Fix commenting on non-utf8 encoded files (#11916) (#11950)

* Add comment on non-unicode line to force fail

Signed-off-by: Andrew Thornton <art27@cantab.net>

* Just quote/unquote patch

Signed-off-by: Andrew Thornton <art27@cantab.net>

Co-authored-by: zeripath <art27@cantab.net>

1 files changed, 31 insertions, 1 deletions

diff --git a/models/issue_comment.go b/models/issue_comment.go
index bb3c4be7ba..28f45b59f6 100644
--- a/models/issue_comment.go
+++ b/models/issue_comment.go
@@ -9,7 +9,9 @@ package models
 import (
 	"fmt"
 	"regexp"
+	"strconv"
 	"strings"
+	"unicode/utf8"
 
 	"code.gitea.io/gitea/modules/git"
 	"code.gitea.io/gitea/modules/log"
@@ -139,7 +141,8 @@ type Comment struct {
 	RenderedContent string `xorm:"-"`
 
 	// Path represents the 4 lines of code cemented by this comment
-	Patch string `xorm:"TEXT"`
+	Patch       string `xorm:"-"`
+	PatchQuoted string `xorm:"TEXT patch"`
 
 	CreatedUnix timeutil.TimeStamp `xorm:"INDEX created"`
 	UpdatedUnix timeutil.TimeStamp `xorm:"INDEX updated"`
@@ -183,6 +186,33 @@ func (c *Comment) loadIssue(e Engine) (err error) {
 	return
 }
 
+// BeforeInsert will be invoked by XORM before inserting a record
+func (c *Comment) BeforeInsert() {
+	c.PatchQuoted = c.Patch
+	if !utf8.ValidString(c.Patch) {
+		c.PatchQuoted = strconv.Quote(c.Patch)
+	}
+}
+
+// BeforeUpdate will be invoked by XORM before updating a record
+func (c *Comment) BeforeUpdate() {
+	c.PatchQuoted = c.Patch
+	if !utf8.ValidString(c.Patch) {
+		c.PatchQuoted = strconv.Quote(c.Patch)
+	}
+}
+
+// AfterLoad is invoked from XORM after setting the values of all fields of this object.
+func (c *Comment) AfterLoad(session *xorm.Session) {
+	c.Patch = c.PatchQuoted
+	if len(c.PatchQuoted) > 0 && c.PatchQuoted[0] == '"' {
+		unquoted, err := strconv.Unquote(c.PatchQuoted)
+		if err == nil {
+			c.Patch = unquoted
+		}
+	}
+}
+
 func (c *Comment) loadPoster(e Engine) (err error) {
 	if c.PosterID <= 0 || c.Poster != nil {
 		return nil