On Wed, Jan 08, 2025 at 03:21:42PM +0100, Helmut Grohne wrote: > Since I was failing to get this right earlier, I figured I could as well > write some tests. As a result, I'm attaching a shell script. It > constructs a few demo packages and attempts a hopefully exhaustive > combination of using them and dpkg-divert by writing its experiments to > the working directory. Running it should be fairly boring. If you happen > to review it, I appreciate a response.
Despite multiple requests to actually attach my tests, nobody seems to have read them and spot the missing test cases. And of course they were failing. The case of upgrading from a single, aliased diversion to a duplicated diversion was not covered. In this case, we want the canonicalized diversion target to be used by the canonicalized diversion and therefore, the aliased diversion must change its diversion target. Failure to do so results in file loss and/or left-over files. I'm attaching an updated version that tests the transition from aliased diversion to duplicated diversion both in the case of the divertee being absent and present. It's a lot of branches now, but I don't quite see how we could do it with less. Again, I appreciate someone actually reading the test cases and telling me where I am wrong such that I don't have to redo too many redone patches. At least the one I submitted for bfh-container fails an additional test case. Helmut
m18_diversions.sh
Description: Bourne shell script
