Control: forwarded -1 https://github.com/montanaflynn/stats/issues/33
Hi Nicoo and Guinness, On Tue, Dec 15, 2020 at 6:57 AM nicoo <ni...@debian.org> wrote: > Hi Anthony, > > On Tue, Dec 15, 2020 at 12:38:55PM +0000, Anthony Fok wrote: > > Fix FTBFS by skipping test on non-amd64 architectures > > > > until test failures due to FMA (fused multiply-add) differences are > > resolved. > > See https://github.com/montanaflynn/stats/issues/33 > > I believe that fixing the testsuite upstream would be a much better approach, > as > discussed in the upstream issue. I can probably do that and sumbit a fix > upstream, and Guiness also offered to help. Thank you so much for both of you for helping out! Indeed, fixing the testsuite upstream is a much better and complete solution. I tried experimenting with tolerance() and veryclose() etc. but got a bit stuck, and was feeling really sleepy, and, recalling how long it took me to solve a similar issue in Hugo back in October 2019: https://github.com/gohugoio/hugo/issues/6387 I decided to be lazy and simply disabled the tests on non-amd64 architectures... <embarrassed, grin, duck, run> I am so glad you both offered to help, and went ahead and fixed the issue so quickly and properly. Kudos and thanks to both of you! Cheers, Anthony
