[teststabilization] RFR: 8229481: sun/net/www/protocol/https/ChunkedOutputStream.java failed with a SSLException
Michael McMahon
michael.x.mcmahon at oracle.com
Thu Aug 15 12:57:14 UTC 2019
Would it be clearer if test0() of chunkedoutputstream was renamed to
testPlainText() and it is given
the authority component only. So, it would create the URL itself and it
would be clearer that it
is creating a http url?
Minor comment typo in "veryfy"
Otherwise, looks good.
- Michael.
On 14/08/2019 16:00, Daniel Fuchs wrote:
> Michael reminded me that the first byte read off a TLS connection
> is supposed to be 0x16.
>
> So here is a cleaner alternative: a
>
> http://cr.openjdk.java.net/~dfuchs/webrev_8229481/webrev.02
>
> best regards,
>
> -- daniel
>
> On 13/08/2019 16:31, Daniel Fuchs wrote:
>> Hi,
>>
>> Please find below a fix for:
>>
>> 8229481: sun/net/www/protocol/https/ChunkedOutputStream.java
>> failed with a SSLException
>> [1] https://bugs.openjdk.java.net/browse/JDK-8229481
>>
>> webrev:
>> [2] http://cr.openjdk.java.net/~dfuchs/webrev_8229481/webrev.01/
>>
>> The test was observed failing once with a server side exception:
>>
>> javax.net.ssl.SSLException: Unrecognized SSL message, plaintext
>> connection? (see [1])
>>
>> The most probable cause is that the test server received some
>> rogue connection from another entity.
>>
>> The patch adds a hook to the server code that the test
>> code can override to simply ignore such rogue connections.
>>
>> best regards,
>>
>> -- daniel
>
More information about the net-dev
mailing list