fix(GODT-2582): Fix hash for messages

When hashing child parts, do not hash the boundary parameter of the `Content-Type` header field as this is unique to each message.
ProtonMail · Apr 26, 2023 · aed4b8a · aed4b8a
1 parent 8ec759b
commit aed4b8a
Show file tree

Hide file tree

Showing 5 changed files with 118 additions and 61 deletions.
diff --git a/rfc822/hash.go b/rfc822/hash.go
@@ -4,6 +4,8 @@ import (
 	"bytes"
 	"crypto/sha256"
 	"encoding/base64"
+	"github.com/sirupsen/logrus"
+	"strings"
 )
 
 // GetMessageHash returns the hash of the given message.
@@ -60,8 +62,28 @@ func GetMessageHash(b []byte) (string, error) {
 			return err
 		}
 
-		if _, err := h.Write([]byte(header.Get("Content-Type"))); err != nil {
-			return err
+		contentType := header.Get("Content-Type")
+		mimeType, values, err := ParseMIMEType(contentType)
+		if err != nil {
+			logrus.Warnf("Message contains invalid mime type: %v", contentType)
+		} else {
+			if _, err := h.Write([]byte(mimeType)); err != nil {
+				return err
+			}
+
+			for k, v := range values {
+				if strings.EqualFold(k, "boundary") {
+					continue
+				}
+
+				if _, err := h.Write([]byte(k)); err != nil {
+					return err
+				}
+
+				if _, err := h.Write([]byte(v)); err != nil {
+					return err
+				}
+			}
 		}
 
 		if _, err := h.Write([]byte(header.Get("Content-Disposition"))); err != nil {

diff --git a/tests/recovery_mailbox_test.go b/tests/recovery_mailbox_test.go
@@ -249,8 +249,8 @@ func TestRecoveryMailboxOnlyReportsOnFirstDedupedMessage(t *testing.T) {
 		status, err := client.Select("INBOX", false)
 		require.NoError(t, err)
 		require.Equal(t, uint32(0), status.Messages)
-		require.Error(t, doAppendWithClientFromFile(t, client, "INBOX", "testdata/multipart-mixed.eml", time.Now()))
-		require.Error(t, doAppendWithClientFromFile(t, client, "INBOX", "testdata/multipart-mixed2.eml", time.Now()))
+		require.Error(t, doAppendWithClientFromFile(t, client, "INBOX", "testdata/original.eml", time.Now()))
+		require.Error(t, doAppendWithClientFromFile(t, client, "INBOX", "testdata/duplicate.eml", time.Now()))
 
 		{
 			status, err := client.Status(ids.GluonRecoveryMailboxName, []goimap.StatusItem{goimap.StatusMessages})

diff --git a/tests/testdata/duplicate.eml b/tests/testdata/duplicate.eml
@@ -0,0 +1,46 @@
+From: "Testing" <[email protected]>
+To: "Test 123" <[email protected]>
+Subject: Small doc attachment VALID v5
+Date: Wed, 26 Apr 2023 08:25:16 +0200
+MIME-Version: 1.0
+Content-Type: multipart/mixed;
+	boundary="----=_NextPart_000_0022_01D97836.53531060"
+X-Mailer: Microsoft Outlook 16.0
+Content-Language: en-us
+X-OlkEid: 00000000225333D1866CB24A8049706F3452C40E0700C3B68E10F77511CEB4CD00AA00BBB6E600000000000C00003BB97E4DCE360843BCF25AA31956C0E2000000000408000048AD3E35A4DC1A47ABA5F573BB76D09A
+
+This is a multipart message in MIME format.
+
+------=_NextPart_000_0022_01D97836.53531060
+Content-Type: multipart/alternative;
+	boundary="----=_NextPart_001_0023_01D97836.53531060"
+
+
+------=_NextPart_001_0023_01D97836.53531060
+Content-Type: text/plain;
+	charset="utf-8"
+Content-Transfer-Encoding: 7bit
+
+Simple Body
+
+
+------=_NextPart_001_0023_01D97836.53531060
+Content-Type: text/html;
+	boundary="----=_NextPart_000_003E_01D97831.767F38C0";
+	charset="utf-8"
+Content-Transfer-Encoding: quoted-printable
+
+<div style=3D"font-family: Arial, sans-serif; font-size: 14px; color: =
+rgb(0, 0, 0);">Simple Body<br></div>
+------=_NextPart_001_0023_01D97836.53531060--
+
+------=_NextPart_000_0022_01D97836.53531060
+Content-Type: application/vnd.openxmlformats-officedocument.wordprocessingml.document;
+	name="test.docx"
+Content-Transfer-Encoding: base64
+Content-Disposition: attachment;
+	filename="test.docx"
+
+UEsDBBQABgAIAAAAIQDfpNJsWgEAACAFAAATAAgCW0NvbnRlbnRfVHlwZXNdLnhtbCCiBAIooAAC
+
+------=_NextPart_000_0022_01D97836.53531060--
diff --git a/tests/testdata/multipart-mixed2.eml b/tests/testdata/multipart-mixed2.eml
diff --git a/tests/testdata/original.eml b/tests/testdata/original.eml
@@ -0,0 +1,46 @@
+From: "Testing" <[email protected]>
+To: "Test 123" <[email protected]>
+Subject: Small doc attachment VALID v5
+Date: Wed, 26 Apr 2023 08:25:16 +0200
+MIME-Version: 1.0
+Content-Type: multipart/mixed;
+	boundary="----=_NextPart_000_003E_01D97831.767F38C0"
+X-Mailer: Microsoft Outlook 16.0
+Content-Language: en-us
+X-OlkEid: 00000000225333D1866CB24A8049706F3452C40E0700C3B68E10F77511CEB4CD00AA00BBB6E600000000000C00003BB97E4DCE360843BCF25AA31956C0E20000000004030000DD8E24308ACA7243B56646E28B20359C
+
+This is a multipart message in MIME format.
+
+------=_NextPart_000_003E_01D97831.767F38C0
+Content-Type: multipart/alternative;
+	boundary="----=_NextPart_001_003F_01D97831.767F38C0"
+
+
+------=_NextPart_001_003F_01D97831.767F38C0
+Content-Type: text/plain;
+	charset="utf-8"
+Content-Transfer-Encoding: 7bit
+
+Simple Body
+
+
+------=_NextPart_001_003F_01D97831.767F38C0
+Content-Type: text/html;
+	boundary="b1_CZjzIg4XCgPBIDFY9EOw4CTGxwvt0c6cIuThfhnqmVc";
+	charset="utf-8"
+Content-Transfer-Encoding: quoted-printable
+
+<div style=3D"font-family: Arial, sans-serif; font-size: 14px; color: =
+rgb(0, 0, 0);">Simple Body<br></div>
+------=_NextPart_001_003F_01D97831.767F38C0--
+
+------=_NextPart_000_003E_01D97831.767F38C0
+Content-Type: application/vnd.openxmlformats-officedocument.wordprocessingml.document;
+	name="test.docx"
+Content-Transfer-Encoding: base64
+Content-Disposition: attachment;
+	filename="test.docx"
+
+UEsDBBQABgAIAAAAIQDfpNJsWgEAACAFAAATAAgCW0NvbnRlbnRfVHlwZXNdLnhtbCCiBAIooAAC
+
+------=_NextPart_000_003E_01D97831.767F38C0--