Re: [boost] [multi_index] regression with Intel 8.0 for Linux.

From: Joaquín Mª López Muñoz (joaquin_at_[hidden])
Date: 2004-07-26 10:22:25

Richard Hadsell ha escrito:

> Joaquín Mª López Muñoz wrote:
>
> >One out of the 18 tests is making ICC 8.0 go wild, please see
> >http://tinyurl.com/6u4og
> >
> >
> Do you have the latest patch (l_cc_pc_8.0.066_pe070.1.tar.gz)?  It was
> announced this weekend.  If not, it would be nice to eliminate known
> compiler bugs before pursuing the problem.
>

Wait, seems like this is a spurious problem.
Tests have just been updated (13:21:03 UTC). test_update works OK
again, and it is now test_capacity that fails with the same segfault
on compilation.
I've searched for this problem throughout the reports, the following
tests are affected:

dynamic_bitset - dyn_bitset_unit_tests4 - intel-8.0-linux
math - octonion_test - intel-7.1-linux
multi_index - test_capacity - intel-8.0-linux

(note that the second crash is with Intel 7.1 instead of 8.0)
Seems to me like the machine is low on resources and Intel compiler
occasionally gives up. I cannot verify this, of course, but it is the only
reasonable explanation I can come up with.

Joaquín M López Muñoz
Telefónica, Investigación y Desarrollo