Replace cycle detection algorithm with Tarjan's strongly connected components algorithm #49
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I noticed that this plugin was taking significant time (almost a second for a 5500 module project) to check for cycles in hot rebuilds. Tarjan's strongly connected components algorithm as I understand is the fastest way to find dependency cycles. Its running time is linear in the number of vertices (modules) + edges (dependencies).
My best guess is that the existing algorithm is quadratic in the number of modules, because for each module in the compilation, all its dependencies are traversed recursively, which in the worst case also visits each module in the compilation.
I haven't done benchmarks, but in the same project after this change, this plugin is no longer perceptible in the ProgressPlugin output.
I also changed the output:
a.js -> b.js
even ifa.js
does not directly importb.js
, which is confusing. If all modules in a cycle are excluded, no warning is generated (I added a test to enforce this)