Currently, translations are handled by changing what symbol phonetic characters map to. This limits the characters we can output to those in the symbol map.
The plain text in some dialogues (words such as "Cathedral" or "Danger" which are shown in Latin script) seems to use a normal font (see LocalizationFontSet). It should be possible to override MixedTextBuilder::breakLineIntoSegments to return only localizable text elements where normal text is prefixed by a special indicator, then overriding the Localization::localizationDictionary getter so the returned dictionary contains the special localization text.
Currently, translations are handled by changing what symbol phonetic characters map to. This limits the characters we can output to those in the symbol map.
The plain text in some dialogues (words such as "Cathedral" or "Danger" which are shown in Latin script) seems to use a normal font (see
LocalizationFontSet). It should be possible to overrideMixedTextBuilder::breakLineIntoSegmentsto return only localizable text elements where normal text is prefixed by a special indicator, then overriding theLocalization::localizationDictionarygetter so the returned dictionary contains the special localization text.