- Refactor update/delete endpoints to use installed item's own ID instead of marketplace itemId
- Add SHA-256 checksum comparison to detect local entity modifications before update
- Support force update to skip checksum check when user confirms overwrite
- Add per-type update methods (widget, dashboard, calculated field, rule chain)
- Add entityId field to CalculatedFieldInstalledItemDescriptor
- Replace checkForUpdates with getItemsPublishedVersions API
- Replace getInstalledItemInfos with getInstalledItemIds returning List<UUID>
- Remove unused IotHubInstalledItemInfo class and resolver
- Set metadata version from saved rule chain in install/update flows
- Add entity-modified confirmation dialog in update UI
- Full entity stack: IotHubInstalledItem with ID, descriptors, JPA entity, DAO (extends JpaAbstractDao), service
- Controller endpoints: install, list (paginated), list infos, get by itemId, delete
- Install flow saves descriptor with entity IDs, delete cascades to installed entities
- Frontend installed items page with search, pagination, type chips, action buttons
- Browse page shows "Installed" label on cards, detail dialog shows "Open details" for installed items
- Installed item infos fetched on browse page init to track installation state
In Jackson 2.18.x, EXISTING_PROPERTY type info combined with the no-arg
@JsonIgnoreProperties causes the triggerType discriminator field to be
silently excluded from the serialized JSON. When the server then tries to
deserialize the POST body for /api/notification/rule, Jackson cannot find
triggerType and throws "missing type id property 'triggerType'", resulting
in a 500 for NotificationEdgeTest.testNotificationRule.
Fix by:
1. Adding @JsonProperty("triggerType") to force the field into normal bean
serialization, overriding any suppression by the type info machinery.
2. Replacing the no-arg @JsonIgnoreProperties with @JsonIgnoreProperties(
ignoreUnknown = true) so unknown properties are ignored rather than
causing errors (e.g. for forward compatibility).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Using the same connection for both SCAN cursor iteration and GET value
fetches causes Jedis 5.x response-ordering corruption: the SCAN response
parser receives a GET response (byte[]) where it expects a List, and vice
versa, resulting in ClassCastException on startup.
Fix: open two connections per getAll() call — one dedicated to the scan
cursor and one for value fetches — eliminating any interleaving.
Affected: TbRedisLwM2MClientStore, TbRedisLwM2MModelConfigStore,
TbLwM2mRedisRegistrationStore.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Tests that read shared static state (e.g. testAllowedUrls with 8.8.8.8)
could run concurrently with tests that mutate it (e.g. testAdditionalBlockedSingleIp),
causing intermittent failures. Class-level @ResourceLock serializes all tests.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Move TestDbCallbackExecutor from rule-engine test sources to
common/util main sources as DirectListeningExecutor, making it
available to all modules. Convert to an enum singleton since the
executor is stateless. Widen JpaAbstractDaoListeningExecutorService
service field type from JpaExecutorService to ListeningExecutor to
allow injecting DirectListeningExecutor in tests. Fix
AbstractChunkedAggregationTimeseriesDaoTest NPE by injecting the
direct executor into the spy.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
When a TbRuleChainInputNode has `forwardMsgToDefaultRuleChain=true` and the originator's
default rule chain is the same as the rule chain containing this node, the message enters
an infinite loop: the node forwards to the default rule chain, which routes back to the
same node, which forwards again, causing unbounded recursion and 100% CPU on rule-engine.
Fix: detect the loop in DefaultTbContext.input() by checking whether the calling rule node
is already present in the message's return stack (TbMsgProcessingCtx). On the second+
iteration the stack already contains the (ruleChainId, ruleNodeId) pair of the node,
so the call is a cycle. In that case tellFailure() is called with a descriptive message
and a WARN log is emitted instead of re-enqueuing the message.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Delay and deduplication rule nodes were creating brand new TbMsg objects
instead of copying the original, which reset the ruleNodeExecCounter to 0.
This allowed bypassing the maxRuleNodeExecutionsPerMessage limit.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Resolve the commit once in onUpdate() and reuse the cached RevCommit
for listFiles and getFileContent operations, instead of resolving
the branch ref on every call. Added RevCommit-accepting overloads
to GitRepository for listFilesAtCommit and getFileContentAtCommit.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Adds page-level byte-size tracking in TbResultSet.allRows() using
ExecutionInfo.getResponseSizeInBytes() to fail early when accumulated
result set size exceeds the configurable limit (default 50MB).
Also fixes pre-existing bugs where onFailure callbacks in
CassandraBaseTimeseriesDao only logged errors but never completed
the future, causing callers to hang indefinitely on failures.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Increased timing tolerance gap from 500ms to 1000ms in both
testRateLimitWithGreedyRefill and testRateLimitWithIntervalRefill
to prevent ConditionTimeoutException caused by scheduling jitter.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Igor Kulikov
Viacheslav Klimov
Sergey Matvienko
dashevchenko
Dmytro Skarzhynets
Volodymyr Babak
Nikita Mazurenko
Andrii Landiak
nickAS21
IrynaMatveieva
Artem Barysh