Upgrade LangChain4j to 1.15.1-TB1 and migrate GitHub Models to the
official OpenAI integration (OpenAiOfficialChatModel), enabling JSON
Schema (strict) structured output for GitHub Models and exposing JSON
Schema as a response format for Anthropic and Amazon Bedrock.
- Rename model capability flags to supportsSchemalessJsonOutput() and
supportsJsonSchemaOutput(), declared explicitly per provider
- Push response-format capability checks into TbResponseFormat and
simplify response-format handling in TbAiNode
- Remove obsolete manual JSON escaping for GitHub Models, which now
double-escaped requests under the official OpenAI SDK
- Drop the now-obsolete opennlp-tools version pin (the upgraded
LangChain4j fork ships opennlp-tools >= 2.5.9)
- Update the AI rule node UI to offer response formats per provider,
refresh available model lists, and migrate toggle options to @if
- Expand configurer tests to verify per-provider model configuration
- Provide a scoped Angular ErrorHandler on the dynamic component's
injector in HtmlContainerWidgetComponent so template-runtime
exceptions raised inside user-authored Angular templates flow
through handleWidgetException instead of crashing the global
Angular ErrorHandler.
- Add `ctx.invokeAction($event, actionName, additionalParams?)` to
WidgetActionsApi / WidgetComponent and a `WidgetDestroyCallback` +
ctx.registerDestroyCallback() API on WidgetContext (callbacks run
in registration order on destroy, errors per-callback are caught
so one bad cleanup doesn't break the rest).
- Surface widget actions on the HTML Container basic config: render
tb-widget-actions-panel with the new strokedPanel input below
the html-container settings, and register a default "JavaScript"
multi-action source on the html_container widget JSON.
- Expand the widget-completion docs for `registerDestroyCallback`
(when it fires, what to use it for, lifecycle semantics, exact
callback signature `() => void`) and for `invokeAction`'s
`additionalParams` arg (forwarded to the configured JS action
handler, common payload examples).
Netty 4.1.133.Final introduced a regression in MqttDecoder while fixing
CVE-2026-44248: when multiple MQTT packets are present in the same
cumulation buffer, the per-message size check used the total buffer size
instead of the current packet's declared remaining length. Valid in-limit
packets get rejected with TooLongFrameException("message length exceeds
65536: <small number>"). Fixed upstream by netty/netty#16787 and ported
to 4.1 as netty/netty@30f8f284db, released in 4.1.134.Final.
Exposes the dashboard entities-limit check as a reusable public method
so callers outside the validator can enforce the same limit without
duplicating the apiLimitService logic.
Dmytro Skarzhynets
Igor Kulikov
Viacheslav Klimov
dashevchenko
Vladyslav Prykhodko
Andrii Landiak
Maksym Tsymbarov