On Fri, Feb 04, 2022 at 12:49:27AM +0200, Nir Soffer wrote:
On Thu, Feb 3, 2022 at 10:26 PM Eric Blake <eblake@redhat.com> wrote:
Like a lot of the C examples, the aio copy test ignores read and write errors in the completion callback, which can cause silent data corruption. The failure in the test is not critical, but this is a bad example that may be copied by developers to a real application.
@@ -36,6 +36,7 @@ def asynch_copy(src, dst): # This callback is called when any pread from the source # has completed. def read_completed(buf, offset, error): + assert not error
This works for the test, since the test is not compiled to pyc file, which removes the asserts (like C -DNODBUG) by default when building rpms.
If someone will copy this to a real application they will have no error checking.
I would use:
if error != 0: raise RuntimeError(f"read: {os.strerror(error)}")
Which also teaches the reader that error is an errno value.
Will do (I trust your python expertise more than my own ;) -- Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3266 Virtualization: qemu.org | libvirt.org