The current jdk9/hs-comp forest is broken (original) (raw)

Volker Simonis volker.simonis at gmail.com
Fri Feb 14 02:41:26 PST 2014

On Thu, Feb 13, 2014 at 8:53 PM, Vladimir Kozlov <vladimir.kozlov at oracle.com> wrote:

"The script" is cron job which runs once per week on Friday to push forest changes from hotspot main repo jdk9/hs into our groups repos. It is very simple and bailout if a push requires a merge. I think it bailed out after push into hs-comp/jdk failed and did not finished push into the rest repositories. I see nashorn/ repo was also not updated. I am just speculating. Alejandro should know.

Volker, can you wait 1 day? Tomorrow will be another merge from hs/ to hs-comp/ which will fail again. I will do final merge after it.

That's fine for me. Please don't forget to also push:

http://cr.openjdk.java.net/~simonis/webrevs/8034797

into jdk9/hs-comp after you pulled from hs/ because that pull will bring in "8028280: ParkEvent leak when running modified runThese which only loads classes" which must be fixed by 8034797 on AIX.

And it would be nice if you could give us a chance to check port in jdk9/hs-comp before pulling it down to jdk9/dev to avoid a broken port in jdk9/dev.

Thanks, Volker

Thanks, Vladimir

On 2/13/14 10:08 AM, Volker Simonis wrote:

On Thursday, February 13, 2014, Vladimir Kozlov <vladimir.kozlov at oracle.com <mailto:vladimir.kozlov at oracle.com>> wrote: Alejandro's auto script pushed 8026773 changes into hs-comp forest. We forgot to disable it. I think 8026773 was not pushed into all repos because those repos have new ppc changes: 8033154 - autoscript skips such repos. But 8033154 is only in jdk not in langtools. So this wouldn't explain why langtools hasn't been updated. And just out of interest - what is "Alejandro's auto script" all about? What does it do and how does it work? I haven't been aware of this until now. I will pull the rest of forest today. I don't see any conflicts. OK, thanks. Thanks, Vladimir On 2/13/14 5:30 AM, Volker Simonis wrote: Hi, currently the jdk9/hs-comp forst can not be build. The build fails with the following build error in the Corba build: Compiling 141 files for BUILDIDLJ _(/share/software/Java/jdk1.7.009/bin/java -Xms64M -Xmx1100M -XX:ThreadStackSize=1536 _"-Xbootclasspath/p:/tmp/_output-hs-comp/langtools/dist/interimlangtools.jar" _-cp /tmp/output-hs-comp/langtools/dist/interimlangtools.jar com.sun.tools.javac.Main -source 7 -target 7 -bootclasspath _"/share/software/Java/jdk1.7._009/jre/lib/rt.jar:/share/_software/Java/jdk1.7.009/lib/tools.jar" _-Xlint:all,-deprecation,-_unchecked,-serial,-_fallthrough,-cast,-rawtypes,-static,-dep-ann -implicit:none -sourcepath _"/tmp/hs-comp/corba/src/share/classes" -d _/tmp/output-hs-comp/corba/idljclasses _@/tmp/output-hs-comp/corba/idljclasses/the.BUILDIDLJ_batch.tmp && /bin/mv _/tmp/output-hs-comp/corba/idljclasses/the.BUILDIDLJ_batch.tmp _/tmp/output-hs-comp/corba/idljclasses/the.BUILDIDLJ_batch) Error: Could not find or load main class com.sun.tools.javac.Main I think this is related to "8026773: Failing compilation in Corba does not fail the build" which (I suppose) must be applied synchronously to the top-level repository as well as to the jdk/, corba/ and langtools/ repositories/. But for some reason, this change has only be pulled into the toplevel and the corba/ repositories in the hs-comp forest. Updating the top-level and corba repositories to the change right before 8026773 will result in a sane forest. Regards, Volker

More information about the hotspot-compiler-dev mailing list