htmx:responseError and hx-sync queue behavior for shared swap target

Answer first

With hx-sync="queue" on requests that share the same swap target, an htmx:responseError from one trigger does not reset the queue. The failed request is treated as finished and the next queued request is allowed to start. The default sync mode is parallel, so when no explicit hx-sync is set on a new trigger there is no cross-source abort; new requests run concurrently with in-flight requests to the same target.

Confirmed facts

• hx-sync="queue" serializes requests to the same target. The next request only starts after the previous request completes or fails.
• htmx:responseError is emitted for non-2xx responses. The event is informational; htmx does not cancel further processing based on it.
• In queue mode a response error is considered completion, so the queue proceeds to the next pending request.
• Default hx-sync is parallel. No built-in response versioning or ordering guarantees are provided.
• hx-sync="abort" cancels any currently active request to the target before starting the new one. This must be declared on the triggering element.

Likely explanation, not a specification guarantee

htmx tracks sync state per sync key, which is derived from the target selector and the hx-sync value on the initiating element. An error event does not clear the sync state or pause the queue. The library does not provide automatic cross-source abort. If different source elements use different hx-sync values, or no hx-sync at all, they will not coordinate with each other unless they share the same sync key and mode.

Minimal steps for this case

1. Confirm the sync key is identical for all sources. Check that each source uses the same target selector and the same hx-sync value, e.g. hx-sync="[id=panel] queue".
2. Verify queue behavior in devtools Network tab. Trigger two requests rapidly; requests should be sequential, not parallel.
3. Force a 500 response for the first request and observe via the htmx event log whether htmx:responseError fires and whether the second queued request is dispatched after the error.
4. If you need abort-on-new-trigger across sources, set hx-sync="[id=panel] abort" on the elements that should preempt in-flight work, or set it on a common parent so it applies consistently.

Uncertainty note: exact queue release timing on network errors vs HTTP error responses can vary with htmx version and with custom event handling that calls htmx:abort. Current verification is recommended.

One diagnostic detail that changes the recommendation: are all sources declaring hx-sync on themselves, or is hx-sync declared only on the shared target element? The sync mode is read from the initiating element, not inherited from the target, so placement determines whether cross-source coordination actually applies.