History log of /libcore/luni/src/main/java/org/apache/harmony/security/provider/cert/Cache.java
Revision Date Author Comments
fb0ec0e650bf8be35acb0d47da0311a7c446aa33 14-Jan-2011 Elliott Hughes <enh@google.com> Remove useless android-changed comments.

I've changed useful ones to regular comments or TODOs, as appropriate.

I've left ones in code like java.util.concurrent where we really are
tracking an upstream source, making the change markers useful.

I've left a handful of others where I intend to actually investigate
the implied TODOs before deciding how to resolve them.

Change-Id: Iaf71059b818596351cf8ee5a3cf3c85586051fa6
cdeb809350d8c1a14a96924bf01febfa82a8b5b6 23-Sep-2010 Brian Carlstrom <bdc@google.com> Fix ArrayIndexOutOfBoundsExceptions in cert Cache on zero filled array input

The Harmony cert Cache has a long[] where each long is a combination
of a hash and an one-based index into another table containing the
cached values. The cache is searched with Arrays.binarySearch, which
should never find an actual hit, since even a hash hit will look like
a miss since the input hash doesn't contain the one-based index and
entries in the table do. However, this approach has the property that
both hits and misses give the same location in the array, which is
subsequently checked for a real hit/miss with a mask.

However, the hash of a byte array filled with zeroes was zero, which
is found by Arrays.binarySearch in unused slots. Unfortunately, the
code never expects a direct hit, so when it does uses -index-1 to find
the slot to check for hit/miss, it ends up with a negative number,
causing the ArrayIndexOutOfBoundsExceptions.

The solution is ensure that when the hash function returns zero to
simply treat it as a miss. It should not be true for any non-trival
legal input.

Bug: 2753594
Change-Id: I2ee282cc28f22a0ca26da311ae683edf548c67a6
f33eae7e84eb6d3b0f4e86b59605bb3de73009f3 13-May-2010 Elliott Hughes <enh@google.com> Remove all trailing whitespace from the dalvik team-maintained parts of libcore.

Gentlemen, you may now set your editors to "strip trailing whitespace"...

Change-Id: I85b2f6c80e5fbef1af6cab11789790b078c11b1b
fd6bb3510c2f94d636f3572dcf5f7f4dcd1a2726 13-May-2010 Elliott Hughes <enh@google.com> Remove //$NON-NLS-\d$ cruft.

Mostly done by perl(1), with manual cleanup of the few misspelled instances.
This makes our trailing whitespace slightly worse, but I'll fix all that with
a follow-on change.

Change-Id: I0b4ca98819be6f9519c4ba980d759bd1ee1a0303
cec4dd4b1d33f78997603d0f89c0d0e56e64dbcd 26-Apr-2010 Peter Hallam <peterhal@google.com> merge more modules into luni