migrate: check errors in stripWalletAssignmentFKs#674
Merged
Conversation
2 tasks
parkan
added a commit
that referenced
this pull request
Apr 23, 2026
## Summary PR #669 added \`//go:generate make -C sol bytecode\` to \`util/testutil/fvm_precompiles.go\`, which calls \`forge build\` + \`forge inspect\`. The devcontainer's Dockerfile only copied \`anvil\` from the Foundry image, not \`forge\`. Effect: CI's \"Generate swagger code\" step runs \`go generate ./...\` whenever codegen-gated paths (\`api/\`, \`handler/\`, \`cmd/\`, \`storagesystem/\`, \`docs/gen/\`, \`model/\`, \`singularity.go\`, \`docgen.sh\`) change. On those PRs, \`make -C sol bytecode\` fails with \`forge: No such file or directory\`. Currently blocks #670, #673, #674. ## Test plan - [ ] CI devcontainer build includes \`/usr/local/bin/forge\` - [ ] \`go generate ./...\` succeeds inside the devcontainer
a01e2ed to
9a44992
Compare
4 tasks
parkan
added a commit
that referenced
this pull request
Apr 23, 2026
## Summary \`forge inspect\` appends a CBOR metadata section to the deployedBytecode: an IPFS hash of the compiler metadata JSON plus the solc version tag. That IPFS hash drifts between forge versions even with \`solc\` pinned in \`foundry.toml\`. Locally (forge 1.5.1) it matches what was committed in #669; in CI (\`foundry:latest\`, newer) \`go generate ./...\` produces different trailing bytes and \`git diff --exit-code\` fails. Fix: \`bytecode_hash = \"none\"\` + \`cbor_metadata = false\` -- solc omits the metadata section entirely. Output is deterministic across forge versions and solc patches. Regenerated the three .txt files against the new config; runtime bytecode is unchanged (FVM precompile mocks don't consult the metadata). Currently blocks #670, #673, #674 after #675 unblocked the missing-\`forge\` issue. ## Test plan - [x] \`make -C util/testutil/sol clean bytecode\` produces the committed output under forge 1.5.1 - [x] \`go test ./util/testutil/...\` pass - [x] \`go test ./service/dealpusher/...\` pass (DDO / PDP mocks exercise the stripped bytecode) - [ ] CI
Three dialect branches silently discarded errors from the statements that rebuild or strip wallet_assignments FKs. sqlite: INSERT and DROP were unchecked. A failed INSERT (schema drift, constraint) followed by successful DROP + RENAME would wipe the table while reporting success. wallet_assignments carries preparation_id -> actor_id links that export-keys needs to reconstruct preparation.wallet_id, so losing them silently breaks legacy wallet migration. postgres: DROP CONSTRAINT IF EXISTS errors (permissions, lock contention) were dropped. Behavior unchanged on happy path; real errors now surface. mysql: DROP FOREIGN KEY has no IF EXISTS before 8.0.19, so checking information_schema.table_constraints makes the path idempotent. A re-run previously failed silently; it now skips the drop cleanly.
9a44992 to
fb9f64d
Compare
anjor
approved these changes
Apr 24, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Three dialect branches of `stripWalletAssignmentFKs` were discarding `.Error` from their `db.Exec` calls.
sqlite (potentially destructive): INSERT into `wallet_assignments_tmp` and `DROP TABLE wallet_assignments` were unchecked. A silent INSERT failure followed by a successful DROP + RENAME would wipe the table while reporting success. The data is load-bearing -- `export-keys` reads `wallet_assignments` to rebuild `preparation.wallet_id` after the schema change.
postgres: errors from `DROP CONSTRAINT IF EXISTS` (permissions, lock contention) were swallowed; now propagated. Happy path unchanged.
mysql: `DROP FOREIGN KEY` has no `IF EXISTS` before 8.0.19, so a re-run failed and the error was swallowed. Added an `information_schema.table_constraints` check to skip the drop cleanly when the FK is already gone.
Test plan