replayanalysis

Replay analysis sample

Shows replay-safe orchestration patterns and how to verify them with cmd/orchestratorvet.

Features

• Good runnable orchestration uses deterministic counterparts: ctx.CurrentTimeUtc, ctx.NewGuid, ctx.CreateTimer, and activities for side effects.
• Intentional bad fixtures live under testdata/bad, outside normal go test ./... package discovery.
• Integration validation can build the existing vet tool and run positive/negative checks without adding a module or root dependency.

Prerequisites

• DTS_CONNECTION_STRING for DTS.
• Existing cmd/orchestratorvet sources.

Run

# from the repository root
go run ./samples/replayanalysis

Build and run the analyzer:

# from the repository root
(cd cmd/orchestratorvet && go build -o ../../samples/replayanalysis/orchestratorvet-local .)
go vet -vettool="$PWD/samples/replayanalysis/orchestratorvet-local" ./samples/replayanalysis
! go vet -vettool="$PWD/samples/replayanalysis/orchestratorvet-local" ./samples/replayanalysis/testdata/bad
rm -f samples/replayanalysis/orchestratorvet-local

The negative command must report hazards such as wall-clock time, HTTP I/O, goroutines/channels/select, random UUIDs, and stdout logging.

Suggested fixes

To preview suggested fixes, copy the bad fixture to a scratch file inside the workspace and run go vet -fix against that copy. Do not modify repository fixtures during integration validation.

Expected outcome

The runnable sample exits 0 and prints:

SAMPLE_OK replayanalysis

The good vet command exits 0. The bad fixture vet command exits nonzero with the expected diagnostics.

Cleanup

The sample recursively terminates/purges its orchestration before worker shutdown. It creates no external resources.

Required validation variants

Integration validation should run the go run, positive vet, and negative vet commands above.