This commit is contained in:
Umputun
2021-04-02 19:41:16 -05:00
parent 8ef87f400a
commit 961b8aacd9
5 changed files with 34 additions and 6 deletions
+6
View File
@@ -67,6 +67,12 @@ func TestComment_Sanitize(t *testing.T) {
out: Comment{Text: "blah blah",
Locator: Locator{URL: "/p/2021/03/23/prep-747/#remark42__comment-1b365913-7056-4920-b9ad-01304bdda085"}},
},
{
inp: Comment{Text: "<scrİpt>&lt;img src=x onerror=alert(1)&gt;",
Locator: Locator{URL: "/p/2021/03/23/prep-747/#remark42__comment-1b365913-7056-4920-b9ad-01304bdda085"}},
out: Comment{Text: "&lt;img src=x onerror=alert(1)&gt;",
Locator: Locator{URL: "/p/2021/03/23/prep-747/#remark42__comment-1b365913-7056-4920-b9ad-01304bdda085"}},
},
}
for n, tt := range tbl {
+1 -1
View File
@@ -23,7 +23,7 @@ require (
github.com/gorilla/feeds v1.1.1
github.com/hashicorp/go-multierror v1.1.0
github.com/kyokomi/emoji v2.2.1+incompatible
github.com/microcosm-cc/bluemonday v1.0.4
github.com/microcosm-cc/bluemonday v1.0.5
github.com/pkg/errors v0.9.1
github.com/rakyll/statik v0.1.7
github.com/rs/xid v1.2.1
+2
View File
@@ -179,6 +179,8 @@ github.com/mattn/go-isatty v0.0.8/go.mod h1:Iq45c/XA43vh69/j3iqttzPXn0bhXyGjM0Hd
github.com/microcosm-cc/bluemonday v1.0.2/go.mod h1:iVP4YcDBq+n/5fb23BhYFvIMq/leAFZyRl6bYmGDlGc=
github.com/microcosm-cc/bluemonday v1.0.4 h1:p0L+CTpo/PLFdkoPcJemLXG+fpMD7pYOoDEq1axMbGg=
github.com/microcosm-cc/bluemonday v1.0.4/go.mod h1:8iwZnFn2CDDNZ0r6UXhF4xawGvzaqzCRa1n3/lO3W2w=
github.com/microcosm-cc/bluemonday v1.0.5 h1:cF59UCKMmmUgqN1baLvqU/B1ZsMori+duLVTLpgiG3w=
github.com/microcosm-cc/bluemonday v1.0.5/go.mod h1:8iwZnFn2CDDNZ0r6UXhF4xawGvzaqzCRa1n3/lO3W2w=
github.com/mitchellh/mapstructure v1.1.2/go.mod h1:FVVH3fgwuzCH5S8UJGiWEs2h04kUh9fWfEaFds41c1Y=
github.com/montanaflynn/stats v0.0.0-20171201202039-1bf9dbcd8cbe/go.mod h1:wL8QJuTMNUDYhXwkmfOly8iTdp5TEcJFWZD2D7SIkUc=
github.com/moul/http2curl v1.0.0 h1:dRMWoAtb+ePxMlLkrCbAqh4TlPHXvoGUSQ323/9Zahs=
+24 -4
View File
@@ -229,7 +229,7 @@ func (p *Policy) sanitize(r io.Reader) *bytes.Buffer {
case html.StartTagToken:
mostRecentlyStartedToken = strings.ToLower(token.Data)
mostRecentlyStartedToken = normaliseElementName(token.Data)
aps, ok := p.elsAndAttrs[token.Data]
if !ok {
@@ -272,7 +272,7 @@ func (p *Policy) sanitize(r io.Reader) *bytes.Buffer {
case html.EndTagToken:
if mostRecentlyStartedToken == strings.ToLower(token.Data) {
if mostRecentlyStartedToken == normaliseElementName(token.Data) {
mostRecentlyStartedToken = ""
}
@@ -350,11 +350,11 @@ func (p *Policy) sanitize(r io.Reader) *bytes.Buffer {
if !skipElementContent {
switch mostRecentlyStartedToken {
case "script":
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)
case "style":
case `style`:
// not encouraged, but if a policy allows CSS styles we
// should not HTML escape it as that would break the output
buff.WriteString(token.Data)
@@ -887,3 +887,23 @@ func (p *Policy) matchRegex(elementName string) (map[string]attrPolicy, bool) {
}
return aps, matched
}
// normaliseElementName takes a HTML element like <script> which is user input
// and returns a lower case version of it that is immune to UTF-8 to ASCII
// conversion tricks (like the use of upper case cyrillic i scrİpt which a
// strings.ToLower would convert to script). Instead this func will preserve
// all non-ASCII as their escaped equivalent, i.e. \u0130 which reveals the
// characters when lower cased
func normaliseElementName(str string) string {
// that useful QuoteToASCII put quote marks at the start and end
// so those are trimmed off
return strings.TrimSuffix(
strings.TrimPrefix(
strings.ToLower(
strconv.QuoteToASCII(str),
),
`"`),
`"`,
)
}
+1 -1
View File
@@ -168,7 +168,7 @@ github.com/klauspost/compress/zstd/internal/xxhash
# github.com/kyokomi/emoji v2.2.1+incompatible
## explicit
github.com/kyokomi/emoji
# github.com/microcosm-cc/bluemonday v1.0.4
# github.com/microcosm-cc/bluemonday v1.0.5
## explicit
github.com/microcosm-cc/bluemonday
# github.com/nullrocks/identicon v0.0.0-20180626043057-7875f45b0022