[8u] TLSv1.3 RFR: 8245653: Remove 8u TLS tests
Martin Balao
mbalao at redhat.com
Tue Aug 11 15:30:25 UTC 2020
Hi Alexey,
On 8/11/20 11:11 AM, Alexey Bakhtin wrote:
>>
>> * test/jdk/ProblemList.txt (modified)
>> * do we need to black-list those in JDK-8?
> Agree. It should be updated but not at this step.
>
>> * test/jdk/com/sun/jndi/ldap/DeadSSLLdapTimeoutTest.java (modified)
>> * I suggest including this change.
> OK. Make sense. Removed in this step
>> * test/jdk/java/net/httpclient/MockServer.java (modified)
>> * Not strictly necessary.
>> * test/jdk/sun/security/ec/TestEC.java (modified)
>> * Do we need this change?
> OK. Update of this test is included into the next step, but it would be better to remove it and add in the next step. Removed
>> * test/jdk/sun/security/pkcs11/KeyStore/ClientAuth.java (modified)
>> * Do we need this change?
>> * test/jdk/sun/security/pkcs11/KeyStore/ClientAuth.sh (modified)
>> * Do we need this change?
> Agree. Tests removed
>> * test/jdk/sun/security/tools/keytool/PrintSSL.java (modified)
>> * Looks like nice to have.
> Agree. Test removed
>> If you agree, we can handle this in a separated step. Also, in Step 11 I
>> expect nothing but new files from JDK 11.0.7 without any changes. If you
>> need modifications on top of these files, please include them on a
>> separated step -so it's much easier for me to review-.
>>
> I think we can handle it in this step and add new version of the tests in Step 11
>
I've some concerns regarding the file-replacement approach for these
test files out of SSL categories. The reason is that together with the
file-replacement we are bringing other changes not related to SSL, which
may cause test failures or miss-documentation. For example, TestEC.java
introduces a reference to 8168078 which is not even fixed in JDK-8 and a
run-command with different arguments -which I guess is to test 8168078-.
Another example is ClientAuth.java, which changes the way in which the
SunPKCS11 security provider instance is created and loaded -is the
mechanism used supported in JDK-8?-. In ClientAuth.sh most of the
changes are tabs replacement but there are a few other that we would
need to analyze. PrintSSL seems to bring many changes, which I'm certain
that will require rework because 'jdk.test.lib.process.OutputAnalyzer'
import is not the same in JDK-8. printssl.sh deletion was unintended?
Even if we create a new step to revert these changes, I believe it's
easier and more clean (and easier) to apply 8196584 changes affecting
these files in a separated step, particularly because these changes are
very small.
As a result, I'm inclined to approve your Webrev.03. Does it make sense
to you?
Thanks,
Martin.-
More information about the jdk8u-dev
mailing list