@WG-prioritization/alerts issue #76507 has been requested for prioritization.
So, it looks like there is an undefined behaviour on AVR architecture (Tier 3) when linking in the same address space
data segments. By quickly glancing at the linked issues and the linked thread on the users forum, my (admittedly limited) understanding is that this is not something that be very easily "fixed", it's kind of how these architectures work. I think OP is just asking to put a sort of "safeguard". Happy to receive more insights on this issue.
In short, given there's also a list of linked issues around the same theme that seem not to be critical, I'd warrant a
P-medium (if not
it's kind of how these architectures work
The reason I tagged as I-unsound is because it lets you cause UB from safe rust. This is a property of the language, not of the architecture (it just happens to only occur on AVR). I'm ok with P-low based on the linked issues, but I would prefer to have it I-nominated I think so it has more eyes on it.
This is an issue on all architectures fwiw, but it's been known for a long time that linkage-controlling attributes can cause problems like this
speaking of which, it would be great if someone could review https://github.com/rust-lang/rust/pull/72209, it's been sitting alone for a month
Lint additions need to be T-lang approved AFAIK
seems a little odd, but I'll post in #t-lang
Joshua Nelson said:
I'm ok with P-low based on the linked issues, but I would prefer to have it I-nominated I think so it has more eyes on it.
completely agree :thumbs_up:
and thanks for correcting my (ignorant) take on the issue :-)
P-low, but feel free to adjust if needed
Issue #76507's prioritization request has been removed.