Switch CI install to mamba#1073
Conversation
sroet
left a comment
There was a problem hiding this comment.
LGTM, one leftover question. Feel free to merge without ;)
| if [ ! -z "$OPS_ENV" ] | ||
| then | ||
| conda create -q -y --name $OPS_ENV conda future pyyaml python=$CONDA_PY | ||
| $INSTALL_CMD --name $OPS_ENV conda future pyyaml python=$CONDA_PY |
There was a problem hiding this comment.
only leftover comment from the other PR:
"Is the conda install here necessary?
| $INSTALL_CMD --name $OPS_ENV conda future pyyaml python=$CONDA_PY | |
| $INSTALL_CMD --name $OPS_ENV future pyyaml python=$CONDA_PY |
Codecov Report
@@ Coverage Diff @@
## master #1073 +/- ##
=======================================
Coverage 81.57% 81.57%
=======================================
Files 140 140
Lines 15416 15416
=======================================
Hits 12576 12576
Misses 2840 2840 Continue to review full report at Codecov.
|
There was some stuff in there that went back to parsing conda recipees, which is now far from relevant.
|
I restructured the part in question a little more thoroughly than just the recommended change. That The script is used by the @sroet : If you have a chance to take another look, please do. If not, I'm pretty sure the changes are fine, and I'll aim to merge this before the next overnight scheduled tests run. |
|
Whoops. forgot to merge this before I went to bed last night! |
Third (and, I hope, final!) version of this. See also #1072 and #1071 for the history of trying to debug this problem.
Main points, quoting from previous:
However, that change resulted in a segfault that appears to have been because a test didn't close a storage before removing the file (which is fixed in this PR).
In addition to fixing that segfault and switching to
mambafor installation, this PR: