In the past we used to require 8-bit encodings for translations (typically iso-8859-#), with corresponding localised bds fonts in the themes to support those translations.
When we switched the translations encoding to Unicode, some leftovers were left behind, most notably all the localised bds fonts that were no longer used, but still bloated the themes.
This pull request:
The main benefit is that the theme files are noticeably smaller. But this is simplifies a bit the
I tried to do the changes step by step in a logical order so that each commit is easy to understand. This should also help to find the issue if this introduces regressions.
I am not completely confident with all the changes, so I would appreciate another pair of eyes looking at this before this gets merged.
The text was updated successfully, but these errors were encountered:
We used to have different charsets depending on the language and a different set of fonts for each one. So themes could support some language and not others depending on the fonts it contained. Now all languages are using UTF-32, except English. So the only case that can fail is when not using English and the Theme does not have TTF fonts.
This was only used to be part of the name under which loaded TTF fonts are stored in the FontMan font map. But since all the languages now use the same charset (UTF-32), there is no longuer a need to include the charset in the name. We still need to differentiate between bdf fonts that can only be used for English and TTF fonts, but this is is the case due to the font size being included in the name. And otherwise we could just add a hardcoded "-ttf" in the name.
This function supported loading alternative bdf fonts when we were using 8 bits encodings for the translations. Now that we are using unicode for all language, this is no longer needed.
Now all translations use unicode.