From maintainers-request at octave dot org Tue Mar 21 16:34:55 2006 Subject: Re: dist target, ufsparse and metis From: David Bateman To: octave maintainers mailing list CC: "John W. Eaton" , "Christophe Prud'homme" Date: Tue, 21 Mar 2006 23:29:34 +0100 John, I'm the wrong side of my corporate firewall and so am responding from your mail on the website. For the need for octave itself with all dependencies to build to dist file, thats a logical consequence of the design choice of having octave itself build the figures rather than store them. I can send you the built figures if you like, and with these and the stamp-eps, etc files being touched, you should get able to force the figures into the manual. It is normal the NPARTITION disables everything in the Partition directory. Cholmod still uses the COLAMD and AMD package for the matrix reordering. The constrained COLAMD and AMD packages in cholmod_ccolamd.c and cholmod_csymamd.c are only needed to allow metis to place additional constrains on the graph partitioning. There are RPM for Fedora that Quentin maintains based on an RPM I developed that is available at http://www.dbateman.org/?page=octave though at the moment debian is only linking to a umfpack 4.4 package and so misses CXSparse for the QR and dmperm function, CHOLMOD for the Cholesky factorization and the constrained ccolamd and csymamd functions. Christope Prud'homme who sent a mail recently to the list about eigs is responsible for umfpack in debian, and I've informed him about the need for ufsparse for octave 2.9, and the fact of the new umfpack upstream release (the version in the latest UFsparse is 4.6), so I hope something will be done here. I have no experience with making debian packages, otherwise I'd probably try and propose a package, though I hope the RM spec-file in my package on the site above will give sufficient information to a debian developer to allow ufsparse to be built... I'm not sure what is being done for cygwin, though for mingw I'm building all of the dependencies, including ufsparse. Regards David -- David Bateman David dot Bateman at motorola dot com Motorola Labs - Paris +33 1 69 35 48 04 (Ph) Parc Les Algorithmes, Commune de St Aubin +33 6 72 01 06 33 (Mob) 91193 Gif-Sur-Yvette FRANCE +33 1 69 35 77 01 (Fax) The information contained in this communication has been classified as: [x] General Business Information [ ] Motorola Internal Use Only [ ] Motorola Confidential Proprietary