Ignore failures of RLS on aarch64 Windows

[Mark-Simulacrum]

We've been putting this patch on beta/stable for multiple cycles now, it likely makes sense to just put it on master as well to avoid the extra cherry pick on each beta promotion. I've filed #86606 to track actually fixing this but I somewhat doubt it'll happen in the short term.

r? @pietroalbini

[pietroalbini]

I'm all for moving the workaround to master instead of cherry-picking it to beta every single time, but I'm not too happy with this specific fix landing on master. The code that disables RLS is deep into rustbuild, and I'm worried someone will forget to remove it once RLS starts building again on that target.

Personally I'd be more happy with a generic BROKEN_TOOLS=rls environment variable we can set on CI, as that's a more visible place we'll be less likely to forget about.

[Mark-Simulacrum]

I would prefer to avoid that. Hooks like that add additional complexity to the whole system as they're external configuration we always have to think about; this fix is local. I'm happy to add a note to the relevant issue that this code was added as a stopgap and add a FIXME to the code being added; I think that's more likely to not let us forget this.

[Mark-Simulacrum]

Updated with a FIXME and added a note to the issue.

[Xanewok]

FWIW it looks like the build will be fixed on aarch64 Windows once rust-lang/rls#1741 is merged.

[pietroalbini]

@Xanewok do you think that PR is going to be merged in this cycle?

[Xanewok]

Is the cut around July 27th? If so then yes, I'd say this PR is going to be merged then.