1 // Copyright 2010 The Go Authors. All rights reserved.
2 // Use of this source code is governed by a BSD-style
3 // license that can be found in the LICENSE file.
4 5 /*
6 Package html implements an HTML5-compliant tokenizer and parser.
7 8 Tokenization is done by creating a Tokenizer for an io.Reader r. It is the
9 caller's responsibility to ensure that r provides UTF-8 encoded HTML.
10 11 z := html.NewTokenizer(r)
12 13 Given a Tokenizer z, the HTML is tokenized by repeatedly calling z.Next(),
14 which parses the next token and returns its type, or an error:
15 16 for {
17 tt := z.Next()
18 if tt == html.ErrorToken {
19 // ...
20 return ...
21 }
22 // Process the current token.
23 }
24 25 There are two APIs for retrieving the current token. The high-level API is to
26 call Token; the low-level API is to call Text or TagName / TagAttr. Both APIs
27 allow optionally calling Raw after Next but before Token, Text, TagName, or
28 TagAttr. In EBNF notation, the valid call sequence per token is:
29 30 Next {Raw} [ Token | Text | TagName {TagAttr} ]
31 32 Token returns an independent data structure that completely describes a token.
33 Entities (such as "<") are unescaped, tag names and attribute keys are
34 lower-cased, and attributes are collected into a []Attribute. For example:
35 36 for {
37 if z.Next() == html.ErrorToken {
38 // Returning io.EOF indicates success.
39 return z.Err()
40 }
41 emitToken(z.Token())
42 }
43 44 The low-level API performs fewer allocations and copies, but the contents of
45 the []byte values returned by Text, TagName and TagAttr may change on the next
46 call to Next. For example, to extract an HTML page's anchor text:
47 48 depth := 0
49 for {
50 tt := z.Next()
51 switch tt {
52 case html.ErrorToken:
53 return z.Err()
54 case html.TextToken:
55 if depth > 0 {
56 // emitBytes should copy the []byte it receives,
57 // if it doesn't process it immediately.
58 emitBytes(z.Text())
59 }
60 case html.StartTagToken, html.EndTagToken:
61 tn, _ := z.TagName()
62 if len(tn) == 1 && tn[0] == 'a' {
63 if tt == html.StartTagToken {
64 depth++
65 } else {
66 depth--
67 }
68 }
69 }
70 }
71 72 Parsing is done by calling Parse with an io.Reader, which returns the root of
73 the parse tree (the document element) as a *Node. It is the caller's
74 responsibility to ensure that the Reader provides UTF-8 encoded HTML. For
75 example, to process each anchor node in depth-first order:
76 77 doc, err := html.Parse(r)
78 if err != nil {
79 // ...
80 }
81 for n := range doc.Descendants() {
82 if n.Type == html.ElementNode && n.Data == "a" {
83 // Do something with n...
84 }
85 }
86 87 The relevant specifications include:
88 https://html.spec.whatwg.org/multipage/syntax.html and
89 https://html.spec.whatwg.org/multipage/syntax.html#tokenization
90 91 # Security Considerations
92 93 Care should be taken when parsing and interpreting HTML, whether full documents
94 or fragments, within the framework of the HTML specification, especially with
95 regard to untrusted inputs.
96 97 This package provides both a tokenizer and a parser, which implement the
98 tokenization, and tokenization and tree construction stages of the WHATWG HTML
99 parsing specification respectively. While the tokenizer parses and normalizes
100 individual HTML tokens, only the parser constructs the DOM tree from the
101 tokenized HTML, as described in the tree construction stage of the
102 specification, dynamically modifying or extending the document's DOM tree.
103 104 If your use case requires semantically well-formed HTML documents, as defined by
105 the WHATWG specification, the parser should be used rather than the tokenizer.
106 107 In security contexts, if trust decisions are being made using the tokenized or
108 parsed content, the input must be re-serialized (for instance by using Render or
109 Token.String) in order for those trust decisions to hold, as the process of
110 tokenization or parsing may alter the content.
111 */
112 package html // import "golang.org/x/net/html"
113 114 // The tokenization algorithm implemented by this package is not a line-by-line
115 // transliteration of the relatively verbose state-machine in the WHATWG
116 // specification. A more direct approach is used instead, where the program
117 // counter implies the state, such as whether it is tokenizing a tag or a text
118 // node. Specification compliance is verified by checking expected and actual
119 // outputs over a test suite rather than aiming for algorithmic fidelity.
120 121 // TODO(nigeltao): Does a DOM API belong in this package or a separate one?
122 // TODO(nigeltao): How does parsing interact with a JavaScript engine?
123