[OMPI users] Corrupt documentation file in vampirtrace/opari

[number.cruncher]

When building a 1.3.4 RPM on Fedora 11, I get the following error about a 
unreadable compressed PostScript file:

error: Recognition of file 
"/home/build/rpmbuild/BUILDROOT/openmpi-1.3.4-1.x86_64/opt/build/share/vampirtrace/doc/opari/lacsi01.ps.gz"
 failed: mode 100644 zlib: invalid distance too far backempty (gzip compressed 
data, was "lacsi01.ps", from Unix, last modified: Thu Aug 30 16:43:24 2001, max 
compression)

I can't read this file; can it be removed or corrected?
Thanks,
Simon

Re: [OMPI users] memcpy overlap in ompi_ddt_copy_content_same_ddt and glibc 2.12

[e-mail number.cruncher]

On 10 November 2010 17:25, Jed Brown  wrote:
>
> Is the memcpy-back code ever executed when called as memcpy()?  I can't
> imagine why it would be, but it would make plenty of sense to use it inside
> memmove when the destination is at a higher address than the source.
> Jed

Oh yes. And, after further reading, it seems I'm not alone. There's an
article at http://lwn.net/Articles/414467/ about the new glibc changes
and a bug at https://bugzilla.redhat.com/show_bug.cgi?id=638477
relating directly to memcpy. The Fedora bug is particularly insightful
with some interesting contributions from Linus Torvalds himself,
amongst other.

In short, someone from Intel submitted a glibc patch that does faster
memcpy's on e.g. Intel i7, respects the ISO C definition, but does
things backwards. I think any software that ignores the ISO warning
"If copying takes place between objects that overlap, the behavior is
undefined" needs fixing.