From e79a807a8461a73bd66146d816f635b66e198c89 Mon Sep 17 00:00:00 2001
From: coldWater <254244460@qq.com>
Date: Thu, 14 Mar 2024 10:51:55 +0800
Subject: [PATCH] Refactor markup/csv: don't read all to memory (#29760)

---
 modules/markup/csv/csv.go      | 63 ++++++++++++++++++++++++++--------
 modules/markup/csv/csv_test.go | 10 ++++++
 2 files changed, 58 insertions(+), 15 deletions(-)

diff --git a/modules/markup/csv/csv.go b/modules/markup/csv/csv.go
index 12458e954..570c4f470 100644
--- a/modules/markup/csv/csv.go
+++ b/modules/markup/csv/csv.go
@@ -77,29 +77,62 @@ func writeField(w io.Writer, element, class, field string) error {
 }
 
 // Render implements markup.Renderer
-func (Renderer) Render(ctx *markup.RenderContext, input io.Reader, output io.Writer) error {
+func (r Renderer) Render(ctx *markup.RenderContext, input io.Reader, output io.Writer) error {
 	tmpBlock := bufio.NewWriter(output)
+	maxSize := setting.UI.CSV.MaxFileSize
 
-	// FIXME: don't read all to memory
-	rawBytes, err := io.ReadAll(input)
+	if maxSize == 0 {
+		return r.tableRender(ctx, input, tmpBlock)
+	}
+
+	rawBytes, err := io.ReadAll(io.LimitReader(input, maxSize+1))
 	if err != nil {
 		return err
 	}
 
-	if setting.UI.CSV.MaxFileSize != 0 && setting.UI.CSV.MaxFileSize < int64(len(rawBytes)) {
-		if _, err := tmpBlock.WriteString("<pre>"); err != nil {
-			return err
-		}
-		if _, err := tmpBlock.WriteString(html.EscapeString(string(rawBytes))); err != nil {
-			return err
-		}
-		if _, err := tmpBlock.WriteString("</pre>"); err != nil {
-			return err
-		}
-		return tmpBlock.Flush()
+	if int64(len(rawBytes)) <= maxSize {
+		return r.tableRender(ctx, bytes.NewReader(rawBytes), tmpBlock)
+	}
+	return r.fallbackRender(io.MultiReader(bytes.NewReader(rawBytes), input), tmpBlock)
+}
+
+func (Renderer) fallbackRender(input io.Reader, tmpBlock *bufio.Writer) error {
+	_, err := tmpBlock.WriteString("<pre>")
+	if err != nil {
+		return err
 	}
 
-	rd, err := csv.CreateReaderAndDetermineDelimiter(ctx, bytes.NewReader(rawBytes))
+	scan := bufio.NewScanner(input)
+	scan.Split(bufio.ScanRunes)
+	for scan.Scan() {
+		switch scan.Text() {
+		case `&`:
+			_, err = tmpBlock.WriteString("&amp;")
+		case `'`:
+			_, err = tmpBlock.WriteString("&#39;") // "&#39;" is shorter than "&apos;" and apos was not in HTML until HTML5.
+		case `<`:
+			_, err = tmpBlock.WriteString("&lt;")
+		case `>`:
+			_, err = tmpBlock.WriteString("&gt;")
+		case `"`:
+			_, err = tmpBlock.WriteString("&#34;") // "&#34;" is shorter than "&quot;".
+		default:
+			_, err = tmpBlock.Write(scan.Bytes())
+		}
+		if err != nil {
+			return err
+		}
+	}
+
+	_, err = tmpBlock.WriteString("</pre>")
+	if err != nil {
+		return err
+	}
+	return tmpBlock.Flush()
+}
+
+func (Renderer) tableRender(ctx *markup.RenderContext, input io.Reader, tmpBlock *bufio.Writer) error {
+	rd, err := csv.CreateReaderAndDetermineDelimiter(ctx, input)
 	if err != nil {
 		return err
 	}
diff --git a/modules/markup/csv/csv_test.go b/modules/markup/csv/csv_test.go
index 8c07184b2..3d12be477 100644
--- a/modules/markup/csv/csv_test.go
+++ b/modules/markup/csv/csv_test.go
@@ -4,6 +4,8 @@
 package markup
 
 import (
+	"bufio"
+	"bytes"
 	"strings"
 	"testing"
 
@@ -29,4 +31,12 @@ func TestRenderCSV(t *testing.T) {
 		assert.NoError(t, err)
 		assert.EqualValues(t, v, buf.String())
 	}
+
+	t.Run("fallbackRender", func(t *testing.T) {
+		var buf bytes.Buffer
+		err := render.fallbackRender(strings.NewReader("1,<a>\n2,<b>"), bufio.NewWriter(&buf))
+		assert.NoError(t, err)
+		want := "<pre>1,&lt;a&gt;\n2,&lt;b&gt;</pre>"
+		assert.Equal(t, want, buf.String())
+	})
 }