This is also relevant to the discussion about making lrs standard - David Avis has once again promptly responded to my email - in short he says the ETH address (fukuda at ifor.math.ethz.ch) for Komei Fukuda should be correct; he also cc'd this email to Komei. Here is his response.
Hi Marshall, Thanks for your helpful email. Here are some responses: Marshall Hampton wrote: > > One issue that has been raised is automated testing. You have some test > examples already, so I > have added a "make check" to the make file. Micheal Abshoff, our release > manager, also had some > suggestions for improving the makefile. He has a lot of experience porting > the large amount of > code in Sage to a variety of platforms. He would prefer to not maintain a > seperate codebase, so > one of his questions is whether you might be willing to distribute our > version of the makefile. I > do not think we would alter the program itself at all. I am not sure which test problems you are using. I have a small set of 7 described on P. 3 of the article "Computational Results" on the lrs home page. I can send you these if you like. Note lrs and cdd share a common file format, so you can use them for cdd also if you wish. I am more than happy to try your makefile, and to distribute it with lrs if I can understand it :-) > > A seperate but related issue that you might be able to help with is cddlib. > Cddlib has been > standard in Sage for years (and is why I heard about Sage in the first place) > but it also has a > patched makefile and I think some other alterations. Some of the Sage > developers have written to > Fukuda but he has never responded. Do you have a current email or other > contact information for > him? He has moved quite a bit, but is now in Zurich. His current email is the cc above. > > Finally, if you have a chance to take the polytope code in Sage for a spin > sometime I would > appreciate any feedback you might have. I have been slowly adding > functionality over the last > year. Recently for example I added some Gale transform methods to the > polyhedra object, but I do > not use those in my own research so I am not sure what else would be useful. Will do. I'm currently in Japan with a borrowed windows laptop, but I see you have a windows interface. I'll ask to have it installed in on our (linux) machines at McGill where I do most of my serious computing. best, david On Jan 27, 4:18 pm, mabshoff <mabsh...@googlemail.com> wrote: > On Jan 26, 8:01 am, Sebastien Barthelemy > > > > <sebastien.barthel...@gmail.com> wrote: > > 2009/1/26 mabshoff <mabsh...@googlemail.com>: > > > > There are several fixes for Solaris and improvements to build shared > > > libraries in our tree: > > > > * allfaces.c.diff makes cddlib compile on Solaris 10 on x86 and Sparc > > > since some variable names collide with numerical constants > > > * #3304 - Make cddlib Debian package use shared library > > > These would be nice to have since otherwise we will ship a rather > > > dirtied up 0.94f compared to vanilla upstream sources. Given the > > > linker problem this thread was created due I am now leaning toward > > > updating to 0.94f and against policy patch inside src directly since > > > adding the files to patches and copying them over is pointless when > > > they are autogenerated files. > > > I just updated the spkg to 0.94f, keeping the allfaces.c patch, but without > > integrating #3304. This seems to work and I'll continue to use this version > > for now. > > Good. > > > If I can contribute it, please tell me how. > > Post a link to the new spkg at the update ticket for cddlib in trac. I > will then review it. Change the summary line of that ticket to "[with > spkg, needs review] ...." > > > Do you think it can make it to 3.3? > > Absolutely. > > > Cheers, > > Sébastien > > Cheers, > > Michael --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
