fix where clauses in type aliases and allow empty where clauses

[m4rch3n1ng]

according to the reference, type aliases can also have a where_clause in front of the = in addition to one after it. this is fixed in 42dce44.

and also according to the reference, where clauses can also be empty. i didn't just change the sepBy1 to a sepBy, because that would allow a trailing comma without any rules, which is not allowed. this is fixed in f2be460.

diff for #229:

this is gonna be so diff trust me

--- .tmp/list.txt	2025-02-28 16:16:12.002933806 +0100
+++ .tmp/list.txt.type-alias-where-before-assignment.optional-where-clause	2025-03-01 01:13:03.203344726 +0100
@@ -21,7 +21,6 @@
 tests/mir-opt/coroutine_tiny.rs
 tests/mir-opt/tail_call_drops.rs
 tests/pretty/delegation.rs
-tests/pretty/issue-25031.rs
 tests/pretty/macro.rs
 tests/pretty/stmt_expr_attributes.rs
 tests/pretty/yeet-expr.rs
@@ -42,7 +41,6 @@
 tests/ui/associated-type-bounds/return-type-notation/path-constrained-in-method.rs
 tests/ui/associated-type-bounds/return-type-notation/path-self-qself.rs
 tests/ui/associated-type-bounds/return-type-notation/path-works.rs
-tests/ui/associated-type-bounds/type-alias.rs
 tests/ui/associated-types/default-associated-types.rs
 tests/ui/associated-types/issue-25339.rs
 tests/ui/associated-types/issue-32350.rs
@@ -145,7 +143,6 @@
 tests/ui/feature-gates/feature-gate-trivial_bounds-lint.rs
 tests/ui/for-loop-while/loop-break-value.rs
 tests/ui/generic-associated-types/collections.rs
-tests/ui/generic-associated-types/issue-93341.rs
 tests/ui/generic-const-items/assoc-const-AnonConst-ice-108220.rs
 tests/ui/generic-const-items/associated-const-equality.rs
 tests/ui/generic-const-items/basic.rs
@@ -175,7 +172,6 @@
 tests/ui/inline-const/const-match-pat-inference.rs
 tests/ui/inline-const/const-match-pat-range.rs
 tests/ui/inline-const/pat-unsafe.rs
-tests/ui/issues/issue-22471.rs
 tests/ui/issues/issue-26186.rs
 tests/ui/issues/issue-36116.rs
 tests/ui/issues/issue-37051.rs
@@ -313,11 +309,9 @@
 tests/ui/try-block/try-is-identifier-edition2015.rs
 tests/ui/try-trait/yeet-for-option.rs
 tests/ui/try-trait/yeet-for-result.rs
-tests/ui/type-alias-impl-trait/closure_parent_substs.rs
 tests/ui/type-alias-impl-trait/issue-57611-trait-alias.rs
 tests/ui/type-alias-impl-trait/issue-58662-coroutine-with-lifetime.rs
 tests/ui/type-alias-impl-trait/issue-58662-simplified.rs
-tests/ui/type/type-alias-bounds.rs
 tests/ui/type_length_limit.rs
 tests/ui/typeck/auxiliary/tdticc_coherence_lib.rs
 tests/ui/underscore-imports/hygiene-2.rs
@@ -325,5 +319,3 @@
 tests/ui/underscore-method-after-integer.rs
 tests/ui/unsafe-binders/expr.rs
 tests/ui/unsafe-binders/simple.rs
-tests/ui/where-clauses/where-clause-placement-assoc-type-in-impl.rs
-tests/ui/where-clauses/where-clause-placement-assoc-type-in-trait.rs

[m4rch3n1ng]

rebased this and the other pr

[maxbrunsfeld]

Sorry you have to rebase so much. At some point, we'll change the workflow so generated files are not checked in, at least on the default branch. Right now it kind of sucks for people who are making lots of PRs.

[m4rch3n1ng]

lol it's fine, don't have much to do right now anyway and it's not actually that annoying (edit: i'm more worried about your ci minutes lol)