From d54c86abfba35f48cad78bb99b1d199afe8c2095 Mon Sep 17 00:00:00 2001 From: Jeff McCune Date: Tue, 16 Apr 2013 11:27:30 -0400 Subject: [PATCH] (#14522) Force /proc/self/status encoding to valid UTF-8 (try 2) Without this patch the previous attempt to fix issue 14522 is insufficient because the String#encode method is a no-op when the source and destination encodings are the same encoding, even if there are invalid byte sequences. This is a problem because we're still getting unhandled `invalid byte sequence in UTF-8` errors running the specs. This patch addresses the problem by changing encoding from UTF-8 to UTF-16, replacing all invalid byte sequences with the default unicode string of "uFFFD" We then convert back to UTF-8 to guarantee only valid byte sequences remain. --- lib/facter/util/virtual.rb | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/lib/facter/util/virtual.rb b/lib/facter/util/virtual.rb index 8223b41..65cf58a 100644 --- a/lib/facter/util/virtual.rb +++ b/lib/facter/util/virtual.rb @@ -63,7 +63,8 @@ def self.vserver? return false unless FileTest.exists?("/proc/self/status") txt = File.open("/proc/self/status", "rb").read if txt.respond_to?(:encode!) - txt.encode!('UTF-8', 'UTF-8', :invalid => :replace) + txt.encode!('UTF-16', 'UTF-8', :invalid => :replace) + txt.encode!('UTF-8', 'UTF-16') end return true if txt =~ /^(s_context|VxID):[[:blank:]]*[0-9]/ return false -- 1.8.1.6