Could this also explain what we see in some of our nightlies? The builds broke on some platforms for us (e.g. linux ppcle). It seems that "@LIBZ_CFLAGS@" does not get expanded correctly and finds its way into the makefile, confusing the compiler:
( /sapmnt/appl_sw/gcc-4.8/bin/gcc ... -g1 -fPIC @LIBZ_CFLAGS@ ... -o /net/ usr.work/openjdk/nb/linuxppc64le/nightly/output-jdk-fastdebug/support/native/java.base/libzip/Adler32.o /net/ usr.work/openjdk/nb/linuxppc64le/nightly/jdk/src/java.base/share/native/libzip/Adler32.c gcc: error: @LIBZ_CFLAGS@: No such file or directory ? On Tue, Feb 6, 2018 at 10:47 PM, Tim Bell <tim.b...@oracle.com> wrote: > Erik: > > A small typo managed to sneak itself into my previous patch, causing the >> build to fail on Windows. >> >> Bug: https://bugs.openjdk.java.net/browse/JDK-8196911 >> > [...snip...] > >> - if test "x$USER_EXTERNAL_LIBZ" = "xfalse"; then >> + if test "x$USE_EXTERNAL_LIBZ" = "xfalse"; then >> > > Looks good. I missed that one myself as a reviewer. > > /Tim > >
