path-has-extension? requires suffix to be strictly shorter than name

[mbutterick]

As seen here. Thereby producing the perplexing result:

(path-has-extension? (string->path "foobar") #"") ; #true
(path-has-extension? (string->path "foobar") #"bar") ; #true
(path-has-extension? (string->path "foobar") #"oobar") ; #true
(path-has-extension? (string->path "foobar") #"foobar") ; #false 

If the comparator were >= rather than >, then the last example would be #true.

[mflatt]

The "not exactly the same as" constraint is explicit in the specification of path-has-extension?. The intent is to avoid treating at dot at the beginning of a path as an extension separator.

[mbutterick]

OK. I remember the earlier discussion at #1307. I thought, erroneously, that a notion of dot-prefixing was still embedded in the idea of an “extension”.