fix(evm): query `state_snapshot.storage` in `ForkDbStateSnapshot::storage_ref` by gutonosa · Pull Request #14007 · foundry-rs/foundry

gutonosa · 2026-03-30T15:24:18Z

ForkDbStateSnapshot::storage_ref was skipping state_snapshot.storage entirely. The lookup path went from local.cache straight to the remote RPC backend, missing any storage data that SharedBackend had previously fetched and persisted in the BlockchainDb. This caused stale or incorrect values to be returned when reverting to a state snapshot, since the RPC would return current on-chain state rather than the snapshotted state.

aligns storage_ref with basic_ref, which already queries state_snapshot as the second lookup layer. Replaced the redundant get_storage() helper (which only re-checked local.cache) with storage_from_snapshot_or_backend(), giving the correct lookup order: local.cache → state_snapshot.storage → SharedBackend / RPC.

figtracer

let's add a test to verify the behavior

gutonosa · 2026-05-05T08:37:27Z

let's add a test to verify the behavior

sorry for holding a long time, rebase adn add the test

mablr

lgtm, thank you.

Pending another review + conflict resolution

figtracer

thanks! pls solve conflict

…rage_ref`

gutonosa · 2026-05-05T09:59:47Z

solved @mablr @figtracer

@zerosnacks

* fix(cheatcodes): fix solc 0.8.35 error keyword warning (#14509) fix(cheatcodes): avoid solc error keyword warning (cherry picked from commit f0bd9bb) * chore: bump compiler version to 0.8.35 for tests (#14524) * bump compiler version for tests * fix CI: update gas snapshots and bump svm-rs to 0.5.25 Amp-Thread-ID: https://ampcode.com/threads/T-019dddbd-7f28-72cc-965b-d5a23ecb37ce Co-authored-by: Amp <amp@ampcode.com> * make fmt Amp-Thread-ID: https://ampcode.com/threads/T-019dddbd-7f28-72cc-965b-d5a23ecb37ce Co-authored-by: Amp <amp@ampcode.com> * fix: add blockTimestamp to LegacyTransactionResult for testRpcTransactionByHash DRPC now returns a blockTimestamp field, which shifts ABI decoding offsets and breaks the test. Amp-Thread-ID: https://ampcode.com/threads/T-019dddbd-7f28-72cc-965b-d5a23ecb37ce Co-authored-by: Amp <amp@ampcode.com> --------- Co-authored-by: Amp <amp@ampcode.com> (cherry picked from commit 470d048) * fix(ci): use `PATH_USD` fallback fee token in Mail templates (#14546) (cherry picked from commit e8d0d89) * fix(forge): ignore ETH_RPC_URL for test forking (#14555) Co-authored-by: zerosnacks <95942363+zerosnacks@users.noreply.github.com> Co-authored-by: Mablr <59505383+mablr@users.noreply.github.com> (cherry picked from commit 6c9bbea) * fix(cli): fix jsonwebtoken panic (#14562) `cast` panicked with this message coming from jsonwebtoken: ``` Call CryptoProvider::install_default() before this point to select a provider manually, or make sure exactly one of the 'rust_crypto' and 'aws_lc_rs' features is enabled. See the documentation of the CryptoProvider type for more information. ``` This seemingly was introduced with the bump of jsonwebtoken to 10. Now it requires you to pick one backend used by default controlled by the compile time cargo features or call `CryptoProvider::install_default()` at the beginning. I realized that probably it would be better to just select the feature and I picked `aws_lc_rs` as it seems to be increasingly a default and we already are using the C toolchain. Co-authored-by: zerosnacks <95942363+zerosnacks@users.noreply.github.com> (cherry picked from commit 6e40f56) * ci: sign release archives, docker images, and publish SBOMs (#14563) - release.yml: emit per-archive sha256 + SPDX SBOM (Syft), cosign keyless sign-blob of the archive, and use actions/attest@v4 for both build provenance and SBOM attestations. Upload all artifacts to the draft release. - docker-publish.yml: enable BuildKit SBOM, capture the build digest, cosign keyless sign each pushed tag, and publish a Sigstore-signed SLSA provenance attestation via actions/attest with push-to-registry. - SECURITY.md: document how external users verify archives and the docker image (gh attestation, cosign, plain sha256, buildx imagetools). - README.md: link to the new verification section. (cherry picked from commit 3e7f0ed) * fix(ci): increase permissions for the enhanced attestation writing (#14584) * increase permissions for artifact writing * apply write permissions to release-docker (cherry picked from commit 1fd6466) * fix(forge): encode Tempo creates as AA calls (#14585) (cherry picked from commit 8b44ae6) * fix(cli): fix release version strings for immutable tags, bump to 1.7.1 (#14496) * Fix release version metadata for immutable tags Amp-Thread-ID: https://ampcode.com/threads/T-019dd617-b29f-7409-8523-9858a1504f17 Co-authored-by: Amp <amp@ampcode.com> * Derive nightly release suffix from commit SHA Amp-Thread-ID: https://ampcode.com/threads/T-019dd617-b29f-7409-8523-9858a1504f17 Co-authored-by: Amp <amp@ampcode.com> * Apply suggestion from @zerosnacks * Apply suggestion from @zerosnacks * Apply suggestion from @zerosnacks * bump to v1.7.1 * avoid appending whole sha hash, not necessary, handle version cmp correctly. after v1.7.1 release we need to bump to v1.7.2 for nightlies following it to compare correctly * Make foundryVersionCmp tolerate new version format and add tests - Strip both pre-release ('-nightly', '-dev') and build metadata ('+<sha>.<ts>.<profile>') from SEMVER_VERSION before comparison so the cheatcode keeps working for tagged releases (which have no '-' separator). - Extract strip_semver_metadata helper and add Rust unit tests covering all SEMVER_VERSION shapes, version_cmp ordering, and parse_version rejection of pre-release/build/garbage input. - Extend the Solidity test suite for vm.getFoundryVersion()/foundryVersionCmp/foundryVersionAtLeast: validate MAJOR.MINOR.PATCH parseability, build profile value, cmp/atLeast invariant, and error paths for invalid user-supplied versions. Amp-Thread-ID: https://ampcode.com/threads/T-019dd971-fcb7-7149-9680-f0134130844c Co-authored-by: Amp <amp@ampcode.com> * fix(test): drop view from solidity tests using assert helpers and fix fmt - assertTrue/assertEq aren't view, so testGetFoundryVersionBuildProfile and testFoundryVersionCmpAndAtLeastAreConsistent can't be view either. - Collapse the buildType assertion onto one line to satisfy forge fmt. Amp-Thread-ID: https://ampcode.com/threads/T-019dd971-fcb7-7149-9680-f0134130844c Co-authored-by: Amp <amp@ampcode.com> * test(version): assert build profile is non-empty instead of debug|release The dist profile (used for distributed release binaries) is also valid; just require non-empty so any future profile works. Amp-Thread-ID: https://ampcode.com/threads/T-019dd971-fcb7-7149-9680-f0134130844c Co-authored-by: Amp <amp@ampcode.com> * Normalize nightly-<sha> to nightly in release_version Ensures tarball and Docker nightly artifacts produce the same version string. The commit identifier is already included in the SemVer build metadata (after `+`), so collapsing `nightly-<sha>` to `nightly` avoids duplicating the SHA in the pre-release tag. Co-authored-by: Amp <amp@ampcode.com> Amp-Thread-ID: https://ampcode.com/threads/T-019df79e-d4c9-707c-85eb-2efbf59160b3 --------- Co-authored-by: Centaur AI <ai@centaur.local> Co-authored-by: Amp <amp@ampcode.com> Co-authored-by: zerosnacks <95942363+zerosnacks@users.noreply.github.com> Co-authored-by: zerosnacks <zerosnacks@protonmail.com> (cherry picked from commit a70d4aa) * fix(evm): query `state_snapshot.storage` in `ForkDbStateSnapshot::storage_ref` (#14007) * fix(evm): query `state_snapshot.storage` in `ForkDbStateSnapshot::storage_ref` * test(evm): cover `ForkDbStateSnapshot::storage_ref` snapshot lookup (cherry picked from commit fa5429a) * fix(cast): consistent `--json` output for `keychain` subcommands (#14590) - `keychain rl`: wrap remaining limit in `{"remaining":"..."}` object instead of emitting a bare JSON string - `keychain policy add-call`: emit `{"status":"already_present","target":"..."}` when the rule already exists, instead of plain text - `send_keychain_tx`: wrap sponsor hash in `{"sponsor_hash":"0x..."}` object when --tempo.print-sponsor-hash is used with --json Add CLI tests covering the rl and sponsor-hash JSON output shapes. (cherry picked from commit 7916719) * fix(cheatcodes): transfer value for payable mock calls (#14547) * test: updated tests * fix: execute value transfer * test: improve * imp: review item * test: vm.prank test * imp: moved mocked-call handling after prank application --------- Co-authored-by: Mablr <59505383+mablr@users.noreply.github.com> (cherry picked from commit f2e09d2) * fix(forge): `--fuzz-seed` parameter is not effective in `forge coverage` (#14610) fix --fuzz-seed not effective in forge coverage (cherry picked from commit bd9cf55) * fix(cheatcodes): enforce `expectRevert` reverter address for CREATE frames (#14615) * fix(cheatcodes): enforce `expectRevert` reverter address for CREATE frames The reverter address argument to `vm.expectRevert` was silently ignored when the innermost reverting frame was a CREATE (top-level or nested), because create_end never populated `expected_revert.reverted_by`. Mirror call_end's logic in create_end: when the outcome reverts and a reverter address is expected, record outcome.address (revm guarantees this is Some(would-be address) whenever the constructor executed). Adds positive regression tests for top-level and nested-CREATE reverts, and a negative regression test asserting wrong-reverter now fails. Co-authored-by: Amp <amp@ampcode.com> * improve coverage * add Derek's suggested test cases * fix: forge fmt for ExpectRevert.t.sol Amp-Thread-ID: https://ampcode.com/threads/T-019dfdc5-5414-70b6-9f49-cb5797a37a29 Co-authored-by: Amp <amp@ampcode.com> --------- Co-authored-by: Amp <amp@ampcode.com> Co-authored-by: zerosnacks <95942363+zerosnacks@users.noreply.github.com> (cherry picked from commit 85cd75c) * fix(cheatcodes): preserve reverts with `expectEmit` (#14619) * test: added regression test * fix: re-order revert handling * refactor: simplify * lint: fmt * polish: tighten comment, extend test with revert reason and custom error Amp-Thread-ID: https://ampcode.com/threads/T-019dfd96-7a03-7249-8c10-af20ee2729f5 Co-authored-by: Amp <amp@ampcode.com> --------- Co-authored-by: zerosnacks <95942363+zerosnacks@users.noreply.github.com> Co-authored-by: Amp <amp@ampcode.com> (cherry picked from commit 2e8ae6c) * fix(evm): preserve `CREATE` address on revert in isolation mode (#14637) (cherry picked from commit 9584c6c) * chore(clippy): fix for_kv_map and useless_borrows_in_formatting (#14554) * chore(clippy): fix for_kv_map and useless_borrows_in_formatting Amp-Thread-ID: https://ampcode.com/threads/T-019df0f9-62e7-74b8-bd5e-da2acce678fb Co-authored-by: Amp <amp@ampcode.com> * chore(clippy): drop redundant borrows in cheatcodes assert formatters Amp-Thread-ID: https://ampcode.com/threads/T-019df0f9-62e7-74b8-bd5e-da2acce678fb Co-authored-by: Amp <amp@ampcode.com> --------- Co-authored-by: Amp <amp@ampcode.com> (cherry picked from commit 84d3873) * fix(ext): bump forge-std commit in external integration tests (#14504) bump forge commit (cherry picked from commit 024820c) * chore(tests): bump forge-std version (#14510) chore: bump forge-std version used for tests Co-authored-by: zerosnacks <95942363+zerosnacks@users.noreply.github.com> (cherry picked from commit 5766929) * fix(config): suppress spurious unknown-key warning for `network` (#14534) * fix(config): suppress spurious unknown-key warning for `network` The `network` field in `NetworkConfigs` had `skip_serializing_if = "Option::is_none"`, so it was missing from the serialized default `Config`. The warnings provider builds its allowed-keys set from that serialized default, so a valid `network = "tempo"` entry in foundry.toml triggered: Warning: Found unknown `network` config for profile `default` defined in foundry.toml. Replace `skip_serializing_if` with `#[serde(default)]` so `network` appears (as null) in the default config and is recognized as a known profile key. TOML output is unchanged (None is omitted by toml-rs); only the JSON snapshot in `test_default_config` needed updating. Amp-Thread-ID: https://ampcode.com/threads/T-019ddec8-2f04-7255-b8bf-70c1ce2a996d Co-authored-by: Amp <amp@ampcode.com> * test(config): add regression tests for network/tempo unknown-key warnings Cover both the new `network = "tempo"` form and the legacy `tempo = true` alias to ensure neither triggers UnknownKey warnings. Amp-Thread-ID: https://ampcode.com/threads/T-019ddec8-2f04-7255-b8bf-70c1ce2a996d Co-authored-by: Amp <amp@ampcode.com> * fix(config): canonicalize network field on serialization Previously, with `network = "tempo"` set, `forge config` would emit both: network = "tempo" tempo = false which is misleading — the legacy boolean alias contradicts the resolved network. Conversely, with `tempo = true` (legacy), the output omitted `network` entirely, hiding the actual resolved network. Make `NetworkConfigs` use a custom `Serialize` impl that: - Always emits the *resolved* network as the canonical `network = "..."` field. - Never emits the legacy `tempo` / `optimism` boolean aliases (they are still accepted as deserialize-only aliases for backward compatibility). - Continues to emit `celo` and `bypass_prevrandao` as before. This ensures consistent output regardless of input form: network = "tempo" (canonical) → network = "tempo" tempo = true (legacy) → network = "tempo" optimism = true (legacy) → network = "optimism" Both legacy keys are added to `BACKWARD_COMPATIBLE_KEYS` in the warnings provider so they don't trigger unknown-key warnings on input. Test snapshots updated accordingly. Amp-Thread-ID: https://ampcode.com/threads/T-019ddec8-2f04-7255-b8bf-70c1ce2a996d Co-authored-by: Amp <amp@ampcode.com> * style: cargo fmt Amp-Thread-ID: https://ampcode.com/threads/T-019ddec8-2f04-7255-b8bf-70c1ce2a996d Co-authored-by: Amp <amp@ampcode.com> --------- Co-authored-by: Amp <amp@ampcode.com> (cherry picked from commit 95bf101) * feat(forge): use network = "tempo" and add rpc_endpoints in tempo template (#14528) * feat(forge): use network = "tempo" and add rpc_endpoints in tempo template Amp-Thread-ID: https://ampcode.com/threads/T-019dddff-1855-752f-b136-bd0465462ac8 Co-authored-by: Amp <amp@ampcode.com> * feat(forge): add eth_rpc_url = "tempo" to tempo template Amp-Thread-ID: https://ampcode.com/threads/T-019dddff-1855-752f-b136-bd0465462ac8 Co-authored-by: Amp <amp@ampcode.com> * style: cargo fmt Amp-Thread-ID: https://ampcode.com/threads/T-019dddff-1855-752f-b136-bd0465462ac8 Co-authored-by: Amp <amp@ampcode.com> * fix: drop eth_rpc_url from tempo template (would silently fork all tests) Amp-Thread-ID: https://ampcode.com/threads/T-019dddff-1855-752f-b136-bd0465462ac8 Co-authored-by: Amp <amp@ampcode.com> * refactor: build tempo template foundry.toml in a single toml serialize Amp-Thread-ID: https://ampcode.com/threads/T-019dddff-1855-752f-b136-bd0465462ac8 Co-authored-by: Amp <amp@ampcode.com> --------- Co-authored-by: Amp <amp@ampcode.com> (cherry picked from commit 323bceb) * ci(tempo): align ci-tempo workflow with master, drop fork-schedule test step Backport of CI Tempo workflow updates (#14501, #14537, #14556) so the v1.7.1 release branch matches master: - Devnet checks run on PR and push (re-enabled) - Devnet wallet tests run on PR and push - Testnet/mainnet checks moved to nightly schedule only - Nightly failure issue creation The 'Check Tempo fork schedule compatibility' step is intentionally omitted: the underlying tempo::tests::test_fork_schedule_parses_configured_rpcs test was added in #14485 (Tempo deps bump) which is not part of v1.7.1. Amp-Thread-ID: https://ampcode.com/threads/T-019e01cd-cea5-72e8-8338-5f0594e06335 Co-authored-by: Amp <amp@ampcode.com> --------- Co-authored-by: figtracer <me@figtracer.com> Co-authored-by: Amp <amp@ampcode.com> Co-authored-by: Mablr <59505383+mablr@users.noreply.github.com> Co-authored-by: Sergei Shulepov <s.pepyakin@gmail.com> Co-authored-by: grandizzy <38490174+grandizzy@users.noreply.github.com> Co-authored-by: Derek Cofausper <256792747+decofe@users.noreply.github.com> Co-authored-by: Centaur AI <ai@centaur.local> Co-authored-by: zerosnacks <zerosnacks@protonmail.com> Co-authored-by: Nikki <gutonosa@protonmail.com> Co-authored-by: srdtrk <59252793+srdtrk@users.noreply.github.com> Co-authored-by: lazymio <mio@lazym.io> Co-authored-by: stevencartavia <112043913+stevencartavia@users.noreply.github.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

gutonosa requested review from 0xrusowsky, DaniPopes, grandizzy, mattsse, onbjerg and zerosnacks as code owners March 30, 2026 15:24

github-project-automation Bot added this to Foundry Mar 30, 2026

figtracer reviewed Apr 23, 2026

View reviewed changes

gutonosa requested review from mablr and stevencartavia as code owners May 5, 2026 08:35

gutonosa requested a review from figtracer May 5, 2026 08:37

mablr previously approved these changes May 5, 2026

View reviewed changes

figtracer previously approved these changes May 5, 2026

View reviewed changes

gutonosa added 2 commits May 5, 2026 17:49

fix(evm): query state_snapshot.storage in `ForkDbStateSnapshot::sto…

5ffb102

…rage_ref`

test(evm): cover ForkDbStateSnapshot::storage_ref snapshot lookup

1eb3af0

gutonosa dismissed stale reviews from figtracer and mablr via 1eb3af0 May 5, 2026 09:59

gutonosa force-pushed the evm branch from 4ec4345 to 1eb3af0 Compare May 5, 2026 09:59

gutonosa requested review from figtracer and mablr May 5, 2026 09:59

mattsse approved these changes May 5, 2026

View reviewed changes

figtracer approved these changes May 5, 2026

View reviewed changes

figtracer enabled auto-merge (squash) May 5, 2026 11:00

figtracer merged commit fa5429a into foundry-rs:master May 5, 2026
17 of 18 checks passed

github-project-automation Bot moved this to Done in Foundry May 5, 2026

zerosnacks mentioned this pull request May 7, 2026

chore(release): v1.7.1 #14640

Merged

mablr mentioned this pull request Jul 3, 2026

fix(evm): read snapshot storage from state_snapshot instead of local cache #13556

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fix(evm): query `state_snapshot.storage` in `ForkDbStateSnapshot::storage_ref`#14007

fix(evm): query `state_snapshot.storage` in `ForkDbStateSnapshot::storage_ref`#14007
figtracer merged 2 commits into
foundry-rs:masterfrom
gutonosa:evm

gutonosa commented Mar 30, 2026 •

edited

Loading

Uh oh!

figtracer left a comment

Uh oh!

gutonosa commented May 5, 2026

Uh oh!

mablr left a comment •

edited

Loading

Uh oh!

figtracer left a comment

Uh oh!

gutonosa commented May 5, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Uh oh!

Conversation

gutonosa commented Mar 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

figtracer left a comment

Choose a reason for hiding this comment

Uh oh!

gutonosa commented May 5, 2026

Uh oh!

mablr left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

figtracer left a comment

Choose a reason for hiding this comment

Uh oh!

gutonosa commented May 5, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

gutonosa commented Mar 30, 2026 •

edited

Loading

mablr left a comment •

edited

Loading