Package compatibility and environment issues with respect to new version

New to Julia
1

Hello all,

I have been using Julia (v1.9.3) in JupyterLab through the Anaconda distribution, and all my projects and files were functioning correctly in the global environment. However, after updating to Julia v1.11.2, I started encountering significant issues, particularly with package compatibility. For example, during precompilation, the following errors occurred:

✗ OrdinaryDiffEq
✗ StochasticDiffEq
✗ Trebuchet
✗ DiffEqSensitivity
8 dependencies successfully precompiled in 287 seconds. 258 already precompiled.
1 dependency had output during precompilation:
┌ WebSockets
│  WARNING: could not import Logging.termlength into WebSockets
│  WARNING: could not import Logging.showvalue into WebSockets

In an attempt to resolve the issues, I reverted back to Julia v1.9.3. However, after the downgrade, the Jupyter kernel started dying and reloading repeatedly, making it impossible to run any projects.

I am now looking for a solution to either fix the compatibility issues in Julia v1.11.2 or restore a stable working environment with Julia v1.9.3 in JupyterLab.

Note: At the moment, I have 2 versions installed side by side, and I have installed the julia from the microsoft via winget, which was a standalone. The status of my IJulia is IJulia v1.26.0. The status at the moment is:

(@v1.9) pkg> st
Status `C:\Users\HP\.julia\environments\v1.9\Project.toml`
  [fbb218c0] BSON v0.3.9
  [31a5f54b] Debugger v0.7.10
  [41bf760c] DiffEqSensitivity v6.79.0
⌅ [587475ba] Flux v0.13.17
  [f6369f11] ForwardDiff v0.10.38
  [7073ff75] IJulia v1.26.0
  [429524aa] Optim v1.10.0
⌅ [1dea7af3] OrdinaryDiffEq v6.51.2
  [91a5bcdd] Plots v1.40.9
  [49802e3a] ProgressBars v1.5.1
⌃ [c3572dad] Sundials v4.20.1
  [ddb6d928] YAML v0.4.12
Info Packages marked with ⌃ and ⌅ have new versions available, but those with ⌅ are restricted by compatibility constraints from upgrading. To see why use `status --outdated`

My version info

julia> versioninfo()
Julia Version 1.9.3
Commit bed2cd540a (2023-08-24 14:43 UTC)
Build Info:
  Official https://julialang.org/ release
Platform Info:
  OS: Windows (x86_64-w64-mingw32)
  CPU: 8 × AMD Ryzen 5 3550H with Radeon Vega Mobile Gfx
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-14.0.6 (ORCJIT, znver1)
  Threads: 1 on 8 virtual cores

Kindly help me out in sorting out this issue. I am kind of overwhelmed with being not able to figure out.

1 Like
2

I’ve never tried using Julia “through the Anaconda distribution” and have no idea what additional issues this might bring, so I would recommend:

  1. Install juliaup to install and manage your Julia versions
  2. Add the LTS channel juliaup add lts - this should install Julia 1.10.7
  3. Start Julia by just typing julia in the terminal (or julia +lts if just julia doesn’t give you 1.10.7)
  4. Install IJulia ] add IJulia
  5. Start a Jupyter notebook via using IJulia; notebook()
  6. In the notebook, add the packages you need to the notebook environment (in a Jupyter code cell do ] add OrdinaryDiffEq, Plots, ...
  7. Run your code
  8. Profit
1 Like
3

I agree that you should not use anaconda, but use juliaup to install Julia.

If you have package compatibility issues, please always use a per-project environment:

mkdir myproject
cd myproject
julia --project=.

and then in Julia add your packages one by one, e.g.

using Pkg
pkg"add IJulia"
pkg"add Plots"
...

In most cases this will just work, if not please report the first package that failed to be installed successfully and the output of Pkg.status().

1 Like