On Fri, 2018-09-28 at 08:48 -0700, Erik Joelsson wrote: > Hello, > > Looks ok to me. > > It will look confusing on linux-x64 where it first says: > > checking if zgc should be built... yes > > But then quietly disables it anyway. Not a big issue and probably not > worth fixing given that zero on that platform is a rather rare > configuration (unless for testing purposes).
Hmm, wouldn't this only show up on a configure like this? $ configure --with-jvm-features=zgc --with-jvm-variants=zero [...] If so, who would really do this? So IMO, not worth fixing. Thanks, Severin > /Erik > > > On 2018-09-28 02:49, Aleksey Shipilev wrote: > > RFE: > > https://bugs.openjdk.java.net/browse/JDK-8211268 > > > > Webrev: > > http://cr.openjdk.java.net/~shade/8211268/webrev.01/ > > > > Not all GCs are supported by Zero, because it requires implementing the GC > > barriers in C++ > > interpreter. Now that we have jvm features that can selectively disable > > particular GCs, we can only > > build supported GCs. This avoids surprise build failures like these: > > https://bugs.openjdk.java.net/browse/JDK-8211228 > > > > Testing: x86_64 Zero build, running GCBasher with the rest of GCs > > > > Thanks, > > -Aleksey > > > >
