
.as(not null), a
, b
)
2028 private fun next_pos
: Int is abstract
2031 # A span code token.
2032 class TokenCodeSingle
2035 redef fun next_pos
do return 0
2038 # A doubled span code token.
2039 class TokenCodeDouble
2042 redef fun next_pos
do return 1
2045 # A link or image token.
2046 # This class is mainly used to factorize work between images and links.
2047 abstract class TokenLinkOrImage
2051 var link
: nullable Text = null
2054 var name
: nullable Text = null
2057 var comment
: nullable Text = null
2059 # Is the link construct an abbreviation?
2060 var is_abbrev
= false
2062 redef fun emit
(v
) do
2063 var tmp
= new FlatBuffer
2064 var b
= check_link
(v
, tmp
, pos
, self)
2073 # Emit the hyperlink as link or image.
2074 private fun emit_hyper
(v
: MarkdownEmitter) is abstract
2076 # Check if the link is a valid link.
2077 private fun check_link
(v
: MarkdownEmitter, out
: FlatBuffer, start
: Int, token
: Token): Int do
2078 var md
= v
.current_text
2080 if token
isa TokenLink then
2085 var tmp
= new FlatBuffer
2086 pos
= md
.read_md_link_id
(tmp
, pos
)
2087 if pos
< start
then return -1
2091 pos
= md
.skip_spaces
(pos
)
2093 var tid
= name
.write_to_string
.to_lower
2094 if v
.processor
.link_refs
.has_key
(tid
) then
2095 var lr
= v
.processor
.link_refs
[tid
]
2096 is_abbrev
= lr
.is_abbrev
2103 else if md
[pos
] == '(' then
2105 pos
= md
.skip_spaces
(pos
)
2106 if pos
< start
then return -1
2107 tmp
= new FlatBuffer
2108 var use_lt
= md
[pos
] == '<'
2110 pos
= md
.read_until
(tmp
, pos
+ 1, '>')
2112 pos
= md
.read_md_link
(tmp
, pos
)
2114 if pos
< start
then return -1
2115 if use_lt
then pos
+= 1
2116 link
= tmp
.write_to_string
2117 if md
[pos
] == ' ' then
2118 pos
= md
.skip_spaces
(pos
)
2119 if pos
> start
and md
[pos
] == '"' then
2121 tmp
= new FlatBuffer
2122 pos
= md
.read_until
(tmp
, pos
, '"')
2123 if pos
< start
then return -1
2124 comment
= tmp
.write_to_string
2126 pos
= md
.skip_spaces
(pos
)
2127 if pos
== -1 then return -1
2130 if pos
< start
then return -1
2131 if md
[pos
] != ')' then return -1
2132 else if md
[pos
] == '[' then
2134 tmp
= new FlatBuffer
2135 pos
= md
.read_raw_until
(tmp
, pos
, ']')
2136 if pos
< start
then return -1
2138 if tmp
.length
> 0 then
2143 var tid
= id
.write_to_string
.to_lower
2144 if v
.processor
.link_refs
.has_key
(tid
) then
2145 var lr
= v
.processor
.link_refs
[tid
]
2150 var tid
= name
.write_to_string
.replace
("\n", " ").to_lower
2151 if v
.processor
.link_refs
.has_key
(tid
) then
2152 var lr
= v
.processor
.link_refs
[tid
]
2160 if link
== null then return -1
2165 # A markdown link token.
2167 super TokenLinkOrImage
2169 redef fun emit_hyper
(v
) do
2170 if is_abbrev
and comment
!= null then
2171 v
.decorator
.add_abbr
(v
, name
.as(not null), comment
.as(not null))
2173 v
.decorator
.add_link
(v
, link
.as(not null), name
.as(not null), comment
)
2178 # A markdown image token.
2180 super TokenLinkOrImage
2182 redef fun emit_hyper
(v
) do
2183 v
.decorator
.add_image
(v
, link
.as(not null), name
.as(not null), comment
)
2191 redef fun emit
(v
) do
2192 var tmp
= new FlatBuffer
2193 var b
= check_html
(v
, tmp
, v
.current_text
.as(not null), v
.current_pos
)
2198 v
.decorator
.escape_char
(v
, char
)
2202 # Is the HTML valid?
2203 # Also take care of link and mailto shortcuts.
2204 private fun check_html
(v
: MarkdownEmitter, out
: FlatBuffer, md
: Text, start
: Int): Int do
2205 # check for auto links
2206 var tmp
= new FlatBuffer
2207 var pos
= md
.read_until
(tmp
, start
+ 1, ':', ' ', '>', '\n')
2208 if pos
!= -1 and md
[pos
] == ':' and tmp
.is_link_prefix
then
2209 pos
= md
.read_until
(tmp
, pos
, '>')
2211 var link
= tmp
.write_to_string
2212 v
.decorator
.add_link
(v
, link
, link
, null)
2216 # TODO check for mailto
2217 # check for inline html
2218 if start
+ 2 < md
.length
then
2219 return md
.read_xml
(out
, start
, true)
2225 # An HTML entity token.
2229 redef fun emit
(v
) do
2230 var tmp
= new FlatBuffer
2231 var b
= check_entity
(tmp
, v
.current_text
.as(not null), pos
)
2236 v
.decorator
.escape_char
(v
, char
)
2240 # Is the entity valid?
2241 private fun check_entity
(out
: FlatBuffer, md
: Text, start
: Int): Int do
2242 var pos
= md
.read_until
(out
, start
, ';')
2243 if pos
< 0 or out
.length
< 3 then
2246 if out
[1] == '#' then
2247 if out
[2] == 'x' or out
[2] == 'X' then
2248 if out
.length
< 4 then return -1
2249 for i
in [3..out
.length
[ do
2251 if (c
< '0' or c
> '9') and (c
< 'a' and c
> 'f') and (c
< 'A' and c
> 'F') then
2256 for i
in [2..out
.length
[ do
2258 if c
< '0' or c
> '9' then return -1
2263 for i
in [1..out
.length
[ do
2265 if not c
.is_digit
and not c
.is_letter
then return -1
2268 # TODO check entity is valid
2269 # if out.is_entity then
2279 # A markdown escape token.
2283 redef fun emit
(v
) do
2285 v
.addc v
.current_text
[v
.current_pos
]
2289 # A markdown strike token.
2291 # Extended mode only (see `MarkdownProcessor::ext_mode`)
2295 redef fun emit
(v
) do
2296 var tmp
= v
.push_buffer
2297 var b
= v
.emit_text_until
(v
.current_text
.as(not null), pos
+ 2, self)
2300 v
.decorator
.add_strike
(v
, tmp
)
2301 v
.current_pos
= b
+ 1
2310 # Get the position of the next non-space character.
2311 private fun skip_spaces
(start
: Int): Int do
2313 while pos
> -1 and pos
< length
and (self[pos
] == ' ' or self[pos
] == '\n') do
2316 if pos
< length
then return pos
2320 # Read `self` until `nend` and append it to the `out` buffer.
2321 # Escape markdown special chars.
2322 private fun read_until
(out
: FlatBuffer, start
: Int, nend
: Char...): Int do
2324 while pos
< length
do
2326 if c
== '\\' and pos
+ 1 < length
then
2327 pos
= escape
(out
, self[pos
+ 1], pos
)
2329 var end_reached
= false
2336 if end_reached
then break
2341 if pos
== length
then return -1
2345 # Read `self` as raw text until `nend` and append it to the `out` buffer.
2346 # No escape is made.
2347 private fun read_raw_until
(out
: FlatBuffer, start
: Int, nend
: Char...): Int do
2349 while pos
< length
do
2351 var end_reached
= false
2358 if end_reached
then break
2362 if pos
== length
then return -1
2366 # Read `self` as XML until `to` and append it to the `out` buffer.
2367 # Escape HTML special chars.
2368 private fun read_xml_until
(out
: FlatBuffer, from
: Int, to
: Char...): Int do
2371 var str_char
: nullable Char = null
2372 while pos
< length
do
2378 if pos
< length
then
2384 if c
== str_char
then
2391 if c
== '"' or c
== '\'' then
2396 var end_reached = false
2397 for n in [0..to.length[ do
2403 if end_reached then break
2408 if pos == length then return -1
2412 # Read `self` as XML and append it to the `out` buffer.
2413 # Safe mode can be activated to limit reading to valid xml.
2414 private fun read_xml(out: FlatBuffer, start: Int, safe_mode: Bool): Int do
2417 var is_close_tag = false
2418 if start + 1 >= length then return -1
2419 if self[start + 1] == '/' then
2422 else if self[start + 1] == '!' then
2426 is_close_tag = false
2430 var tmp = new FlatBuffer
2431 pos = read_xml_until(tmp, pos, ' ', '/', '>')
2432 if pos == -1 then return -1
2433 var tag = tmp.write_to_string.trim.to_lower
2434 if not tag.is_valid_html_tag then
2437 else if tag.is_html_unsafe then
2440 if is_close_tag then out.add '/'
2444 if is_close_tag then out.add '/'
2449 if is_close_tag then out.add '/'
2450 pos = read_xml_until(out, pos, ' ', '/', '>')
2452 if pos == -1 then return -1
2453 pos = read_xml_until(out, pos, '/', '>')
2454 if pos == -1 then return -1
2455 if self[pos] == '/' then
2457 pos = self.read_xml_until(out, pos + 1, '>')
2458 if pos == -1 then return -1
2460 if self[pos] == '>' then
2471 # Read a markdown link address and append it to the `out` buffer.
2472 private fun read_md_link(out: FlatBuffer, start: Int): Int do
2475 while pos < length do
2477 if c == '\\
' and pos + 1 < length then
2478 pos = escape(out, self[pos + 1], pos)
2480 var end_reached = false
2483 else if c == ' ' then
2484 if counter == 1 then end_reached = true
2485 else if c == ')' then
2487 if counter == 0 then end_reached = true
2489 if end_reached then break
2494 if pos == length then return -1
2498 # Read a markdown link text and append it to the `out` buffer.
2499 private fun read_md_link_id(out: FlatBuffer, start: Int): Int do
2502 while pos < length do
2504 var end_reached = false
2508 else if c == ']' then
2510 if counter == 0 then
2518 if end_reached then break
2521 if pos == length then return -1
2525 # Extract the XML tag name from a XML tag.
2526 private fun xml_tag: String do
2527 var tpl = new FlatBuffer
2529 if pos < length and self[1] == '/' then pos += 1
2530 while pos < length - 1 and (self[pos].is_digit or self[pos].is_letter) do
2534 return tpl.write_to_string.to_lower
2537 private fun is_valid_html_tag: Bool do
2538 if is_empty then return false
2540 if not c.is_alpha then return false
2545 # Read and escape the markdown contained in `self`.
2546 private fun escape(out: FlatBuffer, c: Char, pos: Int): Int do
2547 if c == '\\
' or c == '[' or c == ']' or c == '(' or c == ')' or c == '{' or
2548 c == '}' or c == '#' or c == '"' or c == '\'' or c == '.' or c == '<' or
2549 c
== '>' or c
== '*' or c
== '+' or c
== '-' or c
== '_' or c
== '!' or
2550 c
== '`' or c
== '~' or c
== '^' then
2558 # Extract string found at end of fence opening.
2559 private fun meta_from_fence
: nullable Text do
2560 for i
in [0..chars
.length
[ do
2562 if c
!= ' ' and c
!= '`' and c
!= '~' then
2563 return substring_from
(i
).trim
2569 # Is `self` an unsafe HTML element?
2570 private fun is_html_unsafe
: Bool do return html_unsafe_tags
.has
(self.write_to_string
)
2572 # Is `self` a HRML block element?
2573 private fun is_html_block
: Bool do return html_block_tags
.has
(self.write_to_string
)
2575 # Is `self` a link prefix?
2576 private fun is_link_prefix
: Bool do return html_link_prefixes
.has
(self.write_to_string
)
2578 private fun html_unsafe_tags
: Array[String] do return once
["applet", "head", "body", "frame", "frameset", "iframe", "script", "object"]
2580 private fun html_block_tags
: Array[String] do return once
["address", "article", "aside", "audio", "blockquote", "canvas", "dd", "div", "dl", "fieldset", "figcaption", "figure", "footer", "form", "h1", "h2", "h3", "h4", "h5", "h6", "header", "hgroup", "hr", "noscript", "ol", "output", "p", "pre", "section", "table", "tfoot", "ul", "video"]
2582 private fun html_link_prefixes
: Array[String] do return once
["http", "https", "ftp", "ftps"]
2587 # Parse `self` as markdown and return the HTML representation
2589 # var md = "**Hello World!**"
2590 # var html = md.md_to_html
2591 # assert html == "<p><strong>Hello World!</strong></p>\n"
2592 fun md_to_html
: Writable do
2593 var processor
= new MarkdownProcessor
2594 return processor
.process
(self)