bors [Tue, 26 Jan 2021 06:58:04 +0000 (06:58 +0000)]
Auto merge of #6469 - matthiaskrgr:clippy_dev_crater, r=flip1995
add "cargo dev crater" to run clippy on a fixed set of crates and diff the lint warnings
`cargo dev crater` now does the following:
build clippy in debug mode
for a fixed set of crates:
download and extract the crate
run compiled clippy on the crate
dump the warnings into a file that is inside the repo
We can then do a "git diff" and see what effects our clippy changes had on a tiny fraction of the rust ecosystem and can see when an change unexpectedly added or silenced a lot of warnings.
Checking all the crates took less than 5 minutes on my system.
Should help with https://github.com/rust-lang/rust-clippy/issues/6429
---
*Please write a short comment explaining your change (or "none" for internal only changes)*
changelog: extend cargo dev to run clippy against a fixed set of crates and compare warnings
I'm not sure if the guide is written in the best way. Style suggestions are appreciated. :upside_down_face:
---
Again a big **thank you** for everyone who helped to collect the abbreviation list over on [zulip]. I had a lot of fun, and it was also very informative. Keep up the good work :upside_down_face:
bors [Fri, 22 Jan 2021 22:28:41 +0000 (22:28 +0000)]
Auto merge of #6591 - camsteffen:manual-filter-map, r=llogiq
`manual_filter_map` and `manual_find_map`
changelog: Add `manual_filter_map` and replace `find_map` with `manual_find_map`
Replaces #6453
Fixes #3188
Fixes #4193
~Depends on #6567 (to fix an internal lint false positive)~
This replaces `filter_map` and `find_map` with `manual_filter_map` and `manual_find_map` respectively. However, `filter_map` is left in place since it is used for a variety of other cases. See discussion in #6453.
xFrednet [Thu, 21 Jan 2021 23:19:22 +0000 (00:19 +0100)]
Added documentation for common abbreviations
This list was created as a collaborative effort on Zulip and the [thread] is definitely worth a read as we had quite some fun. A big **THANK YOU** goes out to everyone who participated you make this project fun to work on!!!
The Zulip [thread]: https://rust-lang.zulipchat.com/#narrow/stream/257328-clippy/topic/Common.20abbreviations.20in.20basics.2Emd/near/223548065
bors [Thu, 21 Jan 2021 14:12:43 +0000 (14:12 +0000)]
Auto merge of #6611 - pastchick3:master, r=flip1995
Fix the reversed suggestion message of `stable_sort_primitive`.
Now Clippy emits `stable_sort_primitive` warning as follows:
```
warning: used sort instead of sort_unstable to sort primitive type `usize`
--> src\asm.rs:41:13
|
41 | self.successors.sort();
| ^^^^^^^^^^^^^^^^^^^^^^ help: try: `self.successors.sort_unstable()`
|
= note: `#[warn(clippy::stable_sort_primitive)]` on by default
= help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#stable_sort_primitive
```
I think the position of `sort` and `sort_unstable` in the first line should be reversed.
changelog: Fix the reversed suggestion message of `stable_sort_primitive`.
bors [Wed, 20 Jan 2021 21:35:11 +0000 (21:35 +0000)]
Auto merge of #6567 - camsteffen:path-to-res-enum, r=Manishearth
Fix path_to_res for enum inherent items
changelog: none
I tried to add `Option::is_some` to the paths but got a false positive from the invalid paths lint. Turns out the `path_to_res` function does not find inherent impls for enums. I fixed this and took the liberty to do some additional cleanup in the method.
bors [Wed, 20 Jan 2021 07:24:34 +0000 (07:24 +0000)]
Auto merge of #6578 - MarijnS95:size-in-element-count-divide-by-byte-size, r=flip1995
size_of_in_element_count: Disable lint on division by byte-size
Fixes #6511
It is fairly common to divide some length in bytes by the byte-size of a single element before creating a `from_raw_parts` slice or similar operation. This lint would erroneously disallow such expressions.
Just in case, instead of simply disabling this lint in the RHS of a division, keep track of the inversion and enable it again on recursive division.
---
changelog: Do not trigger size_of_in_element_count when dividing by element size
Marijn Suijten [Tue, 12 Jan 2021 10:35:44 +0000 (11:35 +0100)]
size_of_in_element_count: Disable lint on division by byte-size
It is fairly common to divide some length in bytes by the byte-size of a
single element before creating a `from_raw_parts` slice or similar
operation. This lint would erroneously disallow such expressions.
Just in case, instead of simply disabling this lint in the RHS of a
division, keep track of the inversion and enable it again on recursive
division.
Marijn Suijten [Tue, 19 Jan 2021 18:20:26 +0000 (19:20 +0100)]
size_of_in_element_count: Separate test file in expressions and functions
An upcoming test case for new expresssion variants make the stderr file
go over 200 lines. Split this test case in two to have a clear
distinction between checking whether the lint is still applying on
all the functions with member counts as argument, versus validating
various member-count expressions that may or may not be invalid.
bors [Tue, 19 Jan 2021 08:01:56 +0000 (08:01 +0000)]
Auto merge of #6608 - rail-rain:note_on_as_conversions, r=phansch
Add a note to `as_conversions`
I have seen a couple of examples where there are some misunderstandings of `as_conversions` ([1](https://github.com/rust-lang/rust-clippy/issues/5890#issuecomment-671852546), [2](https://github.com/rust-lang/rust-clippy/issues/6116#issuecomment-704251710) and [3](https://github.com/rust-lang/rust-clippy/issues/6428)). This PR adds the note that explains its purpose and relationship with other `as` related casts. Open question: should I list every related lints for discoverbility, or just suggest how to find these? I chose the former because there's no way to list only and all `as` related lints (e.g. on All the Clippt Lints, 'cast' includes some noises, but `cast_` excludes some) even though I cannot guarantee the list will be updated to include future changes.
---
changelog: Add a note to the document of `as_conversions`
This will trigger on any type which implements `Index<RangeFull>` that returns the input type. This would be a false positive if the implementation does something other than return itself, but I'm not sure why you would ever want to do that.