(ensure_utf8) introduce it (copied & slightly altered from rpmtools)

so that we never crash on garbaged UTF-8

1 files changed, 14 insertions, 0 deletions

diff --git a/Rpmdrake/formatting.pm b/Rpmdrake/formatting.pm
index bf8aa987..8b1ab479 100644
--- a/Rpmdrake/formatting.pm
+++ b/Rpmdrake/formatting.pm
@@ -35,6 +35,7 @@ use Exporter;
 our @ISA = qw(Exporter);
 our @EXPORT = qw(
                     $spacing
+                    ensure_utf8
                     format_changelog_changelogs
                     format_changelog_string
                     format_field
@@ -50,6 +51,19 @@ our @EXPORT = qw(
                     urpm_name);
 
 
+# from rpmtools, #37482:
+sub ensure_utf8 {
+    my ($s) = @_;
+    require Encode;
+    Encode::_utf8_on($s); #- this is done on the copy
+    if (!Encode::is_utf8($s, 1)) {
+        Encode::_utf8_off($_[0]);
+        Encode::from_to($_[0], 'iso-8859-15', 'utf8'); # most probable
+    }
+    Encode::_utf8_on($_[0]); #- now we know it is valid utf8
+    $_[0];
+}
+
 sub rpm_description {
     my ($description) = @_;
     utf8::decode($description);