Skip to content
jscyclesjscyclesjscycles

Comparison with other tools

  • madge draws dependency graphs as images, reads webpack configuration, and lists every cycle. jscycles does none of these. It does report package cycles in monorepos, follows tsconfig extends chains and workspace aliases, and fails rather than reporting clean when a file cannot be parsed.
  • dpdm prints dependency trees from given entry files and can skip type dependencies (-T).
  • eslint-plugin-import no-cycle reports cycles inline in the editor, at a higher cost per lint run.
  • Nx module boundaries enforces dependency rules between projects in Nx workspaces.

Results differ between tools because their resolution rules differ. Compare them on your own repository before switching.

Performance

Section titled “Performance”

Median of five runs after one warm-up, on an Apple M3 Pro (12 cores) running macOS, analyzing each repository’s packages directory with every source extension enabled:

Repository Source files jscycles 0.2.0 madge 8.0.0 dpdm 3.14.0
nest 1,068 0.08 s 2.04 s 0.86 s
typeorm 3,609 0.28 s 2.97 s 1.48 s
excalidraw 631 0.11 s 2.63 s 0.87 s
ts-morph 1,171 0.08 s 1.64 s 0.83 s

The tools count cycles differently (see what it finds), so their cycle counts are not comparable. Reproduce with python3 scripts/compare_tools.py in a checkout; results depend on hardware and file cache.