Has anybody thought about how these changes will affect the builds on Windows?
I know that JEP 138 mentions a multi-core Linux machine as main target, but I think we should not completely forget about the Windows build because: - as of today, the Windows build are already significantly slower than the current Linux build - the Windows build already has very "esoteric" dependencies like special 'make' and 'cygwin' versions which are hard to fulfill (adding new ones like autoconf will probably worsen this situation) Further enhancing the *unix build is of course a nice thing but on the other hand that will further increase the gap to the Windows build. Moreover I'm afraid that the usage of autoconf will make it even harder than it is today to build on Windows. Currently the Windows build can be done with either MKS or Cygwin but using autoconf may limit this to Cygwin only which is not only considerably slower than MKS but also has sever problems on Windows 7 (see this mail thread: http://old.nabble.com/Is-anyone-able-to-build-on-Win-7-td33196055.html#a33199600) On the other hand changes like removing and compacting of makefiles, the elimination of make file recursion and especially the enhancements of JEP 139 (javac enhancements) will be highly beneficial on Windows as well. So I just wanted to suggest to not completely loose the Windows build out of sight and especially to make sure that the introduction of autoconf will have no additional negative impact on the Windows build. Regards, Volker On Tue, Feb 21, 2012 at 11:10 PM, <mark.reinh...@oracle.com> wrote: > Posted: http://openjdk.java.net/jeps/138 > > - Mark
