Originally, a simple substr of the identifier (unprocessed) from the (processed) baseName was used to slice off the part of a full path to a node_modules file. This ended up with incorrect paths to modules.
Now, we actually split then rejoin the path more reliably.
Also adds some test fixtures and does some futzing to baseName for synthetic modules.
Originally, a simple
substrof theidentifier(unprocessed) from the (processed)baseNamewas used to slice off the part of a full path to a node_modules file. This ended up with incorrect paths to modules.Now, we actually split then rejoin the path more reliably.
Also adds some test fixtures and does some futzing to baseName for synthetic modules.