Closed MarcMush closed 2 months ago
Hmm. The constant factor in this comparison started as 40, which I found not to be high enough to pass on a julia 1.6 prerelease. Later, this constant was lowered to 8, and later it was increased to 9. It seems that all of these changes have made it into ProgressMeter.jl 1.7.1, though. The two commits to master
since then do not seem suspicious to me.
The performance test fails on julia-1.7.0-rc1 and ProgressMeter#master (doesn't happen with ProgressMeter 1.7.1)
tested on windows and wsl