#!/bin/sh
#
# Copyright (c) 2005 Jonas Fonseca
#

test_description='Test the DOM configuration module

This test checks that the normalization performed by the DOM configuration
is done correctly.
'

. "$TEST_LIB"

test_normalize_output_equals () {
	desc="$1"; shift
	config="$1"; shift
	src="$1"; shift
	out="$1"; shift

	URI="test:$(normalize "$desc")"

	sgml-parser --src "$src" --normalize "$config" --uri "$URI" --src "$src" > output
	echo "#document: $URI" > expected
	echo "$out" | sed -n '2,$p' | sed 's/^/  /' >> expected

	test_expect_success "$desc" 'cmp output expected' 
}


## Config strings ###########################################################

NOOP='cdata-sections,comments,element-content-whitespace,entities'
NOCOMMENTS='cdata-sections,element-content-whitespace,entities'
CDATA2TEXT='comments,element-content-whitespace,entities'
ENTITIES='cdata-section,comments,element-content-whitespace'
NOWSTEXT='cdata-section,comments,entities'
NORM1=''


## No-ops ###################################################################

test_normalize_output_equals \
'Normalization no-op.' \
"$NOOP" \
'<roswell>   <![CDATA[|  that  |]]><!-- ends -->  well</roswell>' \
'
element: roswell
  #text:    
  #cdata-section: |  that  |
  #comment:  ends 
  #text:   well'

test_normalize_output_equals \
'Keep comments.' \
"$NOOP" \
'and<!-- comment:1 -->or<!-- comment:2 -->' \
'
#text: and
#comment:  comment:1 
#text: or
#comment:  comment:2 '

test_normalize_output_equals \
'Keep CDATA sections ' \
"$NOOP" \
'<![CDATA[and]]>or<![CDATA[maybe]]>' \
'
#cdata-section: and
#text: or
#cdata-section: maybe'


## Comments #################################################################

test_normalize_output_equals \
'Remove comments. (I)' \
"$NOCOMMENTS" \
"<no><!-- comment -->?</no>" \
'
element: no
  #text: ?'

test_normalize_output_equals \
'Remove comments. (II)' \
"$NOCOMMENTS" \
'<!-- comment:1 -->and<!-- comment:2 -->' \
'
#text: and'

test_normalize_output_equals \
'Remove comments. (III)' \
"$NOCOMMENTS" \
'nothing to see <!-- comment -->here' \
'
#text: nothing to see here'


## Entities #################################################################

# Entities should be shown 'verbatim' here after expansion.

test_normalize_output_equals \
'Expand entities. (I)' \
"$ENTITIES" \
'a&lt;b&gt;c' \
'
#text: a&lt;b&gt;c'

test_normalize_output_equals \
'Expand entities. (II)' \
"$ENTITIES" \
'&bad-entity&good-entity;' \
'
#text: &bad-entity;&good-entity;'

test_normalize_output_equals \
'Expand entities. (III)' \
"$ENTITIES" \
'<a>&b;</a>' \
'
element: a
  #text: &b;'


## CDATA Sections ###########################################################

test_normalize_output_equals \
'Replace CDATA section with text. (I)' \
"$CDATA2TEXT" \
'<![CDATA[a small text snippet]]>' \
'
#text: a small text snippet'

test_normalize_output_equals \
'Replace CDATA section with text. (II)' \
"$CDATA2TEXT" \
'<![CDATA[a small]]> <![CDATA[text snippet]]>' \
'
#text: a small text snippet'

test_normalize_output_equals \
'Replace CDATA section with text. (III)' \
"$CDATA2TEXT" \
'before <![CDATA[and]]> after' \
'
#text: before and after'


## Element Content Whitespace ###############################################

test_normalize_output_equals \
'Remove element content whitespace. (I)' \
"$NOWSTEXT" \
'<a>
	<b>some text</b>
</a>' \
'
element: a
  element: b
    #text: some text' \

# I haven't read the specs about this thing, for now it just blasts all
# space-only text nodes. Probably not the wanted behaviour all of the time.
# --jonas
test_normalize_output_equals \
'Remove element content whitespace. (II)' \
"$NOWSTEXT" \
'<e>space between &this; &that; gets removed</e>' \
'
element: e
  #text: space between 
  entity-reference: this
  entity-reference: that
  #text:  gets removed'


## Mixes ####################################################################

test_normalize_output_equals \
'Normalization mix #1. (I)' \
"$NORM1" \
'before <![CDATA[and]]> after &some;<!--comments--> remain' \
'
#text: before and after &some; remain'

test_normalize_output_equals \
'Normalization mix #1. (II)' \
"$NORM1" \
'<!--a-->b&c;  <![CDATA[d]]>' \
'
#text: b&c;d'


## Special ELinks Extensions ################################################

test_normalize_output_equals \
'Remove unknown (HTML) elements and attributes. (I)' \
"$NOOP,unknown" \
'<html wack="..."><title w00t="...">where?<doit>here!</doit></title></html>' \
'
element: html
  element: title
    #text: where?' \

test_normalize_output_equals \
'Remove unknown (HTML) elements and attributes. (II)' \
"$NOOP,unknown" \
'<x y=""><z></z></x> aint no HTML' \
'
#text:  aint no HTML' \

test_normalize_output_equals \
'Normalize whitespace. (I)' \
"$NOOP,normalize-whitespace" \
'Here    is      a



lot of     useless	space.' \
'
#text: Here is a lot of useless space.' \

test_normalize_output_equals \
'Normalize whitespace. (II)' \
"$CDATA2TEXT,normalize-whitespace" \
'Could we <![CDATA[    please     ]]> read that again?' \
'
#text: Could we please read that again?' \

test_done