RFR: JDK-8146712 jdk/internal/jimage/JImageReadTest.java fails on all platforms

[Andreas Lundblad]

On Sat, Jan 09, 2016 at 03:31:34PM +0000, Alan Bateman wrote:
> On 09/01/2016 15:10, Jim Laskey (Oracle) wrote:
> >https://bugs.openjdk.java.net/browse/JDK-8146712
> >http://cr.openjdk.java.net/~jlaskey/8146712/webrev/index.html <http://cr.openjdk.java.net/~jlaskey/8146712/webrev/index.html>
> >
> >New resources must have been added. Since the native API filters all META-INF, the test should have filtered out all META-INF instead of just META-INF/services .
> >
> >META-INF/jvmci.providers/jdk.vm.ci.hotspot.amd64.AMD64HotSpotJVMCIBackendFactory
> >META-INF/jvmci.providers/jdk.vm.ci.hotspot.sparc.SPARCHotSpotJVMCIBackendFactory
> 
> The change looks okay but puzzling as to what triggered this. Did
> the sjavac changes really tickle this? I assume these resources have
> been there for some time and I wonder if we have a race or some bug
> in the build that meant they weren't there before. That is, I would
> have thought this test should have failed as soon the JVMCI changes
> got to jdk9/dev.
> 
> Also when you push this then you need to remove the test from the
> exclude list. I'm starting to suspect that Joe is collecting a
> bounty for every test that he locks up these days :-)

I'm not familiar with these type of issues, but the recent sjavac changes (and accompanying make file changes) did change how files are included/excluded and I guess it could potentially be related to what resources get included/excluded. As far as I know, these changes should however only be observable if you use --enable-sjavac (and that's not enabled by default yet).

Let me know if there's anything I can do to help diagnose this.

-- Andreas