Both tools use NULL for blanks right now. They should use the blanks from the default config I suppose.
I think originally I liked the idea of the tools being config-independent. But that is broken because the computed languages are incorrect this way.
should we add the kind of --ignore-blanks for the backward compatibility? I don't know how many people depends on it though.
--ignore-blanks might be a good idea. Though, perhaps wait on this, we might want to hardcode blanks and remove them from config completely.
Going back here. the targeted API is something in fcfreetype.c though, those doesn't get involved with FcConfig which has the hardcode blanks. so it's not possible to do it inside APIs.
I still think adding an option to get back the behavior might be a good idea for this.
Added --ignore-blanks option to fc-query and fc-scan. observing blanks by default after this change:
Shall we close then?