Why this tool doesn't support anaconda?

[hongyi-zhao]

Hi,

I tried this tool, it seems it works smoothly with conda, while not works with anaconda.

Any hints?

[tartansandal]

Hi @hongyi-zhao,

In short, this completion code only supports the conda command.

The conda command is for managing local Anaconda/Miniconda installations and package sets.
This is a core component that is always installed.

The anaconda command is for managing Anaconda Cloud accounts and the packages stored there.
This is an additional component that is only optionally installed.

These are two separate commands with different code bases. They would need separate completion implementations.

Of course this project could serve as a template for anyone wanting to write completion code for the anaconda command. :-)

[hongyi-zhao]

Thanks for your notes.

But, I also noticed that the current implementation of your code cannot do the sub-command' arg completions. Say, for the following:

$ conda config <tab>

It will do nothing for completions. In my mind, the completions should be trigged in any level for sub-commands. Any hints for this issue?

Regards

[tartansandal]

Try:

$ conda config --<tab>

the config sub-command is odd in that it only takes option arguments, whereas say, the env sub-command takes both options and arguments. We have to insist on at least one dash - to get consistent behaviour.

Hope this helps.

[hongyi-zhao]

It cann't solve the problem with your suggestion:

$ conda config --<tab>

Then I get the following:
$ conda config --bash: conda config: syntax error in expression (error token is "config")

[tartansandal]

Damn. That is a bug I thought I had squashed 😞
Try the following to reset the completion cache.

$ unset __comp_conda_cache

[hongyi-zhao]

Still failed:

$ unset __comp_conda_cache
$ conda config --<tab>
$ conda config --bash: conda config: syntax error in expression (error token is "config")
$ 

ps. I use the follwong anaconda python with pyenv:

$ conda info 

     active environment : base
    active env location : /home/werner/.pyenv/versions/anaconda3-4.3.0
            shell level : 0
       user config file : /home/werner/.condarc
 populated config files : /home/werner/.condarc
          conda version : 4.8.0
    conda-build version : not installed
         python version : 3.6.0.final.0
       virtual packages : __glibc=2.30
       base environment : /home/werner/.pyenv/versions/anaconda3-4.3.0  (writable)
           channel URLs : https://conda.anaconda.org/anaconda-fusion/linux-64
                          https://conda.anaconda.org/anaconda-fusion/noarch
                          https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/linux-64
                          https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/noarch
                          https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/linux-64
                          https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/noarch
                          https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/r/linux-64
                          https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/r/noarch
          package cache : /home/werner/.pyenv/versions/anaconda3-4.3.0/pkgs
                          /home/werner/.conda/pkgs
       envs directories : /home/werner/.pyenv/versions/anaconda3-4.3.0/envs
                          /home/werner/.conda/envs
               platform : linux-64
             user-agent : conda/4.8.0 requests/2.22.0 CPython/3.6.0 Linux/5.3.0-24-generic ubuntu/19.10 glibc/2.30
                UID:GID : 1000:1000
             netrc file : /home/werner/.netrc
           offline mode : False

[tartansandal]

Thanks that. I might not be able to get back to that until tomorrow, however, I am very keen to figure out what is going wrong and help you out.

[hongyi-zhao]

Thanks a lot, I'm very grad to find the bug and let this tool more perfect, and convenient for users like me.

[tartansandal]

Ah I thought that bug looked old.

While tracking this down, I discovered that the new conda release that you are using (4.8.0) contains my older completion code. This has known bugs and was hardwired into the init bash process rather than being packaged. I will get in touch with the maintainers to to see if I can get it removed ASAP.

Until then, you can try to force the loading of the new improved code with:

source /home/werner/.pyenv/versions/anaconda3-4.3.0/share/bash-completion/completions/conda

Hope this helps.

[hongyi-zhao]

Thanks, I'll try it.

[hongyi-zhao]

Ah I thought that bug looked old.

While tracking this down, I discovered that the new conda release that you are using (4.8.0) contains my older completion code. This has known bugs and was hardwired into the init bash process rather than being packaged. I will get in touch with the maintainers to to see if I can get it removed ASAP.

Until then, you can try to force the loading of the new improved code with:

source /home/werner/.pyenv/versions/anaconda3-4.3.0/share/bash-completion/completions/conda

To install the latest git version of this package, the following command is enough or not:

conda install -c tartansandal conda-bash-completion

Hope this helps.

[tartansandal]

That should work. The same package is now available from conda-forge so the following will also work:

conda install -c conda-forge conda-bash-completion

The long-term plan is to make one of the base packages, so at some point in the future, it will be installed by default.

[hongyi-zhao]

I've learned that the conda-forge is based on several continuous integration tools and can auto build the latest sources from the corresponding repository.

But, whether the following two commands will always obtain the same version:

$ conda install -c tartansandal conda-bash-completion
$ conda install -c conda-forge conda-bash-completion

What's the difference between the above two commands?

From the manpage of conda install, it said that:

  -c CHANNEL, --channel CHANNEL
                        Additional channel to search for packages. These are
                        URLs searched in the order they are given (including
                        file:// for local directories). Then, the defaults or
                        channels from .condarc are searched (unless
                        --override-channels is given). You can use 'defaults'
                        to get the default packages for conda. You can also
                        use any name and the .condarc channel_alias value will
                        be prepended. The default channel_alias is
                        http://conda.anaconda.org/.

From the above description, I can easily understand that the conda-forge is a channel name.
But do you mean tartansandal is a channel name too?

Regards

[tartansandal]

You are 100% correct that 'tartansandal' is a channel name too. Individual users can host there own channels on Anaconda Cloud, e.g., https://anaconda.org/tartansandal. That was my initial attempt at distributing the code and learning how to write conda 'recipes'. Later, this was accepted into conda-forge -- a somewhat convoluted process. Some versions have been published to both channels. New versions will now only be published to conda-forge. The old packages only exist for historical consistency.

[tartansandal]

Also note that the conda developers have been made aware of the issue with the 4.8.0 release that your issue has brought to light (see )conda/conda#9421 (comment)). Thank you very much for taking the time to report it! A new release is being built, but it may take some time for it to surface. I don't really know how long that will be.

[tartansandal]

Closing this issue since it will be resolved by the upcoming patch release of conda and there is a temporary workaround.