Bug: TransferStatusPage regresses to step 1 after failed hash resubmit #86

New issue

Open

opened 2026-03-30 02:54:16 +00:00 by Brouie · 4 comments

Brouie commented

2026-03-30 02:54:16 +00:00

(Migrated from gitlab.com)

After HASH SUBMISSION FAILED on step 2, clicking the resubmit button drops the TransferStatusPage back to step 1 (DEPOSIT ACTIVE, 'Confirming deposit on source chain') instead of staying on step 2 or retrying the submission. The deposit is already confirmed on-chain — page should not regress. Happens both from the Transfer Status page retry button and after manual SUBMIT HASH NOW on the Verify page. Found on v0.1.197/84cf7e4 testing Anvil->LocalTerra 100 TKNA.

PlasticDigits commented

2026-03-30 08:25:30 +00:00

(Migrated from gitlab.com)

mentioned in commit 001afd14ee

mentioned in commit 001afd14ee0e8b2a1aa1cad703dce857020a084c

PlasticDigits commented

2026-03-30 08:47:41 +00:00

(Migrated from gitlab.com)

@brouie Fix landed on feat/solana-integration (commit 001afd1): hash retry no longer resets lifecycle to deposited, so the stepper stays on Submit Hash; we call triggerSubmit() after resetForRetry(). Also tightened step index for retry/synthetic source-only lookups. Please verify on Anvil → LocalTerra when you can.

@brouie Fix landed on `feat/solana-integration` (commit `001afd1`): hash retry no longer resets lifecycle to `deposited`, so the stepper stays on Submit Hash; we call `triggerSubmit()` after `resetForRetry()`. Also tightened step index for retry/synthetic source-only lookups. Please verify on Anvil → LocalTerra when you can.

Brouie commented

2026-04-23 05:08:16 +00:00

(Migrated from gitlab.com)

@PlasticDigits confirming your PDA-keypair hypothesis — QA stack deploy-solana step fails with DeclaredProgramIdMismatch (AnchorError 4100), so E2E verification is blocked.

Attempted this session: make start-qa → deploy-solana loops on 'Error creating IDL account' with program 74o1KKwuUvtrf6ozbjmVrARnoHpQ4WjsQ647TCTKe1mW — declared program id in source doesn't match deployed address. Bridge stack has Anvil + LocalTerra up but Solana step aborts the whole start flow (no SKIP_SOLANA flag in start-qa.sh, solana is baked into docker compose up + deploy-solana).

Code-level review of the fix 001afd1 (pure frontend, 20 lines in TransferStatusPage.tsx):

resetForRetry() no longer writes lifecycle: 'deposited' on retry (root cause of the regression)
void triggerSubmit() called directly after resetForRetry() — cleaner retry flow
currentStepIdx hardened for 3 edge states: retryingHash, submitting-hash with source present, synthetic source-only (source confirmed but dest null)
useCallback dep array updated to include triggerSubmit

Fix is present on main as of this morning's pull, code review looks correct. Actual reproduction of the failed-hash-retry flow deferred until Solana PDA is sorted and QA stack can come up clean (or until you want me to patch start-qa to allow a Solana-less bringup for EVM↔Terra verification).

Leaving open.

@PlasticDigits confirming your PDA-keypair hypothesis — QA stack deploy-solana step fails with `DeclaredProgramIdMismatch` (AnchorError 4100), so E2E verification is blocked. Attempted this session: `make start-qa` → deploy-solana loops on 'Error creating IDL account' with program `74o1KKwuUvtrf6ozbjmVrARnoHpQ4WjsQ647TCTKe1mW` — declared program id in source doesn't match deployed address. Bridge stack has Anvil + LocalTerra up but Solana step aborts the whole start flow (no `SKIP_SOLANA` flag in start-qa.sh, solana is baked into `docker compose up` + `deploy-solana`). Code-level review of the fix `001afd1` (pure frontend, 20 lines in TransferStatusPage.tsx): - `resetForRetry()` no longer writes `lifecycle: 'deposited'` on retry (root cause of the regression) - `void triggerSubmit()` called directly after `resetForRetry()` — cleaner retry flow - `currentStepIdx` hardened for 3 edge states: `retryingHash`, `submitting-hash` with source present, synthetic source-only (source confirmed but dest null) - useCallback dep array updated to include `triggerSubmit` Fix is present on main as of this morning's pull, code review looks correct. Actual reproduction of the failed-hash-retry flow deferred until Solana PDA is sorted and QA stack can come up clean (or until you want me to patch start-qa to allow a Solana-less bringup for EVM↔Terra verification). Leaving open.

PlasticDigits commented

2026-04-23 05:30:37 +00:00

(Migrated from gitlab.com)

Set as low priority