Building on non-std architectures
Dave Pointon
pointo1d at gmail.com
Fri Oct 17 14:08:15 UTC 2014
Hi all,
Further to my previous e-mail, I've implemented a strawman IBM specific fix
(which bears all the hallmarks of working:-) as follows ...
diff --git a/common/autoconf/platform.m4 b/common/autoconf/platform.m4
--- a/common/autoconf/platform.m4
+++ b/common/autoconf/platform.m4
@@ -519,6 +519,12 @@
AC_CHECK_SIZEOF([int *], [1111])
TESTED_TARGET_CPU_BITS=`expr 8 \* $ac_cv_sizeof_int_p`
+
+ if test "x$ZERO_ARCHDEF" = xS390; then
+ if test "x$OPENJDK_TARGET_CPU_BITS" = x31; then
+ TESTED_TARGET_CPU_BITS=`expr $TESTED_TARGET_CPU_BITS - 1`
+ fi
+ fi
if test "x$TESTED_TARGET_CPU_BITS" != "x$OPENJDK_TARGET_CPU_BITS";
then
AC_MSG_ERROR([The tested number of bits in the target
($TESTED_TARGET_CPU_BITS) differs from the number of bits expected to be
found in the target ($OPENJDK_TARGET_CPU_BITS)])
Any advice etc. to make it [the above fix] more generically applicable will
be gratefully received.
TIA & rgds ,
--
Dave Pointon FIAP MBCS
Now I saw, tho' too late, the folly of beginning a work before we count the
cost and before we we judge rightly of our strength to go thro' with it -
Robinson Crusoe
On 1 October 2014 11:44, pointo1d <pointo1d at gmail.com> wrote:
> Hiya Erik ,
>
> On 01/10/14 09:10, Erik Joelsson wrote:
>
>> This happens in common/autoconf/platform.m4 in
>> PLATFORM_SETUP_OPENJDK_TARGET_BITS. We make the assumption that a
>> platform is either 32 or 64 bit. The test that is failing is trying to make
>> sure that what's produced by the compiler matches the target cpu. I think
>> it will need some tweaking to handle 31 bit platforms.
>>
>> /Erik
>>
>> <snip>
>>
>
> TFT, I have to say that was the conclusion at which I arrived by CoP
> yesterday - now to play ...
>
>
> --
> Dave Pointon FIAP MBCS - Contractor engaged by IBM
>
> Now I saw, tho' too late, the folly of beginning a work before we count
> the cost and
> before we we judge rightly of our strength to go thro' with it - Robinson
> Crusoe
>
>
More information about the build-dev
mailing list