fix: move uppercase char check into lowercase method

[dwalluck]

No description provided.

[jeremylong]

LGTM

[jeremylong]

While I'm okay with this change - I question why there is any implementation of toLowerCase() in the code base. This method might be faster for ASCII strings; however, String.toLowerCase() exists, is highly optimized, handles different locals, etc.

[dwalluck]

While I'm okay with this change - I question why there is any implementation of toLowerCase() in the code base. This method might be faster for ASCII strings; however, String.toLowerCase() exists, is highly optimized, handles different locals, etc.

It was to avoid needing to use a Locale as an argument. The previous lack of a Locale in toLowerCase made the lowercase code incorrect on certain other non-English locales. ASCII lowercase should be locale-independent anyway.

The String::toLowerCase for ASCII (actually, Latin1) does some similar bitwise operations, but not the same. Performance should be similar. By the way, Google Guava has something similar https://github.com/google/guava/blob/master/guava/src/com/google/common/base/Ascii.java#L407-L422, so I assume there is a good reason for it.

[ppkarwasz]

I guess this is something the specification should specify. For the general PURL spec, the elements that should be lowercased need also to be ASCII, so there is no problem there.
A problem might appear if some ecosystems allow arbitrary letters in namespaces and names.

[jeremylong]

Merging this - as the discussion should be a question for the spec. The PR is just a minor change in existing code.