> This points to the core problem: we simply have no way to guess whether > people want to do that or not.
We do not have to care. Doctests must pass, whatever the install is. If the local install is broken then we should have a way to detect it, and if we detect it we must not run the doctests automatically. But if the install is not broken and the doctests break, then we must know it. > Here is a possible compromise proposal: > * keep default optional package as they are now > * add a new flag --optional=external (or change --optional=all to do this) > to test all auto-detected-but-not-part-of-Sage packages like mathematica. > * add a corresponding Makefile target running tests with this flag. You know that nobody will ever run those things. The same way that we do not do it today, even though we have a "--optional=mathematica" flag available. That's just code that nobody will run, and our interface will remain as broken as they are now because nobody tests them. Anyway. Like I said on the ticket, I'm tired of arguing with you when you don't want to hear something. Cplex and Gurobi will be tested, and those who care about having a working interface with Maple, Matlab, Octave, Mathematica, Scilab, Macaulay2 and others will have to defend their interests against you. Too bad you fight for a way which will bring no improvement to our software. Nathann -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
