Le 11/11/2013 20:45, Robert Bradshaw a écrit : > On Mon, Nov 11, 2013 at 10:31 AM, Volker Braun <vbraun.n...@gmail.com> wrote: >> On Monday, November 11, 2013 10:09:54 AM UTC-8, Ralf Hemmecke wrote: >>> >>> But I think even if I send that log, the bug is rather that Sage should >>> check "git --version" and then don't start trying to build git at all, >>> no? >> >> >> That is really a feature request, not a bug. Of course it would be nice. If >> you want to work on a configure-style script that queries the system for >> build- and runtime dependencies then please do! Ideally it'll know about >> package management in different distributions to make reasonable suggestions >> about what to install before trying to build from source. And is easily >> extensible. > > It's much better to accidentally build a git that we don't need than > the other way around (especially if version mismatches manifest as > subtle bugs). Also, the git shipped with Sage, if not patched, at > least is well tested in context. Working binary downloads are very > important too, not everyone clones and builds.
This is maddening. All other projects would do like this: (1) check if git is available ; (2) check if the available git has a recent enough version (for other things than git, a full equality on the version might make sense) ; (3) in case it's not a good git, either error out (hard dep) or build a good one (embedded source dep). Only sage won't look around for what is there, and will actively spread doctests around to check the embedded version was really used (there are places where exact versions are checked just for that!) ; and will also on some occasions *patch* the embedded version and use the variation somewhere else for no other reason than making using upstream break (example: disabling threading in ecl was done in the spkg instead of calling the api to disable it in ecl.pyx). "Beware of the subtle bugs!" say the sage developers. The spkg-ed versions have subtle bugs, so using them doesn't magically solve anything. The rest of the world knows about the subtle bugs, and deals with them. They even deal with them better precisely because having people use the code in more various settings makes them easier to understand and fix. To make a comparison which should go directly to the heart of quite a few people around here: when you want to use a theorem, you check the hypotheses then call it by its name; you don't fall down to ZFC and do the full climb upwards! Snark on #sagemath -- 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/groups/opt_out.
