On 29/06/19 4:31 PM, Jaikiran Pai wrote: > Hello Stuart, > > Thank you for the detailed response. Comments inline. > > On 28/06/19 2:48 AM, Stuart Marks wrote: >> On 6/26/19 9:28 PM, Jaikiran Pai wrote: >>> I am looking to contribute a patch for the enhancement noted in >>> https://bugs.openjdk.java.net/browse/JDK-8225763. The change itself >>> looks relatively straightforward to me and here's what I plan to do: >>> >>> 1. Have both java.util.zip.Inflater and java.util.zip.Deflater start >>> implementing the AutoCloseable interface >>> >>> 2. Have the close() method call the end() method >>> >>> 3. Add javadoc to the close() implementation to mention that the end() >>> method gets called >>> >>> 4. Add test(s) to verify that the close indeed actually "ends" the >>> inflater/deflater >>> >>> 5. This is more of a question - do we have to update the class level >>> javadoc of both these classes to make a mention that these classes have >>> started implementing the AutoCloseable starting version 14 (or whichever >>> version this change makes into) of Java? >>> >>> Any other inputs/suggestions on what else might be needed in this >>> change? >> Hi Jaikiran, >> >> Thanks for picking this up. There are some small wrinkles with this, >> but I hope nothing that will prevent it from happening. >> >> >> 2) Alan already mentioned this, but we need to consider the issue of >> idempotent close() or end() carefully. It's not strictly required, but >> it would be nice. I had taken a quick look at end() and I *thought* it >> was idempotent already, but a more careful analysis needs to be done. > I did some checks in the current JDK code. From what I see, the Inflater > and Deflater do not have any subclasses within the JDK itself.
To be clear - I couldn't find any subclasses in the main source code of JDK. There are a couple of subclasses in the tests (ConstructInflaterOutput.java and ConstructDeflaterInput.java), but those don't do much outside of the context of testing. -Jaikiran
