-
Notifications
You must be signed in to change notification settings - Fork 24
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Hard crash for downgrade request (segfault or ... python: free(): invalid next size ...
)
#391
Comments
... python: free(): invalid next size ...
... python: free(): invalid next size ...
)
I can't reproduce the crash with Docker on Apple Silicon, but the solve does take a while to conclude. It ends up in a conflict caused by Tricks to debug this faster:
I don't particularly enjoy the implicit behavior that I imagine something like:
However that information is not available until later in the code path, so we will probably need to change the Spinner stuff a bit. |
Log for podman images --digests quay.io/condaforge/linux-anvil-cos7-x86_64:latest &&
podman run --rm quay.io/condaforge/linux-anvil-cos7-x86_64:latest sh -c \
'conda install -y conda-forge-ci-setup=4 && CONDA_LIBMAMBA_SOLVER_MAX_ATTEMPTS=1 conda install -vvv -d conda-build\<3.27 conda\<23.10' |
Hm I was using an old image:
With the one from two days ago (same as yours) it does reproduce 🤔 Sadly our quay setup removes old manifests: https://quay.io/repository/condaforge/linux-anvil-cos7-x86_64?tab=history |
Fortunately I have it disk and one difference between the two is the conda-libmamba-version. 23.11.1 seems to be the culprit, so now I can take a look :) Thanks! |
Ah, that'll make it much easier to debug for you :). conda install -qy conda-forge-ci-setup=4
conda create -qynconda conda conda-libmamba-solver=23.11.0
conda activate conda
CONDA_LIBMAMBA_SOLVER_MAX_ATTEMPTS=1 "${CONDA_PREFIX}/bin/conda" install -d -p /opt/conda conda-build\<3.27 conda\<23.10
conda install -qy conda-libmamba-solver=23.11.1
CONDA_LIBMAMBA_SOLVER_MAX_ATTEMPTS=1 "${CONDA_PREFIX}/bin/conda" install -d -p /opt/conda conda-build\<3.27 conda\<23.10 an indeed, only the one with |
Found the offending cause. This |
Huh, odd! Do you know why it would fail? |
No clue yet. |
Took a look and in this case, if you add the The interesting part is that #381 does send |
If you want to take a look, these are the logs:
These ones take the task order from #381 and shift things to see what works.
Finally, if you run with #381 as is, it works too: log-ok-with-install-before-update-PR381.txt. This contradicts (B1) above, so the order of specs within a tasks must also be important? |
Unfortunately, I won't be able to look into this further. |
Checklist
What happened?
To reproduce, the
install conda-forge-ci-setup=4
is necessary as well bothconda-build<3.27
andconda<23.10
.The crash either gives a Segfault or backtrace for
*** Error in `/opt/conda/bin/python': free(): invalid next size
.The same install with
mamba
does not crash.Full reproducer:
Conda Info
No response
Conda Config
No response
Conda list
No response
Additional Context
No response
The text was updated successfully, but these errors were encountered: