The gitlab.common-lisp.net system has migrated to a new host. One consequence of this is that the remotes for all cloned repos must now refer to the host gitlab.common-lisp.net rather than simply common-lisp.net.
You can see your current remotes from a cloned directory with git remote -v
If you see common-lisp.net in there without the gitlab. in front of it, then please update your remotes by using e.g. the following command:
git remote set-url origin git@gitlab.common-lisp.net:/{$user|$group}/{$repo}.git (assuming your remote name is origin).
gitlab-runner users: If you are running gitlab-runner on your own infrastructure, you need to update your config to point to https://gitlab.common-lisp.net, if it was previously pointing at simply https://common-lisp.net. The common-lisp.net host can no longer be used interchangeably with gitlab.common-lisp.net as may have worked previously. N.B. you can simply edit your gitlab-runner config and change the hostname; it should not be necessary to change anything with the token which you'll see along with that host url in the gitlab-runner config file.
gitlab.common-lisp.net has recently been migrated and has become pickier about what types of SSH keys it accepts. So if you are using an older keypair (e.g. rsa with less than 2048 bits) then you may need to generate and register a new pair.
Our gitlab now supports the following key types for both SSH login and Git operations:
ED25519: ssh-ed25519, ssh-ed25519-cert-v01@openssh.com
ECDSA: ecdsa-sha2-nistp256, ecdsa-sha2-nistp384, ecdsa-sha2-nistp521, and their certificate versions
RSA: rsa-sha2-512, rsa-sha2-256, and their certificate versions (Note: RSA keys must be at least 2048 bits long)
Security Keys: sk-ssh-ed25519@openssh.com, sk-ecdsa-sha2-nistp256@openssh.com, and their certificate forms
Older RSA keys with less than 2048 bits are no longer accepted. Please ensure your RSA keys meet this requirement.
How to Update: If you need to update or generate new SSH keys, please follow the instructions provided in the GitLab documentation at: Generating SSH KeysThank you for your coöperation in keeping our infrastructure up to date.
Following a heated discussion on #lisp, it seems that the test cases regarding values of iteration values in FINALLY clauses are incorrect. The spec contradicts itself in its wording and therefore we've come to a "conclusion" that it would be best to treat these values as undefined. Therefore it cannot be counted on that these tests are "correct" since they are testing undefined behaviour. It would therefore be preferable to note them as disabled.