This is the mail archive of the java-patches@gcc.gnu.org mailing list for the Java project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Patch for Preview: Cross-Configury Tweaks


Ranjit,

I am running into this issue when using this patch, I try to
do a self-hosted build on mingw32.

>+   if test -x "${builddotdot}/../../gcc/gcj${ac_cv_exeext}"; then
>+      if test x"$build" = x"$host"; then
>+         dir="`cd ${builddotdot}/../../gcc && ${PWDCMD-pwd}`"
>+         GCJ="$dir/gcj -B`${PWDCMD-pwd}`/ -B$dir/"
>+      else
>+         # Crossed-native build (host=target and host!=build)
>+         CANADIAN=yes
>+         NULL_TARGET=no
>+         GCJ="${target_alias}-gcj -B`${PWDCMD-pwd}`/"
>+      fi
>    else
>       CANADIAN=yes

The mingw32 compiler used to do the build was configured (on Linux)
with target = i386-pc-mingw32.

When I do the configure on Win32 for the compiler I'm about to build,
config.guess reports build=i686-pc-mingw32.

Therefore, since build!=host, we are assuming a crossed-native
build and the wrong gcj, etc. get picked up.

To remedy this, I explicitly set BUILD=mingw32 rather than
BUILD=`$GCC_SRC_DIR/config.guess`.

What is your take on this? Are the semantics of a compiler
building a compiler such that config.guess should not be used
under these circumstances, or is the patch's logic faulty?

-- Mohan
http://www.thisiscool.com/
http://www.animalsong.org/





Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]