Kristoffer Carlsson
Kristoffer Carlsson
I wonder if the 2 arg show method should perhaps be removed?
Maybe the Julia ambiguity checker became better?
Just a small comment. Does it make sense to put all the daily reports under a common `daily` folder in BaseBenchmarkReports? I know that using Github as data storage is...
Or allow for an extra keyword (`primary`?) that the `vs` thing is compared against. That way you could compare x and y where x and y can be branches, commits...
I could do a PDE solver (FEM) benchmark. A funny thing about PDEs is that one solving the simplest possible (regular grid with diffusion) is a couple of lines while...
We have some data files in this repo https://github.com/JuliaCI/BaseBenchmarks.jl/tree/master/src/problem/data.
Doesn't even seem to be from this PR though?
Passes now lul... Oh, it passes even with the stackoverflow :package: Should probably check that the plot is properly produced: https://travis-ci.org/JuliaCI/BaseBenchmarks.jl/jobs/623400233#L67244-L67251.
We have a manifest for that?
Wouldn't it make sense to check in a separate manifest for these specific tests? Otherwise it becomes hard to upgrade only the dependencies of BaseBenchmarks.jl itself.