| commit | f2297f595285082c7b4b859024fe14e14b73a497 | [log] [tgz] |
|---|---|---|
| author | Philippe Suter <psuter@us.ibm.com> | Wed Dec 07 14:04:10 2016 -0500 |
| committer | Rodric Rabbah <rabbah@us.ibm.com> | Wed Dec 14 13:15:14 2016 -0500 |
| tree | bae989891c68fadd4c4007f5dd2e59732ad1602e | |
| parent | ba67939a8343349e17351125cbdc6ce32697028c [diff] |
Measure bytesize of strings as if encoded in UTF-8. Previous implementation took the conservative estimate that every char is two bytes. While this may fit the JVM model, it violates user expectations that, for instance, base64-encoded binary files take 1 byte per character. Adapted tests to new way of measuring String sizes.