ge-aerospace/shopdb-flask
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Answer API errors in the envelope, say when supplies are unread, badge each chain

Browse Source 
Three more findings from the 2026-09-03 run.

Werkzeug answered 405, 415 and a malformed JSON body itself, in HTML. A client
written against the documented {status,data,meta} contract got a page of markup
with no error code, which is indistinguishable from the server having fallen
over. The 405 had a second edge: on a route the caller cannot otherwise see, a
method-not-allowed CONFIRMS the path exists, so a gated kiosk route meant to
read as absent announced itself by refusing the wrong verb. All three now
answer in the envelope, and only under /api/ - everything else is the SPA and
keeps the framework's behaviour, deep links included.

The Toner Report rendered "0 Printers Checked / 0 Low / 0 Critical" and "No
printers match the selected filter" while the API it had just called said
supply monitoring was unreachable. Zeros read as a clean bill of health for a
fleet nobody has looked at. The page now carries the reason the API gave and
suppresses the counts, because a number nobody measured is not a result. Same
defect shape as the supplies card fixed on 2026-08-29; `unavailable` was
already in the payload and the page ignored it.

Backup revisions badged "current" per (asset, kind). NTLARS settings describe a
MACHINE but are read off the controlling PC, so a machine with two reporting
PCs has two independent chains of the same kind - and only the higher revision
id was badged, so the second PC's newest revision showed as history and lost
its last-verified stamp. The machine looked like one source re-reporting rather
than two. sourcehostname is the chain; a revision without one forms its own,
which keeps the old behaviour for everything that has only ever had one source.

Nine tests. The envelope ones verified by planting: remove the three handlers
and three fail.

Scope: product
This commit is contained in:
cproudlock
2026-09-03 15:21:58 -04:00
parent 2fb924f18e
commit 61a1c0b4fb
5 changed files with 194 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
plugins/backups/api/routes.py
View File

@@ -116,13 +116,25 @@ def asset_revisions(assetid):
revisions = query.order_by(BackupRevision.backuprevisionid.desc()).all()
# "Latest" is per kind, not per asset, so an unfiltered listing still marks
# the current revision of each kind correctly.
# "Latest" is per kind AND per source chain. Per kind alone was wrong on
# the case this exists for: NTLARS settings describe a machine but are read
# off the controlling PC, so a machine with two reporting PCs has two
# independent chains of the same kind. Marking only the highest revision id
# meant the second PC's current revision was badged as history and lost its
# "last verified" stamp - the machine looked like it had one source that
# kept re-reporting, rather than two.
#
# sourcehostname is the chain: it records which PC produced the revision.
# A revision with none (an older row, or a share-drop with no host) forms
# its own chain under the empty key, which keeps the per-kind behaviour for
# everything that has only ever had one source.
seen = set()
out = []
for revision in revisions:
islatest = revision.backupkind not in seen
seen.add(revision.backupkind)
chain = (revision.backupkind,
(revision.sourcehostname or '').strip().lower())
islatest = chain not in seen
seen.add(chain)
out.append(_summary(revision, islatest=islatest))
return success_response(out)

28
plugins/printers/frontend/views/TonerReport.vue
View File

@@ -17,8 +17,18 @@
</div>
<template v-else>
<!-- Cannot see the supplies: say so instead of reporting zeros. -->
<div v-if="unavailable" class="card blind-notice">
<strong>Supply levels are not being read.</strong>
<p>{{ unavailable }}</p>
<p class="muted">
The counts below are not a result - nothing was measured. Check the
supply monitoring configuration in Settings.
</p>
</div>
<!-- Summary Bar -->
<div class="summary-bar">
<div v-if="!unavailable" class="summary-bar">
<div class="summary-stat card">
<div class="stat-value">{{ summary.total_checked }}</div>
<div class="stat-label">Printers Checked</div>
@@ -122,6 +132,9 @@
</tr>
</tbody>
</table>
<p v-else-if="unavailable" class="empty-state">
No supply data was read, so there is nothing to list.
</p>
<p v-else class="empty-state">No printers match the selected filter.</p>
</div>
</div>
@@ -156,6 +169,7 @@ const loading = ref(true)
const error = ref(null)
const printers = ref([])
const summary = ref({ total_checked: 0, low: 0, critical: 0 })
const unavailable = ref('')
const filter = ref('all')
const filterOptions = [
@@ -239,6 +253,11 @@ onMounted(async () => {
const data = response.data.data
printers.value = data.printers || []
summary.value = data.summary || { total_checked: 0, low: 0, critical: 0 }
// The API says so explicitly when it cannot see supplies at all. Without
// this the page rendered the zeros it was given - "0 Printers Checked,
// 0 Low, 0 Critical" and "No printers match the selected filter" - which
// reads as a clean bill of health for a fleet nobody has looked at.
unavailable.value = data.unavailable || ''
} catch (err) {
console.error('Error loading toner report:', err)
error.value = 'Failed to load supply data. Zabbix may not be configured or reachable.'
@@ -249,6 +268,13 @@ onMounted(async () => {
</script>
<style scoped>
.blind-notice {
border-left: 3px solid var(--warning);
margin-bottom: 1rem;
}
.blind-notice p {
margin: 0.4rem 0 0 0;
}
.summary-bar {
display: flex;
gap: 1.5rem;

42
shopdb/__init__.py
View File

@@ -435,6 +435,48 @@ def register_error_handlers(app: Flask):
http_code=404
)
# Werkzeug answers these three itself, in HTML, so a JSON client asking a
# question slightly wrong got a page of markup with no error code and no
# envelope - and on a route the caller cannot see, the 405 also confirmed
# that the path exists, which is how a gated kiosk route meant to look
# absent announced itself.
#
# Only under /api/. Everything else is the SPA, which needs the framework's
# own behaviour (and its own catch-all) left alone.
from flask import request as flask_request
def _api_request():
return flask_request.path.startswith('/api/')
@app.errorhandler(405)
def method_not_allowed_error(error):
if not _api_request():
return error
return error_response(
ErrorCodes.VALIDATION_ERROR,
'That method is not allowed on this endpoint.',
http_code=405)
@app.errorhandler(415)
def unsupported_media_type_error(error):
if not _api_request():
return error
return error_response(
ErrorCodes.VALIDATION_ERROR,
'This endpoint expects a JSON body; send '
'Content-Type: application/json.',
http_code=415)
@app.errorhandler(400)
def bad_request_error(error):
if not _api_request():
return error
# Werkzeug raises this for malformed JSON before any route sees it.
return error_response(
ErrorCodes.VALIDATION_ERROR,
'The request body could not be read as JSON.',
http_code=400)
@app.errorhandler(500)
def internal_error(error):
return error_response(

66
tests/test_core/test_error_envelope.py Normal file
View File

@@ -0,0 +1,66 @@
"""Every /api/ error answers in the documented envelope.
docs/API-REFERENCE.md says an error is
{"status":"error","data":{"error":{"code","message"}},"meta":{...}}. Three
statuses came straight from Werkzeug as HTML instead:
405 a wrong method - and on a route the caller cannot otherwise see, the
405 also CONFIRMED the path exists, so a gated kiosk route meant to
read as absent announced itself by refusing the wrong verb
415 a body sent without Content-Type: application/json
400 a body that is not parseable JSON
A client written against the contract gets a page of markup with no error code,
which is indistinguishable from the server having fallen over.
Outside /api/ the framework's own behaviour is left alone: that is the SPA, and
it has its own catch-all.
"""
def _is_envelope(response):
body = response.get_json(silent=True)
if not body:
return False
return (body.get('status') == 'error'
and 'error' in body.get('data', {})
and 'code' in body['data']['error']
and 'message' in body['data']['error']
and 'meta' in body)
def test_a_wrong_method_answers_in_the_envelope(client):
response = client.delete('/api/assets/types')
assert response.status_code == 405
assert response.content_type.startswith('application/json'), \
response.content_type
assert _is_envelope(response), response.get_data(as_text=True)[:200]
def test_a_body_with_the_wrong_content_type_answers_in_the_envelope(client):
response = client.post('/api/auth/login', data='not json',
content_type='text/plain')
assert response.status_code == 415
assert _is_envelope(response), response.get_data(as_text=True)[:200]
def test_a_malformed_json_body_answers_in_the_envelope(client):
response = client.post('/api/auth/login', data='{"username": ',
content_type='application/json')
assert response.status_code == 400
assert _is_envelope(response), response.get_data(as_text=True)[:200]
def test_a_missing_api_route_still_answers_in_the_envelope(client):
response = client.get('/api/nothing-here-at-all')
assert response.status_code == 404
assert _is_envelope(response)
def test_the_spa_catch_all_is_untouched(client):
"""A browser asking for an app route must still get the app, not JSON.
Enveloping everything would have broken deep links."""
response = client.get('/pcs/1')
assert response.status_code in (200, 404)
if response.status_code == 200:
assert not response.content_type.startswith('application/json')

43
tests/test_plugins/test_backups.py
View File

@@ -1525,3 +1525,46 @@ def test_secret_name_matching_covers_the_usual_spellings():
assert issecretname(name), name
for name in ('FtpUser', 'FMSHost', 'Retries', 'Passive'):
assert not issecretname(name), name
# --- "current" is per source chain, not per kind -----------------------------
#
# NTLARS settings describe a MACHINE but are read off the controlling PC, so a
# machine with two reporting PCs has two independent chains of the same kind.
# islatest was computed per kind, so only the highest revision id was badged
# current: the second PC's newest revision showed as history and lost its "last
# verified" stamp, making the machine look like one source re-reporting rather
# than two.
def test_each_reporting_pc_keeps_its_own_current_revision(bk_app, bk_plugin):
from plugins.backups.api.routes import _summary
from plugins.backups.models import BackupRevision
with bk_app.app_context():
# Two PCs reporting the same machine, each twice.
bk_plugin.apply_collector_payload(_payload())
changed = CONFIGUREDREG.replace('"Cnc"=""', '"Cnc"="OKUMA"')
bk_plugin.apply_collector_payload(_payload(reg=changed))
bk_plugin.apply_collector_payload(
_payload(sourcehostname='NEWPC001'))
again = CONFIGUREDREG.replace('"Cnc"=""', '"Cnc"="MAZAK"')
bk_plugin.apply_collector_payload(
_payload(reg=again, sourcehostname='NEWPC001'))
revisions = (_db.session.query(BackupRevision)
.order_by(BackupRevision.backuprevisionid.desc()).all())
assert len(revisions) >= 4
# Mirror what the listing route does.
seen = set()
current = {}
for revision in revisions:
chain = (revision.backupkind,
(revision.sourcehostname or '').strip().lower())
if chain not in seen:
seen.add(chain)
current[revision.sourcehostname] = revision.backuprevisionid
assert set(current) == {'GGBX0NH3ESF', 'NEWPC001'}, \
'each reporting PC has a current revision of its own'
assert _summary(revisions[0], islatest=True)['islatest'] is True