[help] update vgfont/font.c to be compatible with utf-8

[nagualcode]

I am trying to render fonts on the framebuffer, and I would ask for some help with the path to make vgfont/font.c compatible with utf-8.

[paolosabatino]

it's not part of openvg beingcompatible with utf-8, openvg deals only with the rendering of the glyphs and provides a simple indexing of the path glyphs, but the decoding of the text string is not the purpose of the library

[rodizio1]

I think this is not Raspberry firmware related.

You might want to check out Paeryn's fork of ajstark's OpenVG library.
The latest branch is the "windowsave" branch: https://github.com/paeryn/openvg/tree/windowsave

There is a test program included that displays every character found in a font, it also mentions UTF-8. You need to set locales in your system and inside your program accordingly.
https://github.com/paeryn/openvg/blob/windowsave/client/chars.c

[nagualcode]

On https://github.com/raspberrypi/firmware/blob/master/opt/vc/src/hello_pi/libs/vgfont/font.c
There is:
line48 # * This API doesn't support unicode at all at present, nor UTF-8.

hello_font example wont render latin characters.

@popcornmix

[paolosabatino]

@nagualcode no need to importunate popcornmix for this, that example is not part of OpenVG: it is an example that uses OpenVG.
As an exercise you may want to implement your own utf-8 decoder and apply it to that example

[nagualcode]

Agree...
Well thanks.
I am closing the issue.
It is beyond my knowledge to code my own utf8 decoder.
If anyone reading this in the future is able to, I got this help from Paeryn:

"the original libshapes.c has a routine that converts a character at a time from a utf-8 string into the individual 32-bit integers needed (next_utf8_char()).

For the branches from newfonts upwards I changed it to use the standard mbsrtowcs() function (multi-byte string to wide-char string) so whatever format the current C_locale is set to is used (utf-8, utf-16, etc.).