How to Output Unicode Strings on the Windows Console

Question

there are already a few questions relating to this problem. I think my question is a bit different because I don\'t have an actual problem, I\'m only asking out of academic

Answer 1

To answer your first question, you can output Unicode strings to the Windows console using _setmode. Specific details regarding this can be found on Michael Kaplan's blog. By default, the console is not Unicode (UCS-2/UTF-16). It works in an Ansi (locale/code page) manner and must specifically be configured to use Unicode.

Also, you have to change the console font, as the default font only supports Ansi characters. There are some minor exceptions here, such as zero-extended ASCII characters, but printing actual Unicode characters requires the use of _setmode.

In Windows, everything is UTF-16. Regardless of whether you're dealing with the kernel, the graphics subsystem, the filesystem or whatever, you're passing UTF-16 strings. There are no locales or charsets in the Unix sense.

This is not completely true. While the underlying core of Windows does use Unicode, there is a huge amount of interoperability that comes into play that lets Windows interact with a wide variety of software.

Consider notepad (yes, notepad is far from a core component, but it gets my point across). Notepad has the ability to read files that contain Ansi (your current code page), Unicode or UTF-8. You might consider notepad to be a Unicode application, but that is not entirely accurate.

A better example is drivers. Drivers can be written in either Unicode or Ansi. It really depends on the nature of the interface. To further this point, Microsoft provides the StrSafe library, which was specifically written with Kernel-mode drivers in mind, and it includes both Unicode and Ansi versions. While the drivers are either Ansi or Unicode, the Windows kernel must interact with them - correctly - regardless of whatever form they take.

The further away you get from the core of Windows, the more interoperability comes into play. This includes code pages and locales. You have to remember that not all software is written with Unicode in mind. Visual C++ 2010 still has the ability to build using Ansi, Multi-Byte or Unicode. This includes the use of code pages and locales, which are part of the C/C++ standard.

However, I think this represents a design flaw in the Windows API

the following two articles discuss this fairly well.

• Conventional wisdom is retarded, aka What the @#%&* is _O_U16TEXT?
• Header files are not retarded, aka What the @#%&* is _O_WTEXT?

So my questions are: What to do in this situation? And why isn't this problem solved even in Microsoft's own libraries? Both the .NET Framework and the C and C++ libraries seem to adhere to the obsolete codepage model. How would you design the Windows API or an application framework to circumvent this issue?

On this point, I think you are looking at Windows in hindsight. Unicode did not come first, ASCII did. After ASCII, came code pages. After code pages, came DBCS. After DBCS came MBCS (and eventually UTF-8). After UTF-8, came Unicode (UTF-16/UCS-2).

Each of these technologies was incorporated into the Windows OS over the years. Each building on the last, but without breaking each other. Software was written with each of these in mind. While it may not seem like it sometimes, Microsoft puts a huge amount of effort into not breaking software it didn't write. Even now, you can write new software that takes advantage of any of these technologies and it will work.

The real answer here is "compatibility". Microsoft still uses these technologies and so do many other companies. There are an untold number of programs, components and libraries which have not been updated (or ever will be updated) to use Unicode. Even as newer technologies arise - like .NET - the older technologies must stick around. At the very least for interoperability.

For example, say you have a DLL that you need to interact with from .NET, but this DLL was written using Ansi (single byte code page localized). To make it worse, you don't have the source for the DLL. The only answer here is to use those obsolete features.