Recover LINE auth errors at runtime#188
Conversation
|
Warning Review limit reached
More reviews will be available in 30 minutes and 8 seconds. Learn how PR review limits work. To continue reviewing without waiting, enable usage-based billing in the billing tab. ⌛ How to resolve this issue?After more reviews become available, a review can be triggered using the To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based credits. 🚦 How do rate limits work?CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability. For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window. Please see our Fair Usage Limits Policy for further information. ℹ️ Review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (1)
📝 WalkthroughWalkthroughAdds LINE auth classification, shared retry-on-recovery call helpers, and recovery state on ChangesLINE Auth Recovery and Client Reuse
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
There was a problem hiding this comment.
Pull request overview
Note
Copilot was unable to run its full agentic suite in this review.
This PR introduces centralized token/auth recovery for LINE API calls and adds structured helpers to classify auth-related errors, then refactors connector operations to use the new recovery path.
Changes:
- Added
line.IsRefreshRequired,line.IsLoggedOut,line.IsUnauthorizedStatus, andline.IsAuthErrorplus unit tests. - Implemented generic
callLineWithRecovery+LineClient.callLinewrappers and refactored multiple connector handlers to use them. - Added concurrency controls and a short “recent recovery” window to avoid repeated token recovery attempts.
Reviewed changes
Copilot reviewed 10 out of 11 changed files in this pull request and generated 4 comments.
Show a summary per file
| File | Description |
|---|---|
| pkg/line/errors.go | Adds helper predicates to detect auth/unauthorized/refresh-required errors. |
| pkg/line/errors_test.go | Adds unit tests for new error classification helpers. |
| pkg/connector/auth_recovery.go | Introduces generic LINE call wrapper with token recovery + retry. |
| pkg/connector/auth_recovery_test.go | Adds tests for retry/recovery behavior and concurrency serialization. |
| pkg/connector/client.go | Routes refresh detection through line helpers and serializes recovery with mutex/window. |
| pkg/connector/send_message.go | Refactors message sending/uploads to use recovery wrapper instead of direct client calls. |
| pkg/connector/userinfo.go | Uses lc.callLine for read receipts. |
| pkg/connector/reaction.go | Uses lc.callLine for reactions and cancellations. |
| pkg/connector/handlers/handler.go | Switches recovery trigger from substring "401" to line.IsUnauthorizedStatus. |
| pkg/connector/handlers/file.go | Retries OBS download after token recovery. |
| .gitignore | Stops ignoring *_test.go so tests can be committed. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (2)
pkg/line/errors.go (1)
22-24: ⚡ Quick winHarden auth classifiers against case-variant error strings.
These predicates drive token-recovery retry; case-sensitive matching can miss recoverable auth errors if upstream text casing drifts.
Suggested patch
func IsRefreshRequired(err error) bool { if err == nil { return false } - msg := err.Error() + msg := strings.ToLower(err.Error()) return strings.Contains(msg, "\"code\":119") || - strings.Contains(msg, "Access token refresh required") + strings.Contains(msg, "access token refresh required") } func IsLoggedOut(err error) bool { if err == nil { return false } - return strings.Contains(err.Error(), "V3_TOKEN_CLIENT_LOGGED_OUT") + return strings.Contains(strings.ToLower(err.Error()), "v3_token_client_logged_out") } func IsUnauthorizedStatus(err error) bool { if err == nil { return false } - msg := err.Error() - return strings.Contains(msg, "API error 401") || - strings.Contains(msg, "API error 403") || - strings.Contains(msg, "HTTP 401") || - strings.Contains(msg, "HTTP 403") || - strings.Contains(msg, "SSE error: 401") || - strings.Contains(msg, "SSE error: 403") || - strings.Contains(msg, "OBS upload failed (401)") || - strings.Contains(msg, "OBS upload failed (403)") || - strings.Contains(msg, "OBS download failed (401)") || - strings.Contains(msg, "OBS download failed (403)") + msg := strings.ToLower(err.Error()) + return strings.Contains(msg, "api error 401") || + strings.Contains(msg, "api error 403") || + strings.Contains(msg, "http 401") || + strings.Contains(msg, "http 403") || + strings.Contains(msg, "sse error: 401") || + strings.Contains(msg, "sse error: 403") || + strings.Contains(msg, "obs upload failed (401)") || + strings.Contains(msg, "obs upload failed (403)") || + strings.Contains(msg, "obs download failed (401)") || + strings.Contains(msg, "obs download failed (403)") }Also applies to: 31-31, 38-48
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@pkg/line/errors.go` around lines 22 - 24, The error string matching in this function is case-sensitive, which can cause recoverable auth errors to be missed if the upstream error messages have different casing. Convert the error message to lowercase using strings.ToLower() before performing the strings.Contains() checks against both the "code":119 pattern and the "Access token refresh required" string. Apply the same case-insensitive matching fix to all other similar error classification predicates mentioned in the file at the referenced line ranges.pkg/connector/auth_recovery.go (1)
24-27: ⚡ Quick winPreserve the triggering auth error when recovery fails.
The returned error currently drops the original LINE call failure context, which makes incident triage harder.
Suggested patch
- return client, zero, fmt.Errorf("failed to recover token after LINE auth error: %w", errRecover) + return client, zero, fmt.Errorf("LINE auth call failed (%v); token recovery failed: %w", err, errRecover)🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@pkg/connector/auth_recovery.go` around lines 24 - 27, The error returned when token recovery fails only preserves the recovery error from errRecover but loses the original LINE authentication error that triggered the recovery attempt. Modify the recovery error handling block to capture the original authentication error before calling deps.recover(ctx), and then when errRecover is not nil, wrap both errors together in the returned error using fmt.Errorf or error wrapping to ensure both the original LINE auth error and the recovery error are included in the error chain for better incident triage.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@pkg/connector/client.go`:
- Around line 33-34: The AccessToken field is read concurrently in
auth_recovery.go (in the newClient closures at Lines 43 and 55) while being
mutated in refresh/login paths, creating a data race condition. Add a new
tokenMu sync.RWMutex field to the LineClient struct to synchronize all token
access. Create a getAccessToken() method that acquires a read lock and returns
the AccessToken value. Update auth_recovery.go to use this getAccessToken()
getter instead of directly accessing AccessToken. Guard all token write
operations in refreshAndSave and tryLogin methods by acquiring a write lock on
tokenMu before modifying AccessToken. This ensures consistent synchronization of
token state across concurrent operations.
---
Nitpick comments:
In `@pkg/connector/auth_recovery.go`:
- Around line 24-27: The error returned when token recovery fails only preserves
the recovery error from errRecover but loses the original LINE authentication
error that triggered the recovery attempt. Modify the recovery error handling
block to capture the original authentication error before calling
deps.recover(ctx), and then when errRecover is not nil, wrap both errors
together in the returned error using fmt.Errorf or error wrapping to ensure both
the original LINE auth error and the recovery error are included in the error
chain for better incident triage.
In `@pkg/line/errors.go`:
- Around line 22-24: The error string matching in this function is
case-sensitive, which can cause recoverable auth errors to be missed if the
upstream error messages have different casing. Convert the error message to
lowercase using strings.ToLower() before performing the strings.Contains()
checks against both the "code":119 pattern and the "Access token refresh
required" string. Apply the same case-insensitive matching fix to all other
similar error classification predicates mentioned in the file at the referenced
line ranges.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: a074534e-d455-4510-acf7-d316680eba28
📒 Files selected for processing (11)
.gitignorepkg/connector/auth_recovery.gopkg/connector/auth_recovery_test.gopkg/connector/client.gopkg/connector/handlers/file.gopkg/connector/handlers/handler.gopkg/connector/reaction.gopkg/connector/send_message.gopkg/connector/userinfo.gopkg/line/errors.gopkg/line/errors_test.go
💤 Files with no reviewable changes (1)
- .gitignore
📜 Review details
🧰 Additional context used
📓 Path-based instructions (2)
**/*.go
📄 CodeRabbit inference engine (AGENTS.md)
**/*.go: Usego fmtfor code formatting across all Go files
Usegoimportswith-local "github.com/highesttt/matrix-line-messenger"flag to group project-local imports correctly
Usezerologfor logging throughout the codebase
Do not useMsgfin logging; useMsgwith structured fields instead
UseStringerinterface where applicable in Go code
Files:
pkg/connector/reaction.gopkg/connector/handlers/file.gopkg/connector/userinfo.gopkg/line/errors_test.gopkg/connector/auth_recovery_test.gopkg/connector/auth_recovery.gopkg/connector/handlers/handler.gopkg/line/errors.gopkg/connector/client.gopkg/connector/send_message.go
**/!(ltsm)/**/*.go
📄 CodeRabbit inference engine (AGENTS.md)
**/!(ltsm)/**/*.go: Runstaticcheckon all Go files excludingpkg/ltsmpackage (transpiled WASM code)
Rungo veton all Go files excludingpkg/ltsmpackage (transpiled WASM code)
Files:
pkg/connector/reaction.gopkg/connector/handlers/file.gopkg/connector/userinfo.gopkg/line/errors_test.gopkg/connector/auth_recovery_test.gopkg/connector/auth_recovery.gopkg/connector/handlers/handler.gopkg/line/errors.gopkg/connector/client.gopkg/connector/send_message.go
🔇 Additional comments (9)
pkg/connector/auth_recovery_test.go (1)
21-120: LGTM!Also applies to: 122-133, 135-149, 151-194
pkg/line/errors_test.go (1)
8-84: LGTM!pkg/connector/handlers/handler.go (1)
35-35: LGTM!pkg/connector/send_message.go (1)
34-52: LGTM!Also applies to: 226-228, 240-242, 328-330, 411-413, 421-423, 449-451, 513-515, 657-657, 682-684, 719-719, 746-746, 770-772, 783-785, 837-839, 858-862, 866-869, 879-882
pkg/connector/userinfo.go (1)
40-43: LGTM!pkg/line/errors.go (1)
1-8: Run the required Go hygiene checks for this PR cohort.Please verify formatting, import grouping, and static analysis compliance across changed Go packages (excluding
pkg/ltsm). Run the following checks:
gofmtfor code formattinggoimports -local "github.com/highesttt/matrix-line-messenger"for import groupingrg -n --type=go '\.Msgf\s*\('to scan for prohibitedMsgfusage in zerologgo vetandstaticcheckon all packages exceptpkg/ltsmAs per coding guidelines, ensure compliance with
go fmt,goimportswith local grouping,Msg(notMsgf) for structured logging, and static analysis for all Go code outside the WASM package.pkg/connector/handlers/file.go (1)
41-44: LGTM!pkg/connector/reaction.go (1)
369-372: LGTM!Also applies to: 397-399
pkg/connector/auth_recovery.go (1)
1-8: Run mandatory Go quality checks locally before merging.Per coding guidelines, the following checks are required for all Go files (excluding
pkg/ltsm):
go fmtfor formattinggoimports -local "github.com/highesttt/matrix-line-messenger"for import groupinggo vetfor static analysisstaticcheckfor lintingThese tools must be executed in your development environment to verify compliance before this change is merged.
|
@clins1994 Could you check this one out and get back to me in a week to see if that fixes your issue |
|
@Adri11334 thanks for working on it. i'll leave this running on my bridge so we can verify. let's check back in 7 days cc @highesttt |
Awesome. Thank you so much! Looking forward to seeing your feedback om this |
|
@Adri11334 @highesttt good news is it seems to be working so the key evidence happened about 6 days after deploy, and by June 25 the container had passed the ~7-day mark without requiring manual login |
|
@Adri11334 can you pull from main and make sure it compiles and still runs, just making sure nothing breaks. |
|
Pulled latest
Everything passes on my side. |
There was a problem hiding this comment.
Actionable comments posted: 3
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@pkg/connector/client.go`:
- Around line 347-355: FullSync still starts message delivery before portal
creation, so the new portal-first ordering is bypassed. Update the fullSync path
in pkg/connector/sync.go so syncChats runs and completes before syncDMChats and
prefetchMessages are launched, matching the ordering used in Connect and
preventing first-message portal creation from making existing membership look
like a fresh join.
In `@pkg/connector/handle_message.go`:
- Around line 117-153: The hidden join is being emitted twice for a first-seen
sender because ensureGroupMessageSenderKnown already queues it and the Message
PreHandleFunc in handleMessage repeats it once the portal exists. Update the
flow so ensureGroupMessageSenderKnown stays cache-only and remove the duplicate
hidden member update from the Message path, keeping the actual join emission in
one place such as hiddenJoinGroupMessageSender.
In `@pkg/connector/userinfo.go`:
- Around line 319-320: The LINE user-id lookup path in the contact search flow
still skips refresh-required/logged-out recovery; update the logic around
client.FindContactByUserid in the search routine to detect token-expired
failures, trigger the same token recovery/re-login path used by the later
contact-list lookup, and then retry the exact user-id search before returning no
result. Use the existing search method and login recovery helpers in
lc.newClient(), FindContactByUserid, and the surrounding contact lookup flow so
non-contact LINE IDs are retried after token recovery.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: fb65fbd5-86ec-4bd2-b319-ade85fbe8fcc
📒 Files selected for processing (14)
pkg/connector/auth_recovery.gopkg/connector/auth_recovery_test.gopkg/connector/client.gopkg/connector/creategroup.gopkg/connector/e2ee_keys.gopkg/connector/handle_message.gopkg/connector/handlers/audio.gopkg/connector/handlers/file.gopkg/connector/handlers/video.gopkg/connector/send_message.gopkg/connector/sync.gopkg/connector/userinfo.gopkg/line/errors.gopkg/line/errors_test.go
💤 Files with no reviewable changes (3)
- pkg/connector/handlers/file.go
- pkg/connector/handlers/video.go
- pkg/connector/handlers/audio.go
🚧 Files skipped from review as they are similar to previous changes (5)
- pkg/line/errors.go
- pkg/line/errors_test.go
- pkg/connector/auth_recovery_test.go
- pkg/connector/auth_recovery.go
- pkg/connector/send_message.go
📜 Review details
🧰 Additional context used
📓 Path-based instructions (2)
**/*.go
📄 CodeRabbit inference engine (AGENTS.md)
**/*.go: Usego fmtfor code formatting across all Go files
Usegoimportswith-local "github.com/highesttt/matrix-line-messenger"flag to group project-local imports correctly
Usezerologfor logging throughout the codebase
Do not useMsgfin logging; useMsgwith structured fields instead
UseStringerinterface where applicable in Go code
Files:
pkg/connector/handle_message.gopkg/connector/creategroup.gopkg/connector/userinfo.gopkg/connector/e2ee_keys.gopkg/connector/client.gopkg/connector/sync.go
**/!(ltsm)/**/*.go
📄 CodeRabbit inference engine (AGENTS.md)
**/!(ltsm)/**/*.go: Runstaticcheckon all Go files excludingpkg/ltsmpackage (transpiled WASM code)
Rungo veton all Go files excludingpkg/ltsmpackage (transpiled WASM code)
Files:
pkg/connector/handle_message.gopkg/connector/creategroup.gopkg/connector/userinfo.gopkg/connector/e2ee_keys.gopkg/connector/client.gopkg/connector/sync.go
🔇 Additional comments (1)
pkg/connector/sync.go (1)
524-535: 🎯 Functional CorrectnessImmediate post-queue lookup can miss new portals
This only works ifQueueRemoteEvent(...)makes the portal visible before returning. If portal creation is deferred,GetExistingPortalByKey(...)can returnnilhere,pendingPortalCreateswill skip that chat, andwaitForGroupPortalCreates(...)won’t block prefetch/SSE for it.
There was a problem hiding this comment.
Caution
Inline review comments failed to post. This is likely due to GitHub's internal server error or limits when posting large numbers of comments. If you are seeing this consistently it is likely a permissions issue. Please check "Moderation" -> "Code review limits" under your organization settings.
Actionable comments posted: 3
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@pkg/connector/client.go`:
- Around line 347-355: FullSync still starts message delivery before portal
creation, so the new portal-first ordering is bypassed. Update the fullSync path
in pkg/connector/sync.go so syncChats runs and completes before syncDMChats and
prefetchMessages are launched, matching the ordering used in Connect and
preventing first-message portal creation from making existing membership look
like a fresh join.
In `@pkg/connector/handle_message.go`:
- Around line 117-153: The hidden join is being emitted twice for a first-seen
sender because ensureGroupMessageSenderKnown already queues it and the Message
PreHandleFunc in handleMessage repeats it once the portal exists. Update the
flow so ensureGroupMessageSenderKnown stays cache-only and remove the duplicate
hidden member update from the Message path, keeping the actual join emission in
one place such as hiddenJoinGroupMessageSender.
In `@pkg/connector/userinfo.go`:
- Around line 319-320: The LINE user-id lookup path in the contact search flow
still skips refresh-required/logged-out recovery; update the logic around
client.FindContactByUserid in the search routine to detect token-expired
failures, trigger the same token recovery/re-login path used by the later
contact-list lookup, and then retry the exact user-id search before returning no
result. Use the existing search method and login recovery helpers in
lc.newClient(), FindContactByUserid, and the surrounding contact lookup flow so
non-contact LINE IDs are retried after token recovery.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: fb65fbd5-86ec-4bd2-b319-ade85fbe8fcc
📒 Files selected for processing (14)
pkg/connector/auth_recovery.gopkg/connector/auth_recovery_test.gopkg/connector/client.gopkg/connector/creategroup.gopkg/connector/e2ee_keys.gopkg/connector/handle_message.gopkg/connector/handlers/audio.gopkg/connector/handlers/file.gopkg/connector/handlers/video.gopkg/connector/send_message.gopkg/connector/sync.gopkg/connector/userinfo.gopkg/line/errors.gopkg/line/errors_test.go
💤 Files with no reviewable changes (3)
- pkg/connector/handlers/file.go
- pkg/connector/handlers/video.go
- pkg/connector/handlers/audio.go
🚧 Files skipped from review as they are similar to previous changes (5)
- pkg/line/errors.go
- pkg/line/errors_test.go
- pkg/connector/auth_recovery_test.go
- pkg/connector/auth_recovery.go
- pkg/connector/send_message.go
📜 Review details
🔇 Additional comments (1)
pkg/connector/sync.go (1)
524-535: 🎯 Functional CorrectnessImmediate post-queue lookup can miss new portals
This only works ifQueueRemoteEvent(...)makes the portal visible before returning. If portal creation is deferred,GetExistingPortalByKey(...)can returnnilhere,pendingPortalCreateswill skip that chat, andwaitForGroupPortalCreates(...)won’t block prefetch/SSE for it.
🛑 Comments failed to post (3)
pkg/connector/client.go (1)
347-355: 🎯 Functional Correctness | 🟠 Major | 🏗️ Heavy lift
FullSync still bypasses the new portal-first ordering.
Connectnow blocks onsyncChatsbefore prefetch/polling, butpkg/connector/sync.goLine 1121-1124 still startssyncChats,syncDMChats, andprefetchMessagestogether onfullSync. That means a later full sync can still let the first delivered group message create the portal and make existing membership look like a fresh join again.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@pkg/connector/client.go` around lines 347 - 355, FullSync still starts message delivery before portal creation, so the new portal-first ordering is bypassed. Update the fullSync path in pkg/connector/sync.go so syncChats runs and completes before syncDMChats and prefetchMessages are launched, matching the ordering used in Connect and preventing first-message portal creation from making existing membership look like a fresh join.pkg/connector/handle_message.go (1)
117-153: 🎯 Functional Correctness | 🟡 Minor | ⚡ Quick win
Avoid emitting the hidden join twice for a newly seen sender.
ensureGroupMessageSenderKnown(...)already queues a join when the sender is first added to the cache, andPreHandleFuncemits the same hidden join again once the portal is available. Existing group portals will process both. Make the early path cache-only, and keep the actual hidden member update in one place.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@pkg/connector/handle_message.go` around lines 117 - 153, The hidden join is being emitted twice for a first-seen sender because ensureGroupMessageSenderKnown already queues it and the Message PreHandleFunc in handleMessage repeats it once the portal exists. Update the flow so ensureGroupMessageSenderKnown stays cache-only and remove the duplicate hidden member update from the Message path, keeping the actual join emission in one place such as hiddenJoinGroupMessageSender.pkg/connector/userinfo.go (1)
319-320: 🎯 Functional Correctness | 🟠 Major | ⚡ Quick win
Retry exact LINE user-id searches after token recovery.
FindContactByUserid(...)is the one lookup path here that still ignores refresh-required/logged-out failures. When the token expires, searches for non-contact LINE IDs will return nothing even though the later contact-list path recovers correctly.Suggested fix
if lowerQuery != "" { client := lc.newClient() contact, err := client.FindContactByUserid(lowerQuery) + if err != nil && (lc.isRefreshRequired(err) || lc.isLoggedOut(err)) { + if errRecover := lc.recoverToken(ctx); errRecover == nil { + client = lc.newClient() + contact, err = client.FindContactByUserid(lowerQuery) + } + } if err == nil && contact != nil && contact.Mid != "" { if r := lc.midToResolveIdentifier(ctx, contact.Mid); r != nil { results = append(results, r)📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.client := lc.newClient() contact, err := client.FindContactByUserid(lowerQuery) if err != nil && (lc.isRefreshRequired(err) || lc.isLoggedOut(err)) { if errRecover := lc.recoverToken(ctx); errRecover == nil { client = lc.newClient() contact, err = client.FindContactByUserid(lowerQuery) } } if err == nil && contact != nil && contact.Mid != "" {🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@pkg/connector/userinfo.go` around lines 319 - 320, The LINE user-id lookup path in the contact search flow still skips refresh-required/logged-out recovery; update the logic around client.FindContactByUserid in the search routine to detect token-expired failures, trigger the same token recovery/re-login path used by the later contact-list lookup, and then retry the exact user-id search before returning no result. Use the existing search method and login recovery helpers in lc.newClient(), FindContactByUserid, and the surrounding contact lookup flow so non-contact LINE IDs are retried after token recovery.
Summary
Fixes runtime LINE auth recovery for API paths that can currently return expired-token errors directly while the bridge is already running.
When LINE returns an auth error such as
TalkExceptioncode 119 /Access token refresh required, the bridge now attempts token recovery, recreates the LINE client with the recovered token, and retries the failed call once.Fixes #163.
Changes
*_test.goignore so the new auth recovery regression tests are trackedTesting
GOCACHE=/private/tmp/line-gocache GOWORK=off go test ./pkg/line ./pkg/connector ./pkg/connector/handlersdocker run --rm -v "$PWD":/build -w /build alpine:3.23 sh -lc 'apk add --no-cache go git build-base olm-dev staticcheck && go test ./... && go vet $(go list ./... | grep -v /ltsm) && staticcheck $(go list ./... | grep -v /ltsm)'GOCACHE=/private/tmp/line-gocache GOWORK=off go test -race ./pkg/connectordocker build -t matrix-line-auth-recovery .