# Improved XML highlighter by: Brian Candler <B.Candler@pobox.com>
# Incorporated some Ant logic from Christian Nicolai (http://mycrobase.de)

=Idle
=Bad
=Tag		+DefinedFunction +DefinedIdent
=TagName	+Tag
=TagEnd		+Tag
=Attr		+Type
=Constant
=String		+Constant
=Entity		+Escape
=StringEntity	+StringEscape +Entity
=Namespace	+Comment
=Decl		+Preproc +Tag
=Comment
=CommentEdge	+Comment
=PIStart	+Preproc +Tag
=PIBody		+Comment
=PIEnd		+Preproc +Tag
=CdataEdge	+Statement
=CdataBody	+String
=Keyword

# http://www.w3.org/TR/2004/REC-xml-20040204/
#
# NOTE: For UNICODE compatibility, the ranges
#	"A-Za-z_:"		-- first character of Name
#	"A-Za-z0-9._:-"		-- subsequent characters of Name
# ought to be replaced with some appropriate Unicode character classes
#
# Actually the standard tells you which characters to use, see:
#  http://www.w3.org/TR/xml11/#NT-NameStartChar

:init Idle
	*		idle noeat call=.xml()

.subr xml

:content Idle
	*		content
	"<"		tag		recolor=-1
	"&"		entityref	recolor=-1

# > is allowed
#	">"		error		noeat recolor=-1


# In several contexts a space is an error, and since a coloured space is
# not visible, we colour the next non-space character as well.

:error Bad
	*		error_visible	noeat recolor=-1
	" \t\r\n"	error

:error_visible Bad
	*		content

# Matched: &

:entityref Entity
	*		error		noeat recolor=-1
	"-.:\c"		entityref
	"#"		entityref
	";"		content

# Matched: <

:tag Tag
	*		error		noeat recolor=-1
	"/"		end_tag		recolor=-2 mark
	":\i"		start_or_empty_tag recolor=-1 mark buffer
	"?"		pi_start	recolor=-2
	"!"		decl		recolor=-2 buffer

# Matched: </

:end_tag TagEnd
	*		error		noeat recolor=-1
	":\i"		end_tag2	recolor=-1

:end_tag2 TagEnd
	*		error		noeat recolor=-1
	"-.:\c"		end_tag2
	" \t\r\n"	end_tag2_sp	recolor=-1
	">"		close_tag_end	noeat

:end_tag2_sp TagEnd
	*		end_tag3	noeat

:end_tag3 TagEnd
	*		error		noeat recolor=-1
	" \t\r\n"	end_tag3_sp	recolor=-1
	">"		close_tag_end	noeat

:end_tag3_sp TagEnd
	*		end_tag_3	noeat

# Matched: <tag

:start_or_empty_tag TagName
.ifdef ant
	*		decide_tag_type noeat strings
	"target"	kw
done
.else
	*		decide_tag_type noeat
.endif
	":"		tag_namespace	markend recolormark recolor=-1
	"-.\c"		start_or_empty_tag

:kw Keyword
	*		decide_tag_type noeat

:decide_tag_type TagName
	*		error		noeat recolor=-1
	"/"		empty_tag	mark
	" \t\r\n"	start_or_empty_tag_sp	markend buffer recolor=-1
	">"		close_tag	noeat

:tag_namespace Namespace
	*		start_or_empty_tag noeat buffer

:start_or_empty_tag_sp Idle
	*		tag_space	noeat

# Matched: <tag/

:empty_tag Tag
	*		error		noeat recolor=-1
	">"		close_tag	noeat

# Matched: <tag SPACE

:tag_space Attr
	*		error		noeat recolor=-1
	" \t\r\n"	tag_space_sp	recolor=-1
	":\i"		attr		mark
	"/"		empty_tag	recolor=-1
	">"		close_tag	noeat

:tag_space_sp Idle
	*		tag_space	mark noeat

# Matched: <tag attr

:attr Attr
	*		error		noeat recolor=-1
	":"		attr_namespace	markend recolormark recolor=-1
	"-.\c" 		attr
	" \t\r\n"	attr_sp		recolor=-1
	"="		string

:attr_namespace Namespace
	*		attr		noeat

:attr_sp Idle
	*		attr_before	noeat

:attr_before Attr
	*		error		noeat recolor=-1
	" \t\r\n"	attr_sp		recolor=-1
	"="		string

# Matched: <tag attr=

:string Attr
	*		error		noeat recolor=-1
	" \t\r\n"	string_sp	recolor=-1
	"\""		astring		recolor=-1
	"'"		bstring		recolor=-1

:string_sp Idle
	*		string		noeat

# Matched: <tag attr="

:astring String string
	*		astring
	"<"		error		noeat recolor=-1
	"&"		achar		recolor=-1
	"\""		endstring
.ifdef ant
	"$"		astring		call=ant.antvar() recolor=-1
.endif

#	">"		error		noeat recolor=-1


:bstring String string
	*		bstring
	"<"		error		noeat recolor=-1
	"&"		bchar		recolor=-1
	"'"		endstring
.ifdef ant
	"$"		bstring		call=ant.maybe_antvar()
.endif

#	">"		error		noeat recolor=-1


:achar StringEntity string
	*		error		noeat recolor=-1
	"-.:\c"		achar
	"#"		achar
	";"		astring

:bchar Entity
	*		error		noeat recolor=-1
	"-.:\c"		bchar
	"#"		bchar
	";"		bstring

# Matched: <tag attr="foo"

:endstring Attr
	*		error		noeat recolor=-1
	" \t\r\n"	tag_space_sp
	"/"		empty_tag	recolor=-1
	">"		close_tag	noeat

# This state is just to recolor the final ">" at the end of <tag attr="val">

:close_tag Tag
	*		content

:close_tag_end TagEnd
	*		content

# Matched: <?

:pi_start PIStart
	*		pi		noeat recolor=-1

:pi PIBody
	*		pi
	"?"		pi2

:pi2 PIBody
	*		pi
	">"		pi_end		recolor=-2

:pi_end PIEnd
	*		content		noeat recolor=-1

# Matched: <!

:decl Decl
	*		decl		strings
	"!--"		comment_start	recolor=-5
	"![CDATA["	cdata_start	recolor=-10
done
#	"<"		decl_nest
	">"		content

# We allow one level of <...> nesting within declarations

:decl_nest Decl
	*		decl_nest
	">"		decl

# Matched: <!--

:comment_start CommentEdge
	*		comment		noeat

:comment Comment comment
	*		comment
	"BFHNTX"	comment		noeat call=comment_todo.comment_todo()
	"-"		comment2

:comment2 Comment comment
	*		comment
	"-"		comment3

:comment3 Comment comment
	*		comment_error	noeat recolor=-3
	">"		comment_end	recolor=-3

:comment_end CommentEdge
	*		content		noeat recolor=-1

# For compatibility, the string "--" (double-hyphen) MUST NOT occur within
# comments. [http://www.w3.org/TR/REC-xml/ section 2.5]

:comment_error Bad
	*		comment
	"-"		comment_error
	">"		comment_end	recolor=-3

# Matched: <![CDATA[

:cdata_start CdataEdge
	*		cdata		noeat

:cdata CdataBody
	*		cdata
	"]"		cdata2

:cdata2 CdataBody
	*		cdata
	"]"		cdata3

:cdata3 CdataBody
	*		cdata
	"]"		cdata3
	">"		cdata_end	recolor=-3

:cdata_end CdataEdge
	*		content		noeat recolor=-1

.end