RFR: 8160730: [JVMCI] compiler selection should work without -Djvmci.Compiler

[Christian Thalinger]

> On Jul 1, 2016, at 12:28 PM, Doug Simon <doug.simon at oracle.com> wrote:
> 
> Currently, JVMCI compiler selection requires setting the jvmci.Compiler system property. This webrev removes the requirement for explicitly setting the jvmci.Compiler system property if there is only one compiler factory available or if the set of available compiler factories has an inherent selection preference total ordering. It also revives 
> JvmciNotifyInstallEventTest and modifies it to test that auto-selection of a JVMCI compiler works in the case where there’s exactly one compiler factory available.
> 
> 
> https://bugs.openjdk.java.net/browse/JDK-8160730
> http://cr.openjdk.java.net/~dnsimon/8160730/ <http://cr.openjdk.java.net/~dnsimon/8160730/>

src/jdk.vm.ci/share/classes/jdk.vm.ci.hotspot/src/jdk/vm/ci/hotspot/HotSpotJVMCICompilerConfig.java

 <>+         * Compares two compiler factories and returns -1 if {@code o1} should be auto-selected over
+         * {@code o2}, -1 if {@code o1} should be auto-selected over {@code o2} or 0 if
The -1 case is documented twice and 1 is not.

Otherwise this looks good.  I like that change; needing to specify the compiler is annoying.

> 
> -Doug

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/attachments/20160701/906d3e25/attachment.html>