Preserve BOM in web editor (#28935) (#28959)

Backport #28935 by @silverwind

The `ToUTF8*` functions were stripping BOM, while BOM is actually valid
in UTF8, so the stripping must be optional depending on use case. This
does:

- Add a options struct to all `ToUTF8*` functions, that by default will
strip BOM to preserve existing behaviour
- Remove `ToUTF8` function, it was dead code
- Rename `ToUTF8WithErr` to `ToUTF8`
- Preserve BOM in Monaco Editor
- Remove a unnecessary newline in the textarea value. Browsers did
ignore it, it seems but it's better not to rely on this behaviour.

Fixes: https://github.com/go-gitea/gitea/issues/28743
Related: https://github.com/go-gitea/gitea/issues/6716 which seems to
have once introduced a mechanism that strips and re-adds the BOM, but
from what I can tell, this mechanism was removed at some point after
that PR.

Co-authored-by: silverwind <me@silverwind.io>

modules/charset/charset_test.go

@@ -30,15 +30,15 @@ func resetDefaultCharsetsOrder() {
 	}
 }
 
-func TestRemoveBOMIfPresent(t *testing.T) {
-	res := RemoveBOMIfPresent([]byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba})
+func TestMaybeRemoveBOM(t *testing.T) {
+	res := MaybeRemoveBOM([]byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, ConvertOpts{})
 	assert.Equal(t, []byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, res)
 
-	res = RemoveBOMIfPresent([]byte{0xef, 0xbb, 0xbf, 0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba})
+	res = MaybeRemoveBOM([]byte{0xef, 0xbb, 0xbf, 0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, ConvertOpts{})
 	assert.Equal(t, []byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, res)
 }
 
-func TestToUTF8WithErr(t *testing.T) {
+func TestToUTF8(t *testing.T) {
 	resetDefaultCharsetsOrder()
 	var res string
 	var err error
@@ -47,53 +47,53 @@ func TestToUTF8WithErr(t *testing.T) {
 	// locale, so some conversions might behave differently. For that reason, we don't
 	// depend on particular conversions but in expected behaviors.
 
-	res, err = ToUTF8WithErr([]byte{0x41, 0x42, 0x43})
+	res, err = ToUTF8([]byte{0x41, 0x42, 0x43}, ConvertOpts{})
 	assert.NoError(t, err)
 	assert.Equal(t, "ABC", res)
 
 	// "áéíóú"
-	res, err = ToUTF8WithErr([]byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba})
+	res, err = ToUTF8([]byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, ConvertOpts{})
 	assert.NoError(t, err)
 	assert.Equal(t, []byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, []byte(res))
 
 	// "áéíóú"
-	res, err = ToUTF8WithErr([]byte{
+	res, err = ToUTF8([]byte{
 		0xef, 0xbb, 0xbf, 0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3,
 		0xc3, 0xba,
-	})
+	}, ConvertOpts{})
 	assert.NoError(t, err)
 	assert.Equal(t, []byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, []byte(res))
 
-	res, err = ToUTF8WithErr([]byte{
+	res, err = ToUTF8([]byte{
 		0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63,
 		0xF3, 0x6D, 0x6F, 0x20, 0xF1, 0x6F, 0x73, 0x41, 0x41, 0x41, 0x2e,
-	})
+	}, ConvertOpts{})
 	assert.NoError(t, err)
 	stringMustStartWith(t, "Hola,", res)
 	stringMustEndWith(t, "AAA.", res)
 
-	res, err = ToUTF8WithErr([]byte{
+	res, err = ToUTF8([]byte{
 		0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63,
 		0xF3, 0x6D, 0x6F, 0x20, 0x07, 0xA4, 0x6F, 0x73, 0x41, 0x41, 0x41, 0x2e,
-	})
+	}, ConvertOpts{})
 	assert.NoError(t, err)
 	stringMustStartWith(t, "Hola,", res)
 	stringMustEndWith(t, "AAA.", res)
 
-	res, err = ToUTF8WithErr([]byte{
+	res, err = ToUTF8([]byte{
 		0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63,
 		0xF3, 0x6D, 0x6F, 0x20, 0x81, 0xA4, 0x6F, 0x73, 0x41, 0x41, 0x41, 0x2e,
-	})
+	}, ConvertOpts{})
 	assert.NoError(t, err)
 	stringMustStartWith(t, "Hola,", res)
 	stringMustEndWith(t, "AAA.", res)
 
 	// Japanese (Shift-JIS)
 	// 日属秘ぞしちゅ。
-	res, err = ToUTF8WithErr([]byte{
+	res, err = ToUTF8([]byte{
 		0x93, 0xFA, 0x91, 0xAE, 0x94, 0xE9, 0x82, 0xBC, 0x82, 0xB5, 0x82,
 		0xBF, 0x82, 0xE3, 0x81, 0x42,
-	})
+	}, ConvertOpts{})
 	assert.NoError(t, err)
 	assert.Equal(t, []byte{
 		0xE6, 0x97, 0xA5, 0xE5, 0xB1, 0x9E, 0xE7, 0xA7, 0x98, 0xE3,
@@ -101,7 +101,7 @@ func TestToUTF8WithErr(t *testing.T) {
 	},
 		[]byte(res))
 
-	res, err = ToUTF8WithErr([]byte{0x00, 0x00, 0x00, 0x00})
+	res, err = ToUTF8([]byte{0x00, 0x00, 0x00, 0x00}, ConvertOpts{})
 	assert.NoError(t, err)
 	assert.Equal(t, []byte{0x00, 0x00, 0x00, 0x00}, []byte(res))
 }
@@ -109,22 +109,22 @@ func TestToUTF8WithErr(t *testing.T) {
 func TestToUTF8WithFallback(t *testing.T) {
 	resetDefaultCharsetsOrder()
 	// "ABC"
-	res := ToUTF8WithFallback([]byte{0x41, 0x42, 0x43})
+	res := ToUTF8WithFallback([]byte{0x41, 0x42, 0x43}, ConvertOpts{})
 	assert.Equal(t, []byte{0x41, 0x42, 0x43}, res)
 
 	// "áéíóú"
-	res = ToUTF8WithFallback([]byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba})
+	res = ToUTF8WithFallback([]byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, ConvertOpts{})
 	assert.Equal(t, []byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, res)
 
 	// UTF8 BOM + "áéíóú"
-	res = ToUTF8WithFallback([]byte{0xef, 0xbb, 0xbf, 0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba})
+	res = ToUTF8WithFallback([]byte{0xef, 0xbb, 0xbf, 0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, ConvertOpts{})
 	assert.Equal(t, []byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, res)
 
 	// "Hola, así cómo ños"
 	res = ToUTF8WithFallback([]byte{
 		0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63,
 		0xF3, 0x6D, 0x6F, 0x20, 0xF1, 0x6F, 0x73,
-	})
+	}, ConvertOpts{})
 	assert.Equal(t, []byte{
 		0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xC3, 0xAD, 0x20, 0x63,
 		0xC3, 0xB3, 0x6D, 0x6F, 0x20, 0xC3, 0xB1, 0x6F, 0x73,
@@ -133,126 +133,65 @@ func TestToUTF8WithFallback(t *testing.T) {
 	// "Hola, así cómo "
 	minmatch := []byte{0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xC3, 0xAD, 0x20, 0x63, 0xC3, 0xB3, 0x6D, 0x6F, 0x20}
 
-	res = ToUTF8WithFallback([]byte{0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63, 0xF3, 0x6D, 0x6F, 0x20, 0x07, 0xA4, 0x6F, 0x73})
+	res = ToUTF8WithFallback([]byte{0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63, 0xF3, 0x6D, 0x6F, 0x20, 0x07, 0xA4, 0x6F, 0x73}, ConvertOpts{})
 	// Do not fail for differences in invalid cases, as the library might change the conversion criteria for those
 	assert.Equal(t, minmatch, res[0:len(minmatch)])
 
-	res = ToUTF8WithFallback([]byte{0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63, 0xF3, 0x6D, 0x6F, 0x20, 0x81, 0xA4, 0x6F, 0x73})
+	res = ToUTF8WithFallback([]byte{0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63, 0xF3, 0x6D, 0x6F, 0x20, 0x81, 0xA4, 0x6F, 0x73}, ConvertOpts{})
 	// Do not fail for differences in invalid cases, as the library might change the conversion criteria for those
 	assert.Equal(t, minmatch, res[0:len(minmatch)])
 
 	// Japanese (Shift-JIS)
 	// "日属秘ぞしちゅ。"
-	res = ToUTF8WithFallback([]byte{0x93, 0xFA, 0x91, 0xAE, 0x94, 0xE9, 0x82, 0xBC, 0x82, 0xB5, 0x82, 0xBF, 0x82, 0xE3, 0x81, 0x42})
+	res = ToUTF8WithFallback([]byte{0x93, 0xFA, 0x91, 0xAE, 0x94, 0xE9, 0x82, 0xBC, 0x82, 0xB5, 0x82, 0xBF, 0x82, 0xE3, 0x81, 0x42}, ConvertOpts{})
 	assert.Equal(t, []byte{
 		0xE6, 0x97, 0xA5, 0xE5, 0xB1, 0x9E, 0xE7, 0xA7, 0x98, 0xE3,
 		0x81, 0x9E, 0xE3, 0x81, 0x97, 0xE3, 0x81, 0xA1, 0xE3, 0x82, 0x85, 0xE3, 0x80, 0x82,
 	}, res)
 
-	res = ToUTF8WithFallback([]byte{0x00, 0x00, 0x00, 0x00})
+	res = ToUTF8WithFallback([]byte{0x00, 0x00, 0x00, 0x00}, ConvertOpts{})
 	assert.Equal(t, []byte{0x00, 0x00, 0x00, 0x00}, res)
 }
 
-func TestToUTF8(t *testing.T) {
-	resetDefaultCharsetsOrder()
-	// Note: golang compiler seems so behave differently depending on the current
-	// locale, so some conversions might behave differently. For that reason, we don't
-	// depend on particular conversions but in expected behaviors.
-
-	res := ToUTF8(string([]byte{0x41, 0x42, 0x43}))
-	assert.Equal(t, "ABC", res)
-
-	// "áéíóú"
-	res = ToUTF8(string([]byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}))
-	assert.Equal(t, []byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, []byte(res))
-
-	// BOM + "áéíóú"
-	res = ToUTF8(string([]byte{
-		0xef, 0xbb, 0xbf, 0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3,
-		0xc3, 0xba,
-	}))
-	assert.Equal(t, []byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, []byte(res))
-
-	// Latin1
-	// Hola, así cómo ños
-	res = ToUTF8(string([]byte{
-		0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63,
-		0xF3, 0x6D, 0x6F, 0x20, 0xF1, 0x6F, 0x73,
-	}))
-	assert.Equal(t, []byte{
-		0x48, 0x6f, 0x6c, 0x61, 0x2c, 0x20, 0x61, 0x73, 0xc3, 0xad, 0x20, 0x63,
-		0xc3, 0xb3, 0x6d, 0x6f, 0x20, 0xc3, 0xb1, 0x6f, 0x73,
-	}, []byte(res))
-
-	// Latin1
-	// Hola, así cómo \x07ños
-	res = ToUTF8(string([]byte{
-		0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63,
-		0xF3, 0x6D, 0x6F, 0x20, 0x07, 0xA4, 0x6F, 0x73,
-	}))
-	// Hola,
-	bytesMustStartWith(t, []byte{0x48, 0x6F, 0x6C, 0x61, 0x2C}, []byte(res))
-
-	// This test FAILS
-	// res = ToUTF8("Hola, así cómo \x81ños")
-	// Do not fail for differences in invalid cases, as the library might change the conversion criteria for those
-	// assert.Regexp(t, "^Hola, así cómo", res)
-
-	// Japanese (Shift-JIS)
-	// 日属秘ぞしちゅ。
-	res = ToUTF8(string([]byte{
-		0x93, 0xFA, 0x91, 0xAE, 0x94, 0xE9, 0x82, 0xBC, 0x82, 0xB5, 0x82,
-		0xBF, 0x82, 0xE3, 0x81, 0x42,
-	}))
-	assert.Equal(t, []byte{
-		0xE6, 0x97, 0xA5, 0xE5, 0xB1, 0x9E, 0xE7, 0xA7, 0x98, 0xE3,
-		0x81, 0x9E, 0xE3, 0x81, 0x97, 0xE3, 0x81, 0xA1, 0xE3, 0x82, 0x85, 0xE3, 0x80, 0x82,
-	},
-		[]byte(res))
-
-	res = ToUTF8("\x00\x00\x00\x00")
-	assert.Equal(t, []byte{0x00, 0x00, 0x00, 0x00}, []byte(res))
-}
-
 func TestToUTF8DropErrors(t *testing.T) {
 	resetDefaultCharsetsOrder()
 	// "ABC"
-	res := ToUTF8DropErrors([]byte{0x41, 0x42, 0x43})
+	res := ToUTF8DropErrors([]byte{0x41, 0x42, 0x43}, ConvertOpts{})
 	assert.Equal(t, []byte{0x41, 0x42, 0x43}, res)
 
 	// "áéíóú"
-	res = ToUTF8DropErrors([]byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba})
+	res = ToUTF8DropErrors([]byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, ConvertOpts{})
 	assert.Equal(t, []byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, res)
 
 	// UTF8 BOM + "áéíóú"
-	res = ToUTF8DropErrors([]byte{0xef, 0xbb, 0xbf, 0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba})
+	res = ToUTF8DropErrors([]byte{0xef, 0xbb, 0xbf, 0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, ConvertOpts{})
 	assert.Equal(t, []byte{0xc3, 0xa1, 0xc3, 0xa9, 0xc3, 0xad, 0xc3, 0xb3, 0xc3, 0xba}, res)
 
 	// "Hola, así cómo ños"
-	res = ToUTF8DropErrors([]byte{0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63, 0xF3, 0x6D, 0x6F, 0x20, 0xF1, 0x6F, 0x73})
+	res = ToUTF8DropErrors([]byte{0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63, 0xF3, 0x6D, 0x6F, 0x20, 0xF1, 0x6F, 0x73}, ConvertOpts{})
 	assert.Equal(t, []byte{0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73}, res[:8])
 	assert.Equal(t, []byte{0x73}, res[len(res)-1:])
 
 	// "Hola, así cómo "
 	minmatch := []byte{0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xC3, 0xAD, 0x20, 0x63, 0xC3, 0xB3, 0x6D, 0x6F, 0x20}
 
-	res = ToUTF8DropErrors([]byte{0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63, 0xF3, 0x6D, 0x6F, 0x20, 0x07, 0xA4, 0x6F, 0x73})
+	res = ToUTF8DropErrors([]byte{0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63, 0xF3, 0x6D, 0x6F, 0x20, 0x07, 0xA4, 0x6F, 0x73}, ConvertOpts{})
 	// Do not fail for differences in invalid cases, as the library might change the conversion criteria for those
 	assert.Equal(t, minmatch, res[0:len(minmatch)])
 
-	res = ToUTF8DropErrors([]byte{0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63, 0xF3, 0x6D, 0x6F, 0x20, 0x81, 0xA4, 0x6F, 0x73})
+	res = ToUTF8DropErrors([]byte{0x48, 0x6F, 0x6C, 0x61, 0x2C, 0x20, 0x61, 0x73, 0xED, 0x20, 0x63, 0xF3, 0x6D, 0x6F, 0x20, 0x81, 0xA4, 0x6F, 0x73}, ConvertOpts{})
 	// Do not fail for differences in invalid cases, as the library might change the conversion criteria for those
 	assert.Equal(t, minmatch, res[0:len(minmatch)])
 
 	// Japanese (Shift-JIS)
 	// "日属秘ぞしちゅ。"
-	res = ToUTF8DropErrors([]byte{0x93, 0xFA, 0x91, 0xAE, 0x94, 0xE9, 0x82, 0xBC, 0x82, 0xB5, 0x82, 0xBF, 0x82, 0xE3, 0x81, 0x42})
+	res = ToUTF8DropErrors([]byte{0x93, 0xFA, 0x91, 0xAE, 0x94, 0xE9, 0x82, 0xBC, 0x82, 0xB5, 0x82, 0xBF, 0x82, 0xE3, 0x81, 0x42}, ConvertOpts{})
 	assert.Equal(t, []byte{
 		0xE6, 0x97, 0xA5, 0xE5, 0xB1, 0x9E, 0xE7, 0xA7, 0x98, 0xE3,
 		0x81, 0x9E, 0xE3, 0x81, 0x97, 0xE3, 0x81, 0xA1, 0xE3, 0x82, 0x85, 0xE3, 0x80, 0x82,
 	}, res)
 
-	res = ToUTF8DropErrors([]byte{0x00, 0x00, 0x00, 0x00})
+	res = ToUTF8DropErrors([]byte{0x00, 0x00, 0x00, 0x00}, ConvertOpts{})
 	assert.Equal(t, []byte{0x00, 0x00, 0x00, 0x00}, res)
 }
 
@@ -302,10 +241,6 @@ func stringMustEndWith(t *testing.T, expected, value string) {
 	assert.Equal(t, expected, value[len(value)-len(expected):])
 }
 
-func bytesMustStartWith(t *testing.T, expected, value []byte) {
-	assert.Equal(t, expected, value[:len(expected)])
-}
-
 func TestToUTF8WithFallbackReader(t *testing.T) {
 	resetDefaultCharsetsOrder()
 
@@ -317,7 +252,7 @@ func TestToUTF8WithFallbackReader(t *testing.T) {
 		}
 		input = input[:testLen]
 		input += "// Выключаем"
-		rd := ToUTF8WithFallbackReader(bytes.NewReader([]byte(input)))
+		rd := ToUTF8WithFallbackReader(bytes.NewReader([]byte(input)), ConvertOpts{})
 		r, _ := io.ReadAll(rd)
 		assert.EqualValuesf(t, input, string(r), "testing string len=%d", testLen)
 	}