TEXTINPUT does not recognize some keys in the layout.

[CicTec]

Hi,

It seems that TEXTINPUT does not correctly recognize some keys with particular symbols or accented letters.

Here is the image of the keyboard and layout used (Italian):

The characters marked in red are returned with code -61/-62 or 194/195 bitted as unsigned char.

Some examples:

• To obtain the character "£" you must press SHIFT + the key
• To obtain the character "è" just press the relative key
• To obtain the character "é" you must press SHIFT + the key
• To obtain the character "[" you must press ALT GR + the key

Below is the output of the test:

Below is the source of the test:

SDL_Event e;

SDL_StartTextInput();

while(SDL_PollEvent(&e)) {
  switch(e.type) {
    case SDL_TEXTINPUT: {
      printf("[INPUT]: e.text.text[0]:             %ld\n", e.text.text[0]);
      printf("[INPUT]: e.text.text[0]:             %c\n",  e.text.text[0]);
      printf("[INPUT]: e.text.text[0] & 0xFF: %ld\n", e.text.text[0] & 0xFF);
      printf("[INPUT]: e.text.text[0] & 0xFF: %c\n",  e.text.text[0] & 0xFF);
      break;
    }

    case SDL_QUIT: {
      goto MainLoop_Finalize;
    }
  }
}
MainLoop_Finalize:
  SDL_StopTextInput();

All the other keys respond correctly.

The version used for the test is SDL 2.30.8.

[po-87]

UTF-8 uses multiple chars to represent £, è and é.

Try printing the entire e.text.text as a string:

printf("%s\n", e.text.text);

[CicTec]

UTF-8 uses multiple chars to represent £, è and é.

Try printing the entire e.text.text as a string:

printf("%s\n", e.text.text);

I need to recover just the first character as in the example and the corresponding ASCII code, since all the characters indicated are within the first 256.

[po-87]

ASCII is a 7-bit encoding so it only goes up to 127. The ASCII characters are encoded the same way in UTF-8 (which is the encoding used for e.text.text) but for the rest it uses two or more chars to represent one character.

[CicTec]

The standard indicates that the first 256 characters must be those of ISO8859-1, the characters I indicated fall within this standard, so they should result up to code 255 (1 byte) all the others as you say are represented with 2 or more bytes.

Does SDL implement unicode support different from the standard?

[po-87]

The Unicode standard specifies that the first 256 code points are identical to ISO8859-1 but this doesn't say anything about how they are stored. That depends on the encoding. For the UTF-8 encoding, code points above 127 uses two or more bytes.

[slouken]

That's correct, the text event uses UTF-8 which is a multi-byte encoding. You can use SDL_StepUTF8() if you want to convert from UTF-8 to codepoints.

Also note that the text event might have multiple characters in it, all encoded in UTF-8.

[CicTec]

That's correct, the text event uses UTF-8 which is a multi-byte encoding. You can use SDL_StepUTF8() if you want to convert from UTF-8 to codepoints.

Also note that the text event might have multiple characters in it, all encoded in UTF-8.

Hi, thanks, let's see if the function is useful for the purpose.

[CicTec]

That's correct, the text event uses UTF-8 which is a multi-byte encoding. You can use SDL_StepUTF8() if you want to convert from UTF-8 to codepoints.
Also note that the text event might have multiple characters in it, all encoded in UTF-8.

Hi, thanks, let's see if the function is useful for the purpose.

This function It doesn't exist in version 2.30.8 that I use. Is there any other equivalent function that I can use?