1
1
mirror of https://github.com/go-gitea/gitea synced 2024-12-25 18:14:28 +00:00
gitea/modules/charset/escape.go
wxiaoguang a3f403f438
Add option to disable ambiguous unicode characters detection (#28454) (#28499)
Backport #28454 (the only conflict is caused by some comments)

* Close #24483
* Close #28123
* Close #23682
* Close #23149
2023-12-18 12:20:37 +08:00

45 lines
1.6 KiB
Go

// Copyright 2022 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
//go:generate go run invisible/generate.go -v -o ./invisible_gen.go
//go:generate go run ambiguous/generate.go -v -o ./ambiguous_gen.go ambiguous/ambiguous.json
package charset
import (
"html/template"
"io"
"strings"
"code.gitea.io/gitea/modules/log"
"code.gitea.io/gitea/modules/setting"
"code.gitea.io/gitea/modules/translation"
)
// RuneNBSP is the codepoint for NBSP
const RuneNBSP = 0xa0
// EscapeControlHTML escapes the unicode control sequences in a provided html document
func EscapeControlHTML(html template.HTML, locale translation.Locale, allowed ...rune) (escaped *EscapeStatus, output template.HTML) {
sb := &strings.Builder{}
escaped, _ = EscapeControlReader(strings.NewReader(string(html)), sb, locale, allowed...) // err has been handled in EscapeControlReader
return escaped, template.HTML(sb.String())
}
// EscapeControlReader escapes the unicode control sequences in a provided reader of HTML content and writer in a locale and returns the findings as an EscapeStatus
func EscapeControlReader(reader io.Reader, writer io.Writer, locale translation.Locale, allowed ...rune) (escaped *EscapeStatus, err error) {
if !setting.UI.AmbiguousUnicodeDetection {
_, err = io.Copy(writer, reader)
return &EscapeStatus{}, err
}
outputStream := &HTMLStreamerWriter{Writer: writer}
streamer := NewEscapeStreamer(locale, outputStream, allowed...).(*escapeStreamer)
if err = StreamHTML(reader, streamer); err != nil {
streamer.escaped.HasError = true
log.Error("Error whilst escaping: %v", err)
}
return streamer.escaped, err
}