Hi David,
(1) Deflater.setDictionary(ByteBuffer) missed a "@since 11"
(2) infalte(...) may not need to catch DFE twice, better (?) to just
update the inputPos/input
at outsider catch as
if (input == null)
this.inputPos = inputPos + inputConsumed;
else
input.position(inputPos + inputConsumed;
the if/else might be repetitive, but better than two layers of same
DFE catch?
-sherman
On 3/30/18, 9:54 AM, David Lloyd wrote:
Thanks!
On Fri, Mar 30, 2018 at 11:52 AM, Xueming Shen<xueming.s...@oracle.com> wrote:
On 3/30/18, 7:07 AM, Alan Bateman wrote:
On 29/03/2018 13:18, David Lloyd wrote:
:
OK great. In that case, I think all feedback has been accounted for,
and this should be ready to go AFAIK.
I skimmed through the patch attached to your last mail. I also saw
Sherman's mail offering to look at the existing wording about the flush
marker. So I think this the API is good and we should get the CSR submitted.
I'm less sure about the tests. The patch modifies FlaterTest but it's not
clear that is covers all the scenarios, ReadOnlyBufferException just one
that comes to mind. Maybe we could identify additional tests while the CSR
is in progress.
-Alan
CSR: https://bugs.openjdk.java.net/browse/JDK-8200527
