From 348ed379209aacfe6c1f870a3a68e06b39a51f06 Mon Sep 17 00:00:00 2001
From: "Bryn M. Reeves" <bmr@redhat.com>
Date: Tue, 9 Dec 2014 17:30:06 +0000
Subject: [PATCH 21/93] [global] make all utf-8 handling use errors='ignore'

Stop playing whack-a-mole with unicode handling and ignore all
invalid characters. It's not really possible to ensure that we
always get strict utf-8 data from the system - e.g. dmesg on
systems with broken BIOS strings will often spit undecodable
byte sequences.

Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
---
 sos/archive.py          | 4 ++--
 sos/plugins/__init__.py | 5 +++--
 sos/utilities.py        | 7 +++++--
 3 files changed, 10 insertions(+), 6 deletions(-)

diff --git a/sos/archive.py b/sos/archive.py
index 6063625..0e019bf 100644
--- a/sos/archive.py
+++ b/sos/archive.py
@@ -406,9 +406,9 @@ class TarFileArchive(FileCacheArchive):
                           close_fds=True)
                 stdout, stderr = p.communicate()
                 if stdout:
-                    self.log_info(stdout.decode('utf-8'))
+                    self.log_info(stdout.decode('utf-8', 'ignore'))
                 if stderr:
-                    self.log_error(stderr.decode('utf-8'))
+                    self.log_error(stderr.decode('utf-8', 'ignore'))
                 self._suffix += suffix
                 return self.name()
             except Exception as e:
diff --git a/sos/plugins/__init__.py b/sos/plugins/__init__.py
index b289144..4fd85be 100644
--- a/sos/plugins/__init__.py
+++ b/sos/plugins/__init__.py
@@ -542,7 +542,7 @@ class Plugin(object):
     def add_string_as_file(self, content, filename):
         """Add a string to the archive as a file named `filename`"""
         self.copy_strings.append((content, filename))
-        content = "..." + (content.splitlines()[0]).decode('utf8')
+        content = "..." + (content.splitlines()[0]).decode('utf8', 'ignore')
         self._log_debug("added string '%s' as '%s'" % (content, filename))
 
     def get_cmd_output_now(self, exe, suggest_filename=None,
@@ -611,7 +611,8 @@ class Plugin(object):
 
     def _collect_strings(self):
         for string, file_name in self.copy_strings:
-            content = "..." + (string.splitlines()[0]).decode('utf8')
+            content = "..."
+            content += (string.splitlines()[0]).decode('utf8', 'ignore')
             self._log_info("collecting string '%s' as '%s'"
                            % (content, file_name))
             try:
diff --git a/sos/utilities.py b/sos/utilities.py
index 7e8cd7e..8cb4ed6 100644
--- a/sos/utilities.py
+++ b/sos/utilities.py
@@ -140,7 +140,7 @@ def sos_get_command_output(command, timeout=300, runat=None):
 
     # shlex.split() reacts badly to unicode on older python runtimes.
     if not six.PY3:
-        command = command.encode('utf-8')
+        command = command.encode('utf-8', 'ignore')
     args = shlex.split(command)
     try:
         p = Popen(args, shell=False, stdout=PIPE, stderr=STDOUT,
@@ -159,7 +159,10 @@ def sos_get_command_output(command, timeout=300, runat=None):
     if p.returncode == 126 or p.returncode == 127:
         stdout = six.binary_type(b"")
 
-    return {'status': p.returncode, 'output': stdout.decode('utf-8')}
+    return {
+        'status': p.returncode,
+        'output': stdout.decode('utf-8', 'ignore')
+    }
 
 
 def import_module(module_fqname, superclasses=None):
-- 
1.9.3