4 files changed, 29 insertions, 44 deletions

diff --git a/lib/PublicInbox/ContentId.pm b/lib/PublicInbox/ContentId.pm
index b1d27eb8..dd3155be 100644
--- a/lib/PublicInbox/ContentId.pm
+++ b/lib/PublicInbox/ContentId.pm
@@ -75,14 +75,7 @@ sub content_digest ($) {
                 }
                 $dig->add("b\0");
                 my $ct = $part->content_type || 'text/plain';
-                my $s = eval { $part->body_str };
-                if ($@ && $ct =~ m!\btext/plain\b!i) {
-                        # Try to assume UTF-8 because Alpine
-                        # seems to do wacky things and set
-                        # charset=X-UNKNOWN
-                        $part->charset_set('UTF-8');
-                        $s = eval { $part->body_str };
-                }
+                my ($s, undef) = msg_part_text($part, $ct);
                 if (defined $s) {
                         $s =~ s/\r\n/\n/gs;
                         $s =~ s/\s*\z//s;
diff --git a/lib/PublicInbox/MsgIter.pm b/lib/PublicInbox/MsgIter.pm
index a795f617..9e2d797f 100644
--- a/lib/PublicInbox/MsgIter.pm
+++ b/lib/PublicInbox/MsgIter.pm
@@ -5,7 +5,7 @@ package PublicInbox::MsgIter;
 use strict;
 use warnings;
 use base qw(Exporter);
-our @EXPORT = qw(msg_iter);
+our @EXPORT = qw(msg_iter msg_part_text);
 use PublicInbox::MIME;
 
 # Like Email::MIME::walk_parts, but this is:
@@ -34,4 +34,27 @@ sub msg_iter ($$) {
         }
 }
 
+sub msg_part_text ($$) {
+        my ($part, $ct) = @_;
+
+        my $s = eval { $part->body_str };
+        my $err = $@;
+
+        # text/plain is the default, multipart/mixed happened a few
+        # times when it should not have been:
+        #   <87llgalspt.fsf@free.fr>
+        #   <200308111450.h7BEoOu20077@mail.osdl.org>
+        if ($ct =~ m!\btext/plain\b!i || $ct =~ m!\bmultipart/mixed\b!i) {
+                # Try to assume UTF-8 because Alpine seems to
+                # do wacky things and set charset=X-UNKNOWN
+                $part->charset_set('UTF-8');
+                $s = eval { $part->body_str };
+
+                # If forcing charset=UTF-8 failed,
+                # caller will warn further down...
+                $s = $part->body if $@;
+        }
+        ($s, $err);
+}
+
 1;
diff --git a/lib/PublicInbox/SearchIdx.pm b/lib/PublicInbox/SearchIdx.pm
index ca832ad3..76f3f33a 100644
--- a/lib/PublicInbox/SearchIdx.pm
+++ b/lib/PublicInbox/SearchIdx.pm
@@ -305,19 +305,7 @@ sub add_xapian ($$$$$) {
                         $self->index_text($fn, 1, 'XFN');
                 }
 
-                return if $ct =~ m!\btext/x?html\b!i;
-
-                my $s = eval { $part->body_str };
-                if ($@) {
-                        if ($ct =~ m!\btext/plain\b!i) {
-                                # Try to assume UTF-8 because Alpine
-                                # seems to do wacky things and set
-                                # charset=X-UNKNOWN
-                                $part->charset_set('UTF-8');
-                                $s = eval { $part->body_str };
-                                $s = $part->body if $@;
-                        }
-                }
+                my ($s, undef) = msg_part_text($part, $ct);
                 defined $s or return;
 
                 my (@orig, @quot);
diff --git a/lib/PublicInbox/View.pm b/lib/PublicInbox/View.pm
index 86acd824..bb49c035 100644
--- a/lib/PublicInbox/View.pm
+++ b/lib/PublicInbox/View.pm
@@ -543,33 +543,14 @@ sub add_text_body {
         my ($part, $depth) = @$p; # attachment @idx is unused
         my $ct = $part->content_type || 'text/plain';
         my $fn = $part->filename;
+        my ($s, $err) = msg_part_text($part, $ct);
 
-        if ($ct =~ m!\btext/x?html\b!i) {
-                return attach_link($upfx, $ct, $p, $fn);
-        }
-
-        my $s = eval { $part->body_str };
-
-        # badly-encoded message? tell the world about it!
-        my $err = $@;
-        if ($err) {
-                if ($ct =~ m!\btext/plain\b!i) {
-                        # Try to assume UTF-8 because Alpine seems to
-                        # do wacky things and set charset=X-UNKNOWN
-                        $part->charset_set('UTF-8');
-                        $s = eval { $part->body_str };
-
-                        # If forcing charset=UTF-8 failed,
-                        # attach_link will warn further down...
-                        $s = $part->body if $@;
-                } else {
-                        return attach_link($upfx, $ct, $p, $fn);
-                }
-        }
+        return attach_link($upfx, $ct, $p, $fn) unless defined $s;
 
         my @lines = split(/^/m, $s);
         $s = '';
         if (defined($fn) || $depth > 0 || $err) {
+                # badly-encoded message with $err? tell the world about it!
                 $s .= attach_link($upfx, $ct, $p, $fn, $err);
                 $s .= "\n";
         }