Dust off `make_full_schema` and add CI using it to show schema diffs. by reivilibre · Pull Request #20027 · element-hq/synapse

reivilibre · 2026-07-31T16:32:08Z

It seems a lot of time in our trial tests goes towards setting up the database. (The same is probably true of Complement too)

We haven't done a full schema for about 20 schema versions, so no surprise!

As a result, I want to produce a full schema soon.

In this PR I dust off make_full_schema.sh (which seems to have broken after some SQLite changes)
and add a CI workflow that runs it (producing a diff) when someone changes the schema.

The CI workflow also adds a sticky comment showing the diff on the schema, so you can better appreciate the final effect of a change.

Original commit schedule, with full messages:

make_full_schema: Use --nosys to avoid dumping sqlite_sequence
make_full_schema: don't rely on pip
Not present on my machine
make_full_schema: tweak example to use new minimum postgres version
~~make_full_schema: no longer drop sqlite internal tables~~

Moved to Remove broken DROP statements for SQLite in make_full_schema script. #20028
CI: add schema diff 'sticky comment'
make_full_schema: improve documentation

Dead changes:

I wanted to make it possible to generate a versioned full schema without the manual work,
but you can't run the background updates without essentially starting up a homeserver,
at which point it might fail because you haven't run all the deltas yet.

There's no actual good way to do this, short of deleting the latest deltas (+ tweaking code to not crash without them)
or rolling back in the git history.

Backed out these changes, but they're preserved here.

update_synapse_database: allow updating to an arbitrary schema version
make_full_schema: don't update past the requested schema version
Revert "update_synapse_database: allow updating to an arbitrary schema version"
This reverts commit b3fbb8144ace9c933e013d881f7bd92a02d21745.
Revert "make_full_schema: don't update past the requested schema version"
This reverts commit 70cceba43d0131e7276182725454f01a5b1684b1.

github-actions · 2026-07-31T16:54:21Z

Schema Diff

Please check that this looks as expected!

common (sqlite)

Unchanged

common (postgres)

--- /tmp/schema_diff_fb4wkuc9/before/common/full_schemas/9999/full.sql.postgres	2026-08-11 10:31:33.463625039 +0000
+++ /tmp/schema_diff_fb4wkuc9/after/common/full_schemas/9999/full.sql.postgres	2026-08-11 10:31:23.422978143 +0000
@@ -1,12 +1,8 @@
-\restrict hycDXNaQTcWwg2zqH4FcLG9rBhV4Ygg65vUDfQLBogFe1Xa1SEpewRYdLy4YUul
 CREATE TABLE background_updates (
     update_name text NOT NULL,
     progress_json text NOT NULL,
     depends_on text,
     ordering integer DEFAULT 0 NOT NULL
 );
 ALTER TABLE ONLY background_updates
     ADD CONSTRAINT background_updates_uniqueness UNIQUE (update_name);
-\unrestrict hycDXNaQTcWwg2zqH4FcLG9rBhV4Ygg65vUDfQLBogFe1Xa1SEpewRYdLy4YUul
-\restrict ONJJNiONvZYwI63NakJ0mdnmi5JcTLLN9A7UaapE9aBdVb7xyjMMxfUXUg5YvSb
-\unrestrict ONJJNiONvZYwI63NakJ0mdnmi5JcTLLN9A7UaapE9aBdVb7xyjMMxfUXUg5YvSb

main (sqlite)

--- /tmp/schema_diff_fb4wkuc9/before/main/full_schemas/9999/full.sql.sqlite	2026-08-11 10:31:33.291631088 +0000
+++ /tmp/schema_diff_fb4wkuc9/after/main/full_schemas/9999/full.sql.sqlite	2026-08-11 10:31:23.282983061 +0000
@@ -972,21 +972,20 @@
 );
 CREATE UNIQUE INDEX sliding_sync_membership_snapshots_event_stream_ordering ON sliding_sync_membership_snapshots(event_stream_ordering);
 CREATE TABLE sliding_sync_connections(
     connection_key INTEGER PRIMARY KEY AUTOINCREMENT,
     user_id TEXT NOT NULL,
     -- Generally the device ID, but may be something else for e.g. puppeted accounts.
     effective_device_id TEXT NOT NULL,
     conn_id TEXT NOT NULL,
     created_ts BIGINT NOT NULL
 , last_used_ts BIGINT);
-CREATE TABLE sqlite_sequence(name,seq);
 CREATE INDEX sliding_sync_connections_idx ON sliding_sync_connections(user_id, effective_device_id, conn_id);
 CREATE INDEX sliding_sync_connections_ts_idx ON sliding_sync_connections(created_ts);
 CREATE TABLE sliding_sync_connection_positions(
     connection_position INTEGER PRIMARY KEY AUTOINCREMENT,
     connection_key BIGINT NOT NULL REFERENCES sliding_sync_connections(connection_key) ON DELETE CASCADE,
     created_ts BIGINT NOT NULL
 );
 CREATE INDEX sliding_sync_connection_positions_key ON sliding_sync_connection_positions(connection_key);
 CREATE INDEX sliding_sync_connection_positions_ts_idx ON sliding_sync_connection_positions(created_ts);
 CREATE TABLE sliding_sync_connection_required_state(
@@ -1238,13 +1237,13 @@
 /* WARNING: Script requires that SQLITE_DBCONFIG_DEFENSIVE be disabled */
 INSERT INTO appservice_stream_position VALUES('X',0);
 INSERT INTO federation_stream_position VALUES('federation',-1,'master');
 INSERT INTO federation_stream_position VALUES('events',-1,'master');
 INSERT INTO event_push_summary_stream_ordering VALUES('X',0);
 INSERT INTO user_directory_stream_pos VALUES('X',-1);
 INSERT INTO stats_incremental_position VALUES('X',1);
 INSERT INTO event_push_summary_last_receipt_stream_id VALUES('X',0);
 INSERT INTO device_lists_changes_converted_stream_position VALUES('X',1,'',NULL);
 INSERT INTO room_forgetter_stream_pos VALUES('X',0);
-INSERT INTO scheduled_tasks VALUES('delete_old_otks_task','delete_old_otks','scheduled',1786444289000,NULL,NULL,NULL,NULL);
+INSERT INTO scheduled_tasks VALUES('delete_old_otks_task','delete_old_otks','scheduled',1786444279000,NULL,NULL,NULL,NULL);
 INSERT INTO delayed_events_stream_pos VALUES('X',0);
 INSERT INTO device_lists_changes_in_room_max_pruned_stream_id VALUES('X',0);

main (postgres)

--- /tmp/schema_diff_fb4wkuc9/before/main/full_schemas/9999/full.sql.postgres	2026-08-11 10:31:33.832612060 +0000
+++ /tmp/schema_diff_fb4wkuc9/after/main/full_schemas/9999/full.sql.postgres	2026-08-11 10:31:23.867962514 +0000
@@ -1,11 +1,10 @@
-\restrict KlVdCFGPDN7Tl5bUxj9VjzkaYOE3oMUUj4nFerNDZxkiALAPgfwWLZbi34N1pu0
 CREATE FUNCTION check_event_stream_ordering() RETURNS trigger
     LANGUAGE plpgsql
     AS $$
             BEGIN
                 IF EXISTS (
                     SELECT 1 FROM events
                     WHERE events.event_id = NEW.event_id
                        AND events.stream_ordering != NEW.event_stream_ordering
                 ) THEN
                     RAISE EXCEPTION 'Incorrect event_stream_ordering';
@@ -1938,22 +1937,20 @@
 ALTER TABLE ONLY un_partial_stated_event_stream
     ADD CONSTRAINT un_partial_stated_event_stream_event_id_fkey FOREIGN KEY (event_id) REFERENCES events(event_id) ON DELETE CASCADE;
 ALTER TABLE ONLY un_partial_stated_room_stream
     ADD CONSTRAINT un_partial_stated_room_stream_room_id_fkey FOREIGN KEY (room_id) REFERENCES rooms(room_id) ON DELETE CASCADE;
 ALTER TABLE ONLY users_to_send_full_presence_to
     ADD CONSTRAINT users_to_send_full_presence_to_user_id_fkey FOREIGN KEY (user_id) REFERENCES users(name);
 ALTER TABLE ONLY worker_read_write_locks
     ADD CONSTRAINT worker_read_write_locks_lock_name_lock_key_write_lock_fkey FOREIGN KEY (lock_name, lock_key, write_lock) REFERENCES worker_read_write_locks_mode(lock_name, lock_key, write_lock);
 ALTER TABLE ONLY worker_read_write_locks_mode
     ADD CONSTRAINT worker_read_write_locks_mode_foreign FOREIGN KEY (lock_name, lock_key, token) REFERENCES worker_read_write_locks(lock_name, lock_key, token) DEFERRABLE INITIALLY DEFERRED;
-\unrestrict KlVdCFGPDN7Tl5bUxj9VjzkaYOE3oMUUj4nFerNDZxkiALAPgfwWLZbi34N1pu0
-\restrict 8e7NoW4Yo8Oxpo1j8DiQOGeYFfo6w6PRdjyESr3OeEz4O1naGLaWXJchjJ2IEnP
 INSERT INTO appservice_stream_position VALUES ('X', 0);
 INSERT INTO delayed_events_stream_pos VALUES ('X', 0);
 INSERT INTO device_lists_changes_converted_stream_position VALUES ('X', 1, '', NULL);
 INSERT INTO device_lists_changes_in_room_max_pruned_stream_id VALUES ('X', 0);
 INSERT INTO event_push_summary_last_receipt_stream_id VALUES ('X', 0);
 INSERT INTO event_push_summary_stream_ordering VALUES ('X', 0);
 INSERT INTO federation_stream_position VALUES ('federation', -1, 'master');
 INSERT INTO federation_stream_position VALUES ('events', -1, 'master');
 INSERT INTO room_forgetter_stream_pos VALUES ('X', 0);
 INSERT INTO scheduled_tasks VALUES ('delete_old_otks_task', 'delete_old_otks', 'scheduled', 990794562000, NULL, NULL, NULL, NULL);
@@ -1976,11 +1973,10 @@
 SELECT pg_catalog.setval('quarantined_media_id_seq', 1, true);
 SELECT pg_catalog.setval('receipts_sequence', 1, true);
 SELECT pg_catalog.setval('sliding_sync_connection_positions_connection_position_seq', 1, false);
 SELECT pg_catalog.setval('sliding_sync_connection_required_state_required_state_id_seq', 1, false);
 SELECT pg_catalog.setval('sliding_sync_connections_connection_key_seq', 1, false);
 SELECT pg_catalog.setval('sticky_events_sequence', 1, true);
 SELECT pg_catalog.setval('thread_subscriptions_sequence', 2, true);
 SELECT pg_catalog.setval('un_partial_stated_event_stream_sequence', 1, true);
 SELECT pg_catalog.setval('un_partial_stated_room_stream_sequence', 1, true);
 SELECT pg_catalog.setval('user_id_seq', 1, false);
-\unrestrict 8e7NoW4Yo8Oxpo1j8DiQOGeYFfo6w6PRdjyESr3OeEz4O1naGLaWXJchjJ2IEnP

state (sqlite)

--- /tmp/schema_diff_fb4wkuc9/before/state/full_schemas/9999/full.sql.sqlite	2026-08-11 10:31:33.293358852 +0000
+++ /tmp/schema_diff_fb4wkuc9/after/state/full_schemas/9999/full.sql.sqlite	2026-08-11 10:31:23.286052638 +0000
@@ -16,19 +16,18 @@
 );
 CREATE INDEX state_group_edges_prev_idx ON state_group_edges (prev_state_group);
 CREATE INDEX state_groups_state_type_idx ON state_groups_state (state_group, type, state_key);
 CREATE INDEX state_groups_room_id_idx ON state_groups (room_id) ;
 CREATE UNIQUE INDEX state_group_edges_unique_idx ON state_group_edges (state_group, prev_state_group) ;
 CREATE TABLE state_groups_pending_deletion (
     sequence_number INTEGER PRIMARY KEY AUTOINCREMENT,
     state_group BIGINT NOT NULL,
     insertion_ts BIGINT NOT NULL
 );
-CREATE TABLE sqlite_sequence(name,seq);
 CREATE UNIQUE INDEX state_groups_pending_deletion_state_group ON state_groups_pending_deletion(state_group);
 CREATE INDEX state_groups_pending_deletion_insertion_ts ON state_groups_pending_deletion(insertion_ts);
 CREATE TABLE state_groups_persisting (
     state_group BIGINT NOT NULL,
     instance_name TEXT NOT NULL,
     PRIMARY KEY (state_group, instance_name)
 );
 CREATE INDEX state_groups_persisting_instance_name ON state_groups_persisting(instance_name);

state (postgres)

--- /tmp/schema_diff_fb4wkuc9/before/state/full_schemas/9999/full.sql.postgres	2026-08-11 10:31:34.009605834 +0000
+++ /tmp/schema_diff_fb4wkuc9/after/state/full_schemas/9999/full.sql.postgres	2026-08-11 10:31:24.045956263 +0000
@@ -1,11 +1,10 @@
-\restrict NY9nMB6Pz6E8GlHW1faNoOjNx5aoA0LG0npd8dlsalJSjD0PLPeZgM90nIfuemd
 CREATE TABLE state_group_edges (
     state_group bigint NOT NULL,
     prev_state_group bigint NOT NULL
 );
 CREATE SEQUENCE state_group_id_seq
     START WITH 1
     INCREMENT BY 1
     NO MINVALUE
     NO MAXVALUE
     CACHE 1;
@@ -45,15 +44,12 @@
     ADD CONSTRAINT state_groups_persisting_pkey PRIMARY KEY (state_group, instance_name);
 ALTER TABLE ONLY state_groups
     ADD CONSTRAINT state_groups_pkey PRIMARY KEY (id);
 CREATE INDEX state_group_edges_prev_idx ON state_group_edges USING btree (prev_state_group);
 CREATE UNIQUE INDEX state_group_edges_unique_idx ON state_group_edges USING btree (state_group, prev_state_group);
 CREATE INDEX state_groups_pending_deletion_insertion_ts ON state_groups_pending_deletion USING btree (insertion_ts);
 CREATE UNIQUE INDEX state_groups_pending_deletion_state_group ON state_groups_pending_deletion USING btree (state_group);
 CREATE INDEX state_groups_persisting_instance_name ON state_groups_persisting USING btree (instance_name);
 CREATE INDEX state_groups_room_id_idx ON state_groups USING btree (room_id);
 CREATE INDEX state_groups_state_type_idx ON state_groups_state USING btree (state_group, type, state_key);
-\unrestrict NY9nMB6Pz6E8GlHW1faNoOjNx5aoA0LG0npd8dlsalJSjD0PLPeZgM90nIfuemd
-\restrict z3RYkophnxWGphTiuMvOOXYe47YYBsq25yfpwAUixRz3flP7UYo3bmD9YUjm0ye
 SELECT pg_catalog.setval('state_group_id_seq', 1, false);
 SELECT pg_catalog.setval('state_groups_pending_deletion_sequence_number_seq', 1, false);
-\unrestrict z3RYkophnxWGphTiuMvOOXYe47YYBsq25yfpwAUixRz3flP7UYo3bmD9YUjm0ye

…pt. (#20028) Broken out of #20027 because I'd like to have it land first. > It seems a lot of time in our trial tests goes towards setting up the database. (The same is probably true of Complement too) > > We haven't done a full schema for about 20 schema versions, so no surprise! > > As a result, I want to produce a full schema soon. When running `make_full_schema.sh`, these drop statements now cause the error: ``` Parse error near line 2: table event_search_content may not be dropped Parse error near line 3: table event_search_segments may not be dropped Parse error near line 4: table event_search_segdir may not be dropped Parse error near line 5: table event_search_docsize may not be dropped Parse error near line 6: table event_search_stat may not be dropped Parse error near line 7: table user_directory_search_content may not be dropped Parse error near line 8: table user_directory_search_segments may not be dropped Parse error near line 9: table user_directory_search_segdir may not be dropped Parse error near line 10: table user_directory_search_docsize may not be dropped Parse error near line 11: table user_directory_search_stat may not be dropped ``` It seems SQLite has cracked down on code that edits the internal tables. Because SQLite dumps the schema with `CREATE TABLE IF NOT EXISTS` for these virtual tables, it's harmless to leave them in the schema dump. --------- Signed-off-by: Olivier 'reivilibre' <oliverw@matrix.org>

Not present on my machine

…a version" This reverts commit b3fbb8144ace9c933e013d881f7bd92a02d21745.

…ion" This reverts commit 70cceba43d0131e7276182725454f01a5b1684b1.

Signed-off-by: Olivier 'reivilibre <oliverw@matrix.org>

This reverts commit 22adfd4.

reivilibre · 2026-08-06T14:29:24Z

@@ -0,0 +1,225 @@
+#!/usr/bin/env python3


This was mostly LLM'd with a few minor tweaks. Since it's not production code, I don't think it's too problematic.

It looks 'fine' to me, I'm sure it could be better-structured, but other than the few tweaks I have made I'm not sure there's anything I'm rushing to tweak.

MadLittleMods · 2026-08-06T20:46:16Z

-pg_dump --format=plain --data-only --inserts --no-tablespaces --no-acl --no-owner "$POSTGRES_MAIN_DB_NAME"   | cleanup_pg_schema >> "$OUTPUT_DIR/main/full_schemas/$SCHEMA_NUMBER/full.sql.postgres"
-pg_dump --format=plain --schema-only         --no-tablespaces --no-acl --no-owner "$POSTGRES_STATE_DB_NAME"  | cleanup_pg_schema  > "$OUTPUT_DIR/state/full_schemas/$SCHEMA_NUMBER/full.sql.postgres"
-pg_dump --format=plain --data-only --inserts --no-tablespaces --no-acl --no-owner "$POSTGRES_STATE_DB_NAME"  | cleanup_pg_schema >> "$OUTPUT_DIR/state/full_schemas/$SCHEMA_NUMBER/full.sql.postgres"
+pg_dump --restrict-key=REMOVEME --format=plain --schema-only         --no-tablespaces --no-acl --no-owner "$POSTGRES_COMMON_DB_NAME" | cleanup_pg_schema  > "$OUTPUT_DIR/common/full_schemas/$SCHEMA_NUMBER/full.sql.postgres"


Why new --restrict-key=REMOVEME flag necessary?

We set this as normally it's a random string but we want a consistent predictable string to match the find/replace in cleanup_pg_schema above. We should comment about that that detail.

Previous versions of Postgres didn't have this \restrict/\unrestrict mechanism, it was added for https://www.postgresql.org/support/security/CVE-2025-8714/

MadLittleMods · 2026-08-06T20:50:32Z

+  # - Remove the `\restrict` and `\unrestrict` psql meta-commands.
+  #   We don't run the pg_dump output through psql so no meta-commands are
+  #   supported and so the security feature (which doesn't apply anyway
+  #   as we trust the source database) is not relevant.


Why do these even come into play?

I don't see \restrict/\unrestrict anywhere

Explanation for myself: pg_dump will output \restrict mysecret/\unrestrict mysecret commands. It's security feature to avoid poisoned data running arbitrary commands. Any psql meta commands after \restrict mysecret does not work until you \unrestrict mysecret

Why do we care about disabling this? Seems like it couldn't hurt.

It's a psql meta-command only interpreted by psql and not (AFAIK) supported by Postgres itself or psycopg2. (This is sort of what the comment at this line is trying to say, but if you missed it then it might be worth a reword)

I think the comment is sufficient but I wasn't following on the first pass 👍

MadLittleMods · 2026-08-06T21:06:06Z

+  contents: read
+
+jobs:
+  # Posts a GitHub PR comment that shows what the effective change to the schema is.


Why do we care about having this?

Two things really:

an excuse to run make_full_schema regularly so we hopefully spot it breaking

as a review aid for schema changes — particularly when we make schema changes in a background update, or more weird ones where we have to do a complex rewrite for SQLite (example) — it's nice to see the diff of the end result and have another opportunity to spot a problem

Lol, I didn't intend for this, but this schema diff review aid would also have spotted that the following schema file does nothing:

https://github.com/element-hq/synapse/blob/release-v1.157/synapse/storage/schema/main/delta/78/03_remove_unused_indexes_user_filters.py

MadLittleMods · 2026-08-06T21:07:19Z

+        with:
+          header: schema-diff
+          path: ${{ runner.temp }}/schema_diff.md
+          ignore_empty: true


ignore_empty: true prevents it from posting a comment for PRs that don't change the database at all?

Naw, it's redundant here as it shouldn't be empty. Removed

prevents it from posting a comment for PRs that don't change the database at all?

Do we always post comments? (even for PRs which don't change any database stuff)

(seems annoying if so)

Only when you touch a schema delta, full schema, or the schema diff workflow: https://github.com/element-hq/synapse/pull/20027/files/8ab5389c0bc1c6eae241ebb61667b93da149a03c#diff-589b31c74e9c8e8b7c6808f05a82bcd2a1ca18a00320e396afb36bd507cb1411R6

I didn't want to waste CI time otherwise.
I realise that it will miss cases where you touch the code concerning a background update that could have an overall impact on the schema, but at the moment I think I'm fine with that (to avoid triggering it unnecessarily), I think that case is sufficiently rare (if you're fixing a background update, you're also likely to put in a schema delta to re-trigger it or something, to fix up servers that already ran the broken one).

MadLittleMods · 2026-08-06T21:08:34Z

@@ -0,0 +1,225 @@
+#!/usr/bin/env python3


(didn't review)

MadLittleMods · 2026-08-10T16:32:08Z

-
-NB at the time of writing, this script predates the split into separate `state`/`main`
-databases so will require updates to handle that correctly.
+Run `poetry run make_full_schema.sh -h` for usage instructions.


I'd prefer a snippet I can run:

Suggested change

Run `poetry run make_full_schema.sh -h` for usage instructions.

Run `poetry run scripts-dev/make_full_schema.sh -h` for usage instructions.

Good principle to live by

MadLittleMods · 2026-08-10T16:32:55Z

-NB at the time of writing, this script predates the split into separate `state`/`main`
-databases so will require updates to handle that correctly.


Were these updates made so this is no longer relevant?

The comment was stale: this was addressed in matrix-org/synapse#13792

MadLittleMods · 2026-08-10T16:41:45Z

+  # - Remove the `\restrict` and `\unrestrict` psql meta-commands.
+  #   We don't run the pg_dump output through psql so no meta-commands are
+  #   supported and so the security feature (which doesn't apply anyway
+  #   as we trust the source database) is not relevant.


I think the comment is sufficient but I wasn't following on the first pass 👍

reivilibre · 2026-08-11T10:29:43Z

Cheers for the review @MadLittleMods 🐑

Issue spotted in: #20098 Follows: #20027 We already use `faketime` for Postgres, but I forgot that the SQLite schema delta would have the same problem and somehow tuned it out of the preview diff on the original PR. --------- Signed-off-by: Olivier 'reivilibre <oliverw@matrix.org>

…0117) Follows: #20027 Noticed in #20003 The problem was that `--no-root` prevents reinstalling the root project (Synapse). However, since we just did a `git checkout`, we need to reinstall the root project in case the Rust code changed, as the `poetry install` command is what causes the Rust to be recompiled. Doing otherwise causes the 'Rust module outdated' error when importing `synapse`. --------- Signed-off-by: Olivier 'reivilibre <oliverw@matrix.org>

reivilibre mentioned this pull request Jul 31, 2026

Remove broken DROP statements for SQLite in make_full_schema script. #20028

Merged

reivilibre force-pushed the rei/full_schema branch from 2440b34 to 838a876 Compare July 31, 2026 17:05

reivilibre added 3 commits August 6, 2026 10:17

make_full_schema: Use --nosys to avoid dumping sqlite_sequence

5b22c3d

make_full_schema: don't rely on pip

27271f3

Not present on my machine

make_full_schema: tweak example to use new minimum postgres version

e37d487

reivilibre force-pushed the rei/full_schema branch 4 times, most recently from 73e3899 to 115db87 Compare August 6, 2026 10:09

reivilibre and others added 9 commits August 6, 2026 11:17

CI: add schema diff 'sticky comment'

3186073

make_full_schema: improve documentation

ecc249d

update_synapse_database: allow updating to an arbitrary schema version

6b3e076

make_full_schema: don't update past the requested schema version

f27255d

Revert "update_synapse_database: allow updating to an arbitrary schem…

2fd7060

…a version" This reverts commit b3fbb8144ace9c933e013d881f7bd92a02d21745.

Revert "make_full_schema: don't update past the requested schema vers…

95c1e52

…ion" This reverts commit 70cceba43d0131e7276182725454f01a5b1684b1.

Newsfile

f7cf1ed

Signed-off-by: Olivier 'reivilibre <oliverw@matrix.org>

make_full_schema: remove \(un)restrict psql meta-commands

b7235d9

fixup! CI: add schema diff 'sticky comment'

871b9cb

reivilibre force-pushed the rei/full_schema branch 2 times, most recently from df240e9 to b153f65 Compare August 6, 2026 13:44

CI: Use faketime on Python for schema stability

22adfd4

reivilibre force-pushed the rei/full_schema branch from b153f65 to 7a9bd79 Compare August 6, 2026 13:55

reivilibre added 2 commits August 6, 2026 15:15

CI: Use a faketime Postgres database

d029c42

Revert "CI: Use faketime on Python for schema stability"

8ab5389

This reverts commit 22adfd4.

reivilibre force-pushed the rei/full_schema branch from 7a9bd79 to 8ab5389 Compare August 6, 2026 14:15

reivilibre commented Aug 6, 2026

View reviewed changes

reivilibre marked this pull request as ready for review August 6, 2026 14:31

reivilibre requested a review from a team as a code owner August 6, 2026 14:31

MadLittleMods added A-Database A-Testing tests labels Aug 6, 2026

MadLittleMods reviewed Aug 6, 2026

View reviewed changes

reivilibre added 6 commits August 10, 2026 11:37

Comment restrict-key

a4d9bc8

Comment caller is to use poetry run

4b0b8a2

ignore_empty is unnecessary

457c885

Comment why we do the comment

0fcd622

Signpost to poetry run

f9fdfa4

Reword comment: if you want to build..

2cb1487

reivilibre requested a review from MadLittleMods August 10, 2026 10:51

MadLittleMods approved these changes Aug 10, 2026

View reviewed changes

reivilibre added 3 commits August 11, 2026 11:22

Add instruction to Schema Diff comment

9444ecc

Expand --nosys comment

8cf84c5

Cosmetic tweaks

6465da9

reivilibre merged commit d80a4e6 into develop Aug 11, 2026
46 checks passed

reivilibre deleted the rei/full_schema branch August 11, 2026 11:56

reivilibre mentioned this pull request Aug 13, 2026

Fix the schema diff CI not using faketime for SQLite. #20099

Merged

This was referenced Aug 13, 2026

Add CI that checks that old schemas are upgradable and give the same final schema. #20102

Draft

Fix the schema diff CI breaking when the Rust module was changed. #20117

Merged

	Run `poetry run make_full_schema.sh -h` for usage instructions.
	Run `poetry run scripts-dev/make_full_schema.sh -h` for usage instructions.

		NB at the time of writing, this script predates the split into separate `state`/`main`
		databases so will require updates to handle that correctly.

Conversation

reivilibre commented Jul 31, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions Bot commented Jul 31, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Schema Diff

common (sqlite)

common (postgres)

main (sqlite)

main (postgres)

state (sqlite)

state (postgres)

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

reivilibre Aug 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

reivilibre Aug 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

reivilibre Aug 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

MadLittleMods Aug 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

reivilibre commented Aug 11, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

reivilibre commented Jul 31, 2026 •

edited

Loading

github-actions Bot commented Jul 31, 2026 •

edited

Loading

reivilibre Aug 10, 2026 •

edited

Loading

reivilibre Aug 10, 2026 •

edited

Loading

reivilibre Aug 10, 2026 •

edited

Loading

MadLittleMods Aug 10, 2026 •

edited

Loading