Bug Rendering UNICODE in XBMC/KODI Incorrectly
#25
(2021-05-22, 12:10)deadrat Wrote: I came to the forum with the same issue as the user @Cyberrule

I tried installing several Tamil Fonts and nothing seems to fix the issue. 

Could it have something to do with the Character Encoding? 

Under character set in Kodi settings, there is no option for me to select "Full Unicode" which is listed in this page


But I have no idea about the technicalities. :-( 

If I remember correctly, this issue with rendering Tamil and Malayalam texts used to happen with many different applications (such as XFCE, LXDE, Chromium, etc,) but that was almost a decade ago. 

I hope it gets fixed soon.

The technicality is that some scripts are considered "complex".  For these scripts it is insufficient to simply look up a unicode codepoint in a font file and get the glyph(s) and render them.  Complex scripts need to have a set of rules for rendering characters, and to support these rules font files need to have 2 open-type font tables, GSUB and GPOS, which the rules use to modify or substitute the glyphs pulled from the font file and then rendered.

Kodi uses the "freetype" open source library to access font files.  In the past freetype had some capability to process complex scripts, but a decision was taken by the freetype devs to remove the code for dealing with complex scripts.  That code forms the basis for a second open source project that goes by the name of HarfBuzz.  Kodi does not currently use the HarfBuzz library.  The good news is that there is development underway to add HarfBuzz to Kodi and use it to shape complex scripts.  The bad news is this is a Kodi version 20 project so if/when it gets completed it will only be available in Kodi nightly builds until such time as Kodi 20 is released.

scott s.
.
Reply


Messages In This Thread
RE: Rendering UNICODE in XBMC/KODI Incorrectly - by scott967 - 2021-05-23, 11:55
Logout Mark Read Team Forum Stats Members Help
Rendering UNICODE in XBMC/KODI Incorrectly0