C "memcpy" shim: ensure the pointers are valid

[RalfJung]

Also add tests for some other shims that already behave correctly

Fixes #2966

[RalfJung]

@bors r+

[bors]

📌 Commit 2625865 has been approved by RalfJung

It is now in the queue for this repository.

[bors]

⌛ Testing commit 2625865 with merge b220309...

[bors]

💔 Test failed - checks-actions

[oli-obk]

I don't get this CI failure. memrchr exists (it works locally, and it doesn't have any cargo features/cfgs/anything).

[oli-obk]

Ooooh, that wasn't clear from the ui_test annotation: this is for a cross compilation target.

[oli-obk]

@bors r=RalfJung,oli-obk

[bors]

📌 Commit 0613689 has been approved by RalfJung,oli-obk

It is now in the queue for this repository.

[bors]

⌛ Testing commit 0613689 with merge fc6bd08...

[RalfJung]

Btw, what is this in the test output?

   error file=tests/fail/shims/memrchr_null.stderr,line=0,title=actual output differs from expected::```diff%0A1 unchanged lines skipped%0A-error: Undefined Behavior: memory access failed: null pointer is a dangling pointer (it has no provenance)%0A+error[E0425]: cannot find function `memrchr` in crate `libc`%0A   --> $DIR/memrchr_null.rs:LL:CC%0A    |%0A LL |         libc::memrchr(ptr::null(), 0, 0);%0A-   |         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ memory access failed: null pointer is a dangling pointer (it has no provenance)%0A+   |               ^^^^^^^%0A    |%0A-   = help: this indicates a bug in the program: it performed an invalid operation, and caused Undefined Behavior%0A-   = help: see https://doc.rust-lang.org/nightly/reference/behavior-considered-undefined.html for further information%0A-   = note: BACKTRACE:%0A-   = note: inside `main` at $DIR/memrchr_null.rs:LL:CC%0A+  ::: CARGO_REGISTRY/.../mod.rs:LL:CC%0A+   |%0A+LL |     pub fn memchr(cx: *const c_void, c: c_int, n: size_t) -> *mut c_void;%0A+   |     -------------------------------------------------------------------- similarly named function `memchr` defined here%0A+   |%0A+help: a function with a similar name exists%0A+   |%0A+LL |         libc::memchr(ptr::null(), 0, 0);%0A+   |               ~~~~~~%0A+help: consider importing this function%0A+   |%0A+LL + use core::slice::memchr::memrchr;%0A+   |%0A+help: if you import `memrchr`, refer to it directly%0A+   |%0A+LL -         libc::memrchr(ptr::null(), 0, 0);%0A+LL +         memrchr(ptr::null(), 0, 0);%0A+   |%0A %0A-note: some details are omitted, run with `MIRIFLAGS=-Zmiri-backtrace=full` for a verbose backtrace%0A-%0A error: aborting due to previous error%0A %0A+For more information about this error, try `rustc --explain E0425`.%0A %0A```

That's a very long line repeating the entire diff... some kind of debug output?

[oli-obk]

that's just github actions printing the github actions commands

[RalfJung]

Looks like the blessed output now says "cannot find function"? That doesn't seem right?^^

[oli-obk]

yea 🤦 I forgot to bless normally again

@bors r=RalfJung,oli-obk

[bors]

📌 Commit a9773f5 has been approved by RalfJung,oli-obk

It is now in the queue for this repository.

[bors]

⌛ Testing commit a9773f5 with merge ee6623f...

[RalfJung]

No need to have that wrong bless in the history at all...

@bors retry r+

[bors]

📌 Commit 852d74f has been approved by RalfJung

It is now in the queue for this repository.

[bors]

⌛ Testing commit 852d74f with merge 8b9be80...

[oli-obk]

No need to have that wrong bless in the history at all...

the gh pr checkout command does not allow force pushing 🙃

[bors]

☀️ Test successful - checks-actions
Approved by: RalfJung
Pushing 8b9be80 to master...