rustc: Add official support for weak failure

[alexcrichton]

This commit is part of the ongoing libstd facade efforts (cc #13851). The
compiler now recognizes some language items as "extern { fn foo(...); }" and
will automatically perform the following actions:

1. The foreign function has a pre-defined name.
2. The crate and downstream crates can only be built as rlibs until a crate
   defines the lang item itself.
3. The actual lang item has a pre-defined name.

This is essentially nicer compiler support for the hokey
core-depends-on-std-failure scheme today, but it is implemented the same way.
The details are a little more hidden under the covers.

In addition to failure, this commit promotes the eh_personality and
rust_stack_exhausted functions to official lang items. The compiler can generate
calls to these functions, causing linkage errors if they are left undefined. The
checking for these items is not as precise as it could be. Crates compiling with
-Z no-landing-pads will not need the eh_personality lang item, and crates
compiling with no split stacks won't need the stack exhausted lang item. For
ease, however, these items are checked for presence in all final outputs of the
compiler.

It is quite easy to define dummy versions of the functions necessary:

#[lang = "stack_exhausted"]
extern fn stack_exhausted() { /* ... */ }

#[lang = "eh_personality"]
extern fn eh_personality() { /* ... */ }

cc #11922, rust_stack_exhausted is now a lang item
cc #13851, libcollections is blocked on eh_personality becoming weak

[brson]

Why do the lang items themselves need to be extern fns? Is it because the weak symbol is the same as the lang item symbol and the names must be exactly the same? Given the current system, I might prefer that the lang items are allowed to continue being normal rust functions but are inlined into a wrapper with a known symbol name, but if we convert intrinsics to lang items then there will be additional precedent for having extern lang items (I still might like the wrappers though so the lang items don't have to artifically use a specific name).

Finally, if we are going to demand that certain Rust-specific symbols always exist, I would prefer they start with __rust_ to make it absolutely clear these are implementation details.

[luqmana]

Should this be extern "C" fn?

[alexcrichton]

The default ABI for an extern fn is not "C", so the string should be optional.

[huonw]

Was "not" a typo?

extern fn foo() {}
let _: extern "C" fn() = foo;

compiles fine.

[alexcrichton]

Oh whoops, yes, I mean to say that the default ABI is "C"

[bharrisau]

Not sure if I've missed something here, but with this change shared libs are now defining rust_eh_personality ? So if I try to link a binary using two rust shared libs I'll get linker errors for multiple definitions. Is there a work around?

[huonw]

Are you hitting that in practice? (If so, file a bug.)