diff options
| author | Christophe Besson <cbesson@gmail.com> | 2026-09-07 21:35:57 +0200 |
|---|---|---|
| committer | Christophe Besson <cbesson@gmail.com> | 2026-09-07 21:35:57 +0200 |
| commit | f7b1111c250902315e454c155c147e58d69ee9f0 (patch) | |
| tree | c2bfa30464b13b6c5da65c3d702042f2b8f0c491 /packages/meshbay-node/src/meshbay_node | |
| parent | 3bd31db9d4fa2352f1095dcc630c77915d0774b8 (diff) | |
| download | meshbay-f7b1111c250902315e454c155c147e58d69ee9f0.tar.gz | |
fix(chat): the operator was missing from the roster they host
Found on two live machines within minutes of deploying: every message from the
person running the node arrived at every other member under "this account is
using a key you have not seen before".
An operator's authority is node-wide and is recorded in `members` with an
**empty** group_id — `is_authorized` has always said so, in a clause written for
exactly that. `group_devices` spelled the rule out a second time as
`WHERE m.group_id = ?`, which excludes them. So the operator was absent from the
roster relayed to members, no chain could reach their device key, and Tier 2
reported the most ordinary event there is — the operator talking in their own
group — as a key substitution.
A notice that fires on normal use is worse than no notice: it is the one people
learn to dismiss, and §4.8 budgets exactly one for the whole feature. That makes
this a defect in the property, not only in a query.
The clause now lives once, as `_MEMBER_OF_GROUP`, shared by both callers so they
cannot drift again. `DISTINCT` because an operator who is also an explicit member
of the group matches both halves of it. `get_member` is untouched: it is a raw
lookup and its callers already fall back to `get_member("", user_id)` themselves.
Three tests, and the first fails against the old query with the reported
symptom: the operator appears in the roster of a group they host and
`is_authorized` agrees; an operator who is also a member is listed once; a
revoked one comes back through neither.
No stored state to clean up — nothing was written to a client's pins when the
account was missing, so the notice stops as soon as the node serves the roster
correctly.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TZZxYjz8YeWRz13xDi8LJr
Diffstat (limited to 'packages/meshbay-node/src/meshbay_node')
| -rw-r--r-- | packages/meshbay-node/src/meshbay_node/roster.py | 39 |
1 files changed, 30 insertions, 9 deletions
diff --git a/packages/meshbay-node/src/meshbay_node/roster.py b/packages/meshbay-node/src/meshbay_node/roster.py index 2288ae4..3d3a143 100644 --- a/packages/meshbay-node/src/meshbay_node/roster.py +++ b/packages/meshbay-node/src/meshbay_node/roster.py @@ -418,6 +418,16 @@ class Roster: - it is scoped to *this* group. A person in two groups on one node is not disclosed to the second by being in the first. + "Member of this group" is `_MEMBER_OF_GROUP`, shared with + `is_authorized` — **the operator belongs to every group this node + hosts**, with their authority recorded under an empty group_id. Spelling + that out a second time here is exactly what went wrong first time: + the operator was missing from the roster, so every one of their messages + reached other members as a key nobody could vouch for. + + `DISTINCT` because an operator who is *also* an explicit member of the + group matches both halves of that clause. + `add_sig`/`add_nonce`/`add_ts` are empty for a device pinned before the evidence was kept, and for the first device of any account — which an operator code admitted, not a countersignature. Both read as @@ -426,13 +436,12 @@ class Roster: """ assert self._db async with self._db.execute( - "SELECT i.user_id, i.username, i.pk_ed25519, i.pk_x25519, " - " i.added_by_pk, i.add_sig, i.add_nonce, i.add_ts, i.pinned_at " - "FROM identities i " - "JOIN members m ON m.user_id = i.user_id " - "WHERE m.group_id = ? AND m.status = 'active' " - " AND i.revoked_at IS NULL " - "ORDER BY i.user_id, i.pinned_at", (group_id,) + f"SELECT DISTINCT i.user_id, i.username, i.pk_ed25519, i.pk_x25519, " + f" i.added_by_pk, i.add_sig, i.add_nonce, i.add_ts, i.pinned_at " + f"FROM identities i " + f"JOIN members m ON m.user_id = i.user_id " + f"WHERE i.revoked_at IS NULL AND m.{self._MEMBER_OF_GROUP} " + f"ORDER BY i.user_id, i.pinned_at", (group_id,) ) as cur: rows = await cur.fetchall() return [ @@ -546,6 +555,18 @@ class Roster: async def has_operator(self) -> bool: return bool(await self.operator_pks()) + # Who counts as a member of a group, in SQL, in **one** place. + # + # The operator's authority is node-wide and is stored with an empty + # group_id, so "belongs to this group" is not `group_id = ?`. Writing that + # clause a second time is how `group_devices` came to omit the operator + # from the roster it relays — and the symptom was a member seeing "this + # account is using a key you have not seen before" on every single message + # from the person running the node. Found on a live pair of machines, not + # by a test. + _MEMBER_OF_GROUP = ("status = 'active' AND " + "(group_id = ? OR (group_id = '' AND role = 'operator'))") + async def is_authorized(self, group_id: str, user_id: str) -> bool: """ May this person be handed the group key? @@ -560,8 +581,8 @@ class Roster: """ assert self._db async with self._db.execute( - "SELECT 1 FROM members WHERE user_id = ? AND status = 'active' " - "AND (group_id = ? OR (group_id = '' AND role = 'operator')) LIMIT 1", + f"SELECT 1 FROM members WHERE user_id = ? AND {self._MEMBER_OF_GROUP} " + f"LIMIT 1", (user_id, group_id), ) as cur: return await cur.fetchone() is not None |