Movechekcpointtestpackage by evisdren · Pull Request #557 · entireio/cli

evisdren · 2026-02-28T03:18:29Z

No description provided.

Move growth_model_test.go and push_perf_test.go from package checkpoint_test to package checkpoint, and remove self-import in push_perf_test.go. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Entire-Checkpoint: 16571a9186d3

cursor · 2026-02-28T03:18:34Z

PR Summary

Medium Risk
Build-tagged tests add GitHub-pushing and repo-deleting performance harnesses; while not in default test runs, they can be destructive and require elevated gh permissions if executed.

Overview
Adds three new build-tagged test suites for capacity/performance analysis.

growth_model_test.go logs projected per-repo/platform storage growth, push-time estimates, and size-threshold timelines based on configurable checkpoint size/rate assumptions.

push_perf_test.go seeds repos with real transcripts, creates/deletes a temporary GitHub repo via gh, and measures first vs incremental push times plus Trace2 phase breakdowns.

commit_hook_profile_test.go adds a deep PostCommit profiler that times key sub-steps (overlap checks, PrepareCondensation components, batch write/finalize) across many seeded sessions.

^{Written by Cursor Bugbot for commit 1a38675. Configure here.}

Copilot

Pull request overview

Adds opt-in (build-tagged) performance/profiling tests to measure commit-hook behavior, Git push timings, and projected checkpoint growth.

Changes:

Introduces a deep PostCommit profiling test (hookperf) with per-session timing breakdowns.
Adds push performance + trace2 profiling tests (pushperf) that seed repos with real transcript blobs and measure push phases.
Adds a growth projection “model” test (growthmodel) for estimating storage and push-time scaling.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 12 comments.

File	Description
cmd/entire/cli/strategy/commit_hook_profile_test.go	Adds `hookperf` profiling test that decomposes PostCommit into timed sub-steps.
cmd/entire/cli/checkpoint/push_perf_test.go	Adds `pushperf` push timing + trace2 profiling tests and helpers to seed repos using real transcripts.
cmd/entire/cli/checkpoint/growth_model_test.go	Adds `growthmodel` projection test printing tables for storage/push-time growth assumptions.
.claude/worktrees/cached-spinning-nebula	Updates referenced subproject commit.

Copilot · 2026-02-28T03:20:58Z

cmd/entire/cli/checkpoint/push_perf_test.go

+	// Log progress for large scenarios.
+	logInterval := progressInterval(count, transcriptSize)
+
+	for i := range count {


for i := range count does not compile in Go (you can’t range over an int). Use an indexed loop (e.g., for i := 0; i < count; i++ { ... }).

Suggested change

for i := range count {

for i := 0; i < count; i++ {

Copilot · 2026-02-28T03:20:59Z

cmd/entire/cli/strategy/commit_hook_profile_test.go

+		headTree:      headTree,
+		shadowTree:    shadowTree,
+		parentTree:    parentTree,
+		hasParentTree: true,


hasParentTree is always set to true even when parentTree is nil (e.g., initial commit). This can lead to incorrect overlap behavior or nil dereferences inside filesOverlapWithContent. Set hasParentTree based on whether parentTree != nil (and/or early-return when there is no parent).

Suggested change

hasParentTree: true,

hasParentTree: parentTree != nil,

Copilot · 2026-02-28T03:20:59Z

cmd/entire/cli/strategy/commit_hook_profile_test.go

+	ref := shadowRef
+	hasShadowBranch := ref != nil
+	if !hasShadowBranch {
+		refName := plumbing.NewBranchReferenceName(shadowBranchName)
+		var err error
+		ref, err = repo.Reference(refName, true)
+		hasShadowBranch = err == nil
+	}


ref can still be nil if the shadow branch doesn’t exist and repo.Reference fails, but it’s passed unconditionally to calculateSessionAttributions. Guard the attribution step (and any downstream use) on hasShadowBranch or return early when ref == nil.

Copilot · 2026-02-28T03:20:59Z

cmd/entire/cli/strategy/commit_hook_profile_test.go

+
+	// Sub-step 4: Attribution
+	ts3 := time.Now()
+	_ = calculateSessionAttributions(ctx, repo, ref, &ExtractedSessionData{FilesTouched: state.FilesTouched}, state, attributionOpts{headTree: headTree, headTreeHashes: headTreeHashes})


ref can still be nil if the shadow branch doesn’t exist and repo.Reference fails, but it’s passed unconditionally to calculateSessionAttributions. Guard the attribution step (and any downstream use) on hasShadowBranch or return early when ref == nil.

Suggested change

_ = calculateSessionAttributions(ctx, repo, ref, &ExtractedSessionData{FilesTouched: state.FilesTouched}, state, attributionOpts{headTree: headTree, headTreeHashes: headTreeHashes})

if hasShadowBranch && ref != nil {

_ = calculateSessionAttributions(ctx, repo, ref, &ExtractedSessionData{FilesTouched: state.FilesTouched}, state, attributionOpts{headTree: headTree, headTreeHashes: headTreeHashes})

}

Copilot · 2026-02-28T03:21:00Z

cmd/entire/cli/strategy/commit_hook_profile_test.go

+			shadowBranchesToDelete := make(map[string]struct{})
+
+			// Precompute headTree hash index (same as real PostCommit does)
+			headTreeHashes, _ := BuildTreeHashIndex(ctx, headTree)


The error from BuildTreeHashIndex is ignored. Since this is a test/profiler, failing fast on errors makes results more trustworthy; consider handling the error and t.Fatalf so profiling doesn’t silently proceed with incomplete data.

Suggested change

headTreeHashes, _ := BuildTreeHashIndex(ctx, headTree)

headTreeHashes, err := BuildTreeHashIndex(ctx, headTree)

if err != nil {

t.Fatalf("BuildTreeHashIndex failed: %v", err)

}

Copilot · 2026-02-28T03:21:00Z

cmd/entire/cli/checkpoint/push_perf_test.go

+	firstPush := []scenario{
+		// Small sessions
+		{"10cp_100KB", 10, 100 * 1024},
+		{"50cp_500KB", 50, 500 * 1024},
+		// Average-sized sessions (~2.8MB real-world avg)
+		{"50cp_2500KB", 50, 2500 * 1024},
+		{"100cp_2500KB", 100, 2500 * 1024},
+		{"200cp_2500KB", 200, 2500 * 1024},
+		// Heavy sessions
+		{"50cp_10MB", 50, 10 * 1024 * 1024},
+		// Near-max sessions
+		{"10cp_50MB", 10, 50 * 1024 * 1024},
+	}


The firstPush scenario list is duplicated between TestPushPerformance and TestPushProfile. Consider defining shared scenario slices (or a helper that returns them) to keep the two tests in sync and reduce maintenance overhead.

Copilot · 2026-02-28T03:21:00Z

cmd/entire/cli/checkpoint/push_perf_test.go

+	firstPush := []scenario{
+		{"10cp_100KB", 10, 100 * 1024},
+		{"50cp_500KB", 50, 500 * 1024},
+		{"50cp_2500KB", 50, 2500 * 1024},
+		{"100cp_2500KB", 100, 2500 * 1024},
+		{"200cp_2500KB", 200, 2500 * 1024},
+		{"50cp_10MB", 50, 10 * 1024 * 1024},
+		{"10cp_50MB", 10, 50 * 1024 * 1024},
+	}


The firstPush scenario list is duplicated between TestPushPerformance and TestPushProfile. Consider defining shared scenario slices (or a helper that returns them) to keep the two tests in sync and reduce maintenance overhead.

Copilot · 2026-02-28T03:21:00Z

cmd/entire/cli/strategy/commit_hook_profile_test.go

+// profilePrepareCondensation runs PrepareCondensation step by step with timing.
+// This replicates the logic in manual_commit_condensation.go but with instrumentation.


The comments claim this replicates PrepareCondensation, but some steps are effectively no-ops (simulate the filtering loop, and the summary timing measures an empty block). Either update the comments to reflect that these are placeholders, or implement the actual logic you intend to measure so the reported breakdown is meaningful.

Suggested change

// profilePrepareCondensation runs PrepareCondensation step by step with timing.

// This replicates the logic in manual_commit_condensation.go but with instrumentation.

// profilePrepareCondensation runs the PrepareCondensation-related steps with timing.

// It is a profiling helper that approximates the logic in manual_commit_condensation.go,

// and some sub-steps may be simplified or implemented as placeholders.

Copilot · 2026-02-28T03:21:01Z

cmd/entire/cli/strategy/commit_hook_profile_test.go

+	// Sub-step 3: File filtering
+	ts2 := time.Now()
+	if len(committedFiles) > 0 && len(state.FilesTouched) > 0 {
+		for _, f := range state.FilesTouched {
+			_ = f // simulate the filtering loop
+			if _, ok := committedFiles[f]; ok {
+				_ = ok
+			}
+		}
+	}


The comments claim this replicates PrepareCondensation, but some steps are effectively no-ops (simulate the filtering loop, and the summary timing measures an empty block). Either update the comments to reflect that these are placeholders, or implement the actual logic you intend to measure so the reported breakdown is meaningful.

Copilot · 2026-02-28T03:21:01Z

cmd/entire/cli/strategy/commit_hook_profile_test.go

+	// Sub-step 5: Summary generation (disabled in test — summarize needs API key)
+	// Just measure the check
+	ts4 := time.Now()
+	// settings.IsSummarizeEnabled() check only
+	prof.pcSummary = time.Since(ts4)


The comments claim this replicates PrepareCondensation, but some steps are effectively no-ops (simulate the filtering loop, and the summary timing measures an empty block). Either update the comments to reflect that these are placeholders, or implement the actual logic you intend to measure so the reported breakdown is meaningful.

evisdren and others added 2 commits February 27, 2026 19:17

Fix checkpoint test files to use correct package name

a81a7d2

Move growth_model_test.go and push_perf_test.go from package checkpoint_test to package checkpoint, and remove self-import in push_perf_test.go. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Entire-Checkpoint: 16571a9186d3

update packages

1a38675

Copilot AI review requested due to automatic review settings February 28, 2026 03:18

evisdren requested a review from a team as a code owner February 28, 2026 03:18

evisdren closed this Feb 28, 2026

evisdren deleted the movechekcpointtestpackage branch February 28, 2026 03:20

Copilot AI reviewed Feb 28, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Movechekcpointtestpackage#557

Movechekcpointtestpackage#557
evisdren wants to merge 2 commits intomainfrom
movechekcpointtestpackage

evisdren commented Feb 28, 2026

Uh oh!

cursor bot commented Feb 28, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Feb 28, 2026

Uh oh!

Copilot AI Feb 28, 2026

Uh oh!

Copilot AI Feb 28, 2026

Uh oh!

Copilot AI Feb 28, 2026

Uh oh!

Copilot AI Feb 28, 2026

Uh oh!

Copilot AI Feb 28, 2026

Uh oh!

Copilot AI Feb 28, 2026

Uh oh!

Copilot AI Feb 28, 2026

Uh oh!

Copilot AI Feb 28, 2026

Uh oh!

Copilot AI Feb 28, 2026

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

2 participants

-			headTreeHashes, _ := BuildTreeHashIndex(ctx, headTree)
+			headTreeHashes, err := BuildTreeHashIndex(ctx, headTree)
+			if err != nil {
+				t.Fatalf("BuildTreeHashIndex failed: %v", err)
+			}

		// profilePrepareCondensation runs PrepareCondensation step by step with timing.
		// This replicates the logic in manual_commit_condensation.go but with instrumentation.

-// profilePrepareCondensation runs PrepareCondensation step by step with timing.
-// This replicates the logic in manual_commit_condensation.go but with instrumentation.
+// profilePrepareCondensation runs the PrepareCondensation-related steps with timing.
+// It is a profiling helper that approximates the logic in manual_commit_condensation.go,
+// and some sub-steps may be simplified or implemented as placeholders.

Conversation

evisdren commented Feb 28, 2026

Uh oh!

cursor bot commented Feb 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR Summary

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Feb 28, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 28, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 28, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 28, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 28, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 28, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 28, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 28, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 28, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 28, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

2 participants

cursor bot commented Feb 28, 2026 •

edited

Loading