[PPL-devel] [Fwd: Re: PPL on a Power6 running AIX]

[Roberto Bagnara]

-------- Original Message --------
Subject: Re: [PPL-devel] PPL on a Power6 running AIX
Date: Thu, 10 Sep 2009 06:56:57 -0600
From: Davide Del Vento <ddvento at mail.ucar.edu>
To: Roberto Bagnara <bagnara at cs.unipr.it>
References: <4AA7F256.10007 at ucar.edu> <4AA8A0AB.4070001 at cs.unipr.it>

Thanks a lot for your message!
xlc and xlC are the IBM's compilers for c and c++ respectively.
They do support several standards:
http://publib.boulder.ibm.com/infocenter/lnxpcomp/v8v101/index.jsp?topic=/com.ibm.xlcpp8l.doc/getstart/overview/gs_ovr.htm

and sometimes are more nitpicking than gcc about the standards,
as I believe it's happening here. I'll try to figure out
(I am not fluent about C++ specs) and let you know.

Compiling with gcc is possible, but there will be several
linking problems with other libraries that I eventually
need to link together (gcc binaries are not-linkable with
xlc binaries).

Thanks again
;Dav

On Thu, 10 Sep 2009 08:46:03 +0200
  Roberto Bagnara <bagnara at cs.unipr.it> wrote:
> Davide Del Vento wrote:
>> Folks,
>> I am trying to compile the latest PPL on bluefire, an IBM Power6 
>>running
>> AIX 5.3, with IBM's compiler xlc/xlC
>> (http://www.cisl.ucar.edu/docs/bluefire/be_quickstart.html#software)
>> 
>> Unfortunately, while the configure script is happy (see config.log 
>>for
>> details), the compilation fails as described in the attached 
>>make.log
>> and I have no clues about it.
>> 
>> Note that before it was failing with the following one, which I 
>>believe
>> was due to a different make version being used, but might be the 
>>first
>> culprit (I have the make.log for this too but I'm not sending it 
>>right now):
>> 
>> /usr/bin/sed -f ./ppl-config.sed  >ppl-config.h
>> "../ppl-config.h", line 1.0: 1540-0809 (W) The source file is empty.
>> 
>> Do you have any recommendations about what I can do to fix the 
>>issue?
> 
> Hi Davide,
> 
> looking at config.log, I gather you are using C and C++ compilers
> called xlc and xlC, respectively.  We never compiled the PPL using 
>those,
> but this should not be a problem as long as they implement a large 
>enough
> fragment of the C and C++ standards.
> 
> Looking at make.log, I see something strange is happening.  The 
>first errors,
> that is,
> 
> "Row.defs.hh", line 88.46: 1540-0063 (S) The text "Row_Impl_Handler" 
>is unexpected.
> "Linear_Row.defs.hh", line 124.52: 1540-0063 (S) The text "Row" is 
>unexpected.
> "Linear_Expression.defs.hh", line 236.60: 1540-0063 (S) The text 
>"Linear_Row" is unexpected.
> "Linear_System.defs.hh", line 54.55: 1540-0063 (S) The text "Matrix" 
>is unexpected.
> "Constraint_System.defs.hh", line 127.60: 1540-0063 (S) The text 
>"Linear_System" is unexpected.
> "Constraint.defs.hh", line 260.53: 1540-0063 (S) The text 
>"Linear_Row" is unexpected.
> "Congruence_System.defs.hh", line 120.60: 1540-0063 (S) The text 
>"Matrix" is unexpected.
> 
> would be explained if you were including some funny header file 
>doing
> 
> #define private
> #define public
> 
> I know, this is crazy, but how else could the text 
>"Row_Impl_Handler"
> be unexpected in the lines
> 
> class Parma_Polyhedra_Library::Row : private Row_Impl_Handler {
> 
> and
> 
> class Parma_Polyhedra_Library::Linear_Row : public Row {
> 
> ?
> 
> Do you have the option of compiling with GCC 4.0.3 or later version?
> Otherwise, please use the -E option of your compiler to inspect the
> preprocessor output, and check how the two lines above are 
>transformed.
> 
>> PS: I speak Italian, but I'm not sure if this list is read just by
>> Italian, despite its .it domain.
> 
> This is an English-language mailing list indeed.
> All the best,
> 
>   Roberto
> 
> -- 
> Prof. Roberto Bagnara
> Computer Science Group
> Department of Mathematics, University of Parma, Italy
> http://www.cs.unipr.it/~bagnara/
> mailto:bagnara at cs.unipr.it


-- 
Prof. Roberto Bagnara
Computer Science Group
Department of Mathematics, University of Parma, Italy
http://www.cs.unipr.it/~bagnara/
mailto:bagnara at cs.unipr.it