Skip to content

development.md

Latest commit

 

History

History
218 lines (147 loc) · 7.39 KB

development.md

File metadata and controls

218 lines (147 loc) · 7.39 KB
title description
Development Guide
How to build, test, lint, and contribute to go-session.

Development Guide

Prerequisites

  • Go 1.26 or later -- the module requires Go 1.26 (go.mod). The benchmark suite uses b.Loop(), introduced in Go 1.25.
  • github.com/stretchr/testify -- test-only dependency, fetched automatically by go test.
  • vhs (github.com/charmbracelet/vhs) -- optional, required only for RenderMP4. Install with go install github.com/charmbracelet/vhs@latest.
  • golangci-lint -- optional, for running the full lint suite. Configuration is in .golangci.yml.

Build and Test

# Run all tests
go test ./...

# Run a single test by name
go test -v -run TestParseTranscript_ToolCalls_Good

# Run with race detector
go test -race ./...

# Run benchmarks
go test -bench=. -benchmem ./...

# Vet the package
go vet ./...

# Format code
gofmt -w .

# Lint (requires golangci-lint)
golangci-lint run ./...

# Check test coverage
go test -coverprofile=coverage.out ./...
go tool cover -html=coverage.out

If this module is part of the Go workspace at ~/Code/go.work, you can also use the core CLI:

core go test
core go qa          # fmt + vet + lint + test
core go cov --open  # coverage with HTML report

Test Structure

All tests are in the session package (white-box). Test files are co-located with their corresponding source files:

Source file Test file What it covers
parser.go parser_test.go ParseTranscript, ParseTranscriptReader, ListSessions, ListSessionsSeq, FetchSession, extractToolInput, extractResultContent, truncate, shortID, formatDuration, edge cases (malformed JSON, truncated lines, binary garbage, null bytes, 5 MiB lines)
analytics.go analytics_test.go Analyse (nil, empty, single tool, mixed tools with errors, latency calculations, token estimation), FormatAnalytics
html.go html_test.go RenderHTML (basic session, empty session, error events, XSS protection, invalid path, label-per-tool-type)
video.go video_test.go generateTape (basic, skips non-tool events, failed commands, long output truncation, Task/Edit/Write events, empty session, empty command), extractCommand, RenderMP4 error path
search.go search_test.go Search and SearchSeq (empty directory, no matches, single/multiple matches, case-insensitive, output matching, skips non-tool events, ignores non-JSONL files, skips malformed sessions)
-- bench_test.go Performance benchmarks for parsing, listing, and searching

Test Naming Convention

Tests use a _Good, _Bad, _Ugly suffix pattern:

  • _Good: happy path; valid inputs, expected successful output.
  • _Bad: expected error conditions or graceful degradation (malformed input, missing optional data).
  • _Ugly: panic-inducing or extreme edge cases (missing files, nil input, binary garbage, path errors).

Test Helpers

parser_test.go defines helpers for building synthetic JSONL content without external fixtures:

// Fixed epoch: 2026-02-20 10:00:00 UTC, offset by seconds
ts(offsetSec int) string

// Marshal arbitrary map to a single JSONL line
jsonlLine(m map[string]any) string

// Convenience builders for specific entry types
userTextEntry(timestamp, text string) string
assistantTextEntry(timestamp, text string) string
toolUseEntry(timestamp, toolName, toolID string, input map[string]any) string
toolResultEntry(timestamp, toolUseID string, content any, isError bool) string

// Write lines to a temp .jsonl file, return the file path
writeJSONL(t *testing.T, dir string, name string, lines ...string) string

All test output uses t.TempDir(), which Go cleans up automatically after each test.

Benchmarks

The benchmark suite generates synthetic JSONL files with a realistic distribution of tool types (Bash, Read, Edit, Grep, Glob):

Benchmark File size Tool pairs
BenchmarkParseTranscript ~2.2 MB 5,000
BenchmarkParseTranscript_Large ~11 MB 25,000
BenchmarkListSessions 20 files, 100 pairs each --
BenchmarkSearch 10 files, 500 pairs each --

Run with:

go test -bench=. -benchmem ./...

Coverage Targets

The current statement coverage is 90.9%. New contributions should maintain or improve this figure. When adding a function, add tests covering at minimum:

  • The success path.
  • Nil or zero-value inputs where applicable.
  • At least one error path.

Coding Standards

Language

UK English throughout all source code comments, documentation, and commit messages. Examples: colour, organisation, licence, initialise, centre.

Formatting and Lint

Code must be formatted with gofmt. The project uses golangci-lint with the following linters enabled (see .golangci.yml):

  • govet, errcheck, staticcheck, unused, gosimple
  • ineffassign, typecheck, gocritic, gofmt

Both go vet ./... and golangci-lint run ./... must be clean before committing.

Types and Declarations

  • Use explicit types on struct fields and function signatures.
  • Avoid interface{} in public APIs; use typed parameters where possible.
  • Handle all errors explicitly; do not use blank _ for error returns in non-test code.

File Headers

Source files should carry the SPDX licence identifier:

// SPDX-Licence-Identifier: EUPL-1.2
package session

Licence

EUPL-1.2. All new source files must include the SPDX header. By contributing, you agree that your contributions will be licensed under the European Union Public Licence 1.2.

Commit Guidelines

Use conventional commits:

type(scope): description

Common types: feat, fix, test, refactor, docs, chore.

Examples:

feat(parser): add ParseTranscriptReader for streaming parse
fix(html): escape data-text attribute value
test(analytics): add latency calculation edge cases
docs(architecture): update scanner buffer size

All commits must include the co-author trailer:

Co-Authored-By: Virgil <virgil@lethean.io>

go test ./... must pass before committing.

Adding a New Tool Type

  1. Define an input struct in parser.go:

    type myToolInput struct {
    SomeField string `json:"some_field"`
}

  2. Add a case "MyTool": branch in extractToolInput that unmarshals the struct and returns a human-readable string.

  3. Add a corresponding case in html.go's input label logic (inside RenderHTML) if the label should differ from the default "Command". For example, if MyTool targets a URL, use "Target".

  4. Add a case in video.go's generateTape switch if the tool should appear in VHS tape output.

  5. Add tests in parser_test.go:

    • A TestExtractToolInput_MyTool_Good test for extractToolInput.
    • An integration test using toolUseEntry + toolResultEntry to exercise the full parse pipeline.

Adding Analytics Fields

analytics.go is a pure computation layer with no I/O. To add a new metric:

  1. Add the field to the SessionAnalytics struct.
  2. Populate it in the Analyse function's event iteration loop.
  3. Add a row to FormatAnalytics if it should appear in CLI output.
  4. Add a test case in analytics_test.go.

Module Path and Go Workspace

The module path is dappco.re/go/core/session. If this package is used within a Go workspace, add it with:

go work use ./go-session
go work sync