Closed atheriel closed 2 years ago
I think we may skip a bit too much now... reading the md5 passworded files errors, but the subsequent reads should not.
Maybe we should run the tests, and verify that under FIPS we get a sensible error message, not a crash or something.
I have merged the first commit, thanks for catching that.
OK I manually added most of the other fixes. Thanks!
This PR actually has two parts:
With these changes the test suite runs on my test FIPS system, which is CentOS-based:
Related to #96.