if the charset is unknown, assume UTF-8

[debbugs.git] / Debbugs / UTF8.pm

diff --git a/Debbugs/UTF8.pm b/Debbugs/UTF8.pm
index c90cedf48109f0c6301b1e87ebdc964e28d1a170..01351f3668f04e2e33704eb672300006bf656579 100644 (file)
--- a/Debbugs/UTF8.pm
+++ b/Debbugs/UTF8.pm
@@ -28,7 +28,7 @@ charsets to UTF8.
 use warnings;
 use strict;
 use vars qw($VERSION $DEBUG %EXPORT_TAGS @EXPORT_OK @EXPORT);
-use base qw(Exporter);
+use Exporter qw(import);
 
 BEGIN{
      $VERSION = 1.00;
@@ -68,7 +68,7 @@ our $depth = 0;
 sub encode_utf8_structure {
     ++$depth;
     my @ret;
-    for my $_ (@_) {
+    for $_ (@_) {
        if (ref($_) eq 'HASH') {
            push @ret, {encode_utf8_structure(%{$depth == 1 ? dclone($_):$_})};
        }
@@ -143,8 +143,6 @@ sub decode_utf8_safely{
 
 =cut
 
-our %iconv_converters;
-
 sub convert_to_utf8 {
     my ($data,$charset,$internal_call) = @_;
     $internal_call //= 0;
@@ -156,28 +154,26 @@ sub convert_to_utf8 {
     if ($charset eq 'RAW') {
         croak("Charset must not be raw when calling convert_to_utf8");
     }
-    if (not defined $iconv_converters{$charset}) {
-        eval {
-            $iconv_converters{$charset} = Text::Iconv->new($charset,"UTF-8") or
-                die "Unable to create converter for '$charset'";
-        };
-        if ($@) {
-            return undef if $internal_call;
-            warn $@;
-            # We weren't able to create the converter, so use Encode
-            # instead
-            return __fallback_convert_to_utf8($data,$charset);
-        }
+    ## if the charset is unknown or unknown 8 bit, assume that it's UTF-8.
+    if ($charset =~ /unknown/i) {
+       $charset = 'UTF-8'
     }
-    if (not defined $iconv_converters{$charset}) {
+    my $iconv_converter;
+    eval {
+        $iconv_converter = Text::Iconv->new($charset,"UTF-8") or
+            die "Unable to create converter for '$charset'";
+    };
+    if ($@) {
         return undef if $internal_call;
-        warn "The converter for $charset wasn't created properly somehow!";
+        warn $@;
+        # We weren't able to create the converter, so use Encode
+        # instead
         return __fallback_convert_to_utf8($data,$charset);
     }
-    my $converted_data = $iconv_converters{$charset}->convert($data);
+    my $converted_data = $iconv_converter->convert($data);
     # if the conversion failed, retval will be undefined or perhaps
     # -1.
-    my $retval = $iconv_converters{$charset}->retval();
+    my $retval = $iconv_converter->retval();
     if (not defined $retval or
         $retval < 0
        ) {
@@ -187,11 +183,9 @@ sub convert_to_utf8 {
             # if there's an à (0xC3), it's probably something
             # horrible, and we shouldn't try to convert it.
             if (defined $call_back_data and $call_back_data !~ /\x{C3}/) {
-                warn "failed to convert to utf8 (charset: $charset, data: $data), but succeeded with ISO8859-1: ".encode_utf8($call_back_data);
                 return $call_back_data;
             }
         }
-        warn "failed to convert to utf8 (charset: $charset, data: $data)";
         # Fallback to encode, which will probably also fail.
         return __fallback_convert_to_utf8($data,$charset);
     }
@@ -210,6 +204,10 @@ sub __fallback_convert_to_utf8 {
      }
      # lets assume everything that doesn't have a charset is utf8
      $charset //= 'utf8';
+     ## if the charset is unknown, assume it's UTF-8
+     if ($charset =~ /unknown/i) {
+        $charset = 'utf8';
+     }
      my $result;
      eval {
         $result = decode($charset,$data,0);