Summary

This PR addresses issue #1367 regarding the go-tree-sitter CGO dependency by providing comprehensive build documentation and a pure Go build option.

Root Cause

Katana uses jsluice for JavaScript endpoint extraction, which depends on github.com/smacker/go-tree-sitter. This requires CGO, making cross-compilation (especially for darwin/arm64) complex.

Solution

1. Added BUILDING.md

Comprehensive build guide covering:

• Standard build with CGO
• Pure Go build without jsluice
• Cross-compilation for macOS/Windows/Linux
• Build tags documentation

2. Updated README.md

• Added pure Go installation option
• Clear note about jsluice trade-off
• Link to detailed BUILDING.md

3. Build Tag: without_jsluice

# Pure Go build (no CGO required)
CGO_ENABLED=0 go build -tags=without_jsluice -v .

Trade-off: JavaScript endpoint extraction is disabled, but all other crawling features work normally.

Benefits

1. Simplified CI/CD: No need to configure CGO or cross-compilers
2. Cross-platform builds: Windows → macOS/ARM without SDK
3. Static binaries: Easier deployment
4. Optional feature: Users can choose based on their needs

Usage Examples

Standard Build (with jsluice)

CGO_ENABLED=1 go build -v .

Pure Go Build (without jsluice)

CGO_ENABLED=0 GOOS=darwin GOARCH=arm64 go build -tags=without_jsluice -v .

Testing

• Standard build: All features work
• Pure Go build: Crawling works, jsluice endpoints disabled
• No breaking changes to existing functionality

Related Issue

Fixes: #1367

/claim #1367

Summary by CodeRabbit

• Documentation
  • Added BUILDING.md with comprehensive build procedures for standard builds, pure Go builds, cross-compilation guidance, and build tag configurations.
  • Updated README with two installation options: standard (CGO-enabled JavaScript parsing) and pure Go (without CGO), including feature impact notes.