Bump go dependencies#2898
Conversation
Assisted-By: docker-agent
Also bumps github.com/aws/aws-sdk-go-v2/credentials from v1.19.16 to v1.19.17. Assisted-By: docker-agent
Assisted-By: docker-agent
…50fd to v0.0.0-20260526131538-fc97bf12bbdb The new version validates GitHub token CRC32 checksums, so the\nfake tokens used as redaction inputs in tests had to be regenerated\nwith correct base62 CRC32 suffixes. Assisted-By: docker-agent
…portcullistest Both pkg/hooks/builtins/redact_secrets_test.go and pkg/sandbox/kit/ kit_test.go grew an identical CRC32-suffix computation to produce a synthetic, portcullis-detectable GitHub PAT without hard-coding the full 40-char ghp_ literal (which would trip GitHub secret-scanning push protection). Move that computation to a single helper — portcullistest.FakeGitHubPAT — parameterised by a 30-char body, and have both tests call it.
docker-agent
left a comment
docker-agent
left a comment
There was a problem hiding this comment.
Assessment: 🟢 APPROVE
Reviewed 6 files across dependency bumps and a new portcullistest helper package. No bugs found.
Dependency bumps (go.mod / go.sum): All changes are minor/patch version increments — alecthomas/chroma v2.25.0, aws-sdk-go-v2/config v1.32.18, aws-sdk-go-v2/credentials v1.19.17, docker/portcullis (new pseudo-version), junegunn/fzf v0.73.1. The new indirect dependency dlclark/regexp2/v2 v2.1.0 is the expected transitive pull-in from chroma v2.25.0.
FakeGitHubPAT helper (pkg/internal/portcullistest/fakegithubpat.go): The CRC32 base62 encoding is correct — 62^6 ≈ 56.8B comfortably covers the full 32-bit CRC32 range, the big-endian loop is correct, and uint64 is used to prevent intermediate overflow. The 30-char body length guard is correct for ASCII inputs.
Test refactoring (redact_secrets_test.go, kit_test.go): Converting const literals to fakeGitHubPAT()/fakeGitHubToken() calls preserves all test semantics. The reuse of a single token value for both files in TestPrintSummary is intentional and correct (the test asserts 2 secrets redacted).
⚠️ Note: the automated review pipeline completed analysis but the drafter returnedreview_complete: false. All key areas were covered in the analysis; no findings were produced.
4 participants
No description provided.