RFR (M) JDK-8202605: Standardize on ClassLoaderData::loader_name() throughout the VM to obtain a class loader's name

[Lois Foltan]

Please review this updated webrev that address review comments received.

http://cr.openjdk.java.net/~lfoltan/bug_jdk8202605.1/webrev/

Thanks,
Lois

On 6/13/2018 6:58 PM, Lois Foltan wrote:
> Please review this change to standardize on how to obtain a class 
> loader's name within the VM.  SystemDictionary::loader_name() methods 
> have been removed in favor of ClassLoaderData::loader_name().
>
> Since the loader name is largely used in the VM for display purposes 
> (error messages, logging, jcmd, JFR) this change also adopts a new 
> format to append to a class loader's name its identityHashCode and if 
> the loader has not been explicitly named it's qualified class name is 
> used instead.
>
> 391 /**
> 392 * If the defining loader has a name explicitly set then
> 393 * '<loader-name>' @<id>
> 394 * If the defining loader has no name then
> 395 * <qualified-class-name> @<id>
> 396 * If it's built-in loader then omit `@<id>` as there is only one 
> instance.
> 397 */
>
> The names for the builtin loaders are 'bootstrap', 'app' and 'platform'.
>
> open webrev at http://cr.openjdk.java.net/~lfoltan/bug_jdk8202605/webrev/
> bug link at https://bugs.openjdk.java.net/browse/JDK-8202605
>
> Testing: hs-tier(1-2), jdk-tier(1-2) complete
>                hs-tier(3-5), jdk-tier(3) in progress
>
> Thanks,
> Lois
>