Remove sporadicly failing test

[matthiasbeyer]

This is obviously not how it is done. This patch will be merged to master so that we can work on master without issues, but a PR will be filed to revert this patch immediately... with a fix, if someone can come up with one.

---

This currently blocks the 0.6.0 release. I know that this is not how it is done, but I don't see a better way.

[matthiasbeyer]

bors merge=petreeftime

[bors]

Build failed:

• CI

[petreeftime]

I think the issue might be due to some shared resource between tests. All tests execute on their own thread inside a single program and their order of execution is not guaranteed, so I think order of execution of tests or some sort of race condition might play a role in the observed bug, due to some resource which is incorrectly shared between them.

[matthiasbeyer]

Hm... but the tests execute subprocesses and read and write their stdin/stdout... I don't see where are shared resources, tbh.

[petreeftime]

rexpect/src/process.rs

Line 89 in 59213cd

let master_fd = posix_openpt(OFlag::O_RDWR)?;

On the call to posix_openpt, I wonder if the O_NOCTTY flag might be required, the docs seem to imply it changes the master for the calling process.

[petreeftime]

I've tested this and it seems to pass tests on my repo.

https://github.com/petreeftime/rexpect/actions/runs/4779091507

[petreeftime]

Let me create a PR for the fix.

[petreeftime]

Attempted fix in #101

[ekarlsn]

This approach has my blessing. But it seems like more tests have the issue.

[matthiasbeyer]

Rebased to latest master, to see what's happening.

[matthiasbeyer]

bors try

[bors]

try

Build failed:

• CI