On 12/10/12 12:21, Thomas Kahle wrote: > Hi, > > On 21:42 Thu 11 Oct 2012, Francois Bissey wrote: >> I don't experience this problem at all. ecls-12.2.7 cause some other >> problems with some tests. You are the first person to report this kind >> of problems. Could you provide more info, such use flags for ecls and >> maxima. Bigger chunk of the maxima compile failure. gcc used... > > Here's what I do to reproduce: > > I'm on a mostly stable amd64 system, but with gcc-4.6.3 (see > emerge.info) > > 1) emerge =dev-lisp/ecls-12.2.1 with USE="X emacs threads unicode" -> Works > 2) emerge =sci-mathematics/maxima-5.26.0 with USE="X ecls emacs latex nls tk > unicode" > -> fails with attached build.log > > Let me just assume without trying that it works with USE="-ecls", but > that's not relevant for me as I want sage which has > =sci-mathematics/maxima-5.26.0[ecls] as a dep. > > 3) emerge =dev-lisp/ecls-12.7.1 with USE="X emacs threads unicode" -> Works > 4) emerge =sci-mathematics/maxima-5.26.0 with USE="X ecls emacs latex nls tk > unicode" -> Works > 5) emerge =sage-5.3 -> Works > > I conclude that my original problem with sage not working was due to > some upgrade breaking either maxima or ecls (but not picked up by > revdep-rebuild!). When trying to rebuild I ran into the build failures > described above, but the combination of ecls-12.7.1 and maxima-5.26.0 > builds and works with sage so far. >
OK ecls is fine sbcl isn't. I would like you to try without unicode for both ecls and maxima. ecls with unicode can cause trouble in sage we thought we had a solution with Paulo (formerly from Mandriva) but it is not fool-proof. Francois
