Better handling of calls to RepositoryEntityLinks.slash(…) [DATAREST-76]

Oliver Drotbohm opened DATAREST-76 and commented

Currently the implementation of RepositoryEntityLinks.slash(…) has a few shortcomings. It currently uses builder.pathSegment(…) which means that Strings starting with a slash will be prepended as is. This might result in duplicate slashes and query parameters and fragment included in the given String not discovered as such.

It's probably better to build a UriComponents from the String representation of the given object and add the relevant parts to the current builder:

UriCopmponents components = UriComponentsBuilder.fromUriString(path).build();
for (String pathSegment : components.getPathSegments()) {
  builder.pathSegment(pathSegment);
}

builder.query(components.getQuery());
builder.fragments(components.getFragment());

No further details from DATAREST-76

Start at RepositoryEntityLinks.slash(…) and trace its current pathSegment handling, then inspect the UriComponents and UriComponentsBuilder APIs mentioned in the issue. Done means leading slashes do not create duplicates and query parameters and fragments in the supplied String are handled separately by the resulting URI.