On 1/17/22 14:09, Andrea Bolognani wrote:
On Mon, Jan 17, 2022 at 01:39:31PM +0100, Michal Prívozník wrote:
On 1/17/22 13:37, Andrea Bolognani wrote:
On Mon, Jan 17, 2022 at 01:11:29PM +0100, Michal Prívozník wrote:
On 1/14/22 17:49, Andrea Bolognani wrote:
This all works, but I wonder if we couldn't just create a trivial shell script that behaves minimally the way we expect dnsmasq to, and change our virFindFileInPath() mock so that it returns the absolute path to it? That way we wouldn't need to implement any additional mocking and the code would end up being much simpler. Diff below.
I thought that we should avoid shell for new contributions:
Fair enough. Python version below.
#!/usr/bin/env python3
import sys
output = { "--version": "Dnsmasq version 2.67", "--help": "--bind-dynamic\n--ra-param", }
if len(sys.argv) != 2 or sys.argv[1] not in output: print("invalid usage") sys.exit(1)
print(output[sys.argv[1]])
And what exactly is the point? I'm failing to see why this would be any better than mocking virCommand. Can you elaborate please?
I believe the diffstats speak for themselves :)
$ git diff --stat df09e45310..64325fa9ef tests/networkmock.c | 16 ++++++++++++++++ tests/networkxml2conftest.c | 38 +++++++++++++++++++++++++++++++++++++- 2 files changed, 53 insertions(+), 1 deletion(-)
$ git diff --stat df09e45310..2b64fb492b tests/networkmock.c | 2 +- tests/networkxml2conftest.c | 2 +- tests/virdnsmasqmock.py | 14 ++++++++++++++ 3 files changed, 16 insertions(+), 2 deletions(-)
Alright, let's switch to python. I'm starting to not care about these patches fate so I won't object. Michal