Don't check [print/write]_with_newline on raw strings
Some tests for #3778 and some maybe-not-the-greatest code that passes those tests!
I didn't run `fmt` because a) it doesn't seem to install on nightly for me, and b) on stable it wanted to apply formatting to over 90 files. Happy to make any tweaks though!
I suspect this contribution may require more than just tweaks. I'm still sort of new to rust so it may not be idiomatic, and the specific approach I took feels a little heavy-handed and brittle. I'm happy to make changes with some guidance, or equally happy if this gives a starting place for someone else to do it better :)
This test doesn't reproduce the ICE since it only happens, when the macro is defined in another file.
Currently we can't add tests with multiple files AFAIK
Also using the auxiliary folder didn't help
Add a lint to warn on `T: Drop` bounds
**What it does:** Checks for generics with `std::ops::Drop` as bounds.
**Why is this bad?** `Drop` bounds do not really accomplish anything.
A type may have compiler-generated drop glue without implementing the
`Drop` trait itself. The `Drop` trait also only has one method,
`Drop::drop`, and that function is by fiat not callable in user code.
So there is really no use case for using `Drop` in trait bounds.
**Known problems:** None.
**Example:**
```rust
fn foo<T: Drop>() {}
```
Fixes#3773
Both regular strings and raw strings can contain literal newlines. This commit
extends the lint to also warn about terminating strings with these.
Behaviour handling for raw strings is also moved into `check_newlines` by
passing in the `is_raw` boolean from `check_tts` as
[suggested](https://github.com/rust-lang/rust-clippy/pull/3781#pullrequestreview-204663732)
Update why transmute_int_to_float is bad
As suggested in #3550, this PR changes the reason why using `transmute` from an integer to a float is not recommended. Effectively, `from_bits` uses `transmute` underneath, but the former is preferred.
Pass tests for #3778, {print,write}_with_newline false positive
This change guards the lint from checking newlines with a sort of complicated
check to see if it's a raw string. Raw strings shouldn't be newline-checked,
since r"\n" is literally \-n, not a newline. I think it's ok not to check for
_literal_ newlines at the end of raw strings, but maybe that's debatable.
I... don't think this code is that great. I wanted to write the check after
`check_tts`, but that was too late -- raw string type info is lost (or I
couldn't find it). Putting it inside `check_tts` feels heavy-duty and the check
itself feels like a brittle reach possibly into places it shouldn't.
Maybe someone can fix this up :)
Literal `\n` characters (not a newline) in a `r"raw"` string should not
fail the lint.
This affects both write_with_newline and print_with_newline, so it is added in
both places.
I also copied a missing test case from write_with_newline over to
print_with_newline and added a note that one of those tests is supposed to
fail.
**What it does:** Checks for generics with `std::ops::Drop` as bounds.
**Why is this bad?** `Drop` bounds do not really accomplish anything.
A type may have compiler-generated drop glue without implementing the
`Drop` trait itself. The `Drop` trait also only has one method,
`Drop::drop`, and that function is by fiat not callable in user code.
So there is really no use case for using `Drop` in trait bounds.
**Known problems:** None.
**Example:**
```rust
fn foo<T: Drop>() {}
```
Fix ICE #3717 in lint implicit_hasher
Fixes#3717
This fixes the ICE. We lose some information in a very specific case though. But less information if better than an ICE. For an example see the test file.
Does anyone know, if there's another way to get the `ty::Ty` of a `hir::Expr`?
