Closed ccoffrin closed 7 years ago
The MathProgBase version bounds are okay. Can't immediately explain the bizarre failure on 0.6.
Cool. Maybe we should wait on v0.6 for now.
Any ideas why the Windows build is failing? Seems like something around how binaries are packaged? https://ci.appveyor.com/project/tkelman/coinoptservices-jl/build/1.0.669/job/qei1utyyvd1y1l1v
Let's see if @tkelman has a chance to take a look
On 0.6 it looks like I'm missing an esc
somewhere, probably a macro hygiene issue
I'm going to assume the windows failure is not related to this PR.
We should be aware that runtests.jl is currently only testing two instances. JuMP is not calling this solver in its tests (https://github.com/JuliaOpt/CoinOptServices.jl/issues/26).
Probably need to switch the WinRPM source from opensuse 13.2 to 42.2, but it's not able to build there because of https://github.com/JuliaOpt/CoinOptServices.jl/issues/27.
Standard stuff. Dropping Julia v0.4 support. The version lb on MPB might need to be updated, please advise.
Also, would be open to working on v0.6 compatibility in this PR, but need a hint on why this is breaking? See, https://travis-ci.org/ccoffrin/CoinOptServices.jl/jobs/217027340.