Closed dragly closed 2 years ago
Thanks for reporting! I can confirm the issue and the fix. It also affects resize
. I will prepare a 0.17.1 release with a fix.
Since this is a memory safety issue, I wonder whether we should yank 0.17.0 from crates.io? cc @davidhewitt
(We should probably add tests running under ASAN to the repository with all the unsafe code. Too bad Miri cannot be expected to handle our FFI interactions.)
Was this a new bug in 0.17.0? Seems reasonable to yank if so once 0.17.1 released. Do you need me to do that?
Was this a new bug in 0.17.0? Seems reasonable to yank if so once 0.17.1 released. Do you need me to do that?
Yes, this was introduced during copy-editing and hence is new in 0.17.0. I have not tried yanking and will message you if it fails. Here, I first wanted to know your opinion/advice whether yanking is reasonable response.
0.17.1 released and 0.17.0 yanked.
Great! Thanks for the incredibly rapid response and fix! :)
Running the reshape example as a test fails in release mode.
Steps to reproduce
Append the following to
test/array.rs
:Run the following command:
Output:
Workaround
My guess is that something goes wrong when pointing to memory that perhaps is optimized out in release mode.
The following workaround "fixes" the issue:
src/array.rs
:It also seems like other tests are failing in release mode.
System information
Reproduced with