RFR [11u]: 8160768: Add capability to custom resolve host/domain names within the default JNDI LDAP provider

[Langer, Christoph]

Hi Goetz,

thanks for your review.

I tried to address your concerns regarding the class comment for java.naming/share/classes/com/sun/jndi/ldap/LdapDnsProviderServiceInternal.java and it's hopefully a bit easier to read and understand now:
http://cr.openjdk.java.net/~clanger/webrevs/8160768.11u.1/

Ok for you?

Thanks
Christoph

> -----Original Message-----
> From: Lindenmaier, Goetz <goetz.lindenmaier at sap.com>
> Sent: Freitag, 28. August 2020 17:15
> To: Langer, Christoph <christoph.langer at sap.com>; jdk-updates-
> dev at openjdk.java.net
> Cc: Osipov, Michael <michael.osipov at siemens.com>
> Subject: RE: RFR [11u]: 8160768: Add capability to custom resolve
> host/domain names within the default JNDI LDAP provider
> 
> Hi Christoph,
> 
> You just pinged for a review, but I was looking at it anyways.
> 
> I think this looks good. The basic functionality is the same,
> you added functionality to access LdapDnsProviderService
> without having it in the same module.
> Our jck tests are green with the change as you implemented it.
> 
> One nit, I think the comment here is too complicated, maybe
> shorten the sentences a bit?
> http://cr.openjdk.java.net/~clanger/webrevs/8160768.11u.0/src/java.namin
> g/share/classes/com/sun/jndi/ldap/LdapDnsProviderServiceInternal.java.ht
> ml
> 
>   34  * It is implemented by jdk.naming.ldap and the singleton instance is
> registered
>   35  * in LdapCtxFactory to which provides access to the service mechanism
>   36  * of jdk.naming.ldap/com.sun.jndi.ldap.spi for module java.naming.
> -->
>   34  * It is implemented by jdk.naming.ldap. LdapDnsProviderService. A
> singleton instance is registered
>   35  * in LdapCtxFactory. This singleton provides access to the service
> mechanism
>   36  * of jdk.naming.ldap/com.sun.jndi.ldap.spi for module java.naming.
> 
> Best regards,
>   Goetz.
> 
> > -----Original Message-----
> > From: jdk-updates-dev <jdk-updates-dev-retn at openjdk.java.net> On
> Behalf
> > Of Langer, Christoph
> > Sent: Tuesday, August 11, 2020 5:52 PM
> > To: jdk-updates-dev at openjdk.java.net
> > Cc: Osipov, Michael <michael.osipov at siemens.com>
> > Subject: [CAUTION] RFR [11u]: 8160768: Add capability to custom resolve
> > host/domain names within the default JNDI LDAP provider
> >
> > Hi,
> >
> > I've been working on backporting JDK- 8160768: "Add capability to custom
> > resolve host/domain names within the default JNDI LDAP provider" to
> > OpenJDK 11u. This seems to be quite an important enhancement for usage
> in
> > more complex LDAP setups so there is a certain demand for this backport.
> > Oracle has backported it to their JDK8 and 11 releases already.
> >
> > Backporting is not so trivial, though, as the original change introduced a
> new
> > package with two classes in the javax namespace: javax.naming.ldap.spi
> with
> > classes LdapDnsProvider and LdapDnsProviderResult. To facilitate
> > backporting without having to change the spec, this backport creates
> > package com.sun.jndi.ldap.spi instead of javax.naming.ldap.spi to provide
> > the new service interface. So people wanting to leverage the feature in a
> JDK
> > < 12 will have to build against the com.sun.jndi.ldap.spi implementation.
> > Furthermore, we can't publicly export package com.sun.jndi.ldap.spi from
> > module java.naming as this is not allowed for a module in the java
> > namespace. So I introduced a new module jdk.naming.ldap for being able
> to
> > export the new service interface. The service is hooked into java.naming's
> > class LdapCtxFactory by reflective lookup of class
> > com.sun.jndi.ldap.dns.LdapDnsProviderService and triggering its
> initialization.
> > This in turn will register the LdapDnsProviderServiceInternal singleton for
> > accessing com.sun.jndi.ldap.spi service implementations from java.naming.
> If
> > a JDK was jlink'ed without module jdk.naming.ldap, the missing
> > LdapDnsProvider facility would then just be ignored quietly.
> >
> > In my backport I also include the adaptations to the test from the already
> > backported change of JDK-8139965 to improve test stability.
> >
> > I had to create a backport CSR as well for which I'd also need a review.
> >
> > Bug: https://bugs.openjdk.java.net/browse/JDK-8160768
> > Original CSR: https://bugs.openjdk.java.net/browse/JDK-8192975
> > Original Change: http://hg.openjdk.java.net/jdk/jdk/rev/a609d549992a
> > Backport CSR: https://bugs.openjdk.java.net/browse/JDK-8251380
> > Webrev: http://cr.openjdk.java.net/~clanger/webrevs/8160768.11u.0/
> >
> > Thanks in advance for looking at this.
> >
> > Best regards
> > Christoph