Don't treat BOM escape sequence as hidden character. (#18909) (#18910)

* Don't treat BOM escape sequence as hidden character. (#18909)

Backport #18909
This commit is contained in:
Gusted 2022-02-26 22:15:04 +00:00 committed by GitHub
parent df35049196
commit 4fb718d405
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 27 additions and 5 deletions

View file

@ -63,6 +63,7 @@ func EscapeControlBytes(text []byte) (EscapeStatus, []byte) {
func EscapeControlReader(text io.Reader, output io.Writer) (escaped EscapeStatus, err error) {
buf := make([]byte, 4096)
readStart := 0
runeCount := 0
var n int
var writePos int
@ -79,6 +80,8 @@ readingloop:
for i < len(bs) {
r, size := utf8.DecodeRune(bs[i:])
runeCount++
// Now handle the codepoints
switch {
case r == utf8.RuneError:
@ -113,6 +116,8 @@ readingloop:
lineHasRTLScript = false
lineHasLTRScript = false
case runeCount == 1 && r == 0xFEFF: // UTF BOM
// the first BOM is safe
case r == '\r' || r == '\t' || r == ' ':
// These are acceptable control characters and space characters
case unicode.IsSpace(r):
@ -144,7 +149,8 @@ readingloop:
return
}
writePos = i + size
case unicode.Is(unicode.C, r):
// 65279 == BOM rune.
case unicode.Is(unicode.C, r) && r != rune(65279):
escaped.Escaped = true
escaped.HasControls = true
if writePos < i {

View file

@ -129,6 +129,14 @@ then resh (ר), and finally heh (ה) (which should appear leftmost).`,
"\n" + `if access_level != "user<span class="escaped-code-point" data-escaped="[U+202E]"><span class="char">` + "\u202e" + `</span></span> <span class="escaped-code-point" data-escaped="[U+2066]"><span class="char">` + "\u2066" + `</span></span>// Check if admin<span class="escaped-code-point" data-escaped="[U+2069]"><span class="char">` + "\u2069" + `</span></span> <span class="escaped-code-point" data-escaped="[U+2066]"><span class="char">` + "\u2066" + `</span></span>" {` + "\n",
status: EscapeStatus{Escaped: true, HasBIDI: true, BadBIDI: true, HasLTRScript: true, HasRTLScript: true},
},
{
// UTF-8/16/32 all use the same codepoint for BOM
// Gitea could read UTF-16/32 content and convert into UTF-8 internally then render it, so we only process UTF-8 internally
name: "UTF BOM",
text: "\xef\xbb\xbftest",
result: "\xef\xbb\xbftest",
status: EscapeStatus{HasLTRScript: true},
},
}
func TestEscapeControlString(t *testing.T) {
@ -163,10 +171,18 @@ func TestEscapeControlReader(t *testing.T) {
// lets add some control characters to the tests
tests := make([]escapeControlTest, 0, len(escapeControlTests)*3)
copy(tests, escapeControlTests)
// if there is a BOM, we should keep the BOM
addPrefix := func(prefix, s string) string {
if strings.HasPrefix(s, "\xef\xbb\xbf") {
return s[:3] + prefix + s[3:]
}
return prefix + s
}
for _, test := range escapeControlTests {
test.name += " (+Control)"
test.text = "\u001E" + test.text
test.result = `<span class="escaped-code-point" data-escaped="[U+001E]"><span class="char">` + "\u001e" + `</span></span>` + test.result
test.text = addPrefix("\u001E", test.text)
test.result = addPrefix(`<span class="escaped-code-point" data-escaped="[U+001E]"><span class="char">`+"\u001e"+`</span></span>`, test.result)
test.status.Escaped = true
test.status.HasControls = true
tests = append(tests, test)
@ -174,8 +190,8 @@ func TestEscapeControlReader(t *testing.T) {
for _, test := range escapeControlTests {
test.name += " (+Mark)"
test.text = "\u0300" + test.text
test.result = `<span class="escaped-code-point" data-escaped="[U+0300]"><span class="char">` + "\u0300" + `</span></span>` + test.result
test.text = addPrefix("\u0300", test.text)
test.result = addPrefix(`<span class="escaped-code-point" data-escaped="[U+0300]"><span class="char">`+"\u0300"+`</span></span>`, test.result)
test.status.Escaped = true
test.status.HasMarks = true
tests = append(tests, test)