* feat: add missing OpenAI/Claude/Gemini request fields and responses stream options
* fix: skip field filtering when request passthrough is enabled
* fix: include subscription in personal sidebar module controls
* feat: gate Claude inference_geo passthrough behind channel setting and add field docs
- Introduced a new function to collect invalid status code entries from the status code mapping.
- Updated the EditChannelModal to display an error message if invalid status codes are detected.
- Enhanced localization files to include new error messages for invalid status codes in multiple languages.
- Removed unused styles from the RiskAcknowledgementModal for cleaner UI.
- Added TaskTimeoutMinutes constant to configure the timeout duration for asynchronous tasks.
- Implemented sweepTimedOutTasks function to identify and handle unfinished tasks that exceed the timeout limit, marking them as failed and processing refunds if applicable.
- Enhanced task polling loop to include the new timeout handling logic, ensuring timely cleanup of stale tasks.
1. Async task model redirection (aligned with sync tasks):
- Integrate ModelMappedHelper in RelayTaskSubmit after model name
determination, populating OriginModelName / UpstreamModelName on RelayInfo.
- All task adaptors now send UpstreamModelName to upstream providers:
- Gemini & Vertex: BuildRequestURL uses UpstreamModelName.
- Doubao & Ali: BuildRequestBody conditionally overwrites body.Model.
- Vidu, Kling, Hailuo, Jimeng: convertToRequestPayload accepts RelayInfo
and unconditionally uses info.UpstreamModelName.
- Sora: BuildRequestBody parses JSON and multipart bodies to replace
the "model" field with UpstreamModelName.
- Frontend log visibility: LogTaskConsumption and taskBillingOther now
emit is_model_mapped / upstream_model_name in the "other" JSON field.
- Billing safety: RecalculateTaskQuotaByTokens reads model name from
BillingContext.OriginModelName (via taskModelName) instead of
task.Data["model"], preventing billing leaks from upstream model names.
2. Per-call billing (TaskPricePatches lifecycle):
- Rename TaskBillingContext.ModelName → OriginModelName; add PerCallBilling
bool field, populated from TaskPricePatches at submission time.
- settleTaskBillingOnComplete short-circuits when PerCallBilling is true,
skipping both adaptor adjustments and token-based recalculation.
- Remove ModelName from TaskSubmitResult; use relayInfo.OriginModelName
consistently in controller/relay.go for billing context and logging.
3. Multipart retry boundary mismatch fix:
- Root cause: after Sora (or OpenAI audio) rebuilds a multipart body with a
new boundary and overwrites c.Request.Header["Content-Type"], subsequent
calls to ParseMultipartFormReusable on retry would parse the cached
original body with the wrong boundary, causing "NextPart: EOF".
- Fix: ParseMultipartFormReusable now caches the original Content-Type in
gin context key "_original_multipart_ct" on first call and reuses it for
all subsequent parses, making multipart parsing retry-safe globally.
- Sora adaptor reverted to the standard pattern (direct header set/get),
which is now safe thanks to the root fix.
4. Tests:
- task_billing_test.go: update makeTask to use OriginModelName; add
PerCallBilling settlement tests (skip adaptor adjust, skip token recalc);
add non-per-call adaptor adjustment test with refund verification.
- Updated UpdateWithStatus method to use Model().Select("*").Updates() for conditional updates, preventing GORM's INSERT fallback.
- Introduced comprehensive integration tests for UpdateWithStatus, covering scenarios for winning and losing CAS updates, as well as concurrent updates.
- Added task_cas_test.go to validate the new behavior and ensure data integrity during concurrent state transitions.
Replace all bare task.Update() (DB.Save) calls with UpdateWithStatus(),
which adds a WHERE status = ? guard to prevent concurrent processes from
overwriting each other's state transitions.
Key changes:
model/task.go:
- Add taskSnapshot struct with Equal() method for change detection
- Add Snapshot() method to capture pre-update state
- Add UpdateWithStatus(fromStatus) using DB.Where().Save() for CAS
semantics with full-struct save (no explicit field listing needed)
model/midjourney.go:
- Add UpdateWithStatus(fromStatus string) with same CAS pattern
service/task_polling.go (updateVideoSingleTask):
- Snapshot before processing upstream response; skip DB write if unchanged
- Terminal transitions (SUCCESS/FAILURE) use UpdateWithStatus CAS:
billing/refund only executes if this process wins the transition
- Non-terminal updates also use UpdateWithStatus to prevent overwriting
a concurrent terminal transition back to IN_PROGRESS
- Defer settleTaskBillingOnComplete to after CAS check (shouldSettle flag)
relay/relay_task.go (tryRealtimeFetch):
- Add snapshot + change detection; use UpdateWithStatus for CAS safety
controller/midjourney.go (UpdateMidjourneyTaskBulk):
- Capture preStatus before mutations; use UpdateWithStatus CAS
- Gate refund (IncreaseUserQuota) on CAS success (won && shouldReturnQuota)
This prevents the multi-instance race condition where:
1. Instance A reads task (IN_PROGRESS), fetches upstream (still IN_PROGRESS)
2. Instance B reads same task, fetches upstream (now SUCCESS), writes SUCCESS
3. Instance A's bare Save() overwrites SUCCESS back to IN_PROGRESS
- Enhanced the RelayTask function to utilize a locked channel when available, allowing for better reuse during retries.
- Updated error handling to ensure proper context setup for the selected channel.
- Clarified comments in ResolveOriginTask regarding channel locking and retry behavior.
- Introduced a new field in TaskRelayInfo to store the locked channel object, improving type safety and reducing import cycles.
- Updated the remix handling in ResolveOriginTask to prioritize extracting OtherRatios from the BillingContext of the original task if available.
- Retained the previous logic for extracting seconds and size from task data as a fallback.
- Improved clarity and maintainability of the remix logic by separating the new and old approaches.
- Renamed RelayTask function to RelayTaskFetch for clarity.
- Updated routing in relay-router.go and video-router.go to use RelayTaskFetch for fetch operations.
- Enhanced error handling in RelayTaskFetch function.
- Adjusted task data conversion in TaskAdaptor to include task ID.
Add three billing lifecycle methods to the TaskAdaptor interface:
- EstimateBilling: compute OtherRatios from user request before pricing
- AdjustBillingOnSubmit: adjust ratios from upstream submit response
- AdjustBillingOnComplete: determine final quota at task terminal state
Introduce BaseBilling as embeddable no-op default for adaptors without
custom billing. Move Sora/Ali OtherRatios logic from shared validation
into per-adaptor EstimateBilling implementations.
Add TaskBillingContext to persist pricing params (model_price, group_ratio,
other_ratios) in task private data for async polling settlement.
Extract RecalculateTaskQuota as a general-purpose delta settlement
function and unify polling billing via settleTaskBillingOnComplete
(adaptor-first, then token-based fallback).
Restructure the task relay system for better separation of concerns:
- Extract task billing into service/task_billing.go with unified settlement flow
- Move task polling loop from controller to service/task_polling.go (supports Suno + video platforms)
- Split RelayTask into fetch/submit paths with dedicated retry logic (taskSubmitWithRetry)
- Add TaskDto, TaskResponse generics, and FetchReq to dto/task.go
- Add taskcommon/helpers.go for shared task adaptor utilities
- Remove controller/task_video.go (logic consolidated into service layer)
- Update all task adaptors (ali, doubao, gemini, hailuo, jimeng, kling, sora, suno, vertex, vidu)
- Simplify frontend task logs to use new TaskDto response format
Custom OAuth providers redirect to /oauth/{slug} after authorization,
but only hardcoded provider routes (github, discord, oidc, linuxdo)
existed in the frontend router, causing a 404 for custom providers.
Clients like OpenClaw send input_text content blocks (a Responses API
type) through /v1/messages. The Claude-to-OpenAI converter silently
drops unknown types, so the message arrives empty at the upstream,
causing "Invalid value: 'input_text'" errors.
Map input_text to text since they share the same structure.
