[BUG] alias: type == _ generates invalid cpp1

[realgdman]

Minimal reproduce
alias5: type == _;
passes succssfully cpp2, but generates invalid cpp1:
using alias5 = auto;

Command lines including which C++ compiler you are using
cppfront example.cpp2 -p
(build from latest mai. f83ca9b with clang++-15)
Then cpp1 compiled with
clang++-15 -Icppfront/include $1.cpp -std=c++20 -o $1

Actual result/error
Cppfront passes with
example.cpp2... ok (all Cpp2, passes safety checks)
But then clang fails with
error: 'auto' not allowed in type alias using alias5 = auto;

Expected result
Diagnostic at cpp2 side. (I guess that`s strategy, to catch whatever possible cpp2 relateed before relying on cpp1 compiler?)

Additional context
This is extreme case, because clang says auto is never allowed.
There is other case
alias6: type == foo when foo is not defined generates correct
using alias6 = foo;
But I guess this can be separate issue, for should cppfront check if alias goes to something undefined.

[JohelEGP]

But I guess this can be separate issue, for should cppfront check if alias goes to something undefined.

This is better left off to the C++1 compiler to diagnose. Otherwise, cppfront would have to be a C++1 compiler to know if it's ill-formed.

[filipsajdak]

We can diagnose cases like:

alias5: type == _;

[filipsajdak]

Change that provides diagnostics for a wildcard: #360

[realgdman]

Great.
BTW what do you think about cases like that? Currently they're passed successfully to cpp1 and fail there.
alias8: == ();
alias9: == (int);