Re: [ublas] CVS updates (please update and report regressions)

From: Gunter Winkler (guwi17_at_[hidden])
Date: 2007-03-23 08:45:04

Am Freitag, 23. März 2007 11:09 schrieb Neal Becker:

> 1) While those interested in high performance computing will
> generally want this change, it will break old code that assumes that
> a
> ublas::vector<Complex> is initialized to all zero (none of us write
> code like this, of course).

I put a note at the relaese notes page.

> 2) The constructor bypass uses partial specialization.  I think we
> are safe to assume all supported compilers have this?  If not we can
> add the appropriate ifdef.

There are already more locations where partial specializations is 
required. So we can assume that this is available.

mfg
Gunter