From: Joey Hess <joey@gnu.kitenet.net>
Date: Mon, 9 Mar 2009 18:18:55 +0000 (-0400)
Subject: git: Manually decode git output from utf-8, avoids warning messages on invalidly... 
X-Git-Url: https://sipb.mit.edu/gitweb.cgi/ikiwiki.git/commitdiff_plain/4ac0b2953131d7a53562ab8918c8e5a49952d8ac?hp=7bd0d536442493a8898981a8a784bf16126085c5

git: Manually decode git output from utf-8, avoids warning messages on invalidly encoded output.
---

diff --git a/IkiWiki/Plugin/git.pm b/IkiWiki/Plugin/git.pm
index 12f3a74cb..b386ab71b 100644
--- a/IkiWiki/Plugin/git.pm
+++ b/IkiWiki/Plugin/git.pm
@@ -136,14 +136,16 @@ sub safe_git (&@) {
 	}
 	# In parent.
 
+	# git output is probably utf-8 encoded, but may contain
+	# other encodings or invalidly encoded stuff. So do not rely
+	# on the normal utf-8 IO layer, decode it by hand.
+	binmode($OUT);
+
 	my @lines;
 	while (<$OUT>) {
+		$_=decode_utf8($_, 0);
+
 		chomp;
-		
-		# check for invalid utf-8, and toss it back to avoid crashes
-		if (! utf8::valid($_)) {
-			$_=encode_utf8($_);
-		}
 
 		push @lines, $_;
 	}
diff --git a/debian/changelog b/debian/changelog
index a7e7544c6..77b5d9fa6 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,6 +1,8 @@
 ikiwiki (3.08) UNRELEASED; urgency=low
 
   * git: Fix utf-8 encoding of author names.
+  * git: Manually decode git output from utf-8, avoids
+    warning messages on invalidly encoded output.
 
  -- Joey Hess <joeyh@debian.org>  Mon, 09 Mar 2009 14:00:21 -0400