Add Lint PathBufPushOverwrite #3954

andrehjr · 2019-04-13T16:55:31Z

This is a very simple Lint that checks if push is being called with a Root Path. Because that can make it overwrite the previous path.

I used std::path::Path to check if it's root, in order to keep it working across windows/linux environments instead of checking for '/'. Is that alright?

On the if_chain! block, Is there a way to make it short while getting the value of the first argument? I got the example from other lints.

Note that this is first Lint, I hope I got everything covered 🚀

JamesHinshelwood · 2019-04-13T18:14:35Z

clippy_lints/src/path_buf_push_overwrite.rs

+            if let ExprKind::Lit(ref lit) = get_index_arg.node;
+            if let LitKind::Str(ref r, _) = lit.node;
+            then {
+                if Path::new(&r.as_str()).has_root() {


Would is_absolute be better than has_root here? The docs for push seem to suggest the path will be replaced if the other path is absolute.
Also, could this if be part of the if_chain! above?

That makes sense! Seems like is_absolute is better. As It's only equivalent to is_root on Unix. I'll make the changes.

Copying from below: #3954 (comment)

I'm not sure about the is_absolute vs has_root thing. ::push("\windows") on windows will result in c:\windows, while it doesn't matter what the previous path was (except the prefix). But is_absolute does not cover this case.

flip1995

Thanks for the contribution!

Basic implementation LGTM. Some minor things need to be changed.

Also AppVeyor is not happy with the unix-like paths. probably because of the is_absolute vs has_root thing.

clippy_lints/src/path_buf_push_overwrite.rs

flip1995 · 2019-04-15T07:45:30Z

clippy_lints/src/path_buf_push_overwrite.rs

+            then {
+                span_help_and_lint(cx, PATH_BUF_PUSH_OVERWRITE, expr.span,
+                "Calling `push` with '/' (file system root) can overwrite the previous path definition",
+                "Remove the '/' (file system root) from the start of the path instead");


You should be able to use span_lint_and_sugg. You can build the suggestion with https://doc.rust-lang.org/std/primitive.str.html#method.trim_left_matches.

Also formatting:

span_lint_and_sugg( cx, LINT, span, // ... );

clippy_lints/src/path_buf_push_overwrite.rs

flip1995 · 2019-04-16T07:37:01Z

This LGTM now. I just would like to see, that you change the span_help_and_lint with span_lint_and_sugg: #3954 (comment)

If you need help with this just let me know! :)

andrehjr · 2019-04-16T11:43:09Z

Sure! I fixed the other issues first, I'll do the span_lint_and_sugg today! 🚀

Another thing that I left for today, is that one of the builds on travisci failing because of this doctest:

thread 'main' panicked at 'assertion failed: `(left == right)`
  left: `"/foo\\bar"`,
 right: `"/foo/bar"`', doctest.rs:7:1

https://travis-ci.com/rust-lang/rust-clippy/jobs/193169548

Is the best alternative here to add a if cfg!(windows) and have both assert! versions for linux/windows ?

flip1995 · 2019-04-16T12:09:38Z

Try to write the assert like this, instead of comparing strings (like in the push documentation):

assert_eq!(x, PathBuf::from("/bar"));

andrehjr · 2019-04-17T01:45:30Z

I 'almost' got it, added the suggestion and rustfix.

I tried using Component::RootDir from the Path in order to replace the the Root correctly for both win/linux use cases. It did work for linux, but it's failing now on windows. I'll get a windows vm to debug this better, or find a better way to do it. Thanks for the reviews so far @flip1995 !

clippy_lints/src/path_buf_push_overwrite.rs

flip1995 · 2019-04-17T12:31:11Z

I'll get a windows vm to debug this better, or find a better way to do it.

Stupid (?) idea: You could setup a test repo with travis for windows. Then you push your clippy binaries there and debug it in Travis. Not sure if that or the windows VM is more work.

andrehjr · 2019-04-18T00:18:24Z

@flip1995 that was a very good idea indeed! Much better than getting a VM just for that. Thanks so much! :D

In the end, I think the best impl to remove 'root' paths is the following: path.trim_start_matches(|c| c == '/' || c == '\\')) when path.has_root() => true.

Digging more on how this is being tested on Rust And seeing how windows, unix and others, look for their MAIN_SEPARATOR.

And also because calling components() on a path Component::RootDir changes the root from '/' to '\' on windows.

flip1995 · 2019-04-18T07:26:22Z

This LGTM now!

Could you squash some commits and drop 53e34fa? This commit is already implemented in #3968 and would lead to conflicts. Once #3968 is merged you probably have to rebase (Sorry).

bors · 2019-04-18T07:50:42Z

☔ The latest upstream changes (presumably #3968) made this pull request unmergeable. Please resolve the merge conflicts.

andrehjr · 2019-04-18T11:06:56Z

Awesome 🚀 no worries, I was going to rebase it anyway!

flip1995 · 2019-04-18T11:29:40Z

Thanks!

@bors r+

bors · 2019-04-18T11:29:41Z

📌 Commit 7e9cb5b has been approved by flip1995

bors · 2019-04-18T11:29:48Z

⌛ Testing commit 7e9cb5b with merge c6e43b1...

Add Lint PathBufPushOverwrite Closes #3923 This is a very simple Lint that checks if push is being called with a Root Path. Because that can make it overwrite the previous path. I used std::path::Path to check if it's root, in order to keep it working across windows/linux environments instead of checking for '/'. Is that alright? On the `if_chain!` block, Is there a way to make it short while getting the value of the first argument? I got the example from other lints. Note that this is first Lint, I hope I got everything covered 🚀

bors · 2019-04-18T12:53:34Z

☀️ Test successful - checks-travis, status-appveyor
Approved by: flip1995
Pushing c6e43b1 to master...

flip1995 added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties label Apr 13, 2019

JamesHinshelwood reviewed Apr 13, 2019

View reviewed changes

flip1995 requested changes Apr 15, 2019

View reviewed changes

flip1995 added S-waiting-on-author Status: This is awaiting some action from the author. (Use `@rustbot ready` to update this status) and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties labels Apr 15, 2019

andrehjr force-pushed the add-lint-path-buf-overwrite branch from 78c53f5 to 53e34fa Compare April 15, 2019 22:45

flip1995 reviewed Apr 17, 2019

View reviewed changes

clippy_lints/src/path_buf_push_overwrite.rs Outdated Show resolved Hide resolved

flip1995 approved these changes Apr 18, 2019

View reviewed changes

andrehjr force-pushed the add-lint-path-buf-overwrite branch from 645683d to 6d4e682 Compare April 18, 2019 11:03

Add lint PathBufPushOverwrite

7e9cb5b

andrehjr force-pushed the add-lint-path-buf-overwrite branch from 6d4e682 to 7e9cb5b Compare April 18, 2019 11:06

bors merged commit 7e9cb5b into rust-lang:master Apr 18, 2019

andrehjr deleted the add-lint-path-buf-overwrite branch April 18, 2019 13:10

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add Lint PathBufPushOverwrite #3954

Add Lint PathBufPushOverwrite #3954

andrehjr commented Apr 13, 2019

JamesHinshelwood Apr 13, 2019

andrehjr Apr 13, 2019

flip1995 Apr 15, 2019

flip1995 left a comment •

edited

Loading

flip1995 Apr 15, 2019

flip1995 commented Apr 16, 2019

andrehjr commented Apr 16, 2019

flip1995 commented Apr 16, 2019

andrehjr commented Apr 17, 2019

flip1995 commented Apr 17, 2019

andrehjr commented Apr 18, 2019 •

edited

Loading

flip1995 commented Apr 18, 2019

bors commented Apr 18, 2019

andrehjr commented Apr 18, 2019

flip1995 commented Apr 18, 2019

bors commented Apr 18, 2019

bors commented Apr 18, 2019

bors commented Apr 18, 2019

Add Lint PathBufPushOverwrite #3954

Add Lint PathBufPushOverwrite #3954

Conversation

andrehjr commented Apr 13, 2019

JamesHinshelwood Apr 13, 2019

Choose a reason for hiding this comment

andrehjr Apr 13, 2019

Choose a reason for hiding this comment

flip1995 Apr 15, 2019

Choose a reason for hiding this comment

flip1995 left a comment • edited Loading

Choose a reason for hiding this comment

flip1995 Apr 15, 2019

Choose a reason for hiding this comment

flip1995 commented Apr 16, 2019

andrehjr commented Apr 16, 2019

flip1995 commented Apr 16, 2019

andrehjr commented Apr 17, 2019

flip1995 commented Apr 17, 2019

andrehjr commented Apr 18, 2019 • edited Loading

flip1995 commented Apr 18, 2019

bors commented Apr 18, 2019

andrehjr commented Apr 18, 2019

flip1995 commented Apr 18, 2019

bors commented Apr 18, 2019

bors commented Apr 18, 2019

bors commented Apr 18, 2019

flip1995 left a comment •

edited

Loading

andrehjr commented Apr 18, 2019 •

edited

Loading