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 var meta
= block
.meta
763 v
.add
"<pre class=\""
764 append_value(v, meta)
770 v
.add
"</code></pre>\n"
773 redef fun add_blockquote
(v
, block
) do
774 v
.add
"<blockquote>\n"
776 v
.add
"</blockquote>\n"
779 redef fun add_unorderedlist
(v
, block
) do
785 redef fun add_orderedlist
(v
, block
) do
791 redef fun add_listitem
(v
, block
) do
797 redef fun add_em
(v
, text
) do
803 redef fun add_strong
(v
, text
) do
809 redef fun add_strike
(v
, text
) do
815 redef fun add_image
(v
, link
, name
, comment
) do
817 append_value(v, link)
819 append_value(v, name)
821 if comment
!= null and not comment
.is_empty
then
823 append_value(v, comment)
829 redef fun add_link
(v
, link
, name
, comment
) do
831 append_value(v, link)
833 if comment
!= null and not comment
.is_empty
then
835 append_value(v, comment)
843 redef fun add_abbr
(v
, name
, comment
) do
844 v
.add
"<abbr title=\""
845 append_value(v, comment)
851 redef fun add_span_code
(v
, text
, from
, to
) do
853 append_code
(v
, text
, from
, to
)
857 redef fun add_line_break
(v
) do
861 redef fun append_value
(v
, text
) do for c
in text
do escape_char
(v
, c
)
863 redef fun escape_char
(v
, c
) do
866 else if c
== '<' then
868 else if c
== '>' then
870 else if c
== '"' then
872 else if c
== '\'' then
879 redef fun append_code(v, buffer, from, to) do
880 for i in [from..to[ do
884 else if c == '<' then
886 else if c == '>' then
894 redef fun strip_id(txt) do
896 var b = new FlatBuffer
901 if not c.is_letter and
903 not allowed_id_chars.has(c) then continue
909 # check for multiple id definitions
910 if headlines.has_key(key) then
913 while headlines.has_key(key) do
921 private var allowed_id_chars: Array[Char] = ['-', '_
', ':', '.']
924 # Location in a Markdown input.
927 # Starting line number (starting from 1).
930 # Starting column number (starting from 1).
931 var column_start: Int
933 # Stopping line number (starting from 1).
936 # Stopping column number (starting from 1).
939 redef fun to_s do return "{line_start},{column_start}--{line_end},{column_end}"
941 # Return a copy of `self`.
942 fun copy: MDLocation do
943 return new MDLocation(line_start, column_start, line_end, column_end)
947 # A block of markdown lines.
948 # A `MDBlock` can contains lines and/or sub-blocks.
951 # Position of `self` in the input.
952 var location: MDLocation
956 var kind: Block = new BlockNone(self) is writable
959 var first_line: nullable MDLine = null is writable
962 var last_line: nullable MDLine = null is writable
964 # First sub-block if any.
965 var first_block: nullable MDBlock = null is writable
967 # Last sub-block if any.
968 var last_block: nullable MDBlock = null is writable
970 # Previous block if any.
971 var prev: nullable MDBlock = null is writable
974 var next: nullable MDBlock = null is writable
976 # Does this block contain subblocks?
977 fun has_blocks: Bool do return first_block != null
980 fun count_blocks: Int do
982 var block = first_block
983 while block != null do
990 # Does this block contain lines?
991 fun has_lines: Bool do return first_line != null
994 fun count_lines: Int do
996 var line = first_line
997 while line != null do
1004 # Split `self` creating a new sub-block having `line` has `last_line`.
1005 fun split(line: MDLine): MDBlock do
1006 # location for new block
1007 var new_loc = new MDLocation(
1008 first_line.location.line_start,
1009 first_line.location.column_start,
1010 line.location.line_end,
1011 line.location.column_end)
1013 var block = new MDBlock(new_loc)
1014 block.first_line = first_line
1015 block.last_line = line
1016 first_line = line.next
1018 if first_line == null then
1021 first_line.prev = null
1022 # update current block loc
1023 location.line_start = first_line.location.line_start
1024 location.column_start = first_line.location.column_start
1026 if first_block == null then
1030 last_block.next = block
1036 # Add a `line` to this block.
1037 fun add_line(line: MDLine) do
1038 if last_line == null then
1042 last_line.next_empty = line.is_empty
1043 line.prev_empty = last_line.is_empty
1044 line.prev = last_line
1045 last_line.next = line
1050 # Remove `line` from this block.
1051 fun remove_line(line: MDLine) do
1052 if line.prev == null then
1053 first_line = line.next
1055 line.prev.next = line.next
1057 if line.next == null then
1058 last_line = line.prev
1060 line.next.prev = line.prev
1066 # Remove leading empty lines.
1067 fun remove_leading_empty_lines: Bool do
1068 var was_empty = false
1069 var line = first_line
1070 while line != null and line.is_empty do
1078 # Remove trailing empty lines.
1079 fun remove_trailing_empty_lines: Bool do
1080 var was_empty = false
1081 var line = last_line
1082 while line != null and line.is_empty do
1090 # Remove leading and trailing empty lines.
1091 fun remove_surrounding_empty_lines: Bool do
1092 var was_empty = false
1093 if remove_leading_empty_lines then was_empty = true
1094 if remove_trailing_empty_lines then was_empty = true
1098 # Remove list markers and up to 4 leading spaces.
1099 # Used to clean nested lists.
1100 fun remove_list_indent(v: MarkdownProcessor) do
1101 var line = first_line
1102 while line != null do
1103 if not line.is_empty then
1104 var kind = v.line_kind(line)
1105 if kind isa LineList then
1106 line.value = kind.extract_value(line)
1108 line.value = line.value.substring_from(line.leading.min(4))
1110 line.leading = line.process_leading
1116 # Collect block line text.
1118 var text = new FlatBuffer
1119 var line = first_line
1120 while line != null do
1121 if not line.is_empty then
1122 text.append line.text
1127 return text.write_to_string
1131 # Representation of a markdown block in the AST.
1132 # Each `Block` is linked to a `MDBlock` that contains mardown code.
1133 abstract class Block
1135 # The markdown block `self` is related to.
1138 # Output `self` using `v.decorator`.
1139 fun emit(v: MarkdownEmitter) do v.emit_in(self)
1141 # Emit the containts of `self`, lines or blocks.
1142 fun emit_in(v: MarkdownEmitter) do
1143 block.remove_surrounding_empty_lines
1144 if block.has_lines then
1151 # Emit lines contained in `block`.
1152 fun emit_lines(v: MarkdownEmitter) do
1153 var tpl = v.push_buffer
1154 var line = block.first_line
1155 while line != null do
1156 if not line.is_empty then
1157 v.add line.value.substring(line.leading, line.value.length - line.trailing)
1158 if line.trailing >= 2 then v.decorator.add_line_break(v)
1160 if line.next != null then
1169 # Emit sub-blocks contained in `block`.
1170 fun emit_blocks(v: MarkdownEmitter) do
1171 var block = self.block.first_block
1172 while block != null do
1173 v.push_loc(block.location)
1180 # The raw content of the block as a multi-line string.
1181 fun raw_content: String do
1182 var infence = self isa BlockFence
1183 var text = new FlatBuffer
1184 var line = self.block.first_line
1185 while line != null do
1186 if not line.is_empty then
1187 var str = line.value
1188 if not infence and str.has_prefix(" ") then
1189 text.append str.substring(4, str.length - line.trailing)
1197 return text.write_to_string
1201 # A block without any markdown specificities.
1203 # Actually use the same implementation than `BlockCode`,
1204 # this class is only used for typing purposes.
1209 # A markdown blockquote.
1213 redef fun emit(v) do v.decorator.add_blockquote(v, self)
1215 # Remove blockquote markers.
1216 private fun remove_block_quote_prefix(block: MDBlock) do
1217 var line = block.first_line
1218 while line != null do
1219 if not line.is_empty then
1220 if line.value[line.leading] == '>' then
1221 var rem = line.leading + 1
1222 if line.leading + 1 < line.value.length and
1223 line.value[line.leading + 1] == ' ' then
1226 line.value = line.value.substring_from(rem)
1227 line.leading = line.process_leading
1235 # A markdown code block.
1239 # Any string found after fence token.
1240 var meta: nullable Text
1242 # Number of char to skip at the beginning of the line.
1244 # Block code lines start at 4 spaces.
1245 protected var line_start = 4
1247 redef fun emit(v) do v.decorator.add_code(v, self)
1249 redef fun emit_lines(v) do
1250 var line = block.first_line
1251 while line != null do
1252 if not line.is_empty then
1253 v.decorator.append_code(v, line.value, line_start, line.value.length)
1261 # A markdown code-fence block.
1263 # Actually use the same implementation than `BlockCode`,
1264 # this class is only used for typing purposes.
1268 # Fence code lines start at 0 spaces.
1269 redef var line_start = 0
1272 # A markdown headline.
1276 redef fun emit(v) do
1277 var loc = block.location.copy
1278 loc.column_start += start
1280 v.decorator.add_headline(v, self)
1284 private var start = 0
1286 # Depth of the headline used to determine the headline level.
1289 # Remove healine marks from lines contained in `self`.
1290 private fun transform_headline(block: MDBlock) do
1291 if depth > 0 then return
1293 var line = block.first_line
1294 if line.is_empty then return
1295 var start = line.leading
1296 while start < line.value.length and line.value[start] == '#' do
1300 while start
< line
.value
.length
and line
.value
[start
] == ' ' do
1303 if start
>= line
.value
.length
then
1304 line
.is_empty
= true
1306 var nend
= line
.value
.length
- line
.trailing
- 1
1307 while line
.value
[nend
] == '#' do nend
-= 1
1308 while line
.value
[nend
] == ' ' do nend
-= 1
1309 line
.value
= line
.value
.substring
(start
, nend
- start
+ 1)
1314 depth
= level
.min
(6)
1318 # A markdown list item block.
1322 redef fun emit
(v
) do v
.decorator
.add_listitem
(v
, self)
1325 # A markdown list block.
1326 # Can be either an ordered or unordered list, this class is mainly used to factorize code.
1327 abstract class BlockList
1330 # Split list block into list items sub-blocks.
1331 private fun init_block
(v
: MarkdownProcessor) do
1332 var line
= block
.first_line
1334 while line
!= null do
1335 var t
= v
.line_kind
(line
)
1336 if t
isa LineList or
1337 (not line
.is_empty
and (line
.prev_empty
and line
.leading
== 0 and
1338 not (t
isa LineList))) then
1339 var sblock
= block
.split
(line
.prev
.as(not null))
1340 sblock
.kind
= new BlockListItem(sblock
)
1344 var sblock
= block
.split
(block
.last_line
.as(not null))
1345 sblock
.kind
= new BlockListItem(sblock
)
1348 # Expand list items as paragraphs if needed.
1349 private fun expand_paragraphs
(block
: MDBlock) do
1350 var outer
= block
.first_block
1351 var inner
: nullable MDBlock
1352 var has_paragraph
= false
1353 while outer
!= null and not has_paragraph
do
1354 if outer
.kind
isa BlockListItem then
1355 inner
= outer
.first_block
1356 while inner
!= null and not has_paragraph
do
1357 if inner
.kind
isa BlockParagraph then
1358 has_paragraph
= true
1365 if has_paragraph
then
1366 outer
= block
.first_block
1367 while outer
!= null do
1368 if outer
.kind
isa BlockListItem then
1369 inner
= outer
.first_block
1370 while inner
!= null do
1371 if inner
.kind
isa BlockNone then
1372 inner
.kind
= new BlockParagraph(inner
)
1383 # A markdown ordered list.
1384 class BlockOrderedList
1387 redef fun emit
(v
) do v
.decorator
.add_orderedlist
(v
, self)
1390 # A markdown unordred list.
1391 class BlockUnorderedList
1394 redef fun emit
(v
) do v
.decorator
.add_unorderedlist
(v
, self)
1397 # A markdown paragraph block.
1398 class BlockParagraph
1401 redef fun emit
(v
) do v
.decorator
.add_paragraph
(v
, self)
1408 redef fun emit
(v
) do v
.decorator
.add_ruler
(v
, self)
1411 # Xml blocks that can be found in markdown markup.
1415 redef fun emit_lines
(v
) do
1416 var line
= block
.first_line
1417 while line
!= null do
1418 if not line
.is_empty
then v
.add line
.value
1428 # Location of `self` in the original input.
1429 var location
: MDLocation
1431 # Text contained in this line.
1432 var value
: String is writable
1434 # Is this line empty?
1435 # Lines containing only spaces are considered empty.
1436 var is_empty
: Bool = true is writable
1438 # Previous line in `MDBlock` or null if first line.
1439 var prev
: nullable MDLine = null is writable
1441 # Next line in `MDBlock` or null if last line.
1442 var next
: nullable MDLine = null is writable
1444 # Is the previous line empty?
1445 var prev_empty
: Bool = false is writable
1447 # Is the next line empty?
1448 var next_empty
: Bool = false is writable
1450 # Initialize a new MDLine from its string value
1452 self.leading
= process_leading
1453 if leading
!= value
.length
then
1454 self.is_empty
= false
1455 self.trailing
= process_trailing
1459 # Set `value` as an empty String and update `leading`, `trailing` and is_`empty`.
1465 if prev
!= null then prev
.next_empty
= true
1466 if next
!= null then next
.prev_empty
= true
1469 # Number or leading spaces on this line.
1470 var leading
: Int = 0 is writable
1472 # Compute `leading` depending on `value`.
1473 fun process_leading
: Int do
1475 var value
= self.value
1476 while count
< value
.length
and value
[count
] == ' ' do count
+= 1
1477 if leading
== value
.length
then clear
1481 # Number of trailing spaces on this line.
1482 var trailing
: Int = 0 is writable
1484 # Compute `trailing` depending on `value`.
1485 fun process_trailing
: Int do
1487 var value
= self.value
1488 while value
[value
.length
- count
- 1] == ' ' do
1494 # Count the amount of `ch` in this line.
1495 # Return A value > 0 if this line only consists of `ch` end spaces.
1496 fun count_chars
(ch
: Char): Int do
1512 # Count the amount of `ch` at the start of this line ignoring spaces.
1513 fun count_chars_start
(ch
: Char): Int do
1528 # Last XML line if any.
1529 private var xml_end_line
: nullable MDLine = null
1531 # Does `value` contains valid XML markup?
1532 private fun check_html
: Bool do
1533 var tags
= new Array[String]
1534 var tmp
= new FlatBuffer
1536 if pos
+ 1 < value
.length
and value
[pos
+ 1] == '!' then
1537 if read_xml_comment
(self, pos
) > 0 then return true
1539 pos
= value
.read_xml
(tmp
, pos
, false)
1543 if not tag
.is_html_block
then
1551 var line
: nullable MDLine = self
1552 while line
!= null do
1553 while pos
< line
.value
.length
and line
.value
[pos
] != '<' do
1556 if pos
>= line
.value
.length
then
1557 if pos
- 2 >= 0 and line
.value
[pos
- 2] == '/' then
1559 if tags
.is_empty
then
1567 tmp
= new FlatBuffer
1568 var new_pos
= line
.value
.read_xml
(tmp
, pos
, false)
1571 if tag
.is_html_block
and not tag
== "hr" then
1572 if tmp
[1] == '/' then
1573 if tags
.last
!= tag
then
1581 if tags
.is_empty
then
1591 return tags
.is_empty
1596 # Read a XML comment.
1597 # Used by `check_html`.
1598 private fun read_xml_comment
(first_line
: MDLine, start
: Int): Int do
1599 var line
: nullable MDLine = first_line
1600 if start
+ 3 < line
.value
.length
then
1601 if line
.value
[2] == '-' and line
.value
[3] == '-' then
1603 while line
!= null do
1604 while pos
< line
.value
.length
and line
.value
[pos
] != '-' do
1607 if pos
== line
.value
.length
then
1611 if pos
+ 2 < line
.value
.length
then
1612 if line
.value
[pos
+ 1] == '-' and line
.value
[pos
+ 2] == '>' then
1613 first_line
.xml_end_line
= line
1625 # Extract the text of `self` without leading and trailing.
1626 fun text
: String do return value
.substring
(leading
, value
.length
- trailing
)
1633 # See `MarkdownProcessor::recurse`.
1634 fun process
(v
: MarkdownProcessor) is abstract
1637 # An empty markdown line.
1641 redef fun process
(v
) do
1642 v
.current_line
= v
.current_line
.next
1646 # A non-specific markdown construction.
1647 # Mainly used as part of another line construct such as paragraphs or lists.
1651 redef fun process
(v
) do
1652 var line
= v
.current_line
1654 var was_empty
= line
.prev_empty
1655 while line
!= null and not line
.is_empty
do
1656 var t
= v
.line_kind
(line
)
1657 if (v
.in_list
or v
.ext_mode
) and t
isa LineList then
1660 if v
.ext_mode
and (t
isa LineCode or t
isa LineFence) then
1663 if t
isa LineHeadline or t
isa LineHeadline1 or t
isa LineHeadline2 or
1664 t
isa LineHR or t
isa LineBlockquote or t
isa LineXML then
1670 if line
!= null and not line
.is_empty
then
1671 var block
= v
.current_block
.split
(line
.prev
.as(not null))
1672 if v
.in_list
and not was_empty
then
1673 block
.kind
= new BlockNone(block
)
1675 block
.kind
= new BlockParagraph(block
)
1677 v
.current_block
.remove_leading_empty_lines
1680 if line
!= null then
1681 block
= v
.current_block
.split
(line
)
1683 block
= v
.current_block
.split
(v
.current_block
.last_line
.as(not null))
1685 if v
.in_list
and (line
== null or not line
.is_empty
) and not was_empty
then
1686 block
.kind
= new BlockNone(block
)
1688 block
.kind
= new BlockParagraph(block
)
1690 v
.current_block
.remove_leading_empty_lines
1692 v
.current_line
= v
.current_block
.first_line
1696 # A line of markdown code.
1700 redef fun process
(v
) do
1701 var line
= v
.current_line
1703 while line
!= null and (line
.is_empty
or v
.line_kind
(line
) isa LineCode) do
1706 # split at block end line
1708 if line
!= null then
1709 block
= v
.current_block
.split
(line
.prev
.as(not null))
1711 block
= v
.current_block
.split
(v
.current_block
.last_line
.as(not null))
1713 block
.kind
= new BlockCode(block
)
1714 block
.remove_surrounding_empty_lines
1715 v
.current_line
= v
.current_block
.first_line
1719 # A line of raw XML.
1723 redef fun process
(v
) do
1724 var line
= v
.current_line
1725 var prev
= line
.prev
1726 if prev
!= null then v
.current_block
.split
(prev
)
1727 var block
= v
.current_block
.split
(line
.xml_end_line
.as(not null))
1728 block
.kind
= new BlockXML(block
)
1729 v
.current_block
.remove_leading_empty_lines
1730 v
.current_line
= v
.current_block
.first_line
1734 # A markdown blockquote line.
1735 class LineBlockquote
1738 redef fun process
(v
) do
1739 var line
= v
.current_line
1741 while line
!= null do
1742 if not line
.is_empty
and (line
.prev_empty
and
1743 line
.leading
== 0 and
1744 not v
.line_kind
(line
) isa LineBlockquote) then break
1749 if line
!= null then
1750 block
= v
.current_block
.split
(line
.prev
.as(not null))
1752 block
= v
.current_block
.split
(v
.current_block
.last_line
.as(not null))
1754 var kind
= new BlockQuote(block
)
1756 block
.remove_surrounding_empty_lines
1757 kind
.remove_block_quote_prefix
(block
)
1758 v
.current_line
= line
1759 v
.recurse
(block
, false)
1760 v
.current_line
= v
.current_block
.first_line
1764 # A markdown ruler line.
1768 redef fun process
(v
) do
1769 var line
= v
.current_line
1770 if line
.prev
!= null then v
.current_block
.split
(line
.prev
.as(not null))
1771 var block
= v
.current_block
.split
(line
.as(not null))
1772 block
.kind
= new BlockRuler(block
)
1773 v
.current_block
.remove_leading_empty_lines
1774 v
.current_line
= v
.current_block
.first_line
1778 # A markdown fence code line.
1782 redef fun process
(v
) do
1784 var line
= v
.current_line
.next
1785 while line
!= null do
1786 if v
.line_kind
(line
) isa LineFence then break
1789 if line
!= null then
1794 if line
!= null then
1795 block
= v
.current_block
.split
(line
.prev
.as(not null))
1797 block
= v
.current_block
.split
(v
.current_block
.last_line
.as(not null))
1799 block
.remove_surrounding_empty_lines
1800 var meta
= block
.first_line
.value
.meta_from_fence
1801 block
.kind
= new BlockFence(block
, meta
)
1802 block
.first_line
.clear
1803 var last
= block
.last_line
1804 if last
!= null and v
.line_kind
(last
) isa LineFence then
1805 block
.last_line
.clear
1807 block
.remove_surrounding_empty_lines
1808 v
.current_line
= line
1812 # A markdown headline.
1816 redef fun process
(v
) do
1817 var line
= v
.current_line
1818 var lprev
= line
.prev
1819 if lprev
!= null then v
.current_block
.split
(lprev
)
1820 var block
= v
.current_block
.split
(line
.as(not null))
1821 var kind
= new BlockHeadline(block
)
1823 kind
.transform_headline
(block
)
1824 v
.current_block
.remove_leading_empty_lines
1825 v
.current_line
= v
.current_block
.first_line
1829 # A markdown headline of level 1.
1833 redef fun process
(v
) do
1834 var line
= v
.current_line
1835 var lprev
= line
.prev
1836 if lprev
!= null then v
.current_block
.split
(lprev
)
1838 var block
= v
.current_block
.split
(line
.as(not null))
1839 var kind
= new BlockHeadline(block
)
1841 kind
.transform_headline
(block
)
1843 v
.current_block
.remove_leading_empty_lines
1844 v
.current_line
= v
.current_block
.first_line
1848 # A markdown headline of level 2.
1852 redef fun process
(v
) do
1853 var line
= v
.current_line
1854 var lprev
= line
.prev
1855 if lprev
!= null then v
.current_block
.split
(lprev
)
1857 var block
= v
.current_block
.split
(line
.as(not null))
1858 var kind
= new BlockHeadline(block
)
1860 kind
.transform_headline
(block
)
1862 v
.current_block
.remove_leading_empty_lines
1863 v
.current_line
= v
.current_block
.first_line
1867 # A markdown list line.
1868 # Mainly used to factorize code between ordered and unordered lists.
1869 abstract class LineList
1872 redef fun process
(v
) do
1873 var line
= v
.current_line
1875 while line
!= null do
1876 var t
= v
.line_kind
(line
)
1877 if not line
.is_empty
and (line
.prev_empty
and line
.leading
== 0 and
1878 not t
isa LineList) then break
1883 if line
!= null then
1884 list
= v
.current_block
.split
(line
.prev
.as(not null))
1886 list
= v
.current_block
.split
(v
.current_block
.last_line
.as(not null))
1888 var kind
= block_kind
(list
)
1890 list
.first_line
.prev_empty
= false
1891 list
.last_line
.next_empty
= false
1892 list
.remove_surrounding_empty_lines
1893 list
.first_line
.prev_empty
= false
1894 list
.last_line
.next_empty
= false
1896 var block
= list
.first_block
1897 while block
!= null do
1898 block
.remove_list_indent
(v
)
1899 v
.recurse
(block
, true)
1902 kind
.expand_paragraphs
(list
)
1903 v
.current_line
= line
1906 # Create a new block kind based on this line.
1907 protected fun block_kind
(block
: MDBlock): BlockList is abstract
1909 # Extract string value from `MDLine`.
1910 protected fun extract_value
(line
: MDLine): String is abstract
1913 # An ordered list line.
1917 redef fun block_kind
(block
) do return new BlockOrderedList(block
)
1919 redef fun extract_value
(line
) do
1920 return line
.value
.substring_from
(line
.value
.index_of
('.') + 2)
1924 # An unordered list line.
1928 redef fun block_kind
(block
) do return new BlockUnorderedList(block
)
1930 redef fun extract_value
(line
) do
1931 return line
.value
.substring_from
(line
.leading
+ 2)
1935 # A token represent a character in the markdown input.
1936 # Some tokens have a specific markup behaviour that is handled here.
1937 abstract class Token
1939 # Location of `self` in the original input.
1940 var location
: MDLocation
1942 # Position of `self` in input independant from lines.
1945 # Character found at `pos` in the markdown input.
1948 # Output that token using `MarkdownEmitter::decorator`.
1949 fun emit
(v
: MarkdownEmitter) do v
.decorator
.add_char
(v
, char
)
1952 # A token without a specific meaning.
1957 # An emphasis token.
1958 abstract class TokenEm
1961 redef fun emit
(v
) do
1962 var tmp
= v
.push_buffer
1963 var b
= v
.emit_text_until
(v
.current_text
.as(not null), pos
+ 1, self)
1966 v
.decorator
.add_em
(v
, tmp
)
1974 # An emphasis star token.
1979 # An emphasis underscore token.
1980 class TokenEmUnderscore
1985 abstract class TokenStrong
1988 redef fun emit
(v
) do
1989 var tmp
= v
.push_buffer
1990 var b
= v
.emit_text_until
(v
.current_text
.as(not null), pos
+ 2, self)
1993 v
.decorator
.add_strong
(v
, tmp
)
1994 v
.current_pos
= b
+ 1
2001 # A strong star token.
2002 class TokenStrongStar
2006 # A strong underscore token.
2007 class TokenStrongUnderscore
2012 # This class is mainly used to factorize work between single and double quoted span codes.
2013 abstract class TokenCode
2016 redef fun emit
(v
) do
2017 var a
= pos
+ next_pos
+ 1
2018 var b
= v
.processor
.find_token
(v
.current_text
.as(not null), a
, self)
2020 v
.current_pos
= b
+ next_pos
2021 while a
< b
and v
.current_text
[a
] == ' ' do a
+= 1
2023 while v
.current_text
[b
- 1] == ' ' do b
-= 1
2024 v
.decorator
.add_span_code
(v
, v
.current_text
.as(not null), a
, b
)
2031 private fun next_pos
: Int is abstract
2034 # A span code token.
2035 class TokenCodeSingle
2038 redef fun next_pos
do return 0
2041 # A doubled span code token.
2042 class TokenCodeDouble
2045 redef fun next_pos
do return 1
2048 # A link or image token.
2049 # This class is mainly used to factorize work between images and links.
2050 abstract class TokenLinkOrImage
2054 var link
: nullable Text = null
2057 var name
: nullable Text = null
2060 var comment
: nullable Text = null
2062 # Is the link construct an abbreviation?
2063 var is_abbrev
= false
2065 redef fun emit
(v
) do
2066 var tmp
= new FlatBuffer
2067 var b
= check_link
(v
, tmp
, pos
, self)
2076 # Emit the hyperlink as link or image.
2077 private fun emit_hyper
(v
: MarkdownEmitter) is abstract
2079 # Check if the link is a valid link.
2080 private fun check_link
(v
: MarkdownEmitter, out
: FlatBuffer, start
: Int, token
: Token): Int do
2081 var md
= v
.current_text
2083 if token
isa TokenLink then
2088 var tmp
= new FlatBuffer
2089 pos
= md
.read_md_link_id
(tmp
, pos
)
2090 if pos
< start
then return -1
2094 pos
= md
.skip_spaces
(pos
)
2096 var tid
= name
.write_to_string
.to_lower
2097 if v
.processor
.link_refs
.has_key
(tid
) then
2098 var lr
= v
.processor
.link_refs
[tid
]
2099 is_abbrev
= lr
.is_abbrev
2106 else if md
[pos
] == '(' then
2108 pos
= md
.skip_spaces
(pos
)
2109 if pos
< start
then return -1
2110 tmp
= new FlatBuffer
2111 var use_lt
= md
[pos
] == '<'
2113 pos
= md
.read_until
(tmp
, pos
+ 1, '>')
2115 pos
= md
.read_md_link
(tmp
, pos
)
2117 if pos
< start
then return -1
2118 if use_lt
then pos
+= 1
2119 link
= tmp
.write_to_string
2120 if md
[pos
] == ' ' then
2121 pos
= md
.skip_spaces
(pos
)
2122 if pos
> start
and md
[pos
] == '"' then
2124 tmp
= new FlatBuffer
2125 pos
= md
.read_until
(tmp
, pos
, '"')
2126 if pos
< start
then return -1
2127 comment
= tmp
.write_to_string
2129 pos
= md
.skip_spaces
(pos
)
2130 if pos
== -1 then return -1
2133 if pos
< start
then return -1
2134 if md
[pos
] != ')' then return -1
2135 else if md
[pos
] == '[' then
2137 tmp
= new FlatBuffer
2138 pos
= md
.read_raw_until
(tmp
, pos
, ']')
2139 if pos
< start
then return -1
2141 if tmp
.length
> 0 then
2146 var tid
= id
.write_to_string
.to_lower
2147 if v
.processor
.link_refs
.has_key
(tid
) then
2148 var lr
= v
.processor
.link_refs
[tid
]
2153 var tid
= name
.write_to_string
.replace
("\n", " ").to_lower
2154 if v
.processor
.link_refs
.has_key
(tid
) then
2155 var lr
= v
.processor
.link_refs
[tid
]
2163 if link
== null then return -1
2168 # A markdown link token.
2170 super TokenLinkOrImage
2172 redef fun emit_hyper
(v
) do
2173 if is_abbrev
and comment
!= null then
2174 v
.decorator
.add_abbr
(v
, name
.as(not null), comment
.as(not null))
2176 v
.decorator
.add_link
(v
, link
.as(not null), name
.as(not null), comment
)
2181 # A markdown image token.
2183 super TokenLinkOrImage
2185 redef fun emit_hyper
(v
) do
2186 v
.decorator
.add_image
(v
, link
.as(not null), name
.as(not null), comment
)
2194 redef fun emit
(v
) do
2195 var tmp
= new FlatBuffer
2196 var b
= check_html
(v
, tmp
, v
.current_text
.as(not null), v
.current_pos
)
2201 v
.decorator
.escape_char
(v
, char
)
2205 # Is the HTML valid?
2206 # Also take care of link and mailto shortcuts.
2207 private fun check_html
(v
: MarkdownEmitter, out
: FlatBuffer, md
: Text, start
: Int): Int do
2208 # check for auto links
2209 var tmp
= new FlatBuffer
2210 var pos
= md
.read_until
(tmp
, start
+ 1, ':', ' ', '>', '\n')
2211 if pos
!= -1 and md
[pos
] == ':' and tmp
.is_link_prefix
then
2212 pos
= md
.read_until
(tmp
, pos
, '>')
2214 var link
= tmp
.write_to_string
2215 v
.decorator
.add_link
(v
, link
, link
, null)
2219 # TODO check for mailto
2220 # check for inline html
2221 if start
+ 2 < md
.length
then
2222 return md
.read_xml
(out
, start
, true)
2228 # An HTML entity token.
2232 redef fun emit
(v
) do
2233 var tmp
= new FlatBuffer
2234 var b
= check_entity
(tmp
, v
.current_text
.as(not null), pos
)
2239 v
.decorator
.escape_char
(v
, char
)
2243 # Is the entity valid?
2244 private fun check_entity
(out
: FlatBuffer, md
: Text, start
: Int): Int do
2245 var pos
= md
.read_until
(out
, start
, ';')
2246 if pos
< 0 or out
.length
< 3 then
2249 if out
[1] == '#' then
2250 if out
[2] == 'x' or out
[2] == 'X' then
2251 if out
.length
< 4 then return -1
2252 for i
in [3..out
.length
[ do
2254 if (c
< '0' or c
> '9') and (c
< 'a' and c
> 'f') and (c
< 'A' and c
> 'F') then
2259 for i
in [2..out
.length
[ do
2261 if c
< '0' or c
> '9' then return -1
2266 for i
in [1..out
.length
[ do
2268 if not c
.is_digit
and not c
.is_letter
then return -1
2271 # TODO check entity is valid
2272 # if out.is_entity then
2282 # A markdown escape token.
2286 redef fun emit
(v
) do
2288 v
.addc v
.current_text
[v
.current_pos
]
2292 # A markdown strike token.
2294 # Extended mode only (see `MarkdownProcessor::ext_mode`)
2298 redef fun emit
(v
) do
2299 var tmp
= v
.push_buffer
2300 var b
= v
.emit_text_until
(v
.current_text
.as(not null), pos
+ 2, self)
2303 v
.decorator
.add_strike
(v
, tmp
)
2304 v
.current_pos
= b
+ 1
2313 # Get the position of the next non-space character.
2314 private fun skip_spaces
(start
: Int): Int do
2316 while pos
> -1 and pos
< length
and (self[pos
] == ' ' or self[pos
] == '\n') do
2319 if pos
< length
then return pos
2323 # Read `self` until `nend` and append it to the `out` buffer.
2324 # Escape markdown special chars.
2325 private fun read_until
(out
: FlatBuffer, start
: Int, nend
: Char...): Int do
2327 while pos
< length
do
2329 if c
== '\\' and pos
+ 1 < length
then
2330 pos
= escape
(out
, self[pos
+ 1], pos
)
2332 var end_reached
= false
2339 if end_reached
then break
2344 if pos
== length
then return -1
2348 # Read `self` as raw text until `nend` and append it to the `out` buffer.
2349 # No escape is made.
2350 private fun read_raw_until
(out
: FlatBuffer, start
: Int, nend
: Char...): Int do
2352 while pos
< length
do
2354 var end_reached
= false
2361 if end_reached
then break
2365 if pos
== length
then return -1
2369 # Read `self` as XML until `to` and append it to the `out` buffer.
2370 # Escape HTML special chars.
2371 private fun read_xml_until
(out
: FlatBuffer, from
: Int, to
: Char...): Int do
2374 var str_char
: nullable Char = null
2375 while pos
< length
do
2381 if pos
< length
then
2387 if c
== str_char
then
2394 if c
== '"' or c
== '\'' then
2399 var end_reached = false
2400 for n in [0..to.length[ do
2406 if end_reached then break
2411 if pos == length then return -1
2415 # Read `self` as XML and append it to the `out` buffer.
2416 # Safe mode can be activated to limit reading to valid xml.
2417 private fun read_xml(out: FlatBuffer, start: Int, safe_mode: Bool): Int do
2420 var is_close_tag = false
2421 if start + 1 >= length then return -1
2422 if self[start + 1] == '/' then
2425 else if self[start + 1] == '!' then
2429 is_close_tag = false
2433 var tmp = new FlatBuffer
2434 pos = read_xml_until(tmp, pos, ' ', '/', '>')
2435 if pos == -1 then return -1
2436 var tag = tmp.write_to_string.trim.to_lower
2437 if not tag.is_valid_html_tag then
2440 else if tag.is_html_unsafe then
2443 if is_close_tag then out.add '/'
2447 if is_close_tag then out.add '/'
2452 if is_close_tag then out.add '/'
2453 pos = read_xml_until(out, pos, ' ', '/', '>')
2455 if pos == -1 then return -1
2456 pos = read_xml_until(out, pos, '/', '>')
2457 if pos == -1 then return -1
2458 if self[pos] == '/' then
2460 pos = self.read_xml_until(out, pos + 1, '>')
2461 if pos == -1 then return -1
2463 if self[pos] == '>' then
2474 # Read a markdown link address and append it to the `out` buffer.
2475 private fun read_md_link(out: FlatBuffer, start: Int): Int do
2478 while pos < length do
2480 if c == '\\
' and pos + 1 < length then
2481 pos = escape(out, self[pos + 1], pos)
2483 var end_reached = false
2486 else if c == ' ' then
2487 if counter == 1 then end_reached = true
2488 else if c == ')' then
2490 if counter == 0 then end_reached = true
2492 if end_reached then break
2497 if pos == length then return -1
2501 # Read a markdown link text and append it to the `out` buffer.
2502 private fun read_md_link_id(out: FlatBuffer, start: Int): Int do
2505 while pos < length do
2507 var end_reached = false
2511 else if c == ']' then
2513 if counter == 0 then
2521 if end_reached then break
2524 if pos == length then return -1
2528 # Extract the XML tag name from a XML tag.
2529 private fun xml_tag: String do
2530 var tpl = new FlatBuffer
2532 if pos < length and self[1] == '/' then pos += 1
2533 while pos < length - 1 and (self[pos].is_digit or self[pos].is_letter) do
2537 return tpl.write_to_string.to_lower
2540 private fun is_valid_html_tag: Bool do
2541 if is_empty then return false
2543 if not c.is_alpha then return false
2548 # Read and escape the markdown contained in `self`.
2549 private fun escape(out: FlatBuffer, c: Char, pos: Int): Int do
2550 if c == '\\
' or c == '[' or c == ']' or c == '(' or c == ')' or c == '{' or
2551 c == '}' or c == '#' or c == '"' or c == '\'' or c == '.' or c == '<' or
2552 c
== '>' or c
== '*' or c
== '+' or c
== '-' or c
== '_' or c
== '!' or
2553 c
== '`' or c
== '~' or c
== '^' then
2561 # Extract string found at end of fence opening.
2562 private fun meta_from_fence
: nullable Text do
2563 for i
in [0..chars
.length
[ do
2565 if c
!= ' ' and c
!= '`' and c
!= '~' then
2566 return substring_from
(i
).trim
2572 # Is `self` an unsafe HTML element?
2573 private fun is_html_unsafe
: Bool do return html_unsafe_tags
.has
(self.write_to_string
)
2575 # Is `self` a HRML block element?
2576 private fun is_html_block
: Bool do return html_block_tags
.has
(self.write_to_string
)
2578 # Is `self` a link prefix?
2579 private fun is_link_prefix
: Bool do return html_link_prefixes
.has
(self.write_to_string
)
2581 private fun html_unsafe_tags
: Array[String] do return once
["applet", "head", "body", "frame", "frameset", "iframe", "script", "object"]
2583 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"]
2585 private fun html_link_prefixes
: Array[String] do return once
["http", "https", "ftp", "ftps"]
2590 # Parse `self` as markdown and return the HTML representation
2592 # var md = "**Hello World!**"
2593 # var html = md.md_to_html
2594 # assert html == "<p><strong>Hello World!</strong></p>\n"
2595 fun md_to_html
: Writable do
2596 var processor
= new MarkdownProcessor
2597 return processor
.process
(self)