Remove redundant parameter of get_next_char

[the-be-to]

The parameter agindex in get_next_char is not used

[HinTak]

While I agree with you this API is poorly implemented - the parameter is a "pass by reference" OUTPUT parameter, changing API with a long history seems Risky.

[the-be-to]

While I agree with you this API is poorly implemented - the parameter is a "pass by reference" OUTPUT parameter, changing API with a long history seems Risky.

In the current implementation agindex is immediately reassigned so it doesn't even function as an output parameter here. Also, it seems inconsistent to have such a parameter in get_next_char as FT_Get_First_Char takes a similar parameter but its wrapper get_first_char abstracts it and returns the value directly.

[HinTak]

It seems to be returned as "(return, agindex)", I think? I don't disagree with the idea of the pull, but an API change of a frequently used API needs a bit more thought.

[the-be-to]

Both FT_Get_First_Char and FT_Get_Next_Char return a character code directly and a glyph index via a pointer parameter agindex. The wrappers get_first_char and get_next_char indeed return a tuple of the two values, so the parameter agindex would not be needed in either function, but it is nevertheless present in get_next_char, while it is absent in get_first_char. Aside from that inconsistency, the first line of get_next_char assigns a new FT_UInt to agindex, so whatever was passed in that parameter is lost and the subsequent call to FT_Get_Next_Char modifies a new object, leaving the original argument unchanged.