Pascal case primitives

[adamski]

This properly addresses our use case where we want to the default of e.g. MyType______c_char to be MyTypeCChar. The previous PR did not take the c_char type or similar into account, only capitalising the first letter for e.g. MyTypeU32.

[adamski]

So I was thinking about this, and my feeling is that the previous (now merged) PR that allows for different separators, is possibly not that useful (at least I can't think of a use case where someone would want a different character to _).

What we ultimately want is to be able to PascalCase generated types/typedefs. So my proposal is to merge the two (remove underscores and PascalCase primitives), to something like rename_types = "PascalCase". Perhaps a remove_underscores bool in addition to a rename_types rule?

[emilio]

remove_underscores + rename_types would be the same as mangling_separator = "" and rename_types, right? If so, I think there's no need to change the existing option unless you feel really strongly about it.

[adamski]

remove_underscores + rename_types would be the same as mangling_separator = "" and rename_types, right?

Yes, although I'm also suggesting they could be merged into one option.

If so, I think there's no need to change the existing option unless you feel really strongly about it.

OK, I'm just struggling to think of a use case where someone would want a different mangle_seperator to _...

[adamski]

@emilio please review the updated changes when you have time. Thank you.

[emilio]

This looks generally good, but I have some nits and one error in to_str. I'm still not a fan of removing an existing option for mangle_separator somewhat arbitrarily, but I guess it's relatively new so it can be fine.

Can you squash the commits too?

[emilio]

Thanks! This looks fine. I have a few nits but I addressed them in a follow-up commit in #575.

In particular, the rename_types should probably be on its own subsection of the configuration, otherwise being under [export] seems like you'd rename all the types in here which is not what you're doing. I also added some docs and some integration tests.

[emilio]

Closing because I merged this in #575