Cannot create custom image in Julia 1.10: Method overwriting is not permitted during Module precompilation

[denizyuret]

Following the instructions at https://pyjulia.readthedocs.io/en/latest/sysimage.html with Julia 1.10.0 I get the error given in the subject. Here is the full output: https://gist.github.com/denizyuret/dcf47516b0379ce1f7fce8069610d039.

The instructions seem to work with Julia 1.9.4. However when trying to start julia with the new sys.so we get a different error: https://gist.github.com/denizyuret/62bac89ea0ac1f27e8bc0077d282c497.

[mkitti]

Do you have __precompile__(false) somewhere in your source files?

[denizyuret]

I have no source files, just trying to create a plain sys.so that works with conda.

[mkitti]

I'm not sure I understand. If you have no new source code, there should not be a reason to create a new system image. What are your objectives? Do you encounter a problem otherwise?

Also have you considered using juliacall? I just deployed a package to here:
https://anaconda.org/julialang/pyjuliacall

conda install julialang::pyjuliacall

[denizyuret]

I may be misunderstanding, but the pyjulia documentation says it doesn’t work by default with conda because the conda python binary is statically linked (which I confirmed for my setup). Among the several workarounds suggested is generating a new sys.so image: https://pyjulia.readthedocs.io/en/latest/troubleshooting.html#your-python-interpreter-is-statically-linked-to-libpython I will check out pyjuliacall.

…

On Sun, Feb 4, 2024 at 9:08 PM Mark Kittisopikul ***@***.***> wrote: I'm not sure I understand. If you have no new source code, there should not be a reason to create a new system image. What are your objectives? Do you encounter a problem otherwise? Also have you considered using juliacall? I just deployed a package to here: https://anaconda.org/julialang/pyjuliacall conda install julialang::pyjuliacall — Reply to this email directly, view it on GitHub <#555 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAN43JUEF4AXH27QNYYJ42LYR7FBXAVCNFSM6AAAAABCYU2IBSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMRVHA3DQMZRGA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[mkitti]

I think most settle for the following when using pyjulia.

from julia.api import Julia
jl = Julia(compiled_modules=False)

However, I think pyjuliacall probably works better in this situation. Let me know if that's not the case.

[mkitti]

If pyjuliacall does work better, I would appreciate your feedback on conda-forge/staged-recipes#25097