feat: add CI for docker

Docker and Compose support is provided through a standalone repo at
https://github.com/chatmail/docker, add reusable Docker build/test CI
jobs to staging workflows and a Docker docs stub pointing to the
chatmail/docker repository.

Requires CHATMAIL_DOCKER_DISPATCH_TOKEN secret in relay repo settings
(fine-grained PAT with contents:write on chatmail/docker).

.github/workflows/ci.yaml

@@ -15,78 +15,28 @@ jobs:
         with:
           ref: ${{ github.event.pull_request.head.sha }}
       - name: download filtermail
-        run: curl -L https://github.com/chatmail/filtermail/releases/download/v0.5.2/filtermail-x86_64 -o /usr/local/bin/filtermail && chmod +x /usr/local/bin/filtermail
-      - name: run chatmaild tests
+        run: curl -L https://github.com/chatmail/filtermail/releases/download/v0.6.1/filtermail-x86_64 -o /usr/local/bin/filtermail && chmod +x /usr/local/bin/filtermail
+      - name: run chatmaild tests 
         working-directory: chatmaild
         run: pipx run tox
 
   scripts:
-    name: deploy-chatmail tests
+    name: deploy-chatmail tests 
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
 
-      - name: initenv
+      - name: initenv 
         run: scripts/initenv.sh
 
       - name: append venv/bin to PATH
         run: echo venv/bin >>$GITHUB_PATH
 
-      - name: run formatting checks
-        run: cmdeploy fmt -v
+      - name: run formatting checks 
+        run: cmdeploy fmt -v 
 
-      - name: run deploy-chatmail offline tests
-        run: pytest --pyargs cmdeploy
+      - name: run deploy-chatmail offline tests 
+        run: pytest --pyargs cmdeploy 
 
-  lxc-test:
-    name: LXC deploy and test
-    runs-on: ubuntu-24.04
-    timeout-minutes: 30
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          ref: ${{ github.event.pull_request.head.sha }}
-
-      - name: install incus
-        run: |
-          # zabbly is the official incus community packages source
-          curl -fsSL https://pkgs.zabbly.com/key.asc \
-            | sudo gpg --dearmor -o /etc/apt/keyrings/zabbly.gpg
-          sudo sh -c 'cat <<EOF > /etc/apt/sources.list.d/zabbly-incus-stable.sources
-          Enabled: yes
-          Types: deb
-          URIs: https://pkgs.zabbly.com/incus/stable
-          Suites: $(. /etc/os-release && echo ${VERSION_CODENAME})
-          Components: main
-          Architectures: $(dpkg --print-architecture)
-          Signed-By: /etc/apt/keyrings/zabbly.gpg
-          EOF'
-          sudo apt-get update
-          sudo apt-get install -y incus
-
-      - name: initialise incus
-        run: |
-          sudo systemctl stop docker.socket docker || true
-          sudo iptables -P FORWARD ACCEPT
-          sudo sysctl -w fs.inotify.max_user_instances=65535
-          sudo sysctl -w fs.inotify.max_user_watches=65535
-          sudo incus admin init --minimal
-          sudo usermod -aG incus-admin "$USER"
-
-      - name: initenv
-        run: scripts/initenv.sh
-
-      - name: append venv/bin to PATH
-        run: echo venv/bin >>$GITHUB_PATH
-
-      - name: lxc-test
-        run: sg incus-admin -c 'cmdeploy lxc-test'
-
-      - name: export images for cache
-        if: always()
-        run: |
-          for alias in localchat-base localchat-ns localchat-test0 localchat-test1; do
-            if ! [ -f /tmp/$alias.tar.gz ]; then
-              sg incus-admin -c "incus image export $alias /tmp/$alias" || true
-            fi
-          done
+      # all other cmdeploy commands require a staging server 
+      # see https://github.com/deltachat/chatmail/issues/100

.github/workflows/docker-deploy.yaml

@@ -0,0 +1,125 @@
+name: Docker deploy
+
+on:
+  workflow_call:
+    inputs:
+      staging_host:
+        required: true
+        type: string
+        description: 'SSH hostname (e.g. staging2.testrun.org)'
+      mail_domain:
+        required: true
+        type: string
+        description: 'MAIL_DOMAIN for docker compose'
+      zone_file:
+        required: true
+        type: string
+        description: 'Default zone file basename (e.g. staging.testrun.org-default.zone)'
+
+jobs:
+  deploy-docker:
+    name: Docker deploy on ${{ inputs.staging_host }}
+    runs-on: ubuntu-latest
+    timeout-minutes: 20
+    environment:
+      name: ${{ inputs.staging_host }}
+      url: https://${{ inputs.staging_host }}/
+    concurrency: ${{ inputs.staging_host }}
+    env:
+      VPS: root@${{ inputs.staging_host }}
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: true
+      - name: Setup SSH
+        run: |
+          mkdir ~/.ssh
+          echo "${{ secrets.STAGING_SSH_KEY }}" >> ~/.ssh/id_ed25519
+          chmod 600 ~/.ssh/id_ed25519
+          ssh-keyscan ${{ inputs.staging_host }} > ~/.ssh/known_hosts
+          # Reuse TCP connection for all subsequent ssh/scp calls
+          echo -e "Host ${{ inputs.staging_host }}\n  ControlMaster auto\n  ControlPath ~/.ssh/ctrl-%r@%h:%p\n  ControlPersist 10m" >> ~/.ssh/config
+
+      - name: stop bare services, install Docker, prepare mounts
+        run: |
+          ssh $VPS bash -s <<'EOF'
+          systemctl stop postfix dovecot nginx opendkim unbound filtermail doveauth chatmail-metadata iroh-relay mtail fcgiwrap acmetool 2>/dev/null || true
+          curl -fsSL https://download.docker.com/linux/debian/gpg -o /etc/apt/keyrings/docker.asc && chmod a+r /etc/apt/keyrings/docker.asc
+          echo "deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.asc] https://download.docker.com/linux/debian $(. /etc/os-release && echo $VERSION_CODENAME) stable" > /etc/apt/sources.list.d/docker.list
+          apt-get update
+          apt-get install -y docker-ce docker-ce-cli containerd.io docker-compose-plugin
+          mkdir -p /srv/chatmail/certs /srv/chatmail/dkim
+          cp -a /var/lib/acme/. /srv/chatmail/certs/ || true
+          cp -a /etc/dkimkeys/. /srv/chatmail/dkim/ || true
+          cp /etc/chatmail/chatmail.ini /srv/chatmail/chatmail.ini
+          EOF
+
+      - name: deploy with Docker
+        run: |
+          SHORT_SHA=$(echo "${{ github.sha }}" | cut -c1-7)
+          GHCR_IMAGE="ghcr.io/chatmail/docker:sha-${SHORT_SHA}"
+          rsync -avz --exclude='.git' --exclude='venv' --exclude='__pycache__' ./ $VPS:/srv/chatmail/relay/
+          echo "${{ secrets.GITHUB_TOKEN }}" | ssh $VPS "docker login ghcr.io -u ${{ github.actor }} --password-stdin && \
+            docker pull ${GHCR_IMAGE} && \
+            cd /srv/chatmail/relay && CHATMAIL_IMAGE=${GHCR_IMAGE} MAIL_DOMAIN=${{ inputs.mail_domain }} docker compose -f docker/docker-compose.yaml -f docker/docker-compose.ci.yaml up -d"
+
+      - name: wait for container healthy
+        run: |
+          ssh $VPS 'docker exec chatmail journalctl -f --no-pager' &
+          LOG_PID=$!
+          trap "kill $LOG_PID 2>/dev/null || true" EXIT
+          for i in $(seq 1 60); do
+            status=$(ssh $VPS 'docker inspect --format={{.State.Health.Status}} chatmail 2>/dev/null' || echo "missing")
+            echo "  [$i/60] status=$status"
+            if [ "$status" = "healthy" ]; then
+              echo "Container is healthy."
+              exit 0
+            fi
+            if [ "$status" = "unhealthy" ]; then
+              echo "Container is unhealthy!"
+              break
+            fi
+            sleep 5
+          done
+          echo "Container did not become healthy."
+          kill $LOG_PID 2>/dev/null || true
+          ssh $VPS bash -s <<'EOF'
+          echo "--- failed units ---"
+          docker exec chatmail systemctl --failed --no-pager || true
+          echo "--- service logs ---"
+          docker exec chatmail journalctl -u dovecot -u postfix -u nginx -u unbound --no-pager -n 50 || true
+          echo "--- listening ports ---"
+          docker exec chatmail ss -tlnp || true
+          echo "--- chatmail.ini ---"
+          docker exec chatmail cat /etc/chatmail/chatmail.ini || true
+          EOF
+          exit 1
+
+      - name: show container state
+        run: |
+          ssh $VPS bash -s <<'EOF'
+          echo "--- listening ports ---"
+          docker exec chatmail ss -tlnp
+          echo "--- chatmail.ini ---"
+          docker exec chatmail cat /etc/chatmail/chatmail.ini
+          EOF
+
+      - name: Docker integration tests
+        run: ssh $VPS 'docker exec chatmail cmdeploy test --slow --ssh-host @local'
+
+      - name: Docker DNS
+        run: |
+          git checkout .github/workflows/${{ inputs.zone_file }}
+          ssh $VPS bash -s <<'EOF'
+          docker exec chatmail chown opendkim:opendkim -R /etc/dkimkeys
+          docker exec chatmail cmdeploy dns --ssh-host @local --zonefile /opt/chatmail/staging.zone --verbose
+          docker cp chatmail:/opt/chatmail/staging.zone /tmp/staging.zone
+          EOF
+          scp $VPS:/tmp/staging.zone staging-generated.zone
+          cat staging-generated.zone >> .github/workflows/${{ inputs.zone_file }}
+          cat .github/workflows/${{ inputs.zone_file }}
+          scp .github/workflows/${{ inputs.zone_file }} root@ns.testrun.org:/etc/nsd/${{ inputs.staging_host }}.zone
+          ssh root@ns.testrun.org "nsd-checkzone ${{ inputs.staging_host }} /etc/nsd/${{ inputs.staging_host }}.zone && systemctl reload nsd"
+
+      - name: Docker final DNS check
+        run: ssh $VPS 'docker exec chatmail cmdeploy dns -v --ssh-host @local'

.github/workflows/staging-ipv4.testrun.org-default.zone

@@ -0,0 +1,20 @@
+;; Zone file for staging-ipv4.testrun.org
+
+$ORIGIN staging-ipv4.testrun.org.
+$TTL 300
+
+@  IN  SOA     ns.testrun.org. root.nine.testrun.org (
+  2023010101 ; Serial
+  7200       ; Refresh
+  3600       ; Retry
+  1209600    ; Expire
+  3600       ; Negative response caching TTL
+)
+
+;; Nameservers.
+@  IN NS ns.testrun.org.
+
+;; DNS records.
+@       IN      A       37.27.95.249
+mta-sts.staging-ipv4.testrun.org.    CNAME   staging-ipv4.testrun.org.
+www.staging-ipv4.testrun.org.        CNAME   staging-ipv4.testrun.org.

.github/workflows/staging.testrun.org-default.zone

@@ -0,0 +1,21 @@
+;; Zone file for staging2.testrun.org
+
+$ORIGIN staging2.testrun.org.
+$TTL 300
+
+@  IN  SOA     ns.testrun.org. root.nine.testrun.org (
+  2023010101 ; Serial
+  7200       ; Refresh
+  3600       ; Retry
+  1209600    ; Expire
+  3600       ; Negative response caching TTL
+)
+
+;; Nameservers.
+@  IN NS ns.testrun.org.
+
+;; DNS records.
+@       IN      A       37.27.24.139
+mta-sts.staging2.testrun.org.    CNAME   staging2.testrun.org.
+www.staging2.testrun.org.        CNAME   staging2.testrun.org.
+

.github/workflows/test-and-deploy-ipv4only.yaml

@@ -0,0 +1,121 @@
+name: deploy on staging-ipv4.testrun.org, and run tests
+
+on:
+  push:
+    branches:
+      - main
+      - j4n/docker-pr
+  pull_request:
+    paths-ignore:
+      - 'scripts/**'
+      - '**/README.md'
+      - 'CHANGELOG.md'
+      - 'LICENSE'
+
+jobs:
+  trigger-docker-build:
+    if: github.event_name == 'push'
+    uses: ./.github/workflows/trigger-docker-build.yaml
+    secrets: inherit
+
+  deploy:
+    name: deploy on staging-ipv4.testrun.org, and run tests
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    environment:
+      name: staging-ipv4.testrun.org
+      url: https://staging-ipv4.testrun.org/
+    concurrency: staging-ipv4.testrun.org
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: true
+
+      - name: prepare SSH
+        run: |
+          mkdir ~/.ssh
+          echo "${{ secrets.STAGING_SSH_KEY }}" >> ~/.ssh/id_ed25519
+          chmod 600 ~/.ssh/id_ed25519
+          ssh-keyscan staging-ipv4.testrun.org > ~/.ssh/known_hosts
+          # save previous acme & dkim state
+          rsync -avz root@staging-ipv4.testrun.org:/var/lib/acme acme-ipv4 || true
+          rsync -avz root@staging-ipv4.testrun.org:/etc/dkimkeys dkimkeys-ipv4 || true
+          # store previous acme & dkim state on ns.testrun.org, if it contains useful certs
+          if [ -f dkimkeys-ipv4/dkimkeys/opendkim.private ]; then rsync -avz -e "ssh -o StrictHostKeyChecking=accept-new" dkimkeys-ipv4 root@ns.testrun.org:/tmp/ || true; fi
+          if [ "$(ls -A acme-ipv4/acme/certs)" ]; then rsync -avz -e "ssh -o StrictHostKeyChecking=accept-new" acme-ipv4 root@ns.testrun.org:/tmp/ || true; fi
+          # make sure CAA record isn't set
+          scp -o StrictHostKeyChecking=accept-new .github/workflows/staging-ipv4.testrun.org-default.zone root@ns.testrun.org:/etc/nsd/staging-ipv4.testrun.org.zone
+          ssh root@ns.testrun.org sed -i '/CAA/d' /etc/nsd/staging-ipv4.testrun.org.zone
+          ssh root@ns.testrun.org nsd-checkzone staging-ipv4.testrun.org /etc/nsd/staging-ipv4.testrun.org.zone
+          ssh root@ns.testrun.org systemctl reload nsd
+
+      - name: rebuild staging-ipv4.testrun.org to have a clean VPS
+        run: |
+            curl -X POST \
+            -H "Authorization: Bearer ${{ secrets.HETZNER_API_TOKEN }}" \
+            -H "Content-Type: application/json" \
+            -d '{"image":"debian-12"}' \
+            "https://api.hetzner.cloud/v1/servers/${{ secrets.STAGING_IPV4_SERVER_ID }}/actions/rebuild"
+
+      - run: scripts/initenv.sh
+
+      - name: append venv/bin to PATH
+        run: echo venv/bin >>$GITHUB_PATH
+
+      - name: upload TLS cert after rebuilding
+        run: |
+          echo " --- wait until staging-ipv4.testrun.org VPS is rebuilt --- "
+          rm ~/.ssh/known_hosts
+          while ! ssh -o ConnectTimeout=180 -o StrictHostKeyChecking=accept-new -v root@staging-ipv4.testrun.org id -u ; do sleep 1 ; done
+          ssh -o StrictHostKeyChecking=accept-new -v root@staging-ipv4.testrun.org id -u
+          # download acme & dkim state from ns.testrun.org
+          rsync -e "ssh -o StrictHostKeyChecking=accept-new" -avz root@ns.testrun.org:/tmp/acme-ipv4/acme acme-restore || true
+          rsync -avz root@ns.testrun.org:/tmp/dkimkeys-ipv4/dkimkeys dkimkeys-restore || true
+          # restore acme & dkim state to staging-ipv4.testrun.org
+          rsync -avz acme-restore/acme root@staging-ipv4.testrun.org:/var/lib/ || true
+          rsync -avz dkimkeys-restore/dkimkeys root@staging-ipv4.testrun.org:/etc/ || true
+          ssh -o StrictHostKeyChecking=accept-new -v root@staging-ipv4.testrun.org chown root:root -R /var/lib/acme || true
+
+      - name: run deploy-chatmail offline tests
+        run: pytest --pyargs cmdeploy
+
+      - name: setup dependencies
+        run: |
+          ssh root@staging-ipv4.testrun.org apt update
+          ssh root@staging-ipv4.testrun.org apt install -y git python3.11-venv python3-dev gcc
+          ssh root@staging-ipv4.testrun.org git clone https://github.com/chatmail/relay
+          ssh root@staging-ipv4.testrun.org "cd relay && git checkout " ${{ github.head_ref }}
+          ssh root@staging-ipv4.testrun.org "cd relay && scripts/initenv.sh"
+
+      - name: initialize config
+        run: |
+          ssh root@staging-ipv4.testrun.org "cd relay && scripts/cmdeploy init staging-ipv4.testrun.org"
+          ssh root@staging-ipv4.testrun.org "sed -i 's#disable_ipv6 = False#disable_ipv6 = True#' relay/chatmail.ini"
+          ssh root@staging-ipv4.testrun.org "sed -i 's/#\s*mtail_address/mtail_address/' relay/chatmail.ini"
+
+      - run: ssh root@staging-ipv4.testrun.org "cd relay && scripts/cmdeploy run --verbose --skip-dns-check --ssh-host localhost"
+
+      - name: set DNS entries
+        run: |
+          ssh root@staging-ipv4.testrun.org "cd relay && scripts/cmdeploy dns --zonefile staging-generated.zone --ssh-host localhost"
+          ssh root@staging-ipv4.testrun.org cat relay/staging-generated.zone >> .github/workflows/staging-ipv4.testrun.org-default.zone
+          cat .github/workflows/staging-ipv4.testrun.org-default.zone
+          scp .github/workflows/staging-ipv4.testrun.org-default.zone root@ns.testrun.org:/etc/nsd/staging-ipv4.testrun.org.zone
+          ssh root@ns.testrun.org nsd-checkzone staging-ipv4.testrun.org /etc/nsd/staging-ipv4.testrun.org.zone
+          ssh root@ns.testrun.org systemctl reload nsd
+
+      - name: cmdeploy test
+        run: ssh root@staging-ipv4.testrun.org "cd relay && CHATMAIL_DOMAIN2=ci-chatmail.testrun.org scripts/cmdeploy test --slow --ssh-host localhost"
+
+      - name: cmdeploy dns
+        run: ssh root@staging-ipv4.testrun.org "cd relay && scripts/cmdeploy dns -v --ssh-host localhost"
+
+  deploy-docker:
+    needs: [deploy, trigger-docker-build]
+    if: github.event_name == 'push'
+    uses: ./.github/workflows/docker-deploy.yaml
+    with:
+      staging_host: staging-ipv4.testrun.org
+      mail_domain: staging-ipv4.testrun.org
+      zone_file: staging-ipv4.testrun.org-default.zone
+    secrets: inherit

.github/workflows/test-and-deploy.yaml

@@ -0,0 +1,112 @@
+name: deploy on staging2.testrun.org, and run tests
+
+on:
+  push:
+    branches:
+      - main
+      - j4n/docker-pr
+  pull_request:
+    paths-ignore:
+      - 'scripts/**'
+      - '**/README.md'
+      - 'CHANGELOG.md'
+      - 'LICENSE'
+
+jobs:
+  trigger-docker-build:
+    if: github.event_name == 'push'
+    uses: ./.github/workflows/trigger-docker-build.yaml
+    secrets: inherit
+
+  deploy:
+    name: deploy on staging2.testrun.org, and run tests
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    environment:
+      name: staging2.testrun.org
+      url: https://staging2.testrun.org/
+    concurrency: staging2.testrun.org
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: prepare SSH
+        run: |
+          mkdir ~/.ssh
+          echo "${{ secrets.STAGING_SSH_KEY }}" >> ~/.ssh/id_ed25519
+          chmod 600 ~/.ssh/id_ed25519
+          ssh-keyscan staging2.testrun.org > ~/.ssh/known_hosts
+          # save previous acme & dkim state
+          rsync -avz root@staging2.testrun.org:/var/lib/acme . || true
+          rsync -avz root@staging2.testrun.org:/etc/dkimkeys . || true
+          # store previous acme & dkim state on ns.testrun.org, if it contains useful certs
+          if [ -f dkimkeys/opendkim.private ]; then rsync -avz -e "ssh -o StrictHostKeyChecking=accept-new" dkimkeys root@ns.testrun.org:/tmp/ || true; fi
+          if [ "$(ls -A acme/certs)" ]; then rsync -avz -e "ssh -o StrictHostKeyChecking=accept-new" acme root@ns.testrun.org:/tmp/ || true; fi
+          # make sure CAA record isn't set
+          scp -o StrictHostKeyChecking=accept-new .github/workflows/staging.testrun.org-default.zone root@ns.testrun.org:/etc/nsd/staging2.testrun.org.zone
+          ssh root@ns.testrun.org sed -i '/CAA/d' /etc/nsd/staging2.testrun.org.zone
+          ssh root@ns.testrun.org nsd-checkzone staging2.testrun.org /etc/nsd/staging2.testrun.org.zone
+          ssh root@ns.testrun.org systemctl reload nsd
+
+      - name: rebuild staging2.testrun.org to have a clean VPS
+        run: |
+            curl -X POST \
+            -H "Authorization: Bearer ${{ secrets.HETZNER_API_TOKEN }}" \
+            -H "Content-Type: application/json" \
+            -d '{"image":"debian-12"}' \
+            "https://api.hetzner.cloud/v1/servers/${{ secrets.STAGING_SERVER_ID }}/actions/rebuild"
+
+      - run: scripts/initenv.sh
+
+      - name: append venv/bin to PATH
+        run: echo venv/bin >>$GITHUB_PATH
+
+      - name: upload TLS cert after rebuilding
+        run: |
+          echo " --- wait until staging2.testrun.org VPS is rebuilt --- "
+          rm ~/.ssh/known_hosts
+          while ! ssh -o ConnectTimeout=180 -o StrictHostKeyChecking=accept-new -v root@staging2.testrun.org id -u ; do sleep 1 ; done
+          ssh -o StrictHostKeyChecking=accept-new -v root@staging2.testrun.org id -u
+          # download acme & dkim state from ns.testrun.org
+          rsync -e "ssh -o StrictHostKeyChecking=accept-new" -avz root@ns.testrun.org:/tmp/acme acme-restore || true
+          rsync -avz root@ns.testrun.org:/tmp/dkimkeys dkimkeys-restore || true
+          # restore acme & dkim state to staging2.testrun.org
+          rsync -avz acme-restore/acme root@staging2.testrun.org:/var/lib/ || true
+          rsync -avz dkimkeys-restore/dkimkeys root@staging2.testrun.org:/etc/ || true
+          ssh -o StrictHostKeyChecking=accept-new -v root@staging2.testrun.org chown root:root -R /var/lib/acme || true
+
+      - name: add hpk42 key to staging server
+        run: ssh root@staging2.testrun.org 'curl -s https://github.com/hpk42.keys >> .ssh/authorized_keys'
+
+      - name: run deploy-chatmail offline tests 
+        run: pytest --pyargs cmdeploy 
+
+      - run: |
+          cmdeploy init staging2.testrun.org
+          sed -i 's/#\s*mtail_address/mtail_address/' chatmail.ini
+
+      - run: cmdeploy run --verbose --skip-dns-check
+
+      - name: set DNS entries
+        run: |
+          cmdeploy dns --zonefile staging-generated.zone --verbose
+          cat staging-generated.zone >> .github/workflows/staging.testrun.org-default.zone
+          cat .github/workflows/staging.testrun.org-default.zone
+          scp .github/workflows/staging.testrun.org-default.zone root@ns.testrun.org:/etc/nsd/staging2.testrun.org.zone
+          ssh root@ns.testrun.org nsd-checkzone staging2.testrun.org /etc/nsd/staging2.testrun.org.zone
+          ssh root@ns.testrun.org systemctl reload nsd
+
+      - name: cmdeploy test
+        run: CHATMAIL_DOMAIN2=ci-chatmail.testrun.org cmdeploy test --slow
+
+      - name: cmdeploy dns
+        run: cmdeploy dns -v
+
+  deploy-docker:
+    needs: [deploy, trigger-docker-build]
+    if: github.event_name == 'push'
+    uses: ./.github/workflows/docker-deploy.yaml
+    with:
+      staging_host: staging2.testrun.org
+      mail_domain: staging2.testrun.org
+      zone_file: staging.testrun.org-default.zone
+    secrets: inherit

.github/workflows/trigger-docker-build.yaml

@@ -0,0 +1,24 @@
+name: Trigger Docker image build
+
+on:
+  workflow_call:
+
+jobs:
+  trigger-docker-build:
+    name: Trigger Docker image build
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/github-script@v7
+        with:
+          github-token: ${{ secrets.CHATMAIL_DOCKER_DISPATCH_TOKEN }}
+          script: |
+            await github.rest.repos.createDispatchEvent({
+              owner: 'chatmail',
+              repo: 'docker',
+              event_type: 'relay-updated',
+              client_payload: {
+                relay_ref: context.ref,
+                relay_sha: context.sha,
+                relay_sha_short: context.sha.slice(0, 7)
+              }
+            })

.gitignore

@@ -5,7 +5,6 @@ __pycache__/
 *.swp
 *qr-*.png
 chatmail*.ini
-lxconfigs/
 
 
 # C extensions

chatmaild/pyproject.toml

@@ -6,10 +6,7 @@ build-backend = "setuptools.build_meta"
 name = "chatmaild"
 version = "0.3"
 dependencies = [
-  "aiosmtpd",
   "iniconfig",
-  "deltachat-rpc-server",
-  "deltachat-rpc-client",
   "filelock",
   "requests",
   "crypt-r >= 3.13.1 ; python_version >= '3.11'",
@@ -70,6 +67,7 @@ commands =
 deps = pytest 
        pdbpp 
        pytest-localserver
+       aiosmtpd
        execnet
 commands = pytest -v -rsXx {posargs}
 """

chatmaild/src/chatmaild/config.py

@@ -38,6 +38,7 @@ class Config:
         self.filtermail_smtp_port_incoming = int(
             params.get("filtermail_smtp_port_incoming", "10081")
         )
+        self.filtermail_http_port = int(params.get("filtermail_http_port", "10082"))
         self.postfix_reinject_port = int(params.get("postfix_reinject_port", "10025"))
         self.postfix_reinject_port_incoming = int(
             params.get("postfix_reinject_port_incoming", "10026")

chatmaild/src/chatmaild/tests/plugin.py

@@ -85,13 +85,13 @@ def mockout():
         captured_green = []
         captured_plain = []
 
-        def red(self, msg, **kw):
+        def red(self, msg):
             self.captured_red.append(msg)
 
-        def green(self, msg, **kw):
+        def green(self, msg):
             self.captured_green.append(msg)
 
-        def print(self, msg="", **kw):
+        def __call__(self, msg):
             self.captured_plain.append(msg)
 
     return MockOut()

cmdeploy/pyproject.toml

@@ -10,7 +10,6 @@ dependencies = [
   "pillow",
   "qrcode",
   "markdown",
-  "pytest",
   "setuptools>=68",
   "termcolor",
   "build",
@@ -21,6 +20,7 @@ dependencies = [
   "execnet",
   "imap_tools",
   "deltachat-rpc-client",
+  "deltachat-rpc-server",
 ]
 
 [project.scripts]

cmdeploy/src/cmdeploy/basedeploy.py

@@ -3,6 +3,8 @@ import io
 import os
 from contextlib import contextmanager
 
+from pyinfra import host
+from pyinfra.facts.server import Command
 from pyinfra.operations import files, server, systemd
 
 
@@ -11,6 +13,17 @@ def has_systemd():
     return os.path.isdir("/run/systemd/system")
 
 
+def is_in_container() -> bool:
+    """Return True if running inside a container (Docker, LXC, etc.)."""
+    return (
+        host.get_fact(
+            Command,
+            "systemd-detect-virt --container --quiet 2>/dev/null && echo yes || true",
+        )
+        == "yes"
+    )
+
+
 @contextmanager
 def blocked_service_startup():
     """Prevent services from auto-starting during package installation.

cmdeploy/src/cmdeploy/cmdeploy.py

@@ -15,27 +15,10 @@ from pathlib import Path
 import pyinfra
 from chatmaild.config import read_config, write_initial_config
 from packaging import version
+from termcolor import colored
 
 from . import dns, remote
-from .lxc.cli import (
-    lxc_start_cmd,
-    lxc_start_cmd_options,
-    lxc_status_cmd,
-    lxc_status_cmd_options,
-    lxc_stop_cmd,
-    lxc_stop_cmd_options,
-    lxc_test_cmd,
-    lxc_test_cmd_options,
-)
-from .lxc.incus import DNSConfigurationError
-from .sshexec import (
-    LocalExec,
-    SSHExec,
-    resolve_host_from_ssh_config,
-    resolve_key_from_ssh_config,
-)
-from .util import Out
-from .www import main as webdev_main
+from .sshexec import LocalExec, SSHExec
 
 #
 # cmdeploy sub commands and options
@@ -99,21 +82,18 @@ def run_cmd_options(parser):
         help="disable checks nslookup for dns",
     )
     add_ssh_host_option(parser)
-    add_ssh_config_option(parser)
 
 
 def run_cmd(args, out):
     """Deploy chatmail services on the remote server."""
 
     ssh_host = args.ssh_host if args.ssh_host else args.config.mail_domain
-    sshexec = get_sshexec(ssh_host, ssh_config=args.ssh_config)
+    sshexec = get_sshexec(ssh_host)
     require_iroh = args.config.enable_iroh_relay
     strict_tls = args.config.tls_cert_mode == "acme"
     if not args.dns_check_disabled:
         remote_data = dns.get_initial_remote_data(sshexec, args.config.mail_domain)
-        if not dns.check_initial_remote_data(
-            remote_data, strict_tls=strict_tls, print=out.red
-        ):
+        if not dns.check_initial_remote_data(remote_data, strict_tls=strict_tls, print=out.red):
             return 1
 
     env = os.environ.copy()
@@ -124,28 +104,11 @@ def run_cmd(args, out):
     if not args.dns_check_disabled:
         env["CHATMAIL_ADDR_V4"] = remote_data.get("A") or ""
         env["CHATMAIL_ADDR_V6"] = remote_data.get("AAAA") or ""
-    env["DEBIAN_FRONTEND"] = "noninteractive"
-    env["TERM"] = "linux"
     deploy_path = importlib.resources.files(__package__).joinpath("run.py").resolve()
     pyinf = "pyinfra --dry" if args.dry_run else "pyinfra"
 
     cmd = f"{pyinf} --ssh-user root {ssh_host} {deploy_path} -y"
-    ssh_config = args.ssh_config
-    if ssh_config:
-        ssh_config = str(Path(ssh_config).resolve())
-
-        # Use pyinfra's native SSH data keys to configure the connection directly
-        # rather than relying on paramiko config parsing (see also sshexec.py)
-        ip = resolve_host_from_ssh_config(ssh_host, ssh_config)
-        key = resolve_key_from_ssh_config(ssh_host, ssh_config)
-        data_args = f"--data ssh_hostname={ip} --data ssh_known_hosts_file=/dev/null"
-        if key:
-            data_args += f" --data ssh_key={key}"
-        cmd = f"{pyinf} --ssh-user root {ssh_host} {deploy_path} -y {data_args}"
-    if ssh_host in ["localhost", "@docker"]:
-        if ssh_host == "@docker":
-            env["CHATMAIL_NOPORTCHECK"] = "True"
-            env["CHATMAIL_NOSYSCTL"] = "True"
+    if ssh_host == "localhost":
         cmd = f"{pyinf} @local {deploy_path} -y"
 
     if version.parse(pyinfra.__version__) < version.parse("3"):
@@ -153,17 +116,10 @@ def run_cmd(args, out):
         return 1
 
     try:
-        ret = out.shell(cmd, env=env)
-        if ret:
-            out.red("Deploy failed")
-            return 1
+        out.check_call(cmd, env=env)
         if args.website_only:
             out.green("Website deployment completed.")
-        elif (
-            not args.dns_check_disabled
-            and strict_tls
-            and not remote_data["acme_account_url"]
-        ):
+        elif not args.dns_check_disabled and strict_tls and not remote_data["acme_account_url"]:
             out.red("Deploy completed but letsencrypt not configured")
             out.red("Run 'cmdeploy run' again")
         else:
@@ -180,16 +136,15 @@ def dns_cmd_options(parser):
         dest="zonefile",
         type=pathlib.Path,
         default=None,
-        help="write DNS records in standard BIND format to the given file",
+        help="write out a zonefile",
     )
     add_ssh_host_option(parser)
-    add_ssh_config_option(parser)
 
 
 def dns_cmd(args, out):
     """Check DNS entries and optionally generate dns zone file."""
     ssh_host = args.ssh_host if args.ssh_host else args.config.mail_domain
-    sshexec = get_sshexec(ssh_host, verbose=args.verbose, ssh_config=args.ssh_config)
+    sshexec = get_sshexec(ssh_host, verbose=args.verbose)
     tls_cert_mode = args.config.tls_cert_mode
     strict_tls = tls_cert_mode == "acme"
     remote_data = dns.get_initial_remote_data(sshexec, args.config.mail_domain)
@@ -220,14 +175,13 @@ def dns_cmd(args, out):
 
 def status_cmd_options(parser):
     add_ssh_host_option(parser)
-    add_ssh_config_option(parser)
 
 
 def status_cmd(args, out):
     """Display status for online chatmail instance."""
 
     ssh_host = args.ssh_host if args.ssh_host else args.config.mail_domain
-    sshexec = get_sshexec(ssh_host, verbose=args.verbose, ssh_config=args.ssh_config)
+    sshexec = get_sshexec(ssh_host, verbose=args.verbose)
 
     out.green(f"chatmail domain: {args.config.mail_domain}")
     if args.config.privacy_mail:
@@ -247,14 +201,15 @@ def test_cmd_options(parser):
         help="also run slow tests",
     )
     add_ssh_host_option(parser)
-    add_ssh_config_option(parser)
 
 
 def test_cmd(args, out):
     """Run local and online tests for chatmail deployment."""
 
     env = os.environ.copy()
-    env["CHATMAIL_INI"] = str(args.inipath.resolve())
+    env["CHATMAIL_INI"] = str(args.inipath.absolute())
+    if args.ssh_host:
+        env["CHATMAIL_SSH"] = args.ssh_host
 
     pytest_path = shutil.which("pytest")
     pytest_args = [
@@ -268,11 +223,7 @@ def test_cmd(args, out):
     ]
     if args.slow:
         pytest_args.append("--slow")
-    if args.ssh_host:
-        pytest_args.extend(["--ssh-host", args.ssh_host])
-    if args.ssh_config:
-        pytest_args.extend(["--ssh-config", str(Path(args.ssh_config).resolve())])
-    ret = out.shell(" ".join(pytest_args), env=env)
+    ret = out.run_ret(pytest_args, env=env)
     return ret
 
 
@@ -309,8 +260,8 @@ def fmt_cmd(args, out):
     format_args.extend(sources)
     check_args.extend(sources)
 
-    out.shell(" ".join(format_args), quiet=not args.verbose)
-    out.shell(" ".join(check_args), quiet=not args.verbose)
+    out.check_call(" ".join(format_args), quiet=not args.verbose)
+    out.check_call(" ".join(check_args), quiet=not args.verbose)
 
 
 def bench_cmd(args, out):
@@ -323,7 +274,9 @@ def bench_cmd(args, out):
 
 def webdev_cmd(args, out):
     """Run local web development loop for static web pages."""
-    webdev_main()
+    from .www import main
+
+    main()
 
 
 #
@@ -331,25 +284,41 @@ def webdev_cmd(args, out):
 #
 
 
+class Out:
+    """Convenience output printer providing coloring."""
+
+    def red(self, msg, file=sys.stderr):
+        print(colored(msg, "red"), file=file)
+
+    def green(self, msg, file=sys.stderr):
+        print(colored(msg, "green"), file=file)
+
+    def __call__(self, msg, red=False, green=False, file=sys.stdout):
+        color = "red" if red else ("green" if green else None)
+        print(colored(msg, color), file=file)
+
+    def check_call(self, arg, env=None, quiet=False):
+        if not quiet:
+            self(f"[$ {arg}]", file=sys.stderr)
+        return subprocess.check_call(arg, shell=True, env=env)
+
+    def run_ret(self, args, env=None, quiet=False):
+        if not quiet:
+            cmdstring = " ".join(args)
+            self(f"[$ {cmdstring}]", file=sys.stderr)
+        proc = subprocess.run(args, env=env, check=False)
+        return proc.returncode
+
+
 def add_ssh_host_option(parser):
     parser.add_argument(
         "--ssh-host",
         dest="ssh_host",
-        help="Run commands on 'localhost', via '@docker', or on a specific SSH host "
+        help="Run commands on 'localhost' or on a specific SSH host "
         "instead of chatmail.ini's mail_domain.",
     )
 
 
-def add_ssh_config_option(parser):
-    parser.add_argument(
-        "--ssh-config",
-        dest="ssh_config",
-        type=Path,
-        default=None,
-        help="Path to an SSH config file (e.g. lxconfigs/ssh-config).",
-    )
-
-
 def add_config_option(parser):
     parser.add_argument(
         "--config",
@@ -359,26 +328,25 @@ def add_config_option(parser):
         type=Path,
         help="path to the chatmail.ini file",
     )
+    parser.add_argument(
+        "--verbose",
+        "-v",
+        dest="verbose",
+        action="store_true",
+        default=False,
+        help="provide verbose logging",
+    )
 
 
-def add_subcommand(subparsers, func, add_config=True):
+def add_subcommand(subparsers, func):
     name = func.__name__
     assert name.endswith("_cmd")
-    name = name[:-4].replace("_", "-")
+    name = name[:-4]
     doc = func.__doc__.strip()
     help = doc.split("\n")[0].strip(".")
     p = subparsers.add_parser(name, description=doc, help=help)
     p.set_defaults(func=func)
-    if add_config:
-        add_config_option(p)
-    p.add_argument(
-        "-v",
-        "--verbose",
-        dest="verbose",
-        action="count",
-        default=0,
-        help="increase verbosity (can be repeated: -v, -vv)",
-    )
+    add_config_option(p)
     return p
 
 
@@ -387,60 +355,43 @@ Setup your chatmail server configuration and
 deploy it via SSH to your remote location.
 """
 
-# Explicit subcommand registry: (cmd_func, options_func_or_None, needs_config).
-# LXC commands don't need a chatmail.ini (no config); all others do.
-SUBCOMMANDS = [
-    (init_cmd, init_cmd_options, True),
-    (run_cmd, run_cmd_options, True),
-    (dns_cmd, dns_cmd_options, True),
-    (status_cmd, status_cmd_options, True),
-    (test_cmd, test_cmd_options, True),
-    (fmt_cmd, fmt_cmd_options, True),
-    (bench_cmd, None, True),
-    (webdev_cmd, None, True),
-    (lxc_start_cmd, lxc_start_cmd_options, False),
-    (lxc_stop_cmd, lxc_stop_cmd_options, False),
-    (lxc_status_cmd, lxc_status_cmd_options, False),
-    (lxc_test_cmd, lxc_test_cmd_options, False),
-]
-
 
 def get_parser():
     """Return an ArgumentParser for the 'cmdeploy' CLI"""
 
     parser = argparse.ArgumentParser(description=description.strip())
-    parser.set_defaults(func=None, inipath=None)
     subparsers = parser.add_subparsers(title="subcommands")
 
-    for func, addopts, needs_config in SUBCOMMANDS:
-        subparser = add_subcommand(subparsers, func, add_config=needs_config)
-        if addopts is not None:
-            addopts(subparser)
+    # find all subcommands in the module namespace
+    glob = globals()
+    for name, func in glob.items():
+        if name.endswith("_cmd"):
+            subparser = add_subcommand(subparsers, func)
+            addopts = glob.get(name + "_options")
+            if addopts is not None:
+                addopts(subparser)
 
     return parser
 
 
-def get_sshexec(ssh_host: str, verbose=True, ssh_config=None):
+def get_sshexec(ssh_host: str, verbose=True):
     if ssh_host in ["localhost", "@local"]:
-        return LocalExec(verbose, docker=False)
-    elif ssh_host == "@docker":
-        return LocalExec(verbose, docker=True)
+        return LocalExec(verbose)
     if verbose:
         print(f"[ssh] login to {ssh_host}")
-    return SSHExec(ssh_host, verbose=verbose, ssh_config=ssh_config)
+    return SSHExec(ssh_host, verbose=verbose)
 
 
 def main(args=None):
     """Provide main entry point for 'cmdeploy' CLI invocation."""
     parser = get_parser()
     args = parser.parse_args(args=args)
-    if args.func is None:
+    if not hasattr(args, "func"):
         return parser.parse_args(["-h"])
 
-    out = Out(sepchar="\u2501", verbosity=args.verbose)
+    out = Out()
     kwargs = {}
-
-    if args.inipath is not None and args.func.__name__ not in ("init_cmd", "fmt_cmd"):
+    if args.func.__name__ not in ("init_cmd", "fmt_cmd"):
         if not args.inipath.exists():
             out.red(f"expecting {args.inipath} to exist, run init first?")
             raise SystemExit(1)
@@ -455,9 +406,6 @@ def main(args=None):
         if res is None:
             res = 0
         return res
-    except DNSConfigurationError as exc:
-        out.red(str(exc))
-        return 1
     except KeyboardInterrupt:
         out.red("KeyboardInterrupt")
         sys.exit(130)

cmdeploy/src/cmdeploy/deployers.py

@@ -2,7 +2,6 @@
 Chat Mail pyinfra deploy.
 """
 
-import os
 import shutil
 import subprocess
 import sys
@@ -17,6 +16,8 @@ from pyinfra.facts.files import Sha256File
 from pyinfra.facts.systemd import SystemdEnabled
 from pyinfra.operations import apt, files, pip, server, systemd
 
+from cmdeploy.cmdeploy import Out
+
 from .acmetool import AcmetoolDeployer
 from .basedeploy import (
     Deployer,
@@ -26,6 +27,7 @@ from .basedeploy import (
     configure_remote_units,
     get_resource,
     has_systemd,
+    is_in_container,
 )
 from .dovecot.deployer import DovecotDeployer
 from .external.deployer import ExternalTlsDeployer
@@ -35,7 +37,6 @@ from .nginx.deployer import NginxDeployer
 from .opendkim.deployer import OpendkimDeployer
 from .postfix.deployer import PostfixDeployer
 from .selfsigned.deployer import SelfSignedTlsDeployer
-from .util import Out, get_version_string
 from .www import build_webpages, find_merge_conflict, get_paths
 
 
@@ -254,14 +255,8 @@ class WebsiteDeployer(Deployer):
                     logger.warning("Web page build failed, skipping website deployment")
                     return
             # if it is not a hugo page, upload it as is
-            # pyinfra files.rsync (experimental) causes problems with ssh-config configuration
-            # the stable files.sync should do
-            files.sync(
-                src=str(www_path),
-                dest="/var/www/html",
-                user="www-data",
-                group="www-data",
-                delete=True,
+            files.rsync(
+                f"{www_path}/", "/var/www/html", flags=["-avz", "--chown=www-data"]
             )
 
 
@@ -463,8 +458,9 @@ class ChatmailDeployer(Deployer):
         ("iroh", None, None),
     ]
 
-    def __init__(self, mail_domain):
-        self.mail_domain = mail_domain
+    def __init__(self, config):
+        self.config = config
+        self.mail_domain = config.mail_domain
 
     def install(self):
         files.put(
@@ -487,12 +483,18 @@ class ChatmailDeployer(Deployer):
             name="Install rsync",
             packages=["rsync"],
         )
-        apt.packages(
-            name="Ensure cron is installed",
-            packages=["cron"],
-        )
 
     def configure(self):
+        # metadata crashes if the mailboxes dir does not exist
+        files.directory(
+            name="Ensure vmail mailbox directory exists",
+            path=str(self.config.mailboxes_dir),
+            user="vmail",
+            group="vmail",
+            mode="700",
+            present=True,
+        )
+
         # This file is used by auth proxy.
         # https://wiki.debian.org/EtcMailName
         server.shell(
@@ -521,9 +523,17 @@ class FcgiwrapDeployer(Deployer):
 
 class GithashDeployer(Deployer):
     def activate(self):
+        try:
+            git_hash = subprocess.check_output(["git", "rev-parse", "HEAD"]).decode()
+        except Exception:
+            git_hash = "unknown\n"
+        try:
+            git_diff = subprocess.check_output(["git", "diff"]).decode()
+        except Exception:
+            git_diff = ""
         files.put(
             name="Upload chatmail relay git commit hash",
-            src=StringIO(get_version_string()),
+            src=StringIO(git_hash + git_diff),
             dest="/etc/chatmail-version",
             mode="700",
         )
@@ -567,20 +577,14 @@ def deploy_chatmail(config_path: Path, disable_mail: bool, website_only: bool) -
         )
 
     # Check if mtail_address interface is available (if configured)
-    if config.mtail_address and config.mtail_address not in (
-        "127.0.0.1",
-        "::1",
-        "localhost",
-    ):
+    if config.mtail_address and config.mtail_address not in ('127.0.0.1', '::1', 'localhost'):
         ipv4_addrs = host.get_fact(hardware.Ipv4Addrs)
         all_addresses = [addr for addrs in ipv4_addrs.values() for addr in addrs]
         if config.mtail_address not in all_addresses:
-            Out().red(
-                f"Deploy failed: mtail_address {config.mtail_address} is not available (VPN up?).\n"
-            )
+            Out().red(f"Deploy failed: mtail_address {config.mtail_address} is not available (VPN up?).\n")
             exit(1)
 
-    if not os.environ.get("CHATMAIL_NOPORTCHECK"):
+    if not is_in_container():
         port_services = [
             (["master", "smtpd"], 25),
             ("unbound", 53),
@@ -620,7 +624,7 @@ def deploy_chatmail(config_path: Path, disable_mail: bool, website_only: bool) -
     tls_deployer = get_tls_deployer(config, mail_domain)
 
     all_deployers = [
-        ChatmailDeployer(mail_domain),
+        ChatmailDeployer(config),
         LegacyRemoveDeployer(),
         FiltermailDeployer(),
         JournaldDeployer(),

cmdeploy/src/cmdeploy/dns.py

@@ -4,11 +4,7 @@ from . import remote
 
 
 def parse_zone_records(text):
-    """Yield ``(name, ttl, rtype, rdata)`` from standard BIND-format text.
-
-    Skips comment lines (starting with ``;``) and blank lines.
-    Each record line must have the format ``name TTL IN type rdata``.
-    """
+    """Yield ``(name, ttl, rtype, rdata)`` from standard BIND-format text."""
     for raw_line in text.splitlines():
         line = raw_line.strip()
         if not line or line.startswith(";"):
@@ -47,33 +43,36 @@ def get_filled_zone_file(remote_data):
         remote_data["sts_id"] = datetime.datetime.now().strftime("%Y%m%d%H%M")
 
     d = remote_data["mail_domain"]
+
+    def append_record(name, rtype, rdata, ttl=3600):
+        lines.append(f"{name:<40} {ttl:<6} IN  {rtype:<5}  {rdata}")
+
     lines = ["; Required DNS entries"]
     if remote_data.get("A"):
-        lines.append(f"{d}.  3600  IN  A  {remote_data['A']}")
+        append_record(f"{d}.", "A", remote_data["A"])
     if remote_data.get("AAAA"):
-        lines.append(f"{d}.  3600  IN  AAAA  {remote_data['AAAA']}")
-    lines.append(f"{d}.  3600  IN  MX  10 {d}.")
+        append_record(f"{d}.", "AAAA", remote_data["AAAA"])
+    append_record(f"{d}.", "MX", f"10 {d}.")
     if remote_data.get("strict_tls"):
-        lines.append(
-            f'_mta-sts.{d}.  3600  IN  TXT  "v=STSv1; id={remote_data["sts_id"]}"'
-        )
-        lines.append(f"mta-sts.{d}.  3600  IN  CNAME  {d}.")
-    lines.append(f"www.{d}.  3600  IN  CNAME  {d}.")
+        append_record(f"_mta-sts.{d}.", "TXT", f'"v=STSv1; id={remote_data["sts_id"]}"')
+        append_record(f"mta-sts.{d}.", "CNAME", f"{d}.")
+    append_record(f"www.{d}.", "CNAME", f"{d}.")
     lines.append(remote_data["dkim_entry"])
     lines.append("")
     lines.append("; Recommended DNS entries")
-    lines.append(f'{d}.  3600  IN  TXT  "v=spf1 a ~all"')
-    lines.append(f'_dmarc.{d}.  3600  IN  TXT  "v=DMARC1;p=reject;adkim=s;aspf=s"')
+    append_record(f"{d}.", "TXT", '"v=spf1 a ~all"')
+    append_record(f"_dmarc.{d}.", "TXT", '"v=DMARC1;p=reject;adkim=s;aspf=s"')
     if remote_data.get("acme_account_url"):
-        lines.append(
-            f"{d}.  3600  IN  CAA  0 issue"
-            f' "letsencrypt.org;accounturi={remote_data["acme_account_url"]}"'
+        append_record(
+            f"{d}.",
+            "CAA",
+            f'0 issue "letsencrypt.org;accounturi={remote_data["acme_account_url"]}"',
         )
-    lines.append(f'_adsp._domainkey.{d}.  3600  IN  TXT  "dkim=discardable"')
-    lines.append(f"_submission._tcp.{d}.  3600  IN  SRV  0 1 587 {d}.")
-    lines.append(f"_submissions._tcp.{d}.  3600  IN  SRV  0 1 465 {d}.")
-    lines.append(f"_imap._tcp.{d}.  3600  IN  SRV  0 1 143 {d}.")
-    lines.append(f"_imaps._tcp.{d}.  3600  IN  SRV  0 1 993 {d}.")
+    append_record(f"_adsp._domainkey.{d}.", "TXT", '"dkim=discardable"')
+    append_record(f"_submission._tcp.{d}.", "SRV", f"0 1 587 {d}.")
+    append_record(f"_submissions._tcp.{d}.", "SRV", f"0 1 465 {d}.")
+    append_record(f"_imap._tcp.{d}.", "SRV", f"0 1 143 {d}.")
+    append_record(f"_imaps._tcp.{d}.", "SRV", f"0 1 993 {d}.")
     lines.append("")
     return "\n".join(lines)
 
@@ -91,19 +90,19 @@ def check_full_zone(sshexec, remote_data, out, zonefile) -> int:
     if required_diff:
         out.red("Please set required DNS entries at your DNS provider:\n")
         for line in required_diff:
-            out.print(line)
-        out.print()
+            out(line)
+        out("")
         returncode = 1
     if remote_data.get("dkim_entry") in required_diff:
-        out.print(
+        out(
             "If the DKIM entry above does not work with your DNS provider,"
             " you can try this one:\n"
         )
-        out.print(remote_data.get("web_dkim_entry") + "\n")
+        out(remote_data.get("web_dkim_entry") + "\n")
     if recommended_diff:
-        out.print("WARNING: these recommended DNS entries are not set:\n")
+        out("WARNING: these recommended DNS entries are not set:\n")
         for line in recommended_diff:
-            out.print(line)
+            out(line)
 
     if not (recommended_diff or required_diff):
         out.green("Great! All your DNS entries are verified and correct.")

cmdeploy/src/cmdeploy/dovecot/deployer.py

@@ -1,9 +1,10 @@
+import io
 import urllib.request
 
 from chatmaild.config import Config
 from pyinfra import host
-from pyinfra.facts.server import Arch, Command, Sysctl
-from pyinfra.facts.systemd import SystemdEnabled
+from pyinfra.facts.deb import DebPackages
+from pyinfra.facts.server import Arch, Sysctl
 from pyinfra.operations import apt, files, server, systemd
 
 from cmdeploy.basedeploy import (
@@ -12,9 +13,21 @@ from cmdeploy.basedeploy import (
     blocked_service_startup,
     configure_remote_units,
     get_resource,
-    has_systemd,
+    is_in_container,
 )
 
+DOVECOT_ARCHIVE_VERSION = "2.3.21+dfsg1-3"
+DOVECOT_PACKAGE_VERSION = f"1:{DOVECOT_ARCHIVE_VERSION}"
+
+DOVECOT_SHA256 = {
+    ("core", "amd64"): "dd060706f52a306fa863d874717210b9fe10536c824afe1790eec247ded5b27d",
+    ("core", "arm64"): "e7548e8a82929722e973629ecc40fcfa886894cef3db88f23535149e7f730dc9",
+    ("imapd", "amd64"): "8d8dc6fc00bbb6cdb25d345844f41ce2f1c53f764b79a838eb2a03103eebfa86",
+    ("imapd", "arm64"): "178fa877ddd5df9930e8308b518f4b07df10e759050725f8217a0c1fb3fd707f",
+    ("lmtpd", "amd64"): "2f69ba5e35363de50962d42cccbfe4ed8495265044e244007d7ccddad77513ab",
+    ("lmtpd", "arm64"): "89f52fb36524f5877a177dff4a713ba771fd3f91f22ed0af7238d495e143b38f",
+}
+
 
 class DovecotDeployer(Deployer):
     daemon_reload = False
@@ -26,17 +39,43 @@ class DovecotDeployer(Deployer):
 
     def install(self):
         arch = host.get_fact(Arch)
-        if has_systemd() and "dovecot.service" in host.get_fact(SystemdEnabled):
-            return  # already installed and running
-
         with blocked_service_startup():
-            _install_dovecot_package("core", arch)
-            _install_dovecot_package("imapd", arch)
-            _install_dovecot_package("lmtpd", arch)
+            debs = []
+            for pkg in ("core", "imapd", "lmtpd"):
+                deb, changed = _download_dovecot_package(pkg, arch)
+                self.need_restart |= changed
+                if deb:
+                    debs.append(deb)
+            if debs:
+                deb_list = " ".join(debs)
+                # First dpkg may fail on missing dependencies (stderr suppressed);
+                # apt-get --fix-broken pulls them in, then dpkg retries cleanly.
+                server.shell(
+                    name="Install dovecot packages",
+                    commands=[
+                        f"dpkg --force-confdef --force-confold -i {deb_list} 2> /dev/null || true",
+                        "DEBIAN_FRONTEND=noninteractive apt-get -y --fix-broken install",
+                        f"dpkg --force-confdef --force-confold -i {deb_list}",
+                    ],
+                )
+                self.need_restart = True
+        files.put(
+            name="Pin dovecot packages to block Debian dist-upgrades",
+            src=io.StringIO(
+                "Package: dovecot-*\n"
+                "Pin: version *\n"
+                "Pin-Priority: -1\n"
+            ),
+            dest="/etc/apt/preferences.d/pin-dovecot",
+            user="root",
+            group="root",
+            mode="644",
+        )
 
     def configure(self):
         configure_remote_units(self.config.mail_domain, self.units)
-        self.need_restart, self.daemon_reload = _configure_dovecot(self.config)
+        config_restart, self.daemon_reload = _configure_dovecot(self.config)
+        self.need_restart |= config_restart
 
     def activate(self):
         activate_remote_units(self.units)
@@ -65,43 +104,40 @@ def _pick_url(primary, fallback):
         return fallback
 
 
-def _install_dovecot_package(package: str, arch: str):
+def _download_dovecot_package(package: str, arch: str) -> tuple[str | None, bool]:
+    """Download a dovecot .deb if needed, return (path, changed)."""
     arch = "amd64" if arch == "x86_64" else arch
     arch = "arm64" if arch == "aarch64" else arch
-    primary_url = f"https://download.delta.chat/dovecot/dovecot-{package}_2.3.21%2Bdfsg1-3_{arch}.deb"
-    fallback_url = f"https://github.com/chatmail/dovecot/releases/download/upstream%2F2.3.21%2Bdfsg1/dovecot-{package}_2.3.21%2Bdfsg1-3_{arch}.deb"
-    url = _pick_url(primary_url, fallback_url)
-    deb_filename = "/root/" + url.split("/")[-1]
 
-    match (package, arch):
-        case ("core", "amd64"):
-            sha256 = "dd060706f52a306fa863d874717210b9fe10536c824afe1790eec247ded5b27d"
-        case ("core", "arm64"):
-            sha256 = "e7548e8a82929722e973629ecc40fcfa886894cef3db88f23535149e7f730dc9"
-        case ("imapd", "amd64"):
-            sha256 = "8d8dc6fc00bbb6cdb25d345844f41ce2f1c53f764b79a838eb2a03103eebfa86"
-        case ("imapd", "arm64"):
-            sha256 = "178fa877ddd5df9930e8308b518f4b07df10e759050725f8217a0c1fb3fd707f"
-        case ("lmtpd", "amd64"):
-            sha256 = "2f69ba5e35363de50962d42cccbfe4ed8495265044e244007d7ccddad77513ab"
-        case ("lmtpd", "arm64"):
-            sha256 = "89f52fb36524f5877a177dff4a713ba771fd3f91f22ed0af7238d495e143b38f"
-        case _:
-            apt.packages(packages=[f"dovecot-{package}"])
-            return
+    pkg_name = f"dovecot-{package}"
+    sha256 = DOVECOT_SHA256.get((package, arch))
+    if sha256 is None:
+        op = apt.packages(packages=[pkg_name])
+        return None, bool(getattr(op, "changed", False))
+
+    installed_versions = host.get_fact(DebPackages).get(pkg_name, [])
+    if DOVECOT_PACKAGE_VERSION in installed_versions:
+        return None, False
+
+    url_version = DOVECOT_ARCHIVE_VERSION.replace("+", "%2B")
+    deb_base = f"{pkg_name}_{url_version}_{arch}.deb"
+    primary_url = f"https://download.delta.chat/dovecot/{deb_base}"
+    fallback_url = f"https://github.com/chatmail/dovecot/releases/download/upstream%2F{url_version}/{deb_base}"
+    url = _pick_url(primary_url, fallback_url)
+    deb_filename = f"/root/{deb_base}"
 
     files.download(
-        name=f"Download dovecot-{package}",
+        name=f"Download {pkg_name}",
         src=url,
         dest=deb_filename,
         sha256sum=sha256,
         cache_time=60 * 60 * 24 * 365 * 10,  # never redownload the package
     )
 
-    apt.deb(name=f"Install dovecot-{package}", src=deb_filename)
+    return deb_filename, True
 
 
-def _configure_dovecot(config: Config, debug: bool = False) -> (bool, bool):
+def _configure_dovecot(config: Config, debug: bool = False) -> tuple[bool, bool]:
     """Configures Dovecot IMAP server."""
     need_restart = False
     daemon_reload = False
@@ -136,7 +172,7 @@ def _configure_dovecot(config: Config, debug: bool = False) -> (bool, bool):
 
     # as per https://doc.dovecot.org/2.3/configuration_manual/os/
     # it is recommended to set the following inotify limits
-    can_modify = host.get_fact(Command, "systemd-detect-virt -c || true") == "none"
+    can_modify = not is_in_container()
     for name in ("max_user_instances", "max_user_watches"):
         key = f"fs.inotify.{name}"
         value = host.get_fact(Sysctl)[key]
@@ -144,8 +180,8 @@ def _configure_dovecot(config: Config, debug: bool = False) -> (bool, bool):
             continue
         if not can_modify:
             print(
-                "\n!!!! refusing to attempt sysctl setting in shared-kernel containers\n"
-                f"!!!! dovecot: sysctl {key!r}={value}, should be >65535 for production setups\n"
+                "\n!!!! refusing to attempt sysctl setting in containers\n"
+                f"!!!! dovecot: sysctl {key!r}={value}, should be >65534 for production setups\n"
                 "!!!!"
             )
             continue

cmdeploy/src/cmdeploy/filtermail/deployer.py

@@ -14,10 +14,10 @@ class FiltermailDeployer(Deployer):
 
     def install(self):
         arch = host.get_fact(facts.server.Arch)
-        url = f"https://github.com/chatmail/filtermail/releases/download/v0.5.2/filtermail-{arch}"
+        url = f"https://github.com/chatmail/filtermail/releases/download/v0.6.1/filtermail-{arch}"
         sha256sum = {
-            "x86_64": "ce24ca0075aa445510291d775fb3aea8f4411818c7b885ae51a0fe18c5f789ce",
-            "aarch64": "c5d783eefa5332db3d97a0e6a23917d72849e3eb45da3d16ce908a9b4e5a797d",
+            "x86_64": "48b3fb80c092d00b9b0a0ef77a8673496da3b9aed5ec1851e1df936d5589d62f",
+            "aarch64": "c65bd5f45df187d3d65d6965a285583a3be0f44a6916ff12909ff9a8d702c22e",
         }[arch]
         self.need_restart |= files.download(
             name="Download filtermail",

cmdeploy/src/cmdeploy/lxc/cli.py

@@ -1,454 +0,0 @@
-"""lxc-start/stop/status/test subcommands for testing with local containers."""
-
-import os
-import time
-
-from ..util import get_git_hash, get_version_string, shell
-from .incus import RELAY_IMAGE_ALIAS, Incus, RelayContainer
-
-RELAY_NAMES = ("test0", "test1")
-
-
-# -------------------------------------------------------------------
-# lxc-start
-# -------------------------------------------------------------------
-
-
-def lxc_start_cmd_options(parser):
-    _add_name_args(
-        parser,
-        help_text="User relay name(s) to create (default: test0).",
-    )
-    parser.add_argument(
-        "--ipv4-only",
-        dest="ipv4_only",
-        action="store_true",
-        help="Create an IPv4-only container.",
-    )
-    parser.add_argument(
-        "--run",
-        action="store_true",
-        help="Run 'cmdeploy run' on each container after starting it.",
-    )
-
-
-def lxc_start_cmd(args, out):
-    """Create/Ensure and start LXC relay and DNS containers."""
-
-    with out.section("Preparing container setup"):
-        _lxc_start_cmd(args, out)
-
-
-def _lxc_start_cmd(args, out):
-    ix = Incus(out)
-    sub = out.new_prefixed_out()
-    out.green("Ensuring base image ...")
-    ix.ensure_base_image()
-    out.green("Ensuring DNS container (ns-localchat) ...")
-    dns_ct = ix.get_dns_container()
-    dns_ct.ensure()
-    sub.print(f"DNS container IP: {dns_ct.ipv4}")
-
-    names = args.names if args.names else RELAY_NAMES
-    relays = list(ix.get_container(n) for n in names)
-    for ct in relays:
-        out.green(f"Ensuring container {ct.name!r} ({ct.domain}) ...")
-        ct.ensure()
-        ip = ct.ipv4
-
-        sub.print("Configuring container hostname ...")
-        ct.configure_hosts(ip)
-
-        sub.print(f"Writing {ct.ini.name} ...")
-        ct.write_ini(disable_ipv6=args.ipv4_only)
-        sub.print(f"Config: {ct.ini}")
-        if args.ipv4_only:
-            ct.disable_ipv6()
-            ipv6 = None
-        else:
-            output = ct.bash(
-                "ip -6 addr show scope global -deprecated"
-                " | grep -oP '(?<=inet6 )[^/]+'",
-                check=False,
-            )
-            ipv6 = output.strip() if output else None
-        sub.print(f"{_format_addrs(ip, ipv6)}")
-
-        sub.green(f"Container {ct.name!r} ready: {ct.domain} -> {ip}")
-        out.print()
-
-    # Reset DNS zones only for the containers we just started
-    started_cnames = {ct.name for ct in relays}
-    managed = ix.list_managed()
-    started = [c for c in managed if c["name"] in started_cnames]
-
-    if started:
-        out.print(
-            f"Resetting DNS zones for {len(started)} domain(s) (A + AAAA records) ..."
-        )
-        dns_ct.reset_dns_records(dns_ct.ipv4, started)
-
-        for ct in relays:
-            if ct.name in started_cnames:
-                sub.print(f"Configuring DNS in {ct.name} ...")
-                ct.configure_dns(dns_ct.ipv4)
-
-    # Generate the unified SSH config
-    out.green("Writing ssh-config ...")
-    ssh_cfg = ix.write_ssh_config()
-    sub.print(f"{ssh_cfg}")
-
-    # Verify SSH via the generated config
-    for ct in relays:
-        sub.print(f"Verifying SSH to {ct.name} via ssh-config ...")
-        if ct.verify_ssh(ssh_cfg):
-            sub.print(f"SSH OK: ssh -F lxconfigs/ssh-config {ct.domain}")
-        else:
-            sub.red(f"WARNING: SSH verification failed for {ct.name}")
-
-    # Print integration suggestions
-    ssh_cfg = ix.ssh_config_path
-    if not ix.check_ssh_include():
-        sub.green(
-            "\n(Optional) To use containers from any SSH client, add to ~/.ssh/config:"
-        )
-        sub.green(f"    Include {ssh_cfg}")
-
-    # Optionally run cmdeploy run on each relay
-    if args.run:
-        for ct in relays:
-            with out.section(f"cmdeploy run: {ct.sname} ({ct.domain})"):
-                ret = _run_cmdeploy("run", ct, ix, out, extra=["--skip-dns-check"])
-                if ret:
-                    out.red(f"Deploy to {ct.sname} failed (exit {ret})")
-                    return ret
-
-
-# -------------------------------------------------------------------
-# lxc-stop
-# -------------------------------------------------------------------
-
-
-def lxc_stop_cmd_options(parser):
-    parser.add_argument(
-        "--destroy",
-        action="store_true",
-        help="Delete containers and their config files after stopping.",
-    )
-    parser.add_argument(
-        "--destroy-all",
-        dest="destroy_all",
-        action="store_true",
-        help="Like --destroy, but also remove the ns-localchat DNS container.",
-    )
-    _add_name_args(
-        parser,
-        help_text="Container name(s) to stop (default: test0 + test1).",
-    )
-
-
-def lxc_stop_cmd(args, out):
-    """Stop (and optionally destroy) local LXC relay containers."""
-    ix = Incus(out)
-    names = args.names or RELAY_NAMES
-    destroy = args.destroy or args.destroy_all
-
-    for ct in map(ix.get_container, names):
-        if destroy:
-            out.green(f"Destroying container {ct.name!r} ...")
-            ct.destroy()
-        else:
-            out.green(f"Stopping container {ct.name!r} ...")
-            ct.stop(force=True)
-
-    if args.destroy_all:
-        dns_ct = ix.get_dns_container()
-        out.green(f"Destroying DNS container {dns_ct.name!r} ...")
-        dns_ct.destroy()
-        ix.delete_images()
-
-    if destroy:
-        ix.write_ssh_config()
-        out.green("LXC containers destroyed.")
-    else:
-        out.green("LXC containers stopped.")
-
-
-# -------------------------------------------------------------------
-# lxc-test
-# -------------------------------------------------------------------
-
-
-def lxc_test_cmd_options(parser):
-    parser.add_argument(
-        "--one",
-        action="store_true",
-        help="Only deploy and test against test0 (skip test1).",
-    )
-
-
-def lxc_test_cmd(args, out):
-    """Run full LXC pipeline: start, deploy, DNS, zone files, and tests.
-
-    All commands run directly on the host using
-    ``--ssh-config lxconfigs/ssh-config`` for SSH access.
-    """
-    ix = Incus(out)
-    t_total = time.time()
-    relay_names = list(RELAY_NAMES)
-    if args.one:
-        relay_names = relay_names[:1]
-
-    local_hash = get_git_hash()
-
-    # Per-relay: start, deploy, then snapshot the first relay as a
-    # reusable image so the second relay launches pre-deployed.
-    ipv4_only_flags = {RELAY_NAMES[0]: False, RELAY_NAMES[1]: True}
-
-    for ct in map(ix.get_container, relay_names):
-        name = ct.sname
-        ipv4_only = ipv4_only_flags.get(name, False)
-        v_flag = " -" + "v" * out.verbosity if out.verbosity > 0 else ""
-        start_cmd = f"cmdeploy lxc-start{v_flag} {name}"
-        if ipv4_only:
-            start_cmd += " --ipv4-only"
-        with out.section(f"cmdeploy lxc-start: {name}"):
-            ret = out.shell(start_cmd, cwd=str(ix.project_root))
-            if ret:
-                return ret
-
-        status = _deploy_status(ct, local_hash, ix)
-        with out.section(f"cmdeploy run: {name}"):
-            if "IN-SYNC" in status:
-                out.print(f"{name} is {status}, skipping")
-            else:
-                ret = _run_cmdeploy("run", ct, ix, out, extra=["--skip-dns-check"])
-                if ret:
-                    out.red(f"Deploy to {name} failed (exit {ret})")
-                    return ret
-
-        # Snapshot the first relay so subsequent ones launch pre-deployed
-        if not ix.find_image([RELAY_IMAGE_ALIAS]):
-            with out.section("lxc-test: caching relay image"):
-                ct.publish_as_relay_image()
-
-    for ct in map(ix.get_container, relay_names):
-        with out.section(f"cmdeploy dns: {ct.sname} ({ct.domain})"):
-            ret = _run_cmdeploy("dns", ct, ix, out, extra=["--zonefile", str(ct.zone)])
-            if ret:
-                out.red(f"DNS for {ct.sname} failed (exit {ret})")
-                return ret
-
-    with out.section(f"lxc-test: loading DNS zones {' & '.join(relay_names)}"):
-        dns_ct = ix.get_dns_container()
-        for ct in map(ix.get_container, relay_names):
-            if ct.zone.exists():
-                zone_data = ct.zone.read_text()
-                out.print(f"Loading {ct.zone} into PowerDNS ...")
-                dns_ct.set_dns_records(zone_data)
-
-    with out.section("cmdeploy test"):
-        first = ix.get_container(relay_names[0])
-        env = None
-        if len(relay_names) > 1:
-            env = os.environ.copy()
-            env["CHATMAIL_DOMAIN2"] = ix.get_container(relay_names[1]).domain
-        ret = _run_cmdeploy("test", first, ix, out, **({"env": env} if env else {}))
-        if ret:
-            out.red(f"Tests failed (exit {ret})")
-            return ret
-
-    elapsed = time.time() - t_total
-    out.section_line(f"lxc-test complete ({elapsed:.1f}s)")
-    if out.section_timings:
-        out.print("Section timings:")
-        for name, secs in out.section_timings:
-            out.print(f"  {name:.<50s} {secs:5.1f}s")
-        out.print(f"  {'total':.<50s} {elapsed:5.1f}s")
-    out.section_timings.clear()
-    return 0
-
-
-# -------------------------------------------------------------------
-# lxc-status
-# -------------------------------------------------------------------
-
-
-def lxc_status_cmd_options(parser):
-    pass
-
-
-def lxc_status_cmd(args, out):
-    """Show status of local LXC chatmail containers."""
-    ix = Incus(out)
-    containers = ix.list_managed()
-    if not containers:
-        out.red("No LXC containers found.  Run 'cmdeploy lxc-start' first.")
-        return 1
-
-    local_hash = get_git_hash()
-
-    # Get storage pool path for display
-    storage_path = None
-    data = ix.run_json(["storage", "show", "default"], check=False)
-    if data:
-        storage_path = data.get("config", {}).get("source")
-    msg = "Container status"
-    if storage_path:
-        msg += f": {storage_path}"
-    out.section_line(msg)
-
-    dns_ip = None
-    for c in containers:
-        _print_container_status(out, c, ix, local_hash)
-        if c["name"] == ix.get_dns_container().name:
-            dns_ip = c["ip"]
-
-    out.section_line("Host ssh and DNS configuration")
-    _print_ssh_status(out, ix)
-    _print_dns_forwarding_status(out, dns_ip)
-    return 0
-
-
-def _print_container_status(out, c, ix, local_hash):
-    """Print name/status, domain/IPs, and RAM for one container."""
-    cname = c["name"]
-    is_running = c.get("status") == "Running"
-    ct = ix.get_container(cname)
-
-    # First line: name + running/STOPPED + deploy status
-    if not is_running:
-        tag = "STOPPED"
-    elif not isinstance(ct, RelayContainer):
-        tag = "running"
-    else:
-        tag = f"running {_deploy_status(ct, local_hash, ix)}"
-    out.print(f"{cname:20s} {tag}")
-
-    # Second line: domain, IPv4, IPv6
-    domain = c.get("domain", "")
-    ip = c.get("ip") or "?"
-    ipv6 = c.get("ipv6")
-    out.print(f"{domain:20s} {_format_addrs(ip, ipv6)}")
-
-    # Third line: RAM (RSS), config
-    detail_out = out.new_prefixed_out(" " * 21)
-    try:
-        used, total = ct.rss_mib()
-    except Exception:
-        ram_str = "RSS ?"
-    else:
-        ram_str = f"RSS {used}/{total} MiB ({used * 100 // total}%)"
-
-    if isinstance(ct, RelayContainer):
-        detail = f"{ram_str}, config: {os.path.relpath(ct.ini)}"
-    else:
-        detail = ram_str
-
-    detail_out.print(detail)
-    out.print()
-
-
-def _print_ssh_status(out, ix):
-    """Print SSH integration status."""
-    ssh_cfg = ix.ssh_config_path
-    if ix.check_ssh_include():
-        out.green("SSH: ~/.ssh/config includes lxconfigs/ssh-config ✓")
-    else:
-        out.red("SSH: ~/.ssh/config does NOT include lxconfigs/ssh-config")
-        sub = out.new_prefixed_out()
-        sub.print("Add to ~/.ssh/config:")
-        sub.print(f"    Include {ssh_cfg}")
-
-
-def _print_dns_forwarding_status(out, dns_ip):
-    """Print host DNS forwarding status for .localchat."""
-    sub = out.new_prefixed_out()
-    if not dns_ip:
-        out.red("DNS: ns-localchat container not found")
-        return
-    try:
-        rv = shell("resolvectl status incusbr0")
-        dns_ok = dns_ip in rv.stdout and "localchat" in rv.stdout
-    except Exception:
-        dns_ok = None
-    if dns_ok is True:
-        out.green(f"DNS: .localchat forwarding to {dns_ip} ✓")
-    elif dns_ok is False:
-        out.red("DNS: .localchat forwarding NOT configured")
-        sub.print("Run:")
-        sub.print(f"    sudo resolvectl dns incusbr0 {dns_ip}")
-        sub.print("    sudo resolvectl domain incusbr0 ~localchat")
-    else:
-        sub.print("DNS: .localchat forwarding status UNKNOWN")
-
-
-# -------------------------------------------------------------------
-# Internal helpers
-# -------------------------------------------------------------------
-
-
-def _format_addrs(ip, ipv6=None):
-    parts = [f"IPv4 {ip}"]
-    if ipv6:
-        parts.append(f"IPv6 {ipv6}")
-    return ", ".join(parts)
-
-
-def _deploy_status(ct, local_hash, ix):
-    """Return a human-readable deploy status string.
-
-    Compares the full deployed version (hash + diff) against
-    the local state built by :func:`~cmdeploy.util.get_version_string`.
-    """
-    deployed = ct.deployed_version()
-    if deployed is None:
-        return "NOT DEPLOYED"
-
-    # A container launched from the relay image has the same
-    # git hash but a different domain — always redeploy.
-    deployed_domain = ct.deployed_domain()
-    if deployed_domain and deployed_domain != ct.domain:
-        return f"DOMAIN-MISMATCH (deployed: {deployed_domain})"
-
-    deployed_lines = deployed.splitlines()
-    deployed_hash = deployed_lines[0] if deployed_lines else ""
-    short = deployed_hash[:12]
-
-    if not local_hash:
-        return f"UNKNOWN (deployed: {short})"
-
-    local_short = local_hash[:12]
-    if deployed_hash != local_hash:
-        return f"STALE (deployed: {short}, local: {local_short})"
-
-    # Hash matches — check for uncommitted diffs
-    local_version = get_version_string()
-    if deployed != local_version:
-        return f"DIRTY ({local_short}, undeployed changes)"
-
-    return f"IN-SYNC ({short})"
-
-
-def _add_name_args(parser, help_text):
-    parser.add_argument("names", nargs="*", metavar="NAME", help=help_text)
-
-
-def _run_cmdeploy(subcmd, ct, ix, out, extra=None, **kwargs):
-    """Run ``cmdeploy <subcmd>`` with standard --config/--ssh flags.
-
-    *ct* is a Container (uses ``ct.ini`` and ``ct.domain``).
-    Returns the subprocess exit code.
-    """
-    extra_str = " ".join(extra) if extra else ""
-    v_flag = " -" + "v" * out.verbosity if out.verbosity > 0 else ""
-    cmd = f"""
-        cmdeploy {subcmd}{v_flag}
-        --config {ct.ini}
-        --ssh-config {ix.ssh_config_path}
-        --ssh-host {ct.domain}
-        {extra_str}
-    """
-    if "cwd" not in kwargs:
-        kwargs["cwd"] = str(ix.project_root)
-    return out.shell(cmd, **kwargs)

cmdeploy/src/cmdeploy/lxc/incus.py

@@ -1,739 +0,0 @@
-"""Core Incus operations for local chatmail LXC containers."""
-
-import json
-import subprocess
-import textwrap
-import time
-from pathlib import Path
-
-from ..util import shell
-
-LABEL_KEY = "user.localchat-managed"
-SSH_KEY_NAME = "id_localchat"
-DOMAIN_SUFFIX = ".localchat"
-UPSTREAM_IMAGE = "images:debian/12"
-BASE_IMAGE_ALIAS = "localchat-base"
-BASE_SETUP_NAME = "localchat-base-setup"
-RELAY_IMAGE_ALIAS = "localchat-relay"
-
-DNS_CONTAINER_NAME = "ns-localchat"
-DNS_DOMAIN = "ns.localchat"
-
-
-class DNSConfigurationError(Exception):
-    """Raised when the DNS container is not reachable or not answering."""
-
-
-def _extract_ip(net_data, family="inet"):
-    """Extract the first global-scope IP of *family* from network state data.
-
-    *net_data* is the ``state.network`` dict from ``incus list --format=json``.
-    *family* is ``"inet"`` for IPv4 or ``"inet6"`` for IPv6.
-    Returns the address string, or None.
-    """
-    for iface_name, iface in net_data.items():
-        if iface_name == "lo":
-            continue
-        for addr in iface.get("addresses", []):
-            if addr["family"] == family and addr["scope"] == "global":
-                return addr["address"]
-    return None
-
-
-class Incus:
-    """Gateway for all Incus container operations.
-
-    Instantiated once per CLI command and passed around so that
-    all modules share a single entry point for Incus interactions.
-    """
-
-    def __init__(self, out):
-        self.out = out
-        self.project_root = Path(__file__).resolve().parent.parent.parent.parent.parent
-        self.lxconfigs_dir = self.project_root / "lxconfigs"
-        self.lxconfigs_dir.mkdir(exist_ok=True)
-        self.ssh_key_path = self.lxconfigs_dir / SSH_KEY_NAME
-        if not self.ssh_key_path.exists():
-            shell(
-                f"ssh-keygen -t ed25519 -f {self.ssh_key_path} -N '' -C localchat",
-                check=True,
-            )
-        self.ssh_config_path = self.lxconfigs_dir / "ssh-config"
-
-    def write_ssh_config(self):
-        """Write ``lxconfigs/ssh-config`` mapping all containers to their IPs.
-
-        Each Host block maps the container name, the domain name, and the
-        short relay name (e.g. ``_test0``) to the container's IP, using the
-        shared localchat SSH key.  Returns the path to the file.
-        """
-        containers = self.list_managed()
-        key_path = self.ssh_key_path
-        lines = ["# Auto-generated by cmdeploy lxc-start — do not edit\n"]
-        for c in containers:
-            hosts = [c["name"]]
-            domain = c.get("domain", "")
-            if domain and domain != c["name"]:
-                hosts.append(domain)
-                short = domain.split(".")[0]
-                if short and short not in hosts:
-                    hosts.append(short)
-            lines.append(f"\nHost {' '.join(hosts)}\n")
-            lines.append(f"    Hostname {c['ip']}\n")
-            lines.append("    User root\n")
-            lines.append(f"    IdentityFile {key_path}\n")
-            lines.append("    IdentitiesOnly yes\n")
-            lines.append("    StrictHostKeyChecking accept-new\n")
-            lines.append("    UserKnownHostsFile /dev/null\n")
-            lines.append("    LogLevel ERROR\n")
-        path = self.ssh_config_path
-        path.write_text("".join(lines))
-        return path
-
-    def check_ssh_include(self):
-        """Check if the user's ~/.ssh/config already includes our ssh-config."""
-        user_ssh_config = Path.home() / ".ssh" / "config"
-        if not user_ssh_config.exists():
-            return False
-        lines = user_ssh_config.read_text().splitlines()
-        target = f"include {self.ssh_config_path}".lower()
-        return any(line.strip().lower() == target for line in lines)
-
-    def run(self, args, check=True, capture=True, input=None):
-        """Run an incus command.
-
-        When *capture* is True and *verbosity* >= 1, output is streamed
-        to the terminal line-by-line while also being captured for
-        later return via result.stdout.
-        """
-        cmd = ["incus"] + list(args)
-        sub = self.out.new_prefixed_out("  ")
-
-        if not capture:
-            # Simple case: let subprocess handle streams (no capture)
-            if self.out.verbosity >= 1:
-                sub.print(f"$ {' '.join(cmd)}")
-            return subprocess.run(
-                cmd, text=True, input=input, check=check, stdout=None, stderr=None
-            )
-
-        # Capture case: we may need to stream while capturing
-        if sub.verbosity >= 1:
-            cmd_lines = " ".join(cmd).splitlines()
-            sub.print(f"$ {cmd_lines.pop(0)}")
-            if sub.verbosity >= 2:
-                for line in cmd_lines:
-                    sub.print(f"  {line}")
-
-        proc = subprocess.Popen(
-            cmd,
-            text=True,
-            stdin=subprocess.PIPE if input else None,
-            stdout=subprocess.PIPE,
-            stderr=subprocess.STDOUT,
-        )
-
-        stdout_lines = []
-        if input:
-            proc.stdin.write(input)
-            proc.stdin.close()
-
-        for line in proc.stdout:
-            stdout_lines.append(line)
-            if sub.verbosity >= 2:
-                sub.print(f"  > {line.rstrip()}")
-
-        ret = proc.wait()
-        stdout = "".join(stdout_lines)
-        if check and ret != 0:
-            for line in stdout.splitlines():
-                if sub.verbosity < 1:  # and we haven't printed it yet
-                    sub.red(line)
-            raise subprocess.CalledProcessError(ret, cmd, output=stdout)
-
-        return subprocess.CompletedProcess(cmd, ret, stdout=stdout)
-
-    def run_json(self, args, check=True):
-        """Run an incus command with ``--format=json``.
-
-        Returns the parsed JSON on success.
-        When *check* is True raises ``subprocess.CalledProcessError``
-        on non-zero exit; when False returns *None* instead.
-        """
-        result = self.run(
-            list(args) + ["--format=json"],
-            check=check,
-        )
-        if result.returncode != 0:
-            return None
-        return json.loads(result.stdout)
-
-    def run_output(self, args, check=True):
-        """Run an incus command and return its stripped stdout.
-
-        When *check* is False, returns *None* on non-zero exit
-        instead of raising.
-        """
-        result = self.run(args, check=check)
-        if result.returncode != 0:
-            return None
-        return result.stdout.strip()
-
-    def find_image(self, aliases):
-        """Return the first alias from *aliases* that exists, else None."""
-        images = self.run_json(["image", "list"], check=False) or []
-        existing = {a.get("name") for img in images for a in img.get("aliases", [])}
-        for alias in aliases:
-            if alias in existing:
-                return alias
-        return None
-
-    def delete_images(self):
-        """Delete the cached base and relay images."""
-        for alias in (RELAY_IMAGE_ALIAS, BASE_IMAGE_ALIAS):
-            self.run(["image", "delete", alias], check=False)  # ok if absent
-
-    def list_managed(self):
-        """Return list of dicts with name, ip, ipv6, domain, status, memory_usage."""
-        containers = []
-        for ct in self.run_json(["list"]):
-            config = ct.get("config", {})
-            if config.get(LABEL_KEY) != "true":
-                continue
-            name = ct["name"]
-            state = ct.get("state", {})
-            net = state.get("network") or {}
-            containers.append(
-                {
-                    "name": name,
-                    "ip": _extract_ip(net, "inet"),
-                    "ipv6": _extract_ip(net, "inet6"),
-                    "domain": config.get(
-                        "user.localchat-domain", f"{name}{DOMAIN_SUFFIX}"
-                    ),
-                    "status": ct.get("status", "Unknown"),
-                    "memory_usage": state.get("memory", {}).get("usage", 0),
-                }
-            )
-        return containers
-
-    def ensure_base_image(self):
-        """Build and cache a base image with openssh and the SSH key.
-
-        The image is published as a local incus image with alias
-        'localchat-base'.  Subsequent container launches use this
-        image instead of the upstream Debian 12, skipping the
-        slow apt-get install step.
-        Returns the image alias.
-        """
-        if self.find_image([BASE_IMAGE_ALIAS]):
-            self.out.print(f"  Base image '{BASE_IMAGE_ALIAS}' already cached.")
-            return BASE_IMAGE_ALIAS
-
-        self.out.print("  Building base image (one-time setup) ...")
-
-        self.run(["delete", BASE_SETUP_NAME, "--force"], check=False)
-        self.run(["image", "delete", BASE_IMAGE_ALIAS], check=False)
-        self.run(["launch", UPSTREAM_IMAGE, BASE_SETUP_NAME])
-
-        ct = Container(self, BASE_SETUP_NAME)
-        ct.wait_ready()
-
-        key_path = self.ssh_key_path
-        pub_key = key_path.with_suffix(".pub").read_text().strip()
-        ct.bash(f"""
-            echo 'nameserver 9.9.9.9' > /etc/resolv.conf
-            apt-get -o DPkg::Lock::Timeout=60 update
-            DEBIAN_FRONTEND=noninteractive apt-get install -y openssh-server python3
-            systemctl enable ssh
-            apt-get clean
-            mkdir -p /root/.ssh
-            chmod 700 /root/.ssh
-            echo '{pub_key}' > /root/.ssh/authorized_keys
-            chmod 600 /root/.ssh/authorized_keys
-        """)
-
-        self.run(["stop", BASE_SETUP_NAME])
-        self.run(["publish", BASE_SETUP_NAME, f"--alias={BASE_IMAGE_ALIAS}"])
-        self.run(["delete", BASE_SETUP_NAME, "--force"])
-        self.out.print(f"  Base image '{BASE_IMAGE_ALIAS}' ready.")
-        return BASE_IMAGE_ALIAS
-
-    def get_container(self, name):
-        """Return a container handle for the given name.
-
-        Accepts both short relay names (``test0``) and full Incus
-        container names (``test0-localchat``).  Returns
-        ``DNSContainer`` for the DNS container and
-        ``RelayContainer`` for everything else.
-        """
-        if name == DNS_CONTAINER_NAME:
-            return DNSContainer(self)
-        return RelayContainer(self, name.removesuffix("-localchat"))
-
-    def get_dns_container(self):
-        """Return a DNSContainer handle."""
-        return DNSContainer(self)
-
-
-class Container:
-    """The base container handle wraps all interactions with incus."""
-
-    def __init__(self, incus, name, domain=None, memory="200MiB"):
-        self.incus = incus
-        self.out = incus.out
-        self.name = name
-        self.domain = domain or f"{name}{DOMAIN_SUFFIX}"
-        self.memory = memory
-        self.ipv4 = None
-        self.ipv6 = None
-
-    def bash(self, script, check=True):
-        """Returns stdout from executing ``bash -ec <script>`` inside this container.
-
-        *script* is dedented and stripped so callers can use triple-quoted strings.
-        When *check* is False, returns *None* on non-zero exit instead of raising.
-        """
-        script = textwrap.dedent(script).strip()
-        cmd = ["exec", self.name, "--", "bash", "-ec", script]
-        return self.incus.run_output(cmd, check=check)
-
-    def run_cmd(self, *args, check=True):
-        """Return stdout from running a command directly in the container (no shell).
-
-        When *check* is False, returns *None* on non-zero exit instead of raising.
-        """
-        return self.incus.run_output(
-            ["exec", self.name, "--", *args],
-            check=check,
-        )
-
-    def start(self):
-        self.incus.run(["start", self.name])
-
-    def stop(self, force=False):
-        cmd = ["stop", self.name]
-        if force:
-            cmd.append("--force")
-        self.incus.run(cmd, check=False)
-
-    def launch(self):
-        """Launch from the best available image, return the alias used."""
-        image = self.incus.find_image([RELAY_IMAGE_ALIAS, BASE_IMAGE_ALIAS])
-        if not image:
-            raise RuntimeError(
-                f"No base image '{BASE_IMAGE_ALIAS}' found. "
-                "Call ensure_base_image() before launching containers."
-            )
-        self.out.print(f"  Launching from '{image}' image ...")
-        cfg = []
-        cfg += ("-c", f"{LABEL_KEY}=true")
-        cfg += ("-c", f"user.localchat-domain={self.domain}")
-        cfg += ("-c", f"limits.memory={self.memory}")
-        self.incus.run(["launch", image, self.name, *cfg])
-        return image
-
-    def ensure(self):
-        """Create/start this container from the cached base image.
-
-        On first call, builds the base image (~30s).
-        Subsequent containers launch in ~2s from the cached image.
-        Returns ``self`` for chaining.
-        """
-        data = self.incus.run_json(["list", self.name], check=False) or []
-
-        existing = [c for c in data if c["name"] == self.name]
-        if existing:
-            if existing[0]["status"] != "Running":
-                self.start()
-        else:
-            self.launch()
-        self.wait_ready()
-        return self
-
-    def destroy(self):
-        """Stop, delete, and clean up config files."""
-        self.stop(force=True)
-        self.incus.run(["delete", self.name, "--force"], check=False)
-
-    def push_file_content(self, dest_path, content):
-        """Write *content* to *dest_path* inside the container.
-
-        *content* is dedented and stripped so callers can use
-        indented triple-quoted strings.
-        """
-        content = textwrap.dedent(content).strip() + "\n"
-        self.incus.run(
-            ["file", "push", "-", f"{self.name}{dest_path}"],
-            input=content,
-        )
-        self.bash(f"chmod 644 {dest_path}")
-
-    def wait_ready(self, timeout=60):
-        """Wait until the container is running with an IPv4 address.
-
-        Sets ``self.ipv4`` and ``self.ipv6`` (may be *None*),
-        or raises ``TimeoutError``.
-        """
-        deadline = time.time() + timeout
-        while time.time() < deadline:
-            data = self.incus.run_json(
-                ["list", self.name],
-                check=False,
-            )
-            if data and data[0].get("status") == "Running":
-                net = data[0].get("state", {}).get("network", {})
-                self.ipv4 = _extract_ip(net, "inet")
-                self.ipv6 = _extract_ip(net, "inet6")
-                if self.ipv4:
-                    return
-            time.sleep(1)
-        raise TimeoutError(
-            f"Container {self.name!r} did not become ready within {timeout}s"
-        )
-
-    def rss_mib(self):
-        """Return ``(used, total)`` memory from container (or None if unobtainable)."""
-        output = self.bash("free -m", check=False)
-        if output:
-            for line in output.splitlines():
-                if line.startswith("Mem:"):
-                    parts = line.split()
-                    return int(parts[2]), int(parts[1])
-
-
-class RelayContainer(Container):
-    """Container handle for a chatmail relay.
-
-    Accepts the short relay name (e.g. ``test0``) and derives
-    the Incus container name and mail domain automatically.
-    """
-
-    def __init__(self, incus, name):
-        super().__init__(
-            incus,
-            f"{name}-localchat",
-            domain=f"_{name}{DOMAIN_SUFFIX}",
-            memory="600MiB",
-        )
-        self.sname = name
-        self.ini = incus.lxconfigs_dir / f"chatmail-{name}.ini"
-        self.zone = incus.lxconfigs_dir / f"{name}.zone"
-
-    def launch(self):
-        """Launch (from a potentially cached image) and clear inherited chatmail-version."""
-        image = super().launch()
-        self.bash("rm -f /etc/chatmail-version")
-        return image
-
-    def destroy(self):
-        """Stop, delete, and clean up config files."""
-        super().destroy()
-        if self.ini.exists():
-            self.ini.unlink()
-
-    def disable_ipv6(self):
-        """Disable IPv6 inside the container via sysctl."""
-        # incus provides net.* virtualization for LXC containers so that
-        # these sysctls only affect the container's network namespace.
-        self.bash("""
-            sysctl -w net.ipv6.conf.all.disable_ipv6=1
-            sysctl -w net.ipv6.conf.default.disable_ipv6=1
-            mkdir -p /etc/sysctl.d
-            printf 'net.ipv6.conf.all.disable_ipv6=1\\n
-            net.ipv6.conf.default.disable_ipv6=1\\n'
-            > /etc/sysctl.d/99-disable-ipv6.conf
-        """)
-
-    def configure_hosts(self, ip):
-        """Set hostname and /etc/hosts inside the container."""
-        self.bash(f"""
-            echo '{self.name}' > /etc/hostname
-            hostname {self.name}
-            sed -i '/ {self.domain}$/d' /etc/hosts
-            echo '{ip} {self.name} {self.domain}' >> /etc/hosts
-        """)
-
-    def publish_as_relay_image(self):
-        """Publish this container as a reusable relay image.
-
-        Stops the container, 'publishes' it as 'localchat-relay', then restarts it.
-        """
-        if self.incus.find_image([RELAY_IMAGE_ALIAS]):
-            return
-        self.out.print(
-            f"  Locally caching {self.name!r} as '{RELAY_IMAGE_ALIAS}' image ..."
-        )
-        self.incus.run(
-            ["publish", self.name, f"--alias={RELAY_IMAGE_ALIAS}", "--force"]
-        )
-        self.wait_ready()
-        self.out.print(f"  Relay image '{RELAY_IMAGE_ALIAS}' ready.")
-
-    def deployed_version(self):
-        """Read /etc/chatmail-version, or None if absent."""
-        return self.bash("cat /etc/chatmail-version", check=False)
-
-    def deployed_domain(self):
-        """Read the domain deployed on the container (postfix myhostname)."""
-        return self.bash(
-            "postconf -h myhostname 2>/dev/null",
-            check=False,
-        )
-
-    def verify_ssh(self, ssh_config):
-        """Verify SSH connectivity to this container."""
-        cmd = f"ssh -F {ssh_config} -o ConnectTimeout=10 root@{self.domain} hostname"
-        return shell(cmd, timeout=15).returncode == 0
-
-    def configure_dns(self, dns_ip):
-        """Point this container's resolver at *dns_ip* and verify DNS is reachable."""
-        self.bash(f"""
-            systemctl disable --now systemd-resolved 2>/dev/null || true
-            rm -f /etc/resolv.conf
-            echo 'nameserver {dns_ip}' > /etc/resolv.conf
-            mkdir -p /etc/unbound/unbound.conf.d
-            printf 'server:\\n  domain-insecure: "localchat"\\n\\n
-            forward-zone:\\n  name: "localchat"\\n
-              forward-addr: {dns_ip}\\n'
-            > /etc/unbound/unbound.conf.d/localchat-forward.conf
-            systemctl restart unbound 2>/dev/null || true
-        """)
-        self._wait_dns_reachable(dns_ip)
-
-    def _wait_dns_reachable(self, dns_ip, timeout=10):
-        """Poll until *dns_ip* answers a DNS query from this container."""
-        if self.bash("which dig", check=False) is None:
-            self.bash(
-                "DEBIAN_FRONTEND=noninteractive "
-                "apt-get install -y dnsutils 2>/dev/null || true"
-            )
-        deadline = time.time() + timeout
-        while time.time() < deadline:
-            result = self.bash(
-                f"dig @{dns_ip} . SOA +short +time=1 +tries=1",
-                check=False,
-            )
-            if result and result.strip():
-                return
-            time.sleep(0.5)
-        raise DNSConfigurationError(
-            f"DNS at {dns_ip} not reachable from {self.name} after {timeout}s"
-        )
-
-    def write_ini(self, disable_ipv6=False):
-        """Generate a chatmail.ini config file in lxconfigs/."""
-        from chatmaild.config import write_initial_config
-
-        overrides = {
-            "max_user_send_per_minute": 600,
-            "max_user_send_burst_size": 100,
-            "mtail_address": "127.0.0.1",
-        }
-        if disable_ipv6:
-            overrides["disable_ipv6"] = "True"
-        write_initial_config(self.ini, self.domain, overrides)
-        return self.ini
-
-
-class DNSContainer(Container):
-    """Container handle for the PowerDNS name server.
-
-    Manages the authoritative and recursive DNS services required for
-    name resolution in the local testing environment.
-    """
-
-    def __init__(self, incus):
-        super().__init__(incus, DNS_CONTAINER_NAME, domain=DNS_DOMAIN)
-
-    def pdnsutil(self, *args, check=True):
-        """Run ``pdnsutil <args>`` inside the DNS container."""
-        return self.run_cmd("pdnsutil", *args, check=check)
-
-    def replace_rrset(self, zone, name, rtype, ttl, rdata):
-        """Shortcut for ``pdnsutil replace-rrset``."""
-        self.pdnsutil("replace-rrset", zone, name, rtype, ttl, rdata)
-
-    def restart_services(self):
-        """Restart pdns and pdns-recursor, then wait until DNS is answering."""
-        self.bash("""
-            systemctl restart pdns
-            systemctl restart pdns-recursor || true
-        """)
-        self._wait_dns_ready()
-
-    def _wait_dns_ready(self, timeout=10):
-        """Poll until the recursor answers a query on port 53."""
-        deadline = time.time() + timeout
-        while time.time() < deadline:
-            result = self.bash(
-                "dig @127.0.0.1 . SOA +short +time=1 +tries=1",
-                check=False,
-            )
-            if result and result.strip():
-                return
-            time.sleep(0.5)
-        raise DNSConfigurationError(f"DNS recursor not answering after {timeout}s")
-
-    def ensure(self):
-        """Create the DNS container with PowerDNS if needed.
-
-        Calls ``super().ensure()`` to create/start the container
-        and set up SSH, then installs PowerDNS and configures
-        the Incus bridge to use this container as DNS.
-        """
-        super().ensure()
-        self._install_powerdns()
-        self.incus.run(
-            ["network", "set", "incusbr0", "dns.mode=none"],
-            check=False,
-        )
-        self.incus.run(
-            ["network", "set", "incusbr0", f"raw.dnsmasq=dhcp-option=6,{self.ipv4}"],
-            check=False,
-        )
-
-    def destroy(self):
-        """Stop, delete, and reset bridge DNS config."""
-        super().destroy()
-        self.incus.run(["network", "unset", "incusbr0", "dns.mode"], check=False)
-        self.incus.run(["network", "unset", "incusbr0", "raw.dnsmasq"], check=False)
-
-    def _install_powerdns(self):
-        """Install and configure PowerDNS if not already present."""
-        if self.run_cmd("which", "pdns_server", check=False) is not None:
-            return
-
-        self.bash("""
-            systemctl disable --now systemd-resolved 2>/dev/null || true
-            rm -f /etc/resolv.conf
-            echo 'nameserver 9.9.9.9' > /etc/resolv.conf
-
-            # Block automatic service startup during package installation
-            printf '#!/bin/sh\\nexit 101\\n' > /usr/sbin/policy-rc.d
-            chmod +x /usr/sbin/policy-rc.d
-
-            apt-get -o DPkg::Lock::Timeout=60 update
-            DEBIAN_FRONTEND=noninteractive apt-get install -y \
-                pdns-server pdns-backend-sqlite3 sqlite3 pdns-recursor dnsutils
-
-            # Remove the startup block
-            rm /usr/sbin/policy-rc.d
-
-            systemctl stop pdns pdns-recursor || true
-            mkdir -p /var/lib/powerdns
-            sqlite3 /var/lib/powerdns/pdns.sqlite3 \
-                </usr/share/doc/pdns-backend-sqlite3/schema.sqlite3.sql
-            chown -R pdns:pdns /var/lib/powerdns
-        """)
-
-        self.push_file_content(
-            "/etc/powerdns/pdns.conf",
-            """
-            launch=gsqlite3
-            gsqlite3-database=/var/lib/powerdns/pdns.sqlite3
-            local-address=127.0.0.1
-            local-port=5353
-            """,
-        )
-
-        self.push_file_content(
-            "/etc/powerdns/recursor.conf",
-            """
-            local-address=0.0.0.0
-            local-port=53
-            forward-zones=localchat=127.0.0.1:5353
-            forward-zones-recurse=.=9.9.9.9;149.112.112.112
-            allow-from=0.0.0.0/0
-            dont-query=
-            dnssec=off
-            """,
-        )
-
-        self.bash("""
-            systemctl start pdns
-            systemctl start pdns-recursor
-            echo 'nameserver 127.0.0.1' > /etc/resolv.conf
-        """)
-        self._wait_dns_ready()
-
-    def reset_dns_records(self, dns_ip, domains):
-        """Create DNS zones with initial A records via pdnsutil.
-
-        Only sets SOA, NS, and A records as the minimal set
-        needed for SSH connectivity.  Full records (MX, TXT, SRV,
-        CNAME, DKIM) are added later by ``cmdeploy dns``.
-
-        Args:
-            dns_ip: IP of the DNS container
-            domains: list of dicts with 'name', 'domain', 'ip'
-        """
-        for d in domains:
-            domain = d["domain"]
-            ip = d["ip"]
-            self.out.print(f"  {domain} -> {ip}")
-
-            # Delete and recreate zone fresh (removes stale records)
-            self.pdnsutil("delete-zone", domain, check=False)
-            self.pdnsutil("create-zone", domain, f"ns.{domain}")
-
-            serial = str(int(time.time()))
-            soa = f"ns.{domain} hostmaster.{domain} {serial} 3600 900 604800 300"
-            self.replace_rrset(domain, ".", "SOA", "3600", soa)
-            self.replace_rrset(domain, ".", "NS", "3600", f"ns.{domain}.")
-            self.replace_rrset(domain, ".", "A", "3600", ip)
-            self.replace_rrset(domain, "ns", "A", "3600", dns_ip)
-
-            # AAAA (domain -> container IPv6, if available)
-            ipv6 = d.get("ipv6")
-            if ipv6:
-                self.replace_rrset(domain, ".", "AAAA", "3600", ipv6)
-                self.out.print(f"    zone reset: SOA, NS, A, AAAA  ({ip}, {ipv6})")
-            else:
-                # Remove any stale AAAA record
-                self.pdnsutil("delete-rrset", domain, ".", "AAAA", check=False)
-                self.out.print(f"    zone reset: SOA, NS, A  ({ip}, IPv4-only)")
-
-        self.restart_services()
-
-    def set_dns_records(self, text):
-        """Add or overwrite DNS records from standard BIND format.
-
-        Uses ``cmdeploy.dns.parse_zone_records`` to parse.
-        Zones are created automatically from the record names.
-        """
-        from ..dns import parse_zone_records
-
-        zones_seen = set()
-
-        for name, ttl, rtype, rdata in parse_zone_records(text):
-            # Derive zone from name: find top-level .localchat domain
-            name_parts = name.split(".")
-            zone = name  # fallback
-            for i in range(len(name_parts) - 1):
-                if name_parts[i + 1 :] == ["localchat"]:
-                    zone = ".".join(name_parts[i:])
-                    break
-
-            # Create zone if first time seeing it
-            if zone not in zones_seen:
-                self.pdnsutil(
-                    "create-zone",
-                    zone,
-                    f"ns.{zone}",
-                    check=False,
-                )
-                zones_seen.add(zone)
-
-            # Figure out the record name relative to zone
-            if name == zone:
-                relative = "."
-            elif name.endswith(f".{zone}"):
-                relative = name[: -(len(zone) + 1)]
-            else:
-                relative = name
-
-            self.replace_rrset(zone, relative, rtype, ttl, rdata)
-
-        if zones_seen:
-            self.restart_services()

cmdeploy/src/cmdeploy/nginx/nginx.conf.j2

@@ -73,6 +73,10 @@ http {
 
 		access_log syslog:server=unix:/dev/log,facility=local7;
 
+		location /mxdeliv/ {
+		    proxy_pass http://127.0.0.1:{{ config.filtermail_http_port }};
+		}
+
 		location / {
 			# First attempt to serve request as file, then
 			# as directory, then fall back to displaying a 404.

cmdeploy/src/cmdeploy/postfix/main.cf.j2

@@ -20,7 +20,7 @@ smtpd_tls_key_file={{ config.tls_key_path }}
 smtpd_tls_security_level=may
 
 smtp_tls_CApath=/etc/ssl/certs
-smtp_tls_security_level={{ "verify" if config.tls_cert_mode == "acme" else "encrypt" }}
+smtp_tls_security_level=verify
 # Send SNI extension when connecting to other servers.
 # <https://www.postfix.org/postconf.5.html#smtp_tls_servername>
 smtp_tls_servername = hostname
@@ -88,6 +88,22 @@ inet_protocols = ipv4
 inet_protocols = all
 {% endif %}
 
+# Postfix does not try IPv4 and IPv6 connections
+# concurrently as of version 3.7.11.
+#
+# When relay has both A (IPv4) and AAAA (IPv6) records,
+# but broken IPv6 connectivity,
+# every second message is delayed by the connection timeout
+# <https://www.postfix.org/postconf.5.html#smtp_connect_timeout>
+# which defaults to 30 seconds. Reducing timeouts is not a solution
+# as this will result in a failure to connect to slow servers.
+#
+# As a workaround we always prefer IPv4 when it is available.
+# 
+# The setting is documented at
+# <https://www.postfix.org/postconf.5.html#smtp_address_preference>
+smtp_address_preference=ipv4
+
 virtual_transport = lmtp:unix:private/dovecot-lmtp
 virtual_mailbox_domains = {{ config.mail_domain }}
 lmtp_header_checks = regexp:/etc/postfix/lmtp_header_cleanup

cmdeploy/src/cmdeploy/remote/rdns.py

@@ -57,9 +57,10 @@ def get_dkim_entry(mail_domain, pre_command, dkim_selector):
     dkim_value_raw = f"v=DKIM1;k=rsa;p={dkim_pubkey};s=email;t=s"
     dkim_value = '" "'.join(re.findall(".{1,255}", dkim_value_raw))
     web_dkim_value = "".join(re.findall(".{1,255}", dkim_value_raw))
+    name = f"{dkim_selector}._domainkey.{mail_domain}."
     return (
-        f'{dkim_selector}._domainkey.{mail_domain}.  3600  IN  TXT  "{dkim_value}"',
-        f'{dkim_selector}._domainkey.{mail_domain}.  3600  IN  TXT  "{web_dkim_value}"',
+        f'{name:<40} 3600   IN  TXT    "{dkim_value}"',
+        f'{name:<40} 3600   IN  TXT    "{web_dkim_value}"',
     )
 
 
@@ -94,11 +95,9 @@ def check_zonefile(zonefile, verbose=True):
         if not zf_line.strip() or zf_line.startswith(";"):
             continue
         print(f"dns-checking {zf_line!r}") if verbose else log_progress("")
-        parts = zf_line.split(None, 4)
-        zf_domain = parts[0].rstrip(".")
-        # parts[1]=TTL, parts[2]=IN, parts[3]=type, parts[4]=rdata
-        zf_typ = parts[3]
-        zf_value = parts[4].strip()
+        zf_domain, _ttl, _in, zf_typ, zf_value = zf_line.split(None, 4)
+        zf_domain = zf_domain.rstrip(".")
+        zf_value = zf_value.strip()
         query_value = query_dns(zf_typ, zf_domain)
         if zf_value != query_value:
             assert zf_typ in ("A", "AAAA", "CNAME", "CAA", "SRV", "MX", "TXT"), zf_line

cmdeploy/src/cmdeploy/selfsigned/deployer.py

@@ -12,27 +12,15 @@ def openssl_selfsigned_args(domain, cert_path, key_path, days=36500):
     ``www.<domain>`` and ``mta-sts.<domain>``.
     """
     return [
-        "openssl",
-        "req",
-        "-x509",
-        "-newkey",
-        "ec",
-        "-pkeyopt",
-        "ec_paramgen_curve:P-256",
-        "-noenc",
-        "-days",
-        str(days),
-        "-keyout",
-        str(key_path),
-        "-out",
-        str(cert_path),
-        "-subj",
-        f"/CN={domain}",
+        "openssl", "req", "-x509",
+        "-newkey", "ec", "-pkeyopt", "ec_paramgen_curve:P-256",
+        "-noenc", "-days", str(days),
+        "-keyout", str(key_path),
+        "-out", str(cert_path),
+        "-subj", f"/CN={domain}",
         # Mark as end-entity cert so it cannot be used as a CA to sign others.
-        "-addext",
-        "basicConstraints=critical,CA:FALSE",
-        "-addext",
-        "extendedKeyUsage=serverAuth,clientAuth",
+        "-addext", "basicConstraints=critical,CA:FALSE",
+        "-addext", "extendedKeyUsage=serverAuth,clientAuth",
         "-addext",
         f"subjectAltName=DNS:{domain},DNS:www.{domain},DNS:mta-sts.{domain}",
     ]
@@ -54,9 +42,7 @@ class SelfSignedTlsDeployer(Deployer):
 
     def configure(self):
         args = openssl_selfsigned_args(
-            self.mail_domain,
-            self.cert_path,
-            self.key_path,
+            self.mail_domain, self.cert_path, self.key_path,
         )
         cmd = shlex.join(args)
         server.shell(

cmdeploy/src/cmdeploy/sshexec.py

@@ -49,13 +49,8 @@ class SSHExec:
     RemoteError = execnet.RemoteError
     FuncError = FuncError
 
-    def __init__(
-        self, host, verbose=False, python="python3", timeout=60, ssh_config=None
-    ):
-        spec = f"ssh=root@{host}//python={python}"
-        if ssh_config:
-            spec += f"//ssh_config={ssh_config}"
-        self.gateway = execnet.makegateway(spec)
+    def __init__(self, host, verbose=False, python="python3", timeout=60):
+        self.gateway = execnet.makegateway(f"ssh=root@{host}//python={python}")
         self._remote_cmdloop_channel = bootstrap_remote(self.gateway, remote)
         self.timeout = timeout
         self.verbose = verbose
@@ -92,9 +87,8 @@ class SSHExec:
 class LocalExec:
     FuncError = FuncError
 
-    def __init__(self, verbose=False, docker=False):
+    def __init__(self, verbose=False):
         self.verbose = verbose
-        self.docker = docker
 
     def __call__(self, call, kwargs=None, log_callback=None):
         if kwargs is None:
@@ -106,10 +100,6 @@ class LocalExec:
         if not title:
             title = call.__name__
         where = "locally"
-        if self.docker:
-            if call == remote.rdns.perform_initial_checks:
-                kwargs["pre_command"] = "docker exec chatmail "
-                where = "in docker"
         if self.verbose:
             print_stderr(f"Running {where}: {title}(**{kwargs})")
             return self(call, kwargs, log_callback=print_stderr)
@@ -118,46 +108,3 @@ class LocalExec:
             res = self(call, kwargs, log_callback=remote.rshell.log_progress)
             print_stderr()
             return res
-
-
-# pyinfra exposes a ``ssh_config_file`` data key that *should* let
-# paramiko parse an SSH config file directly.  In practice it silently
-# fails to connect (zero hosts / zero operations), so we resolve the
-# hostname and identity-file ourselves and pass them via
-# ``--data ssh_hostname`` / ``--data ssh_key`` instead.
-# Execnet uses ssh natively (and not paramiko) and doesn't have this problem.
-
-
-def _get_from_ssh_config(host, ssh_config_path, key):
-    """Internal helper to parse a value for a specific key from ssh-config."""
-    current_hosts = []
-    found_value = None
-    with open(ssh_config_path) as f:
-        for raw_line in f:
-            line = raw_line.strip()
-            if not line or line.startswith("#"):
-                continue
-            parts = line.split(None, 1)
-            if not parts:
-                continue
-            directive = parts[0].lower()
-            if directive == "host":
-                if host in current_hosts and found_value:
-                    return found_value
-                current_hosts = parts[1].split()
-                found_value = None
-            elif directive == key.lower():
-                found_value = parts[1]
-    if host in current_hosts and found_value:
-        return found_value
-    return None
-
-
-def resolve_host_from_ssh_config(host, ssh_config_path):
-    """Resolve a host alias to its IP from an ssh-config file."""
-    return _get_from_ssh_config(host, ssh_config_path, "Hostname") or host
-
-
-def resolve_key_from_ssh_config(host, ssh_config_path):
-    """Resolve a host alias to its IdentityFile from an ssh-config file."""
-    return _get_from_ssh_config(host, ssh_config_path, "IdentityFile")

cmdeploy/src/cmdeploy/tests/data/zftest.zone

@@ -1,18 +1,18 @@
 ; Required DNS entries
-zftest.testrun.org.  3600  IN  A  135.181.204.127
-zftest.testrun.org.  3600  IN  AAAA  2a01:4f9:c012:52f4::1
-zftest.testrun.org.  3600  IN  MX  10 zftest.testrun.org.
-_mta-sts.zftest.testrun.org.  3600  IN  TXT  "v=STSv1; id=202403211706"
-mta-sts.zftest.testrun.org.  3600  IN  CNAME  zftest.testrun.org.
-www.zftest.testrun.org.  3600  IN  CNAME  zftest.testrun.org.
-opendkim._domainkey.zftest.testrun.org.  3600  IN  TXT  "v=DKIM1;k=rsa;p=MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEAoYt82CVUyz2ouaqjX2kB+5J80knAyoOU3MGU5aWppmwUwwTvj/oSTSpkc5JMtVTRmKKr8NUDWAL1Yw7dfGqqPHdHfwwjS3BIvDzYx+hzgtz62RnfNgV+/2MAoNpfX7cAFIHdRzEHNtwugc3RDLquqPoupAE3Y2YRw2T5zG5fILh4vwIcJZL5Uq6B92j8wwJqOex" "33n+vm1NKQ9rxo/UsHAmZlJzpooXcG/4igTBxJyJlamVSRR6N7Nul1v//YJb7J6v2o0iPHW6uE0StzKaPPNC2IVosSRFbD9H2oqppltptFSNPlI0E+t0JBWHem6YK7xcugiO3ImMCaaU8g6Jt/wIDAQAB;s=email;t=s"
+zftest.testrun.org.                      3600   IN  A      135.181.204.127
+zftest.testrun.org.                      3600   IN  AAAA   2a01:4f9:c012:52f4::1
+zftest.testrun.org.                      3600   IN  MX     10 zftest.testrun.org.
+_mta-sts.zftest.testrun.org.             3600   IN  TXT    "v=STSv1; id=202403211706"
+mta-sts.zftest.testrun.org.              3600   IN  CNAME  zftest.testrun.org.
+www.zftest.testrun.org.                  3600   IN  CNAME  zftest.testrun.org.
+opendkim._domainkey.zftest.testrun.org.  3600   IN  TXT    "v=DKIM1;k=rsa;p=MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEAoYt82CVUyz2ouaqjX2kB+5J80knAyoOU3MGU5aWppmwUwwTvj/oSTSpkc5JMtVTRmKKr8NUDWAL1Yw7dfGqqPHdHfwwjS3BIvDzYx+hzgtz62RnfNgV+/2MAoNpfX7cAFIHdRzEHNtwugc3RDLquqPoupAE3Y2YRw2T5zG5fILh4vwIcJZL5Uq6B92j8wwJqOex" "33n+vm1NKQ9rxo/UsHAmZlJzpooXcG/4igTBxJyJlamVSRR6N7Nul1v//YJb7J6v2o0iPHW6uE0StzKaPPNC2IVosSRFbD9H2oqppltptFSNPlI0E+t0JBWHem6YK7xcugiO3ImMCaaU8g6Jt/wIDAQAB;s=email;t=s"
 
 ; Recommended DNS entries
-zftest.testrun.org.  3600  IN  TXT  "v=spf1 a ~all"
-_dmarc.zftest.testrun.org.  3600  IN  TXT  "v=DMARC1;p=reject;adkim=s;aspf=s"
-zftest.testrun.org.  3600  IN  CAA  0 issue "letsencrypt.org;accounturi=https://acme-v02.api.letsencrypt.org/acme/acct/1371472956"
-_adsp._domainkey.zftest.testrun.org.  3600  IN  TXT  "dkim=discardable"
-_submission._tcp.zftest.testrun.org.  3600  IN  SRV  0 1 587 zftest.testrun.org.
-_submissions._tcp.zftest.testrun.org.  3600  IN  SRV  0 1 465 zftest.testrun.org.
-_imap._tcp.zftest.testrun.org.  3600  IN  SRV  0 1 143 zftest.testrun.org.
-_imaps._tcp.zftest.testrun.org.  3600  IN  SRV  0 1 993 zftest.testrun.org.
+zftest.testrun.org.                      3600   IN  TXT    "v=spf1 a ~all"
+_dmarc.zftest.testrun.org.               3600   IN  TXT    "v=DMARC1;p=reject;adkim=s;aspf=s"
+zftest.testrun.org.                      3600   IN  CAA    0 issue "letsencrypt.org;accounturi=https://acme-v02.api.letsencrypt.org/acme/acct/1371472956"
+_adsp._domainkey.zftest.testrun.org.     3600   IN  TXT    "dkim=discardable"
+_submission._tcp.zftest.testrun.org.     3600   IN  SRV    0 1 587 zftest.testrun.org.
+_submissions._tcp.zftest.testrun.org.    3600   IN  SRV    0 1 465 zftest.testrun.org.
+_imap._tcp.zftest.testrun.org.           3600   IN  SRV    0 1 143 zftest.testrun.org.
+_imaps._tcp.zftest.testrun.org.          3600   IN  SRV    0 1 993 zftest.testrun.org.

cmdeploy/src/cmdeploy/tests/online/test_0_login.py

@@ -89,7 +89,9 @@ def test_concurrent_logins_same_account(
         assert login_results.get()
 
 
-def test_no_vrfy(chatmail_config):
+def test_no_vrfy(cmfactory, chatmail_config):
+    ac = cmfactory.get_online_account()
+    addr = ac.get_config("addr")
     domain = chatmail_config.mail_domain
 
     s = smtplib.SMTP(domain)
@@ -98,7 +100,7 @@ def test_no_vrfy(chatmail_config):
     s.putcmd("vrfy", f"wrongaddress@{chatmail_config.mail_domain}")
     result = s.getreply()
     print(result)
-    s.putcmd("vrfy", f"echo@{chatmail_config.mail_domain}")
+    s.putcmd("vrfy", addr)
     result2 = s.getreply()
     print(result2)
     assert result[0] == result2[0] == 252

cmdeploy/src/cmdeploy/tests/online/test_0_qr.py

@@ -20,7 +20,7 @@ def test_fastcgi_working(maildomain, chatmail_config):
 
 
 @pytest.mark.filterwarnings("ignore::urllib3.exceptions.InsecureRequestWarning")
-def test_newemail_configure(maildomain, maildomain_ip, rpc, chatmail_config):
+def test_newemail_configure(maildomain, rpc, chatmail_config):
     """Test configuring accounts by scanning a QR code works."""
     url = f"DCACCOUNT:https://{maildomain}/new"
     for i in range(3):
@@ -30,15 +30,12 @@ def test_newemail_configure(maildomain, maildomain_ip, rpc, chatmail_config):
             # set_config_from_qr, so fetch credentials via requests instead
             res = requests.post(f"https://{maildomain}/new", verify=False)
             data = res.json()
-            rpc.add_or_update_transport(
-                account_id,
-                {
-                    "addr": data["email"],
-                    "password": data["password"],
-                    "imapServer": maildomain_ip,
-                    "smtpServer": maildomain_ip,
-                    "certificateChecks": "acceptInvalidCertificates",
-                },
-            )
+            rpc.add_or_update_transport(account_id, {
+                "addr": data["email"],
+                "password": data["password"],
+                "imapServer": maildomain,
+                "smtpServer": maildomain,
+                "certificateChecks": "acceptInvalidCertificates",
+            })
         else:
             rpc.add_transport_from_qr(account_id, url)

cmdeploy/src/cmdeploy/tests/online/test_1_basic.py

@@ -12,9 +12,8 @@ from cmdeploy.cmdeploy import get_sshexec
 
 class TestSSHExecutor:
     @pytest.fixture(scope="class")
-    def sshexec(self, sshdomain, pytestconfig):
-        ssh_config = pytestconfig.getoption("ssh_config")
-        return get_sshexec(sshdomain, ssh_config=ssh_config)
+    def sshexec(self, sshdomain):
+        return get_sshexec(sshdomain)
 
     def test_ls(self, sshexec):
         out = sshexec(call=remote.rdns.shell, kwargs=dict(command="ls"))
@@ -72,6 +71,44 @@ class TestSSHExecutor:
         assert (now - since_date).total_seconds() < 60 * 60 * 51
 
 
+def test_dovecot_main_process_matches_installed_binary(sshdomain):
+    sshexec = get_sshexec(sshdomain)
+    main_pid = int(
+        sshexec(
+            call=remote.rshell.shell,
+            kwargs=dict(
+                command="timeout 10 systemctl show -p MainPID --value dovecot.service"
+            ),
+        ).strip()
+    )
+    assert main_pid != 0, "dovecot.service MainPID is 0 -- service not running?"
+
+    exe = sshexec(
+        call=remote.rshell.shell,
+        kwargs=dict(command=f"timeout 10 readlink /proc/{main_pid}/exe"),
+    ).strip()
+    status_text = sshexec(
+        call=remote.rshell.shell,
+        kwargs=dict(
+            command="timeout 10 systemctl show -p StatusText --value dovecot.service"
+        ),
+    ).strip()
+    installed_version = sshexec(
+        call=remote.rshell.shell, kwargs=dict(command="timeout 10 dovecot --version")
+    ).strip()
+
+    assert not exe.endswith("(deleted)"), (
+        f"running dovecot binary was deleted (stale after upgrade): {exe}"
+    )
+    expected_status_text = f"v{installed_version}"
+    assert status_text == expected_status_text or status_text.startswith(
+        f"{expected_status_text} "
+    ), (
+        f"dovecot status version mismatch: "
+        f"StatusText={status_text!r}, installed={installed_version!r}"
+    )
+
+
 def test_timezone_env(remote):
     for line in remote.iter_output("env"):
         print(line)
@@ -133,10 +170,11 @@ def test_authenticated_from(cmsetup, maildata):
 @pytest.mark.parametrize("from_addr", ["fake@example.org", "fake@testrun.org"])
 def test_reject_missing_dkim(cmsetup, maildata, from_addr):
     domain = cmsetup.maildomain
+    sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+    sock.settimeout(10)
     try:
-        sock = socket.create_connection((domain, 25), timeout=10)
-        sock.close()
-    except (socket.timeout, OSError):
+        sock.connect((domain, 25))
+    except socket.timeout:
         pytest.skip(f"port 25 not reachable for {domain}")
 
     recipient = cmsetup.gen_users(1)[0]

cmdeploy/src/cmdeploy/tests/online/test_2_deltachat.py

@@ -67,7 +67,7 @@ class TestEndToEndDeltaChat:
         assert msg2.get_snapshot().text == "message0"
 
     def test_exceed_quota(
-        self, cmfactory, lp, tmpdir, remote, chatmail_config, sshdomain, pytestconfig
+        self, cmfactory, lp, tmpdir, remote, chatmail_config, sshdomain
     ):
         """This is a very slow test as it needs to upload >100MB of mail data
         before quota is exceeded, and thus depends on the speed of the upload.
@@ -92,9 +92,7 @@ class TestEndToEndDeltaChat:
         lp.sec(f"filling remote inbox for {user}")
         fn = f"7743102289.M843172P2484002.c20,S={quota},W=2398:2,"
         path = chatmail_config.mailboxes_dir.joinpath(user, "cur", fn)
-        sshexec = get_sshexec(
-            sshdomain, ssh_config=pytestconfig.getoption("ssh_config")
-        )
+        sshexec = get_sshexec(sshdomain)
         sshexec(call=rshell.write_numbytes, kwargs=dict(path=str(path), num=120))
         res = sshexec(call=rshell.dovecot_recalc_quota, kwargs=dict(user=user))
         assert res["percent"] >= 100

cmdeploy/src/cmdeploy/tests/online/test_3_status.py

@@ -3,15 +3,12 @@ import os
 from cmdeploy.cmdeploy import main
 
 
-def test_status_cmd(chatmail_config, capsys, request, pytestconfig):
+def test_status_cmd(chatmail_config, capsys, request):
     os.chdir(request.config.invocation_params.dir)
     command = ["status"]
-    ssh_host = pytestconfig.getoption("ssh_host")
-    if ssh_host:
-        command.extend(["--ssh-host", ssh_host])
-    ssh_config = pytestconfig.getoption("ssh_config")
-    if ssh_config:
-        command.extend(["--ssh-config", ssh_config])
+    if os.getenv("CHATMAIL_SSH"):
+        command.append("--ssh-host")
+        command.append(os.getenv("CHATMAIL_SSH"))
     assert main(command) == 0
     status_out = capsys.readouterr()
     print(status_out.out)

cmdeploy/src/cmdeploy/tests/plugin.py

@@ -2,9 +2,7 @@ import imaplib
 import itertools
 import os
 import random
-import re
 import smtplib
-import socket
 import ssl
 import subprocess
 import time
@@ -20,76 +18,6 @@ def pytest_addoption(parser):
     parser.addoption(
         "--slow", action="store_true", default=False, help="also run slow tests"
     )
-    parser.addoption(
-        "--ssh-host",
-        dest="ssh_host",
-        default=None,
-        help="SSH host (overrides mail_domain for SSH operations).",
-    )
-    parser.addoption(
-        "--ssh-config",
-        dest="ssh_config",
-        default=None,
-        help="Path to an SSH config file (e.g. lxconfigs/ssh-config).",
-    )
-
-
-def _parse_ssh_config_hosts(path):
-    """Parse an OpenSSH config file and return a dict of hostname -> IP."""
-    mapping = {}
-    current_names = []
-    for ln in Path(path).read_text().splitlines():
-        line = ln.strip()
-        m = re.match(r"^Host\s+(.+)", line)
-        if m:
-            current_names = m.group(1).split()
-            continue
-        m = re.match(r"^Hostname\s+(\S+)", line)
-        if m and current_names:
-            ip = m.group(1)
-            for name in current_names:
-                mapping[name] = ip
-            current_names = []
-    return mapping
-
-
-_original_getaddrinfo = socket.getaddrinfo
-
-
-def _make_patched_getaddrinfo(host_map):
-    """Return a getaddrinfo that resolves hosts in host_map to their IPs."""
-
-    def patched_getaddrinfo(host, port, family=0, type=0, proto=0, flags=0):
-        if host in host_map:
-            ip = host_map[host]
-            return _original_getaddrinfo(ip, port, family, type, proto, flags)
-        return _original_getaddrinfo(host, port, family, type, proto, flags)
-
-    return patched_getaddrinfo
-
-
-@pytest.fixture(autouse=True, scope="session")
-def _setup_localchat_dns(pytestconfig):
-    """Monkey-patch socket.getaddrinfo to resolve .localchat via ssh-config."""
-    ssh_config = pytestconfig.getoption("ssh_config")
-    if not ssh_config or not Path(ssh_config).exists():
-        yield {}
-        return
-    host_map = _parse_ssh_config_hosts(ssh_config)
-    if not host_map:
-        yield {}
-        return
-    socket.getaddrinfo = _make_patched_getaddrinfo(host_map)
-    try:
-        yield host_map
-    finally:
-        socket.getaddrinfo = _original_getaddrinfo
-
-
-@pytest.fixture(scope="session")
-def ssh_config_host_map(_setup_localchat_dns):
-    """Return the host-name → IP map parsed from ssh-config."""
-    return _setup_localchat_dns
 
 
 def pytest_configure(config):
@@ -107,11 +35,11 @@ def pytest_runtest_setup(item):
 
 
 def _get_chatmail_config():
-    ini = os.environ.get("CHATMAIL_INI")
-    if ini:
-        path = Path(ini).resolve()
-        if path.exists():
-            return read_config(path), path
+    inipath = os.environ.get("CHATMAIL_INI")
+    if inipath:
+        path = Path(inipath).resolve()
+        return read_config(path), path
+
     current = Path().resolve()
     while 1:
         path = current.joinpath("chatmail.ini").resolve()
@@ -138,14 +66,8 @@ def maildomain(chatmail_config):
 
 
 @pytest.fixture(scope="session")
-def sshdomain(maildomain, pytestconfig):
-    return pytestconfig.getoption("ssh_host") or maildomain
-
-
-@pytest.fixture(scope="session")
-def maildomain_ip(maildomain, ssh_config_host_map):
-    """Return the IP for maildomain from ssh-config, or maildomain itself."""
-    return ssh_config_host_map.get(maildomain, maildomain)
+def sshdomain(maildomain):
+    return os.environ.get("CHATMAIL_SSH", maildomain)
 
 
 @pytest.fixture
@@ -389,22 +311,12 @@ from deltachat_rpc_client import DeltaChat, Rpc
 class ChatmailACFactory:
     """RPC-based account factory for chatmail testing."""
 
-    def __init__(
-        self,
-        rpc,
-        maildomain,
-        maildomain_ip,
-        gencreds,
-        chatmail_config,
-        ssh_config_host_map,
-    ):
+    def __init__(self, rpc, maildomain, gencreds, chatmail_config):
         self.dc = DeltaChat(rpc)
         self.rpc = rpc
         self._maildomain = maildomain
-        self._maildomain_ip = maildomain_ip
         self.gencreds = gencreds
         self.chatmail_config = chatmail_config
-        self._ssh_config_host_map = ssh_config_host_map
 
     def _make_transport(self, domain):
         """Build a transport config dict for the given domain."""
@@ -412,13 +324,11 @@ class ChatmailACFactory:
         transport = {
             "addr": addr,
             "password": password,
+            # Setting server explicitly skips requesting autoconfig XML,
+            # see https://datatracker.ietf.org/doc/draft-ietf-mailmaint-autoconfig/
+            "imapServer": domain,
+            "smtpServer": domain,
         }
-        # To support running against local relays without host DNS resolution
-        # we attempt resolving the domain via ssh-config
-        # because otherwise core fails to find the address
-        server = self._ssh_config_host_map.get(domain)
-        if server is not None:
-            transport.update({"imapServer": server, "smtpServer": server})
         if self.chatmail_config.tls_cert_mode == "self":
             transport["certificateChecks"] = "acceptInvalidCertificates"
         return transport
@@ -471,57 +381,62 @@ def rpc(tmp_path_factory):
 
 
 @pytest.fixture
-def cmfactory(
-    rpc, gencreds, maildomain, maildomain_ip, chatmail_config, ssh_config_host_map
-):
+def cmfactory(rpc, gencreds, maildomain, chatmail_config):
     """Return a ChatmailACFactory for creating online Delta Chat accounts."""
     return ChatmailACFactory(
         rpc=rpc,
         maildomain=maildomain,
-        maildomain_ip=maildomain_ip,
         gencreds=gencreds,
         chatmail_config=chatmail_config,
-        ssh_config_host_map=ssh_config_host_map,
     )
 
 
 @pytest.fixture
-def remote(sshdomain, pytestconfig):
-    return Remote(sshdomain, ssh_config=pytestconfig.getoption("ssh_config"))
+def remote(sshdomain):
+    r = Remote(sshdomain)
+    yield r
+    r.close()
 
 
 class Remote:
-    def __init__(self, sshdomain, ssh_config=None):
+    def __init__(self, sshdomain):
         self.sshdomain = sshdomain
-        self.ssh_config = ssh_config
+        self._procs = []
 
     def iter_output(self, logcmd="", ready=None):
         getjournal = "journalctl -f" if not logcmd else logcmd
         print(self.sshdomain)
         match self.sshdomain:
-            case "@local":
-                command = []
-            case "localhost":
-                command = []
-            case _:
-                command = ["ssh"]
-                if self.ssh_config:
-                    command.extend(["-F", self.ssh_config])
-                command.append(f"root@{self.sshdomain}")
+            case "@local": command = []
+            case "localhost": command = []
+            case _: command = ["ssh", f"root@{self.sshdomain}"]
         [command.append(arg) for arg in getjournal.split()]
-        self.popen = subprocess.Popen(
+        popen = subprocess.Popen(
             command,
+            stdin=subprocess.DEVNULL,
             stdout=subprocess.PIPE,
+            stderr=subprocess.DEVNULL,
         )
-        while 1:
-            line = self.popen.stdout.readline()
-            res = line.decode().strip().lower()
-            if not res:
-                break
-            if ready is not None:
-                ready()
-                ready = None
-            yield res
+        self._procs.append(popen)
+        try:
+            while 1:
+                line = popen.stdout.readline()
+                res = line.decode().strip().lower()
+                if not res:
+                    break
+                if ready is not None:
+                    ready()
+                    ready = None
+                yield res
+        finally:
+            popen.terminate()
+            popen.wait()
+
+    def close(self):
+        while self._procs:
+            proc = self._procs.pop()
+            proc.kill()
+            proc.wait()
 
 
 @pytest.fixture

cmdeploy/src/cmdeploy/tests/test_cmdeploy.py

@@ -23,18 +23,19 @@ class TestCmdline:
         run = parser.parse_args(["run"])
         assert init and run
 
-    def test_init_not_overwrite(self, tmp_path, capsys, monkeypatch):
+    def test_init_not_overwrite(self, capsys, tmp_path, monkeypatch):
         monkeypatch.delenv("CHATMAIL_INI", raising=False)
-        monkeypatch.chdir(tmp_path)
-
-        assert main(["init", "chat.example.org"]) == 0
+        inipath = tmp_path / "chatmail.ini"
+        args = ["init", "--config", str(inipath), "chat.example.org"]
+        assert main(args) == 0
         capsys.readouterr()
 
-        assert main(["init", "chat.example.org"]) == 1
+        assert main(args) == 1
         out, err = capsys.readouterr()
         assert "path exists" in out.lower()
 
-        assert main(["init", "chat.example.org", "--force"]) == 0
+        args.insert(1, "--force")
+        assert main(args) == 0
         out, err = capsys.readouterr()
         assert "deleting config file" in out.lower()
 

cmdeploy/src/cmdeploy/tests/test_dns.py

@@ -132,11 +132,28 @@ def test_parse_zone_records():
 
     ; Another comment
     www.some.domain. 3600 IN CNAME some.domain.
+
+    ; Multi-word rdata
+    some.domain. 3600 IN MX 10 mail.some.domain.
+
+    ; DKIM record (single line, multi-word TXT rdata)
+    dkim._domainkey.some.domain. 3600 IN TXT "v=DKIM1;k=rsa;p=MIIBIjANBgkqhkiG" "9w0BAQEFAAOCAQ8AMIIBCgKCAQEA"
+
+    ; Another TXT record
+    _dmarc.some.domain. 3600 IN TXT "v=DMARC1;p=reject"
     """
     records = list(parse_zone_records(text))
     assert records == [
         ("some.domain", "3600", "A", "1.1.1.1"),
         ("www.some.domain", "3600", "CNAME", "some.domain."),
+        ("some.domain", "3600", "MX", "10 mail.some.domain."),
+        (
+            "dkim._domainkey.some.domain",
+            "3600",
+            "TXT",
+            '"v=DKIM1;k=rsa;p=MIIBIjANBgkqhkiG" "9w0BAQEFAAOCAQ8AMIIBCgKCAQEA"',
+        ),
+        ("_dmarc.some.domain", "3600", "TXT", '"v=DMARC1;p=reject"'),
     ]
 
 
@@ -148,7 +165,6 @@ def test_parse_zone_records_invalid_line():
 
 def parse_zonefile_into_dict(zonefile, mockdns_base, only_required=False):
     if only_required:
-        # Only take records before the "; Recommended" section
         zonefile = zonefile.split("; Recommended")[0]
     for name, ttl, rtype, rdata in parse_zone_records(zonefile):
         mockdns_base.setdefault(rtype, {})[name] = rdata

cmdeploy/src/cmdeploy/tests/test_dovecot_deployer.py

@@ -0,0 +1,238 @@
+from contextlib import nullcontext
+from types import SimpleNamespace
+
+import pytest
+
+from cmdeploy.dovecot import deployer as dovecot_deployer
+from pyinfra.facts.deb import DebPackages
+
+
+def make_host(*fact_pairs):
+    """Build a mock host; get_fact(cls) dispatches to the provided facts mapping.
+
+    Args:
+        *fact_pairs: tuples of (fact_class, fact_value) to register
+
+    Returns:
+        SimpleNamespace with get_fact that raises a clear error if an
+        unexpected fact type is requested.
+    """
+    facts = dict(fact_pairs)
+
+    def get_fact(cls):
+        if cls not in facts:
+            registered = ", ".join(c.__name__ for c in facts)
+            raise LookupError(
+                f"unexpected get_fact({cls.__name__}); "
+                f"only registered: {registered}"
+            )
+        return facts[cls]
+
+    return SimpleNamespace(get_fact=get_fact)
+
+
+@pytest.fixture
+def deployer():
+    return dovecot_deployer.DovecotDeployer(
+        SimpleNamespace(mail_domain="chat.example.org"),
+        disable_mail=False,
+    )
+
+
+@pytest.fixture
+def patch_blocked(monkeypatch):
+    monkeypatch.setattr(dovecot_deployer, "blocked_service_startup", nullcontext)
+
+
+@pytest.fixture
+def mock_files_put(monkeypatch):
+    monkeypatch.setattr(
+        dovecot_deployer.files,
+        "put",
+        lambda **kwargs: SimpleNamespace(changed=False),
+    )
+
+
+@pytest.fixture
+def track_shell(monkeypatch):
+    calls = []
+    monkeypatch.setattr(
+        dovecot_deployer.server,
+        "shell",
+        lambda **kwargs: calls.append(kwargs) or SimpleNamespace(changed=False),
+    )
+    return calls
+
+
+def test_download_dovecot_package_skips_epoch_matched_install(monkeypatch):
+    epoch_version = dovecot_deployer.DOVECOT_PACKAGE_VERSION
+    downloads = []
+    monkeypatch.setattr(
+        dovecot_deployer,
+        "host",
+        make_host((DebPackages, {"dovecot-core": [epoch_version]})),
+    )
+    monkeypatch.setattr(
+        dovecot_deployer,
+        "_pick_url",
+        lambda primary, fallback: primary,
+    )
+    monkeypatch.setattr(
+        dovecot_deployer.files,
+        "download",
+        lambda **kwargs: downloads.append(kwargs),
+    )
+
+    deb, changed = dovecot_deployer._download_dovecot_package("core", "amd64")
+
+    assert deb is None, f"expected no deb path when version matches, got {deb!r}"
+    assert changed is False, "should not flag changed when version already installed"
+    assert downloads == [], "should not download when version already installed"
+
+
+def test_download_dovecot_package_uses_archive_version_for_url_and_filename(
+    monkeypatch,
+):
+    downloads = []
+    monkeypatch.setattr(
+        dovecot_deployer,
+        "host",
+        make_host((DebPackages, {})),
+    )
+    monkeypatch.setattr(
+        dovecot_deployer,
+        "_pick_url",
+        lambda primary, fallback: primary,
+    )
+    monkeypatch.setattr(
+        dovecot_deployer.files,
+        "download",
+        lambda **kwargs: downloads.append(kwargs),
+    )
+
+    deb, changed = dovecot_deployer._download_dovecot_package("core", "amd64")
+
+    archive_version = dovecot_deployer.DOVECOT_ARCHIVE_VERSION.replace("+", "%2B")
+    expected_deb = f"/root/dovecot-core_{archive_version}_amd64.deb"
+
+    # Verify the returned path uses archive version, not package version (with epoch)
+    assert changed is True, "should flag changed when package not yet installed"
+    assert deb == expected_deb, f"deb path mismatch: {deb!r} != {expected_deb!r}"
+    assert dovecot_deployer.DOVECOT_PACKAGE_VERSION not in deb, (
+        f"deb path should use archive version (no epoch), got {deb!r}"
+    )
+    assert len(downloads) == 1, "files.download should be called exactly once"
+
+
+def test_install_skips_dpkg_path_when_epoch_matched_packages_present(
+    deployer, patch_blocked, mock_files_put, track_shell, monkeypatch
+):
+    monkeypatch.setattr(
+        dovecot_deployer,
+        "host",
+        make_host(
+            (
+                dovecot_deployer.DebPackages,
+                {
+                    "dovecot-core": [dovecot_deployer.DOVECOT_PACKAGE_VERSION],
+                    "dovecot-imapd": [dovecot_deployer.DOVECOT_PACKAGE_VERSION],
+                    "dovecot-lmtpd": [dovecot_deployer.DOVECOT_PACKAGE_VERSION],
+                },
+            ),
+            (dovecot_deployer.Arch, "x86_64"),
+        ),
+    )
+    downloads = []
+    monkeypatch.setattr(
+        dovecot_deployer.files,
+        "download",
+        lambda **kwargs: downloads.append(kwargs),
+    )
+
+    deployer.install()
+
+    assert downloads == [], "should not download when all packages epoch-matched"
+    assert track_shell == [], "should not run dpkg when all packages epoch-matched"
+    assert deployer.need_restart is False, (
+        "need_restart should be False when nothing changed"
+    )
+
+
+def test_install_unsupported_arch_falls_back_to_apt(
+    deployer, patch_blocked, mock_files_put, track_shell, monkeypatch
+):
+    # For unsupported architectures, all fact lookups return the arch string.
+    monkeypatch.setattr(
+        dovecot_deployer,
+        "host",
+        SimpleNamespace(get_fact=lambda cls: "riscv64"),
+    )
+    apt_calls = []
+
+    # Mirrors apt.packages() return value: OperationMeta with .changed property.
+    # Only lmtpd triggers a change to verify |= accumulation of changed flags.
+    def fake_apt(**kwargs):
+        apt_calls.append(kwargs)
+        changed = "lmtpd" in kwargs["packages"][0]
+        return SimpleNamespace(changed=changed)
+
+    monkeypatch.setattr(dovecot_deployer.apt, "packages", fake_apt)
+
+    deployer.install()
+
+    actual_pkgs = [c["packages"] for c in apt_calls]
+    assert actual_pkgs == [["dovecot-core"], ["dovecot-imapd"], ["dovecot-lmtpd"]], (
+        f"expected apt install of core/imapd/lmtpd, got {actual_pkgs}"
+    )
+    assert track_shell == [], "should not run dpkg for unsupported arch"
+    assert deployer.need_restart is True, (
+        "need_restart should be True when apt installed a package"
+    )
+
+
+def test_install_runs_dpkg_when_packages_need_download(
+    deployer, patch_blocked, mock_files_put, track_shell, monkeypatch
+):
+    monkeypatch.setattr(
+        dovecot_deployer,
+        "host",
+        make_host(
+            (dovecot_deployer.DebPackages, {}),
+            (dovecot_deployer.Arch, "x86_64"),
+        ),
+    )
+    monkeypatch.setattr(
+        dovecot_deployer,
+        "_pick_url",
+        lambda primary, fallback: primary,
+    )
+    monkeypatch.setattr(
+        dovecot_deployer.files,
+        "download",
+        lambda **kwargs: SimpleNamespace(changed=True),
+    )
+
+    deployer.install()
+
+    assert len(track_shell) == 1, (
+        f"expected one server.shell() call for dpkg install, got {len(track_shell)}"
+    )
+    cmds = track_shell[0]["commands"]
+    assert len(cmds) == 3, f"expected 3 dpkg/apt commands, got: {cmds}"
+    assert cmds[0].startswith("dpkg --force-confdef --force-confold -i ")
+    assert "apt-get -y --fix-broken install" in cmds[1]
+    assert cmds[2].startswith("dpkg --force-confdef --force-confold -i ")
+    assert deployer.need_restart is True, (
+        "need_restart should be True after dpkg install"
+    )
+
+
+def test_pick_url_falls_back_on_primary_error(monkeypatch):
+    def raise_error(req, timeout):
+        raise OSError("connection timeout")
+
+    monkeypatch.setattr(dovecot_deployer.urllib.request, "urlopen", raise_error)
+    result = dovecot_deployer._pick_url("http://primary", "http://fallback")
+    assert result == "http://fallback", (
+        f"should fall back when primary fails, got {result!r}"
+    )

cmdeploy/src/cmdeploy/tests/test_lxc.py

@@ -1,174 +0,0 @@
-"""Tests for cmdeploy lxc-* subcommands."""
-
-import shutil
-import subprocess
-import sys
-
-import pytest
-
-from cmdeploy.lxc import cli
-from cmdeploy.lxc.incus import Incus
-from cmdeploy.util import Out
-
-pytestmark = pytest.mark.skipif(
-    not shutil.which("incus"),
-    reason="incus not installed",
-)
-
-
-# ---------------------------------------------------------------------------
-# Fixtures
-# ---------------------------------------------------------------------------
-
-
-@pytest.fixture
-def ix():
-    out = Out()
-    return Incus(out)
-
-
-@pytest.fixture(scope="session")
-def lxc_setup():
-    out = Out()
-    ix = Incus(out)
-    ix.get_dns_container().ensure()
-    return ix.list_managed()
-
-
-@pytest.fixture(scope="session")
-def relay_container(lxc_setup):
-    test_names = {f"{n}-localchat" for n in cli.RELAY_NAMES}
-    relays = [c for c in lxc_setup if c["name"] in test_names and c.get("ip")]
-    if not relays:
-        pytest.skip("no test relay containers running")
-    return relays[0]
-
-
-@pytest.fixture
-def cmdeploy():
-    def run(*args):
-        return subprocess.run(
-            [sys.executable, "-m", "cmdeploy.cmdeploy", *args],
-            capture_output=True,
-            text=True,
-            check=False,
-        )
-
-    return run
-
-
-# ---------------------------------------------------------------------------
-# Tests
-# ---------------------------------------------------------------------------
-
-
-@pytest.mark.parametrize(
-    "subcmd, expected, absent",
-    [
-        (None, ["lxc-start", "lxc-stop", "lxc-test", "lxc-status"], ["lxc-destroy"]),
-        ("lxc-start", ["--ipv4-only", "--run"], ["--config"]),
-        ("lxc-stop", ["--destroy", "--destroy-all"], ["--config"]),
-        ("lxc-test", ["--one"], ["--config"]),
-        ("lxc-status", [], ["--config"]),
-        ("run", ["--ssh-config"], ["--lxc"]),
-        ("dns", ["--ssh-config"], []),
-        ("test", ["--ssh-config"], []),
-        ("status", ["--ssh-config"], []),
-    ],
-)
-def test_help_options(cmdeploy, subcmd, expected, absent):
-    args = [subcmd, "--help"] if subcmd else ["--help"]
-    result = cmdeploy(*args)
-    output = result.stdout + result.stderr
-    assert result.returncode == 0
-    for flag in expected:
-        assert flag in output
-    for flag in absent:
-        assert flag not in output
-
-
-class TestSSHConfig:
-    def test_lxconfigs(self, ix, lxc_setup):
-        d = ix.lxconfigs_dir
-        assert d.name == "lxconfigs"
-        assert d.exists()
-        path = ix.ssh_config_path
-        assert path.name == "ssh-config"
-        assert path.parent.name == "lxconfigs"
-
-    def test_write_ssh_config(self, ix, lxc_setup):
-        path = ix.write_ssh_config()
-        assert path.exists()
-        text = path.read_text()
-
-        for c in lxc_setup:
-            if c.get("ip"):
-                assert c["name"] in text
-                assert f"Hostname {c['ip']}" in text
-
-        assert "User root" in text
-        assert "IdentityFile" in text
-        assert "StrictHostKeyChecking accept-new" in text
-
-
-def test_dns(ix, relay_container):
-    def dig(qname, qtype):
-        ct = ix.get_dns_container()
-        return ct.bash(f"dig @127.0.0.1 {qname} {qtype} +short").strip()
-
-    domain = relay_container["domain"]
-    assert dig(domain, "A") == relay_container["ip"]
-    assert domain in dig(domain, "MX")
-    assert "587" in dig(f"_submission._tcp.{domain}", "SRV")
-
-
-class TestLxcStatus:
-    def test_cli_lxc_status_help(self, cmdeploy):
-        result = cmdeploy("lxc-status", "--help")
-        assert result.returncode == 0
-        assert "status" in result.stdout.lower()
-
-    def test_shows_containers(self, lxc_setup, capsys):
-        class QuietOut(Out):
-            def red(self, msg, **kw):
-                pass
-
-            def green(self, msg, **kw):
-                pass
-
-        ret = cli.lxc_status_cmd(None, QuietOut())
-        assert ret == 0
-        captured = capsys.readouterr().out
-        assert "ns-localchat" in captured
-        assert "running" in captured
-
-    def test_deploy_freshness(self, ix, monkeypatch):
-        ct = ix.get_container("x")
-
-        monkeypatch.setattr(
-            "cmdeploy.lxc.incus.RelayContainer.deployed_version",
-            lambda _self: "abc123def456",
-        )
-        monkeypatch.setattr(
-            "cmdeploy.lxc.incus.RelayContainer.deployed_domain",
-            lambda _self: ct.domain,
-        )
-        monkeypatch.setattr(
-            "cmdeploy.lxc.cli.get_version_string",
-            lambda: "abc123def456",
-        )
-        assert "IN-SYNC" in cli._deploy_status(ct, "abc123def456", ix)
-        assert "STALE" in cli._deploy_status(ct, "other_hash_here", ix)
-
-        # Hash matches but local has uncommitted changes
-        monkeypatch.setattr(
-            "cmdeploy.lxc.cli.get_version_string",
-            lambda: "abc123def456\ndiff --git a/foo",
-        )
-        assert "DIRTY" in cli._deploy_status(ct, "abc123def456", ix)
-
-        monkeypatch.setattr(
-            "cmdeploy.lxc.incus.RelayContainer.deployed_version",
-            lambda _self: None,
-        )
-        assert "NOT DEPLOYED" in cli._deploy_status(ct, "abc123", ix)

cmdeploy/src/cmdeploy/tests/test_util.py

@@ -1,120 +0,0 @@
-import sys
-
-from cmdeploy.util import Out, collapse, get_git_hash, get_version_string, shell
-
-
-class TestOut:
-    def test_prefix_default(self, capsys):
-        out = Out()
-        out.print("hello")
-        assert capsys.readouterr().out == "hello\n"
-
-    def test_prefix_custom(self, capsys):
-        out = Out(prefix=">> ")
-        out.print("hello")
-        assert capsys.readouterr().out == ">> hello\n"
-
-    def test_prefix_print_file(self):
-        import io
-
-        buf = io.StringIO()
-        out = Out(prefix=":: ")
-        out.print("msg", file=buf)
-        assert ":: msg" in buf.getvalue()
-
-    def test_new_prefixed_out(self, capsys):
-        parent = Out(prefix="A")
-        child = parent.new_prefixed_out("B")
-        child.print("x")
-        assert capsys.readouterr().out == "ABx\n"
-        # shares section_timings
-        assert child.section_timings is parent.section_timings
-
-    def test_section_no_auto_indent(self, capsys):
-        out = Out(prefix="")
-        with out.section("test"):
-            out.print("inside")
-        captured = capsys.readouterr().out
-        # "inside" should NOT be indented by section()
-        lines = captured.strip().splitlines()
-        inside_line = [l for l in lines if "inside" in l][0]
-        assert inside_line == "inside"
-
-    def test_section_records_timing(self):
-        out = Out()
-        with out.section("s1"):
-            pass
-        assert len(out.section_timings) == 1
-        assert out.section_timings[0][0] == "s1"
-
-    def test_shell_failure_shows_output(self):
-        """When a shell command fails, its output and exit code are shown."""
-        import subprocess
-
-        result = subprocess.run(
-            [
-                sys.executable,
-                "-c",
-                "from cmdeploy.util import Out; Out(prefix='').shell("
-                "\"echo 'boom on stderr' >&2; exit 42\")",
-            ],
-            capture_output=True,
-            text=True,
-            check=False,
-        )
-        # the command's stderr is merged into stdout by Popen
-        assert "boom on stderr" in result.stdout
-        # Out.red() prints the failure notice to stderr
-        assert "exit code 42" in result.stderr
-
-
-def test_collapse():
-    text = """
-        line 1
-        line 2
-    """
-    assert collapse(text) == "line 1 line 2"
-    assert collapse("  single line  ") == "single line"
-
-
-def test_git_helpers_no_git(tmp_path):
-    # Not a git repo
-    assert get_git_hash(root=tmp_path) is None
-    assert get_version_string(root=tmp_path) == "unknown"
-
-
-def test_git_helpers_empty_repo(tmp_path):
-    shell("git init", cwd=tmp_path, check=True)
-    # No commits yet
-    assert get_git_hash(root=tmp_path) is None
-    assert get_version_string(root=tmp_path) == "unknown"
-
-
-def test_git_helpers_with_commits_and_diffs(tmp_path):
-    shell("git init", cwd=tmp_path, check=True)
-    shell("git config user.email you@example.com", cwd=tmp_path, check=True)
-    shell("git config user.name 'Your Name'", cwd=tmp_path, check=True)
-
-    # First commit
-    path = tmp_path / "file.txt"
-    path.write_text("content")
-    shell("git add file.txt", cwd=tmp_path, check=True)
-    shell("git commit -m initial", cwd=tmp_path, check=True)
-
-    git_hash = get_git_hash(root=tmp_path)
-    assert len(git_hash) >= 7  # usually 40, but git is git
-    assert get_version_string(root=tmp_path) == git_hash
-
-    # Create a diff
-    path.write_text("new content")
-    v = get_version_string(root=tmp_path)
-    assert v.startswith(git_hash + "\n")
-    assert "new content" in v
-    assert not v.endswith("\n")
-
-    # Commit again -> no diff
-    shell("git add file.txt", cwd=tmp_path, check=True)
-    shell("git commit -m second", cwd=tmp_path, check=True)
-    new_hash = get_git_hash(root=tmp_path)
-    assert new_hash != git_hash
-    assert get_version_string(root=tmp_path) == new_hash

cmdeploy/src/cmdeploy/util.py

@@ -1,168 +0,0 @@
-"""Shared utility functions for cmdeploy."""
-
-import os
-import shutil
-import subprocess
-import sys
-import textwrap
-import time
-from contextlib import contextmanager
-from pathlib import Path
-
-from termcolor import colored
-
-
-class Out:
-    """Convenience output printer providing coloring and section formatting."""
-
-    def __init__(self, sepchar="\u2501", prefix="", verbosity=0):
-        self.section_timings = []
-        self.prefix = prefix
-        self.sepchar = sepchar
-        self.verbosity = verbosity
-        env_width = os.environ.get("_CMDEPLOY_WIDTH")
-        if env_width:
-            self.section_width = int(env_width)
-        else:
-            self.section_width = shutil.get_terminal_size((80, 24)).columns
-
-    def new_prefixed_out(self, newprefix="  "):
-        """Return a new Out with an extended prefix,
-        sharing section_timings with the parent.
-        """
-        out = Out(
-            sepchar=self.sepchar,
-            prefix=self.prefix + newprefix,
-            verbosity=self.verbosity,
-        )
-        out.section_timings = self.section_timings
-        return out
-
-    def red(self, msg, file=sys.stderr):
-        print(colored(self.prefix + msg, "red"), file=file, flush=True)
-
-    def green(self, msg, file=sys.stderr):
-        print(colored(self.prefix + msg, "green"), file=file, flush=True)
-
-    def print(self, msg="", **kwargs):
-        """Print to stdout with automatic flush."""
-        if msg:
-            msg = self.prefix + msg
-        print(msg, flush=True, **kwargs)
-
-    def _format_header(self, title):
-        """Return a formatted section header string."""
-        width = self.section_width - len(self.prefix)
-        bar = self.sepchar * (width - len(title) - 5)
-        return f"{self.sepchar * 3} {title} {bar}"
-
-    @contextmanager
-    def section(self, title):
-        """Context manager that prints a section header and records elapsed time."""
-        self.green(self._format_header(title))
-        t0 = time.time()
-        yield
-        elapsed = time.time() - t0
-        self.section_timings.append((title, elapsed))
-
-    def section_line(self, title):
-        """Print a section header without timing."""
-        self.green(self._format_header(title))
-
-    def shell(self, cmd, quiet=False, **kwargs):
-        """Print *cmd*, run it, and re-print its output with the current prefix.
-
-        *cmd* is passed through :func:`collapse`, so callers
-        can use triple-quoted f-strings freely.
-        Stdout and stderr are merged, read line-by-line,
-        and each line is printed with ``self.prefix`` prepended.
-        When the command exits non-zero, a red error line is printed.
-        """
-        cmd = collapse(cmd)
-        if not quiet:
-            self.print(f"$ {cmd}")
-        indent = self.prefix + "  "
-        env = kwargs.pop("env", None)
-        if env is None:
-            env = os.environ.copy()
-        env["_CMDEPLOY_WIDTH"] = str(self.section_width - len(indent))
-        proc = subprocess.Popen(
-            cmd,
-            shell=True,
-            text=True,
-            stdout=subprocess.PIPE,
-            stderr=subprocess.STDOUT,
-            env=env,
-            **kwargs,
-        )
-        for line in proc.stdout:
-            sys.stdout.write(indent + line)
-            sys.stdout.flush()
-        ret = proc.wait()
-        if ret:
-            self.red(f"command failed with exit code {ret}: {cmd}")
-        return ret
-
-
-def _project_root():
-    """Return the project root directory."""
-    return Path(__file__).resolve().parent.parent.parent.parent
-
-
-def collapse(text):
-    """Dedent, join lines, and strip a (triple-quoted) string.
-
-    Handy for writing shell commands across multiple lines::
-
-        cmd = collapse(f\"""
-            cmdeploy run
-            --config {ct.ini}
-            --ssh-host {ct.domain}
-        \""")
-    """
-    return textwrap.dedent(text).replace("\n", " ").strip()
-
-
-def shell(cmd, check=False, **kwargs):
-    """Run a shell command string with sensible defaults.
-
-    *cmd* is passed through :func:`collapse` first, so callers
-    can use triple-quoted f-strings freely.
-    Captures stdout/stderr by default; pass ``capture_output=False``
-    to stream output to the terminal instead.
-    """
-    if "capture_output" not in kwargs and "stdout" not in kwargs:
-        kwargs["capture_output"] = True
-    return subprocess.run(collapse(cmd), shell=True, text=True, check=check, **kwargs)
-
-
-def get_git_hash(root=None):
-    """Return the local HEAD commit hash, or None."""
-    if root is None:
-        root = _project_root()
-    result = shell(
-        "git rev-parse HEAD",
-        cwd=str(root),
-    )
-    if result.returncode == 0:
-        return result.stdout.strip()
-    return None
-
-
-def get_version_string(root=None):
-    """Return ``git_hash\\ngit_diff`` for the local working tree.
-
-    Used by :class:`~cmdeploy.deployers.GithashDeployer` to write
-    ``/etc/chatmail-version`` and by ``lxc-status`` to compare
-    the deployed state against the local checkout.
-    """
-    if root is None:
-        root = _project_root()
-    git_hash = get_git_hash(root=root) or "unknown"
-    try:
-        git_diff = shell("git diff", cwd=str(root)).stdout.strip()
-    except Exception:
-        git_diff = ""
-    if git_diff:
-        return f"{git_hash}\n{git_diff}"
-    return git_hash

doc/source/conf.py

@@ -15,7 +15,7 @@ author = 'chatmail collective'
 
 extensions = [
     #'sphinx.ext.autodoc',
-    #'sphinx.ext.viewcode',
+    #'sphinx.ext.viewdoc',
     'sphinxcontrib.mermaid',
 ]
 

doc/source/getting_started.rst

@@ -98,6 +98,13 @@ steps. Please substitute it with your own domain.
    configure at your DNS provider (it can take some time until they are
    public).
 
+Docker installation
+-------------------
+
+There is experimental support for running chatmail via Docker Compose.
+See the `chatmail/docker README <https://github.com/chatmail/docker>`_
+for full setup instructions.
+
 Other helpful commands
 ----------------------
 

doc/source/index.rst

@@ -16,6 +16,5 @@ Contributions and feedback welcome through the https://github.com/chatmail/relay
     proxy
     migrate
     overview
-    lxc
     related
     faq

doc/source/lxc.rst

@@ -1,271 +0,0 @@
-Local testing with LXC/Incus
-============================
-
-The ``cmdeploy`` tool includes support for running
-chatmail relays inside local
-`Incus <https://linuxcontainers.org/incus/>`_ LXC containers.
-This is meant for development, testing, and CI
-without requiring a remote server.
-LXC system containers are lightweight virtual machines
-that share the host's kernel but run their own init system,
-package manager, and network stack,
-so the cmdeploy deployment scripts work pretty much
-as they would on a real Debian server or cloud VPS.
-
-Prerequisites
--------------
-
-Install `Incus <https://linuxcontainers.org/incus/>`_
-(LXC container manager).
-See the `official installation guide
-<https://linuxcontainers.org/incus/docs/main/installing/>`_
-for full details.
-
-After installing incus, initialise and grant yourself access::
-
-    sudo incus admin init --minimal
-    sudo usermod -aG incus-admin $USER
-
-
-.. caution::
-
-   Adding yourself to ``incus-admin`` grants effective root access
-   to the host: any member can mount host directories into a container
-   and manipulate them as root.
-   This is fine for local testing of your own relay branches,
-   but do **not** use it for production setups
-   or for testing untrusted relay branches from others.
-
-.. warning::
-
-   You **must now log out and back in** (or run ``newgrp incus-admin``)
-   after adding yourself to the group.
-   Without this, all ``cmdeploy lxc-*`` commands
-   will fail with permission errors.
-
-Verify the installation works by running ``incus list``,
-which should print an empty table without errors.
-
-
-Quick start
------------
-
-::
-
-    cd relay
-    scripts/initenv.sh               # bootstrap venv
-    source venv/bin/activate         # activate venv
-    cmdeploy lxc-test                # create containers, deploy, test
-
-The ``lxc-test`` command provides an automated way
-to run the full deployment and test pipeline.
-It executes several ``cmdeploy`` subcommands in sequential steps.
-If a step fails, you can copy-paste the printed command
-and run it manually to debug.
-No host DNS delegation or ``~/.ssh/config`` changes are needed
-because ``lxc-test`` passes the required SSH and DNS options directly.
-
-
-CLI reference
---------------
-
-``lxc-start [--ipv4-only] [--run] [NAME ...]``
-    Create and start containers.
-    Without arguments, creates ``test0-localchat`` and ``ns-localchat`` (DNS).
-    Pass one or more ``NAME`` arguments to create user relay containers instead
-    (e.g. ``cmdeploy lxc-start myrelay``).
-    Use ``--ipv4-only`` to set ``disable_ipv6 = True`` in the generated ``chatmail.ini``,
-    producing an IPv4-only relay.
-    Use ``--run`` to automatically run ``cmdeploy run`` on each container after starting it.
-    Generates ``lxconfigs/ssh-config``.
-    It reuses existing containers and resets DNS zones to minimal records.
-
-``lxc-stop [--destroy] [--destroy-all] [NAME ...]``
-    Stop relay containers.
-    Without arguments, stops ``test0-localchat`` and ``test1-localchat``.
-    Pass ``NAME`` to stop specific containers.
-    Use ``--destroy`` to also delete the containers and their config files.
-    Use ``--destroy-all`` to additionally destroy
-    the ``ns-localchat`` DNS container **and** remove
-    the cached ``localchat-base`` and ``localchat-relay``
-    images, giving a fully clean slate for the next ``lxc-test``.
-    User containers are **never** destroyed unless named explicitly.
-
-``lxc-test [--one]``
-    By default creates, deploys, and tests both ``test0`` and ``test1``
-    for dual-domain federation testing (sets ``CHATMAIL_DOMAIN2=_test1.localchat``).
-    test0 runs dual-stack (IPv4 + IPv6) while test1 runs IPv4-only (``disable_ipv6 = True``).
-    Pass ``--one`` to only deploy and test against ``test0``
-    (skips ``test1``, does not set ``CHATMAIL_DOMAIN2``).
-
-``lxc-status``
-    Show live status of all LXC containers (including the DNS container),
-    deploy freshness (comparing ``/etc/chatmail-version``
-    against local ``git rev-parse HEAD`` and ``git diff``),
-    SSH config inclusion, and host DNS forwarding for ``.localchat``.
-    Reports **IN-SYNC**, **DIRTY** (hash matches but uncommitted changes exist),
-    **STALE** (different commit), or **NOT DEPLOYED**.
-
-
-Container types
------------------
-
-**Test relay containers** (``test0-localchat``, ``test1-localchat``)
-    Created automatically by ``lxc-test``.
-    **test0** has IPv4 and IPv6 configured,
-    **test1** is IPv4-only (``disable_ipv6 = True``).
-
-**User relay containers** (``<name>-localchat``)
-    Created by ``cmdeploy lxc-start <name>``
-    where ``<name>`` does not start with ``test``.
-    These are personal development instances,
-    never touched by ``lxc-stop --destroy`` unless named explicitly.
-
-**DNS container** (``ns-localchat``)
-    Singleton container running PowerDNS.
-    Created automatically when any relay is started.
-
-
-.. _lxc-ssh-config:
-
-SSH configuration
------------------
-
-``cmdeploy lxc-start`` generates ``lxconfigs/ssh-config``,
-a standard OpenSSH config file mapping every container name,
-its domain, and a short alias to the container's IP address::
-
-    Host test0-localchat _test0.localchat _test0
-        Hostname 10.204.0.42
-        User root
-        IdentityFile /path/to/relay/lxconfigs/id_localchat
-        IdentitiesOnly yes
-        StrictHostKeyChecking accept-new
-        UserKnownHostsFile /dev/null
-        LogLevel ERROR
-
-All ``cmdeploy`` commands (``run``, ``dns``, ``status``, ``test``)
-accept ``--ssh-config lxconfigs/ssh-config`` to use this file.
-``lxc-test`` passes it automatically.
-
-**Using containers from the host shell:**
-
-To make ``ssh _test0`` work from any terminal, add one line to ``~/.ssh/config``::
-
-    Include /absolute/path/to/relay/lxconfigs/ssh-config
-
-
-.. _lxc-dns-setup:
-.. _localchat-tld:
-
-``.localchat`` DNS and name resolution
----------------------------------------
-
-All LXC-managed chatmail domains use the ``.localchat`` pseudo-TLD
-(e.g. ``_test0.localchat``, ``_test1.localchat``),
-a non-delegated suffix that exists only within the local PowerDNS infrastructure.
-A dedicated DNS container (``ns-localchat``)
-is created so that local test relays interact
-with DNS similar to a regular public Internet setup.
-On first start, ``cmdeploy lxc-start`` creates this container
-running two `PowerDNS <https://www.powerdns.com/>`_ services:
-
-* **pdns-server** (authoritative) serves ``.localchat``
-  zones from a local SQLite database.
-
-* **pdns-recursor** (recursive) listens on the Incus
-  bridge so all containers can use it.
-  Forwards ``.localchat`` queries to the local
-  authoritative server and everything else to Quad9 (``9.9.9.9``).
-
-After the DNS container is up, ``lxc-start`` configures the Incus bridge
-to advertise its IP via DHCP and disables Incus's own DNS.
-DNS records are then created in two phases matching the "cmdeploy run" deployment flow:
-
-1. **``lxc-start``** resets each relay zone to
-   **SOA, NS, and A** records (plus **AAAA** for dual-stack containers).
-   If host DNS resolution is configured, users can
-   afterwards run ``cmdeploy run --config lxconfigs/chatmail-test0.ini
-   --ssh-config lxconfigs/ssh-config --ssh-host _test0.localchat``.
-   LXC subcommands do not depend on host DNS resolution
-   and resolve addresses via ``lxconfigs/ssh-config``.
-
-2. **``cmdeploy dns --zonefile``** generates a standard
-   BIND-format zone file (MX, TXT/SPF, TXT/DMARC,
-   TXT/MTA-STS, SRV, CNAME, DKIM) and loads it
-   into PowerDNS.
-
-This two-phase approach prevents premature configuration of mail records
-before the relay is actually deployed and running.
-Once ``cmdeploy run`` deploys `Unbound <https://nlnetlabs.nl/projects/unbound/>`_
-inside a relay container, Unbound has a configuration plugin snippet
-that forwards all ``.localchat`` queries to the PowerDNS recursor,
-and lets all other queries go through normal recursive resolution.
-
-
-State outside the repository
------------------------------
-
-All generated configuration by lxc subcommands live in ``lxconfigs/``
-(git-ignored), including the SSH key pair (``id_localchat``),
-per-container ``chatmail-*.ini`` files, zone files, and ``ssh-config``.
-
-The only state *outside* the repository is the Incus containers and images themselves
-(managed via the ``incus`` CLI, labelled with ``user.localchat-managed=true``).
-The Incus image store retains the following snapshot images:
-
-* ``localchat-base``: Debian 12 with openssh-server and Python (built on first run)
-
-* ``localchat-relay``: fully deployed relay snapshot,
-  cached after the first successful ``cmdeploy run``.
-  Subsequent relay containers launch from this image
-  so the deploy step is mostly no-ops (roughly 3× faster than a fresh deploy).
-
-
-.. _lxc-tls:
-
-TLS handling and underscore domains
-------------------------------------
-
-Container domains start with ``_`` (e.g. ``_test0.localchat``).
-As described in :doc:`getting_started` ("Running a relay with self-signed certificates"),
-underscore domains automatically use self-signed TLS
-and ``smtp_tls_security_level = encrypt``.
-This permits cross-relay federation between LXC containers
-without any external certificate authority.
-Delta Chat clients connecting to these relays
-must be configured with
-``certificateChecks = acceptInvalidCertificates``
-(the test fixtures handle this automatically).
-`PR #7926 on chatmail-core <https://github.com/chatmail/core/pull/7926>`_
-is meant to make this special setting unnecessary for chatmail clients
-that are connecting to underscore domains.
-
-
-Known limitations
-------------------
-
-The LXC environment differs from a production
-deployment in several ways:
-
-**No ACME / Let's Encrypt**:
-  Self-signed TLS only (see :ref:`lxc-tls`);
-  ACME code paths are never exercised locally.
-
-**No inbound connections from the internet**:
-  Containers sit on a private Incus bridge and are not port-forwarded.
-  Only the host and other containers on the same bridge can reach them.
-
-**Local federation only**:
-  Cross-relay mail delivery (e.g. test0 → test1) works between containers on the same host,
-  but these relays are invisible to any external mail server.
-
-**DNS is local only**:
-  The ``.localchat`` pseudo-TLD is not resolvable from the wider internet
-  (see :ref:`lxc-dns-setup`).
-
-**IPv6 is ULA-only**:
-  Containers receive IPv6 addresses from the ``fd42:...`` ULA range on the Incus bridge.
-  These are not globally routable, but are sufficient for testing IPv6 service binding
-  (Postfix, Dovecot, Nginx) and DNS AAAA records inside the local environment.
-  test1 runs with ``disable_ipv6 = True`` to exercise the IPv4-only deployment path.