[IMP] runbot_merge: split staging heads out to join tables

Currently the heads of a staging (both staging heads and merged heads)
are just JSON data on the staging itself. Historically this was
convenient as the heads were mostly of use to the staging process, and
thus accessed directly through the staging essentially exclusively.

However this makes finding stagings from merged commits e.g. for
forensic research almost impossible, because querying based on
the *values* of a JSON map is expensive, and indexing it is difficult.

To make this use case more feasible, split the `heads` field into two
join tables, one for the staging heads and one for the merged heads,
this makes looking for stagings by commits much more
efficient (although the queries may not be trivial). Also add two
utility RPC methods, so that it's possible to query stagings
reasonably easily and efficiently based on a set of commits (branch
heads).

related to #768

runbot_merge/__manifest__.py

@@ -1,6 +1,6 @@
 {
     'name': 'merge bot',
-    'version': '1.7',
+    'version': '1.8',
     'depends': ['contacts', 'website'],
     'data': [
         'security/security.xml',

runbot_merge/migrations/15.0.1.8/pre-migration.py

@@ -0,0 +1,6 @@
+from pathlib import Path
+
+def migrate(cr, version):
+    sql = Path(__file__).parent.joinpath('upgrade.sql')\
+        .read_text(encoding='utf-8')
+    cr.execute(sql)

runbot_merge/migrations/15.0.1.8/upgrade.sql

@@ -0,0 +1,62 @@
+CREATE TABLE runbot_merge_stagings_commits (
+    id serial NOT NULL,
+    staging_id integer not null references runbot_merge_stagings (id),
+    commit_id integer not null references runbot_merge_commit (id),
+    repository_id integer not null references runbot_merge_repository (id)
+);
+
+CREATE TABLE runbot_merge_stagings_heads (
+    id serial NOT NULL,
+    staging_id integer NOT NULL REFERENCES runbot_merge_stagings (id),
+    commit_id integer NOT NULL REFERENCES runbot_merge_commit (id),
+    repository_id integer NOT NULL REFERENCES runbot_merge_repository (id)
+);
+
+-- some of the older stagings only have the head, not the commit,
+-- add the commit
+UPDATE runbot_merge_stagings
+  SET heads = heads::jsonb || jsonb_build_object(
+    'odoo/odoo^', heads::json->'odoo/odoo',
+    'odoo/enterprise^', heads::json->'odoo/enterprise'
+  )
+  WHERE heads NOT ILIKE '%^%';
+
+-- some of the stagings have heads which don't exist in the commits table,
+-- because they never got a status from the runbot...
+-- create fake commits so we don't lose heads
+INSERT INTO runbot_merge_commit (sha, statuses, create_uid, create_date, write_uid, write_date)
+    SELECT r.value, '{}', s.create_uid, s.create_date, s.create_uid, s.create_date
+    FROM runbot_merge_stagings s,
+         json_each_text(s.heads::json) r
+ON CONFLICT DO NOTHING;
+
+CREATE TEMPORARY TABLE staging_commits (
+    id integer NOT NULL,
+    repo integer NOT NULL,
+    -- the staging head (may be a dedup, may be the same as commit)
+    head integer NOT NULL,
+    -- the staged commit
+    commit integer NOT NULL
+);
+-- the splatting works entirely off of the staged head
+-- (the one without the ^ suffix), we concat the `^` to get the corresponding
+-- merge head (the actual commit to push to the branch)
+INSERT INTO staging_commits (id, repo, head, commit)
+    SELECT s.id, re.id AS repo, h.id AS head, c.id AS commit
+    FROM runbot_merge_stagings s,
+         json_each_text(s.heads::json) r,
+         runbot_merge_commit h,
+         runbot_merge_commit c,
+         runbot_merge_repository re
+    WHERE r.key NOT ILIKE '%^'
+      AND re.name = r.key
+      AND h.sha = r.value
+      AND c.sha = s.heads::json->>(r.key || '^');
+
+INSERT INTO runbot_merge_stagings_heads (staging_id, repository_id, commit_id)
+SELECT id, repo, head FROM staging_commits;
+
+INSERT INTO runbot_merge_stagings_commits (staging_id, repository_id, commit_id)
+SELECT id, repo, commit FROM staging_commits;
+
+ALTER TABLE runbot_merge_stagings DROP COLUMN heads;