isuruf
commented
2 years ago On linux, you seem to be using defaults channel and not conda-forge at all. geckodriver in defaults is behind conda-forge.
Closed its-ogawa closed 2 years ago
isuruf
commented
2 years ago On linux, you seem to be using defaults channel and not conda-forge at all. geckodriver in defaults is behind conda-forge.
its-ogawa
commented
2 years ago Thanks for the advice.
Does it mean to install without giving conda-forge? Or does it mean to specify defaults instead of conda-forge? Or is it something else?
The above two do not work for me. That is, I only get geckodriver0.26.0.
(py38) [root@<hostname> python]# conda list geckodriver
# packages in environment at /root/anaconda3/envs/py38:
#
# Name Version Build Channel
geckodriver 0.26.0 he1b5a44_0 conda-forge
(py38) [root@<hostname> python]# conda install geckodriver
Collecting package metadata (current_repodata.json): done
Solving environment: done
## Package Plan ##
environment location: /root/anaconda3/envs/py38
added / updated specs:
- geckodriver
The following packages will be SUPERSEDED by a higher-priority channel:
certifi conda-forge::certifi-2021.10.8-py38h5~ --> pkgs/main::certifi-2021.10.8-py38h06a4308_0
Proceed ([y]/n)? y
Preparing transaction: done
Verifying transaction: done
Executing transaction: done
(py38) [root@<hostname> python]# conda list geckodriver
# packages in environment at /root/anaconda3/envs/py38:
#
# Name Version Build Channel
geckodriver 0.26.0 he1b5a44_0 conda-forge(py38) [root@<hostname> python]# conda install -c defaults geckodriver
Collecting package metadata (current_repodata.json): done
Solving environment: done
# All requested packages already installed.I want to get 0.29.0 or 0.30.0 which is published in a compressed file.
isuruf
commented
2 years ago See the installation instructions at https://github.com/conda-forge/geckodriver-feedstock#installing-geckodriver
bollwyvl
commented
2 years ago Indeed: if you're not using conda install -c conda-forge, or have otherwise configured your channels, you won't get up-to-date versions of many packages. We can highly recommend using miniforge as a baseline, which already has conda-forge as its default package source.
Meanwhile: conda-forge is missing geckodriver 0.30.0, should be patched up in the just-started #20... though we build from source, and have to deal with all of the licenses represented by those big fat binaries, so no particular timeline is implied for when it will be up.
its-ogawa
commented
2 years ago @bollwyvl Thanks for the advice.
I have installed miniforge. However, I get the same problem as with conda-forge.
Namely, I cannot get the same version of geckodriver on windows and linux.
The result is as follows.
> wget https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-Windows-x86_64.exe
> Miniforge3-Windows-x86_64.exe
... snip ...
> conda config --set auto_activate_base false
... snip ...
> conda crete -n ipy38 python=3.8
... snip ...
> conda activate ipy38
... snip ...
(ipy38) > conda install geckodriver
Collecting package metadata (current_repodata.json): done
Solving environment: done
## Package Plan ##
environment location: C:\Users\ogawa\Anaconda3\envs\ipy38
added / updated specs:
- geckodriver
The following NEW packages will be INSTALLED:
geckodriver conda-forge/win-64::geckodriver-0.30.0-hdb13177_0
Proceed ([y]/n)?> wget https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge-pypy3-Linux-x86_64.sh
> bash Miniforge-pypy3-Linux-x86_64.sh
... snip ...
> conda config --set auto_activate_base false
... snip ...
> conda crete -n ipy38 python=3.8
... snip ...
> conda activate ipy38
... snip ...
(ipy38) > conda install geckodriver
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: done
## Package Plan ##
environment location: /root/anaconda3/envs/ipy38
added / updated specs:
- geckodriver
The following NEW packages will be INSTALLED:
geckodriver conda-forge/linux-64::geckodriver-0.26.0-he1b5a44_0
Proceed ([y]/n)?@isuruf Thanks for the advice.
I have checked the geckodriver that I can get from conda-forge. I can see 0.30.0. However, what I can get is 0.26.0. Why is this?
(ipy38) [root@<hostname> ]# conda search geckodriver --channel conda-forge
Loading channels: done
# Name Version Build Channel
geckodriver 0.20.1 hfc679d8_0 conda-forge
geckodriver 0.20.1 hfc679d8_1 conda-forge
geckodriver 0.21.0 hfc679d8_0 conda-forge
geckodriver 0.22.0 hfc679d8_0 conda-forge
geckodriver 0.23.0 hf484d3e_1000 conda-forge
geckodriver 0.23.0 hfc679d8_0 conda-forge
geckodriver 0.24.0 hf484d3e_0 conda-forge
geckodriver 0.24.0 hf484d3e_1 conda-forge
geckodriver 0.26.0 he1b5a44_0 conda-forge
geckodriver 0.27.0 he1b5a44_0 conda-forge
geckodriver 0.27.0 he1b5a44_1 conda-forge
geckodriver 0.28.0 h58526e2_0 conda-forge
geckodriver 0.29.0 h3146498_0 conda-forge
geckodriver 0.29.1 h3146498_0 conda-forge
geckodriver 0.30.0 h3146498_0 conda-forge
(ipy38) [root@<hostname> ]# conda install geckodriver
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: done
## Package Plan ##
environment location: /root/anaconda3/envs/ipy38
added / updated specs:
- geckodriver
The following NEW packages will be INSTALLED:
geckodriver conda-forge/linux-64::geckodriver-0.26.0-he1b5a44_0
Proceed ([y]/n)?The solver kind of does what it wants, and is influenced by what is already installed, the state of your cache, etc. You can, however, ask for something more specific.
I highly recommend creating an environment.yml that captures channels, packages and their versions:
channels:
And then:
conda env update -n name-of-env --file environment.yml
If you get hard-to-understand error messages, you can also try installing and using mamba, which tries to be a 1-for-1 replacement for conda, but faster.
its-ogawa
commented
2 years ago I see. So that's how it works. I've created an environment.yml, can you check it for me? environment.zip
Actually, I also contacted firefox engineers and asked them to make the geckodriver installation package available at the following site. https://anaconda.org/conda-forge/geckodriver/files
After running the following, I have confirmed that geckodriver works properly.
# wget https://anaconda.org/conda-forge/geckodriver/0.30.0/download/linux-64/geckodriver-0.30.0-h3146498_0.tar.bz2
... snip ...
# conda install geckodriver-0.30.0-h3146498_0.tar.bz2
... snip ...However, I cannot get the latest geckodriver with conda-forge.
Can you tell me what is affecting this? If the cache is the problem, how do I reset it?
bollwyvl
commented
2 years ago what exactly comes out of:
conda install -c conda-forge geckodriver==0.30.0Thanks for the advice.
The only thing that happens is that the package confrict does not install properly. The result is attached.
bollwyvl
commented
2 years ago Great! "has conflicts" is a lot more to go on than "doesn't install." I'm still somewhat at a loss, though, as i use these packages daily in all kinds of environments.
Please try installing mamba and re-running the commands: it usually gives much clearer errors in the case of a failure.
If that doesn't work, it's worth trying to work up a minimum reproducer, creating a new environment on your system with just:
mamba create -n gh-19-test -c conda-forge geckodriver==0.30...and then add each "leaf" package you care about.
its-ogawa
commented
2 years ago In the end, I got version 0.30.0 by downloading and extracting geckodriver-v0.30.0-linux64.tar.gz. This works, although I am unhappy that I could not install it in the conda environment.
Also, I just tried conda install geckodriver and it seems to get 0.30.0.
@bollwyvl I apologize for taking so long to reply. Thank you for your patience.
Issue: Different versions of packages can be installed on linux and windows
I am using anaconda on linux and windows. I'm using version 4.10.3 of anaconda on both.
I installed geckodriver on both, but the latest versions are different. I installed geckodriver on both, but the latest versions are different: version 0.26.0 for linux and version 0.29.0 for windows.
Why is the latest version I can get different? Is it possible to get 0.29.0 for both?
More to the point, geckodriver has released 0.30.0. You can find it at https://github.com/mozilla/geckodriver/releases
Is it possible to install this from conda?
Thanks in advance for any advice.
Environment (
conda list):Details about
condaand system (conda info):