Open kevindrosendahl opened 5 years ago
a build should contain much more information about why a particular definition is invalid
this includes:
design decision: currently the resolver does a depth first walk of the tree, and when it hits a single error it bails. should it instead just bail from that path down the tree, and try to aggregate as many errors as it can?
a build should contain much more information about why a particular definition is invalid
this includes:
design decision: currently the resolver does a depth first walk of the tree, and when it hits a single error it bails. should it instead just bail from that path down the tree, and try to aggregate as many errors as it can?