1
0
Fork 0
forked from forgejo/forgejo

upgrade to most recent bluemonday (#11007)

* upgrade to most recent bluemonday

* make vendor

* update tests for bluemonday

* update tests for bluemonday

* update tests for bluemonday
This commit is contained in:
techknowlogick 2020-04-07 16:08:47 -04:00 committed by GitHub
parent 4c54477bb5
commit d00ebf445b
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
50 changed files with 4977 additions and 300 deletions

22
vendor/github.com/aymerick/douceur/LICENSE generated vendored Normal file
View file

@ -0,0 +1,22 @@
The MIT License (MIT)
Copyright (c) 2015 Aymerick JEHANNE
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

60
vendor/github.com/aymerick/douceur/css/declaration.go generated vendored Normal file
View file

@ -0,0 +1,60 @@
package css
import "fmt"
// Declaration represents a parsed style property
type Declaration struct {
Property string
Value string
Important bool
}
// NewDeclaration instanciates a new Declaration
func NewDeclaration() *Declaration {
return &Declaration{}
}
// Returns string representation of the Declaration
func (decl *Declaration) String() string {
return decl.StringWithImportant(true)
}
// StringWithImportant returns string representation with optional !important part
func (decl *Declaration) StringWithImportant(option bool) string {
result := fmt.Sprintf("%s: %s", decl.Property, decl.Value)
if option && decl.Important {
result += " !important"
}
result += ";"
return result
}
// Equal returns true if both Declarations are equals
func (decl *Declaration) Equal(other *Declaration) bool {
return (decl.Property == other.Property) && (decl.Value == other.Value) && (decl.Important == other.Important)
}
//
// DeclarationsByProperty
//
// DeclarationsByProperty represents sortable style declarations
type DeclarationsByProperty []*Declaration
// Implements sort.Interface
func (declarations DeclarationsByProperty) Len() int {
return len(declarations)
}
// Implements sort.Interface
func (declarations DeclarationsByProperty) Swap(i, j int) {
declarations[i], declarations[j] = declarations[j], declarations[i]
}
// Implements sort.Interface
func (declarations DeclarationsByProperty) Less(i, j int) bool {
return declarations[i].Property < declarations[j].Property
}

230
vendor/github.com/aymerick/douceur/css/rule.go generated vendored Normal file
View file

@ -0,0 +1,230 @@
package css
import (
"fmt"
"strings"
)
const (
indentSpace = 2
)
// RuleKind represents a Rule kind
type RuleKind int
// Rule kinds
const (
QualifiedRule RuleKind = iota
AtRule
)
// At Rules than have Rules inside their block instead of Declarations
var atRulesWithRulesBlock = []string{
"@document", "@font-feature-values", "@keyframes", "@media", "@supports",
}
// Rule represents a parsed CSS rule
type Rule struct {
Kind RuleKind
// At Rule name (eg: "@media")
Name string
// Raw prelude
Prelude string
// Qualified Rule selectors parsed from prelude
Selectors []string
// Style properties
Declarations []*Declaration
// At Rule embedded rules
Rules []*Rule
// Current rule embedding level
EmbedLevel int
}
// NewRule instanciates a new Rule
func NewRule(kind RuleKind) *Rule {
return &Rule{
Kind: kind,
}
}
// Returns string representation of rule kind
func (kind RuleKind) String() string {
switch kind {
case QualifiedRule:
return "Qualified Rule"
case AtRule:
return "At Rule"
default:
return "WAT"
}
}
// EmbedsRules returns true if this rule embeds another rules
func (rule *Rule) EmbedsRules() bool {
if rule.Kind == AtRule {
for _, atRuleName := range atRulesWithRulesBlock {
if rule.Name == atRuleName {
return true
}
}
}
return false
}
// Equal returns true if both rules are equals
func (rule *Rule) Equal(other *Rule) bool {
if (rule.Kind != other.Kind) ||
(rule.Prelude != other.Prelude) ||
(rule.Name != other.Name) {
return false
}
if (len(rule.Selectors) != len(other.Selectors)) ||
(len(rule.Declarations) != len(other.Declarations)) ||
(len(rule.Rules) != len(other.Rules)) {
return false
}
for i, sel := range rule.Selectors {
if sel != other.Selectors[i] {
return false
}
}
for i, decl := range rule.Declarations {
if !decl.Equal(other.Declarations[i]) {
return false
}
}
for i, rule := range rule.Rules {
if !rule.Equal(other.Rules[i]) {
return false
}
}
return true
}
// Diff returns a string representation of rules differences
func (rule *Rule) Diff(other *Rule) []string {
result := []string{}
if rule.Kind != other.Kind {
result = append(result, fmt.Sprintf("Kind: %s | %s", rule.Kind.String(), other.Kind.String()))
}
if rule.Prelude != other.Prelude {
result = append(result, fmt.Sprintf("Prelude: \"%s\" | \"%s\"", rule.Prelude, other.Prelude))
}
if rule.Name != other.Name {
result = append(result, fmt.Sprintf("Name: \"%s\" | \"%s\"", rule.Name, other.Name))
}
if len(rule.Selectors) != len(other.Selectors) {
result = append(result, fmt.Sprintf("Selectors: %v | %v", strings.Join(rule.Selectors, ", "), strings.Join(other.Selectors, ", ")))
} else {
for i, sel := range rule.Selectors {
if sel != other.Selectors[i] {
result = append(result, fmt.Sprintf("Selector: \"%s\" | \"%s\"", sel, other.Selectors[i]))
}
}
}
if len(rule.Declarations) != len(other.Declarations) {
result = append(result, fmt.Sprintf("Declarations Nb: %d | %d", len(rule.Declarations), len(other.Declarations)))
} else {
for i, decl := range rule.Declarations {
if !decl.Equal(other.Declarations[i]) {
result = append(result, fmt.Sprintf("Declaration: \"%s\" | \"%s\"", decl.String(), other.Declarations[i].String()))
}
}
}
if len(rule.Rules) != len(other.Rules) {
result = append(result, fmt.Sprintf("Rules Nb: %d | %d", len(rule.Rules), len(other.Rules)))
} else {
for i, rule := range rule.Rules {
if !rule.Equal(other.Rules[i]) {
result = append(result, fmt.Sprintf("Rule: \"%s\" | \"%s\"", rule.String(), other.Rules[i].String()))
}
}
}
return result
}
// Returns the string representation of a rule
func (rule *Rule) String() string {
result := ""
if rule.Kind == QualifiedRule {
for i, sel := range rule.Selectors {
if i != 0 {
result += ", "
}
result += sel
}
} else {
// AtRule
result += fmt.Sprintf("%s", rule.Name)
if rule.Prelude != "" {
if result != "" {
result += " "
}
result += fmt.Sprintf("%s", rule.Prelude)
}
}
if (len(rule.Declarations) == 0) && (len(rule.Rules) == 0) {
result += ";"
} else {
result += " {\n"
if rule.EmbedsRules() {
for _, subRule := range rule.Rules {
result += fmt.Sprintf("%s%s\n", rule.indent(), subRule.String())
}
} else {
for _, decl := range rule.Declarations {
result += fmt.Sprintf("%s%s\n", rule.indent(), decl.String())
}
}
result += fmt.Sprintf("%s}", rule.indentEndBlock())
}
return result
}
// Returns identation spaces for declarations and rules
func (rule *Rule) indent() string {
result := ""
for i := 0; i < ((rule.EmbedLevel + 1) * indentSpace); i++ {
result += " "
}
return result
}
// Returns identation spaces for end of block character
func (rule *Rule) indentEndBlock() string {
result := ""
for i := 0; i < (rule.EmbedLevel * indentSpace); i++ {
result += " "
}
return result
}

25
vendor/github.com/aymerick/douceur/css/stylesheet.go generated vendored Normal file
View file

@ -0,0 +1,25 @@
package css
// Stylesheet represents a parsed stylesheet
type Stylesheet struct {
Rules []*Rule
}
// NewStylesheet instanciate a new Stylesheet
func NewStylesheet() *Stylesheet {
return &Stylesheet{}
}
// Returns string representation of the Stylesheet
func (sheet *Stylesheet) String() string {
result := ""
for _, rule := range sheet.Rules {
if result != "" {
result += "\n"
}
result += rule.String()
}
return result
}

22
vendor/github.com/chris-ramon/douceur/LICENSE generated vendored Normal file
View file

@ -0,0 +1,22 @@
The MIT License (MIT)
Copyright (c) 2015 Aymerick JEHANNE
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

409
vendor/github.com/chris-ramon/douceur/parser/parser.go generated vendored Normal file
View file

@ -0,0 +1,409 @@
package parser
import (
"errors"
"fmt"
"regexp"
"strings"
"github.com/gorilla/css/scanner"
"github.com/aymerick/douceur/css"
)
const (
importantSuffixRegexp = `(?i)\s*!important\s*$`
)
var (
importantRegexp *regexp.Regexp
)
// Parser represents a CSS parser
type Parser struct {
scan *scanner.Scanner // Tokenizer
// Tokens parsed but not consumed yet
tokens []*scanner.Token
// Rule embedding level
embedLevel int
}
func init() {
importantRegexp = regexp.MustCompile(importantSuffixRegexp)
}
// NewParser instanciates a new parser
func NewParser(txt string) *Parser {
return &Parser{
scan: scanner.New(txt),
}
}
// Parse parses a whole stylesheet
func Parse(text string) (*css.Stylesheet, error) {
result, err := NewParser(text).ParseStylesheet()
if err != nil {
return nil, err
}
return result, nil
}
// ParseDeclarations parses CSS declarations
func ParseDeclarations(text string) ([]*css.Declaration, error) {
result, err := NewParser(text).ParseDeclarations()
if err != nil {
return nil, err
}
return result, nil
}
// ParseStylesheet parses a stylesheet
func (parser *Parser) ParseStylesheet() (*css.Stylesheet, error) {
result := css.NewStylesheet()
// Parse BOM
if _, err := parser.parseBOM(); err != nil {
return result, err
}
// Parse list of rules
rules, err := parser.ParseRules()
if err != nil {
return result, err
}
result.Rules = rules
return result, nil
}
// ParseRules parses a list of rules
func (parser *Parser) ParseRules() ([]*css.Rule, error) {
result := []*css.Rule{}
inBlock := false
if parser.tokenChar("{") {
// parsing a block of rules
inBlock = true
parser.embedLevel++
parser.shiftToken()
}
for parser.tokenParsable() {
if parser.tokenIgnorable() {
parser.shiftToken()
} else if parser.tokenChar("}") {
if !inBlock {
errMsg := fmt.Sprintf("Unexpected } character: %s", parser.nextToken().String())
return result, errors.New(errMsg)
}
parser.shiftToken()
parser.embedLevel--
// finished
break
} else {
rule, err := parser.ParseRule()
if err != nil {
return result, err
}
rule.EmbedLevel = parser.embedLevel
result = append(result, rule)
}
}
return result, parser.err()
}
// ParseRule parses a rule
func (parser *Parser) ParseRule() (*css.Rule, error) {
if parser.tokenAtKeyword() {
return parser.parseAtRule()
}
return parser.parseQualifiedRule()
}
// ParseDeclarations parses a list of declarations
func (parser *Parser) ParseDeclarations() ([]*css.Declaration, error) {
result := []*css.Declaration{}
if parser.tokenChar("{") {
parser.shiftToken()
}
for parser.tokenParsable() {
if parser.tokenIgnorable() {
parser.shiftToken()
} else if parser.tokenChar("}") {
// end of block
parser.shiftToken()
break
} else {
declaration, err := parser.ParseDeclaration()
if err != nil {
return result, err
}
result = append(result, declaration)
}
}
return result, parser.err()
}
// ParseDeclaration parses a declaration
func (parser *Parser) ParseDeclaration() (*css.Declaration, error) {
result := css.NewDeclaration()
curValue := ""
for parser.tokenParsable() {
if parser.tokenChar(":") {
result.Property = strings.TrimSpace(curValue)
curValue = ""
parser.shiftToken()
} else if parser.tokenChar(";") || parser.tokenChar("}") {
if result.Property == "" {
errMsg := fmt.Sprintf("Unexpected ; character: %s", parser.nextToken().String())
return result, errors.New(errMsg)
}
if importantRegexp.MatchString(curValue) {
result.Important = true
curValue = importantRegexp.ReplaceAllString(curValue, "")
}
result.Value = strings.TrimSpace(curValue)
if parser.tokenChar(";") {
parser.shiftToken()
}
// finished
break
} else {
token := parser.shiftToken()
curValue += token.Value
}
}
// log.Printf("[parsed] Declaration: %s", result.String())
return result, parser.err()
}
// Parse an At Rule
func (parser *Parser) parseAtRule() (*css.Rule, error) {
// parse rule name (eg: "@import")
token := parser.shiftToken()
result := css.NewRule(css.AtRule)
result.Name = token.Value
for parser.tokenParsable() {
if parser.tokenChar(";") {
parser.shiftToken()
// finished
break
} else if parser.tokenChar("{") {
if result.EmbedsRules() {
// parse rules block
rules, err := parser.ParseRules()
if err != nil {
return result, err
}
result.Rules = rules
} else {
// parse declarations block
declarations, err := parser.ParseDeclarations()
if err != nil {
return result, err
}
result.Declarations = declarations
}
// finished
break
} else {
// parse prelude
prelude, err := parser.parsePrelude()
if err != nil {
return result, err
}
result.Prelude = prelude
}
}
// log.Printf("[parsed] Rule: %s", result.String())
return result, parser.err()
}
// Parse a Qualified Rule
func (parser *Parser) parseQualifiedRule() (*css.Rule, error) {
result := css.NewRule(css.QualifiedRule)
for parser.tokenParsable() {
if parser.tokenChar("{") {
if result.Prelude == "" {
errMsg := fmt.Sprintf("Unexpected { character: %s", parser.nextToken().String())
return result, errors.New(errMsg)
}
// parse declarations block
declarations, err := parser.ParseDeclarations()
if err != nil {
return result, err
}
result.Declarations = declarations
// finished
break
} else {
// parse prelude
prelude, err := parser.parsePrelude()
if err != nil {
return result, err
}
result.Prelude = prelude
}
}
result.Selectors = strings.Split(result.Prelude, ",")
for i, sel := range result.Selectors {
result.Selectors[i] = strings.TrimSpace(sel)
}
// log.Printf("[parsed] Rule: %s", result.String())
return result, parser.err()
}
// Parse Rule prelude
func (parser *Parser) parsePrelude() (string, error) {
result := ""
for parser.tokenParsable() && !parser.tokenEndOfPrelude() {
token := parser.shiftToken()
result += token.Value
}
result = strings.TrimSpace(result)
// log.Printf("[parsed] prelude: %s", result)
return result, parser.err()
}
// Parse BOM
func (parser *Parser) parseBOM() (bool, error) {
if parser.nextToken().Type == scanner.TokenBOM {
parser.shiftToken()
return true, nil
}
return false, parser.err()
}
// Returns next token without removing it from tokens buffer
func (parser *Parser) nextToken() *scanner.Token {
if len(parser.tokens) == 0 {
// fetch next token
nextToken := parser.scan.Next()
// log.Printf("[token] %s => %v", nextToken.Type.String(), nextToken.Value)
// queue it
parser.tokens = append(parser.tokens, nextToken)
}
return parser.tokens[0]
}
// Returns next token and remove it from the tokens buffer
func (parser *Parser) shiftToken() *scanner.Token {
var result *scanner.Token
result, parser.tokens = parser.tokens[0], parser.tokens[1:]
return result
}
// Returns tokenizer error, or nil if no error
func (parser *Parser) err() error {
if parser.tokenError() {
token := parser.nextToken()
return fmt.Errorf("Tokenizer error: %s", token.String())
}
return nil
}
// Returns true if next token is Error
func (parser *Parser) tokenError() bool {
return parser.nextToken().Type == scanner.TokenError
}
// Returns true if next token is EOF
func (parser *Parser) tokenEOF() bool {
return parser.nextToken().Type == scanner.TokenEOF
}
// Returns true if next token is a whitespace
func (parser *Parser) tokenWS() bool {
return parser.nextToken().Type == scanner.TokenS
}
// Returns true if next token is a comment
func (parser *Parser) tokenComment() bool {
return parser.nextToken().Type == scanner.TokenComment
}
// Returns true if next token is a CDO or a CDC
func (parser *Parser) tokenCDOorCDC() bool {
switch parser.nextToken().Type {
case scanner.TokenCDO, scanner.TokenCDC:
return true
default:
return false
}
}
// Returns true if next token is ignorable
func (parser *Parser) tokenIgnorable() bool {
return parser.tokenWS() || parser.tokenComment() || parser.tokenCDOorCDC()
}
// Returns true if next token is parsable
func (parser *Parser) tokenParsable() bool {
return !parser.tokenEOF() && !parser.tokenError()
}
// Returns true if next token is an At Rule keyword
func (parser *Parser) tokenAtKeyword() bool {
return parser.nextToken().Type == scanner.TokenAtKeyword
}
// Returns true if next token is given character
func (parser *Parser) tokenChar(value string) bool {
token := parser.nextToken()
return (token.Type == scanner.TokenChar) && (token.Value == value)
}
// Returns true if next token marks the end of a prelude
func (parser *Parser) tokenEndOfPrelude() bool {
return parser.tokenChar(";") || parser.tokenChar("{")
}

27
vendor/github.com/gorilla/css/LICENSE generated vendored Normal file
View file

@ -0,0 +1,27 @@
Copyright (c) 2013, Gorilla web toolkit
All rights reserved.
Redistribution and use in source and binary forms, with or without modification,
are permitted provided that the following conditions are met:
Redistributions of source code must retain the above copyright notice, this
list of conditions and the following disclaimer.
Redistributions in binary form must reproduce the above copyright notice, this
list of conditions and the following disclaimer in the documentation and/or
other materials provided with the distribution.
Neither the name of the {organization} nor the names of its
contributors may be used to endorse or promote products derived from
this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR
ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

33
vendor/github.com/gorilla/css/scanner/doc.go generated vendored Normal file
View file

@ -0,0 +1,33 @@
// Copyright 2012 The Gorilla Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
/*
Package gorilla/css/scanner generates tokens for a CSS3 input.
It follows the CSS3 specification located at:
http://www.w3.org/TR/css3-syntax/
To use it, create a new scanner for a given CSS string and call Next() until
the token returned has type TokenEOF or TokenError:
s := scanner.New(myCSS)
for {
token := s.Next()
if token.Type == scanner.TokenEOF || token.Type == scanner.TokenError {
break
}
// Do something with the token...
}
Following the CSS3 specification, an error can only occur when the scanner
finds an unclosed quote or unclosed comment. In these cases the text becomes
"untokenizable". Everything else is tokenizable and it is up to a parser
to make sense of the token stream (or ignore nonsensical token sequences).
Note: the scanner doesn't perform lexical analysis or, in other words, it
doesn't care about the token context. It is intended to be used by a
lexer or parser.
*/
package scanner

356
vendor/github.com/gorilla/css/scanner/scanner.go generated vendored Normal file
View file

@ -0,0 +1,356 @@
// Copyright 2012 The Gorilla Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package scanner
import (
"fmt"
"regexp"
"strings"
"unicode"
"unicode/utf8"
)
// tokenType identifies the type of lexical tokens.
type tokenType int
// String returns a string representation of the token type.
func (t tokenType) String() string {
return tokenNames[t]
}
// Token represents a token and the corresponding string.
type Token struct {
Type tokenType
Value string
Line int
Column int
}
// String returns a string representation of the token.
func (t *Token) String() string {
if len(t.Value) > 10 {
return fmt.Sprintf("%s (line: %d, column: %d): %.10q...",
t.Type, t.Line, t.Column, t.Value)
}
return fmt.Sprintf("%s (line: %d, column: %d): %q",
t.Type, t.Line, t.Column, t.Value)
}
// All tokens -----------------------------------------------------------------
// The complete list of tokens in CSS3.
const (
// Scanner flags.
TokenError tokenType = iota
TokenEOF
// From now on, only tokens from the CSS specification.
TokenIdent
TokenAtKeyword
TokenString
TokenHash
TokenNumber
TokenPercentage
TokenDimension
TokenURI
TokenUnicodeRange
TokenCDO
TokenCDC
TokenS
TokenComment
TokenFunction
TokenIncludes
TokenDashMatch
TokenPrefixMatch
TokenSuffixMatch
TokenSubstringMatch
TokenChar
TokenBOM
)
// tokenNames maps tokenType's to their names. Used for conversion to string.
var tokenNames = map[tokenType]string{
TokenError: "error",
TokenEOF: "EOF",
TokenIdent: "IDENT",
TokenAtKeyword: "ATKEYWORD",
TokenString: "STRING",
TokenHash: "HASH",
TokenNumber: "NUMBER",
TokenPercentage: "PERCENTAGE",
TokenDimension: "DIMENSION",
TokenURI: "URI",
TokenUnicodeRange: "UNICODE-RANGE",
TokenCDO: "CDO",
TokenCDC: "CDC",
TokenS: "S",
TokenComment: "COMMENT",
TokenFunction: "FUNCTION",
TokenIncludes: "INCLUDES",
TokenDashMatch: "DASHMATCH",
TokenPrefixMatch: "PREFIXMATCH",
TokenSuffixMatch: "SUFFIXMATCH",
TokenSubstringMatch: "SUBSTRINGMATCH",
TokenChar: "CHAR",
TokenBOM: "BOM",
}
// Macros and productions -----------------------------------------------------
// http://www.w3.org/TR/css3-syntax/#tokenization
var macroRegexp = regexp.MustCompile(`\{[a-z]+\}`)
// macros maps macro names to patterns to be expanded.
var macros = map[string]string{
// must be escaped: `\.+*?()|[]{}^$`
"ident": `-?{nmstart}{nmchar}*`,
"name": `{nmchar}+`,
"nmstart": `[a-zA-Z_]|{nonascii}|{escape}`,
"nonascii": "[\u0080-\uD7FF\uE000-\uFFFD\U00010000-\U0010FFFF]",
"unicode": `\\[0-9a-fA-F]{1,6}{wc}?`,
"escape": "{unicode}|\\\\[\u0020-\u007E\u0080-\uD7FF\uE000-\uFFFD\U00010000-\U0010FFFF]",
"nmchar": `[a-zA-Z0-9_-]|{nonascii}|{escape}`,
"num": `[0-9]*\.[0-9]+|[0-9]+`,
"string": `"(?:{stringchar}|')*"|'(?:{stringchar}|")*'`,
"stringchar": `{urlchar}|[ ]|\\{nl}`,
"nl": `[\n\r\f]|\r\n`,
"w": `{wc}*`,
"wc": `[\t\n\f\r ]`,
// urlchar should accept [(ascii characters minus those that need escaping)|{nonascii}|{escape}]
// ASCII characters range = `[\u0020-\u007e]`
// Skip space \u0020 = `[\u0021-\u007e]`
// Skip quotation mark \0022 = `[\u0021\u0023-\u007e]`
// Skip apostrophe \u0027 = `[\u0021\u0023-\u0026\u0028-\u007e]`
// Skip reverse solidus \u005c = `[\u0021\u0023-\u0026\u0028-\u005b\u005d\u007e]`
// Finally, the left square bracket (\u005b) and right (\u005d) needs escaping themselves
"urlchar": "[\u0021\u0023-\u0026\u0028-\\\u005b\\\u005d-\u007E]|{nonascii}|{escape}",
}
// productions maps the list of tokens to patterns to be expanded.
var productions = map[tokenType]string{
// Unused regexps (matched using other methods) are commented out.
TokenIdent: `{ident}`,
TokenAtKeyword: `@{ident}`,
TokenString: `{string}`,
TokenHash: `#{name}`,
TokenNumber: `{num}`,
TokenPercentage: `{num}%`,
TokenDimension: `{num}{ident}`,
TokenURI: `url\({w}(?:{string}|{urlchar}*?){w}\)`,
TokenUnicodeRange: `U\+[0-9A-F\?]{1,6}(?:-[0-9A-F]{1,6})?`,
//TokenCDO: `<!--`,
TokenCDC: `-->`,
TokenS: `{wc}+`,
TokenComment: `/\*[^\*]*[\*]+(?:[^/][^\*]*[\*]+)*/`,
TokenFunction: `{ident}\(`,
//TokenIncludes: `~=`,
//TokenDashMatch: `\|=`,
//TokenPrefixMatch: `\^=`,
//TokenSuffixMatch: `\$=`,
//TokenSubstringMatch: `\*=`,
//TokenChar: `[^"']`,
//TokenBOM: "\uFEFF",
}
// matchers maps the list of tokens to compiled regular expressions.
//
// The map is filled on init() using the macros and productions defined in
// the CSS specification.
var matchers = map[tokenType]*regexp.Regexp{}
// matchOrder is the order to test regexps when first-char shortcuts
// can't be used.
var matchOrder = []tokenType{
TokenURI,
TokenFunction,
TokenUnicodeRange,
TokenIdent,
TokenDimension,
TokenPercentage,
TokenNumber,
TokenCDC,
}
func init() {
// replace macros and compile regexps for productions.
replaceMacro := func(s string) string {
return "(?:" + macros[s[1:len(s)-1]] + ")"
}
for t, s := range productions {
for macroRegexp.MatchString(s) {
s = macroRegexp.ReplaceAllStringFunc(s, replaceMacro)
}
matchers[t] = regexp.MustCompile("^(?:" + s + ")")
}
}
// Scanner --------------------------------------------------------------------
// New returns a new CSS scanner for the given input.
func New(input string) *Scanner {
// Normalize newlines.
input = strings.Replace(input, "\r\n", "\n", -1)
return &Scanner{
input: input,
row: 1,
col: 1,
}
}
// Scanner scans an input and emits tokens following the CSS3 specification.
type Scanner struct {
input string
pos int
row int
col int
err *Token
}
// Next returns the next token from the input.
//
// At the end of the input the token type is TokenEOF.
//
// If the input can't be tokenized the token type is TokenError. This occurs
// in case of unclosed quotation marks or comments.
func (s *Scanner) Next() *Token {
if s.err != nil {
return s.err
}
if s.pos >= len(s.input) {
s.err = &Token{TokenEOF, "", s.row, s.col}
return s.err
}
if s.pos == 0 {
// Test BOM only once, at the beginning of the file.
if strings.HasPrefix(s.input, "\uFEFF") {
return s.emitSimple(TokenBOM, "\uFEFF")
}
}
// There's a lot we can guess based on the first byte so we'll take a
// shortcut before testing multiple regexps.
input := s.input[s.pos:]
switch input[0] {
case '\t', '\n', '\f', '\r', ' ':
// Whitespace.
return s.emitToken(TokenS, matchers[TokenS].FindString(input))
case '.':
// Dot is too common to not have a quick check.
// We'll test if this is a Char; if it is followed by a number it is a
// dimension/percentage/number, and this will be matched later.
if len(input) > 1 && !unicode.IsDigit(rune(input[1])) {
return s.emitSimple(TokenChar, ".")
}
case '#':
// Another common one: Hash or Char.
if match := matchers[TokenHash].FindString(input); match != "" {
return s.emitToken(TokenHash, match)
}
return s.emitSimple(TokenChar, "#")
case '@':
// Another common one: AtKeyword or Char.
if match := matchers[TokenAtKeyword].FindString(input); match != "" {
return s.emitSimple(TokenAtKeyword, match)
}
return s.emitSimple(TokenChar, "@")
case ':', ',', ';', '%', '&', '+', '=', '>', '(', ')', '[', ']', '{', '}':
// More common chars.
return s.emitSimple(TokenChar, string(input[0]))
case '"', '\'':
// String or error.
match := matchers[TokenString].FindString(input)
if match != "" {
return s.emitToken(TokenString, match)
}
s.err = &Token{TokenError, "unclosed quotation mark", s.row, s.col}
return s.err
case '/':
// Comment, error or Char.
if len(input) > 1 && input[1] == '*' {
match := matchers[TokenComment].FindString(input)
if match != "" {
return s.emitToken(TokenComment, match)
} else {
s.err = &Token{TokenError, "unclosed comment", s.row, s.col}
return s.err
}
}
return s.emitSimple(TokenChar, "/")
case '~':
// Includes or Char.
return s.emitPrefixOrChar(TokenIncludes, "~=")
case '|':
// DashMatch or Char.
return s.emitPrefixOrChar(TokenDashMatch, "|=")
case '^':
// PrefixMatch or Char.
return s.emitPrefixOrChar(TokenPrefixMatch, "^=")
case '$':
// SuffixMatch or Char.
return s.emitPrefixOrChar(TokenSuffixMatch, "$=")
case '*':
// SubstringMatch or Char.
return s.emitPrefixOrChar(TokenSubstringMatch, "*=")
case '<':
// CDO or Char.
return s.emitPrefixOrChar(TokenCDO, "<!--")
}
// Test all regexps, in order.
for _, token := range matchOrder {
if match := matchers[token].FindString(input); match != "" {
return s.emitToken(token, match)
}
}
// We already handled unclosed quotation marks and comments,
// so this can only be a Char.
r, width := utf8.DecodeRuneInString(input)
token := &Token{TokenChar, string(r), s.row, s.col}
s.col += width
s.pos += width
return token
}
// updatePosition updates input coordinates based on the consumed text.
func (s *Scanner) updatePosition(text string) {
width := utf8.RuneCountInString(text)
lines := strings.Count(text, "\n")
s.row += lines
if lines == 0 {
s.col += width
} else {
s.col = utf8.RuneCountInString(text[strings.LastIndex(text, "\n"):])
}
s.pos += len(text) // while col is a rune index, pos is a byte index
}
// emitToken returns a Token for the string v and updates the scanner position.
func (s *Scanner) emitToken(t tokenType, v string) *Token {
token := &Token{t, v, s.row, s.col}
s.updatePosition(v)
return token
}
// emitSimple returns a Token for the string v and updates the scanner
// position in a simplified manner.
//
// The string is known to have only ASCII characters and to not have a newline.
func (s *Scanner) emitSimple(t tokenType, v string) *Token {
token := &Token{t, v, s.row, s.col}
s.col += len(v)
s.pos += len(v)
return token
}
// emitPrefixOrChar returns a Token for type t if the current position
// matches the given prefix. Otherwise it returns a Char token using the
// first character from the prefix.
//
// The prefix is known to have only ASCII characters and to not have a newline.
func (s *Scanner) emitPrefixOrChar(t tokenType, prefix string) *Token {
if strings.HasPrefix(s.input[s.pos:], prefix) {
return s.emitSimple(t, prefix)
}
return s.emitSimple(TokenChar, string(prefix[0]))
}

View file

@ -1,22 +1,15 @@
# Compiled Object files, Static and Dynamic libs (Shared Objects)
*.o
*.a
*.so
# Folders
_obj
_test
# Architecture specific extensions/prefixes
*.[568vq]
[568vq].out
*.cgo1.go
*.cgo2.c
_cgo_defun.c
_cgo_gotypes.go
_cgo_export.*
_testmain.go
# Binaries for programs and plugins
*.exe
*.exe~
*.dll
*.so
*.dylib
# Test binary, built with `go test -c`
*.test
# Output of the go coverage tool, specifically when used with LiteIDE
*.out
# goland idea folder
*.idea

View file

@ -1,18 +1,22 @@
language: go
go:
- 1.1
- 1.2
- 1.3
- 1.4
- 1.5
- 1.6
- 1.7
- 1.2.x
- 1.3.x
- 1.4.x
- 1.5.x
- 1.6.x
- 1.7.x
- 1.8.x
- 1.9.x
- 1.10.x
- 1.11.x
- 1.12.x
- tip
matrix:
allow_failures:
- go: tip
fast_finish: true
install:
- go get golang.org/x/net/html
- go get .
script:
- go test -v ./...

View file

@ -6,6 +6,10 @@ Third-party patches are essential for keeping bluemonday secure and offering the
* Make sure you have a [Github account](https://github.com/signup/free)
## Guidelines
1. Do not vendor dependencies. As a security package, were we to vendor dependencies the projects that then vendor bluemonday may not receive the latest security updates to the dependencies. By not vendoring dependencies the project that implements bluemonday will vendor the latest version of any dependent packages. Vendoring is a project problem, not a package problem. bluemonday will be tested against the latest version of dependencies periodically and during any PR/merge.
## Submitting an Issue
* Submit a ticket for your issue, assuming one does not already exist

View file

@ -1,6 +1,7 @@
1. John Graham-Cumming http://jgc.org/
1. Mohammad Gufran https://github.com/Gufran
1. Steven Gutzwiller https://github.com/StevenGutzwiller
1. Andrew Krasichkov @buglloc https://github.com/buglloc
1. Mike Samuel mikesamuel@gmail.com
1. Dmitri Shuralyov shurcooL@gmail.com
1. https://github.com/opennota
1. https://github.com/Gufran
1. https://github.com/opennota

View file

@ -1,4 +1,4 @@
# bluemonday [![Build Status](https://travis-ci.org/microcosm-cc/bluemonday.svg?branch=master)](https://travis-ci.org/microcosm-cc/bluemonday) [![GoDoc](https://godoc.org/github.com/microcosm-cc/bluemonday?status.png)](https://godoc.org/github.com/microcosm-cc/bluemonday)
# bluemonday [![Build Status](https://travis-ci.org/microcosm-cc/bluemonday.svg?branch=master)](https://travis-ci.org/microcosm-cc/bluemonday) [![GoDoc](https://godoc.org/github.com/microcosm-cc/bluemonday?status.png)](https://godoc.org/github.com/microcosm-cc/bluemonday) [![Sourcegraph](https://sourcegraph.com/github.com/microcosm-cc/bluemonday/-/badge.svg)](https://sourcegraph.com/github.com/microcosm-cc/bluemonday?badge)
bluemonday is a HTML sanitizer implemented in Go. It is fast and highly configurable.
@ -58,10 +58,12 @@ We expect to be supplied with well-formatted HTML (closing elements for every ap
### Supported Go Versions
bluemonday is regularly tested against Go 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7 and tip.
bluemonday is tested against Go 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 1.10, 1.11, 1.12, and tip.
We do not support Go 1.0 as we depend on `golang.org/x/net/html` which includes a reference to `io.ErrNoProgress` which did not exist in Go 1.0.
We support Go 1.1 but Travis no longer tests against it.
## Is it production ready?
*Yes*
@ -87,7 +89,11 @@ import (
)
func main() {
// Do this once for each unique policy, and use the policy for the life of the program
// Policy creation/editing is not safe to use in multiple goroutines
p := bluemonday.UGCPolicy()
// The policy can then be used to sanitize lots of input and it is safe to use the policy in multiple goroutines
html := p.Sanitize(
`<a onblur="alert(secret)" href="http://www.google.com">Google</a>`,
)
@ -140,7 +146,7 @@ func main() {
We ship two default policies:
1. `bluemonday.StrictPolicy()` which can be thought of as equivalent to stripping all HTML elements and their attributes as it has nothing on it's whitelist. An example usage scenario would be blog post titles where HTML tags are not expected at all and if they are then the elements *and* the content of the elements should be stripped. This is a *very* strict policy.
1. `bluemonday.StrictPolicy()` which can be thought of as equivalent to stripping all HTML elements and their attributes as it has nothing on its whitelist. An example usage scenario would be blog post titles where HTML tags are not expected at all and if they are then the elements *and* the content of the elements should be stripped. This is a *very* strict policy.
2. `bluemonday.UGCPolicy()` which allows a broad selection of HTML elements and attributes that are safe for user generated content. Note that this policy does *not* whitelist iframes, object, embed, styles, script, etc. An example usage scenario would be blog post bodies where a variety of formatting is expected along with the potential for TABLEs and IMGs.
## Policy Building
@ -163,12 +169,26 @@ To add elements to a policy either add just the elements:
p.AllowElements("b", "strong")
```
Or using a regex:
_Note: if an element is added by name as shown above, any matching regex will be ignored_
It is also recommended to ensure multiple patterns don't overlap as order of execution is not guaranteed and can result in some rules being missed.
```go
p.AllowElementsMatching(regex.MustCompile(`^my-element-`))
```
Or add elements as a virtue of adding an attribute:
```go
// Not the recommended pattern, see the recommendation on using .Matching() below
p.AllowAttrs("nowrap").OnElements("td", "th")
```
Again, this also supports a regex pattern match alternative:
```go
p.AllowAttrs("nowrap").OnElementsMatching(regex.MustCompile(`^my-element-`))
```
Attributes can either be added to all elements:
```go
p.AllowAttrs("dir").Matching(regexp.MustCompile("(?i)rtl|ltr")).Globally()
@ -198,6 +218,49 @@ p := bluemonday.UGCPolicy()
p.AllowElements("fieldset", "select", "option")
```
### Inline CSS
Although it's possible to handle inline CSS using `AllowAttrs` with a `Matching` rule, writing a single monolithic regular expression to safely process all inline CSS which you wish to allow is not a trivial task. Instead of attempting to do so, you can whitelist the `style` attribute on whichever element(s) you desire and use style policies to control and sanitize inline styles.
It is suggested that you use `Matching` (with a suitable regular expression)
`MatchingEnum`, or `MatchingHandler` to ensure each style matches your needs,
but default handlers are supplied for most widely used styles.
Similar to attributes, you can allow specific CSS properties to be set inline:
```go
p.AllowAttrs("style").OnElements("span", "p")
// Allow the 'color' property with valid RGB(A) hex values only (on any element allowed a 'style' attribute)
p.AllowStyles("color").Matching(regexp.MustCompile("(?i)^#([0-9a-f]{3,4}|[0-9a-f]{6}|[0-9a-f]{8})$")).Globally()
```
Additionally, you can allow a CSS property to be set only to an allowed value:
```go
p.AllowAttrs("style").OnElements("span", "p")
// Allow the 'text-decoration' property to be set to 'underline', 'line-through' or 'none'
// on 'span' elements only
p.AllowStyles("text-decoration").MatchingEnum("underline", "line-through", "none").OnElements("span")
```
Or you can specify elements based on a regex patterm match:
```go
p.AllowAttrs("style").OnElementsMatching(regex.MustCompile(`^my-element-`))
// Allow the 'text-decoration' property to be set to 'underline', 'line-through' or 'none'
// on 'span' elements only
p.AllowStyles("text-decoration").MatchingEnum("underline", "line-through", "none").OnElementsMatching(regex.MustCompile(`^my-element-`))
```
If you need more specific checking, you can create a handler that takes in a string and returns a bool to
validate the values for a given property. The string parameter has been
converted to lowercase and unicode code points have been converted.
```go
myHandler := func(value string) bool{
return true
}
p.AllowAttrs("style").OnElements("span", "p")
// Allow the 'color' property with values validated by the handler (on any element allowed a 'style' attribute)
p.AllowStyles("color").MatchingHandler(myHandler).Globally()
```
### Links
Links are difficult beasts to sanitise safely and also one of the biggest attack vectors for malicious content.
@ -232,6 +295,13 @@ Regardless of whether you have enabled parseable URLs, you can force all URLs to
p.RequireNoFollowOnLinks(true)
```
Similarly, you can force all URLs to have "noreferrer" in their rel attribute.
```go
// This applies to "a" "area" "link" elements that have a "href" attribute
p.RequireNoReferrerOnLinks(true)
```
We provide a convenience method that applies all of the above, but you will still need to whitelist the linkable elements for the URL rules to be applied to:
```go
p.AllowStandardURLs()
@ -273,7 +343,7 @@ We also bundle some helpers to simplify policy building:
// Permits the "dir", "id", "lang", "title" attributes globally
p.AllowStandardAttributes()
// Permits the "img" element and it's standard attributes
// Permits the "img" element and its standard attributes
p.AllowImages()
// Permits ordered and unordered lists, and also definition lists
@ -312,7 +382,6 @@ It is not the job of bluemonday to fix your bad HTML, it is merely the job of bl
## TODO
* Add support for CSS sanitisation to allow some CSS properties based on a whitelist, possibly using the [Gorilla CSS3 scanner](http://www.gorillatoolkit.org/pkg/css/scanner)
* Investigate whether devs want to blacklist elements and attributes. This would allow devs to take an existing policy (such as the `bluemonday.UGCPolicy()` ) that encapsulates 90% of what they're looking for but does more than they need, and to remove the extra things they do not want to make it 100% what they want
* Investigate whether devs want a validating HTML mode, in which the HTML elements are not just transformed into a balanced tree (every start tag has a closing tag at the correct depth) but also that elements and character data appear only in their allowed context (i.e. that a `table` element isn't a descendent of a `caption`, that `colgroup`, `thead`, `tbody`, `tfoot` and `tr` are permitted, and that character data is not permitted)

View file

@ -84,7 +84,7 @@ bluemonday is heavily inspired by both the OWASP Java HTML Sanitizer
We ship two default policies, one is bluemonday.StrictPolicy() and can be
thought of as equivalent to stripping all HTML elements and their attributes as
it has nothing on it's whitelist.
it has nothing on its whitelist.
The other is bluemonday.UGCPolicy() and allows a broad selection of HTML
elements and attributes that are safe for user generated content. Note that

10
vendor/github.com/microcosm-cc/bluemonday/go.mod generated vendored Normal file
View file

@ -0,0 +1,10 @@
module github.com/microcosm-cc/bluemonday
go 1.9
require (
github.com/aymerick/douceur v0.2.0 // indirect
github.com/chris-ramon/douceur v0.2.0
github.com/gorilla/css v1.0.0 // indirect
golang.org/x/net v0.0.0-20181220203305-927f97764cc3
)

8
vendor/github.com/microcosm-cc/bluemonday/go.sum generated vendored Normal file
View file

@ -0,0 +1,8 @@
github.com/aymerick/douceur v0.2.0 h1:Mv+mAeH1Q+n9Fr+oyamOlAkUNPWPlA8PPGR0QAaYuPk=
github.com/aymerick/douceur v0.2.0/go.mod h1:wlT5vV2O3h55X9m7iVYN0TBM0NH/MmbLnd30/FjWUq4=
github.com/chris-ramon/douceur v0.2.0 h1:IDMEdxlEUUBYBKE4z/mJnFyVXox+MjuEVDJNN27glkU=
github.com/chris-ramon/douceur v0.2.0/go.mod h1:wDW5xjJdeoMm1mRt4sD4c/LbF/mWdEpRXQKjTR8nIBE=
github.com/gorilla/css v1.0.0 h1:BQqNyPTi50JCFMTw/b67hByjMVXZRwGha6wxVGkeihY=
github.com/gorilla/css v1.0.0/go.mod h1:Dn721qIggHpt4+EFCcTLTU/vk5ySda2ReITrtgBl60c=
golang.org/x/net v0.0.0-20181220203305-927f97764cc3 h1:eH6Eip3UpmR+yM/qI9Ijluzb1bNv/cAU/n+6l8tRSis=
golang.org/x/net v0.0.0-20181220203305-927f97764cc3/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=

2085
vendor/github.com/microcosm-cc/bluemonday/handlers.go generated vendored Normal file

File diff suppressed because it is too large Load diff

View file

@ -135,7 +135,7 @@ func (p *Policy) AllowStandardURLs() {
// Most common URL schemes only
p.AllowURLSchemes("mailto", "http", "https")
// For all anchors we will add rel="nofollow" if it does not already exist
// For linking elements we will add rel="nofollow" if it does not already exist
// This applies to "a" "area" "link"
p.RequireNoFollowOnLinks(true)
}

View file

@ -29,6 +29,8 @@
package bluemonday
//TODO sgutzwiller create map of styles to default handlers
//TODO sgutzwiller create handlers for various attributes
import (
"net/url"
"regexp"
@ -47,21 +49,26 @@ type Policy struct {
// exceptions
initialized bool
// Allows the <!DOCTYPE > tag to exist in the sanitized document
allowDocType bool
// If true then we add spaces when stripping tags, specifically the closing
// tag is replaced by a space character.
addSpaces bool
// When true, add rel="nofollow" to HTML anchors
// When true, add rel="nofollow" to HTML a, area, and link tags
requireNoFollow bool
// When true, add rel="nofollow" to HTML anchors
// When true, add rel="nofollow" to HTML a, area, and link tags
// Will add for href="http://foo"
// Will skip for href="/foo" or href="foo"
requireNoFollowFullyQualifiedLinks bool
// When true, add rel="noreferrer" to HTML a, area, and link tags
requireNoReferrer bool
// When true, add rel="noreferrer" to HTML a, area, and link tags
// Will add for href="http://foo"
// Will skip for href="/foo" or href="foo"
requireNoReferrerFullyQualifiedLinks bool
// When true add target="_blank" to fully qualified links
// Will add for href="http://foo"
// Will skip for href="/foo" or href="foo"
@ -73,12 +80,27 @@ type Policy struct {
// When true, u, _ := url.Parse("url"); !u.IsAbs() is permitted
allowRelativeURLs bool
// When true, allow data attributes.
allowDataAttributes bool
// map[htmlElementName]map[htmlAttributeName]attrPolicy
elsAndAttrs map[string]map[string]attrPolicy
// elsMatchingAndAttrs stores regex based element matches along with attributes
elsMatchingAndAttrs map[*regexp.Regexp]map[string]attrPolicy
// map[htmlAttributeName]attrPolicy
globalAttrs map[string]attrPolicy
// map[htmlElementName]map[cssPropertyName]stylePolicy
elsAndStyles map[string]map[string]stylePolicy
// map[regex]map[cssPropertyName]stylePolicy
elsMatchingAndStyles map[*regexp.Regexp]map[string]stylePolicy
// map[cssPropertyName]stylePolicy
globalStyles map[string]stylePolicy
// If urlPolicy is nil, all URLs with matching schema are allowed.
// Otherwise, only the URLs with matching schema and urlPolicy(url)
// returning true are allowed.
@ -93,6 +115,16 @@ type Policy struct {
// be maintained in the output HTML.
setOfElementsAllowedWithoutAttrs map[string]struct{}
// If an element has had all attributes removed as a result of a policy
// being applied, then the element would be removed from the output.
//
// However some elements are valid and have strong layout meaning without
// any attributes, i.e. <table>.
//
// In this case, any element matching a regular expression will be accepted without
// attributes added.
setOfElementsMatchingAllowedWithoutAttrs []*regexp.Regexp
setOfElementsToSkipContent map[string]struct{}
}
@ -103,6 +135,20 @@ type attrPolicy struct {
regexp *regexp.Regexp
}
type stylePolicy struct {
// handler to validate
handler func(string) bool
// optional pattern to match, when not nil the regexp needs to match
// otherwise the property is removed
regexp *regexp.Regexp
// optional list of allowed property values, for properties which
// have a defined list of allowed values; property will be removed
// if the value is not allowed
enum []string
}
type attrPolicyBuilder struct {
p *Policy
@ -111,13 +157,26 @@ type attrPolicyBuilder struct {
allowEmpty bool
}
type stylePolicyBuilder struct {
p *Policy
propertyNames []string
regexp *regexp.Regexp
enum []string
handler func(string) bool
}
type urlPolicy func(url *url.URL) (allowUrl bool)
// init initializes the maps if this has not been done already
func (p *Policy) init() {
if !p.initialized {
p.elsAndAttrs = make(map[string]map[string]attrPolicy)
p.elsMatchingAndAttrs = make(map[*regexp.Regexp]map[string]attrPolicy)
p.globalAttrs = make(map[string]attrPolicy)
p.elsAndStyles = make(map[string]map[string]stylePolicy)
p.elsMatchingAndStyles = make(map[*regexp.Regexp]map[string]stylePolicy)
p.globalStyles = make(map[string]stylePolicy)
p.allowURLSchemes = make(map[string]urlPolicy)
p.setOfElementsAllowedWithoutAttrs = make(map[string]struct{})
p.setOfElementsToSkipContent = make(map[string]struct{})
@ -161,6 +220,21 @@ func (p *Policy) AllowAttrs(attrNames ...string) *attrPolicyBuilder {
return &abp
}
// AllowDataAttributes whitelists all data attributes. We can't specify the name
// of each attribute exactly as they are customized.
//
// NOTE: These values are not sanitized and applications that evaluate or process
// them without checking and verification of the input may be at risk if this option
// is enabled. This is a 'caveat emptor' option and the person enabling this option
// needs to fully understand the potential impact with regards to whatever application
// will be consuming the sanitized HTML afterwards, i.e. if you know you put a link in a
// data attribute and use that to automatically load some new window then you're giving
// the author of a HTML fragment the means to open a malicious destination automatically.
// Use with care!
func (p *Policy) AllowDataAttributes() {
p.allowDataAttributes = true
}
// AllowNoAttrs says that attributes on element are optional.
//
// The attribute policy is only added to the core policy when OnElements(...)
@ -230,6 +304,30 @@ func (abp *attrPolicyBuilder) OnElements(elements ...string) *Policy {
return abp.p
}
// OnElementsMatching will bind an attribute policy to all elements matching a given regex
// and return the updated policy
func (abp *attrPolicyBuilder) OnElementsMatching(regex *regexp.Regexp) *Policy {
for _, attr := range abp.attrNames {
if _, ok := abp.p.elsMatchingAndAttrs[regex]; !ok {
abp.p.elsMatchingAndAttrs[regex] = make(map[string]attrPolicy)
}
ap := attrPolicy{}
if abp.regexp != nil {
ap.regexp = abp.regexp
}
abp.p.elsMatchingAndAttrs[regex][attr] = ap
}
if abp.allowEmpty {
abp.p.setOfElementsMatchingAllowedWithoutAttrs = append(abp.p.setOfElementsMatchingAllowedWithoutAttrs, regex)
if _, ok := abp.p.elsMatchingAndAttrs[regex]; !ok {
abp.p.elsMatchingAndAttrs[regex] = make(map[string]attrPolicy)
}
}
return abp.p
}
// Globally will bind an attribute policy to all HTML elements and return the
// updated policy
func (abp *attrPolicyBuilder) Globally() *Policy {
@ -250,6 +348,139 @@ func (abp *attrPolicyBuilder) Globally() *Policy {
return abp.p
}
// AllowStyles takes a range of CSS property names and returns a
// style policy builder that allows you to specify the pattern and scope of
// the whitelisted property.
//
// The style policy is only added to the core policy when either Globally()
// or OnElements(...) are called.
func (p *Policy) AllowStyles(propertyNames ...string) *stylePolicyBuilder {
p.init()
abp := stylePolicyBuilder{
p: p,
}
for _, propertyName := range propertyNames {
abp.propertyNames = append(abp.propertyNames, strings.ToLower(propertyName))
}
return &abp
}
// Matching allows a regular expression to be applied to a nascent style
// policy, and returns the style policy. Calling this more than once will
// replace the existing regexp.
func (spb *stylePolicyBuilder) Matching(regex *regexp.Regexp) *stylePolicyBuilder {
spb.regexp = regex
return spb
}
// MatchingEnum allows a list of allowed values to be applied to a nascent style
// policy, and returns the style policy. Calling this more than once will
// replace the existing list of allowed values.
func (spb *stylePolicyBuilder) MatchingEnum(enum ...string) *stylePolicyBuilder {
spb.enum = enum
return spb
}
// MatchingHandler allows a handler to be applied to a nascent style
// policy, and returns the style policy. Calling this more than once will
// replace the existing handler.
func (spb *stylePolicyBuilder) MatchingHandler(handler func(string) bool) *stylePolicyBuilder {
spb.handler = handler
return spb
}
// OnElements will bind a style policy to a given range of HTML elements
// and return the updated policy
func (spb *stylePolicyBuilder) OnElements(elements ...string) *Policy {
for _, element := range elements {
element = strings.ToLower(element)
for _, attr := range spb.propertyNames {
if _, ok := spb.p.elsAndStyles[element]; !ok {
spb.p.elsAndStyles[element] = make(map[string]stylePolicy)
}
sp := stylePolicy{}
if spb.handler != nil {
sp.handler = spb.handler
} else if len(spb.enum) > 0 {
sp.enum = spb.enum
} else if spb.regexp != nil {
sp.regexp = spb.regexp
} else {
sp.handler = getDefaultHandler(attr)
}
spb.p.elsAndStyles[element][attr] = sp
}
}
return spb.p
}
// OnElementsMatching will bind a style policy to any HTML elements matching the pattern
// and return the updated policy
func (spb *stylePolicyBuilder) OnElementsMatching(regex *regexp.Regexp) *Policy {
for _, attr := range spb.propertyNames {
if _, ok := spb.p.elsMatchingAndStyles[regex]; !ok {
spb.p.elsMatchingAndStyles[regex] = make(map[string]stylePolicy)
}
sp := stylePolicy{}
if spb.handler != nil {
sp.handler = spb.handler
} else if len(spb.enum) > 0 {
sp.enum = spb.enum
} else if spb.regexp != nil {
sp.regexp = spb.regexp
} else {
sp.handler = getDefaultHandler(attr)
}
spb.p.elsMatchingAndStyles[regex][attr] = sp
}
return spb.p
}
// Globally will bind a style policy to all HTML elements and return the
// updated policy
func (spb *stylePolicyBuilder) Globally() *Policy {
for _, attr := range spb.propertyNames {
if _, ok := spb.p.globalStyles[attr]; !ok {
spb.p.globalStyles[attr] = stylePolicy{}
}
// Use only one strategy for validating styles, fallback to default
sp := stylePolicy{}
if spb.handler != nil {
sp.handler = spb.handler
} else if len(spb.enum) > 0 {
sp.enum = spb.enum
} else if spb.regexp != nil {
sp.regexp = spb.regexp
} else {
sp.handler = getDefaultHandler(attr)
}
spb.p.globalStyles[attr] = sp
}
return spb.p
}
// AllowElements will append HTML elements to the whitelist without applying an
// attribute policy to those elements (the elements are permitted
// sans-attributes)
@ -267,8 +498,16 @@ func (p *Policy) AllowElements(names ...string) *Policy {
return p
}
// RequireNoFollowOnLinks will result in all <a> tags having a rel="nofollow"
// added to them if one does not already exist
func (p *Policy) AllowElementsMatching(regex *regexp.Regexp) *Policy {
p.init()
if _, ok := p.elsMatchingAndAttrs[regex]; !ok {
p.elsMatchingAndAttrs[regex] = make(map[string]attrPolicy)
}
return p
}
// RequireNoFollowOnLinks will result in all a, area, link tags having a
// rel="nofollow"added to them if one does not already exist
//
// Note: This requires p.RequireParseableURLs(true) and will enable it.
func (p *Policy) RequireNoFollowOnLinks(require bool) *Policy {
@ -279,9 +518,10 @@ func (p *Policy) RequireNoFollowOnLinks(require bool) *Policy {
return p
}
// RequireNoFollowOnFullyQualifiedLinks will result in all <a> tags that point
// to a non-local destination (i.e. starts with a protocol and has a host)
// having a rel="nofollow" added to them if one does not already exist
// RequireNoFollowOnFullyQualifiedLinks will result in all a, area, and link
// tags that point to a non-local destination (i.e. starts with a protocol and
// has a host) having a rel="nofollow" added to them if one does not already
// exist
//
// Note: This requires p.RequireParseableURLs(true) and will enable it.
func (p *Policy) RequireNoFollowOnFullyQualifiedLinks(require bool) *Policy {
@ -292,9 +532,35 @@ func (p *Policy) RequireNoFollowOnFullyQualifiedLinks(require bool) *Policy {
return p
}
// AddTargetBlankToFullyQualifiedLinks will result in all <a> tags that point
// to a non-local destination (i.e. starts with a protocol and has a host)
// having a target="_blank" added to them if one does not already exist
// RequireNoReferrerOnLinks will result in all a, area, and link tags having a
// rel="noreferrrer" added to them if one does not already exist
//
// Note: This requires p.RequireParseableURLs(true) and will enable it.
func (p *Policy) RequireNoReferrerOnLinks(require bool) *Policy {
p.requireNoReferrer = require
p.requireParseableURLs = true
return p
}
// RequireNoReferrerOnFullyQualifiedLinks will result in all a, area, and link
// tags that point to a non-local destination (i.e. starts with a protocol and
// has a host) having a rel="noreferrer" added to them if one does not already
// exist
//
// Note: This requires p.RequireParseableURLs(true) and will enable it.
func (p *Policy) RequireNoReferrerOnFullyQualifiedLinks(require bool) *Policy {
p.requireNoReferrerFullyQualifiedLinks = require
p.requireParseableURLs = true
return p
}
// AddTargetBlankToFullyQualifiedLinks will result in all a, area and link tags
// that point to a non-local destination (i.e. starts with a protocol and has a
// host) having a target="_blank" added to them if one does not already exist
//
// Note: This requires p.RequireParseableURLs(true) and will enable it.
func (p *Policy) AddTargetBlankToFullyQualifiedLinks(require bool) *Policy {
@ -369,21 +635,6 @@ func (p *Policy) AllowURLSchemeWithCustomPolicy(
return p
}
// AllowDocType states whether the HTML sanitised by the sanitizer is allowed to
// contain the HTML DocType tag: <!DOCTYPE HTML> or one of it's variants.
//
// The HTML spec only permits one doctype per document, and as you know how you
// are using the output of this, you know best as to whether we should ignore it
// (default) or not.
//
// If you are sanitizing a HTML fragment the default (false) is fine.
func (p *Policy) AllowDocType(allow bool) *Policy {
p.allowDocType = allow
return p
}
// AddSpaceWhenStrippingTag states whether to add a single space " " when
// removing tags that are not whitelisted by the policy.
//
@ -402,7 +653,7 @@ func (p *Policy) AddSpaceWhenStrippingTag(allow bool) *Policy {
}
// SkipElementsContent adds the HTML elements whose tags is needed to be removed
// with it's content.
// with its content.
func (p *Policy) SkipElementsContent(names ...string) *Policy {
p.init()
@ -440,6 +691,7 @@ func (p *Policy) addDefaultElementsWithoutAttrs() {
p.setOfElementsAllowedWithoutAttrs["abbr"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["acronym"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["address"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["article"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["aside"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["audio"] = struct{}{}
@ -451,6 +703,7 @@ func (p *Policy) addDefaultElementsWithoutAttrs() {
p.setOfElementsAllowedWithoutAttrs["button"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["canvas"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["caption"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["center"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["cite"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["code"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["col"] = struct{}{}
@ -484,6 +737,7 @@ func (p *Policy) addDefaultElementsWithoutAttrs() {
p.setOfElementsAllowedWithoutAttrs["kbd"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["li"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["mark"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["marquee"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["nav"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["ol"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["optgroup"] = struct{}{}
@ -496,6 +750,7 @@ func (p *Policy) addDefaultElementsWithoutAttrs() {
p.setOfElementsAllowedWithoutAttrs["ruby"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["s"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["samp"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["script"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["section"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["select"] = struct{}{}
p.setOfElementsAllowedWithoutAttrs["small"] = struct{}{}

View file

@ -33,9 +33,20 @@ import (
"bytes"
"io"
"net/url"
"regexp"
"strconv"
"strings"
"golang.org/x/net/html"
cssparser "github.com/chris-ramon/douceur/parser"
)
var (
dataAttribute = regexp.MustCompile("^data-.+")
dataAttributeXMLPrefix = regexp.MustCompile("^xml.+")
dataAttributeInvalidChars = regexp.MustCompile("[A-Z;]+")
cssUnicodeChar = regexp.MustCompile(`\\[0-9a-f]{1,6} ?`)
)
// Sanitize takes a string that contains a HTML fragment or document and applies
@ -75,6 +86,98 @@ func (p *Policy) SanitizeReader(r io.Reader) *bytes.Buffer {
return p.sanitize(r)
}
const escapedURLChars = "'<>\"\r"
func escapeUrlComponent(val string) string {
w := bytes.NewBufferString("")
i := strings.IndexAny(val, escapedURLChars)
for i != -1 {
if _, err := w.WriteString(val[:i]); err != nil {
return w.String()
}
var esc string
switch val[i] {
case '\'':
// "&#39;" is shorter than "&apos;" and apos was not in HTML until HTML5.
esc = "&#39;"
case '<':
esc = "&lt;"
case '>':
esc = "&gt;"
case '"':
// "&#34;" is shorter than "&quot;".
esc = "&#34;"
case '\r':
esc = "&#13;"
default:
panic("unrecognized escape character")
}
val = val[i+1:]
if _, err := w.WriteString(esc); err != nil {
return w.String()
}
i = strings.IndexAny(val, escapedURLChars)
}
w.WriteString(val)
return w.String()
}
func sanitizedUrl(val string) (string, error) {
u, err := url.Parse(val)
if err != nil {
return "", err
}
// sanitize the url query params
sanitizedQueryValues := make(url.Values, 0)
queryValues := u.Query()
for k, vals := range queryValues {
sk := html.EscapeString(k)
for _, v := range vals {
sv := escapeUrlComponent(v)
sanitizedQueryValues.Set(sk, sv)
}
}
u.RawQuery = sanitizedQueryValues.Encode()
// u.String() will also sanitize host/scheme/user/pass
return u.String(), nil
}
func (p *Policy) writeLinkableBuf(buff *bytes.Buffer, token *html.Token) {
// do not escape multiple query parameters
tokenBuff := bytes.NewBufferString("")
tokenBuff.WriteString("<")
tokenBuff.WriteString(token.Data)
for _, attr := range token.Attr {
tokenBuff.WriteByte(' ')
tokenBuff.WriteString(attr.Key)
tokenBuff.WriteString(`="`)
switch attr.Key {
case "href", "src":
u, ok := p.validURL(attr.Val)
if !ok {
tokenBuff.WriteString(html.EscapeString(attr.Val))
continue
}
u, err := sanitizedUrl(u)
if err == nil {
tokenBuff.WriteString(u)
} else {
// fallthrough
tokenBuff.WriteString(html.EscapeString(attr.Val))
}
default:
// re-apply
tokenBuff.WriteString(html.EscapeString(attr.Val))
}
tokenBuff.WriteByte('"')
}
if token.Type == html.SelfClosingTagToken {
tokenBuff.WriteString("/")
}
tokenBuff.WriteString(">")
buff.WriteString(tokenBuff.String())
}
// Performs the actual sanitization process.
func (p *Policy) sanitize(r io.Reader) *bytes.Buffer {
@ -112,9 +215,13 @@ func (p *Policy) sanitize(r io.Reader) *bytes.Buffer {
switch token.Type {
case html.DoctypeToken:
if p.allowDocType {
buff.WriteString(token.String())
}
// DocType is not handled as there is no safe parsing mechanism
// provided by golang.org/x/net/html for the content, and this can
// be misused to insert HTML tags that are not then sanitized
//
// One might wish to recursively sanitize here using the same policy
// but I will need to do some further testing before considering
// this.
case html.CommentToken:
@ -122,20 +229,23 @@ func (p *Policy) sanitize(r io.Reader) *bytes.Buffer {
case html.StartTagToken:
mostRecentlyStartedToken = token.Data
mostRecentlyStartedToken = strings.ToLower(token.Data)
aps, ok := p.elsAndAttrs[token.Data]
if !ok {
if _, ok := p.setOfElementsToSkipContent[token.Data]; ok {
skipElementContent = true
skippingElementsCount++
aa, matched := p.matchRegex(token.Data)
if !matched {
if _, ok := p.setOfElementsToSkipContent[token.Data]; ok {
skipElementContent = true
skippingElementsCount++
}
if p.addSpaces {
buff.WriteString(" ")
}
break
}
if p.addSpaces {
buff.WriteString(" ")
}
break
aps = aa
}
if len(token.Attr) != 0 {
token.Attr = p.sanitizeAttrs(token.Data, token.Attr, aps)
}
@ -152,11 +262,20 @@ func (p *Policy) sanitize(r io.Reader) *bytes.Buffer {
}
if !skipElementContent {
buff.WriteString(token.String())
// do not escape multiple query parameters
if linkable(token.Data) {
p.writeLinkableBuf(&buff, &token)
} else {
buff.WriteString(token.String())
}
}
case html.EndTagToken:
if mostRecentlyStartedToken == strings.ToLower(token.Data) {
mostRecentlyStartedToken = ""
}
if skipClosingTag && closingTagToSkipStack[len(closingTagToSkipStack)-1] == token.Data {
closingTagToSkipStack = closingTagToSkipStack[:len(closingTagToSkipStack)-1]
if len(closingTagToSkipStack) == 0 {
@ -167,18 +286,27 @@ func (p *Policy) sanitize(r io.Reader) *bytes.Buffer {
}
break
}
if _, ok := p.elsAndAttrs[token.Data]; !ok {
if _, ok := p.setOfElementsToSkipContent[token.Data]; ok {
match := false
for regex := range p.elsMatchingAndAttrs {
if regex.MatchString(token.Data) {
skipElementContent = false
match = true
break
}
}
if _, ok := p.setOfElementsToSkipContent[token.Data]; ok && !match {
skippingElementsCount--
if skippingElementsCount == 0 {
skipElementContent = false
}
}
if p.addSpaces {
buff.WriteString(" ")
if !match {
if p.addSpaces {
buff.WriteString(" ")
}
break
}
break
}
if !skipElementContent {
@ -189,10 +317,14 @@ func (p *Policy) sanitize(r io.Reader) *bytes.Buffer {
aps, ok := p.elsAndAttrs[token.Data]
if !ok {
if p.addSpaces {
buff.WriteString(" ")
aa, matched := p.matchRegex(token.Data)
if !matched {
if p.addSpaces && !matched {
buff.WriteString(" ")
}
break
}
break
aps = aa
}
if len(token.Attr) != 0 {
@ -202,19 +334,23 @@ func (p *Policy) sanitize(r io.Reader) *bytes.Buffer {
if len(token.Attr) == 0 && !p.allowNoAttrs(token.Data) {
if p.addSpaces {
buff.WriteString(" ")
break
}
break
}
if !skipElementContent {
buff.WriteString(token.String())
// do not escape multiple query parameters
if linkable(token.Data) {
p.writeLinkableBuf(&buff, &token)
} else {
buff.WriteString(token.String())
}
}
case html.TextToken:
if !skipElementContent {
switch strings.ToLower(mostRecentlyStartedToken) {
case "javascript":
switch mostRecentlyStartedToken {
case "script":
// not encouraged, but if a policy allows JavaScript we
// should not HTML escape it as that would break the output
buff.WriteString(token.Data)
@ -248,10 +384,47 @@ func (p *Policy) sanitizeAttrs(
return attrs
}
hasStylePolicies := false
sps, elementHasStylePolicies := p.elsAndStyles[elementName]
if len(p.globalStyles) > 0 || (elementHasStylePolicies && len(sps) > 0) {
hasStylePolicies = true
}
// no specific element policy found, look for a pattern match
if !hasStylePolicies{
for k, v := range p.elsMatchingAndStyles{
if k.MatchString(elementName) {
if len(v) > 0{
hasStylePolicies = true
break
}
}
}
}
// Builds a new attribute slice based on the whether the attribute has been
// whitelisted explicitly or globally.
cleanAttrs := []html.Attribute{}
for _, htmlAttr := range attrs {
if p.allowDataAttributes {
// If we see a data attribute, let it through.
if isDataAttribute(htmlAttr.Key) {
cleanAttrs = append(cleanAttrs, htmlAttr)
continue
}
}
// Is this a "style" attribute, and if so, do we need to sanitize it?
if htmlAttr.Key == "style" && hasStylePolicies {
htmlAttr = p.sanitizeStyles(htmlAttr, elementName)
if htmlAttr.Val == "" {
// We've sanitized away any and all styles; don't bother to
// output the style attribute (even if it's allowed)
continue
} else {
cleanAttrs = append(cleanAttrs, htmlAttr)
continue
}
}
// Is there an element specific attribute policy that applies?
if ap, ok := aps[htmlAttr.Key]; ok {
if ap.regexp != nil {
@ -267,6 +440,7 @@ func (p *Policy) sanitizeAttrs(
// Is there a global attribute policy that applies?
if ap, ok := p.globalAttrs[htmlAttr.Key]; ok {
if ap.regexp != nil {
if ap.regexp.MatchString(htmlAttr.Val) {
cleanAttrs = append(cleanAttrs, htmlAttr)
@ -332,6 +506,8 @@ func (p *Policy) sanitizeAttrs(
if (p.requireNoFollow ||
p.requireNoFollowFullyQualifiedLinks ||
p.requireNoReferrer ||
p.requireNoReferrerFullyQualifiedLinks ||
p.addTargetBlankToFullyQualifiedLinks) &&
len(cleanAttrs) > 0 {
@ -359,12 +535,16 @@ func (p *Policy) sanitizeAttrs(
if hrefFound {
var (
noFollowFound bool
noReferrerFound bool
targetBlankFound bool
)
addNoFollow := (p.requireNoFollow ||
externalLink && p.requireNoFollowFullyQualifiedLinks)
addNoReferrer := (p.requireNoReferrer ||
externalLink && p.requireNoReferrerFullyQualifiedLinks)
addTargetBlank := (externalLink &&
p.addTargetBlankToFullyQualifiedLinks)
@ -372,18 +552,18 @@ func (p *Policy) sanitizeAttrs(
for _, htmlAttr := range cleanAttrs {
var appended bool
if htmlAttr.Key == "rel" && addNoFollow {
if htmlAttr.Key == "rel" && (addNoFollow || addNoReferrer) {
if strings.Contains(htmlAttr.Val, "nofollow") {
noFollowFound = true
tmpAttrs = append(tmpAttrs, htmlAttr)
appended = true
} else {
if addNoFollow && !strings.Contains(htmlAttr.Val, "nofollow") {
htmlAttr.Val += " nofollow"
noFollowFound = true
tmpAttrs = append(tmpAttrs, htmlAttr)
appended = true
}
if addNoReferrer && !strings.Contains(htmlAttr.Val, "noreferrer") {
htmlAttr.Val += " noreferrer"
}
noFollowFound = addNoFollow
noReferrerFound = addNoReferrer
tmpAttrs = append(tmpAttrs, htmlAttr)
appended = true
}
if elementName == "a" && htmlAttr.Key == "target" {
@ -402,14 +582,22 @@ func (p *Policy) sanitizeAttrs(
tmpAttrs = append(tmpAttrs, htmlAttr)
}
}
if noFollowFound || targetBlankFound {
if noFollowFound || noReferrerFound || targetBlankFound {
cleanAttrs = tmpAttrs
}
if addNoFollow && !noFollowFound {
if (addNoFollow && !noFollowFound) || (addNoReferrer && !noReferrerFound) {
rel := html.Attribute{}
rel.Key = "rel"
rel.Val = "nofollow"
if addNoFollow {
rel.Val = "nofollow"
}
if addNoReferrer {
if rel.Val != "" {
rel.Val += " "
}
rel.Val += "noreferrer"
}
cleanAttrs = append(cleanAttrs, rel)
}
@ -479,20 +667,112 @@ func (p *Policy) sanitizeAttrs(
return cleanAttrs
}
func (p *Policy) sanitizeStyles(attr html.Attribute, elementName string) html.Attribute {
sps := p.elsAndStyles[elementName]
if len(sps) == 0{
sps = map[string]stylePolicy{}
// check for any matching elements, if we don't already have a policy found
// if multiple matches are found they will be overwritten, it's best
// to not have overlapping matchers
for regex, policies :=range p.elsMatchingAndStyles{
if regex.MatchString(elementName){
for k, v := range policies{
sps[k] = v
}
}
}
}
//Add semi-colon to end to fix parsing issue
if len(attr.Val) > 0 && attr.Val[len(attr.Val)-1] != ';' {
attr.Val = attr.Val + ";"
}
decs, err := cssparser.ParseDeclarations(attr.Val)
if err != nil {
attr.Val = ""
return attr
}
clean := []string{}
prefixes := []string{"-webkit-", "-moz-", "-ms-", "-o-", "mso-", "-xv-", "-atsc-", "-wap-", "-khtml-", "prince-", "-ah-", "-hp-", "-ro-", "-rim-", "-tc-"}
for _, dec := range decs {
addedProperty := false
tempProperty := strings.ToLower(dec.Property)
tempValue := removeUnicode(strings.ToLower(dec.Value))
for _, i := range prefixes {
tempProperty = strings.TrimPrefix(tempProperty, i)
}
if sp, ok := sps[tempProperty]; ok {
if sp.handler != nil {
if sp.handler(tempValue) {
clean = append(clean, dec.Property+": "+dec.Value)
addedProperty = true
}
} else if len(sp.enum) > 0 {
if stringInSlice(tempValue, sp.enum) {
clean = append(clean, dec.Property+": "+dec.Value)
addedProperty = true
}
} else if sp.regexp != nil {
if sp.regexp.MatchString(tempValue) {
clean = append(clean, dec.Property+": "+dec.Value)
addedProperty = true
}
continue
}
}
if sp, ok := p.globalStyles[tempProperty]; ok && !addedProperty {
if sp.handler != nil {
if sp.handler(tempValue) {
clean = append(clean, dec.Property+": "+dec.Value)
}
} else if len(sp.enum) > 0 {
if stringInSlice(tempValue, sp.enum) {
clean = append(clean, dec.Property+": "+dec.Value)
}
} else if sp.regexp != nil {
if sp.regexp.MatchString(tempValue) {
clean = append(clean, dec.Property+": "+dec.Value)
}
continue
}
}
}
if len(clean) > 0 {
attr.Val = strings.Join(clean, "; ")
} else {
attr.Val = ""
}
return attr
}
func (p *Policy) allowNoAttrs(elementName string) bool {
_, ok := p.setOfElementsAllowedWithoutAttrs[elementName]
if !ok {
for _, r := range p.setOfElementsMatchingAllowedWithoutAttrs {
if r.MatchString(elementName) {
ok = true
break
}
}
}
return ok
}
func (p *Policy) validURL(rawurl string) (string, bool) {
if p.requireParseableURLs {
// URLs do not contain whitespace
if strings.Contains(rawurl, " ") ||
// URLs are valid if when space is trimmed the URL is valid
rawurl = strings.TrimSpace(rawurl)
// URLs cannot contain whitespace, unless it is a data-uri
if (strings.Contains(rawurl, " ") ||
strings.Contains(rawurl, "\t") ||
strings.Contains(rawurl, "\n") {
strings.Contains(rawurl, "\n")) &&
!strings.HasPrefix(rawurl, `data:`) {
return "", false
}
// URLs are valid if they parse
u, err := url.Parse(rawurl)
if err != nil {
return "", false
@ -533,3 +813,77 @@ func linkable(elementName string) bool {
return false
}
}
// stringInSlice returns true if needle exists in haystack
func stringInSlice(needle string, haystack []string) bool {
for _, straw := range haystack {
if strings.ToLower(straw) == strings.ToLower(needle) {
return true
}
}
return false
}
func isDataAttribute(val string) bool {
if !dataAttribute.MatchString(val) {
return false
}
rest := strings.Split(val, "data-")
if len(rest) == 1 {
return false
}
// data-xml* is invalid.
if dataAttributeXMLPrefix.MatchString(rest[1]) {
return false
}
// no uppercase or semi-colons allowed.
if dataAttributeInvalidChars.MatchString(rest[1]) {
return false
}
return true
}
func removeUnicode(value string) string {
substitutedValue := value
currentLoc := cssUnicodeChar.FindStringIndex(substitutedValue)
for currentLoc != nil {
character := substitutedValue[currentLoc[0]+1 : currentLoc[1]]
character = strings.TrimSpace(character)
if len(character) < 4 {
character = strings.Repeat("0", 4-len(character)) + character
} else {
for len(character) > 4 {
if character[0] != '0' {
character = ""
break
} else {
character = character[1:]
}
}
}
character = "\\u" + character
translatedChar, err := strconv.Unquote(`"` + character + `"`)
translatedChar = strings.TrimSpace(translatedChar)
if err != nil {
return ""
}
substitutedValue = substitutedValue[0:currentLoc[0]] + translatedChar + substitutedValue[currentLoc[1]:]
currentLoc = cssUnicodeChar.FindStringIndex(substitutedValue)
}
return substitutedValue
}
func (p *Policy) matchRegex(elementName string ) (map[string]attrPolicy, bool) {
aps := make(map[string]attrPolicy, 0)
matched := false
for regex, attrs := range p.elsMatchingAndAttrs {
if regex.MatchString(elementName) {
matched = true
for k, v := range attrs {
aps[k] = v
}
}
}
return aps, matched
}