On Oct 6, 2015 6:06 AM, "Tim Smith" <t...@tim-smith.us> wrote: > > On Tue, Oct 6, 2015 at 5:35 AM, <distutils-sig-requ...@python.org> wrote: > > On Tue, 6 Oct 2015 11:30:00 +0300 > > Ionel Cristian M?rie? <cont...@ionelmc.ro> wrote: > >> On Tue, Oct 6, 2015 at 10:51 AM, Antoine Pitrou <solip...@pitrou.net> wrote: > >> > >> > They should be inside the module. That way, you can check an installed > >> > module is ok by running e.g. "python -m mypackage.tests". Any other > >> > choice makes testing installed modules more cumbersome. > >> > > >> > >> ?Does that really make sense? I haven't heard of any user actually running > >> tests that way. > > Numeric computing may be a special case because of its sensitivity to > compiler/BLAS interactions but numpy and scipy support e.g. > `numpy.test()` and Homebrew has received bug reports from people > running them and getting unexpected results so I believe this is > practiced and useful in some domains.
Yes, and numpy upstream gets bug / support requests on a totally regular basis (weekly-ish?) from end-users who have somehow managed to wedge up their install and either they noticed because they ran the installed test suite, or else they are reporting some bizarre results and we say "can you run python -m 'import numpy; numpy.test()'?". My impression is that there's a non-trivial subset of users who install numpy from binaries and then immediately run the test suite as a matter of course, and as a developer this is something that I'm happy to encourage (yay more test coverage). Adding another step where they have to download something extra would definitely discourage this. I can easily believe that there are other parts of the python world where this is uncommon, but if some people are saying "I've never seen this and it isn't useful" and others are saying "I see this all the time and it's very useful", then this isn't really anecdote versus contradictory anecdote. The conclusion is that we're talking about a thing that happens at least sometimes and is useful at least sometimes. -n
_______________________________________________ Distutils-SIG maillist - Distutils-SIG@python.org https://mail.python.org/mailman/listinfo/distutils-sig
