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 # Work in extended mode (default).
35 # Behavior changes when using extended mode:
37 # * Lists and code blocks end a paragraph
39 # In normal markdown the following:
43 # * and this is not a list
49 # <p>This is a paragraph
50 # * and this is not a list</p>
53 # When using extended mode this changes to:
56 # <p>This is a paragraph</p>
58 # <li>and this is not a list</li>
62 # * Fences code blocks
64 # If you don't want to indent your all your code with 4 spaces,
65 # you can wrap your code in ``` ``` ``` or `~~~`.
71 # print "Hello World!"
77 # If you want to use syntax highlighting tools, most of them need to know what kind
78 # of language they are highlighting.
79 # You can add an optional language identifier after the fence declaration to output
80 # it in the HTML render.
85 # print "# Hello World!".md_to_html
91 # <pre class="nit"><code>import markdown
93 # print "Hello World!".md_to_html
97 # * Underscores (Emphasis)
99 # Underscores in the middle of a word like:
105 # normally produces this:
108 # <p>Con<em>cat</em>this</p>
111 # With extended mode they don't result in emphasis.
114 # <p>Con_cat_this</p>
119 # Like in [GFM](https://help.github.com/articles/github-flavored-markdown),
120 # strikethrought span is marked with `~~`.
129 # <del>Mistaken text.</del>
133 # Disable attaching MDLocation to Tokens
135 # Locations are useful for some tools but they may
136 # cause an important time and space overhead.
139 var no_location
= false is writable
141 # Process the mardown `input` string and return the processed output.
142 fun process
(input
: String): Writable do
149 var parent
= read_lines
(input
)
150 parent
.remove_surrounding_empty_lines
151 recurse
(parent
, false)
152 # output processed text
153 decorator
.headlines
.clear
154 return emit
(parent
.kind
)
157 # Split `input` string into `MDLines` and create a parent `MDBlock` with it.
158 private fun read_lines
(input
: String): MDBlock do
159 var block
= new MDBlock(new MDLocation(1, 1, 1, 1))
160 var value
= new FlatBuffer
166 while i
< input
.length
do
170 while not eol
and i
< input
.length
do
175 else if c
== '\r' then
176 else if c
== '\t' then
177 var np
= pos
+ (4 - (pos
& 3))
190 var loc
= new MDLocation(line_pos
, 1, line_pos
, col_pos
)
191 var line
= new MDLine(loc
, value
.write_to_string
)
192 var is_link_ref
= check_link_ref
(line
)
194 if not is_link_ref
then block
.add_line line
200 # Check if line is a block link definition.
201 # Return `true` if line contains a valid link ref and save it into `link_refs`.
202 private fun check_link_ref
(line
: MDLine): Bool do
204 var is_link_ref
= false
205 var id
= new FlatBuffer
206 var link
= new FlatBuffer
207 var comment
= new FlatBuffer
209 if not line
.is_empty
and line
.leading
< 4 and line
.value
[line
.leading
] == '[' then
210 pos
= line
.leading
+ 1
211 pos
= md
.read_until
(id
, pos
, ']')
212 if not id
.is_empty
and pos
>= 0 and pos
+ 2 < line
.value
.length
then
213 if line
.value
[pos
+ 1] == ':' then
215 pos
= md
.skip_spaces
(pos
)
216 if pos
>= 0 and line
.value
[pos
] == '<' then
218 pos
= md
.read_until
(link
, pos
, '>')
220 else if pos
>= 0 then
221 pos
= md
.read_until
(link
, pos
, ' ', '\n')
223 if not link
.is_empty
then
224 pos
= md
.skip_spaces
(pos
)
225 if pos
> 0 and pos
< line
.value
.length
then
226 var c
= line
.value
[pos
]
227 if c
== '\"' or c
== '\'' or c == '(' then
230 pos = md.read_until(comment, pos, ')')
232 pos = md.read_until(comment, pos, c)
234 if pos > 0 then is_link_ref = true
243 if is_link_ref and not id.is_empty and not link.is_empty then
244 var lr = new LinkRef.with_title(link.write_to_string, comment.write_to_string)
245 add_link_ref(id.write_to_string, lr)
246 if comment.is_empty then last_link_ref = lr
249 comment = new FlatBuffer
250 if not line.is_empty and last_link_ref != null then
252 var c = line.value[pos]
253 if c == '\
"' or c == '\'' or c == '(' then
256 pos = md.read_until(comment, pos, ')')
258 pos = md.read_until(comment, pos, c)
261 var last_link_ref = self.last_link_ref
262 if not comment.is_empty and last_link_ref != null then
263 last_link_ref.title = comment.write_to_string
266 if comment.is_empty then return false
272 # This list will be needed during output to expand links.
273 var link_refs: Map[String, LinkRef] = new HashMap[String, LinkRef]
275 # Last encountered link ref (for multiline definitions)
277 # Markdown allows link refs to be defined over two lines:
280 # [id]: http://example.com/longish/path/to/resource/here
281 # "Optional Title Here"
284 private var last_link_ref: nullable LinkRef = null
286 # Add a link ref to the list
287 fun add_link_ref(key: String, ref: LinkRef) do link_refs[key.to_lower] = ref
289 # Recursively split a `block`.
291 # The block is splitted according to the type of lines it contains.
292 # Some blocks can be splited again recursively like lists.
293 # The `in_list` mode is used to recurse on list and build
294 # nested paragraphs or code blocks.
295 fun recurse(root: MDBlock, in_list: Bool) do
296 var old_mode = self.in_list
297 var old_root = self.current_block
298 self.in_list = in_list
300 var line = root.first_line
301 while line != null and line.is_empty do
303 if line == null then return
308 while current_line != null do
309 line_kind(current_line.as(not null)).process(self)
311 self.in_list = old_mode
312 self.current_block = old_root
315 # Currently processed line.
316 # Used when visiting blocks with `recurse`.
317 var current_line: nullable MDLine = null is writable
319 # Currently processed block.
320 # Used when visiting blocks with `recurse`.
321 var current_block: nullable MDBlock = null is writable
323 # Is the current recursion in list mode?
324 # Used when visiting blocks with `recurse`
325 private var in_list = false
329 fun line_kind(md: MDLine): Line do
331 var leading = md.leading
332 var trailing = md.trailing
333 if md.is_empty then return new LineEmpty
334 if md.leading > 3 then return new LineCode
335 if value[leading] == '#' then return new LineHeadline
336 if value[leading] == '>' then return new LineBlockquote
339 if value.length - leading - trailing > 2 then
340 if value[leading] == '`' and md.count_chars_start('`') >= 3 then
343 if value[leading] == '~' and md.count_chars_start('~') >= 3 then
349 if value.length - leading - trailing > 2 and
350 (value[leading] == '*' or value[leading] == '-' or value[leading] == '_') then
351 if md.count_chars(value[leading]) >= 3 then
356 if value.length - leading >= 2 and value[leading + 1] == ' ' then
357 var c = value[leading]
358 if c == '*' or c == '-' or c == '+' then return new LineUList
361 if value.length - leading >= 3 and value[leading].is_digit then
363 while i < value.length and value[i].is_digit do i += 1
364 if i + 1 < value.length and value[i] == '.' and value[i + 1] == ' ' then
369 if value[leading] == '<' and md.check_html then return new LineXML
372 if next != null and not next.is_empty then
373 if next.count_chars('=') > 0 then
374 return new LineHeadline1
376 if next.count_chars('-') > 0 then
377 return new LineHeadline2
383 # Get the token kind at `pos`.
384 fun token_at(text: Text, pos: Int): Token do
396 if pos + 1 < text.length then
401 if pos + 2 < text.length then
411 loc = new MDLocation(
412 current_loc.line_start,
413 current_loc.column_start + pos,
414 current_loc.line_start,
415 current_loc.column_start + pos)
420 if c0 != ' ' or c2 != ' ' then
421 return new TokenStrongStar(loc, pos, c)
423 return new TokenEmStar(loc, pos, c)
426 if c0 != ' ' or c1 != ' ' then
427 return new TokenEmStar(loc, pos, c)
429 return new TokenNone(loc, pos, c)
431 else if c == '_' then
433 if c0 != ' ' or c2 != ' ' then
434 return new TokenStrongUnderscore(loc, pos, c)
436 return new TokenEmUnderscore(loc, pos, c)
440 if (c0.is_letter or c0.is_digit) and c0 != '_' and
441 (c1.is_letter or c1.is_digit) then
442 return new TokenNone(loc, pos, c)
444 return new TokenEmUnderscore(loc, pos, c)
447 if c0 != ' ' or c1 != ' ' then
448 return new TokenEmUnderscore(loc, pos, c)
450 return new TokenNone(loc, pos, c)
452 else if c == '!' then
453 if c1 == '[' then return new TokenImage(loc, pos, c)
454 return new TokenNone(loc, pos, c)
455 else if c == '[' then
456 return new TokenLink(loc, pos, c)
457 else if c == ']' then
458 return new TokenNone(loc, pos, c)
459 else if c == '`' then
461 return new TokenCodeDouble(loc, pos, c)
463 return new TokenCodeSingle(loc, pos, c)
465 else if c == '\\' then
466 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
467 return new TokenEscape(loc
, pos
, c
)
469 return new TokenNone(loc
, pos
, c
)
471 else if c
== '<' then
472 return new TokenHTML(loc
, pos
, c
)
473 else if c
== '&' then
474 return new TokenEntity(loc
, pos
, c
)
477 if c
== '~' and c1
== '~' then
478 return new TokenStrike(loc
, pos
, c
)
481 return new TokenNone(loc
, pos
, c
)
485 # Find the position of a `token` in `self`.
486 fun find_token
(text
: Text, start
: Int, token
: Token): Int do
488 while pos
< text
.length
do
489 if token_at
(text
, pos
).is_same_type
(token
) then
497 # Kind of decorator used for decoration.
498 type DECORATOR: Decorator
500 # Decorator used for output.
501 # Default is `HTMLDecorator`
502 var decorator
: DECORATOR is writable, lazy
do
503 return new HTMLDecorator
506 # Create a new `MarkdownEmitter` using a custom `decorator`.
507 init with_decorator
(decorator
: DECORATOR) do
508 self.decorator
= decorator
511 # Output `block` using `decorator` in the current buffer.
512 fun emit
(block
: Block): Text do
513 var buffer
= push_buffer
519 # Output the content of `block`.
520 fun emit_in
(block
: Block) do block
.emit_in
(self)
522 # Transform and emit mardown text
523 fun emit_text
(text
: Text) do emit_text_until
(text
, 0, null)
525 # Transform and emit mardown text starting at `start` and
526 # until a token with the same type as `token` is found.
527 # Go until the end of `text` if `token` is null.
528 fun emit_text_until
(text
: Text, start
: Int, token
: nullable Token): Int do
529 var old_text
= current_text
530 var old_pos
= current_pos
533 while current_pos
< text
.length
do
534 if text
[current_pos
] == '\n' then
535 current_loc
.line_start
+= 1
536 current_loc
.column_start
= -current_pos
538 var mt
= token_at
(text
, current_pos
)
539 if (token
!= null and not token
isa TokenNone) and
540 (mt
.is_same_type
(token
) or
541 (token
isa TokenEmStar and mt
isa TokenStrongStar) or
542 (token
isa TokenEmUnderscore and mt
isa TokenStrongUnderscore)) then
548 current_text
= old_text
549 current_pos
= old_pos
553 # Currently processed position in `current_text`.
554 # Used when visiting inline production with `emit_text_until`.
555 private var current_pos
: Int = -1
557 # Currently processed text.
558 # Used when visiting inline production with `emit_text_until`.
559 private var current_text
: nullable Text = null
562 private var buffer_stack
= new List[FlatBuffer]
564 # Push a new buffer on the stack.
565 private fun push_buffer
: FlatBuffer do
566 var buffer
= new FlatBuffer
567 buffer_stack
.add buffer
571 # Pop the last buffer.
572 private fun pop_buffer
do buffer_stack
.pop
574 # Current output buffer.
575 private fun current_buffer
: FlatBuffer do
576 assert not buffer_stack
.is_empty
577 return buffer_stack
.last
581 private var loc_stack
= new List[MDLocation]
583 # Push a new MDLocation on the stack.
584 private fun push_loc
(location
: MDLocation) do loc_stack
.add location
586 # Pop the last buffer.
587 private fun pop_loc
: MDLocation do return loc_stack
.pop
589 # Current output buffer.
590 private fun current_loc
: MDLocation do
591 assert not loc_stack
.is_empty
592 return loc_stack
.last
595 # Append `e` to current buffer.
596 fun add
(e
: Writable) do
598 current_buffer
.append e
600 current_buffer
.append e
.write_to_string
604 # Append `c` to current buffer.
609 # Append a "\n" line break.
610 fun addn
do addc
'\n'
614 # Links that are specified somewhere in the mardown document to be reused as shortcuts.
617 # [1]: http://example.com/ "Optional title"
624 # Optional link title
625 var title
: nullable String = null
627 # Is the link an abreviation?
628 var is_abbrev
= false
630 # Create a link with a title.
631 init with_title
(link
: String, title
: nullable String) do
637 # A `Decorator` is used to emit mardown into a specific format.
638 # Default decorator used is `HTMLDecorator`.
641 # Kind of processor used
642 type PROCESSOR: MarkdownProcessor
644 # Render a single plain char.
646 # Redefine this method to add special escaping for plain text.
647 fun add_char
(v
: PROCESSOR, c
: Char) do v
.addc c
649 # Render a ruler block.
650 fun add_ruler
(v
: PROCESSOR, block
: BlockRuler) is abstract
652 # Render a headline block with corresponding level.
653 fun add_headline
(v
: PROCESSOR, block
: BlockHeadline) is abstract
655 # Render a paragraph block.
656 fun add_paragraph
(v
: PROCESSOR, block
: BlockParagraph) is abstract
658 # Render a code or fence block.
659 fun add_code
(v
: PROCESSOR, block
: BlockCode) is abstract
661 # Render a blockquote.
662 fun add_blockquote
(v
: PROCESSOR, block
: BlockQuote) is abstract
664 # Render an unordered list.
665 fun add_unorderedlist
(v
: PROCESSOR, block
: BlockUnorderedList) is abstract
667 # Render an ordered list.
668 fun add_orderedlist
(v
: PROCESSOR, block
: BlockOrderedList) is abstract
670 # Render a list item.
671 fun add_listitem
(v
: PROCESSOR, block
: BlockListItem) is abstract
673 # Render an emphasis text.
674 fun add_em
(v
: PROCESSOR, text
: Text) is abstract
676 # Render a strong text.
677 fun add_strong
(v
: PROCESSOR, text
: Text) is abstract
679 # Render a strike text.
681 # Extended mode only (see `MarkdownProcessor::ext_mode`)
682 fun add_strike
(v
: PROCESSOR, text
: Text) is abstract
685 fun add_link
(v
: PROCESSOR, link
: Text, name
: Text, comment
: nullable Text) is abstract
688 fun add_image
(v
: PROCESSOR, link
: Text, name
: Text, comment
: nullable Text) is abstract
690 # Render an abbreviation.
691 fun add_abbr
(v
: PROCESSOR, name
: Text, comment
: Text) is abstract
693 # Render a code span reading from a buffer.
694 fun add_span_code
(v
: PROCESSOR, buffer
: Text, from
, to
: Int) is abstract
696 # Render a text and escape it.
697 fun append_value
(v
: PROCESSOR, value
: Text) is abstract
699 # Render code text from buffer and escape it.
700 fun append_code
(v
: PROCESSOR, buffer
: Text, from
, to
: Int) is abstract
702 # Render a character escape.
703 fun escape_char
(v
: PROCESSOR, char
: Char) is abstract
705 # Render a line break
706 fun add_line_break
(v
: PROCESSOR) is abstract
708 # Generate a new html valid id from a `String`.
709 fun strip_id
(txt
: String): String is abstract
711 # Found headlines during the processing labeled by their ids.
712 fun headlines
: ArrayMap[String, HeadLine] is abstract
715 # Class representing a markdown headline.
717 # Unique identifier of this headline.
720 # Text of the headline.
723 # Level of this headline.
725 # According toe the markdown specification, level must be in `[1..6]`.
729 # `Decorator` that outputs HTML.
733 redef var headlines
= new ArrayMap[String, HeadLine]
735 redef fun add_ruler
(v
, block
) do v
.add
"<hr/>\n"
737 redef fun add_headline
(v
, block
) do
739 var line
= block
.block
.first_line
740 if line
== null then return
742 var id
= strip_id
(txt
)
743 var lvl
= block
.depth
744 headlines
[id
] = new HeadLine(id
, txt
, lvl
)
746 v
.add
"<h{lvl} id=\"{id}\
">"
751 redef fun add_paragraph
(v
, block
) do
757 redef fun add_code
(v
, block
) do
758 var meta
= block
.meta
760 v
.add
"<pre class=\""
761 append_value(v, meta)
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.as(not null).location.line_start,
1006 first_line.as(not null).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.as(not null).prev = null
1019 # update current block loc
1020 location.line_start = first_line.as(not null).location.line_start
1021 location.column_start = first_line.as(not null).location.column_start
1023 if first_block == null then
1027 last_block.as(not null).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.as(not null).next_empty = line.is_empty
1040 line.prev_empty = last_line.as(not null).is_empty
1041 line.prev = last_line
1042 last_line.as(not null).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.as(not null).next = line.next
1054 if line.next == null then
1055 last_line = line.prev
1057 line.next.as(not null).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 var block = first_block
1125 while block != null do
1126 text.append block.text
1130 return text.write_to_string
1134 # Representation of a markdown block in the AST.
1135 # Each `Block` is linked to a `MDBlock` that contains mardown code.
1136 abstract class Block
1138 # The markdown block `self` is related to.
1141 # Output `self` using `v.decorator`.
1142 fun emit(v: MarkdownProcessor) do v.emit_in(self)
1144 # Emit the containts of `self`, lines or blocks.
1145 fun emit_in(v: MarkdownProcessor) do
1146 block.remove_surrounding_empty_lines
1147 if block.has_lines then
1154 # Emit lines contained in `block`.
1155 fun emit_lines(v: MarkdownProcessor) do
1156 var tpl = v.push_buffer
1157 var line = block.first_line
1158 while line != null do
1159 if not line.is_empty then
1160 v.add line.value.substring(line.leading, line.value.length - line.trailing)
1161 if line.trailing >= 2 then v.decorator.add_line_break(v)
1163 if line.next != null then
1172 # Emit sub-blocks contained in `block`.
1173 fun emit_blocks(v: MarkdownProcessor) do
1174 var block = self.block.first_block
1175 while block != null do
1176 v.push_loc(block.location)
1183 # The raw content of the block as a multi-line string.
1184 fun raw_content: String do
1185 var infence = self isa BlockFence
1186 var text = new FlatBuffer
1187 var line = self.block.first_line
1188 while line != null do
1189 if not line.is_empty then
1190 var str = line.value
1191 if not infence and str.has_prefix(" ") then
1192 text.append str.substring(4, str.length - line.trailing)
1200 return text.write_to_string
1204 # A block without any markdown specificities.
1206 # Actually use the same implementation than `BlockCode`,
1207 # this class is only used for typing purposes.
1212 # A markdown blockquote.
1216 redef fun emit(v) do v.decorator.add_blockquote(v, self)
1218 # Remove blockquote markers.
1219 private fun remove_block_quote_prefix(block: MDBlock) do
1220 var line = block.first_line
1221 while line != null do
1222 if not line.is_empty then
1223 if line.value[line.leading] == '>' then
1224 var rem = line.leading + 1
1225 if line.leading + 1 < line.value.length and
1226 line.value[line.leading + 1] == ' ' then
1229 line.value = line.value.substring_from(rem)
1230 line.leading = line.process_leading
1238 # A markdown code block.
1242 # Any string found after fence token.
1243 var meta: nullable Text
1245 # Number of char to skip at the beginning of the line.
1247 # Block code lines start at 4 spaces.
1248 protected var line_start = 4
1250 redef fun emit(v) do v.decorator.add_code(v, self)
1252 redef fun emit_lines(v) do
1253 var line = block.first_line
1254 while line != null do
1255 if not line.is_empty then
1256 v.decorator.append_code(v, line.value, line_start, line.value.length)
1264 # A markdown code-fence block.
1266 # Actually use the same implementation than `BlockCode`,
1267 # this class is only used for typing purposes.
1271 # Fence code lines start at 0 spaces.
1272 redef var line_start = 0
1275 # A markdown headline.
1279 redef fun emit(v) do
1280 var loc = block.location.copy
1281 loc.column_start += start
1283 v.decorator.add_headline(v, self)
1287 private var start = 0
1289 # Depth of the headline used to determine the headline level.
1292 # Remove healine marks from lines contained in `self`.
1293 private fun transform_headline(block: MDBlock) do
1294 if depth > 0 then return
1296 var line = block.first_line
1297 if line == null then return
1298 if line.is_empty then return
1299 var start = line.leading
1300 while start < line.value.length and line.value[start] == '#' do
1304 while start
< line
.value
.length
and line
.value
[start
] == ' ' do
1307 if start
>= line
.value
.length
then
1308 line
.is_empty
= true
1310 var nend
= line
.value
.length
- line
.trailing
- 1
1311 while line
.value
[nend
] == '#' do nend
-= 1
1312 while line
.value
[nend
] == ' ' do nend
-= 1
1313 line
.value
= line
.value
.substring
(start
, nend
- start
+ 1)
1318 depth
= level
.min
(6)
1322 # A markdown list item block.
1326 redef fun emit
(v
) do v
.decorator
.add_listitem
(v
, self)
1329 # A markdown list block.
1330 # Can be either an ordered or unordered list, this class is mainly used to factorize code.
1331 abstract class BlockList
1334 # Split list block into list items sub-blocks.
1335 private fun init_block
(v
: MarkdownProcessor) do
1336 var line
= block
.first_line
1337 if line
== null then return
1339 while line
!= null do
1340 var t
= v
.line_kind
(line
)
1341 if t
isa LineList or
1342 (not line
.is_empty
and (line
.prev_empty
and line
.leading
== 0 and
1343 not (t
isa LineList))) then
1344 var sblock
= block
.split
(line
.prev
.as(not null))
1345 sblock
.kind
= new BlockListItem(sblock
)
1349 var sblock
= block
.split
(block
.last_line
.as(not null))
1350 sblock
.kind
= new BlockListItem(sblock
)
1353 # Expand list items as paragraphs if needed.
1354 private fun expand_paragraphs
(block
: MDBlock) do
1355 var outer
= block
.first_block
1356 var inner
: nullable MDBlock
1357 var has_paragraph
= false
1358 while outer
!= null and not has_paragraph
do
1359 if outer
.kind
isa BlockListItem then
1360 inner
= outer
.first_block
1361 while inner
!= null and not has_paragraph
do
1362 if inner
.kind
isa BlockParagraph then
1363 has_paragraph
= true
1370 if has_paragraph
then
1371 outer
= block
.first_block
1372 while outer
!= null do
1373 if outer
.kind
isa BlockListItem then
1374 inner
= outer
.first_block
1375 while inner
!= null do
1376 if inner
.kind
isa BlockNone then
1377 inner
.kind
= new BlockParagraph(inner
)
1388 # A markdown ordered list.
1389 class BlockOrderedList
1392 redef fun emit
(v
) do v
.decorator
.add_orderedlist
(v
, self)
1395 # A markdown unordred list.
1396 class BlockUnorderedList
1399 redef fun emit
(v
) do v
.decorator
.add_unorderedlist
(v
, self)
1402 # A markdown paragraph block.
1403 class BlockParagraph
1406 redef fun emit
(v
) do v
.decorator
.add_paragraph
(v
, self)
1413 redef fun emit
(v
) do v
.decorator
.add_ruler
(v
, self)
1416 # Xml blocks that can be found in markdown markup.
1420 redef fun emit_lines
(v
) do
1421 var line
= block
.first_line
1422 while line
!= null do
1423 if not line
.is_empty
then v
.add line
.value
1433 # Location of `self` in the original input.
1434 var location
: MDLocation
1436 # Text contained in this line.
1437 var value
: String is writable
1439 # Is this line empty?
1440 # Lines containing only spaces are considered empty.
1441 var is_empty
: Bool = true is writable
1443 # Previous line in `MDBlock` or null if first line.
1444 var prev
: nullable MDLine = null is writable
1446 # Next line in `MDBlock` or null if last line.
1447 var next
: nullable MDLine = null is writable
1449 # Is the previous line empty?
1450 var prev_empty
: Bool = false is writable
1452 # Is the next line empty?
1453 var next_empty
: Bool = false is writable
1455 # Initialize a new MDLine from its string value
1457 self.leading
= process_leading
1458 if leading
!= value
.length
then
1459 self.is_empty
= false
1460 self.trailing
= process_trailing
1464 # Set `value` as an empty String and update `leading`, `trailing` and is_`empty`.
1470 if prev
!= null then prev
.as(not null).next_empty
= true
1471 if next
!= null then next
.as(not null).prev_empty
= true
1474 # Number or leading spaces on this line.
1475 var leading
: Int = 0 is writable
1477 # Compute `leading` depending on `value`.
1478 fun process_leading
: Int do
1480 var value
= self.value
1481 while count
< value
.length
and value
[count
] == ' ' do count
+= 1
1482 if leading
== value
.length
then clear
1486 # Number of trailing spaces on this line.
1487 var trailing
: Int = 0 is writable
1489 # Compute `trailing` depending on `value`.
1490 fun process_trailing
: Int do
1492 var value
= self.value
1493 while value
[value
.length
- count
- 1] == ' ' do
1499 # Count the amount of `ch` in this line.
1500 # Return A value > 0 if this line only consists of `ch` end spaces.
1501 fun count_chars
(ch
: Char): Int do
1517 # Count the amount of `ch` at the start of this line ignoring spaces.
1518 fun count_chars_start
(ch
: Char): Int do
1533 # Last XML line if any.
1534 private var xml_end_line
: nullable MDLine = null
1536 # Does `value` contains valid XML markup?
1537 private fun check_html
: Bool do
1538 var tags
= new Array[String]
1539 var tmp
= new FlatBuffer
1541 if pos
+ 1 < value
.length
and value
[pos
+ 1] == '!' then
1542 if read_xml_comment
(self, pos
) > 0 then return true
1544 pos
= value
.read_xml
(tmp
, pos
, false)
1548 if not tag
.is_html_block
then
1556 var line
: nullable MDLine = self
1557 while line
!= null do
1558 while pos
< line
.value
.length
and line
.value
[pos
] != '<' do
1561 if pos
>= line
.value
.length
then
1562 if pos
- 2 >= 0 and line
.value
[pos
- 2] == '/' then
1564 if tags
.is_empty
then
1572 tmp
= new FlatBuffer
1573 var new_pos
= line
.value
.read_xml
(tmp
, pos
, false)
1576 if tag
.is_html_block
and not tag
== "hr" then
1577 if tmp
[1] == '/' then
1578 if tags
.last
!= tag
then
1586 if tags
.is_empty
then
1596 return tags
.is_empty
1601 # Read a XML comment.
1602 # Used by `check_html`.
1603 private fun read_xml_comment
(first_line
: MDLine, start
: Int): Int do
1604 var line
: nullable MDLine = first_line
1605 if start
+ 3 < line
.as(not null).value
.length
then
1606 if line
.as(not null).value
[2] == '-' and line
.as(not null).value
[3] == '-' then
1608 while line
!= null do
1609 while pos
< line
.value
.length
and line
.value
[pos
] != '-' do
1612 if pos
== line
.value
.length
then
1616 if pos
+ 2 < line
.value
.length
then
1617 if line
.value
[pos
+ 1] == '-' and line
.value
[pos
+ 2] == '>' then
1618 first_line
.xml_end_line
= line
1630 # Extract the text of `self` without leading and trailing.
1631 fun text
: String do return value
.substring
(leading
, value
.length
- trailing
)
1638 # See `MarkdownProcessor::recurse`.
1639 fun process
(v
: MarkdownProcessor) is abstract
1642 # An empty markdown line.
1646 redef fun process
(v
) do
1647 v
.current_line
= v
.current_line
.as(not null).next
1651 # A non-specific markdown construction.
1652 # Mainly used as part of another line construct such as paragraphs or lists.
1656 redef fun process
(v
) do
1657 var line
= v
.current_line
1659 var was_empty
= line
.as(not null).prev_empty
1660 while line
!= null and not line
.is_empty
do
1661 var t
= v
.line_kind
(line
)
1662 if (v
.in_list
or v
.ext_mode
) and t
isa LineList then
1665 if v
.ext_mode
and (t
isa LineCode or t
isa LineFence) then
1668 if t
isa LineHeadline or t
isa LineHeadline1 or t
isa LineHeadline2 or
1669 t
isa LineHR or t
isa LineBlockquote or t
isa LineXML then
1675 var current_block
= v
.current_block
.as(not null)
1676 if line
!= null and not line
.is_empty
then
1677 var block
= current_block
.split
(line
.prev
.as(not null))
1678 if v
.in_list
and not was_empty
then
1679 block
.kind
= new BlockNone(block
)
1681 block
.kind
= new BlockParagraph(block
)
1683 current_block
.remove_leading_empty_lines
1686 if line
!= null then
1687 block
= current_block
.split
(line
)
1689 block
= current_block
.split
(current_block
.last_line
.as(not null))
1691 if v
.in_list
and (line
== null or not line
.is_empty
) and not was_empty
then
1692 block
.kind
= new BlockNone(block
)
1694 block
.kind
= new BlockParagraph(block
)
1696 current_block
.remove_leading_empty_lines
1698 v
.current_line
= current_block
.first_line
1702 # A line of markdown code.
1706 redef fun process
(v
) do
1707 var line
= v
.current_line
1709 while line
!= null and (line
.is_empty
or v
.line_kind
(line
) isa LineCode) do
1712 # split at block end line
1713 var current_block
= v
.current_block
.as(not null)
1715 if line
!= null then
1716 block
= current_block
.split
(line
.prev
.as(not null))
1718 block
= current_block
.split
(current_block
.last_line
.as(not null))
1720 block
.kind
= new BlockCode(block
)
1721 block
.remove_surrounding_empty_lines
1722 v
.current_line
= current_block
.first_line
1726 # A line of raw XML.
1730 redef fun process
(v
) do
1731 var line
= v
.current_line
1732 if line
== null then return
1733 var current_block
= v
.current_block
.as(not null)
1734 var prev
= line
.prev
1735 if prev
!= null then current_block
.split
(prev
)
1736 var block
= current_block
.split
(line
.xml_end_line
.as(not null))
1737 block
.kind
= new BlockXML(block
)
1738 current_block
.remove_leading_empty_lines
1739 v
.current_line
= current_block
.first_line
1743 # A markdown blockquote line.
1744 class LineBlockquote
1747 redef fun process
(v
) do
1748 var line
= v
.current_line
1749 var current_block
= v
.current_block
.as(not null)
1751 while line
!= null do
1752 if not line
.is_empty
and (line
.prev_empty
and
1753 line
.leading
== 0 and
1754 not v
.line_kind
(line
) isa LineBlockquote) then break
1759 if line
!= null then
1760 block
= current_block
.split
(line
.prev
.as(not null))
1762 block
= current_block
.split
(current_block
.last_line
.as(not null))
1764 var kind
= new BlockQuote(block
)
1766 block
.remove_surrounding_empty_lines
1767 kind
.remove_block_quote_prefix
(block
)
1768 v
.current_line
= line
1769 v
.recurse
(block
, false)
1770 v
.current_line
= current_block
.first_line
1774 # A markdown ruler line.
1778 redef fun process
(v
) do
1779 var line
= v
.current_line
1780 if line
== null then return
1781 var current_block
= v
.current_block
.as(not null)
1782 if line
.prev
!= null then current_block
.split
(line
.prev
.as(not null))
1783 var block
= current_block
.split
(line
)
1784 block
.kind
= new BlockRuler(block
)
1785 current_block
.remove_leading_empty_lines
1786 v
.current_line
= current_block
.first_line
1790 # A markdown fence code line.
1794 redef fun process
(v
) do
1796 var line
= v
.current_line
.as(not null).next
1797 var current_block
= v
.current_block
.as(not null)
1798 while line
!= null do
1799 if v
.line_kind
(line
) isa LineFence then break
1802 if line
!= null then
1807 if line
!= null then
1808 block
= current_block
.split
(line
.prev
.as(not null))
1810 block
= current_block
.split
(current_block
.last_line
.as(not null))
1812 block
.remove_surrounding_empty_lines
1813 var meta
= block
.first_line
.as(not null).value
.meta_from_fence
1814 block
.kind
= new BlockFence(block
, meta
)
1815 block
.first_line
.as(not null).clear
1816 var last
= block
.last_line
1817 if last
!= null and v
.line_kind
(last
) isa LineFence then
1818 block
.last_line
.as(not null).clear
1820 block
.remove_surrounding_empty_lines
1821 v
.current_line
= line
1825 # A markdown headline.
1829 redef fun process
(v
) do
1830 var line
= v
.current_line
1831 if line
== null then return
1832 var current_block
= v
.current_block
.as(not null)
1833 var lprev
= line
.prev
1834 if lprev
!= null then current_block
.split
(lprev
)
1835 var block
= current_block
.split
(line
)
1836 var kind
= new BlockHeadline(block
)
1838 kind
.transform_headline
(block
)
1839 current_block
.remove_leading_empty_lines
1840 v
.current_line
= current_block
.first_line
1844 # A markdown headline of level 1.
1848 redef fun process
(v
) do
1849 var line
= v
.current_line
1850 if line
== null then return
1851 var current_block
= v
.current_block
.as(not null)
1852 var lprev
= line
.prev
1853 if lprev
!= null then current_block
.split
(lprev
)
1854 line
.next
.as(not null).clear
1855 var block
= current_block
.split
(line
)
1856 var kind
= new BlockHeadline(block
)
1858 kind
.transform_headline
(block
)
1860 current_block
.remove_leading_empty_lines
1861 v
.current_line
= current_block
.first_line
1865 # A markdown headline of level 2.
1869 redef fun process
(v
) do
1870 var line
= v
.current_line
1871 if line
== null then return
1872 var current_block
= v
.current_block
.as(not null)
1873 var lprev
= line
.prev
1874 if lprev
!= null then current_block
.split
(lprev
)
1875 line
.next
.as(not null).clear
1876 var block
= current_block
.split
(line
)
1877 var kind
= new BlockHeadline(block
)
1879 kind
.transform_headline
(block
)
1881 current_block
.remove_leading_empty_lines
1882 v
.current_line
= current_block
.first_line
1886 # A markdown list line.
1887 # Mainly used to factorize code between ordered and unordered lists.
1888 abstract class LineList
1891 redef fun process
(v
) do
1892 var line
= v
.current_line
1894 while line
!= null do
1895 var t
= v
.line_kind
(line
)
1896 if not line
.is_empty
and (line
.prev_empty
and line
.leading
== 0 and
1897 not t
isa LineList) then break
1901 var current_block
= v
.current_block
.as(not null)
1903 if line
!= null then
1904 list
= current_block
.split
(line
.prev
.as(not null))
1906 list
= current_block
.split
(current_block
.last_line
.as(not null))
1908 var kind
= block_kind
(list
)
1910 list
.first_line
.as(not null).prev_empty
= false
1911 list
.last_line
.as(not null).next_empty
= false
1912 list
.remove_surrounding_empty_lines
1913 list
.first_line
.as(not null).prev_empty
= false
1914 list
.last_line
.as(not null).next_empty
= false
1916 var block
= list
.first_block
1917 while block
!= null do
1918 block
.remove_list_indent
(v
)
1919 v
.recurse
(block
, true)
1922 kind
.expand_paragraphs
(list
)
1923 v
.current_line
= line
1926 # Create a new block kind based on this line.
1927 protected fun block_kind
(block
: MDBlock): BlockList is abstract
1929 # Extract string value from `MDLine`.
1930 protected fun extract_value
(line
: MDLine): String is abstract
1933 # An ordered list line.
1937 redef fun block_kind
(block
) do return new BlockOrderedList(block
)
1939 redef fun extract_value
(line
) do
1940 return line
.value
.substring_from
(line
.value
.index_of
('.') + 2)
1944 # An unordered list line.
1948 redef fun block_kind
(block
) do return new BlockUnorderedList(block
)
1950 redef fun extract_value
(line
) do
1951 return line
.value
.substring_from
(line
.leading
+ 2)
1955 # A token represent a character in the markdown input.
1956 # Some tokens have a specific markup behaviour that is handled here.
1957 abstract class Token
1959 # Location of `self` in the original input.
1960 var location
: nullable MDLocation
1962 # Position of `self` in input independant from lines.
1965 # Character found at `pos` in the markdown input.
1968 # Output that token using `MarkdownEmitter::decorator`.
1969 fun emit
(v
: MarkdownProcessor) do v
.decorator
.add_char
(v
, char
)
1972 # A token without a specific meaning.
1977 # An emphasis token.
1978 abstract class TokenEm
1981 redef fun emit
(v
) do
1982 var tmp
= v
.push_buffer
1983 var b
= v
.emit_text_until
(v
.current_text
.as(not null), pos
+ 1, self)
1986 v
.decorator
.add_em
(v
, tmp
)
1994 # An emphasis star token.
1999 # An emphasis underscore token.
2000 class TokenEmUnderscore
2005 abstract class TokenStrong
2008 redef fun emit
(v
) do
2009 var tmp
= v
.push_buffer
2010 var b
= v
.emit_text_until
(v
.current_text
.as(not null), pos
+ 2, self)
2013 v
.decorator
.add_strong
(v
, tmp
)
2014 v
.current_pos
= b
+ 1
2021 # A strong star token.
2022 class TokenStrongStar
2026 # A strong underscore token.
2027 class TokenStrongUnderscore
2032 # This class is mainly used to factorize work between single and double quoted span codes.
2033 abstract class TokenCode
2036 redef fun emit
(v
) do
2037 var current_text
= v
.current_text
.as(not null)
2038 var a
= pos
+ next_pos
+ 1
2039 var b
= v
.find_token
(current_text
, a
, self)
2041 v
.current_pos
= b
+ next_pos
2042 while a
< b
and current_text
[a
] == ' ' do a
+= 1
2044 while current_text
[b
- 1] == ' ' do b
-= 1
2045 v
.decorator
.add_span_code
(v
, current_text
, a
, b
)
2052 private fun next_pos
: Int is abstract
2055 # A span code token.
2056 class TokenCodeSingle
2059 redef fun next_pos
do return 0
2062 # A doubled span code token.
2063 class TokenCodeDouble
2066 redef fun next_pos
do return 1
2069 # A link or image token.
2070 # This class is mainly used to factorize work between images and links.
2071 abstract class TokenLinkOrImage
2075 var link
: nullable Text = null
2078 var name
: nullable Text = null
2081 var comment
: nullable Text = null
2083 # Is the link construct an abbreviation?
2084 var is_abbrev
= false
2086 redef fun emit
(v
) do
2087 var tmp
= new FlatBuffer
2088 var b
= check_link
(v
, tmp
, pos
, self)
2097 # Emit the hyperlink as link or image.
2098 private fun emit_hyper
(v
: MarkdownProcessor) is abstract
2100 # Check if the link is a valid link.
2101 private fun check_link
(v
: MarkdownProcessor, out
: FlatBuffer, start
: Int, token
: Token): Int do
2102 var md
= v
.current_text
2103 if md
== null then return -1
2105 if token
isa TokenLink then
2110 var tmp
= new FlatBuffer
2111 pos
= md
.read_md_link_id
(tmp
, pos
)
2112 if pos
< start
then return -1
2116 pos
= md
.skip_spaces
(pos
)
2118 var tid
= name
.as(not null).write_to_string
.to_lower
2119 if v
.link_refs
.has_key
(tid
) then
2120 var lr
= v
.link_refs
[tid
]
2121 is_abbrev
= lr
.is_abbrev
2128 else if md
[pos
] == '(' then
2130 pos
= md
.skip_spaces
(pos
)
2131 if pos
< start
then return -1
2132 tmp
= new FlatBuffer
2133 var use_lt
= md
[pos
] == '<'
2135 pos
= md
.read_until
(tmp
, pos
+ 1, '>')
2137 pos
= md
.read_md_link
(tmp
, pos
)
2139 if pos
< start
then return -1
2140 if use_lt
then pos
+= 1
2141 link
= tmp
.write_to_string
2142 if md
[pos
] == ' ' then
2143 pos
= md
.skip_spaces
(pos
)
2144 if pos
> start
and md
[pos
] == '"' then
2146 tmp
= new FlatBuffer
2147 pos
= md
.read_until
(tmp
, pos
, '"')
2148 if pos
< start
then return -1
2149 comment
= tmp
.write_to_string
2151 pos
= md
.skip_spaces
(pos
)
2152 if pos
== -1 then return -1
2155 if pos
< start
then return -1
2156 if md
[pos
] != ')' then return -1
2157 else if md
[pos
] == '[' then
2159 tmp
= new FlatBuffer
2160 pos
= md
.read_raw_until
(tmp
, pos
, ']')
2161 if pos
< start
then return -1
2163 if tmp
.length
> 0 then
2168 var tid
= id
.as(not null).write_to_string
.to_lower
2169 if v
.link_refs
.has_key
(tid
) then
2170 var lr
= v
.link_refs
[tid
]
2175 var tid
= name
.as(not null).write_to_string
.replace
("\n", " ").to_lower
2176 if v
.link_refs
.has_key
(tid
) then
2177 var lr
= v
.link_refs
[tid
]
2185 if link
== null then return -1
2190 # A markdown link token.
2192 super TokenLinkOrImage
2194 redef fun emit_hyper
(v
) do
2195 if is_abbrev
and comment
!= null then
2196 v
.decorator
.add_abbr
(v
, name
.as(not null), comment
.as(not null))
2198 v
.decorator
.add_link
(v
, link
.as(not null), name
.as(not null), comment
)
2203 # A markdown image token.
2205 super TokenLinkOrImage
2207 redef fun emit_hyper
(v
) do
2208 v
.decorator
.add_image
(v
, link
.as(not null), name
.as(not null), comment
)
2216 redef fun emit
(v
) do
2217 var tmp
= new FlatBuffer
2218 var b
= check_html
(v
, tmp
, v
.current_text
.as(not null), v
.current_pos
)
2223 v
.decorator
.escape_char
(v
, char
)
2227 # Is the HTML valid?
2228 # Also take care of link and mailto shortcuts.
2229 private fun check_html
(v
: MarkdownProcessor, out
: FlatBuffer, md
: Text, start
: Int): Int do
2230 # check for auto links
2231 var tmp
= new FlatBuffer
2232 var pos
= md
.read_until
(tmp
, start
+ 1, ':', ' ', '>', '\n')
2233 if pos
!= -1 and md
[pos
] == ':' and tmp
.is_link_prefix
then
2234 pos
= md
.read_until
(tmp
, pos
, '>')
2236 var link
= tmp
.write_to_string
2237 v
.decorator
.add_link
(v
, link
, link
, null)
2241 # TODO check for mailto
2242 # check for inline html
2243 if start
+ 2 < md
.length
then
2244 return md
.read_xml
(out
, start
, true)
2250 # An HTML entity token.
2254 redef fun emit
(v
) do
2255 var tmp
= new FlatBuffer
2256 var b
= check_entity
(tmp
, v
.current_text
.as(not null), pos
)
2261 v
.decorator
.escape_char
(v
, char
)
2265 # Is the entity valid?
2266 private fun check_entity
(out
: FlatBuffer, md
: Text, start
: Int): Int do
2267 var pos
= md
.read_until
(out
, start
, ';')
2268 if pos
< 0 or out
.length
< 3 then
2271 if out
[1] == '#' then
2272 if out
[2] == 'x' or out
[2] == 'X' then
2273 if out
.length
< 4 then return -1
2274 for i
in [3..out
.length
[ do
2276 if (c
< '0' or c
> '9') and (c
< 'a' and c
> 'f') and (c
< 'A' and c
> 'F') then
2281 for i
in [2..out
.length
[ do
2283 if c
< '0' or c
> '9' then return -1
2288 for i
in [1..out
.length
[ do
2290 if not c
.is_digit
and not c
.is_letter
then return -1
2293 # TODO check entity is valid
2294 # if out.is_entity then
2304 # A markdown escape token.
2308 redef fun emit
(v
) do
2310 v
.addc v
.current_text
.as(not null)[v
.current_pos
]
2314 # A markdown strike token.
2316 # Extended mode only (see `MarkdownProcessor::ext_mode`)
2320 redef fun emit
(v
) do
2321 var tmp
= v
.push_buffer
2322 var b
= v
.emit_text_until
(v
.current_text
.as(not null), pos
+ 2, self)
2325 v
.decorator
.add_strike
(v
, tmp
)
2326 v
.current_pos
= b
+ 1
2335 # Get the position of the next non-space character.
2336 private fun skip_spaces
(start
: Int): Int do
2338 while pos
> -1 and pos
< length
and (self[pos
] == ' ' or self[pos
] == '\n') do
2341 if pos
< length
then return pos
2345 # Read `self` until `nend` and append it to the `out` buffer.
2346 # Escape markdown special chars.
2347 private fun read_until
(out
: FlatBuffer, start
: Int, nend
: Char...): Int do
2349 while pos
< length
do
2351 if c
== '\\' and pos
+ 1 < length
then
2352 pos
= escape
(out
, self[pos
+ 1], pos
)
2354 for n
in nend
do if c
== n
then break label
2359 if pos
== length
then return -1
2363 # Read `self` as raw text until `nend` and append it to the `out` buffer.
2364 # No escape is made.
2365 private fun read_raw_until
(out
: FlatBuffer, start
: Int, nend
: Char...): Int do
2367 while pos
< length
do
2369 var end_reached
= false
2376 if end_reached
then break
2380 if pos
== length
then return -1
2384 # Read `self` as XML until `to` and append it to the `out` buffer.
2385 # Escape HTML special chars.
2386 private fun read_xml_until
(out
: FlatBuffer, from
: Int, to
: Char...): Int do
2389 var str_char
: nullable Char = null
2390 while pos
< length
do
2396 if pos
< length
then
2402 if c
== str_char
then
2409 if c
== '"' or c
== '\'' then
2414 var end_reached = false
2415 for n in [0..to.length[ do
2421 if end_reached then break
2426 if pos == length then return -1
2430 # Read `self` as XML and append it to the `out` buffer.
2431 # Safe mode can be activated to limit reading to valid xml.
2432 private fun read_xml(out: FlatBuffer, start: Int, safe_mode: Bool): Int do
2435 var is_close_tag = false
2436 if start + 1 >= length then return -1
2437 if self[start + 1] == '/' then
2440 else if self[start + 1] == '!' then
2444 is_close_tag = false
2448 var tmp = new FlatBuffer
2449 pos = read_xml_until(tmp, pos, ' ', '/', '>')
2450 if pos == -1 then return -1
2451 var tag = tmp.write_to_string.trim.to_lower
2452 if not tag.is_valid_html_tag then
2455 else if tag.is_html_unsafe then
2458 if is_close_tag then out.add '/'
2462 if is_close_tag then out.add '/'
2467 if is_close_tag then out.add '/'
2468 pos = read_xml_until(out, pos, ' ', '/', '>')
2470 if pos == -1 then return -1
2471 pos = read_xml_until(out, pos, '/', '>')
2472 if pos == -1 then return -1
2473 if self[pos] == '/' then
2475 pos = self.read_xml_until(out, pos + 1, '>')
2476 if pos == -1 then return -1
2478 if self[pos] == '>' then
2489 # Read a markdown link address and append it to the `out` buffer.
2490 private fun read_md_link(out: FlatBuffer, start: Int): Int do
2493 while pos < length do
2495 if c == '\\
' and pos + 1 < length then
2496 pos = escape(out, self[pos + 1], pos)
2498 var end_reached = false
2501 else if c == ' ' then
2502 if counter == 1 then end_reached = true
2503 else if c == ')' then
2505 if counter == 0 then end_reached = true
2507 if end_reached then break
2512 if pos == length then return -1
2516 # Read a markdown link text and append it to the `out` buffer.
2517 private fun read_md_link_id(out: FlatBuffer, start: Int): Int do
2520 while pos < length do
2522 var end_reached = false
2526 else if c == ']' then
2528 if counter == 0 then
2536 if end_reached then break
2539 if pos == length then return -1
2543 # Extract the XML tag name from a XML tag.
2544 private fun xml_tag: String do
2545 var tpl = new FlatBuffer
2547 if pos < length and self[1] == '/' then pos += 1
2548 while pos < length - 1 and (self[pos].is_digit or self[pos].is_letter) do
2552 return tpl.write_to_string.to_lower
2555 private fun is_valid_html_tag: Bool do
2556 if is_empty then return false
2558 if not c.is_alpha then return false
2563 # Read and escape the markdown contained in `self`.
2564 private fun escape(out: FlatBuffer, c: Char, pos: Int): Int do
2565 if c == '\\
' or c == '[' or c == ']' or c == '(' or c == ')' or c == '{' or
2566 c == '}' or c == '#' or c == '"' or c == '\'' or c == '.' or c == '<' or
2567 c
== '>' or c
== '*' or c
== '+' or c
== '-' or c
== '_' or c
== '!' or
2568 c
== '`' or c
== '~' or c
== '^' then
2576 # Extract string found at end of fence opening.
2577 private fun meta_from_fence
: nullable Text do
2578 for i
in [0..chars
.length
[ do
2580 if c
!= ' ' and c
!= '`' and c
!= '~' then
2581 return substring_from
(i
).trim
2587 # Is `self` an unsafe HTML element?
2588 private fun is_html_unsafe
: Bool do return html_unsafe_tags
.has
(self.write_to_string
)
2590 # Is `self` a HRML block element?
2591 private fun is_html_block
: Bool do return html_block_tags
.has
(self.write_to_string
)
2593 # Is `self` a link prefix?
2594 private fun is_link_prefix
: Bool do return html_link_prefixes
.has
(self.write_to_string
)
2596 private fun html_unsafe_tags
: Array[String] do return once
["applet", "head", "body", "frame", "frameset", "iframe", "script", "object"]
2598 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"]
2600 private fun html_link_prefixes
: Array[String] do return once
["http", "https", "ftp", "ftps"]
2605 # Parse `self` as markdown and return the HTML representation
2607 # var md = "**Hello World!**"
2608 # var html = md.md_to_html
2609 # assert html == "<p><strong>Hello World!</strong></p>\n"
2610 fun md_to_html
: Writable do
2611 var processor
= new MarkdownProcessor
2612 return processor
.process
(self)