Tracking Issue for __rust_no_alloc_shim_is_unstable

[Mark-Simulacrum]

About tracking issues

Tracking issues are used to record the overall progress of implementation.
They are also used as hubs connecting to other relevant issues, e.g., bugs or open design questions.
A tracking issue is however not meant for large scale discussion, questions, or bug reports about a feature.
Instead, open a dedicated issue for the specific matter and add the relevant feature gate label.

Steps

• Implement (done in Support #[global_allocator] without the allocator shim #86844)
• Adjust documentation (see instructions on rustc-dev-guide)
• Stabilization PR (see instructions on rustc-dev-guide)

Unresolved Questions

• Do we want to commit to allowing linking default output (rlib) format with arbitrary linkers? (see Support #[global_allocator] without the allocator shim #86844 (comment))
• TODO, maybe more. Maybe we want some formal RFC etc. before stabilizing this.

Implementation history

• Support #[global_allocator] without the allocator shim #86844
  • default_alloc_error_handler became stable in Stabilize default_alloc_error_handler #102318

[Mark-Simulacrum]

cc @bjorn3 -- wanted to file something official for this internal unstable detail (don't think anything existed already).

[bjorn3]

#102318 is not really useful as the alloc error handler is still part of the allocator shim. #112331 would entirely remove the allocation error handler. After that the allocator shim only contains the __rust_alloc_error_handler_should_panic static when #[global_allocator] is used. Maybe a weak static in liballoc with the default value when -Zoom is not passed would work for that?

[riking]

After #127173, definitions of this symbol must be migrated to include the #[rustc_std_internal_symbol] attribute, it looks like.