Hi Barak, could we agree to keep debian-science in CC to make sure our discussion about f2c is somewhere archived. Considering there are so many clones of f2c and netlib with a code copy of f2c (as well as clapack where I'm fighting desperately with) we should stick to an open discussion to prevent that it will be forgotten and others might create new clones with new features.
If you know a better place for public discussion than Debian Science list please suggest one. On Thu, May 26, 2016 at 01:02:45PM +0100, Barak A. Pearlmutter wrote: > > https://github.com/juanjosegarciaripoll/f2c > > Or https://github.com/barak/f2c-1 which has some minor tweaks. As far as I can see this is the same as https://github.com/juanjosegarciaripoll/f2c So where do we want to let our watch file point to and who is "upstream" of this code? In any case these repositories are missing the files lib/i_ceiling.c lib/i_len_trim.c which are needed to build clapack. I admit I have big hope in a better / more relieable build system since I'm fighting since hours to get clapack build again[1]. I've just compared the build options of f2c inside the cmake based clapack build and the manual makefile and don't find any differences that look relevant - but clapack keeps on failing to build. :-(( If we could get any enhancement I'd be super happy. So if we could agree upon a definite upstream source (which optimally would set some release tags on Github) I'd volunteer to spent some time into this to hopefully get the build finally working. Kind regards Andreas. [1] https://lists.debian.org/debian-mentors/2016/05/msg00646.html -- http://fam-tille.de
