Failed to parse json, Extra data: line 100 column 1 (char 6597) - Could not lock the environment for platform osx-64 #381
Comments
|
I'm not able to reproduce this, it's successfully locking for me. Could you please try deleting your |
|
Or even better, rename it so that in case it's corrupt we can debug how it became corrupt. |
|
I was able to reproduce this error on an M1 using the latest version of Note that |
|
I see from the request JSON output that this is the specification for Looking at the error message, I wonder if the problem is the |
|
Found mamba's traceback in the skipped parts: |
|
roping in @dholth - I think this is conceptually the same error you fixed for libc back in january: conda/conda#12293. In the pre-plugin version, old: https://github.com/conda/conda/blob/60e11b4f1859874c1cb1a3ae324e599e0867a086/conda/core/index.py#L181-L184 |
|
Well, I fixed it in 23.1; can conda-lock's vendored conda be updated? Hopefully we don't have the wrong packages saved to lock files? |
|
@maresb Are you familiar with how Conda was vendored? I see that we only use the vendored Conda for the |
|
Locally, I upgraded conda to |
|
I'm somewhat familiar with the Conda vendoring process. @mariusvniekerk vendored Conda and I vendored Poetry. I was surprised that we use so little of the vendored Conda. Toposort should in principle be easy to replace. For MatchSpec I'm not sure how deeply that goes. But it's easy to get confused... are we relying on the external Conda pacakge or the vendored one? |
|
I think? this is most likely a problem with external's Conda, since the error comes from the subprocess where we call Toposort looks very easy to replace. MatchSpec seems much harder; it imports many things from the core set of modules like |
using the --no-mamba flag worked and successfully locked osx-64. Thanks @srilman |
|
It worked for me as well. |
|
It's great that you found a workaround! However, let's keep this open until we find a proper solution. It would be nice to understand why Mamba is failing and solve the underlying problem. |
|
Also turning off mamba support locked my dependencies, much slower though. Would love mamba. |
|
Is there anything i can provide to help debug this issue? |
|
For me, installing I haven't had a chance to look into this, but I'd be very interested to know why it doesn't work with Conda 23.1 despite the fix on the Conda side. Do we also need to adjust something on our side? |
|
As noted by @croth1 in #381 (comment) it feels like the osx issue is very similar to the one that was fixed by @dholth for glibc in conda/conda#12293. I have opened conda/conda#12715 to hopefully fix the osx one. |
|
This should be fixed now by installing the latest Conda. Ref: sagemath/sage#35986 (comment) |
Checklist
What happened?
I cannot lock the environment for platform osx-64 but I can lock successfully the environment for platforms linux-64 and osx-arm64
command and error message
$ conda-lock -f environment.yml -p osx-64
Locking dependencies for ['osx-64']...
INFO:conda_lock.conda_solver:osx-64 using specs ['python 3.10.', 'mamba', 'pip', 'conda-lock', 'py-xgboost-cpu', 'tensorflow 2.10.']
Failed to parse json, Extra data: line 100 column 1 (char 6597)
Could not lock the environment for platform osx-64
{
"command": "/usr/local/Caskroom/miniforge/base/bin/mamba create --prefix /var/folders/91/x9jf55z17vg9l94bpd_c1qp40000gp/T/tmpz20z5pw4/prefix --dry-run --json --override-channels --channel conda-forge --channel nodefaults --channel file:///var/folders/91/x9jf55z17vg9l94bpd_c1qp40000gp/T/tmpdb9saiyz python 3.10.* mamba pip conda-lock py-xgboost-cpu tensorflow 2.10.*",
environment.yml
Conda Info
Conda Config
Conda list
Additional Context
No response
The text was updated successfully, but these errors were encountered: