You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It looks like the constructor names for LLVM values and types in the python interface don't really have a consistency to them =\ IMO we should address this soon and then stick to whatever naming pattern we prefer.
Options:
Use a consistent suffix for each, e.g., NAME_val and NAME_ty (pros: simple, predictable; cons: a few additional characters here and there)
Use no suffixes, but places values and types in modules which can be imported with whatever prefix is desired (pros: flexible, users decide prefix to import with (if any); cons: val.NAME and ty.NAME might look funny, if no prefix is present from the import it may be difficult to determine meaning from code).
The text was updated successfully, but these errors were encountered:
I think (1) is closest to my ideal, although I don't think we need to use the suffixes for every definition. For instance, I think global_var is fine as it is, and global_var_val would just be unnecessarily redundant.
It looks like the constructor names for LLVM values and types in the python interface don't really have a consistency to them =\ IMO we should address this soon and then stick to whatever naming pattern we prefer.
Options:
NAME_val
andNAME_ty
(pros: simple, predictable; cons: a few additional characters here and there)val.NAME
andty.NAME
might look funny, if no prefix is present from the import it may be difficult to determine meaning from code).The text was updated successfully, but these errors were encountered: