update Miri

r? `@ghost`

avoid transmuting Box when we can just cast raw pointers instead

Always better to avoid a transmute, in particular when the layout assumptions it is making are not clearly documented. :)

Rollup of 5 pull requests

Successful merges:

 - #112751 (rustdoc: Fixes with --test-run-directory and relative paths.)
 - #114749 (Update `mpsc::Sender` doc to reflect that it implements `Sync`)
 - #114876 (Don't ICE in `is_trivially_sized` when encountering late-bound self ty)
 - #114881 (clarify CStr lack of layout guarnatees)
 - #114921 (Remove Folyd from librustdoc static files)

r? `@ghost`
`@rustbot` modify labels: rollup

Remove Folyd from librustdoc static files

r?`@GuillaumeGomez`

clarify CStr lack of layout guarnatees

Follow-up to https://github.com/rust-lang/rust/pull/114800
r? `@cuviper`

Don't ICE in `is_trivially_sized` when encountering late-bound self ty

We can see a bound ty var here:
https://github.com/rust-lang/rust/blob/b531630f4255216fce1400c45976e04f1ab35a84/compiler/rustc_trait_selection/src/traits/query/type_op/prove_predicate.rs#L13-L34

Fixes #114872

Update `mpsc::Sender` doc to reflect that it implements `Sync`

Fixes #114722

rustdoc: Fixes with --test-run-directory and relative paths.

Fixes #112191
Fixes #112210

This fixes some issues with `--test-run-directory` and its interaction with `--runtool` and `--persist-doctests`. Relative directories don't work with `Command::current_dir` very well because it has platform-specific behavior with relative paths. This fixes it by avoiding the use of relative paths.

This is needed because cargo is switching to use `--test-run-directory`, and it uses relative paths when interacting with rustdoc/rustc.

coverage: Anonymize line numbers in `run-coverage` test snapshots

LLVM's coverage reporter always prints line numbers in its coverage reports.

For testing purposes this is slightly inconvenient, because it means that adding or removing a line in a test file causes all subsequent lines in the snapshot to change. That makes it harder to see the actually meaningful changes in the re-blessed snapshot.

---

This change fixes that by adding another normalization pass that replaces all line numbers in the coverage reports with `LL`, which is similar to what UI tests tell the compiler to do when emitting line numbers in error messages.

Remove conditional use of `Sharded` from query caches

`Sharded` is already a zero cost abstraction, so it shouldn't affect the performance of the single thread compiler if LLVM does its job.

r? `@cjgillot`

This makes the test snapshots less sensitive to lines being added/removed.

Only run MaybeInitializedPlaces dataflow once to elaborate drops

This pass allows forward dataflow analyses to modify the CFG depending on the dataflow state. This possibility is used for the `MaybeInitializedPlace` analysis in drop elaboration, to skip the dataflow effect of dead unwinds without having to compute dataflow twice.

Optimize DroplessArena arena allocation

This optimizes `DroplessArena` allocation by always ensuring that it is aligned to `usize` and adding `grow_and_alloc` and `grow_and_alloc_raw`functions which both grow and allocate, reducing code size.

<table><tr><td rowspan="2">Benchmark</td><td colspan="1"><b>Before</b></th><td colspan="2"><b>After</b></th></tr><tr><td align="right">Time</td><td align="right">Time</td><td align="right">%</th></tr><tr><td>🟣 <b>clap</b>:check</td><td align="right">1.6968s</td><td align="right">1.6887s</td><td align="right"> -0.48%</td></tr><tr><td>🟣 <b>hyper</b>:check</td><td align="right">0.2552s</td><td align="right">0.2551s</td><td align="right"> -0.03%</td></tr><tr><td>🟣 <b>regex</b>:check</td><td align="right">0.9613s</td><td align="right">0.9553s</td><td align="right"> -0.62%</td></tr><tr><td>🟣 <b>syn</b>:check</td><td align="right">1.5402s</td><td align="right">1.5374s</td><td align="right"> -0.18%</td></tr><tr><td>🟣 <b>syntex_syntax</b>:check</td><td align="right">5.9175s</td><td align="right">5.8813s</td><td align="right"> -0.61%</td></tr><tr><td>Total</td><td align="right">10.3710s</td><td align="right">10.3178s</td><td align="right"> -0.51%</td></tr><tr><td>Summary</td><td align="right">1.0000s</td><td align="right">0.9962s</td><td align="right"> -0.38%</td></tr></table>

Rollup of 5 pull requests

Successful merges:

 - #113115 (we are migrating to askama)
 - #114784 (Improve `invalid_reference_casting` lint)
 - #114822 (Improve code readability by moving fmt args directly into the string)
 - #114878 (rustc book: make more pleasant to search)
 - #114899 (Add missing Clone/Debug impls to SMIR Trait related tys)

r? `@ghost`
`@rustbot` modify labels: rollup

Add missing Clone/Debug impls to SMIR Trait related tys

r? `@compiler-errors`

rustc book: make more pleasant to search

Improve code readability by moving fmt args directly into the string

There are some of occurrences where I also transformed `write!(f, "{}", x)` into `f.write_str(x.as_str())`.

r? `@notriddle`

Improve `invalid_reference_casting` lint

This PR improves the `invalid_reference_casting` lint:
 - by considering an unlimited number of casts instead only const to mut ptr
 - by also considering ptr-to-integer and integer-to-ptr casts
 - by also taking into account [`ptr::cast`](https://doc.rust-lang.org/std/primitive.pointer.html#method.cast), [`ptr::cast`](https://doc.rust-lang.org/std/primitive.pointer.html#method.cast-1) and [`ptr::cast_const`](https://doc.rust-lang.org/std/primitive.pointer.html#method.cast_const)

Most of this improvements comes from skimming Github Code Search result for [`&mut \*.*as \*const`](https://github.com/search?q=lang%3Arust+%2F%26mut+%5C*.*as+%5C*const%2F&type=code)

r? ``@est31`` (maybe)

we are migrating to askama

see https://github.com/rust-lang/rust/issues/108868

proc_macro: Update docs for `Spacing`

Brings the docs more in line with reality

Avoid unnecessary Vec resize.

If `size > 0` current implementation will first create an empty vec and then push an element into it, which will cause a resize that can be easily avoided.

It's obviously not a big deal, but this also gets rid of `mut` local variable.

If `size > 0` current implementation will first create an empty
vec and then push an element into it, which will cause a resize
that can be easily avoided.

It's obviously not a big deal, but this also gets rid of `mut`
local variable.