On Mon, 2026-03-09 at 18:27 +0700, Arnaud Rebillout wrote: > On 09/03/2026 17:23, Sven Geuer wrote: > > Getting the codename is essential, according to my tests. It won't > > install isc-dhcp-client but dhcpcd-base if debci is invoked like "debci > > setup -s stable". The same applies when setting up an oldstable or > > oldoldstable testbed. > > You're correct, my approach doesn't work for stable, oldstable and the rest. > > I note that there are various places in the template where $release is > matched against a list of codenames, so the template is already bogus in > this regard.
I noticed this also. > I wonder if it's a good idea to fix it by using the > variable distro_codename that you introduced. I would think so. > I'll rework my MR to expand on your patch, give me some minutes (or hours). Cool! > > > @elbrus: I believe the fix doesn't need to apply to trixie, but only to > > > forky onward. Please tell me if I'm mistaken. > > It depends on what release the worker/host-running-the-test is running > > natively. If there are workers running on trixie they need the fix as > > well, it seems to me. > > I don't understand this one. The package isc-dhcp-client is present on > trixie, so why do we need the fix? Or is it missing on trixie/loong64 > (didn't find a way to check, rmadison doesn't tell about loong64)? It makes a difference when running a forky or sid testbed on a trixie- based worker, doesn't it? Correct me if I am wrong. Sven -- GPG Fingerprint 3DF5 E8AA 43FC 9FDF D086 F195 ADF5 0EDA F8AD D585
signature.asc
Description: This is a digitally signed message part
