Consistent `Pallet` and `Module` naming

[shaunxw]

In pallet attribute macro, Module struct is renamed to Pallet, but there are still some places using the name Module. For instance in construct_runtime macro:

substrate/bin/node/runtime/src/lib.rs

Line 1080 in 88014d5

System: frame_system::{Module, Call, Config, Storage, Event<T>},

and DispatchError::Module etc:

substrate/primitives/runtime/src/lib.rs

Line 405 in 88014d5

Module {

They should be renamed to Pallet to make it consistent and less confusing.

[gui1117]

Currently pallet macro do write: type Module<T> = Pallet<T>; in order not to break construct_runtime.

Now we can update construct_runtime to use Pallet struct, make decl_module generate the type alias type Pallet<T> = Module<T>, and deprecated the type alias in pallet macro.

For other occurences of Module we need to renames types and fields.

[gui1117]

there is still DispatchError module variant which could be replaced to Pallet, but at some point DispatchError is not part of FRAME, thus I'm not sure it makes sense to rename.

There is also metadata types which says: ModuleMetadata and ModuleConstantMetadata, etc.. but it is pretty minor IMO

[shaunxw]

IMHO if they refer to the same thing, ideally the same term/naming should be used. I totally agree they are minor and low priority.

[kianenigma]

Both Metadata and DispatchError are none-frame concepts, so I guess it is fine to keep them as they are now.

Although for the metadata, it is reasonable to say that the Metadata of a frame based chain would internally use Pallet etc. But if ModuleMetadata are terms that any runtime would have, they should use Module.

[shaunxw]

I see, so "pallet" is only for inside FRAME, and outside FRAME it still is "runtime modules".