[META] intended breakage

[lcnr]

We're gonna break some crates because there's no good way to support them. I want backports to literally all affected versions.

• Inference hazard due to lazy alias relate: Break it in the old solver? #168
  • bevy ✅ remove reliance on a trait solver inference bug bevyengine/bevy#18840 backport to 0.15: remove reliance on a trait solver bug bevyengine/bevy#19360
  • minijinja 🛠
    • 2.10.2 fix regression with the next-generation trait solver mitsuhiko/minijinja#787
    • 1.0.22 Backport of #787 mitsuhiko/minijinja#790
    • other affected versions found on crater: 0.23 0.27 0.30 0.32 0.24
  • https://crater-reports.s3.amazonaws.com/pr-133502-13/try%233cddd79d4be3c8f1e937d09ca618d192abc69373/gh/LunaAmora.campfire/log.txt
    • remove reliance on old trait solver bug LunaAmora/campfire#1
  • https://crater-reports.s3.amazonaws.com/pr-133502-13/try%233cddd79d4be3c8f1e937d09ca618d192abc69373/gh/navneetankur.bhecs/log.txt
    • unfixed
  • https://crater-reports.s3.amazonaws.com/pr-133502-13/try%233cddd79d4be3c8f1e937d09ca618d192abc69373/gh/Eddieg26.rouge/log.txt
    • unfixed
• uncertain: local projection cache strenghens inference #215
  • stop relying on type inference edge case to support -Znext-solver dyedgreen/uncertain#11

env normalization, rigid normalizable aliases

• (maybe) alias-relate when normalizing in unnormalized env causes overflow #89
  • tachys
    • 0.1
    • 0.2 remove unnecessary where-clauses leptos-rs/leptos#4023
• (maybe) fixed_map: derive(Key) treats alias as rigid #216

candidate preference behavior

• old solver incompletely prefers projection and object candidates #27