Hi Hubert,

Looks like the Sandia-vacpp-10.1 tester ran last night using gcc to build bjam.  It's beginning to look like a problem with xlc at high optimization, does that sound plausible?

-- Noel

On Mar 28, 2012, at 8:52 AM, Hubert Tong wrote:

Hi Noel,

The bjam built with GCC on AIX and Linux PPC does not core immediately.
I have not confirmed if they will work fine for a full run of the tests.

I will try to isolate the issue if time permits.

Thanks,


Hubert Tong

Hubert S K Tong
IBM Canada Ltd, C2/YGH/8200/MKM
8200 Warden Ave, Markham ON L6G 1C7
Canada
E-mail: hstong@ca.ibm.com

<graycol.gif>"Belcourt, K. Noel" ---27-03-2012 09:43:47 PM---Hi Steven, On Mar 27, 2012, at 5:13 PM, Steven Watanabe wrote:

From: "Belcourt, K. Noel" <kbelco@sandia.gov>
To: steven@providere-consulting.com, "Boost.Build developer's and user's list" <boost-build@lists.boost.org>
Date: 27-03-2012 09:43 PM
Subject: Re: [Boost-build] [EXTERNAL] Re: Boost build vacpp on Linux (PPC)
Sent by: boost-build-bounces@lists.boost.org





Hi Steven,

On Mar 27, 2012, at 5:13 PM, Steven Watanabe wrote:

> On 03/27/2012 03:23 PM, Hubert Tong wrote:
>>
>> We have been getting this issue as well on both AIX and Linux on PPC.
>
> Okay, so this appears for xlc/PPC, right?

Yes.

> Does it work for other compilers on PPC?

Yes, it works fine with gcc-4.1.2.  The problem is specific to xlc, as  
in:

./build.sh vacpp

produces

bin.unknownppc/bjam

which cores.  Perhaps Hubert can confirm that gcc builds a good bjam?  
Steven, I lack the time tonight to debug this but will change over our  
tests to use gcc to compile bjam and that should be a short term  
workaround.  I'll attempt to look at this over the weekend and answer  
your questions then.

Thanks for the help.

-- Noel


_______________________________________________
Unsubscribe & other changes:
http://lists.boost.org/mailman/listinfo.cgi/boost-build


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost-build