RFR: 7181175 Enable hotspot builds on Windows with MinGW/MSYS

[Tim Bell]

I wrote:

> Testing a hunch: I renamed my source tree to shorten the length of the 
> file paths.
>
> I have restarted my build in verbose mode, so I should know more soon.

The problem was self-inflicted.  It was not the path length this time.  
It also was not related to the"$(COMPILE_RMIC)" in sa.make.  Removing 
the "" works fine in sa.make.

It was a change I made on top of Volker's suggested fix in 
hotspot/make/windows/makefiles/rules.make

Instead of going with the suggested fix [1]:

RUN_JAVA="$(BootStrapDir)\bin\java"
RUN_JAVAP="$(BootStrapDir)\bin\javap"
RUN_JAVAH="$(BootStrapDir)\bin\javah"
RUN_JAR="$(BootStrapDir)\bin\jar"
COMPILE_JAVAC="$(BootStrapDir)\bin\javac" $(BOOTSTRAP_JAVAC_FLAGS)
COMPILE_RMIC="$(BootStrapDir)\bin\rmic"

I had changed it to:

RUN_JAVA=$(BootStrapDir)/bin/java
RUN_JAVAP=$(BootStrapDir)/bin/javap
RUN_JAVAH=$(BootStrapDir)/bin/javah
RUN_JAR=$(BootStrapDir)/bin/jar
COMPILE_JAVAC=$(BootStrapDir)/bin/javac $(BOOTSTRAP_JAVAC_FLAGS)
COMPILE_RMIC=$(BootStrapDir)/bin/rmic

This is an example where we actually have to preserve the \ in the paths 
to the tools because the RUN_foo commands are fired by nmake and not sh.

My test build is running now.  I'll update the webrev and send out a new 
URL.

Tim
[1]
http://mail.openjdk.java.net/pipermail/build-dev/2012-March/005729.html
http://cr.openjdk.java.net/~simonis/MinGW_MSYS_hotspot.v1/