[META] cycle handling

[lcnr]

• vibe: negative reasoning and solver cycles #122

fixpoint iteration

• trait solver hangs for recursively expanding coinductive cycles #13
• unavoidable overflow due to cycle fixpoint + region constaints #118

initial provisional result

• ambiguity on inductive cycles #20
• inductive cycles as ambig causes unintended breakage #114

undefined behavior uwu

• associated_type_defaults are unsound in the new solver #46
• Unsound associated type outlives bound due to lazy norm #59
• Associated type bounds can be used to prove themselves with trivial where-clauses #62

new cycle semantics

• AliasBound candidates for normalizeable aliases are unsound #6
• projection obligations were accepted in - otherwise coinductive - global cycles #10
• requiring item bounds for coinduction may be breaking #80
• old trait solver provisional cache does not handle changes in the cycle kind #119

proving opaque type item bounds

• normalizing an opaque while proving its item bounds causes overflow errors #143