Evaluate how reference-names with slashes can be properly handled in human-facing URLs

[snazy]

Reference names (branches, tags) can contain slashes. The REST API requires those reference names to be properly escaped, which is fine, because it's intended for machine-to-machine communication.

But users (humans) probably want "nice to read" URLs like https://my-nessie-server-ui/tree/my/project/branch/name/commits in the UI, which reads better than https://my-nessie-server-ui/tree/my%2fproject%2fbranch%2fname/commits

[snazy]

From two experiments against GH:
First, create branch tree/bom -> then a create-branch tree fails. Other way aound:
First, create branch tree -> then a create-branch tree/bom fails.
So there seems to be some path-ish validation.

[dimas-b]

Another point: Even though using %2F in the URL path currently works, it is inconvenient for programmatic use because tools (e.g. rest-assured) will generally re-escape the % if it is passed to them in the URL path. At the same time tools will not escape / charts that are part of the ref name because they lack the knowledge of its special meaning.