forgejo/modules/markup/sanitizer.go
Alexander Scheel ee7df7ba8c Markdown: Sanitizier Configuration (#9075)
* Support custom sanitization policy

Allowing the gitea administrator to configure sanitization policy allows
them to couple external renders and custom templates to support more
markup. In particular, the `pandoc` renderer allows generating KaTeX
annotations, wrapping them in `<span>` elements with class `math` and
either `inline` or `display` (depending on whether or not inline or
block mode was requested).

This iteration gives the administrator whitelisting powers; carefully
crafted regexes will thus let through only the desired attributes
necessary to support their custom markup.

Resolves: #9054

Signed-off-by: Alexander Scheel <alexander.m.scheel@gmail.com>

* Document new sanitization configuration

 - Adds basic documentation to app.ini.sample,
 - Adds an example to the Configuration Cheat Sheet, and
 - Adds extended information to External Renderers section.

Signed-off-by: Alexander Scheel <alexander.m.scheel@gmail.com>

* Drop extraneous length check in newMarkupSanitizer(...)

Signed-off-by: Alexander Scheel <alexander.m.scheel@gmail.com>

* Fix plural ELEMENT and ALLOW_ATTR in docs

These were left over from their initial names. Make them singular to
conform with the current expectations.

Signed-off-by: Alexander Scheel <alexander.m.scheel@gmail.com>
2019-12-07 14:49:04 -05:00

79 lines
2.4 KiB
Go

// Copyright 2017 The Gitea Authors. All rights reserved.
// Copyright 2017 The Gogs Authors. All rights reserved.
// Use of this source code is governed by a MIT-style
// license that can be found in the LICENSE file.
package markup
import (
"regexp"
"sync"
"code.gitea.io/gitea/modules/setting"
"github.com/microcosm-cc/bluemonday"
)
// Sanitizer is a protection wrapper of *bluemonday.Policy which does not allow
// any modification to the underlying policies once it's been created.
type Sanitizer struct {
policy *bluemonday.Policy
init sync.Once
}
var sanitizer = &Sanitizer{}
// NewSanitizer initializes sanitizer with allowed attributes based on settings.
// Multiple calls to this function will only create one instance of Sanitizer during
// entire application lifecycle.
func NewSanitizer() {
sanitizer.init.Do(func() {
ReplaceSanitizer()
})
}
// ReplaceSanitizer replaces the current sanitizer to account for changes in settings
func ReplaceSanitizer() {
sanitizer.policy = bluemonday.UGCPolicy()
// We only want to allow HighlightJS specific classes for code blocks
sanitizer.policy.AllowAttrs("class").Matching(regexp.MustCompile(`^language-\w+$`)).OnElements("code")
// Checkboxes
sanitizer.policy.AllowAttrs("type").Matching(regexp.MustCompile(`^checkbox$`)).OnElements("input")
sanitizer.policy.AllowAttrs("checked", "disabled").OnElements("input")
// Custom URL-Schemes
sanitizer.policy.AllowURLSchemes(setting.Markdown.CustomURLSchemes...)
// Allow keyword markup
sanitizer.policy.AllowAttrs("class").Matching(regexp.MustCompile(`^` + keywordClass + `$`)).OnElements("span")
// Allow <kbd> tags for keyboard shortcut styling
sanitizer.policy.AllowElements("kbd")
// Custom keyword markup
for _, rule := range setting.ExternalSanitizerRules {
if rule.Regexp != nil {
sanitizer.policy.AllowAttrs(rule.AllowAttr).Matching(rule.Regexp).OnElements(rule.Element)
} else {
sanitizer.policy.AllowAttrs(rule.AllowAttr).OnElements(rule.Element)
}
}
}
// Sanitize takes a string that contains a HTML fragment or document and applies policy whitelist.
func Sanitize(s string) string {
NewSanitizer()
return sanitizer.policy.Sanitize(s)
}
// SanitizeBytes takes a []byte slice that contains a HTML fragment or document and applies policy whitelist.
func SanitizeBytes(b []byte) []byte {
if len(b) == 0 {
// nothing to sanitize
return b
}
NewSanitizer()
return sanitizer.policy.SanitizeBytes(b)
}