On Fri, 19 May 2023 15:13:49 GMT, Daniel Fuchs <dfu...@openjdk.org> wrote:
>> Curiously, this method appears to never be called; at least
>>
>>
>> diff --git src/java.base/share/classes/sun/net/www/URLConnection.java
>> src/java.base/share/classes/sun/net/www/URLConnection.java
>> index b3af24c594b..f346cd8869b 100644
>> --- src/java.base/share/classes/sun/net/www/URLConnection.java
>> +++ src/java.base/share/classes/sun/net/www/URLConnection.java
>> @@ -253,14 +253,6 @@ public abstract class URLConnection extends
>> java.net.URLConnection {
>> REMIND */ ;
>> }
>>
>> - /**
>> - * Call this to close the connection and flush any remaining data.
>> - * Overriders must remember to call super.close()
>> - */
>> - public void close() {
>> - url = null;
>> - }
>> -
>> private static HashMap<String,Void> proxiedHosts = new HashMap<>();
>>
>> public static synchronized void setProxiedHost(String host) {
>>
>>
>> does not cause
>>
>>
>> make images
>>
>>
>> to fail, though it seems that the JDK source base does not currently enforce
>> use of `@Override`. (I did not find any overrides from a text search, and if
>> there were some without the annotation, they were apparently not calling
>> `super.close()`.)
>
> Interesting. If it's not called anywhere maybe we could investigate adding
> `throws IOException` to the signature...
That is also an option of course; would be binary compatible. Happy to take
that approach if you think it would be clearer.
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/12871#discussion_r1199090958
