[csw-maintainers] Fwd: java error, core dumped, when building subversion-1.6.11
rupert at opencsw.org
Mon May 10 20:22:47 CEST 2010
On Mon, May 10, 2010 at 15:56, Dagobert Michelsen <dam at opencsw.org> wrote:
> Hi Rupert,
> Am 09.05.2010 um 10:05 schrieb rupert THURNER:
>> ---------- Forwarded message ----------
>> From: Stefan Sperling <stsp at elego.de>
>> Date: Sat, May 8, 2010 at 12:39
>> Subject: Re: java error, core dumped, when building subversion-1.6.11
>> To: "rupert.thurner" <rupert.thurner at gmail.com>
>> Cc: dev at subversion.apache.org
>> On Fri, May 07, 2010 at 10:57:05PM -0700, rupert.thurner wrote:
>>> when i try to build subversion-1.6.11 on solaris 9 the below error
>>> occurs. how could one track that down?
>> /usr/jdk1.6.0_07/bin/javah -force -d subversion/bindings/javahl/
>>> include -classpath subversion/bindings/javahl/classes:
>> gmake: *** [subversion/bindings/javahl/include/
>>> org_tigris_subversion_javahl_PropertyData.h] Illegal Instruction (core
>> My guess is that your java or GNU make binary is not 100% compatible with
>> your processor architecture.
> I am pretty sure that this is not the case - the jdk is the official
> binary from Sun, and the T2 CPU is fully supported.
> I've seen such errors with a self-compiled FreeBSD system, optimized for
>> Athlon XP at compile time, after moving the hard drive into an old
>> dual-Celeron box (this was before dual-core CPUs were available) because
>> the Athlon's mainboard had decided to die.
>> The result was that most programs worked fine, but anything trying to
>> use stuff like advanced floating point instructions was failing with
>> illegal instruction errors. Back then I was lucky enough to be able to
>> re-compile the system in place, this time without optimisations (and I
>> have given up on playing with CFLAGS ever since).
>> So I'd double check with the source of your java / gnu make binaries
>> to make sure they are compatible with your OS and processor.
> Is the corefile really from gnumake?
> i do not know.
a much simpler testcase producing the same result might be compiling berkely
db 5.0. there only configure and a simple java test is involved until the
configure:17407: javac Test.java
configure:17410: $? = 0
configure:17420: java Test
../dist/configure: line 1: 27179 Illegal Instruction (core dumped) $JAVA
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the maintainers