I’m using(what seems to be) an ansi(or ascii??) dll library. I think it is

Question

0

Asked: June 18, 20262026-06-18T05:06:25+00:00 2026-06-18T05:06:25+00:00

I’m using(what seems to be) an ansi(or ascii??) dll library. I think it is

0

I’m using(what seems to be) an ansi(or ascii??) dll library. I think it is such because the header file provided with the lib shows function using char*’s and LPSTR and LPCSTR and structs with char arrays.

This dll is loaded via ::LoadLibrary from a cpp/cli class library that wraps its functionality and exposes it to c#. A c# console app and various other class libs use this cli lib to perform operation.

I can make the cli assembly ether mutibyte or Unicode(which as far as I understand is the same in terms of language support) and c# apps are always Unicode.

This native dll is essentially a broker for a propriety back end server, it passes information back and froth from and to the server.

The issue I’m running into is that the native dll lib will only operate correctly for a particular language if the os locale, for none Unicode apps, its running in is set to that particular language.
I.e. if i want the app to correctly work with Chinese characters, that locale needs to be set. What I find hard to grasp is why does the locale matter for the broker. I understand that if the server is an ansi app if a user wanted to store none Unicode Chinese on it setting the locale on the server would make sense so it would in the client, but not in the middle man that just passes things along. Furthermore the whole thing is getting very confusing.

Is there away to pass Unicode to something like a char array in c++? Wold that even work in this scenario?

Here’s a scenario I’m thinking about:

c# app gets url encoded string
c# app decodes the string and passes it to cli
cli somehow converts the String^(or should it be byte[] at
this point) to char[] and passes it to the native lib

Should this really be possible? In terms of memory layout it should, i mean char is just a byte no?

Am I approaching this the right way? is there there a better way to accomplish cross language support. Mind you the vendor is on record saying that there is no way to mix languages in the api, but that’s not what I’m looking for. I just dont want to have to run an instance of the software on a separate os for each language i want to support.