perf: optimize tracked-file exclusion predicate [CLI-1411]

[basti-snyk]

User description

Summary

• Replace eager O(entries × patterns) index scan with lazy map lookup + early-exit
• Parse .git/index binary directly instead of go-git's full Entry deserialization — reads only file names, skips hashes/timestamps/modes
• Eliminates ~20MB go-git overhead on large repos (100k+ files)

Benchmark (median, flag=on vs flag=off overhead)

Repo	Metric	Before optimization	After optimization
DefinitelyTyped (63k files)	Wall	+60%	+4%
DefinitelyTyped (63k files)	Memory	+73%	≤10% (expected)
cli (84k files)	Wall	+16%	-3%
cli (84k files)	Memory	+59%	+4%
juice-shop (129k files)	Wall	-13%	0%
juice-shop (129k files)	Memory	+5%	-1%

Commits

1. Lazy map + early exit — build map[string]struct{} from index names, early-exit when allMatcher says not excluded, compile 2 matchers instead of 3
2. Direct .git/index parsing — replace go-git PlainOpenWithOptions + Storer.Index() with a lightweight binary parser that reads only file names; handles v2/v3/v4, gitlink files (submodules/worktrees)

Test plan

• Existing TestFileFilter_TrackedFiles* tests pass (real git repos via git.PlainInit)
• New unit tests for readTrackedFileNames (v2, v2 long names, v4 prefix compression)
• New unit tests for findDotGit (found, not found)
• Benchmark on DefinitelyTyped to verify memory reduction
• make lint && make test green

---

PR Type

Enhancement

---

Description

• Replaced dependency on go-git for index parsing with custom logic.

• Optimized tracked file exclusion predicate for performance.

• Introduced utilities for finding the .git directory and parsing index files.

• Added comprehensive tests for new parsing and discovery utilities.

---

Diagram Walkthrough

flowchart LR
  A[FileFilter Context] --> B{Get Tracked Files};
  B -- Old way --> C[go-git Index Loader];
  B -- New way --> D[findDotGit + Custom Index Parser];
  C --> E[All Index Entries Loaded];
  D --> F[Only Tracked File Names Parsed];
  E --> G[Filter Logic (go-git-dependent)];
  F --> H[Filter Logic (custom, performant)];

Loading

File Walkthrough

	Relevant files
Enhancement, dependencies, performance	file_filter.go Custom Git index parsing and optimized exclusion predicate. pkg/utils/file_filter.go Removed the go-git dependency for index operations. Implemented custom binary parsing of .git/index files to read only file names. Added findDotGit function to locate the .git directory and its parent worktree root. Optimized the trackedFileExclusionPredicate to use the custom-parsed tracked file names directly, improving performance. +210/-39
Tests	file_filter_test.go Tests for Git index parsing and .git discovery.                    pkg/utils/file_filter_test.go Added helper functions (gitIndexV2Entry, gitIndexV4Entry, buildGitIndex) for constructing Git index files for testing. Introduced new unit tests for readTrackedFileNames, covering index versions (v2, v4) and prefix compression. Added unit tests for the new findDotGit function, including scenarios where the directory is not found. +101/-0

---

[snyk-io]

✅ Snyk checks have passed. No issues have been found so far.

Status	Scan Engine	Critical	High	Medium	Low	Total (0)
✅	Open Source Security	0	0	0	0	0 issues
✅	Licenses	0	0	0	0	0 issues
✅	Code Security	0	0	0	0	0 issues

💻 Catch issues earlier using the plugins for VS Code, JetBrains IDEs, Visual Studio, and Eclipse.

[snyk-io]

✅ Snyk checks have passed. No issues have been found so far.

Status	Scan Engine	Critical	High	Medium	Low	Total (0)
✅	Open Source Security	0	0	0	0	0 issues
✅	Licenses	0	0	0	0	0 issues
✅	Code Security	0	0	0	0	0 issues
✅	Secrets	0	0	0	0	0 issues

💻 Catch issues earlier using the plugins for VS Code, JetBrains IDEs, Visual Studio, and Eclipse.

[basti-snyk]

/describe

[snyk-pr-review-bot]

PR Description updated to latest commit (acd46f1)