annotate src/luan/modules/parsers/Html.java @ 1279:323743a7f317

add html tag.raw
author Franklin Schmidt <fschmidt@gmail.com>
date Tue, 18 Dec 2018 12:54:55 -0700
parents 9fa8b8389578
children 8d54bcc0b6d3
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
625
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
1 package luan.modules.parsers;
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
2
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
3 import java.util.List;
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
4 import java.util.ArrayList;
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
5 import java.util.Set;
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
6 import java.util.HashSet;
1267
9fa8b8389578 add LuanTable.luan;
Franklin Schmidt <fschmidt@gmail.com>
parents: 1111
diff changeset
7 import luan.LuanState;
625
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
8 import luan.LuanTable;
1111
88b5b81cad4a move Parser to luan.lib.parser
Franklin Schmidt <fschmidt@gmail.com>
parents: 775
diff changeset
9 import luan.lib.parser.Parser;
88b5b81cad4a move Parser to luan.lib.parser
Franklin Schmidt <fschmidt@gmail.com>
parents: 775
diff changeset
10 import luan.lib.parser.ParseException;
625
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
11
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
12
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
13 public final class Html {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
14
1267
9fa8b8389578 add LuanTable.luan;
Franklin Schmidt <fschmidt@gmail.com>
parents: 1111
diff changeset
15 public static LuanTable toList(LuanState luan,String text,LuanTable containerTagsTbl)
9fa8b8389578 add LuanTable.luan;
Franklin Schmidt <fschmidt@gmail.com>
parents: 1111
diff changeset
16 throws ParseException
9fa8b8389578 add LuanTable.luan;
Franklin Schmidt <fschmidt@gmail.com>
parents: 1111
diff changeset
17 {
9fa8b8389578 add LuanTable.luan;
Franklin Schmidt <fschmidt@gmail.com>
parents: 1111
diff changeset
18 return new Html(luan,text,containerTagsTbl).parse();
625
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
19 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
20
1267
9fa8b8389578 add LuanTable.luan;
Franklin Schmidt <fschmidt@gmail.com>
parents: 1111
diff changeset
21 private final LuanState luan;
625
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
22 private final Parser parser;
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
23 private final Set<String> containerTags = new HashSet<String>();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
24
1267
9fa8b8389578 add LuanTable.luan;
Franklin Schmidt <fschmidt@gmail.com>
parents: 1111
diff changeset
25 private Html(LuanState luan,String text,LuanTable containerTagsTbl) {
9fa8b8389578 add LuanTable.luan;
Franklin Schmidt <fschmidt@gmail.com>
parents: 1111
diff changeset
26 this.luan = luan;
625
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
27 this.parser = new Parser(text);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
28 for( Object v : containerTagsTbl.asList() ) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
29 containerTags.add((String)v);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
30 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
31 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
32
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
33 private LuanTable parse() throws ParseException {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
34 List list = new ArrayList();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
35 StringBuilder sb = new StringBuilder();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
36 while( !parser.endOfInput() ) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
37 if( parser.test('<') ) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
38 LuanTable tbl = parseTag();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
39 if( tbl != null ) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
40 String tagName = (String)tbl.rawGet("name");
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
41 if( containerTags.contains(tagName) ) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
42 LuanTable container = parseContainer(tbl);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
43 if( container != null )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
44 tbl = container;
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
45 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
46 if( tbl != null
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
47 || (tbl = parseComment()) != null
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
48 || (tbl = parseCdata()) != null
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
49 ) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
50 if( sb.length() > 0 ) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
51 list.add(sb.toString());
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
52 sb.setLength(0);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
53 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
54 list.add(tbl);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
55 continue;
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
56 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
57 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
58 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
59 sb.append( parser.currentChar() );
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
60 parser.anyChar();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
61 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
62 if( sb.length() > 0 )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
63 list.add(sb.toString());
1267
9fa8b8389578 add LuanTable.luan;
Franklin Schmidt <fschmidt@gmail.com>
parents: 1111
diff changeset
64 return new LuanTable(luan,list);
625
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
65 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
66
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
67 private LuanTable parseComment() {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
68 parser.begin();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
69 if( !parser.match("<!--") )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
70 return parser.failure(null);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
71 int start = parser.currentIndex();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
72 while( !parser.test("-->") ) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
73 if( !parser.anyChar() )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
74 return parser.failure(null);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
75 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
76 String text = parser.textFrom(start);
1267
9fa8b8389578 add LuanTable.luan;
Franklin Schmidt <fschmidt@gmail.com>
parents: 1111
diff changeset
77 LuanTable tbl = new LuanTable(luan);
625
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
78 tbl.rawPut("type","comment");
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
79 tbl.rawPut("text",text);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
80 return parser.success(tbl);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
81 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
82
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
83 private LuanTable parseCdata() {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
84 parser.begin();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
85 if( !parser.match("<![CDATA[") )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
86 return parser.failure(null);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
87 int start = parser.currentIndex();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
88 while( !parser.test("]]>") ) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
89 if( !parser.anyChar() )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
90 return parser.failure(null);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
91 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
92 String text = parser.textFrom(start);
1267
9fa8b8389578 add LuanTable.luan;
Franklin Schmidt <fschmidt@gmail.com>
parents: 1111
diff changeset
93 LuanTable tbl = new LuanTable(luan);
625
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
94 tbl.rawPut("type","cdata");
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
95 tbl.rawPut("text",text);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
96 return parser.success(tbl);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
97 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
98
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
99 private LuanTable parseContainer(LuanTable tag) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
100 String endTagName = '/' + (String)tag.rawGet("name");
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
101 int start = parser.begin();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
102 int end;
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
103 while(true) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
104 if( parser.test('<') ) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
105 end = parser.currentIndex();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
106 LuanTable tag2 = parseTag();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
107 String s = (String)tag2.rawGet("name");
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
108 if( s.equals(endTagName) )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
109 break;
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
110 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
111 if( !parser.anyChar() )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
112 return parser.failure(null);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
113 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
114 String text = parser.text.substring(start,end);
1267
9fa8b8389578 add LuanTable.luan;
Franklin Schmidt <fschmidt@gmail.com>
parents: 1111
diff changeset
115 LuanTable tbl = new LuanTable(luan);
625
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
116 tbl.rawPut("type","container");
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
117 tbl.rawPut("tag",tag);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
118 tbl.rawPut("text",text);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
119 return parser.success(tbl);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
120 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
121
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
122 private LuanTable parseTag() {
1279
323743a7f317 add html tag.raw
Franklin Schmidt <fschmidt@gmail.com>
parents: 1267
diff changeset
123 int tagStart = parser.begin();
625
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
124 if( !parser.match('<') )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
125 return parser.failure(null);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
126 int start = parser.currentIndex();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
127 parser.match('/');
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
128 if( !matchNameChar() )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
129 return parser.failure(null);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
130 while( matchNameChar() );
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
131 String name = parser.textFrom(start).toLowerCase();
1267
9fa8b8389578 add LuanTable.luan;
Franklin Schmidt <fschmidt@gmail.com>
parents: 1111
diff changeset
132 LuanTable attributes = new LuanTable(luan);
625
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
133 String attrName;
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
134 while( (attrName = parseAttrName()) != null ) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
135 String attrValue = parseAttrValue();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
136 attributes.rawPut( attrName, attrValue!=null ? attrValue : true );
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
137 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
138 while( matchSpace() );
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
139 boolean isEmpty = parser.match('/');
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
140 if( !parser.match('>') )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
141 return parser.failure(null);
1279
323743a7f317 add html tag.raw
Franklin Schmidt <fschmidt@gmail.com>
parents: 1267
diff changeset
142 String raw = parser.textFrom(tagStart);
1267
9fa8b8389578 add LuanTable.luan;
Franklin Schmidt <fschmidt@gmail.com>
parents: 1111
diff changeset
143 LuanTable tbl = new LuanTable(luan);
625
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
144 tbl.rawPut("type","tag");
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
145 tbl.rawPut("name",name);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
146 tbl.rawPut("attributes",attributes);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
147 tbl.rawPut("is_empty",isEmpty);
1279
323743a7f317 add html tag.raw
Franklin Schmidt <fschmidt@gmail.com>
parents: 1267
diff changeset
148 tbl.rawPut("raw",raw);
625
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
149 return parser.success(tbl);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
150 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
151
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
152 private String parseAttrName() {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
153 parser.begin();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
154 if( !matchSpace() )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
155 return parser.failure(null);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
156 while( matchSpace() );
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
157 int start = parser.currentIndex();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
158 if( !matchNameChar() )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
159 return parser.failure(null);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
160 while( matchNameChar() );
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
161 String name = parser.textFrom(start);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
162 return parser.success(name);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
163 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
164
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
165 private String parseAttrValue() {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
166 parser.begin();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
167 while( matchSpace() );
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
168 if( !parser.match('=') )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
169 return parser.failure(null);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
170 while( matchSpace() );
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
171 if( parser.anyOf("\"'") ) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
172 char quote = parser.lastChar();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
173 int start = parser.currentIndex();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
174 while( !parser.test(quote) ) {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
175 if( !parser.anyChar() )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
176 return parser.failure(null);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
177 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
178 String value = parser.textFrom(start);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
179 parser.match(quote);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
180 return parser.success(value);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
181 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
182 int start = parser.currentIndex();
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
183 if( !matchValueChar() )
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
184 return parser.failure(null);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
185 while( matchValueChar() );
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
186 String value = parser.textFrom(start);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
187 return parser.success(value);
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
188 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
189
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
190 private boolean matchNameChar() {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
191 return parser.inCharRange('a','z')
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
192 || parser.inCharRange('A','Z')
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
193 || parser.inCharRange('0','9')
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
194 || parser.anyOf("_.-:")
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
195 ;
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
196 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
197
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
198 private boolean matchValueChar() {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
199 return parser.noneOf(" \t\r\n\"'>/=");
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
200 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
201
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
202 private boolean matchSpace() {
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
203 return parser.anyOf(" \t\r\n");
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
204 }
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
205
a3c1e11fb6aa rewrite much of Html to be more understandable;
Franklin Schmidt <fschmidt@gmail.com>
parents:
diff changeset
206 }