Need to use nawk on Solaris to avoid awk limitations

awk on Solaris has limits on input line length that can easily be reached during the build depending on the paths that are being used - see JDK-8006903 for one example.

Specifically this needs to be fixed in ./corba/make/common/shared/Defs-utils.gmk

Additionally when awk fails it does not cause make to fail at that point eg:

(cd /HUDSON/workspace/2-build-solaris-sparcv9@3/jdk8-profiles-staging/2956/build/solaris-sparcv9/corba/gensrc/tmp___com_sun_corba_se_PortableActivationIDL_activation.idl___; find . -type f | sed 's!\./!/HUDSON/workspace/2-build-solaris-sparcv9@3/jdk8-profiles-staging/2956/build/solaris-sparcv9/corba/gensrc/!g' | awk '{ print $1 ": /HUDSON/workspace/2-build-solaris-sparcv9@3/jdk8-profiles-staging/2956/corba/src/share/classes/com/sun/corba/se/PortableActivationIDL/activation.idl" }' > /HUDSON/workspace/2-build-solaris-sparcv9@3/jdk8-profiles-staging/2956/build/solaris-sparcv9/corba/gensrc/com/sun/corba/se/PortableActivationIDL/activation.idl.d)
awk: string too long near line 1
awk: syntax error near line 1
awk: illegal statement near line 1

instead the idl processing continues but then the corba classes never get compiled and classes.jar is not created and you eventually get the import failure in the JDK part of the build - as per JDK-8006903

This needs to be fixed