This is a change I've noticed in versions

2.27.0 WIP

2.29.1 WIP

The difference I am having in these two is that in

2.27.0 WIP

I can register a 16bit char string class and it works perfectly.

in

2.29.1 WIP

registering the same string factory, I am getting odd results of the strings. Lots of asian characters.

Log(CString &)

Log("Testing Script");

gives

2.27.0 WIP = Testing Script

2.29.1 WIP= Testing Script???

Log("Testing Scrip");

gives

2.27.0 WIP = Testing Scrip

2.29.1 WIP= Testing Scrip

hmm, actually found that it works fine up till 13 letters. 8bit chars works as far as I can put them.

Edit: Somehow adding " / " or " \n " or " n " after the first letter or any letters as long as its not the first sign in I am able to put any number of letters after.

I should probably add that this is just a version of std::string with wchar based on the stdstring addon

Edit2: seems adding those extra / \n or n causes the next strings created to mess up. :S

I'm not sure how to address your issue. I have the most current version of AngelScript (revision 1988), and I don't experience any trouble with my 16-bit string literals.

I specify the 16-bit encoding:

engine->SetEngineProperty(asEP_STRING_ENCODING, 1);

I use MFC's CString class as my string class, and I register the string factory as follows:

CString StringFactory(asUINT length, const wchar_t* s)
{
    return CString(s, (int)(length/2));
}

r = engine->RegisterStringFactory("String", asFUNCTIONPR(StringFactory, (asUINT, const wchar_t*), CString), asCALL_CDECL);

yeah, forgot about the length/2 sadly I didn't have that in my constructor so I figured I would just do