Merge branch 'claude/relaxed-faraday-h4zd09' of https://github.com/pablovolenski/vienalatina

The server's main tracks gitea/main while new work lands on a branch on
GitHub, so a bare `git pull` asks Gitea, finds it level, and answers
"Already up to date." That sentence is true about the wrong remote and
reads exactly like there is nothing to fetch — which is how the last
deploy stalled on a script that had never arrived.

Step 8 now names a `github` remote beside `gitea`, and the update recipe
in 11.3 pulls from it by name with --no-edit, pushes to gitea, then
rebuilds. 11.3 previously said plain `git pull`, as though the server
tracked GitHub.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NizVpJ2dwzCbjCrTLCjeHn

apps/board/members.py

@@ -129,6 +129,10 @@ def new():
             # The form says so before it is filled in; offering a ticked checkbox
             # and reporting the problem on submit wastes the work of filling it.
             can_create_accounts=bool(current_app.config.get("ADMIN_TOKEN")),
+            # Same courtesy for mail: without a server configured the invitation
+            # cannot leave the box, and the admin has to pass the link on by
+            # hand. Worth knowing before filling the form rather than after.
+            can_send_mail=mail.configured(),
             gitea_url=current_app.config["GITEA_URL"].rstrip("/"),
         )
 
@@ -175,19 +179,25 @@ def new():
         return redirect(url_for("members.new"))
 
     invite_link = None
+    mail_problem = None
     if create_account:
         link = auth.invite_url(invites.issue(cursor.lastrowid, "invite"))
         try:
             mail.send_invite(email, display_name or login, link)
-        except (mail.MailFailed, mail.MailNotConfigured):
+        except mail.MailNotConfigured:
+            # Nothing is broken — this server has simply never been given a mail
+            # server. Told apart from a failure on purpose: an admin sent looking
+            # for an SMTP error that does not exist is an afternoon wasted.
+            invite_link, mail_problem = link, "unconfigured"
+        except mail.MailFailed:
             # The account exists and the member cannot reach it. Showing the
             # admin the link is the difference between a delayed invitation and
             # a person who simply never gets in.
-            invite_link = link
+            invite_link, mail_problem = link, "failed"
 
     return render_template("member_created.html", login=login,
                            email=email, created=create_account,
-                           invite_link=invite_link,
+                           invite_link=invite_link, mail_problem=mail_problem,
                            role_label=ROLE_LABELS[role])
 
 

apps/board/templates/member_created.html

@@ -12,14 +12,28 @@
 	{# The account exists but the invitation never left the building. Showing the
 	   link is the difference between a delayed invitation and a person who
 	   simply never gets in. #}
-	<p class="flash flash--error">
-		No se pudo enviar el correo. Pásale este enlace por un canal privado —
-		sirve una sola vez y caduca en 7 días.
+	{% if mail_problem == 'unconfigured' %}
+	<p class="flash">
+		Este servidor todavía no envía correo. Pásale este enlace por un canal
+		privado — sirve una sola vez y caduca en 7 días.
 	</p>
+	{% else %}
+	<p class="flash flash--error">
+		No se pudo enviar el correo: el servidor de correo rechazó el envío o no
+		respondió. Pásale este enlace por un canal privado — sirve una sola vez y
+		caduca en 7 días.
+	</p>
+	{% endif %}
 	<p class="password mono">{{ invite_link }}</p>
 	<p class="muted small">
+		{% if mail_problem == 'unconfigured' %}
+		Añade los valores <span class="mono">MAIL_*</span> en
+		<span class="mono">/srv/board/.env</span> para que la próxima invitación
+		salga sola.
+		{% else %}
 		Revisa la configuración de correo del servidor para que la próxima
 		invitación salga sola.
+		{% endif %}
 	</p>
 
 	{% else %}

apps/board/templates/member_new.html

@@ -37,6 +37,16 @@
 	</p>
 	{% endif %}
 
+	{% if can_create_accounts and not can_send_mail %}
+	{# Not an error: the account is still created and the invitation still
+	   issued. But it leaves by hand, and finding that out after filling the
+	   form is how an invitation ends up believed sent and never delivered. #}
+	<p class="muted small">
+		Este servidor todavía no envía correo, así que la invitación no se
+		enviará sola: al terminar te mostraremos el enlace para que se lo pases.
+	</p>
+	{% endif %}
+
 	{% if can_make_admin %}
 	<label for="role">Rol</label>
 	<select id="role" name="role">

apps/board/tests/test_deployment.py

@@ -0,0 +1,46 @@
+"""Guards on the gap between what the app reads and what the server hands it.
+
+This is the one class of bug the rest of the suite cannot see. Every test here
+runs against a Flask config built in-process, so a setting can be documented in
+`.env.example`, read in `app.py`, and still reach the container as nothing at
+all — because Compose does not pass `.env` to a container, it only substitutes
+into the compose file. The symptom is a feature reporting itself unconfigured
+with the settings sitting right there on disk, which is a miserable thing to
+debug from the outside.
+"""
+
+from __future__ import annotations
+
+import re
+from pathlib import Path
+
+INFRA = Path(__file__).resolve().parents[3] / "infra" / "board"
+ENV_EXAMPLE = INFRA / ".env.example"
+COMPOSE = INFRA / "docker-compose.yml"
+
+SETTING = re.compile(r"^([A-Z][A-Z0-9_]*)=", re.MULTILINE)
+
+
+def documented() -> set[str]:
+    return set(SETTING.findall(ENV_EXAMPLE.read_text(encoding="utf-8")))
+
+
+def test_every_documented_setting_reaches_the_container():
+    compose = COMPOSE.read_text(encoding="utf-8")
+    missing = sorted(
+        name for name in documented()
+        if f"${{{name}}}" not in compose and f"${{{name}:-" not in compose
+    )
+    assert not missing, (
+        "these are in .env.example but never substituted into "
+        f"docker-compose.yml, so the container never sees them: {missing}"
+    )
+
+
+def test_the_mail_settings_are_optional_at_the_compose_level():
+    """`${MAIL_HOST}` without a default makes Compose warn on every command for
+    a server that has deliberately not configured mail. Running without it is a
+    supported state, so it must be a quiet one."""
+    compose = COMPOSE.read_text(encoding="utf-8")
+    for name in sorted(n for n in documented() if n.startswith("MAIL_")):
+        assert f"${{{name}:-" in compose, f"{name} has no default in docker-compose.yml"

apps/board/tests/test_invites.py

@@ -262,3 +262,35 @@ def test_linking_an_existing_account_sends_nothing(
         "role": "user", "create_account": "",
     })
     assert outbox == []
+
+
+# --- when the server cannot send at all -----------------------------------
+
+def test_an_unconfigured_server_does_not_blame_the_mail_server(
+        app, client, post, owner_id, sign_in, monkeypatch):
+    """No MAIL_HOST is not a failure, and saying "no se pudo enviar" sends the
+    admin hunting for an SMTP error that was never produced."""
+    monkeypatch.setattr(gitea, "admin_create_user",
+                        lambda login, email, name, password: None)
+    app.config["MAIL_HOST"] = ""
+    sign_in(owner_id)
+
+    page = post("/comunidad/miembros/nuevo", {
+        "login": "maria", "display_name": "María", "email": "m@example.com",
+        "role": "user", "create_account": "on",
+    }).get_data(as_text=True)
+
+    assert "todavía no envía correo" in page
+    assert "No se pudo enviar el correo" not in page
+    assert "/comunidad/invitacion/" in page
+
+
+def test_the_form_warns_before_it_is_filled_in(app, client, owner_id, sign_in):
+    app.config["MAIL_HOST"] = ""
+    sign_in(owner_id)
+    assert "no envía correo" in client.get(
+        "/comunidad/miembros/nuevo").get_data(as_text=True)
+
+    app.config["MAIL_HOST"] = "smtp.example.com"
+    assert "no envía correo" not in client.get(
+        "/comunidad/miembros/nuevo").get_data(as_text=True)

docs/server-setup.md

@@ -193,9 +193,16 @@ your laptop):
 cd ~/vienalatina
 git remote add gitea https://git.vienalatina.com/pablo/vienalatina.git
 git push gitea main
+
+# Name the GitHub remote too, while you are here. The clone in step 5 called it
+# `origin`, but `main` came to track `gitea/main` — so a bare `git pull` asks
+# Gitea, and new work arrives on GitHub. Having both named saves you from typing
+# the URL every time you deploy.
+git remote add github https://github.com/pablovolenski/vienalatina.git
+git remote -v
 ```
 
-(It will ask for your Gitea username/password.)
+(Gitea will ask for your Gitea username/password.)
 
 In Woodpecker (**https://ci.vienalatina.com**):
 
@@ -372,6 +379,51 @@ It cannot be used to take ownership later: that is deliberate, because otherwise
 editing a file on disk would be a quieter route to the top than asking for it.
 Ownership moves only through *Transferir titularidad* inside the app.
 
+#### Updating it later — a pull is not a deploy
+
+The app's code is **inside the image**: `docker/board/Dockerfile` ends with
+`COPY apps /srv/apps`. So pulling new commits into `~/vienalatina` changes
+nothing that is running, and neither does `docker compose up -d
+--force-recreate` — same image tag, same layers, same old code. Everything
+reports success and the server behaves exactly as it did before, which is the
+most expensive kind of nothing.
+
+**And `git pull` on its own will not fetch it.** This is worth knowing before
+it costs you an afternoon: `main` on the server tracks `gitea/main`, while new
+work is pushed to a branch on **GitHub**. So a bare `git pull` asks Gitea,
+finds Gitea level with your local `main`, and answers *"Already up to date."* —
+which is true about the wrong remote, and reads exactly like there is nothing
+to do.
+
+An update is a named pull, a push to Gitea so the build pipeline sees it, and a
+rebuild:
+
+```sh
+cd ~/vienalatina
+git pull --no-edit github <the-branch-name>   # e.g. claude/relaxed-faraday-h4zd09
+git push gitea main
+sudo bash scripts/deploy-board.sh
+```
+
+`--no-edit` accepts the default merge message. Without it git opens an editor,
+which is a strange place to find yourself mid-deploy.
+
+If `github` is not a remote yet, add it once — see the end of step 8:
+
+```sh
+git remote add github https://github.com/pablovolenski/vienalatina.git
+```
+
+That rebuilds the image, copies the compose file across, restarts, and prints
+the log. It never touches `/srv/board/.env` — that file holds the secrets and
+lives only on the server — but it does compare it against `.env.example` and
+name any setting that has appeared in the repository and is missing from yours.
+New settings are always added by hand.
+
+The database schema is applied at start-up with `CREATE TABLE IF NOT EXISTS`,
+so a release that adds a table needs no migration step: the table appears when
+the new code does.
+
 ### 11.4 Route it through Caddy
 
 Add to the `vienalatina.com` block in `/etc/caddy/Caddyfile` (already present in
@@ -622,9 +674,20 @@ perfectly in a mail client fails here, and the error it produces is a timeout
 rather than anything that names the cause.
 
 ```sh
-cd /srv/board && sudo docker compose up -d --force-recreate
+cd ~/vienalatina && sudo bash scripts/deploy-board.sh
 ```
 
+Not `docker compose up -d --force-recreate` on its own: if the code that reads
+these settings arrived in the same pull, that restart runs the old image and
+mail stays unconfigured with the settings sitting right there in `.env`. The
+script rebuilds first.
+
+The variables also have to be listed in `/srv/board/docker-compose.yml`, which
+they now are. Compose does not hand `.env` to a container — it substitutes into
+the compose file — so a setting added to `.env` and not to the compose file is
+read by nobody. `apps/board/tests/test_deployment.py` fails the build if the
+two ever drift apart again.
+
 Test it by adding a member with an address you can read. If the mail cannot be
 sent, the screen says so and shows you the invitation link to pass on by hand —
 the account is created either way, so a mail problem delays somebody rather

infra/board/docker-compose.yml

@@ -35,6 +35,23 @@ services:
       - GITEA_ADMIN_TOKEN=${GITEA_ADMIN_TOKEN:-}
 
       - BOARD_DB=/data/board.db
+
+      # Outgoing mail, for invitations and password resets. Empty MAIL_HOST is
+      # a supported state: the members area still works, but nobody can be
+      # invited or recover a password, and the screens say so.
+      #
+      # Every one of these has to be listed here. Compose does not hand the
+      # contents of .env to the container by itself — it only substitutes into
+      # this file — so a variable added to .env and not added here is read by
+      # nobody, and the app reports mail as unconfigured with the settings
+      # sitting right there on disk.
+      - MAIL_HOST=${MAIL_HOST:-}
+      - MAIL_PORT=${MAIL_PORT:-587}
+      - MAIL_SECURITY=${MAIL_SECURITY:-starttls}
+      - MAIL_USER=${MAIL_USER:-}
+      - MAIL_PASSWORD=${MAIL_PASSWORD:-}
+      - "MAIL_FROM=${MAIL_FROM:-Viena Latina <hola@vienalatina.com>}"
+
     volumes:
       - ./data:/data
     ports:

scripts/deploy-board.sh

@@ -0,0 +1,67 @@
+#!/usr/bin/env bash
+# Put the current checkout of the members area onto the server.
+#
+# The app's code is baked into the image by docker/board/Dockerfile (`COPY apps
+# /srv/apps`), so pulling new commits changes nothing on its own — the container
+# keeps running the code that was in the image when it was built. `docker
+# compose up -d --force-recreate` does not help either: same tag, same layers,
+# same old code. That is a quiet failure, because everything reports success and
+# the site behaves exactly as it did before.
+#
+# This script is the whole update, in the order that matters:
+#
+#   git pull                                  # done by you, first
+#   sudo bash scripts/deploy-board.sh
+#
+# It never touches /srv/board/.env. That file holds the secrets and exists only
+# on the server; the repository has .env.example instead, and the two drifting
+# apart is expected — new settings appear in the example and have to be copied
+# across by hand.
+
+set -euo pipefail
+
+REPO="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
+TARGET="${BOARD_DIR:-/srv/board}"
+IMAGE="vienalatina/board:1"
+
+cd "$REPO"
+
+echo "==> Building $IMAGE from $(git rev-parse --short HEAD)"
+docker build -t "$IMAGE" -f docker/board/Dockerfile .
+
+echo "==> Syncing compose files into $TARGET (never .env)"
+mkdir -p "$TARGET"
+cp "$REPO/infra/board/docker-compose.yml" "$TARGET/docker-compose.yml"
+cp "$REPO/infra/board/.env.example" "$TARGET/.env.example"
+
+if [ ! -f "$TARGET/.env" ]; then
+  echo "No $TARGET/.env — copy .env.example to .env and fill it in first." >&2
+  exit 1
+fi
+
+# Settings that appear in the example and not in the live file. Nothing is
+# copied automatically: some of them are secrets, and a blank line silently
+# added to .env is worse than a line missing loudly.
+missing="$(comm -23 \
+  <(grep -oE '^[A-Z][A-Z0-9_]*=' "$TARGET/.env.example" | sort -u) \
+  <(grep -oE '^[A-Z][A-Z0-9_]*=' "$TARGET/.env"         | sort -u) || true)"
+if [ -n "$missing" ]; then
+  echo
+  echo "!! These settings exist in .env.example but not in your .env:"
+  echo "$missing" | sed 's/^/     /'
+  echo "   Add them to $TARGET/.env and run this again if the feature needs them."
+  echo
+fi
+
+echo "==> Restarting"
+cd "$TARGET"
+docker compose up -d --force-recreate
+
+# The schema is applied at start-up with CREATE TABLE IF NOT EXISTS, so a new
+# table arrives with the new code. If the container is not up a few seconds
+# later it died during that, and the log says why.
+sleep 3
+docker compose ps
+echo
+echo "Recent log:"
+docker compose logs --tail 20 board