| e0840cc531ef6c534797a80b5e229fef5fec8570 |
|
26-Apr-2016 |
Tobias Thierer <tobiast@google.com> |
Fix BigDecimal result for -Long.MIN_VALUE / -1L
BigDecimal produced the result -9223372036854775808 (== -2^64
== Long.MIN_VALUE) instead of 9223372036854775808 (== 2^64).
This was because BigDecimal incorrectly used long arithmetic for
this calculation, which overflows back to Long.MIN_VALUE.
This bug didn't happen in the case where the result scale required
adjustment (diffScale != 0): The bitLength(power of ten) calculation
rounds up to whole bits, and therefore is generous enough not to fall
into this trap in the corner case.
Public bug: https://code.google.com/p/android/issues/detail?id=196555
Bug: 26105053
Change-Id: Ie4b1d08c9074d59d854c664ef17aedcaae24c56e
|
| 1d6c936ab01da702aab6c7e578e2c74cdc90b8a4 |
|
18-Mar-2014 |
Kenny Root <kroot@google.com> |
NativeBN: use bn_correct_top on negative BN
If you use BN_bin2bn, it will call bn_correct_top for you. However,
since this code is poking into the structure of the BIGNUM itself
without using API, we must call it ourselves.
Before the test_negativeValues_superfluousZeroes would fail with this
confusing error:
libcore.java.math.BigIntegerTest#test_negativeValues_superfluousZeros
junit.framework.AssertionFailedError: expected -1374707309 but was -1374707309
at junit.framework.Assert.assertEquals(Assert.java:99)
at libcore.java.math.BigIntegerTest.test_negativeValues_superfluousZeros(BigIntegerTest.java:171)
Bug: 12761797
Change-Id: I724c1313e4febca8ce2bda9d10f9b5ebde96a6b1
|
| dd9d07889d37fd2d8d5052b517150ca31ac9d95a |
|
13-Apr-2013 |
Elliott Hughes <enh@google.com> |
Fix BigInteger small prime generation.
OpenSSL won't generate primes smaller than 16 bits, so do it ourselves
roughly the same way they would.
Also remove pure duplicate tests from OldBigIntegerTest, and add some
new tests to the modern BigIntegerTest.
Change-Id: I3000942a7a4f2359d5cbe28795b4f8dd96715ef8
|
| 495c740217c5ab4ed117c030e9bf72919c47809e |
|
12-Jul-2011 |
Elliott Hughes <enh@google.com> |
Add a failing test for a BigInteger.hashCode bug.
Bug: http://code.google.com/p/android/issues/detail?id=18452
Change-Id: I072c54b16de9630833fdf9107428aa4045f4f937
|
| 99d9f92a991cbbfc58b103b3f1c114fbac5d1188 |
|
08-Sep-2010 |
Elliott Hughes <enh@google.com> |
Fix off-by-one error in BigInteger.valueOf.
Bug: 2981072
Change-Id: Ic85aea22fc4e519355c537376b1526828e2375c5
|
| 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
|