I feel challenged to write this up. So I will put it on my wish list and assign your name to it.
On Aug 4, 2011, at 12:17 PM, Robby Findler wrote: > I think a convention is good. I also think that this kind of > organizational principle (where do tests go? How do you name the > "main" file in some package? etc) is completely appropriate for a > style guide somewhere, so maybe I'm missing something, tho. > > Robby > > On Thu, Aug 4, 2011 at 11:11 AM, Matthias Felleisen > <[email protected]> wrote: >> >> One of the responses to the draft of the Racket style guide contains the >> following paragraph: >> >> >>> There should be unified way to test collections. Let's say I fix >>> something in collect `foo', there should be an obvious way to run >>> `foo''s tests. Currently, the closest we have would be to look in >>> `tests/foo', and see if anything looks like an entry point. >>> Standardizing on test suite entry points would make it easier for >>> people to run tests after fixing bugs in collects that are not their >>> own. Such an entry point could be that each collect `X' is required to >>> have a `tests/X/run.rkt' file that, when run, runs the test suite for >>> `X'. >> >> >> I don't consider this topic appropriate for the style guide. >> But I consider it important enough to bring up for discussion. >> >> QQQ: Is there a policy that spells out testing collections? >> >> QQQ: Should we try to formulate one or leave individual testing styles alone? >> >> >> _________________________________________________ >> For list-related administrative tasks: >> http://lists.racket-lang.org/listinfo/dev >> _________________________________________________ For list-related administrative tasks: http://lists.racket-lang.org/listinfo/dev
