Try to fix 48116 and 48192

[kennytm]

The bug #48116 happens because of a misoptimization of the import_path_to_string function, where a names slice is empty but the !names.is_empty() branch is executed.

rust/src/librustc_resolve/resolve_imports.rs

Lines 1015 to 1042 in 4d2d3fc

	fn import_path_to_string(names: &[SpannedIdent],
	subclass: &ImportDirectiveSubclass,
	span: Span) -> String {
	let pos = names.iter()
	.position(|p| span == p.span && p.node.name != keywords::CrateRoot.name());
	let global = !names.is_empty() && names[0].node.name == keywords::CrateRoot.name();
	if let Some(pos) = pos {
	let names = if global { &names[1..pos + 1] } else { &names[..pos + 1] };
	names_to_string(names)
	} else {
	let names = if global { &names[1..] } else { names };
	if names.is_empty() {
	import_directive_subclass_to_string(subclass)
	} else {
	// Note that this code looks a little wonky, it's currently here to
	// hopefully help debug #48116, but otherwise isn't intended to
	// cause any problems.
	let x = format!(
	"{}::{}",
	names_to_string(names),
	import_directive_subclass_to_string(subclass),
	);
	assert!(!names.is_empty());
	assert!(!x.starts_with("::"));
	return x
	}
	}
	}

Yesterday, @eddyb had locally reproduced the bug, and came across the position function where the assume() call is found to be suspicious. We have not concluded that this assume() causes #48116, but given the reputation of assume(), this seems higher relevant. Here we try to see if commenting it out can fix the errors.

Later @alexcrichton has bisected and found a potential bug in the LLVM side. We are currently testing if reverting that LLVM commit is enough to stop the bug. If true, this PR can be reverted (keep the assume()) and we could backport the LLVM patch instead.

(This PR also includes an earlier commit from #48127 for help debugging ICE happening in compile-fail/parse-fail tests.)

The PR also reverts #48059, which seems to cause #48192.

r? @alexcrichton
cc @eddyb, @arthurprs (#47333)

[alexcrichton]

@bors: r+ p=100

[bors]

📌 Commit 553f82d has been approved by alexcrichton

[alexcrichton]

@bors: r+

[bors]

📌 Commit 71269a5 has been approved by alexcrichton

[bors]

💔 Test failed - status-travis

[kennytm]

@bors r=alexcrichton

[bors]

📌 Commit e0da990 has been approved by alexcrichton

[bors]

⌛ Testing commit e0da990 with merge 3ec5a99...

[bors]

💔 Test failed - status-appveyor

[kennytm]

@bors retry #46903

[kennytm]

(Manually merged since everything else passed)