RecordProgress for DatabaseDB (cli mode) fix to record result for the updated crate name

[azhogin]

Bug in cli mode: record_progress for DatabaseDB stores not updated crate name in results table. While update_crate_version writes updated crate name into experiment_crates table. So, when report is generated, logs for the new crate name can't be found:

[2026-01-18T15:05:30Z ERROR crater::utils] failed to read log of brson/hello-rs/9da3b03b25c2dfa8a8b8a69db503c48dbd08610e on stable
[2026-01-18T15:05:30Z ERROR crater::utils] caused by: failed to read log of brson/hello-rs/9da3b03b25c2dfa8a8b8a69db503c48dbd08610e on stable
[2026-01-18T15:05:30Z ERROR crater::utils] caused by: missing logs

Because results are stored for "brson/hello-rs", not "brson/hello-rs/9da3b03b25c2dfa8a8b8a69db503c48dbd08610e".

In agent mode this problem was already fixed in pub fn store(...):

        let krate = if let Some((old, new)) = &data.version {
            // If we're updating the name of the crate (typically changing the hash we found on
            // github) then we ought to also use that new name for marking the crate as complete.
            // Otherwise, we leave behind the old (unversioned) name and end up running this crate
            // many times, effectively never actually completing it.
            self.update_crate_version(ex, old, new)?;

            // sanity check that the previous name of the crate is the one we intended to run.
            if old.id() != data.result.krate.id() {
                log::warn!(
                    "Storing result under {} despite job intended for {} (with wrong name old={})",
                    new.id(),
                    data.result.krate.id(),
                    old.id(),
                );
            }

            new
        } else {
            &data.result.krate
        };

        self.store_result(
            ex,
            krate,

This PR fixes the problem in cli mode in the similar way.