PythonEngine.PyUnicode_FromString() issue

[Jacek Laskowski 57]

If the data variable contains any character outside the Latin alphabet, such as a Polish diacritical chars, the function returns nil

 

 data: AnsiString = 'ąęść';


  pObj := fPyEngine.PyUnicode_FromString(PAnsiChar(data)); <-- return nil
  fPyModule.SetVar('data', pObj); <-- then this raises an exception

How to properly pass diacritical chars in a string?

 

ps. the function has Unicode in its name, which means it should probably allow arbitrary characters

Edited August 9 by Jacek Laskowski

[David Heffernan 2463]

Why are you expecting ANSI encodings to be of use? Why aren't you using Unicode, either UTF8 or UTF16. 

 

I believe the function you call expects UTF8 which of course you are not passing. Then again I'm not sure why you picked on this function to make a string. 

Edited August 9 by David Heffernan

[pyscripter 828]

P4D provides a function:

 

function TPythonEngine.PyUnicodeFromString(const AString : UnicodeString) : PPyObject;

[Jacek Laskowski 57]

21 hours ago, David Heffernan said:

Why are you expecting ANSI encodings to be of use? Why aren't you using Unicode, either UTF8 or UTF16. 

 

I believe the function you call expects UTF8 which of course you are not passing. Then again I'm not sure why you picked on this function to make a string. 

I passed what the function expected:

 

    PyUnicode_FromString:function (s:PAnsiChar):PPyObject; cdecl;

 

Anyway, unfortunately the module has no documentation, so I also rely on examples from this forum, SO and demos in my experiments. I didn't know that there is a second analogous function with an almost identical name.

 

@pyscripter Thx

 

I am slowly learning, admittedly, the additional difficulty is the lack of knowledge of Python, but AI helps with this.

Edited August 10 by Jacek Laskowski

[David Heffernan 2463]

3 hours ago, Jacek Laskowski said:

unfortunately the module has no documentation

This is a Python API function which is documented. And the documentation states UTF8. You won't see encodings other than UTF8 and UTF16 in the API.