Stop the audit log from being what fails the write

Wave 4, second half. PUT /api/machines/<id> returned 500 and rolled the whole
edit back for any machine with a maintenance date recorded - including a rename
that never touched the date. The Machine edit form could not save such a
machine at all, which is most of them once a maintenance schedule exists.

The route builds {'old': getattr(row, field)} straight off the model, so the
changes dict carries whatever the column holds: a date, a datetime, a Decimal.
AuditLog.changes serialised it with a bare json.dumps, which raises on all
three, and nothing caught it. The audit entry and the edit share a transaction,
so the failure took the edit with it.

json.dumps now has a fallback encoder: dates and datetimes become ISO strings,
Decimals become floats, sets and bytes get something readable, and anything
else becomes its repr rather than an exception. An audit entry that reads a
little awkwardly is worth more than a failed write - the alternative, and what
happened, is losing the change AND the record of it.

Fixed at the audit log rather than in the machines route: every audited route
builds its changes dict the same way, so the next one to audit a date field
would have hit the same wall.

Four tests, including the exact shape the machines route produces. Verified by
planting the bare dumps back: two fail.

Scope: product

tests/test_core/test_strict_write_errors.py

@@ -97,3 +97,44 @@ def test_the_strict_sql_mode_is_actually_set_on_the_mysql_connection():
 
     connect_args = Config.SQLALCHEMY_ENGINE_OPTIONS['connect_args']
     assert 'STRICT_TRANS_TABLES' in connect_args['init_command']
+
+
+# --- the audit log must never be what fails the write ------------------------
+
+def test_audit_changes_serialise_the_types_a_model_actually_holds():
+    """A route builds {'old': getattr(row, field)} straight off the model, so
+    the changes dict carries dates, datetimes and Decimals. json.dumps raises
+    on all three, and the exception took the whole request down: PUT
+    /api/machines/<id> 500d and rolled back for any machine with a maintenance
+    date, including a rename that never touched it."""
+    import datetime
+    import decimal
+
+    from shopdb.core.models.auditlog import AuditLog
+
+    entry = AuditLog()
+    entry.changes = {
+        'nextmaintenancedate': {'old': datetime.date(2026, 12, 15),
+                                'new': '2027-01-05'},
+        'lastseen': {'old': datetime.datetime(2026, 1, 2, 3, 4, 5),
+                     'new': None},
+        'cost': {'old': decimal.Decimal('12.50'), 'new': 13},
+    }
+    stored = entry.changes
+    assert stored['nextmaintenancedate']['old'] == '2026-12-15'
+    assert stored['lastseen']['old'] == '2026-01-02T03:04:05'
+    assert stored['cost']['old'] == 12.5
+
+
+def test_an_unserialisable_value_is_recorded_rather_than_raising():
+    """An audit entry that reads awkwardly beats a failed write: the
+    alternative is losing the change AND the record of it."""
+    from shopdb.core.models.auditlog import AuditLog
+
+    class Odd:
+        def __repr__(self):
+            return '<Odd value>'
+
+    entry = AuditLog()
+    entry.changes = {'thing': {'old': Odd(), 'new': None}}
+    assert entry.changes['thing']['old'] == '<Odd value>'

tests/test_plugins/test_measuringtools.py

@@ -553,3 +553,36 @@ def test_a_mistyped_date_is_refused_at_create_too(client, auth_headers):
     listed = client.get('/api/measuringtools?search=MT-DATE-3',
                         headers=auth_headers).get_json()['data']
     assert listed == [], 'nothing may be created by a refused request'
+
+
+# --- an audited field holding a date must not kill the request ---------------
+#
+# PUT /api/machines/<id> returned 500 and rolled the whole edit back for any
+# machine with a maintenance date: the route builds
+# {'old': getattr(row, field)} straight off the model, and json.dumps raises on
+# a date. A rename that never touched the date failed too. The audit log
+# refusing to record a change must not be what stops the change.
+
+def test_an_edit_is_saved_even_when_an_audited_field_holds_a_date(
+        client, auth_headers):
+    created = client.post('/api/measuringtools', headers=auth_headers, json={
+        'assetnumber': 'MT-AUDIT-1',
+        'name': 'Before rename',
+        'statusid': _status_id(client),
+        'lastcalibrationdate': '2026-01-15',
+        'nextcalibrationdate': '2027-01-15',
+    })
+    assert created.status_code == 201, created.get_json()
+    toolid = created.get_json()['data']['measuringtool']['measuringtoolid']
+
+    # A rename that does not touch the dates. The date is only in the CHANGES
+    # dict because the interval moves with it.
+    renamed = client.put(f'/api/measuringtools/{toolid}', headers=auth_headers,
+                         json={'name': 'After rename',
+                               'nextcalibrationdate': '2028-01-15'})
+    assert renamed.status_code == 200, renamed.get_json()
+
+    after = client.get(f'/api/measuringtools/{toolid}',
+                       headers=auth_headers).get_json()['data']
+    assert after['name'] == 'After rename'
+    assert after['measuringtool']['nextcalibrationdate'] == '2028-01-15'