sahvx655-wq commented on PR #430:
URL: 
https://github.com/apache/commons-validator/pull/430#issuecomment-5143538500

   Sorry about that, the suite was green here on 21 but I hadn't run it against 
the Java 8 job's locale data, which is where it broke. The CI trace shows the 
'separator without symbol' assertion failing for de_DE, fr_FR and sv_SE plus 
the fr_FR percent case: the legacy pre-CLDR data separates the suffix symbol 
with a plain U+0020 space rather than a non-breaking one, and 
AbstractNumberValidator trims the input before parsing, so on Java 8 the bare 
separator degenerates to the valid no-symbol form and parses to 1234.56 instead 
of being rejected. The explicit-pattern sweep pins a non-breaking space, so it 
was unaffected.
   
   The locale-data test now asserts to the trim contract: a separator above 
U+0020 with no symbol must be rejected, while a plain-space one is trimmed away 
and must parse as the no-symbol form. I reproduced the Java 8 failure locally 
with -Djava.locale.providers=COMPAT before changing anything, and after the fix 
both providers are green; the full default goal with -Ddoclint=all also passes, 
3290 tests, checkstyle, pmd and spotbugs all clean.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to