In https://github.com/exercism/exercism.io/issues/2425
@odeits suggests adding the edge case "10convert01". We could also split it up into 10convert and convert10.
Can Ruby be check-marked complete due to exercism/xruby#171
Done.
Consistency as noted in exercism/xruby/issues/174 should be addressed.
How many of these invalid are exception based, rather than 0 based?
I'm not sure--but I'd like to just change the README to allow for both approaches.
You beat me to it. :)
I've just stumbled over a solution in the elixir track which would "convert" the string "22" to 2 * 1 + 2 * 2 = 6, probably we should not only have invalid tests with alphabetic characters but numeric chars also.
Ah, yes, that's a good catch.
I'd be very happy if we get some common test data before fixing this issue. Fixing it now per track and collecting common test data afterwards and then file an issue per track again might create additional and unnecessary noise in the tracks repositories and additional confusion on commenting submissions that are against different versions of the tests.
Since we are in the process of deprecating binary along other exercises in a big sweep (exercism/x-common#279), I think we can close this, can't we?
Yeah, I think we can. Thanks!
Most helpful comment
I'd be very happy if we get some common test data before fixing this issue. Fixing it now per track and collecting common test data afterwards and then file an issue per track again might create additional and unnecessary noise in the tracks repositories and additional confusion on commenting submissions that are against different versions of the tests.