V1.10 appreciation thread

I just want to say thanks for this release. I’ve been using it locally for months now, and going back to v1.9 just to test… man this is a major step forward.

  • Load times are drastically better (order of magnitude for what I do with SciML)
  • Parser messages are clearer
  • Stack traces are no longer infinitely long! They are good and legible!
  • VS Code auto-complete stuff is snappier and more predictive (I believe this has something to do with the parser change?)

Overall really solid release. Nicely done everyone involved!

123 Likes

Agreed. rc2 started working very well on my side and I haven’t looked back.

1 Like

Unless @pfitzseb has done some secret work, I can’t think of anything that would have made any change on that front :slight_smile: The new parser in base is not used (yet) by the VS Code extension.

7 Likes

Completely agree… For someone like me who started with Julia 1.4.0 , getting irritated with loading DiffEq.jl Plots.jl after +20 or +30 seconds or even more on my laptop …to now ~1 or ~2 seconds on same machine with Julia 1.10.0

Its a relief.
At some point i was seriously considering going back to Python for exactly this time to first load. It use to feel bad for seeing instant import numpy or import matplotlib but huge timelag for using DifferentialEquations or using Plots.jl

Also there is something satisfying about seeing parallel precompilation on htop application for mang packages

16 Likes

Just adding on my appreciation too. Together, 1.9 and 1.10 have made such an enormous difference to the overall feel of working in Julia.

7 Likes

Same here. I noticed running Literate notebooks (using Plots.jl , DifferentialEquations.jl, Catalyst.jl, and ModelingToolkit.jl) on GitHub actions becomes significantly faster.

1.9.4: 4m7s
1.10.0: 1m52s

The workflow also showed a trick (hack?) that using GitHub - abelsiqueira/jill: Command line installer of the Julia Language. to install Julia executable will not invalidate the precompiled cache, while using the standard setup-julia action will. The behavior is also demonstrated in GitHub - sosiristseng/julia-precompile-test: Julia precompile tests for GitHub actions. AFAIK the issue will be resolved in Julia 1.11. (relevant thread)

5 Likes

Thanks. Your post made me realize we can fix julia-actions/setup-julia so we can get this fixed on all releases before 1.11

14 Likes

That’s now released in Release v1.9.5: Maintain mtime of julia build to avoid re-precompilation · julia-actions/setup-julia · GitHub

So to get nice caching all you need is something like GitHub - julia-actions/cache: A shortcut action to cache Julia artifacts, packages, and registries.

10 Likes

Thank you. That patch totally fixed the precompile cache invalidation issue. Now, the standard workflow runs as fast as the jill workflow. (GitHub actions run)

3 Likes

Great, thanks for checking

2 Likes

Is there any chance that Julia 1.10 will become the next LTS?
Is there any time frame to know what the next LTS will be?

1 Like

I have no authority on the matter, but my understanding is that the answer is basically “there is a chance, but no promises”

as for time frame, if it is declared LTS that will happen around (and not before) the release of 1.11

6 Likes
3 Likes