# SEE: `String::md_to_html` for a shortcut.
class MarkdownProcessor
- # `MarkdownEmitter` used for ouput.
- var emitter: MarkdownEmitter is noinit, protected writable
-
# Work in extended mode (default).
#
# Behavior changes when using extended mode:
# ~~~
var ext_mode = true
- init do self.emitter = new MarkdownEmitter(self)
+ # Disable attaching MDLocation to Tokens
+ #
+ # Locations are useful for some tools but they may
+ # cause an important time and space overhead.
+ #
+ # Default = `false`
+ var no_location = false is writable
# Process the mardown `input` string and return the processed output.
fun process(input: String): Writable do
parent.remove_surrounding_empty_lines
recurse(parent, false)
# output processed text
- return emitter.emit(parent.kind)
+ decorator.headlines.clear
+ return emit(parent.kind)
end
# Split `input` string into `MDLines` and create a parent `MDBlock` with it.
var c = input[i]
if c == '\n' then
eol = true
+ else if c == '\r' then
else if c == '\t' then
var np = pos + (4 - (pos & 3))
while pos < np do
if not line.is_empty and line.leading < 4 and line.value[line.leading] == '[' then
pos = line.leading + 1
pos = md.read_until(id, pos, ']')
- if not id.is_empty and pos + 2 < line.value.length then
+ if not id.is_empty and pos >= 0 and pos + 2 < line.value.length then
if line.value[pos + 1] == ':' then
pos += 2
pos = md.skip_spaces(pos)
- if line.value[pos] == '<' then
+ if pos >= 0 and line.value[pos] == '<' then
pos += 1
pos = md.read_until(link, pos, '>')
pos += 1
- else
+ else if pos >= 0 then
pos = md.read_until(link, pos, ' ', '\n')
end
if not link.is_empty then
pos = md.read_until(comment, pos, c)
end
end
- if not comment.is_empty then last_link_ref.title = comment.write_to_string
+ var last_link_ref = self.last_link_ref
+ if not comment.is_empty and last_link_ref != null then
+ last_link_ref.title = comment.write_to_string
+ end
end
if comment.is_empty then return false
return true
c2 = ' '
end
- var loc = new MDLocation(
- current_loc.line_start,
- current_loc.column_start + pos,
- current_loc.line_start,
- current_loc.column_start + pos)
+ var loc
+ if no_location then
+ loc = null
+ else
+ loc = new MDLocation(
+ current_loc.line_start,
+ current_loc.column_start + pos,
+ current_loc.line_start,
+ current_loc.column_start + pos)
+ end
if c == '*' then
if c1 == '*' then
end
else if c == '_' then
if c1 == '_' then
- if c0 != ' ' or c2 != ' 'then
+ if c0 != ' ' or c2 != ' ' then
return new TokenStrongUnderscore(loc, pos, c)
else
return new TokenEmUnderscore(loc, pos, c)
return -1
end
- # Location used for next parsed token.
- #
- # This location can be changed by the emitter to adjust with `\n` found
- # in the input.
- private fun current_loc: MDLocation do return emitter.current_loc
-end
-
-# Emit output corresponding to blocks content.
-#
-# Blocks are created by a previous pass in `MarkdownProcessor`.
-# The emitter use a `Decorator` to select the output format.
-class MarkdownEmitter
-
- # Kind of processor used for parsing.
- type PROCESSOR: MarkdownProcessor
-
- # Processor containing link refs.
- var processor: PROCESSOR
-
# Kind of decorator used for decoration.
type DECORATOR: Decorator
end
# Create a new `MarkdownEmitter` using a custom `decorator`.
- init with_decorator(processor: PROCESSOR, decorator: DECORATOR) do
- init processor
+ init with_decorator(decorator: DECORATOR) do
self.decorator = decorator
end
current_loc.line_start += 1
current_loc.column_start = -current_pos
end
- var mt = processor.token_at(text, current_pos)
+ var mt = token_at(text, current_pos)
if (token != null and not token isa TokenNone) and
(mt.is_same_type(token) or
(token isa TokenEmStar and mt isa TokenStrongStar) or
end
# Append `c` to current buffer.
- fun addc(c: Char) do add c.to_s
+ fun addc(c: Char) do
+ current_buffer.add c
+ end
# Append a "\n" line break.
- fun addn do add "\n"
+ fun addn do addc '\n'
end
# A Link Reference.
# Create a link with a title.
init with_title(link: String, title: nullable String) do
- self.link = link
+ init(link)
self.title = title
end
end
# Default decorator used is `HTMLDecorator`.
interface Decorator
- # Kind of emitter used for decoration.
- type EMITTER: MarkdownEmitter
+ # Kind of processor used
+ type PROCESSOR: MarkdownProcessor
# Render a single plain char.
#
# Redefine this method to add special escaping for plain text.
- fun add_char(v: EMITTER, c: Char) do v.addc c
+ fun add_char(v: PROCESSOR, c: Char) do v.addc c
# Render a ruler block.
- fun add_ruler(v: EMITTER, block: BlockRuler) is abstract
+ fun add_ruler(v: PROCESSOR, block: BlockRuler) is abstract
# Render a headline block with corresponding level.
- fun add_headline(v: EMITTER, block: BlockHeadline) is abstract
+ fun add_headline(v: PROCESSOR, block: BlockHeadline) is abstract
# Render a paragraph block.
- fun add_paragraph(v: EMITTER, block: BlockParagraph) is abstract
+ fun add_paragraph(v: PROCESSOR, block: BlockParagraph) is abstract
# Render a code or fence block.
- fun add_code(v: EMITTER, block: BlockCode) is abstract
+ fun add_code(v: PROCESSOR, block: BlockCode) is abstract
# Render a blockquote.
- fun add_blockquote(v: EMITTER, block: BlockQuote) is abstract
+ fun add_blockquote(v: PROCESSOR, block: BlockQuote) is abstract
# Render an unordered list.
- fun add_unorderedlist(v: EMITTER, block: BlockUnorderedList) is abstract
+ fun add_unorderedlist(v: PROCESSOR, block: BlockUnorderedList) is abstract
# Render an ordered list.
- fun add_orderedlist(v: EMITTER, block: BlockOrderedList) is abstract
+ fun add_orderedlist(v: PROCESSOR, block: BlockOrderedList) is abstract
# Render a list item.
- fun add_listitem(v: EMITTER, block: BlockListItem) is abstract
+ fun add_listitem(v: PROCESSOR, block: BlockListItem) is abstract
# Render an emphasis text.
- fun add_em(v: EMITTER, text: Text) is abstract
+ fun add_em(v: PROCESSOR, text: Text) is abstract
# Render a strong text.
- fun add_strong(v: EMITTER, text: Text) is abstract
+ fun add_strong(v: PROCESSOR, text: Text) is abstract
# Render a strike text.
#
# Extended mode only (see `MarkdownProcessor::ext_mode`)
- fun add_strike(v: EMITTER, text: Text) is abstract
+ fun add_strike(v: PROCESSOR, text: Text) is abstract
# Render a link.
- fun add_link(v: EMITTER, link: Text, name: Text, comment: nullable Text) is abstract
+ fun add_link(v: PROCESSOR, link: Text, name: Text, comment: nullable Text) is abstract
# Render an image.
- fun add_image(v: EMITTER, link: Text, name: Text, comment: nullable Text) is abstract
+ fun add_image(v: PROCESSOR, link: Text, name: Text, comment: nullable Text) is abstract
# Render an abbreviation.
- fun add_abbr(v: EMITTER, name: Text, comment: Text) is abstract
+ fun add_abbr(v: PROCESSOR, name: Text, comment: Text) is abstract
# Render a code span reading from a buffer.
- fun add_span_code(v: EMITTER, buffer: Text, from, to: Int) is abstract
+ fun add_span_code(v: PROCESSOR, buffer: Text, from, to: Int) is abstract
# Render a text and escape it.
- fun append_value(v: EMITTER, value: Text) is abstract
+ fun append_value(v: PROCESSOR, value: Text) is abstract
# Render code text from buffer and escape it.
- fun append_code(v: EMITTER, buffer: Text, from, to: Int) is abstract
+ fun append_code(v: PROCESSOR, buffer: Text, from, to: Int) is abstract
# Render a character escape.
- fun escape_char(v: EMITTER, char: Char) is abstract
+ fun escape_char(v: PROCESSOR, char: Char) is abstract
# Render a line break
- fun add_line_break(v: EMITTER) is abstract
+ fun add_line_break(v: PROCESSOR) is abstract
# Generate a new html valid id from a `String`.
fun strip_id(txt: String): String is abstract
redef fun add_headline(v, block) do
# save headline
- var txt = block.block.first_line.value
+ var line = block.block.first_line
+ if line == null then return
+ var txt = line.value
var id = strip_id(txt)
var lvl = block.depth
headlines[id] = new HeadLine(id, txt, lvl)
end
redef fun add_code(v, block) do
- if block isa BlockFence and block.meta != null then
- v.add "<pre class=\"{block.meta.to_s}\"><code>"
+ var meta = block.meta
+ if meta != null then
+ v.add "<pre class=\""
+ append_value(v, meta)
+ v.add "\"><code>"
else
v.add "<pre><code>"
end
fun split(line: MDLine): MDBlock do
# location for new block
var new_loc = new MDLocation(
- first_line.location.line_start,
- first_line.location.column_start,
+ first_line.as(not null).location.line_start,
+ first_line.as(not null).location.column_start,
line.location.line_end,
line.location.column_end)
# create block
if first_line == null then
last_line = null
else
- first_line.prev = null
+ first_line.as(not null).prev = null
# update current block loc
- location.line_start = first_line.location.line_start
- location.column_start = first_line.location.column_start
+ location.line_start = first_line.as(not null).location.line_start
+ location.column_start = first_line.as(not null).location.column_start
end
if first_block == null then
first_block = block
last_block = block
else
- last_block.next = block
+ last_block.as(not null).next = block
last_block = block
end
return block
first_line = line
last_line = line
else
- last_line.next_empty = line.is_empty
- line.prev_empty = last_line.is_empty
+ last_line.as(not null).next_empty = line.is_empty
+ line.prev_empty = last_line.as(not null).is_empty
line.prev = last_line
- last_line.next = line
+ last_line.as(not null).next = line
last_line = line
end
end
if line.prev == null then
first_line = line.next
else
- line.prev.next = line.next
+ line.prev.as(not null).next = line.next
end
if line.next == null then
last_line = line.prev
else
- line.next.prev = line.prev
+ line.next.as(not null).prev = line.prev
end
line.prev = null
line.next = null
text.append "\n"
line = line.next
end
+ var block = first_block
+ while block != null do
+ text.append block.text
+ text.append "\n"
+ block = block.next
+ end
return text.write_to_string
end
end
var block: MDBlock
# Output `self` using `v.decorator`.
- fun emit(v: MarkdownEmitter) do v.emit_in(self)
+ fun emit(v: MarkdownProcessor) do v.emit_in(self)
# Emit the containts of `self`, lines or blocks.
- fun emit_in(v: MarkdownEmitter) do
+ fun emit_in(v: MarkdownProcessor) do
block.remove_surrounding_empty_lines
if block.has_lines then
emit_lines(v)
end
# Emit lines contained in `block`.
- fun emit_lines(v: MarkdownEmitter) do
+ fun emit_lines(v: MarkdownProcessor) do
var tpl = v.push_buffer
var line = block.first_line
while line != null do
end
# Emit sub-blocks contained in `block`.
- fun emit_blocks(v: MarkdownEmitter) do
+ fun emit_blocks(v: MarkdownProcessor) do
var block = self.block.first_block
while block != null do
v.push_loc(block.location)
class BlockCode
super Block
+ # Any string found after fence token.
+ var meta: nullable Text
+
# Number of char to skip at the beginning of the line.
#
# Block code lines start at 4 spaces.
class BlockFence
super BlockCode
- # Any string found after fence token.
- var meta: nullable Text
-
# Fence code lines start at 0 spaces.
redef var line_start = 0
end
if depth > 0 then return
var level = 0
var line = block.first_line
+ if line == null then return
if line.is_empty then return
var start = line.leading
while start < line.value.length and line.value[start] == '#' do
# Split list block into list items sub-blocks.
private fun init_block(v: MarkdownProcessor) do
var line = block.first_line
+ if line == null then return
line = line.next
while line != null do
var t = v.line_kind(line)
leading = 0
trailing = 0
is_empty = true
- if prev != null then prev.next_empty = true
- if next != null then next.prev_empty = true
+ if prev != null then prev.as(not null).next_empty = true
+ if next != null then next.as(not null).prev_empty = true
end
# Number or leading spaces on this line.
# Used by `check_html`.
private fun read_xml_comment(first_line: MDLine, start: Int): Int do
var line: nullable MDLine = first_line
- if start + 3 < line.value.length then
- if line.value[2] == '-' and line.value[3] == '-' then
+ if start + 3 < line.as(not null).value.length then
+ if line.as(not null).value[2] == '-' and line.as(not null).value[3] == '-' then
var pos = start + 4
while line != null do
while pos < line.value.length and line.value[pos] != '-' do
super Line
redef fun process(v) do
- v.current_line = v.current_line.next
+ v.current_line = v.current_line.as(not null).next
end
end
redef fun process(v) do
var line = v.current_line
# go to block end
- var was_empty = line.prev_empty
+ var was_empty = line.as(not null).prev_empty
while line != null and not line.is_empty do
var t = v.line_kind(line)
if (v.in_list or v.ext_mode) and t isa LineList then
line = line.next
end
# build block
+ var current_block = v.current_block.as(not null)
if line != null and not line.is_empty then
- var block = v.current_block.split(line.prev.as(not null))
+ var block = current_block.split(line.prev.as(not null))
if v.in_list and not was_empty then
block.kind = new BlockNone(block)
else
block.kind = new BlockParagraph(block)
end
- v.current_block.remove_leading_empty_lines
+ current_block.remove_leading_empty_lines
else
var block: MDBlock
if line != null then
- block = v.current_block.split(line)
+ block = current_block.split(line)
else
- block = v.current_block.split(v.current_block.last_line.as(not null))
+ block = current_block.split(current_block.last_line.as(not null))
end
if v.in_list and (line == null or not line.is_empty) and not was_empty then
block.kind = new BlockNone(block)
else
block.kind = new BlockParagraph(block)
end
- v.current_block.remove_leading_empty_lines
+ current_block.remove_leading_empty_lines
end
- v.current_line = v.current_block.first_line
+ v.current_line = current_block.first_line
end
end
line = line.next
end
# split at block end line
+ var current_block = v.current_block.as(not null)
var block: MDBlock
if line != null then
- block = v.current_block.split(line.prev.as(not null))
+ block = current_block.split(line.prev.as(not null))
else
- block = v.current_block.split(v.current_block.last_line.as(not null))
+ block = current_block.split(current_block.last_line.as(not null))
end
block.kind = new BlockCode(block)
block.remove_surrounding_empty_lines
- v.current_line = v.current_block.first_line
+ v.current_line = current_block.first_line
end
end
redef fun process(v) do
var line = v.current_line
+ if line == null then return
+ var current_block = v.current_block.as(not null)
var prev = line.prev
- if prev != null then v.current_block.split(prev)
- var block = v.current_block.split(line.xml_end_line.as(not null))
+ if prev != null then current_block.split(prev)
+ var block = current_block.split(line.xml_end_line.as(not null))
block.kind = new BlockXML(block)
- v.current_block.remove_leading_empty_lines
- v.current_line = v.current_block.first_line
+ current_block.remove_leading_empty_lines
+ v.current_line = current_block.first_line
end
end
redef fun process(v) do
var line = v.current_line
+ var current_block = v.current_block.as(not null)
# go to bquote end
while line != null do
if not line.is_empty and (line.prev_empty and
# build sub block
var block: MDBlock
if line != null then
- block = v.current_block.split(line.prev.as(not null))
+ block = current_block.split(line.prev.as(not null))
else
- block = v.current_block.split(v.current_block.last_line.as(not null))
+ block = current_block.split(current_block.last_line.as(not null))
end
var kind = new BlockQuote(block)
block.kind = kind
kind.remove_block_quote_prefix(block)
v.current_line = line
v.recurse(block, false)
- v.current_line = v.current_block.first_line
+ v.current_line = current_block.first_line
end
end
redef fun process(v) do
var line = v.current_line
- if line.prev != null then v.current_block.split(line.prev.as(not null))
- var block = v.current_block.split(line.as(not null))
+ if line == null then return
+ var current_block = v.current_block.as(not null)
+ if line.prev != null then current_block.split(line.prev.as(not null))
+ var block = current_block.split(line)
block.kind = new BlockRuler(block)
- v.current_block.remove_leading_empty_lines
- v.current_line = v.current_block.first_line
+ current_block.remove_leading_empty_lines
+ v.current_line = current_block.first_line
end
end
redef fun process(v) do
# go to fence end
- var line = v.current_line.next
+ var line = v.current_line.as(not null).next
+ var current_block = v.current_block.as(not null)
while line != null do
if v.line_kind(line) isa LineFence then break
line = line.next
# build fence block
var block: MDBlock
if line != null then
- block = v.current_block.split(line.prev.as(not null))
+ block = current_block.split(line.prev.as(not null))
else
- block = v.current_block.split(v.current_block.last_line.as(not null))
+ block = current_block.split(current_block.last_line.as(not null))
end
block.remove_surrounding_empty_lines
- var meta = block.first_line.value.meta_from_fence
+ var meta = block.first_line.as(not null).value.meta_from_fence
block.kind = new BlockFence(block, meta)
- block.first_line.clear
+ block.first_line.as(not null).clear
var last = block.last_line
if last != null and v.line_kind(last) isa LineFence then
- block.last_line.clear
+ block.last_line.as(not null).clear
end
block.remove_surrounding_empty_lines
v.current_line = line
redef fun process(v) do
var line = v.current_line
+ if line == null then return
+ var current_block = v.current_block.as(not null)
var lprev = line.prev
- if lprev != null then v.current_block.split(lprev)
- var block = v.current_block.split(line.as(not null))
+ if lprev != null then current_block.split(lprev)
+ var block = current_block.split(line)
var kind = new BlockHeadline(block)
block.kind = kind
kind.transform_headline(block)
- v.current_block.remove_leading_empty_lines
- v.current_line = v.current_block.first_line
+ current_block.remove_leading_empty_lines
+ v.current_line = current_block.first_line
end
end
redef fun process(v) do
var line = v.current_line
+ if line == null then return
+ var current_block = v.current_block.as(not null)
var lprev = line.prev
- if lprev != null then v.current_block.split(lprev)
- line.next.clear
- var block = v.current_block.split(line.as(not null))
+ if lprev != null then current_block.split(lprev)
+ line.next.as(not null).clear
+ var block = current_block.split(line)
var kind = new BlockHeadline(block)
kind.depth = 1
kind.transform_headline(block)
block.kind = kind
- v.current_block.remove_leading_empty_lines
- v.current_line = v.current_block.first_line
+ current_block.remove_leading_empty_lines
+ v.current_line = current_block.first_line
end
end
redef fun process(v) do
var line = v.current_line
+ if line == null then return
+ var current_block = v.current_block.as(not null)
var lprev = line.prev
- if lprev != null then v.current_block.split(lprev)
- line.next.clear
- var block = v.current_block.split(line.as(not null))
+ if lprev != null then current_block.split(lprev)
+ line.next.as(not null).clear
+ var block = current_block.split(line)
var kind = new BlockHeadline(block)
kind.depth = 2
kind.transform_headline(block)
block.kind = kind
- v.current_block.remove_leading_empty_lines
- v.current_line = v.current_block.first_line
+ current_block.remove_leading_empty_lines
+ v.current_line = current_block.first_line
end
end
line = line.next
end
# build list block
+ var current_block = v.current_block.as(not null)
var list: MDBlock
if line != null then
- list = v.current_block.split(line.prev.as(not null))
+ list = current_block.split(line.prev.as(not null))
else
- list = v.current_block.split(v.current_block.last_line.as(not null))
+ list = current_block.split(current_block.last_line.as(not null))
end
var kind = block_kind(list)
list.kind = kind
- list.first_line.prev_empty = false
- list.last_line.next_empty = false
+ list.first_line.as(not null).prev_empty = false
+ list.last_line.as(not null).next_empty = false
list.remove_surrounding_empty_lines
- list.first_line.prev_empty = false
- list.last_line.next_empty = false
+ list.first_line.as(not null).prev_empty = false
+ list.last_line.as(not null).next_empty = false
kind.init_block(v)
var block = list.first_block
while block != null do
abstract class Token
# Location of `self` in the original input.
- var location: MDLocation
+ var location: nullable MDLocation
# Position of `self` in input independant from lines.
var pos: Int
var char: Char
# Output that token using `MarkdownEmitter::decorator`.
- fun emit(v: MarkdownEmitter) do v.decorator.add_char(v, char)
+ fun emit(v: MarkdownProcessor) do v.decorator.add_char(v, char)
end
# A token without a specific meaning.
super Token
redef fun emit(v) do
+ var current_text = v.current_text.as(not null)
var a = pos + next_pos + 1
- var b = v.processor.find_token(v.current_text.as(not null), a, self)
+ var b = v.find_token(current_text, a, self)
if b > 0 then
v.current_pos = b + next_pos
- while a < b and v.current_text[a] == ' ' do a += 1
+ while a < b and current_text[a] == ' ' do a += 1
if a < b then
- while v.current_text[b - 1] == ' ' do b -= 1
- v.decorator.add_span_code(v, v.current_text.as(not null), a, b)
+ while current_text[b - 1] == ' ' do b -= 1
+ v.decorator.add_span_code(v, current_text, a, b)
end
else
v.addc char
end
# Emit the hyperlink as link or image.
- private fun emit_hyper(v: MarkdownEmitter) is abstract
+ private fun emit_hyper(v: MarkdownProcessor) is abstract
# Check if the link is a valid link.
- private fun check_link(v: MarkdownEmitter, out: FlatBuffer, start: Int, token: Token): Int do
+ private fun check_link(v: MarkdownProcessor, out: FlatBuffer, start: Int, token: Token): Int do
var md = v.current_text
+ if md == null then return -1
var pos
if token isa TokenLink then
pos = start + 1
pos += 1
pos = md.skip_spaces(pos)
if pos < start then
- var tid = name.write_to_string.to_lower
- if v.processor.link_refs.has_key(tid) then
- var lr = v.processor.link_refs[tid]
+ var tid = name.as(not null).write_to_string.to_lower
+ if v.link_refs.has_key(tid) then
+ var lr = v.link_refs[tid]
is_abbrev = lr.is_abbrev
link = lr.link
comment = lr.title
else
id = name
end
- var tid = id.write_to_string.to_lower
- if v.processor.link_refs.has_key(tid) then
- var lr = v.processor.link_refs[tid]
+ var tid = id.as(not null).write_to_string.to_lower
+ if v.link_refs.has_key(tid) then
+ var lr = v.link_refs[tid]
link = lr.link
comment = lr.title
end
else
- var tid = name.write_to_string.replace("\n", " ").to_lower
- if v.processor.link_refs.has_key(tid) then
- var lr = v.processor.link_refs[tid]
+ var tid = name.as(not null).write_to_string.replace("\n", " ").to_lower
+ if v.link_refs.has_key(tid) then
+ var lr = v.link_refs[tid]
link = lr.link
comment = lr.title
pos = old_pos
# Is the HTML valid?
# Also take care of link and mailto shortcuts.
- private fun check_html(v: MarkdownEmitter, out: FlatBuffer, md: Text, start: Int): Int do
+ private fun check_html(v: MarkdownProcessor, out: FlatBuffer, md: Text, start: Int): Int do
# check for auto links
var tmp = new FlatBuffer
var pos = md.read_until(tmp, start + 1, ':', ' ', '>', '\n')
redef fun emit(v) do
v.current_pos += 1
- v.addc v.current_text[v.current_pos]
+ v.addc v.current_text.as(not null)[v.current_pos]
end
end
if c == '\\' and pos + 1 < length then
pos = escape(out, self[pos + 1], pos)
else
- var end_reached = false
- for n in nend do
- if c == n then
- end_reached = true
- break
- end
- end
- if end_reached then break
+ for n in nend do if c == n then break label
out.add c
end
pos += 1
- end
+ end label
if pos == length then return -1
return pos
end