From af1405ed41ca25bcc2967bcf97a0448d5cba208d Mon Sep 17 00:00:00 2001
From: tpowell-progress <104777878+tpowell-progress@users.noreply.github.com>
Date: Wed, 25 Jan 2023 06:13:19 -0800
Subject: Force encoding with timezones that include umlauts if zone encoding
 is IBM437 (#1781)

* If time.zone is encoded in Encoding::IBM437, force to WINDOWS_1252
* Added description of what's going on with the force/encode chain
* Handle umlauts in popen output from Mixlib shellout

Signed-off-by: Thomas Powell <powell@progress.com>
---
 lib/ohai/plugins/timezone.rb | 20 ++++++++++++++++++++
 lib/ohai/plugins/vmware.rb   | 15 +++++++++++++++
 2 files changed, 35 insertions(+)

diff --git a/lib/ohai/plugins/timezone.rb b/lib/ohai/plugins/timezone.rb
index 1d542e43..d2ab47dc 100644
--- a/lib/ohai/plugins/timezone.rb
+++ b/lib/ohai/plugins/timezone.rb
@@ -21,5 +21,25 @@ Ohai.plugin(:Timezone) do
   collect_data(:default) do
     time Mash.new unless time
     time[:timezone] = Time.now.getlocal.zone
+
+    # Windows in German display language outputs LATIN1 bytes for .zone, but marks them as
+    # IBM437, which somehow fails any attempt at conversion to other encodings when
+    # ä is present, as in the timezone name "Mitteleuropäische Zeit" (Central Europe Time)
+    #
+    # Windows-1252 is the legacy encoding for Windows for German that actually
+    # translates (ISO-8859-1 works as well), but going with the more correct
+    # encoding name for Windows' implementation of Latin-1
+    #
+    # References
+    # * [Code Page 437/IBM437](https://en.wikipedia.org/wiki/Code_page_437)
+    # * [ISO/IEC 8859-1](https://en.wikipedia.org/wiki/ISO/IEC_8859-1)
+    # * [Windows-1252](https://en.wikipedia.org/wiki/Windows-1252)
+    if time[:timezone].encoding == Encoding::IBM437
+      # Assume encoding is WINDOWS_1252
+      time[:timezone] = time[:timezone].force_encoding(Encoding::WINDOWS_1252)
+      # Re-encode in UTF_8. Note: If other encodings have problems converting
+      # it might be worth re-encode everything in UTF_8.
+      time[:timezone] = time[:timezone].encode(Encoding::UTF_8)
+    end
   end
 end
diff --git a/lib/ohai/plugins/vmware.rb b/lib/ohai/plugins/vmware.rb
index 250a9e24..438efb41 100644
--- a/lib/ohai/plugins/vmware.rb
+++ b/lib/ohai/plugins/vmware.rb
@@ -53,6 +53,21 @@ Ohai.plugin(:VMware) do
         # to attribute "vmware[:<parameter>]"
         %w{hosttime speed sessionid balloon swap memlimit memres cpures cpulimit}.each do |param|
           vmware[param] = from_cmd([vmtools_path, "stat", param])
+          if param == "hosttime" && vmtools_path.include?("Program Files")
+            # popen and %x return stdout encoded as IBM437 in Windows but in a string marked
+            # UTF-8. The string doesn't throw an exception when encoded to "UTF-8" but
+            # displays [?] character in Windows without this.
+            #
+            # .force_encoding(Encoding::ISO_8859_1) causes the character to be dropped
+            # and .force_encoding(Encoding::Windows_1252) displays the „ character in place
+            # of an ä. .force_encoding(Encoding::IBM437) allows for the correct characters
+            # to be displayed.
+            #
+            # Note:
+            # * this is broken for at least Ruby 2.7 through 3.1.3
+            # * confirmed that this is broken on Windows Server 2022
+            vmware[param] = vmware[param].force_encoding(Encoding::IBM437).encode("UTF-8")
+          end
           if /UpdateInfo failed/.match?(vmware[param])
             vmware[param] = nil
           end
-- 
cgit v1.2.1