I also believe that we should test them. The very fact that there are broken doctests that we did not fix only because we never run them proves it (see comment 7 of #18908).
Also, there are two things that you seem to take for granted: - "[...] does not mean that I care about Sage doctest failures": Why should there be failures in the first place? Doctests must pass, and we will fix those failures. That's the purpose of doctests: detect failures and fix them. - "Maybe I have a version of the external package which is not fully compatible with Sage but which works for me?" Why wouldn't we be able to detect that? What is somebody has such an install and *wants* to use it with Sage? We must be able to raise a proper exception if something does not work, so that (s)he can figure out how to fix his install. "The fact that proprietary packages will probably never be automatically tested on the buildbot for license reasons is also a good reason to not run those tests automatically." "For license reasons"? What's wrong with having Maple installed on a patchbot? We can already have patchbot running non-GPL-compatible softwares like CBC, as it is a new-style optional package. Finally, regardless of whether the patchbots will run the tests, you know as well as I that something which is not tested gets broken. And that's the current state of our interface.* modules, unless we do something about it. Nathann On 16 July 2015 at 12:51, Sébastien Labbé <sla...@gmail.com> wrote: > >> By automatically running those tests, you are forcing people to care >> about stuff that they didn't want to care about. >> > > I don't really care myself about the mathematica interface. But it happens > that mathematica 10.0 is installed on my machine in my lab and that most of > the people in my lab use mathematica rather than sage. They won't test the > mathematica doctests in the sage library since they are not sage > developpers. But, in the next year, I might have a research discussion with > them where they will be using mathematica and I will be using sage. And it > is possible that I will need to use the interface. And if the doctests are > broken, then, I won't be happy. And I won't be able to convince them about > using Sage or that interface. > > Personnaly, I want the mathematica optional tests to be run on my machine > when I run "make ptestlong" if mathematica is available on my machine. > > I we don't agree with this, then, I would like a command like the one > suggested by dimpase [1]: > > make ptest OPTIONAL+=mathematica > > [1] http://trac.sagemath.org/ticket/18904#comment:27 > > Sébastien > > -- > You received this message because you are subscribed to a topic in the > Google Groups "sage-devel" group. > To unsubscribe from this topic, visit > https://groups.google.com/d/topic/sage-devel/M98nD7833KM/unsubscribe. > To unsubscribe from this group and all its topics, 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. -- 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.
