Nexus 42 Release Notes


The following new featues have been added:


System Requirements

MXML XML Parsing Library

Version 2.2.2 or higher of mxml is required. Earlier versions have a bug and the XML API will not work. This package can be downloded in both source and binary rpm form and is also available as part of Fedora Extras. IMPORTANT NOTE: Debian also provides the mxml package, but it based on 2.0 and will not work properly.

Python Interface You will need both the numpy and ctypes modules to be available. These are provided in both the Fedora and EPEL repositories.

Building Notes

NAG F90/F95 Compiler

The NAG compiler needs the -mismatch flag to be specified or else it will not compile NXmodule.f90 This is achieve by running configure with the FCFLAGS environment variable set to contain the flag e.g.

env FCFLAGS="-mismatch" ./configure --with-f90=f95

HDF4 on Intel Macs

There is a problem with the include file, hdfi.h (normally in /usr/local/include). See for details of the modifications necessary to fix it.

New Features

C++ Interface (provided by Freddie Akeroyd and Peter Peterson)

See the doxygen documentation and NeXus API test program. A C++ Stream Like interface is also supported - the idea is to provide an IOSteam like interface and allow you to type

    // create an entry and a data item
    File nf(fname, NXACC_CREATE);
    nf << Group("entry1", "NXentry") << Data("dat1", w, "int_attr", 3);

    File nf1(fname, NXACC_RDWR);
    // add a double_attr to an existing setup
    nf1 >> Group("entry1", "NXentry") >> Data("dat1") << Attr("double_attr", 6.0);

    // read back data items
    File nf2(fname, NXACC_READ);
    nf2 >> Group("entry1", "NXentry") >> Data("dat1", w1, "int_attr", i, "double_attr", d);
    // alternative way to read d1
    nf2 >> Data("dat1") >> Attr("double_attr", d1);

See also the NeXus API test program

IDL Interface (provided by Jussi Kauppila and Mark Koennecke)

There is a new interface to RSI’s Interactive Data Language, IDL for NeXus. This interface has to be considered beta. Nevertheless it is working most of the time. Known issues include:

Python Interface

There is now, thanks to Paul Kienzle, a supported interface for the python scripting language. Arrays are stored in numpy arrays and thus allow for efficient data manipulations.

Changed Features

Known Issues

See the comments on the IDL interface.

Miscellaneous bug fixes

The following items are bugs reported in previous releases and resolved in the 4.2 release.

Upcoming Features