On Mar 27, 2008, at 11:57 PM, Jean-Francois Mertens wrote: > Ben Abbott wrote: >> Yann Clénet <yann.clenet <at> free.fr> writes: >>> I have now a problem installing with fink the package octave-atlas >>> on my new MacBook Pro with preinstalled Leopard 10.5.2 (cf below). >>> >>> Yann >>> >> The problem is related to a change in how gfortran returns values >> of complex functions. > > > > I'm surprised this error is resulting with the atlas version. > > When did you build altas? > Rather : version and revision ? > > > > Indeed, it should not occur, and does not here > (with a "stock" octave-atlas) > >> If you edit octave.info and remove the line "make check", > Was done by the maintainer now... > (I prefer to have signals of problems than to hide them, > but most users probably don't...) > >>> N = 97; >>> mldivide (2*eye(N) + circshift (eye (N), 1), ones (N, 1)); > This too works perfectly here with a "stock" octave-atlas. > > Segfaults are hard to diagnose (remember the last one > about octave-forge, Ben ..) > And BTW, I see the failed test (/sw/share/octave/3.0.0/m/sparse/pcg.m) > is in the "sparse" directory, and contains several refs to sparsity; > suitesparse might not be involved ?? > > JF
It is possible. If Yann is up to the task of helping isolate the problem, we can have him run > test pcg verbose That should tell us what test is actually failing. I suspect the left- divide because that problem has already been identified. However, I'm confused as to why Yann is encountering it ... unless he is using Octave's developer sources form Feb to the mid-Mar. Ben ------------------------------------------------------------------------- Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace _______________________________________________ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
