fnmatch_translate_extended() incorrectly translates wildcard pattern

[ericzundel]

I found that the fnmatch_translate_extended translation of the pattern 'squareup/inventory/adjustment*.proto' to a regular expression did not match the string 'squareup/inventory/adjustment.proto'

This is because the re returned expects at least one character to replace '*' in the expression above:

from twitter.common.dirutil.fileset import fnmatch_translate_extended
pattern = fnmatch_translate_extended('squareup/inventory/adjustment*.proto')
(Pdb) p pattern
u'squareup\\/inventory\\/adjustment([^/]+)\\.proto\\Z(?ms)'

A better translation would be to replace the '+' with '*':

u'squareup\\/inventory\\/adjustment([^/]*)\\.proto\\Z(?ms)'

[ericzundel]

Proposed fix in https://rbcommons.com/s/twitter/r/2351/