Non-western characters in a relevance

Greetings, I’m looking for some guidance on how I can create relevance using non-western characters. I can create relevance in the fixlet debugger that uses non-Western characters that works as I would expect.

exists (values "DisplayName" whose ( (it as string as uppercase = "CITRIX PRESENTATION SERVER クライアント") OR (it as string as uppercase = "CITRIX PRESENTATION SERVER 客户端") OR (it as string as uppercase = "CITRIX PRESENTATION SERVER 用戶端") OR (it as string as uppercase = "КЛИЕНТ CITRIX PRESENTATION SERVER")) of it) of keys of keys "HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstall" of (if x64 of operating system then (x32 registry ; x64 registry) else registry)

However, when I put this into the relevance of a fixlet I get the following error. We are running version 9.5.8

Is there a way to format my relevance so that is can evaluate from a fixlet/task successfully?

Actionscript relevance is a bit more limited than regular relevance in that it is presuming to interact with the “local” codepage and thus has to be limited to that set. This is why this apparent correct relevance is having an issue as the Japanese character does not have a representation in windows-1252

This can also occur even with answers from relevance (which in 9.5 is full UTF8) as the answers may not be in the correct encoding. This decision was made to keep backwards compatibility with the 9.2 and earlier releases and eventually will be corrected by enhancements. Some have already been made with action uses file encoding command added in 9.5.7

https://developer.bigfix.com/action-script/reference/execution/action-uses-file-encoding.html

So, This fixlet worked fine when it was 9.2. We only encountered a problem when we tried to recreate the fixlet after upgrading to 9.5. My problem is when creating the fixlet from scratch or importing the fixlet in 9.5 . How could you anticipate what code pages are loaded at the endpoint?

Is this the expected behavior? Or is this a bug?

This is a expected behavior.
With BigFix 9.5, you can only use characters belong to character set of BigFix server.
See “BigFix interfaces:” part of https://www.ibm.com/support/knowledgecenter/en/SS6MCG_9.5.0/com.ibm.bigfix.doc/Platform/Installation/c_limitations.html.

I think, with 9.2, the language of server, console and client are supposed to be the same and it just happened to work.

As workaround, you can create a UTF-8 encoded file with values in languages and refer line(s) of that file instead of literal.