@WG-prioritization/alerts issue #80706 has been requested for prioritization.
Seems like maybe P-high? Regressed in 1.48. Working code -> ICE
If that code is supposed to compile, then yes definitively. I'd love to see more prominently mentioned somewhere if the expected behaviour is the have code compiling (there is a slight mention at the beginning of this comment), that would raising red flags at a glance.
we prioritize a lot of ICEs as p-mediums with broken code showing weird errors
(in this case, a p-high is probably more correct)
It would be nice if the example were even more minimal. It's hard to see how many different parts need to come together to trigger the issue making it hard to prioritize
the "moving parts" of a reproducible example are the fun of trying to minimize a sample. Touching bits here and there can trigger different behaviour (as comments show). my understanding is that doesnt look easy to shrink down to few lines.
It would really interesting to have mini-workshops or live streams while hunting for mcve on issues :)
Issue #80706's prioritization request has been removed.