The Reader asks each Registry in turn if it 'knows' a filename. If it does, that Registry is then used to retrieve the file 'fullname', 'path', or contents. In the case of LocalRegistry the component is searched for in a more-restrictive way for the three latter cases, allowing for the possibility that a LocalRegistry 'knows' it has a file but can't find it.
This makes the later checks less stringent, fixing #79, which could have the effect of allowing multiple matches for a file, especially if two copies of the same file (with modifications, presumably) exist in the same directory tree.
For now, this possible drawback is not blocking, since if there is a directory structure like
outside of the runtime directory, then one could specify -I /ComponentA or -I /backup/ComponentA (instead of -I /) to get just one copy of the same-named file in the LocalRegistry.
And a similar layout under the runtime directory would require moving or renaming of files.
The
Readerasks eachRegistryin turn if it 'knows' a filename. If it does, thatRegistryis then used to retrieve the file 'fullname', 'path', or contents. In the case ofLocalRegistrythe component is searched for in a more-restrictive way for the three latter cases, allowing for the possibility that aLocalRegistry'knows' it has a file but can't find it.This makes the later checks less stringent, fixing #79, which could have the effect of allowing multiple matches for a file, especially if two copies of the same file (with modifications, presumably) exist in the same directory tree.
For now, this possible drawback is not blocking, since if there is a directory structure like
outside of the runtime directory, then one could specify
-I /ComponentAor-I /backup/ComponentA(instead of-I /) to get just one copy of the same-named file in theLocalRegistry. And a similar layout under the runtime directory would require moving or renaming of files.