diff --git a/gitnexus/src/cli/ai-context.ts b/gitnexus/src/cli/ai-context.ts index 8e18fed7d..42dcb7aa7 100644 --- a/gitnexus/src/cli/ai-context.ts +++ b/gitnexus/src/cli/ai-context.ts @@ -199,7 +199,9 @@ async function fileExists(filePath: string): Promise { async function upsertGitNexusSection( filePath: string, content: string, -): Promise<'created' | 'updated' | 'appended'> { + projectName: string, + stats: RepoStats, +): Promise<'created' | 'updated' | 'appended' | 'preserved'> { const exists = await fileExists(filePath); if (!exists) { @@ -223,7 +225,50 @@ async function upsertGitNexusSection( ); if (startIdx !== -1 && endIdx !== -1 && endIdx > startIdx) { - // Replace existing section + const existingSection = existingContent.substring( + startIdx, + endIdx + GITNEXUS_END_MARKER.length, + ); + + // If the existing section contains , preserve the user's + // custom layout and only update the stats line (node/edge/flow counts). + // This lets teams trim the verbose default template to a lean format without + // having it overwritten on every `gitnexus analyze`. + // + // Note: the keep-marker check operates on `existingSection` (the substring + // between valid section markers identified by findSectionMarkerIndex), so + // a keep marker in user prose OUTSIDE the GitNexus block has no effect. + if (existingSection.includes('')) { + // Build the new stats line from the caller-provided values directly. + // We do NOT re-extract from `content` because: + // (a) first-bold extraction is fragile if the template evolves + // (b) the parenthesized-text fallback can match unrelated tuples + // like `({target: "symbolName", direction: "upstream"})` + // when noStats is set + // Passing projectName + stats explicitly makes the contract obvious. + // noStats controls template generation, not keep-section stat updates — the user opted into a stats line by keeping it. + const newStatsInner = `${stats.nodes || 0} symbols, ${stats.edges || 0} relationships, ${stats.processes || 0} execution flows`; + const statsLine = `Indexed as **${projectName}** (${newStatsInner})`; + + // Match either canonical phrasing at line start (`^` with `m` flag) so we + // cannot replace prose embedded mid-paragraph. Deliberately no `$`: text + // after the closing `)` on the same line (e.g. ". MCP tools.") stays intact. + const statsPattern = /^(?:Indexed as|indexed by GitNexus as) \*\*[^*]+\*\* \([^)]+\)/m; + + if (statsPattern.test(existingSection)) { + const updatedSection = existingSection.replace(statsPattern, statsLine); + const before = existingContent.substring(0, startIdx); + const after = existingContent.substring(endIdx + GITNEXUS_END_MARKER.length); + await fs.writeFile(filePath, (before + updatedSection + after).trim() + '\n', 'utf-8'); + return 'updated'; + } + // Keep marker present but no stats line matched. Section is preserved + // unchanged on disk; return a distinct status so callers/CLI output + // don't mis-report this as 'updated' (which would imply a write). + return 'preserved'; + } + + // No keep marker — replace existing section with full verbose content const before = existingContent.substring(0, startIdx); const after = existingContent.substring(endIdx + GITNEXUS_END_MARKER.length); const newContent = before + content + after; @@ -344,12 +389,12 @@ export async function generateAIContextFiles( if (!options?.skipAgentsMd) { // Create AGENTS.md (standard for Cursor, Windsurf, OpenCode, Cline, etc.) const agentsPath = path.join(repoPath, 'AGENTS.md'); - const agentsResult = await upsertGitNexusSection(agentsPath, content); + const agentsResult = await upsertGitNexusSection(agentsPath, content, projectName, stats); createdFiles.push(`AGENTS.md (${agentsResult})`); // Create CLAUDE.md (for Claude Code) const claudePath = path.join(repoPath, 'CLAUDE.md'); - const claudeResult = await upsertGitNexusSection(claudePath, content); + const claudeResult = await upsertGitNexusSection(claudePath, content, projectName, stats); createdFiles.push(`CLAUDE.md (${claudeResult})`); } else { createdFiles.push('AGENTS.md (skipped via --skip-agents-md)'); diff --git a/gitnexus/src/cli/analyze.ts b/gitnexus/src/cli/analyze.ts index d5a7638f9..a20503bc4 100644 --- a/gitnexus/src/cli/analyze.ts +++ b/gitnexus/src/cli/analyze.ts @@ -117,8 +117,18 @@ export interface AnalyzeOptions { verbose?: boolean; /** Skip AGENTS.md and CLAUDE.md gitnexus block updates. */ skipAgentsMd?: boolean; - /** Omit volatile symbol/relationship counts from AGENTS.md and CLAUDE.md. */ - noStats?: boolean; + /** + * Stats inclusion in AGENTS.md and CLAUDE.md. + * + * Commander.js represents `--no-stats` as `stats: boolean` (default + * `true`; `false` when the user passes `--no-stats`), NOT as + * `noStats: boolean`. Reading the negated form would always be + * `undefined` and the flag would silently no-op (#1477). Consumers + * that want "did the user request --no-stats?" should compare with + * `=== false` to distinguish the explicit-off case from the + * default-on case. + */ + stats?: boolean; /** Skip installing standard GitNexus skill files to .claude/skills/gitnexus/. */ skipSkills?: boolean; /** Pure index mode: skip all file injection (AGENTS.md, CLAUDE.md, skills). */ @@ -449,7 +459,12 @@ export const analyzeCommand = async (inputPath?: string, options?: AnalyzeOption skipGit: options?.skipGit, skipAgentsMd, skipSkills, - noStats: options?.noStats, + // commander.js `.option('--no-stats', …)` registers the flag as + // `options.stats` (boolean, default true; `false` when the user + // passed --no-stats). Reading `options?.noStats` here returns + // undefined every time, so the flag was a no-op on the markdown + // rewrite path before this fix. See #1477. + noStats: options?.stats === false, registryName: options?.name, // Registry-collision bypass — its own CLI flag, intentionally NOT // overloading --force. A user who hits the collision guard should @@ -537,7 +552,13 @@ export const analyzeCommand = async (inputPath?: string, options?: AnalyzeOption processes: s.processes, }, skillResult.skills, - { skipAgentsMd, skipSkills, noStats: options?.noStats }, + { + skipAgentsMd, + skipSkills, + // Mirror runFullAnalysis `noStats` bridge (#1477) — same expression; + // exercised on the `--skills` path by analyze-no-stats-bridge.test.ts. + noStats: options?.stats === false, + }, ); } } catch { diff --git a/gitnexus/src/core/ingestion/markdown-processor.ts b/gitnexus/src/core/ingestion/markdown-processor.ts index dc0e7fee4..b2013ee7a 100644 --- a/gitnexus/src/core/ingestion/markdown-processor.ts +++ b/gitnexus/src/core/ingestion/markdown-processor.ts @@ -36,7 +36,12 @@ export const processMarkdown = ( // Skip if file node doesn't exist (shouldn't happen, structure-processor creates it) if (!graph.getNode(fileNodeId)) continue; - const lines = file.content.split('\n'); + // Normalize CRLF/CR to LF before splitting so that line-end agnostic + // markdown files (Windows-authored, mixed) yield correct headings. + // Without this, splitting on `\n` alone leaves `## Heading\r` on each line; + // `$` in HEADING_RE only matches at end-of-string, while `.+` stops before + // the trailing `\r`, so the line never matches as a heading. + const lines = file.content.split(/\r\n|\r|\n/); // --- Extract headings and build hierarchy --- // First pass: collect all heading positions so we can compute endLine spans diff --git a/gitnexus/test/integration/markdown-processor-crlf.test.ts b/gitnexus/test/integration/markdown-processor-crlf.test.ts new file mode 100644 index 000000000..7a3e91a95 --- /dev/null +++ b/gitnexus/test/integration/markdown-processor-crlf.test.ts @@ -0,0 +1,158 @@ +/** + * Regression test for CRLF-encoded markdown heading extraction. + * + * Files with CRLF line endings (Windows-authored markdown) previously + * produced zero Section nodes because `split('\n')` left a trailing `\r` + * on each line, and the heading regex `/^(#{1,6})\s+(.+)$/` (anchored + * with `$`) failed to match `## Heading\r` because `$` only matches at + * end-of-string while `.+` does not consume the trailing `\r`. + * + * Fix: split on `/\r\n|\r|\n/` so all line-ending conventions are + * normalized at split time. See markdown-processor.ts line 39. + */ + +import { describe, it, expect } from 'vitest'; +import { processMarkdown } from '../../src/core/ingestion/markdown-processor.js'; +import { createKnowledgeGraph } from '../../src/core/graph/graph.js'; +import { generateId } from '../../src/lib/utils.js'; +import type { GraphNode } from 'gitnexus-shared'; +import type { KnowledgeGraph } from '../../src/core/graph/types.js'; + +function getMarkdownSections(graph: KnowledgeGraph, filePath: string): GraphNode[] { + return [...graph.iterNodes()] + .filter((n) => n.label === 'Section' && n.properties.filePath === filePath) + .sort( + (a, b) => + ((a.properties.startLine as number | undefined) ?? 0) - + ((b.properties.startLine as number | undefined) ?? 0), + ); +} + +function expectContainsEdge(graph: KnowledgeGraph, sourceId: string, targetId: string) { + const found = [...graph.iterRelationshipsByType('CONTAINS')].some( + (r) => r.sourceId === sourceId && r.targetId === targetId, + ); + expect(found).toBe(true); +} + +function setupGraphWithFile(filePath: string) { + const graph = createKnowledgeGraph(); + const fileNode: GraphNode = { + id: generateId('File', filePath), + label: 'File', + properties: { name: filePath, filePath }, + }; + graph.addNode(fileNode); + return graph; +} + +describe('markdown-processor CRLF tolerance', () => { + it('extracts headings from LF-encoded markdown (baseline)', () => { + const filePath = 'lf.md'; + const graph = setupGraphWithFile(filePath); + const content = '# Title\nbody line 1\n## Sub\nbody line 2\n### SubSub\nmore\n'; + + const stats = processMarkdown(graph, [{ path: filePath, content }], new Set([filePath])); + + expect(stats.sections).toBe(3); + const sections = getMarkdownSections(graph, filePath); + expect(sections.map((s) => s.properties.name)).toEqual(['Title', 'Sub', 'SubSub']); + expect(sections.map((s) => s.properties.level)).toEqual([1, 2, 3]); + expect(sections.map((s) => s.properties.startLine)).toEqual([1, 3, 5]); + expect(sections.map((s) => s.properties.endLine)).toEqual([7, 7, 7]); + for (const s of sections) { + expect(String(s.properties.name)).not.toMatch(/\r/); + } + const fileId = generateId('File', filePath); + expectContainsEdge(graph, fileId, sections[0]!.id); + expectContainsEdge(graph, sections[0]!.id, sections[1]!.id); + expectContainsEdge(graph, sections[1]!.id, sections[2]!.id); + }); + + it('extracts headings from CRLF-encoded markdown (the regression)', () => { + const filePath = 'crlf.md'; + const graph = setupGraphWithFile(filePath); + const content = '# Title\r\nbody line 1\r\n## Sub\r\nbody line 2\r\n### SubSub\r\nmore\r\n'; + + const stats = processMarkdown(graph, [{ path: filePath, content }], new Set([filePath])); + + // Pre-fix: this returned 0 because `## Sub\r` failed the heading regex. + expect(stats.sections).toBe(3); + const sections = getMarkdownSections(graph, filePath); + expect(sections.map((s) => s.properties.name)).toEqual(['Title', 'Sub', 'SubSub']); + expect(sections.map((s) => s.properties.level)).toEqual([1, 2, 3]); + expect(sections.map((s) => s.properties.startLine)).toEqual([1, 3, 5]); + expect(sections.map((s) => s.properties.endLine)).toEqual([7, 7, 7]); + for (const s of sections) { + expect(String(s.properties.name)).not.toMatch(/\r/); + } + const fileId = generateId('File', filePath); + expectContainsEdge(graph, fileId, sections[0]!.id); + expectContainsEdge(graph, sections[0]!.id, sections[1]!.id); + expectContainsEdge(graph, sections[1]!.id, sections[2]!.id); + }); + + it('extracts headings from CR-only-encoded markdown (old Mac OS Classic)', () => { + const filePath = 'cr.md'; + const graph = setupGraphWithFile(filePath); + const content = '# Title\rbody line 1\r## Sub\rbody line 2\r'; + + const stats = processMarkdown(graph, [{ path: filePath, content }], new Set([filePath])); + + expect(stats.sections).toBe(2); + const sections = getMarkdownSections(graph, filePath); + expect(sections.map((s) => s.properties.name)).toEqual(['Title', 'Sub']); + expect(sections.map((s) => s.properties.level)).toEqual([1, 2]); + expect(sections.map((s) => s.properties.startLine)).toEqual([1, 3]); + expect(sections.map((s) => s.properties.endLine)).toEqual([5, 5]); + for (const s of sections) { + expect(String(s.properties.name)).not.toMatch(/\r/); + } + const fileId = generateId('File', filePath); + expectContainsEdge(graph, fileId, sections[0]!.id); + expectContainsEdge(graph, sections[0]!.id, sections[1]!.id); + }); + + it('extracts headings from mixed CRLF + LF markdown', () => { + const filePath = 'mixed.md'; + const graph = setupGraphWithFile(filePath); + const content = '# LF Title\nbody\r\n## CRLF Sub\r\nmore\n### Trailing LF\nend\n'; + + const stats = processMarkdown(graph, [{ path: filePath, content }], new Set([filePath])); + + expect(stats.sections).toBe(3); + const sections = getMarkdownSections(graph, filePath); + expect(sections.map((s) => s.properties.name)).toEqual(['LF Title', 'CRLF Sub', 'Trailing LF']); + expect(sections.map((s) => s.properties.level)).toEqual([1, 2, 3]); + expect(sections.map((s) => s.properties.startLine)).toEqual([1, 3, 5]); + expect(sections.map((s) => s.properties.endLine)).toEqual([7, 7, 7]); + for (const s of sections) { + expect(String(s.properties.name)).not.toMatch(/\r/); + } + const fileId = generateId('File', filePath); + expectContainsEdge(graph, fileId, sections[0]!.id); + expectContainsEdge(graph, sections[0]!.id, sections[1]!.id); + expectContainsEdge(graph, sections[1]!.id, sections[2]!.id); + }); + + it('reports correct startLine and endLine for CRLF content', () => { + const filePath = 'crlf-lines.md'; + const graph = setupGraphWithFile(filePath); + // Lines 1, 3, 5 are headings (1-indexed) + const content = '# T\r\nbody\r\n## Sub\r\nmore\r\n### SubSub\r\ntail\r\n'; + + processMarkdown(graph, [{ path: filePath, content }], new Set([filePath])); + + const sections = getMarkdownSections(graph, filePath); + const titleSection = sections.find((s) => s.properties.name === 'T'); + const subSection = sections.find((s) => s.properties.name === 'Sub'); + const subSubSection = sections.find((s) => s.properties.name === 'SubSub'); + + expect(titleSection?.properties.startLine).toBe(1); + expect(titleSection?.properties.endLine).toBe(7); + expect(subSection?.properties.startLine).toBe(3); + expect(subSection?.properties.endLine).toBe(7); + expect(subSubSection?.properties.startLine).toBe(5); + expect(subSubSection?.properties.endLine).toBe(7); + }); +}); diff --git a/gitnexus/test/unit/ai-context.test.ts b/gitnexus/test/unit/ai-context.test.ts index 71d7ddbdc..68dee21dd 100644 --- a/gitnexus/test/unit/ai-context.test.ts +++ b/gitnexus/test/unit/ai-context.test.ts @@ -45,6 +45,54 @@ describe('generateAIContextFiles', () => { expect(content).toContain('TestProject'); }); + it('omits volatile counts when noStats option is set (#1477)', async () => { + // Distinct subdir per case so we can assert on a clean slate. + const subDir = await fs.mkdtemp(path.join(os.tmpdir(), 'gn-no-stats-test-')); + const subStorage = path.join(subDir, '.gitnexus'); + await fs.mkdir(subStorage, { recursive: true }); + try { + // Stats values picked to be unmistakable if they leak through. + const stats = { nodes: 12345, edges: 67890, processes: 99 }; + await generateAIContextFiles(subDir, subStorage, 'NoStatsProject', stats, undefined, { + noStats: true, + }); + + for (const f of ['CLAUDE.md', 'AGENTS.md']) { + const content = await fs.readFile(path.join(subDir, f), 'utf-8'); + expect(content).toContain('NoStatsProject'); + // The "(N symbols, N relationships, N execution flows)" + // phrase MUST NOT appear when noStats=true. + expect(content).not.toMatch( + /\(\d+\s+symbols,\s+\d+\s+relationships,\s+\d+\s+execution flows\)/, + ); + // And the distinctive numbers must not leak via any other path. + expect(content).not.toContain('12345'); + expect(content).not.toContain('67890'); + } + } finally { + await fs.rm(subDir, { recursive: true, force: true }); + } + }); + + it('preserves volatile counts when noStats is not set (default)', async () => { + const subDir = await fs.mkdtemp(path.join(os.tmpdir(), 'gn-with-stats-test-')); + const subStorage = path.join(subDir, '.gitnexus'); + await fs.mkdir(subStorage, { recursive: true }); + try { + const stats = { nodes: 12345, edges: 67890, processes: 99 }; + await generateAIContextFiles(subDir, subStorage, 'WithStatsProject', stats); + for (const f of ['CLAUDE.md', 'AGENTS.md']) { + const content = await fs.readFile(path.join(subDir, f), 'utf-8'); + expect(content).toContain('WithStatsProject'); + expect(content).toMatch( + /\(12345\s+symbols,\s+67890\s+relationships,\s+99\s+execution flows\)/, + ); + } + } finally { + await fs.rm(subDir, { recursive: true, force: true }); + } + }); + it('keeps the load-bearing repo-specific sections in the CLAUDE.md block (#856)', async () => { // The trimmed block must still contain everything that is genuinely // unique per repo or load-bearing for the agent: the freshness warning, @@ -123,9 +171,81 @@ describe('generateAIContextFiles', () => { expect(starts).toBe(1); }); + it('preserves custom section when gitnexus:keep is present', async () => { + const claudeMdPath = path.join(tmpDir, 'CLAUDE.md'); + + // Write a custom lean section with keep marker + const customContent = `# My Project + +Some project docs here. + + + +# GitNexus — Code Knowledge Graph + +Indexed as **TestProject** (50 symbols, 100 relationships, 5 execution flows). MCP tools. + +| Tool | Use for | +|------|---------| +| query | Find flows | + +Resources: gitnexus://repo/TestProject/context + +`; + await fs.writeFile(claudeMdPath, customContent, 'utf-8'); + + // Run analyze with new stats — should only update the stats line + const stats = { nodes: 999, edges: 1234, processes: 42 }; + await generateAIContextFiles(tmpDir, storagePath, 'TestProject', stats); + + const result = await fs.readFile(claudeMdPath, 'utf-8'); + + // Stats should be updated + expect(result).toContain('999 symbols'); + expect(result).toContain('1234 relationships'); + expect(result).toContain('42 execution flows'); + expect(result).toContain('. MCP tools.'); + + // Custom layout should be preserved (not replaced with verbose template) + expect(result).toContain(''); + expect(result).toContain('Code Knowledge Graph'); + expect(result).toContain('| query | Find flows |'); + + // Verbose template sections should NOT be present + expect(result).not.toContain('## Always Do'); + expect(result).not.toContain('## Never Do'); + expect(result).not.toContain('## When Debugging'); + + // Non-GitNexus content should be preserved + expect(result).toContain('# My Project'); + expect(result).toContain('Some project docs here.'); + }); + + it('replaces section when no keep marker is present', async () => { + const agentsPath = path.join(tmpDir, 'AGENTS.md'); + + // Write a section WITHOUT keep marker + const content = ` +# GitNexus — Code Intelligence + +Old content here. + +`; + await fs.writeFile(agentsPath, content, 'utf-8'); + + const stats = { nodes: 100, edges: 200, processes: 10 }; + await generateAIContextFiles(tmpDir, storagePath, 'TestProject', stats); + + const result = await fs.readFile(agentsPath, 'utf-8'); + + // Should have the full verbose template + expect(result).toContain('## Always Do'); + expect(result).not.toContain('Old content here'); + }); + it('installs skills files', async () => { const stats = { nodes: 10 }; - const result = await generateAIContextFiles(tmpDir, storagePath, 'TestProject', stats); + await generateAIContextFiles(tmpDir, storagePath, 'TestProject', stats); // Should have installed skill files const skillsDir = path.join(tmpDir, '.claude', 'skills', 'gitnexus'); @@ -371,4 +491,240 @@ describe('generateAIContextFiles', () => { await fs.rm(crlfDir, { recursive: true, force: true }); } }); + + // ────────────────────────────────────────────────────────────────── + // Keep-marker edge cases (added to address PR #1508 review findings) + // ────────────────────────────────────────────────────────────────── + + it('keep marker OUTSIDE the GitNexus section has no effect (#1508 review F5)', async () => { + const dir = await fs.mkdtemp(path.join(os.tmpdir(), 'gn-keep-scope-')); + try { + const claudePath = path.join(dir, 'CLAUDE.md'); + // Keep marker appears in user prose BEFORE the GitNexus section. + // The keep-path must NOT be triggered — full template replacement + // is the correct behavior here, because the marker is not inside + // the generated block. + const fileWithOutOfBandMarker = `# My Project + +A note about markers: they only apply inside the +GitNexus block below, not in prose like this. + + +Old verbose stub here. + +`; + await fs.writeFile(claudePath, fileWithOutOfBandMarker, 'utf-8'); + + const stats = { nodes: 50, edges: 100, processes: 5 }; + await generateAIContextFiles(dir, path.join(dir, '.gitnexus'), 'TestProject', stats); + + const result = await fs.readFile(claudePath, 'utf-8'); + // Section MUST have been fully replaced — keep marker outside section ignored + expect(result).toContain('## Always Do'); + expect(result).not.toContain('Old verbose stub here.'); + // User's prose with the marker reference is preserved untouched + expect(result).toContain('A note about markers'); + } finally { + await fs.rm(dir, { recursive: true, force: true }); + } + }); + + it('AGENTS.md keep path preserves custom layout (#1508 review F5)', async () => { + const dir = await fs.mkdtemp(path.join(os.tmpdir(), 'gn-keep-agents-')); + try { + const agentsPath = path.join(dir, 'AGENTS.md'); + const customAgents = `# AGENTS instructions + +Project-specific agent guidance. + + + +# GitNexus context for AGENTS + +Indexed as **AgentsTest** (10 symbols, 20 relationships, 1 execution flows). + +Use 'query' for finding flows, 'context' for symbol details. + +`; + await fs.writeFile(agentsPath, customAgents, 'utf-8'); + + const stats = { nodes: 777, edges: 888, processes: 9 }; + await generateAIContextFiles(dir, path.join(dir, '.gitnexus'), 'AgentsTest', stats); + + const result = await fs.readFile(agentsPath, 'utf-8'); + // Stats updated + expect(result).toContain('777 symbols'); + expect(result).toContain('888 relationships'); + expect(result).toContain('9 execution flows'); + // Custom layout preserved + expect(result).toContain('# GitNexus context for AGENTS'); + expect(result).toContain("Use 'query' for finding flows"); + // Verbose template NOT injected + expect(result).not.toContain('## Always Do'); + // Non-GitNexus content preserved + expect(result).toContain('# AGENTS instructions'); + expect(result).toContain('Project-specific agent guidance.'); + } finally { + await fs.rm(dir, { recursive: true, force: true }); + } + }); + + it('idempotent: second run with keep marker produces byte-identical output (#1508 review F5)', async () => { + const dir = await fs.mkdtemp(path.join(os.tmpdir(), 'gn-keep-idem-')); + try { + const claudePath = path.join(dir, 'CLAUDE.md'); + const seed = `# Project + + + +Indexed as **Idem** (1 symbols, 2 relationships, 3 execution flows). Custom. + +`; + await fs.writeFile(claudePath, seed, 'utf-8'); + + const stats = { nodes: 99, edges: 100, processes: 7 }; + await generateAIContextFiles(dir, path.join(dir, '.gitnexus'), 'Idem', stats); + const afterFirst = await fs.readFile(claudePath, 'utf-8'); + + await generateAIContextFiles(dir, path.join(dir, '.gitnexus'), 'Idem', stats); + const afterSecond = await fs.readFile(claudePath, 'utf-8'); + + expect(afterSecond).toBe(afterFirst); + } finally { + await fs.rm(dir, { recursive: true, force: true }); + } + }); + + it('CRLF file with keep marker: stats line updates without corrupting content (#1508 review F5)', async () => { + // upsertGitNexusSection writes with .trim() + '\n', so the saved file uses LF + // line endings throughout — CRLF in the seed input is not preserved. + const dir = await fs.mkdtemp(path.join(os.tmpdir(), 'gn-keep-crlf-')); + try { + const claudePath = path.join(dir, 'CLAUDE.md'); + const crlfContent = + '# Project\r\n' + + '\r\n' + + '\r\n' + + '\r\n' + + 'Indexed as **CRLFTest** (5 symbols, 6 relationships, 7 execution flows). Custom CRLF.\r\n' + + '\r\n'; + await fs.writeFile(claudePath, crlfContent, 'utf-8'); + + const stats = { nodes: 50, edges: 60, processes: 7 }; + await generateAIContextFiles(dir, path.join(dir, '.gitnexus'), 'CRLFTest', stats); + + const result = await fs.readFile(claudePath, 'utf-8'); + // Stats updated correctly + expect(result).toContain('50 symbols'); + expect(result).toContain('60 relationships'); + // Custom prose preserved + expect(result).toContain('Custom CRLF'); + // No verbose template injected + expect(result).not.toContain('## Always Do'); + } finally { + await fs.rm(dir, { recursive: true, force: true }); + } + }); + + it('noStats + keep marker: stats line update is NOT corrupted by Always-Do tuple text (#1508 review F3)', async () => { + // Regression guard: with the old fallback regex `\(([^)]+)\)`, when + // noStats=true suppressed the canonical stats line from generated + // content, the fallback matched the FIRST parenthesized text in the + // template, which was `({target: "symbolName", direction: "upstream"})` + // from the Always Do bullet — silently writing that as the stats line. + const dir = await fs.mkdtemp(path.join(os.tmpdir(), 'gn-keep-nostats-')); + try { + const claudePath = path.join(dir, 'CLAUDE.md'); + const seed = ` + +Indexed as **NoStatsTest** (1 symbols, 1 relationships, 1 execution flows). Custom. + +`; + await fs.writeFile(claudePath, seed, 'utf-8'); + + const stats = { nodes: 42, edges: 84, processes: 3 }; + await generateAIContextFiles( + dir, + path.join(dir, '.gitnexus'), + 'NoStatsTest', + stats, + undefined, + { noStats: true }, + ); + + const result = await fs.readFile(claudePath, 'utf-8'); + // Stats line MUST NOT have been corrupted with the Always-Do tuple text + expect(result).not.toMatch(/\(\{target:/); + expect(result).not.toMatch(/direction:\s*"upstream"/); + // Stats line should reflect a sensible numeric update (passed stats) + expect(result).toContain('42 symbols'); + // Custom prose still preserved + expect(result).toContain('Custom.'); + } finally { + await fs.rm(dir, { recursive: true, force: true }); + } + }); + + it("returns 'preserved' (not 'updated') when keep marker is present but no stats line matches (#1508 review F1)", async () => { + // Regression guard for the misleading-return-value bug: previously the + // function returned 'updated' without writing when the keep-section had + // no recognizable stats line, causing CLI output to claim files were + // updated when they were not. + const dir = await fs.mkdtemp(path.join(os.tmpdir(), 'gn-keep-noline-')); + try { + const claudePath = path.join(dir, 'CLAUDE.md'); + // Custom keep-section with NO "Indexed as ..." or "indexed by GitNexus as ..." line + const seed = `# Project + + + +# GitNexus block (custom, no stats line) + +This block intentionally omits the standard stats line. + +`; + await fs.writeFile(claudePath, seed, 'utf-8'); + + const stats = { nodes: 100, edges: 200, processes: 10 }; + const result = await generateAIContextFiles( + dir, + path.join(dir, '.gitnexus'), + 'NoLineTest', + stats, + ); + + // The result manifest should reflect 'preserved', not 'updated' + expect(result.files).toContain('CLAUDE.md (preserved)'); + // File on disk is unchanged + const onDisk = await fs.readFile(claudePath, 'utf-8'); + expect(onDisk).toBe(seed); + } finally { + await fs.rm(dir, { recursive: true, force: true }); + } + }); + + it('project name with markdown-sensitive punctuation lands intact in stats line (#1508 review F5)', async () => { + const dir = await fs.mkdtemp(path.join(os.tmpdir(), 'gn-keep-punct-')); + try { + const claudePath = path.join(dir, 'CLAUDE.md'); + const seed = ` + +Indexed as **placeholder** (1 symbols, 1 relationships, 1 execution flows). Custom. + +`; + await fs.writeFile(claudePath, seed, 'utf-8'); + + // Name with hyphens, dot, and slash — exactly what dp-web4/some-repo + // style names look like + const trickyName = 'dp-web4/some-repo.v2'; + const stats = { nodes: 5, edges: 10, processes: 1 }; + await generateAIContextFiles(dir, path.join(dir, '.gitnexus'), trickyName, stats); + + const result = await fs.readFile(claudePath, 'utf-8'); + // The full name appears in the bold of the stats line, intact + expect(result).toContain(`Indexed as **${trickyName}** (5 symbols`); + } finally { + await fs.rm(dir, { recursive: true, force: true }); + } + }); }); diff --git a/gitnexus/test/unit/analyze-no-stats-bridge.test.ts b/gitnexus/test/unit/analyze-no-stats-bridge.test.ts new file mode 100644 index 000000000..f941141ef --- /dev/null +++ b/gitnexus/test/unit/analyze-no-stats-bridge.test.ts @@ -0,0 +1,140 @@ +import { beforeEach, describe, expect, it, vi } from 'vitest'; + +const { runFullAnalysisMock, generateAIContextFilesMock, generateSkillFilesMock } = vi.hoisted( + () => { + const runFullAnalysisMock = vi.fn(); + const generateAIContextFilesMock = vi.fn(async () => ({ files: [] as string[] })); + const generateSkillFilesMock = vi.fn(async () => ({ + skills: [{ name: 'c', label: 'Community', symbolCount: 1, fileCount: 1 }], + outputPath: '/repo/.claude/skills/generated', + })); + return { runFullAnalysisMock, generateAIContextFilesMock, generateSkillFilesMock }; + }, +); + +vi.mock('../../src/core/run-analyze.js', () => ({ + runFullAnalysis: runFullAnalysisMock, +})); + +vi.mock('../../src/cli/ai-context.js', () => ({ + generateAIContextFiles: generateAIContextFilesMock, +})); + +vi.mock('../../src/cli/skill-gen.js', () => ({ + generateSkillFiles: generateSkillFilesMock, +})); + +vi.mock('../../src/core/lbug/lbug-adapter.js', () => ({ + closeLbug: vi.fn(async () => undefined), +})); + +vi.mock('../../src/storage/repo-manager.js', () => ({ + getStoragePaths: vi.fn(() => ({ storagePath: '.gitnexus', lbugPath: '.gitnexus/lbug' })), + getGlobalRegistryPath: vi.fn(() => 'registry.json'), + RegistryNameCollisionError: class RegistryNameCollisionError extends Error {}, + AnalysisNotFinalizedError: class AnalysisNotFinalizedError extends Error {}, + assertAnalysisFinalized: vi.fn(async () => undefined), +})); + +vi.mock('../../src/storage/git.js', () => ({ + getGitRoot: vi.fn(() => '/repo'), + hasGitDir: vi.fn(() => true), +})); + +vi.mock('../../src/core/ingestion/utils/max-file-size.js', () => ({ + getMaxFileSizeBannerMessage: vi.fn(() => null), +})); + +describe('analyzeCommand commander → runFullAnalysis noStats bridge (#1477)', () => { + beforeEach(() => { + vi.resetModules(); + runFullAnalysisMock.mockReset(); + runFullAnalysisMock.mockResolvedValue({ + repoName: 'repo', + repoPath: '/repo', + stats: {}, + alreadyUpToDate: true, + }); + generateAIContextFilesMock.mockReset(); + generateAIContextFilesMock.mockResolvedValue({ files: [] }); + generateSkillFilesMock.mockReset(); + generateSkillFilesMock.mockResolvedValue({ + skills: [{ name: 'c', label: 'Community', symbolCount: 1, fileCount: 1 }], + outputPath: '/repo/.claude/skills/generated', + }); + process.exitCode = undefined; + process.env.NODE_OPTIONS = `${process.env.NODE_OPTIONS ?? ''} --max-old-space-size=8192`.trim(); + }); + + it('maps commander-shaped stats:false to noStats:true (equivalent to --no-stats)', async () => { + const { analyzeCommand } = await import('../../src/cli/analyze.js'); + + await analyzeCommand(undefined, { stats: false }); + + expect(runFullAnalysisMock).toHaveBeenCalledTimes(1); + const opts = runFullAnalysisMock.mock.calls[0][1]; + expect(opts.noStats).toBe(true); + }); + + it('maps omitted stats to noStats:false (default-on preserved)', async () => { + const { analyzeCommand } = await import('../../src/cli/analyze.js'); + + await analyzeCommand(undefined, {}); + + const opts = runFullAnalysisMock.mock.calls[0][1]; + expect(opts.noStats).toBe(false); + }); + + it('maps explicit stats:true to noStats:false', async () => { + const { analyzeCommand } = await import('../../src/cli/analyze.js'); + + await analyzeCommand(undefined, { stats: true }); + + const opts = runFullAnalysisMock.mock.calls[0][1]; + expect(opts.noStats).toBe(false); + }); + + it('still maps stats:false to noStats:true when skipAgentsMd is set', async () => { + const { analyzeCommand } = await import('../../src/cli/analyze.js'); + + await analyzeCommand(undefined, { stats: false, skipAgentsMd: true }); + + const opts = runFullAnalysisMock.mock.calls[0][1]; + expect(opts.noStats).toBe(true); + expect(opts.skipAgentsMd).toBe(true); + }); + + it('passes stats:false as noStats to generateAIContextFiles on the --skills regeneration path (#1477)', async () => { + runFullAnalysisMock.mockResolvedValueOnce({ + repoName: 'repo', + repoPath: '/repo', + stats: { + files: 1, + nodes: 10, + edges: 20, + communities: 0, + processes: 5, + }, + alreadyUpToDate: false, + pipelineResult: { communityResult: undefined }, + }); + + const exitSpy = vi.spyOn(process, 'exit').mockImplementation(() => undefined as never); + try { + const { analyzeCommand } = await import('../../src/cli/analyze.js'); + + await analyzeCommand(undefined, { skills: true, stats: false }); + + expect(generateSkillFilesMock).toHaveBeenCalledTimes(1); + expect(generateAIContextFilesMock).toHaveBeenCalledTimes(1); + const aiCtxOpts = generateAIContextFilesMock.mock.calls[0]![5]; + expect(aiCtxOpts).toEqual({ + skipAgentsMd: undefined, + skipSkills: undefined, + noStats: true, + }); + } finally { + exitSpy.mockRestore(); + } + }); +});