| 2981b5e8cf7c19dfd85b2088b18b7a6146825317 |
|
03-Feb-2011 |
Elliott Hughes <enh@google.com> |
Improve CharsetDecoder's quality of implementation, like CharsetEncoder.
As I suspected, CharsetDecoder doesn't cope with a character's bytes being
split across multiple writes any more than CharsetEncoder could cope with
halves of a surrogate pair being split across multiple writes. This seems
much more likely to have harmed applications than CharsetEncoder (since
surrogates are rare).
Anyway, ICU does the right thing here too, so the fix is basically the same.
I've also gone through the decoder/encoder code a bit to bring them more
in line with each other.
Bug: 3410124
Change-Id: I151d043e474161e324361cddfc73188ba73fd59c
|
| 54f969b83377b1540efc0180962966df0732a479 |
|
15-Sep-2010 |
Jesse Wilson <jessewilson@google.com> |
Moving Android-authored charset tests into the libcore tree.
Change-Id: I8b3477249ffcd8c6baf34887c502e609d0c0b238
|
| 4557728efb66c455a52b7669a8eefef7a9e54854 |
|
11-Aug-2010 |
Jesse Wilson <jessewilson@google.com> |
Moving tests to be under the libcore.* package.
This is indended to make it easier to run on VMs that restrict the packages
from which application classes can be loaded. For example, on the RI you need
to use the bootclasspath to load these tests.
Change-Id: I52193f35c5fcca18b5a3e1d280505b1e29b388af
|