Closed
Description
Running the same test as #170, but with --kind explicit
results in a different hash on every run:
env_file=/tmp/environment.yml
lock_file=/tmp/conda-explicit.lock
cat << EOF > $env_file
channels:
- conda-forge
dependencies:
- scikit-learn
EOF
function conda_lock_multiple_times() {
rm -rf $lock_file
for i in $(seq 1 4); do
conda-lock lock -k explicit --mamba -p linux-64 -f $env_file --filename-template $lock_file > /dev/null 2>&1
echo "lock file md5sum: $(md5 $lock_file)"
done
}
echo "first run"
conda_lock_multiple_times
echo "second run"
conda_lock_multiple_times
0.13.2
produces consistent output everytime, while 1.0.1
and 1.0.4
produce different hashes on every run. I presume the fix is similar to #181, but on a different code path.
Metadata
Metadata
Assignees
Labels
No labels