From e5b0725824ac9ccf218732238e4b3b525fa6ad46 Mon Sep 17 00:00:00 2001
From: Koni Marti <koni.marti@gmail.com>
Date: Wed, 21 Sep 2022 00:27:58 +0200
Subject: [PATCH] charset: handle unknown charsets more user-friendly

Do not throw an error when the charset is unknown; the message entity
can still be read, but log the error instead.

Reported-by: falsifian
Signed-off-by: Koni Marti <koni.marti@gmail.com>
Acked-by: Robin Jarry <robin@jarry.cc>
---
 lib/messageview.go        |  5 ++---
 worker/imap/fetch.go      |  5 ++++-
 worker/lib/parse.go       | 17 ++++++++++++++++-
 worker/maildir/message.go |  3 +--
 worker/mbox/worker.go     |  3 +--
 worker/notmuch/message.go |  4 +---
 6 files changed, 25 insertions(+), 12 deletions(-)

diff --git a/lib/messageview.go b/lib/messageview.go
index 6ad86c9..ff754a4 100644
--- a/lib/messageview.go
+++ b/lib/messageview.go
@@ -5,7 +5,6 @@ import (
 	"io"
 
 	"github.com/ProtonMail/go-crypto/openpgp"
-	"github.com/emersion/go-message"
 	_ "github.com/emersion/go-message/charset"
 
 	"git.sr.ht/~rjarry/aerc/lib/crypto"
@@ -81,7 +80,7 @@ func NewMessageStoreView(messageInfo *models.MessageInfo,
 				cb(nil, err)
 				return
 			}
-			decrypted, err := message.Read(bytes.NewBuffer(msv.message))
+			decrypted, err := lib.ReadMessage(bytes.NewBuffer(msv.message))
 			if err != nil {
 				cb(nil, err)
 				return
@@ -124,7 +123,7 @@ func (msv *MessageStoreView) FetchBodyPart(part []int, cb func(io.Reader)) {
 	}
 
 	buf := bytes.NewBuffer(msv.message)
-	msg, err := message.Read(buf)
+	msg, err := lib.ReadMessage(buf)
 	if err != nil {
 		panic(err)
 	}
diff --git a/worker/imap/fetch.go b/worker/imap/fetch.go
index f21c6e9..d675ed1 100644
--- a/worker/imap/fetch.go
+++ b/worker/imap/fetch.go
@@ -123,7 +123,10 @@ func (imapw *IMAPWorker) handleFetchMessageBodyPart(
 
 			part, err := message.New(message.Header{Header: h},
 				_msg.GetBody(&partBodySection))
-			if err != nil {
+			if message.IsUnknownCharset(err) {
+				logging.Warnf("unknown charset encountered "+
+					"for uid %d", _msg.Uid)
+			} else if err != nil {
 				return fmt.Errorf("failed to create message reader: %w", err)
 			}
 
diff --git a/worker/lib/parse.go b/worker/lib/parse.go
index 30d06a9..3eaa2ae 100644
--- a/worker/lib/parse.go
+++ b/worker/lib/parse.go
@@ -10,6 +10,7 @@ import (
 	"strings"
 	"time"
 
+	"git.sr.ht/~rjarry/aerc/logging"
 	"git.sr.ht/~rjarry/aerc/models"
 	"github.com/emersion/go-message"
 	_ "github.com/emersion/go-message/charset"
@@ -252,7 +253,7 @@ func MessageInfo(raw RawMessage) (*models.MessageInfo, error) {
 		return nil, err
 	}
 	defer r.Close()
-	msg, err := message.Read(r)
+	msg, err := ReadMessage(r)
 	if err != nil {
 		return nil, fmt.Errorf("could not read message: %w", err)
 	}
@@ -307,3 +308,17 @@ func NewCRLFReader(r io.Reader) io.Reader {
 	}
 	return &buf
 }
+
+// ReadMessage is a wrapper for the message.Read function to read a message
+// from r. The message's encoding and charset are automatically decoded to
+// UTF-8. If an unknown charset is encountered, the error is logged but a nil
+// error is returned since the entity object can still be read.
+func ReadMessage(r io.Reader) (*message.Entity, error) {
+	entity, err := message.Read(r)
+	if message.IsUnknownCharset(err) {
+		logging.Warnf("unknown charset encountered")
+	} else if err != nil {
+		return nil, fmt.Errorf("could not read message: %w", err)
+	}
+	return entity, nil
+}
diff --git a/worker/maildir/message.go b/worker/maildir/message.go
index 81573c5..9cd1331 100644
--- a/worker/maildir/message.go
+++ b/worker/maildir/message.go
@@ -5,7 +5,6 @@ import (
 	"io"
 
 	"github.com/emersion/go-maildir"
-	"github.com/emersion/go-message"
 
 	"git.sr.ht/~rjarry/aerc/models"
 	"git.sr.ht/~rjarry/aerc/worker/lib"
@@ -85,7 +84,7 @@ func (m Message) NewBodyPartReader(requestedParts []int) (io.Reader, error) {
 		return nil, err
 	}
 	defer f.Close()
-	msg, err := message.Read(f)
+	msg, err := lib.ReadMessage(f)
 	if err != nil {
 		return nil, fmt.Errorf("could not read message: %w", err)
 	}
diff --git a/worker/mbox/worker.go b/worker/mbox/worker.go
index 067147e..c5c7082 100644
--- a/worker/mbox/worker.go
+++ b/worker/mbox/worker.go
@@ -15,7 +15,6 @@ import (
 	"git.sr.ht/~rjarry/aerc/worker/handlers"
 	"git.sr.ht/~rjarry/aerc/worker/lib"
 	"git.sr.ht/~rjarry/aerc/worker/types"
-	gomessage "github.com/emersion/go-message"
 )
 
 func init() {
@@ -179,7 +178,7 @@ func (w *mboxWorker) handleMessage(msg types.WorkerMessage) error {
 			break
 		}
 
-		fullMsg, err := gomessage.Read(contentReader)
+		fullMsg, err := lib.ReadMessage(contentReader)
 		if err != nil {
 			reterr = fmt.Errorf("could not read message: %w", err)
 			break
diff --git a/worker/notmuch/message.go b/worker/notmuch/message.go
index a51f78b..59367fd 100644
--- a/worker/notmuch/message.go
+++ b/worker/notmuch/message.go
@@ -11,8 +11,6 @@ import (
 	"git.sr.ht/~rjarry/aerc/models"
 	"git.sr.ht/~rjarry/aerc/worker/lib"
 	notmuch "git.sr.ht/~rjarry/aerc/worker/notmuch/lib"
-	"github.com/emersion/go-message"
-	_ "github.com/emersion/go-message/charset"
 )
 
 type Message struct {
@@ -47,7 +45,7 @@ func (m *Message) NewBodyPartReader(requestedParts []int) (io.Reader, error) {
 		return nil, err
 	}
 	defer f.Close()
-	msg, err := message.Read(f)
+	msg, err := lib.ReadMessage(f)
 	if err != nil {
 		return nil, fmt.Errorf("could not read message: %w", err)
 	}