(no subject)

[weijun.wang]

Hi Martin

I've just changed my Linux box to a x86_64 one, and my first build of  
JDK 7 shows these errors:

...
cc1plus: warnings being treated as errors
/media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/
methodHandles_x86.cpp: In function ‘void  
trace_method_handle_stub(const char*, oopDesc*, intptr_t*, intptr_t*)’:
/media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/
methodHandles_x86.cpp:276: error: format ‘%016lx’ expects type ‘long  
unsigned int’, but argument 3 has type ‘void*’
/media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/
methodHandles_x86.cpp:276: error: format ‘%016lx’ expects type ‘long  
unsigned int’, but argument 4 has type ‘intptr_t*’
make[7]: *** [methodHandles_x86.o] Error 1
...

What am I missing here? and, do you have any private patches to share?

Thanks
Max

[Martin Buchholz-3]

Hi Max,

Sounds to me like you're using a very recent gcc.
My policy has been to use a stodgy gcc,
and let other folks use bleeding edge.
Fedora/Ubuntu hackers are more likely
to dogfood their gcc.... so check out the
icedtea patch set.

I wish that Openjdk release engineering maintained
buildability by a broad range of compilers
through regular testing.

Martin

On Thu, Oct 22, 2009 at 18:05, Max (Weijun) Wang <Weijun.Wang@...> wrote:

> Hi Martin
>
> I've just changed my Linux box to a x86_64 one, and my first build of JDK 7
> shows these errors:
>
> ...
> cc1plus: warnings being treated as errors
> /media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/methodHandles_x86.cpp:
> In function ‘void trace_method_handle_stub(const char*, oopDesc*, intptr_t*,
> intptr_t*)’:
> /media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/methodHandles_x86.cpp:276:
> error: format ‘%016lx’ expects type ‘long unsigned int’, but argument 3 has
> type ‘void*’
> /media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/methodHandles_x86.cpp:276:
> error: format ‘%016lx’ expects type ‘long unsigned int’, but argument 4 has
> type ‘intptr_t*’
> make[7]: *** [methodHandles_x86.o] Error 1
> ...
>
> What am I missing here? and, do you have any private patches to share?
>
> Thanks
> Max
>
>

[weijun.wang]

Martin Buchholz wrote:
> Hi Max,
>
> Sounds to me like you're using a very recent gcc.

Probably, I'm using the gcc of 'apt-get install' on this 9.04.

> My policy has been to use a stodgy gcc,
> and let other folks use bleeding edge.
> Fedora/Ubuntu hackers are more likely
> to dogfood their gcc.... so check out the
> icedtea patch set.

Ok, I'll go have a look.

Thanks
Max

>
> I wish that Openjdk release engineering maintained
> buildability by a broad range of compilers
> through regular testing.
>
> Martin
>
> On Thu, Oct 22, 2009 at 18:05, Max (Weijun) Wang <Weijun.Wang@...> wrote:
>> Hi Martin
>>
>> I've just changed my Linux box to a x86_64 one, and my first build of JDK 7
>> shows these errors:
>>
>> ...
>> cc1plus: warnings being treated as errors
>> /media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/methodHandles_x86.cpp:
>> In function ‘void trace_method_handle_stub(const char*, oopDesc*, intptr_t*,
>> intptr_t*)’:
>> /media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/methodHandles_x86.cpp:276:
>> error: format ‘%016lx’ expects type ‘long unsigned int’, but argument 3 has
>> type ‘void*’
>> /media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/methodHandles_x86.cpp:276:
>> error: format ‘%016lx’ expects type ‘long unsigned int’, but argument 4 has
>> type ‘intptr_t*’
>> make[7]: *** [methodHandles_x86.o] Error 1
>> ...
>>
>> What am I missing here? and, do you have any private patches to share?
>>
>> Thanks
>> Max
>>
>>

[weijun.wang]

Hi Martin

Since the error line is only a printf message, I just remove the line.
And now everything builds fine!

I'll see if the PTR_FORMAT macro is used in other places and why there's
no problem there.

Thanks
Max

Weijun Wang wrote:

>
> Martin Buchholz wrote:
>> Hi Max,
>>
>> Sounds to me like you're using a very recent gcc.
>
> Probably, I'm using the gcc of 'apt-get install' on this 9.04.
>
>> My policy has been to use a stodgy gcc,
>> and let other folks use bleeding edge.
>> Fedora/Ubuntu hackers are more likely
>> to dogfood their gcc.... so check out the
>> icedtea patch set.
>
> Ok, I'll go have a look.
>
> Thanks
> Max
>
>> I wish that Openjdk release engineering maintained
>> buildability by a broad range of compilers
>> through regular testing.
>>
>> Martin
>>
>> On Thu, Oct 22, 2009 at 18:05, Max (Weijun) Wang <Weijun.Wang@...> wrote:
>>> Hi Martin
>>>
>>> I've just changed my Linux box to a x86_64 one, and my first build of JDK 7
>>> shows these errors:
>>>
>>> ...
>>> cc1plus: warnings being treated as errors
>>> /media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/methodHandles_x86.cpp:
>>> In function ‘void trace_method_handle_stub(const char*, oopDesc*, intptr_t*,
>>> intptr_t*)’:
>>> /media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/methodHandles_x86.cpp:276:
>>> error: format ‘%016lx’ expects type ‘long unsigned int’, but argument 3 has
>>> type ‘void*’
>>> /media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/methodHandles_x86.cpp:276:
>>> error: format ‘%016lx’ expects type ‘long unsigned int’, but argument 4 has
>>> type ‘intptr_t*’
>>> make[7]: *** [methodHandles_x86.o] Error 1
>>> ...
>>>
>>> What am I missing here? and, do you have any private patches to share?
>>>
>>> Thanks
>>> Max
>>>
>>>

[John Rose-6]

We talked about this on the bsd-port mailing list.  It may be due to  
new error checks in GCC 4.2.

On Mac, even after those compile-time errors are fixed, GCC 4.2 does  
not build a usable hotspot.  After upgrading to Snow Leopard (10.6), I  
had to back the gcc to 4.0 for building hotspot.

HTH
-- John

On Oct 22, 2009, at 6:05 PM, Max (Weijun) Wang wrote:

> Hi Martin
>
> I've just changed my Linux box to a x86_64 one, and my first build  
> of JDK 7 shows these errors:
>
> ...
> cc1plus: warnings being treated as errors
> /media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/
> methodHandles_x86.cpp: In function ‘void trace_method_handle_stub
> (const char*, oopDesc*, intptr_t*, intptr_t*)’:
> /media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/
> methodHandles_x86.cpp:276: error: format ‘%016lx’ expects type ‘long  
> unsigned int’, but argument 3 has type ‘void*’
> /media/sda6/space/work/myjdk/hotspot/src/cpu/x86/vm/
> methodHandles_x86.cpp:276: error: format ‘%016lx’ expects type ‘long  
> unsigned int’, but argument 4 has type ‘intptr_t*’
> make[7]: *** [methodHandles_x86.o] Error 1
> ...
>
> What am I missing here? and, do you have any private patches to share?
>
> Thanks
> Max
>

[Christian Thalinger-3]

On Fri, 2009-10-23 at 14:32 +0800, Weijun Wang wrote:
> Hi Martin
>
> Since the error line is only a printf message, I just remove the line.
> And now everything builds fine!
>
> I'll see if the PTR_FORMAT macro is used in other places and why there's
> no problem there.

This line is part of the new JSR 292 code and a fix for the build issue
will be commited (with all other JSR 292 stuff) before M5.  Sorry for
the inconvenience.

-- Christian