rexleparser 0.3.3 → 0.4.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- data/lib/rexleparser.rb +68 -80
- metadata +3 -3
data/lib/rexleparser.rb
CHANGED
|
@@ -6,14 +6,11 @@
|
|
|
6
6
|
class RexleParser
|
|
7
7
|
|
|
8
8
|
def initialize(s)
|
|
9
|
-
|
|
10
9
|
super()
|
|
11
10
|
@a = scan_element(s.gsub(/<\?[^>]+>/,'').split(//))
|
|
12
11
|
end
|
|
13
12
|
|
|
14
|
-
def to_a()
|
|
15
|
-
@a
|
|
16
|
-
end
|
|
13
|
+
def to_a() @a end
|
|
17
14
|
|
|
18
15
|
def to_s()
|
|
19
16
|
name, value, attributes, *remaining = @a
|
|
@@ -34,116 +31,107 @@ class RexleParser
|
|
|
34
31
|
|
|
35
32
|
a.shift until a[0] == '<' and a[1] != '/' or a.length < 1
|
|
36
33
|
|
|
37
|
-
|
|
38
|
-
a.shift
|
|
39
|
-
|
|
40
|
-
# CDATA ?
|
|
41
|
-
if a[0..1].join == '![' then
|
|
42
|
-
|
|
43
|
-
name = '!['
|
|
44
|
-
8.times{ a.shift }
|
|
45
|
-
value = ''
|
|
46
|
-
|
|
47
|
-
value << a.shift until a[0..2].join == ']]>' or a.length <= 1
|
|
48
|
-
a.slice!(0,3)
|
|
49
|
-
element = [name, value, {}]
|
|
50
|
-
elsif a[0..2].join == '!--' then
|
|
51
|
-
name = '!-'
|
|
52
|
-
#<![CDATA[
|
|
53
|
-
#<!--
|
|
54
|
-
3.times{ a.shift }
|
|
55
|
-
value = ''
|
|
56
|
-
|
|
57
|
-
value << a.shift until a[0..2].join == '-->' or a.length <= 1
|
|
58
|
-
a.slice!(0,3)
|
|
59
|
-
element = [name, value, {}]
|
|
60
|
-
else
|
|
34
|
+
return unless a.length > 1
|
|
61
35
|
|
|
62
|
-
|
|
63
|
-
name << a.shift
|
|
64
|
-
name << a.shift while a[0] != ' ' and a[0] != '>' and a[0] != '/'
|
|
36
|
+
a.shift
|
|
65
37
|
|
|
66
|
-
|
|
38
|
+
# CDATA ?
|
|
39
|
+
if a[0..1].join == '![' then
|
|
67
40
|
|
|
68
|
-
|
|
69
|
-
|
|
70
|
-
|
|
41
|
+
name = '!['
|
|
42
|
+
8.times{ a.shift }
|
|
43
|
+
value = ''
|
|
71
44
|
|
|
72
|
-
|
|
73
|
-
|
|
45
|
+
value << a.shift until a[0..2].join == ']]>' or a.length <= 1
|
|
46
|
+
a.slice!(0,3)
|
|
47
|
+
element = [name, value, {}]
|
|
48
|
+
elsif a[0..2].join == '!--' then
|
|
49
|
+
name = '!-'
|
|
50
|
+
#<![CDATA[
|
|
51
|
+
#<!--
|
|
52
|
+
3.times{ a.shift }
|
|
53
|
+
value = ''
|
|
74
54
|
|
|
75
|
-
|
|
55
|
+
value << a.shift until a[0..2].join == '-->' or a.length <= 1
|
|
56
|
+
a.slice!(0,3)
|
|
57
|
+
element = [name, value, {}]
|
|
58
|
+
else
|
|
76
59
|
|
|
77
|
-
|
|
78
|
-
|
|
60
|
+
name = ''
|
|
61
|
+
name << a.shift
|
|
62
|
+
name << a.shift while a[0] != ' ' and a[0] != '>' and a[0] != '/'
|
|
79
63
|
|
|
80
|
-
|
|
81
|
-
|
|
82
|
-
element = [name, '', attributes]
|
|
64
|
+
return unless name
|
|
83
65
|
|
|
84
|
-
|
|
85
|
-
|
|
66
|
+
# find the closing tag
|
|
67
|
+
i = a.index('>')
|
|
68
|
+
raw_values = ''
|
|
86
69
|
|
|
87
|
-
|
|
88
|
-
|
|
89
|
-
value, attributes = get_value_and_attributes(raw_values) if raw_values.length > 0
|
|
70
|
+
# is it a self closing tag?
|
|
71
|
+
if a[i-1] == '/' then
|
|
90
72
|
|
|
91
|
-
|
|
92
|
-
|
|
93
|
-
|
|
73
|
+
raw_values << a.shift until (a[0] + a[1..-1].join.strip[0]) == '/>'
|
|
74
|
+
a.shift until a[0] == '<' or a.length < 1
|
|
75
|
+
raw_values.strip!
|
|
94
76
|
|
|
95
|
-
|
|
77
|
+
attributes = get_attributes(raw_values) if raw_values.length > 0
|
|
78
|
+
return [name, '', attributes]
|
|
79
|
+
|
|
80
|
+
else
|
|
96
81
|
|
|
97
|
-
|
|
98
|
-
|
|
82
|
+
raw_values << a.shift until a[0] == '<'
|
|
83
|
+
|
|
84
|
+
if raw_values.length > 0 then
|
|
85
|
+
value, attributes = get_value_and_attribs(raw_values)
|
|
86
|
+
end
|
|
87
|
+
|
|
88
|
+
element = [name, value, attributes]
|
|
89
|
+
tag = a[0, name.length + 3].join
|
|
99
90
|
|
|
100
|
-
|
|
91
|
+
return unless a.length > 0
|
|
92
|
+
|
|
93
|
+
children = tag == ("</%s>" % name) ? false : true
|
|
101
94
|
|
|
102
|
-
|
|
95
|
+
if children == true then
|
|
103
96
|
|
|
104
|
-
|
|
105
|
-
|
|
106
|
-
a.shift until a[0] == '<' or a.length <= 1
|
|
107
|
-
else
|
|
97
|
+
xa = scan_elements(a, element) until (a[0, name.length + 3].join \
|
|
98
|
+
== "</%s>" % [name]) or a.length < 2
|
|
108
99
|
|
|
109
|
-
|
|
110
|
-
|
|
111
|
-
|
|
112
|
-
|
|
100
|
+
xa.shift until xa[0] == '>' or xa.length <= 1
|
|
101
|
+
xa.shift
|
|
102
|
+
after_text = []
|
|
103
|
+
after_text << xa.shift until xa[0] == '<' or xa.length <= 1
|
|
104
|
+
|
|
105
|
+
return after_text.length >= 1 ? [element, after_text.join] : element
|
|
113
106
|
|
|
114
|
-
|
|
115
|
-
#puts 'element: ' + element.inspect
|
|
116
|
-
end
|
|
117
|
-
end
|
|
107
|
+
else
|
|
118
108
|
|
|
119
|
-
|
|
120
|
-
|
|
109
|
+
#check for its end tag
|
|
110
|
+
a.slice!(0, name.length + 3) if a[0, name.length + 3].join \
|
|
111
|
+
== "</%s>" % name
|
|
112
|
+
after_text = []
|
|
113
|
+
after_text << a.shift until a[0] == '<' or a.length <= 1
|
|
121
114
|
|
|
122
|
-
|
|
123
|
-
if remaining.nil? then #or remaining.strip.empty? then
|
|
124
|
-
return element
|
|
125
|
-
else
|
|
126
|
-
return [element, remaining]
|
|
127
|
-
end
|
|
115
|
+
return after_text.length >= 1 ? [element, after_text.join] : element
|
|
128
116
|
|
|
129
117
|
end
|
|
130
118
|
end
|
|
131
|
-
|
|
132
119
|
end
|
|
120
|
+
|
|
133
121
|
end
|
|
134
122
|
|
|
135
123
|
def scan_elements(a, element)
|
|
136
124
|
r = scan_element(a)
|
|
137
125
|
|
|
138
126
|
if r and r[0].is_a?(Array) then
|
|
139
|
-
element = r.inject(element) {|r,x|
|
|
127
|
+
element = r.inject(element) {|r,x| r << x} if r
|
|
140
128
|
elsif r
|
|
141
129
|
element << r
|
|
142
130
|
end
|
|
131
|
+
return a
|
|
143
132
|
end
|
|
144
133
|
|
|
145
|
-
def
|
|
146
|
-
attributes = {}
|
|
134
|
+
def get_value_and_attribs(raw_values)
|
|
147
135
|
|
|
148
136
|
match_found = raw_values.match(/(.*)>([^>]*$)/)
|
|
149
137
|
if match_found then
|
metadata
CHANGED
|
@@ -2,7 +2,7 @@
|
|
|
2
2
|
name: rexleparser
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
4
|
prerelease:
|
|
5
|
-
version: 0.
|
|
5
|
+
version: 0.4.0
|
|
6
6
|
platform: ruby
|
|
7
7
|
authors:
|
|
8
8
|
- James Robertson
|
|
@@ -10,7 +10,7 @@ autorequire:
|
|
|
10
10
|
bindir: bin
|
|
11
11
|
cert_chain: []
|
|
12
12
|
|
|
13
|
-
date: 2012-
|
|
13
|
+
date: 2012-07-14 00:00:00 Z
|
|
14
14
|
dependencies: []
|
|
15
15
|
|
|
16
16
|
description:
|
|
@@ -49,6 +49,6 @@ rubyforge_project:
|
|
|
49
49
|
rubygems_version: 1.8.23
|
|
50
50
|
signing_key:
|
|
51
51
|
specification_version: 3
|
|
52
|
-
summary:
|
|
52
|
+
summary: Rexleparser is an XML parser used by the Rexle gem
|
|
53
53
|
test_files: []
|
|
54
54
|
|