RFR: JDK-8010767: Build fails on OEL6 with 16 cores

Hello, Please review this backport from jdk9 to jdk8u. The problem of hotspot ergonomics wrecking havoc in the build on very large machines is hitting us in jdk8u builds too, when running bootcycle-images. This patch introduces two versions of the JAVA_FLAGS from configure, one for large java

Re: RFR: JDK-8010767: Build fails on OEL6 with 16 cores

On 2014-10-10 15:55, Erik Joelsson wrote: Hello, Please review this backport from jdk9 to jdk8u. The problem of hotspot ergonomics wrecking havoc in the build on very large machines is hitting us in jdk8u builds too, when running bootcycle-images. This patch introduces two versions of the JAV

Re: RFR: JDK-8010767: Build fails on OEL6 with 16 cores

Hi Erik On 10/10/14 07:15, Magnus Ihse Bursie wrote: On 2014-10-10 15:55, Erik Joelsson wrote: Hello, Please review this backport from jdk9 to jdk8u. The problem of hotspot ergonomics wrecking havoc in the build on very large machines is hitting us in jdk8u builds too, when running bootcycle

Unused DEBUG_CLASSFILES variable in makefiles

Hi, The DEBUG_CLASSFILES variable seems to be completely unused but still defined in the makefiles in jdk9 and jdk8u. $ grep -rn DEBUG_CLASSFILES Binary file jaxp/.hg/store/data/make/_makefile.i matches Binary file jaxws/.hg/store/data/make/_makefile.i matches build/config.status:876:S["DEBUG_CLA

GNU Make 4.0

Are there any known issues with using GNU Make 4.0 to build OpenJDK8? I upgraded CygWin in response to paranoia over Bash/Shellshock. Along the way I got Make 4.0. The subsequent build failed very early with: BuildLangtools.gmk:204: *** missing separator. Stop. I rolled the make.exe (and no

Re: GNU Make 4.0

* Salter, Thomas A [2014-10-10 14:23]: > Are there any known issues with using GNU Make 4.0 to build OpenJDK8? Yes. I had to add this patch in Fedora to build with Make 4.0: http://pkgs.fedoraproject.org/cgit/java-1.8.0-openjdk.git/tree/hotspot-build-j-directive.patch Cheers, Omair -- PGP Key: