JDK-8254073, unicode escape preprocessing, and \u005C
John Rose
john.r.rose at oracle.com
Thu Jul 22 22:59:22 UTC 2021
FWIW I think it’s OK to document the long-standing
behavior as correct. Do proceed!
For me this is a lesson in how the most carefully
conceived escape conventions can have unintentional
consequences in dark corners. That is why I think
there’s still room for improvement, not in the JLS,
but in the behavior of the processors of the JLS.
So to track my earlier suggestion for a lint rule to flag
invidious unicode escapes like \u005c and \u0020, I filed:
https://bugs.openjdk.java.net/browse/JDK-8271171
— John
More information about the compiler-dev
mailing list