Improve license parsing error handling

[goneall]

Checks simple license tokens to make sure they are either a listed license or a LicenseRef-

Fixes #227

[pmonks]

@goneall this is probably a separate issue, unrelated to this PR, but I'm not seeing any parsing of AdditionRefs in the 3.x parsing functions. I assume those code paths (if/when they're added) would benefit from similar message updates as the ones here?

[goneall]

@goneall this is probably a separate issue, unrelated to this PR, but I'm not seeing any parsing of AdditionRefs in the 3.x parsing functions. I assume those code paths (if/when they're added) would benefit from similar message updates as the ones here?

Good catch @pmonks - In looking at the license addition code, I realized I'm not taking into account external license refs which is yet another issue.

I'll add a check for the AdditionRef syntax and also implement the external license refs. I'll also add some unit tests to make sure these cases are covered.

[pmonks]

@goneall if it's helpful, here are the regexes I use in my own parser for LicenseRefs & AdditionRefs (the values inside the #" ... " syntax):

; LicenseRefs:
#"(DocumentRef-(?<DocumentRef>[\p{Alnum}-\.]+):)?LicenseRef-(?<LicenseRef>[\p{Alnum}-\.]+)"

; AdditionRefs:
#"(DocumentRef-(?<AdditionDocumentRef>[\p{Alnum}-\.]+):)?AdditionRef-(?<AdditionRef>[\p{Alnum}-\.]+)"

Clojure is hosted on the JVM, so these regexes will also work in Spdx-Java-Library (albeit with escaping as needed for a Java string literal that then gets compiled to a Pattern - Clojure has literal syntax for regexes, which obviates the need for escaping of anything but regex-specific characters).

Oh and note that these regexes make use of named capturing groups for extracting the custom sections of the DocumentRef, LicenseRef and AdditionRef components, but NCGs were only added in Java 7, so depending on the minimum Java version Spdx-Java-Library supports, these regexes might need a little tweaking to switch those to regular (non-named) capturing groups.

[goneall]

@goneall this is probably a separate issue, unrelated to this PR, but I'm not seeing any parsing of AdditionRefs in the 3.x parsing functions. I assume those code paths (if/when they're added) would benefit from similar message updates as the ones here?

Good catch @pmonks - In looking at the license addition code, I realized I'm not taking into account external license refs which is yet another issue.

I'll add a check for the AdditionRef syntax and also implement the external license refs. I'll also add some unit tests to make sure these cases are covered.

Turns out I was checking for external license refs (first line of the method), so I only need to add the additional check for the additionref syntax.

@pmonks - let me know if I missed anything

[pmonks]

@goneall yep I see handling of external LicenseRefs here.

[goneall]

@pmonks @bact - if you could give this another review - I think I addressed all the issues

[bact]

Minor changes, adding 'a'.