jaimergp comments

Results 1520 comments of


                                            jaimergp

Revamp shell logic: default UX should not initialize shells, but spawn new ones

xref https://github.com/conda/conda/pull/15447

Conda debug does not provide a `--debug` option

I see the PR in #5817, but as a workaround you can also set `CONDA_VERBOSITY=3` to enable `conda`'s verbose output. This might not be enough for conda-build specific details.

conda update issue when updating from a no longer supported Python version

You can put python=* in the CLI and the pin will "float", IIRC

BUG: noarch package in multi-output recipe changes CPU architecture mid-build

Intriguing. I would have assumed that https://github.com/conda/conda-build/pull/5350 was responsible for the fix, but that's not released yet :/

Report channels being fetched in CLI

The total size does not need to be reported if not available. The libmamba bars report total downloaded, speed and time spent so far, and that's ok too. @travishathaway mentioned...

Add instructions about splitting header and runtime + run_export> packages

pre-commit.ci autofix

conda run treats executable's '-v' option incorrectly

I checked and it's a tricky one. The usual solution is to use `--` to separate the command being run from the invoker: ``` $ conda run -- python -vhgj...

on a pure ubuntu server: FileNotFoundError: [Errno 2] No such file or directory: 'xdg-mim

You can export CONDA_SHORTCUTS=false for now. Thanks for the report!

Sub-second resolution for repodata.json cache

I observed this with local channels; e.g. `conda create -dnx -c ./path/to/channel`. Is it ok to to put hash of the full timestamp (with microsecond resolution) there?

Sub-second resolution for repodata.json cache

I tried this in some failing tests at https://github.com/conda/conda-libmamba-solver/pull/457 (`CONDA_REPODATA_TTL=0 pytest tests/models/test_prefix_graph.py -k libmamba`) and it doesn't help. Adding a `time.sleep(1)` and the beginning of the test does fix it...