A different Cyrillic character set should be used for Bulgarian localization. Bulgarian cyrillic has its own special characters and some of them looks like Latin script letters, so, I think, there is nothing wrong about the localization on the OS level.
In Bulgarian script, ‘д’ (de-cy) character equilavent is ‘de-cy.loclBGR’ which is shaped as / similar to Latin ‘g’ character. This is also true for ‘т’ (te-cy) character mentioned above. It should be te-cy.loclBGR character which is shaped as / similar to Latin ‘m’.
For more information, you may want to search elsewhere on the internet with the term 'Bulgarian Cyrillic Feature Locl' or similar.