Link sanitizers when creating dynamic libraries on macOS

[tmiasko]

Link sanitizer runtime when creating dynamic libraries on macOS
to resolve sanitizer runtime symbols and avoid failure at link time.

Closes #74571.

[rust-highfive]

r? @kennytm

(rust_highfive has picked a reviewer for you, use r? to override)

[Dylan-DPC-zz]

r? @Mark-Simulacrum

[Mark-Simulacrum]

This seems somewhat reasonable, but it's not clear to me why it wasn't needed for Linux. Do we implicitly link to sanitizers there? Why is that not happening on macOS?

I think that merits some comments in the code.

Regardless, I'm not really familiar enough with sanitizers to do more than rubber stamp this. I guess let's cc @cuviper @nikic and @nagisa as LLVM wg (since that's where our sanitizers come from); can one of you either take over review or suggest an assignee?

[tmiasko]

Added a comment explaining the rationale. The implementation as usual is intended to follow clang, except that we currently don't have any switches to select between static / shared type of runtime on targets were both could be supported.

[cuviper]

The explanation makes sense to me, but I don't know enough about the macOS target to approve this change.

[Mark-Simulacrum]

That explanation feels sufficient to me, so I'm going to @bors r+

That said I'm mostly only approving because this affects sanitizers only which are unstable.

[bors]

📌 Commit 4fbbc81 has been approved by Mark-Simulacrum

[bors]

⌛ Testing commit 4fbbc81 with merge b6396b7...

[bors]

☀️ Test successful - checks-actions, checks-azure
Approved by: Mark-Simulacrum
Pushing b6396b7 to master...